vacuum based photon detectors

126
1 University of California, Los Angeles Department of Physics and Astronomy [email protected] Vacuum based Photon Detectors Katsushi Arisaka 10/28/2012 Katsushi Arisaka, UCLA

Upload: red

Post on 24-Feb-2016

68 views

Category:

Documents


0 download

DESCRIPTION

Vacuum based Photon Detectors. Katsushi Arisaka. University of California, Los Angeles Department of Physics and Astronomy [email protected]. Outline. Concept of Photomultiplier Basic Properties QE, Gain, Time Response Imperfect Behavior of PMT Linearity, Uniformity , Noise… - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Vacuum based Photon Detectors

1

University of California Los Angeles

Department of Physics and Astronomy

arisakaphysicsuclaedu

Vacuum basedPhoton Detectors

Katsushi Arisaka

10282012 Katsushi Arisaka UCLA

2

Outline Concept of Photomultiplier Basic Properties

QE Gain Time Response Imperfect Behavior of PMT

Linearity Uniformity Noisehellip Other Vacuum Devices

Hybrid PDAPD Applications

Energy Resolution Summary

10282012 Katsushi Arisaka UCLA

3

Concept of PMT

10282012 Katsushi Arisaka UCLA

4

PMT (Photomultiplier Tube)

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 51028201211200 of 20rdquo PMTs

Super-Kamiokande

610282012 Katsushi Arisaka UCLA

>

Operation of Head-On Type PMT

signal light-gtphotoelectron photoelectron-gtDy1 electron-gt multiplication

cascade multiplication electric signal from anode

10282012 Katsushi Arisaka UCLA 7

Katsushi Arisaka UCLA 810282012

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

Katsushi Arisaka UCLA 910282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 1010282012

FAQWhy still PMT Why not Silicon

Photodiode

Intrinsically high gainLow noise ndash photon countingFast speedLarge area

butPoor Quantum EfficiencyBulkyExpensive

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 2: Vacuum based Photon Detectors

2

Outline Concept of Photomultiplier Basic Properties

QE Gain Time Response Imperfect Behavior of PMT

Linearity Uniformity Noisehellip Other Vacuum Devices

Hybrid PDAPD Applications

Energy Resolution Summary

10282012 Katsushi Arisaka UCLA

3

Concept of PMT

10282012 Katsushi Arisaka UCLA

4

PMT (Photomultiplier Tube)

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 51028201211200 of 20rdquo PMTs

Super-Kamiokande

610282012 Katsushi Arisaka UCLA

>

Operation of Head-On Type PMT

signal light-gtphotoelectron photoelectron-gtDy1 electron-gt multiplication

cascade multiplication electric signal from anode

10282012 Katsushi Arisaka UCLA 7

Katsushi Arisaka UCLA 810282012

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

Katsushi Arisaka UCLA 910282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 1010282012

FAQWhy still PMT Why not Silicon

Photodiode

Intrinsically high gainLow noise ndash photon countingFast speedLarge area

butPoor Quantum EfficiencyBulkyExpensive

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 3: Vacuum based Photon Detectors

3

Concept of PMT

10282012 Katsushi Arisaka UCLA

4

PMT (Photomultiplier Tube)

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 51028201211200 of 20rdquo PMTs

Super-Kamiokande

610282012 Katsushi Arisaka UCLA

>

Operation of Head-On Type PMT

signal light-gtphotoelectron photoelectron-gtDy1 electron-gt multiplication

cascade multiplication electric signal from anode

10282012 Katsushi Arisaka UCLA 7

Katsushi Arisaka UCLA 810282012

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

Katsushi Arisaka UCLA 910282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 1010282012

FAQWhy still PMT Why not Silicon

Photodiode

Intrinsically high gainLow noise ndash photon countingFast speedLarge area

butPoor Quantum EfficiencyBulkyExpensive

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 4: Vacuum based Photon Detectors

4

PMT (Photomultiplier Tube)

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 51028201211200 of 20rdquo PMTs

Super-Kamiokande

610282012 Katsushi Arisaka UCLA

>

Operation of Head-On Type PMT

signal light-gtphotoelectron photoelectron-gtDy1 electron-gt multiplication

cascade multiplication electric signal from anode

10282012 Katsushi Arisaka UCLA 7

Katsushi Arisaka UCLA 810282012

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

Katsushi Arisaka UCLA 910282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 1010282012

FAQWhy still PMT Why not Silicon

Photodiode

Intrinsically high gainLow noise ndash photon countingFast speedLarge area

butPoor Quantum EfficiencyBulkyExpensive

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 5: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 51028201211200 of 20rdquo PMTs

Super-Kamiokande

610282012 Katsushi Arisaka UCLA

>

Operation of Head-On Type PMT

signal light-gtphotoelectron photoelectron-gtDy1 electron-gt multiplication

cascade multiplication electric signal from anode

10282012 Katsushi Arisaka UCLA 7

Katsushi Arisaka UCLA 810282012

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

Katsushi Arisaka UCLA 910282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 1010282012

FAQWhy still PMT Why not Silicon

Photodiode

Intrinsically high gainLow noise ndash photon countingFast speedLarge area

butPoor Quantum EfficiencyBulkyExpensive

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 6: Vacuum based Photon Detectors

610282012 Katsushi Arisaka UCLA

>

Operation of Head-On Type PMT

signal light-gtphotoelectron photoelectron-gtDy1 electron-gt multiplication

cascade multiplication electric signal from anode

10282012 Katsushi Arisaka UCLA 7

Katsushi Arisaka UCLA 810282012

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

Katsushi Arisaka UCLA 910282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 1010282012

FAQWhy still PMT Why not Silicon

Photodiode

Intrinsically high gainLow noise ndash photon countingFast speedLarge area

butPoor Quantum EfficiencyBulkyExpensive

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 7: Vacuum based Photon Detectors

Operation of Head-On Type PMT

signal light-gtphotoelectron photoelectron-gtDy1 electron-gt multiplication

cascade multiplication electric signal from anode

10282012 Katsushi Arisaka UCLA 7

Katsushi Arisaka UCLA 810282012

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

Katsushi Arisaka UCLA 910282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 1010282012

FAQWhy still PMT Why not Silicon

Photodiode

Intrinsically high gainLow noise ndash photon countingFast speedLarge area

butPoor Quantum EfficiencyBulkyExpensive

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 8: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 810282012

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

Katsushi Arisaka UCLA 910282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 1010282012

FAQWhy still PMT Why not Silicon

Photodiode

Intrinsically high gainLow noise ndash photon countingFast speedLarge area

butPoor Quantum EfficiencyBulkyExpensive

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 9: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 910282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 1010282012

FAQWhy still PMT Why not Silicon

Photodiode

Intrinsically high gainLow noise ndash photon countingFast speedLarge area

butPoor Quantum EfficiencyBulkyExpensive

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 10: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 1010282012

FAQWhy still PMT Why not Silicon

Photodiode

Intrinsically high gainLow noise ndash photon countingFast speedLarge area

butPoor Quantum EfficiencyBulkyExpensive

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 11: Vacuum based Photon Detectors

11

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 12: Vacuum based Photon Detectors

12

Basic Properties

10282012 Katsushi Arisaka UCLA

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 13: Vacuum based Photon Detectors

Outline

Fundamental Parameters of PMT Quantum Efficiency (QE) Photoelectron Collection Efficiency (CE) Gain (G) Excess Noise Factor (ENF)

How to Measure These Parameters

Energy Resolution (E)

10282012 Katsushi Arisaka UCLA 13

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 14: Vacuum based Photon Detectors

14

Quantum Efficiency (QE)

10282012 Katsushi Arisaka UCLA

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 15: Vacuum based Photon Detectors

Quantum Efficiency (QE)Definition

The single most important quantityN

NPhotonsInsident

ronsPhotoelectEmittedQE

pe

)_(

)_(

10282012 Katsushi Arisaka UCLA 15

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 16: Vacuum based Photon Detectors

QE curves of 6 types

01

1

10

100

100 200 300 400 500 600 700 800 900 1000

Wavelength (nm)

QE

()

Cs-Te

GaAsP

Extended RedMultialkali (S-25)

Multialkali (S-20)

Bialkali GaAs

Quantum Efficiency with 6 types of Photocathodes

VUV UV Visible Infra-Red

10282012 Katsushi Arisaka UCLA 16

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 17: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 1710282012

Typical QE

BialkaliSb-Rb-CsSb-K-Cs

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 18: Vacuum based Photon Detectors

18

Transmittance of windows

popular

VisibleUVVUV

More Expensive

Wavelength is Shorter

10282012 Katsushi Arisaka UCLA

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 19: Vacuum based Photon Detectors

FAQ

Why is QE limited to ~40 at best

Competing two factorsbull Absorption of photonbull Emission of photo-electrons

Isotropic emission of photo-electrons

10282012 Katsushi Arisaka UCLA 19

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 20: Vacuum based Photon Detectors

FAQHow can we measure QE

Connect all the dynodes and the anodeSupply more than +100V for 100

collection efficiencyMeasure the cathode current (IC)Compare IC with that of a reference

photon-detector with known QE

10282012 Katsushi Arisaka UCLA 20

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 21: Vacuum based Photon Detectors

UCLA QE System

Reference PMT PMT with unknown QE

Xe Lamp

Source PMT

Monochromator

Integrating Sphere

reference

unknownreferenceunknown I

IQEQE

10282012 21Katsushi Arisaka UCLA

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 22: Vacuum based Photon Detectors

22

UCLA Vacuum UV QE System

PDPMT

W LampD2 Lamp

MonochromatorUCLA

Hamamatsu

10282012 Katsushi Arisaka UCLA

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 23: Vacuum based Photon Detectors

2310282012 Katsushi Arisaka UCLA

Propagation Chain of Absolute Calibration of Photon Detectors

Cryogenic Radiometer

Trap Detector

Pyroelectric Detector

Laser(s)

NIST standard UV Si PD

Reference PMTReal Light Source

Monochromator

UV LEDXe LampLaser(s)

Particle Beam

Real experiments PMTs in our detectors

Light BeamScattered Light

NIST

us

NIST standard UV Si PD

Standard Light Beam

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 24: Vacuum based Photon Detectors

2410282012 Katsushi Arisaka UCLA

NIST High Accuracy Cryogenic Radiometer (HACR)

Photon energy is converted to heat

Heat is compared with resistive (Ohmic) heating

0021 accuracy at 1mW

This is the origin of absolute photon intensity

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 25: Vacuum based Photon Detectors

2510282012 Katsushi Arisaka UCLA

Trap Detector

Front View Bottom View

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 26: Vacuum based Photon Detectors

NIST Standards Quantum efficiencies of typical Si InGaAs and Ge photodiodes

10282012 Katsushi Arisaka UCLA 26

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 27: Vacuum based Photon Detectors

Sk (Cathode Sensitivity)and Skb (Cathode Blue Sensitivity)

Filter for Skb Lump for Sk10282012 Katsushi Arisaka UCLA 27

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 28: Vacuum based Photon Detectors

Collection Efficiency (CE)

Definition

)_()_1___(

ronsPhotoelectEmittedDynodestbycapturedPECE

10282012 Katsushi Arisaka UCLA 28

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 29: Vacuum based Photon Detectors

FAQ

How can we measure Collection Efficiency

Measure the Cathode current (IC)Add 10-5 ND filter in front of PMTMeasure the counting rate of the single

PE (S)Take the ratio of S1610-19 105IC

10282012 Katsushi Arisaka UCLA 29

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 30: Vacuum based Photon Detectors

Detective Quantum Efficiency (DQE)

Definition

bull Often confused as QE by ldquoPhysicistsrdquo

CEQEPhotonsInsident

DynodestbycapturedPEDQE

)_(

)_1___(

10282012 Katsushi Arisaka UCLA 30

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 31: Vacuum based Photon Detectors

FAQHow can we measure Detective QE

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the counting rate of the single PE (S)

Compare S with that of PMT with known DQE

10282012 Katsushi Arisaka UCLA 31

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 32: Vacuum based Photon Detectors

32

Dynode Structure

10282012 Katsushi Arisaka UCLA

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 33: Vacuum based Photon Detectors

PMT Types

lt SIZE gt 12 inch amp 1-18 inchlt Features gt Compact Relatively Cheap

lt SIZE gt 38 inch ~ 20 inchlt Features gt Variety of sizes Direct coupling

lt SIDE-ON TYPE gt lt HEAD-ON TYPE gt

10282012 Katsushi Arisaka UCLA 33

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 34: Vacuum based Photon Detectors

Dynode Structures ndash Side-on vs Head-on

CIRCULAR CAGE

CompactFast time response(mainly for Side-On PMT)

Good CE(Good uniformity)Slow time response

BOX amp GRID

lt SIDE-ON gtlt HEAD-ON gt

10282012 Katsushi Arisaka UCLA 34

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 35: Vacuum based Photon Detectors

LINEAR FOCUSED (CC+BOX)

Fast time responseGood pulse linearity

VENETIAN BLIND

Large dynode areaBetter uniformity

Dynode Structures ndash Linear Focus vs Venetian Blind

Larger DY1 is used in recent new PMTs (Box amp Line)

10282012 Katsushi Arisaka UCLA 35

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 36: Vacuum based Photon Detectors

Metal Channel PMT

CompactFast time responsePosition sensitive

PMT with Metal Channel Dynode

16mm in dia

METAL CHANNEL

Pitch1mm

TO-8 type PMT

10282012 Katsushi Arisaka UCLA 36

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 37: Vacuum based Photon Detectors

37

Fine Mesh PMT

10282012 Katsushi Arisaka UCLA

Fine Mesh

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 38: Vacuum based Photon Detectors

MCP (Micro Channel Plate)

Gain = 100 - 1000

( 5 ndash 10 μm ϕ)

10282012 Katsushi Arisaka UCLA 38

MCP

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 39: Vacuum based Photon Detectors

39

MCP PMT

10282012 Katsushi Arisaka UCLA

MCP PMTImage Intensifier

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 40: Vacuum based Photon Detectors

Principle of Image Intensifier

httpwwwe-radiographynetradtechiintensifierspdf

10282012 Katsushi Arisaka UCLA 40

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 41: Vacuum based Photon Detectors

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

10282012 Katsushi Arisaka UCLA 41

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 42: Vacuum based Photon Detectors

42

Gain of PMT

10282012 Katsushi Arisaka UCLA

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 43: Vacuum based Photon Detectors

Structure of Linear-focus PMT

Mesh Anode Last Dynode

Photo CathodeSecond Last DynodeFirst Dynode

Glass Window

Photons

QE

CE1

2

3

n

N

G = 123 n

E=NQECEG

10282012 Katsushi Arisaka UCLA 43

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 44: Vacuum based Photon Detectors

Secondary electron Emission

HV06

10282012 Katsushi Arisaka UCLA 44

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 45: Vacuum based Photon Detectors

Gain (GP)

Definition by Physicists

nPG 321

(i = Gain of the i-th dynode)

nP HVG 60

10282012 Katsushi Arisaka UCLA 45

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 46: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 4610282012

FAQ

How can we measure the Gain (GP) of our definition

Use a weak pulsed light source (so that gt90 pulse gives the pedestal)

Measure the center of the mass of Single PE charge distribution of the Anode signal (QA)

Take the ratio of QA1610-19

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 47: Vacuum based Photon Detectors

47

Single PE distribution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 48: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 4810282012

Gain (GI)

Definition by Industries

nI CEG 321

(i = Gain of the i-th dynode)

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 49: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 4910282012

FAQHow do manufactures measure the

real Gain (GI)

Measure the Cathode current (IC) Add 10-5 ND filter in front of PMT Measure the Anode current (IA) Take the ratio of IA105IC

PI GCEG

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 50: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 5010282012

Gain vs Voltage Curve

Physicists Definition

GP=δ1bullδ2bullhellip bullδn

Industries Definition

GI=CEbullδ1bullδ2bullhellip bullδn

CE=GIGP~80

GP by UCLA

GI by Photonis

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 51: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 5110282012

270 Auger-SD PMTs HV for G=2105

UCLA vs Photonis

HV varies from PMT to PMT

Photonis is Higher than UCLA (due to CE)

CE varies from PMT to PMT

UCLA

Photonis

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 52: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 5210282012

FAQ

Why is the Gain so different from PMT to PMT at the fixed HV

At given HV each may be 10 different Then Gain could be an order of magnitude

different (G = 123 n)

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 53: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 5310282012

FAQ

What is the maximum allowed HV for stable PMT operation

It can be checked by Dark Current behavior

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 54: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 5410282012

Gain and Dark Current vs HV

ThermalPhotoelectronEmission

LeakageCurrent

FieldEffect

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 55: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 5510282012

Temperature Dependence of Anode Sensitivity

-04oC

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 56: Vacuum based Photon Detectors

56

Two Types of Voltage Divider

lt-HV Operationgt

lt+HV OperationgtPulse operation only

No DC output

10282012 Katsushi Arisaka UCLA

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 57: Vacuum based Photon Detectors

57

Time Response

10282012 Katsushi Arisaka UCLA

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 58: Vacuum based Photon Detectors

58

Time Response

TTSTransit Time Spread

(Variation of Transit Time)

Transit Time

RISE TIME10 to 90

FALL TIME90 to 10

Example ofWaveform

Rise 15 nsFall 27 ns

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 59: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 5910282012

Typical TTS (Transit Time Spread)

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 60: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 6010282012

Transit Time vs HV

Higher VoltageFaster Transit Time

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 61: Vacuum based Photon Detectors

61

Time Properties (R11410)

10282012 Katsushi Arisaka UCLA

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 62: Vacuum based Photon Detectors

6210282012

Time Resolution vs Sensitive Area

HPD

SiPM

Katsushi Arisaka UCLA

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 63: Vacuum based Photon Detectors

63

Imperfect Behavior of PMT

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 64: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 6410282012

Uncertainties Specific to PMTsPMTs are not perfect There are many

issues to be concerned

Non Linearity Cathode and Anode Uniformity Effect of Magnetic Field Temperature Dependence Dark Counts After Pulse Rate Dependence Long-term Stability

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 65: Vacuum based Photon Detectors

65

Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 66: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 6610282012

PMT Non LinearityNon Linearity is the effect of the space charge

mainly between the last and the second last dynode

Mesh Anode Last Dynode

Photo Cathode Second Last Dynode

First Dynode

Glass Window

Photons

QECol

1

2

3

n

N

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 67: Vacuum based Photon Detectors

67

Pulse Linearity

What is Pulse Linearity Relation between radiation energy and PMT output

Devi

atio

n fro

m id

eal l

ine

()

PMT output peak current (mA)PMT output

Radi

atio

n En

ergy

Ligh

t Int

ensi

ty

10282012 Katsushi Arisaka UCLA

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 68: Vacuum based Photon Detectors

68

Block Diagram for Double-Pulsed Mode

Dim 1 Bright 4

Pulse Linearity

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 69: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 69

Optimization of Anode Pulse Linearity

10282012

(The last 3 stages)

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 70: Vacuum based Photon Detectors

70

Linearity at different gainsLow gain (1000V) High gain (1500V)

10282012 Katsushi Arisaka UCLA

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 71: Vacuum based Photon Detectors

71

Uniformity

10282012 Katsushi Arisaka UCLA

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 72: Vacuum based Photon Detectors

72

Anode Uniformity

spot light

SLIT shapeIncident light

Large size ofIncident light

10282012 Katsushi Arisaka UCLA

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 73: Vacuum based Photon Detectors

73

Cathode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 74: Vacuum based Photon Detectors

74

Anode Uniformity (3 inch PMT)

10282012 Katsushi Arisaka UCLA

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 75: Vacuum based Photon Detectors

75

Collection Efficiency (=AnodeCathode)

(KA0044)

10282012 Katsushi Arisaka UCLA

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 76: Vacuum based Photon Detectors

76

Effect of Magnetic Field

10282012 Katsushi Arisaka UCLA

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 77: Vacuum based Photon Detectors

7710282012

Effect of Magnetic Fields

MetalChannel

FineMesh

MCPPMT

SolidState

LinearFocus

HPDAPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 78: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 7810282012

Typical Magnetic Field Effect

Earth B-Field

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 79: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 7910282012

xy

z

Effect of Magnetic Field on Liner-focus 2rdquo PMT

Hamamatsu 2rdquo PMT (R7281-01)

Earth B-Field

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 80: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 8010282012

Edge Effect of Magnetic Shields

For effective shieldingwe need extra mu-metalin front

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 81: Vacuum based Photon Detectors

81

Dark Count

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 82: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 8210282012

Temperature Dependence of Dark Current

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 83: Vacuum based Photon Detectors

83

Dark Count Rate vs Temperature

10282012 Katsushi Arisaka UCLA

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 84: Vacuum based Photon Detectors

84

After Pulse

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 85: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 8510282012

After Pulse (R11410)

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 86: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 8610282012

After Pulse by Helium

Helium Contaminated PMT from MACRO

gt 10

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 87: Vacuum based Photon Detectors

87

Long Term Stability

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 88: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 8810282012

Typical Long-term Stability

From Hamamatsu PMT Handbook

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 89: Vacuum based Photon Detectors

89

Other Vacuum Devices

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 90: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 9010282012

Principle of Silicon Photodiode

Gain = 10 QE ~ 100 Extremely

Stable Large

Dynamic Range

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 91: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 9110282012

APD (Avalanche Photodiode)

High Gain (100-1000) High QE (~70)Then why not replace PMTsDrawbacks

lt2 ENFgt2 Effectively QE lt35 Extremely Sensitive to Temperature and

Voltage change Difficult to manufacture uniform large area

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 92: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 9210282012

In vacuum Silicon Photodiode instead of dynodes

High Gain (1000-3000) we can count 1-5 photoelectrons

Then why not replace PMTs

HPD (Hybrid Photodiode)

Photo Cathode

Silicon PDe-

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 93: Vacuum based Photon Detectors

9310282012

CMS Detector under 4 Tesla

4 TeslaEM Hadron

APD HPDKatsushi Arisaka UCLA

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 94: Vacuum based Photon Detectors

9410282012

CMS HCAL Multi pixel HPD   (DEP)

PIN Diode array

Ceramic feedthrough

Fiber-OpticWindow

Photocathode (-10 kV)

e

19 channel pixel layout

pixel size 54 mm flat-flatgap between pixels 004 mm

34 mm

Katsushi Arisaka UCLA

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 95: Vacuum based Photon Detectors

9510282012

LHCb experiment

Katsushi Arisaka UCLA

RICHRICH

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 96: Vacuum based Photon Detectors

9610282012

The pixel HPD by DEP (for LHCb)

Advantages of this hybrid pixel structurelow noise excellent resolution of single photoelectronshigh channel numberdensity

DEP The Netherlands

Katsushi Arisaka UCLA

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 97: Vacuum based Photon Detectors

10282012 Katsushi Arisaka UCLA 97

Hamamatsu Hybrid APD

Single Channel HAPD

64 Channel HAPD+ Readout

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 98: Vacuum based Photon Detectors

1 2 3 hellip Photo-electron Distribution

10282012 Katsushi Arisaka UCLA 98

12 3

4

5

6 Photo-electrons

True photon counting

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 99: Vacuum based Photon Detectors

Decay Time Measurement by HAPD

10282012 Katsushi Arisaka UCLA 99

Time Resolution = 80 psec

FWHM = 15 ns

No after pulse

Pulse Shape

Decay Time

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 100: Vacuum based Photon Detectors

100

Leica HyD Detector for Confocal Microscope

10282012 Katsushi Arisaka UCLA

HamamatsuCompact HAPDwith GaAsP

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 101: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 10110282012

8 inch HAPD by Hamamatsu

New releaseat NSS 2012

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 102: Vacuum based Photon Detectors

10282012 Katsushi Arisaka UCLA 102

Water Tank

LiquidScinti

Water Tank

Xe 20 ton (10 ton) 40Ar 70 ton (50 ton)

15 m6 m

LiquidScinti

Xe Ar

MAX G3 Dark Matter Detector

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 103: Vacuum based Photon Detectors

Photo Cathode(-6 kV)

APD (0 V)

Quartz

Quartz

Al coating

APD (0 V)

Photo Cathode(-6 kV)

QUPID (QUartz Photon Intensifying Detector)

10282012 103Katsushi Arisaka UCLA

Made by Synthetic Silica only

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 104: Vacuum based Photon Detectors

104

Production Version QUPID

10282012 Katsushi Arisaka UCLA

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 105: Vacuum based Photon Detectors

1 2 and 3 PE Distribution with 2m cable

10282012 Katsushi Arisaka UCLA 105

2 PE

3 PE

1 PE

G = 800 times 200 = 160000

TTS = 160 ps (FWHM)

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 106: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 10610282012

Intevac Electron Bombarded CMOS

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 107: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 10710282012

EBAPS by Intevac

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 108: Vacuum based Photon Detectors

108

Energy Resolution

10282012 Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 109: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 10910282012

Anode Signal (E) Definition

p

pol

IpeI

n

GDQEN

GCQEN

GNGQEN

CEQENE

321

(by Industries)

(by Physicists)

(N = No of Incident Photons)(Npe = No of Photo-electrons)

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 110: Vacuum based Photon Detectors

11010282012

In ideal case

In reality

ndash N Number of incident photonsndash QE Quantum Efficiencyndash CE Collection Efficiencyndash ENF Excess Noise Factor (from Dynodes)ndash ENC Equivalent Noise Charge (Readout Noise)ndash G Gain

Energy Resolution (E)

NN

NE

1

2ENF ENC

E N QE CE N QE CE G

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 111: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 11110282012

Excess Noise Factor (ENF) Definition

In case of PMT

How to measure Set Npe = 10-20 (for nice Gaussian) Measure E of the Gaussian distribution ENF is given by

2

2

Input

OutputENF

n

ENF

21211

1111

CENEENF pe 2

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 112: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 11210282012

Single PE Distribution

of the single PE distribution is given by

Thus ENF is related to Peak to Valley Ratio

1

111

21211_

ENF

E nPESingle

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 113: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 11310282012

Single PE Distribution To see single PE

tune light intensity so that gt90 gives pedestal

If 1 gtgt5 ENFlt14 Clear single PE can be seen

The true position is given by the ldquoCenter of Massrdquo including signal below the pedestal

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 114: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 11410282012

ENF vs PV Ratio of 270 Auger-SD PMTs

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 115: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 11510282012

FAQ

When should we use PMT and when should we use Silicon Photodiode

Depends on intensity of photons Depends on speed of signals

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 116: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 11610282012

Resolution of Hybrid Photodiode (HPD)

HPD can count 1 2 3hellip PE separately 1 gt1000 ENF=10

But it is still suffering from poor QE We can never beat the Poisson statistics

200 300 400 500 600ADC Channel

Pedestal

1 pe NIM A 442 (2000) 164-170

3 pe

2 pe

4 pe

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 117: Vacuum based Photon Detectors

11710282012

Summary Table

QE CE i ENF G ENC E

Ideal 10 10 1000 10 106 0 1N

PMT 035 09 10 13 106 200 38N

PD 07 10 - 10 1 200 14N+(280N)2

APD 07 10 2 20 100 200 29N+(29N)2

HPD 05 09 1000 10 103 200 22N+(04N)2

HAPD 05 09 1000 10 105 200 22N

SiPM 07 04 1000 13 106 1000 43N

VLPC 07 10 1000 10 105 200 14N

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 118: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 11810282012

Energy Resolution vs N

Noof Photons

100 101 102 103 104 105 106 107

Ene

rgy

Res

olut

ion

0001

001

01

1

10

Poisson Limit

Photo Diode

APDHPD

PMT

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 119: Vacuum based Photon Detectors

11910282012

10

15

20

25

30

1 10 100 1000 10000 100000 1000000 10000000Photon

Reso

lutio

n P

oiss

onResolution (over Poisson Limit)

PMT (35 QE)

HPD (50 QE)

APD PD

HAPD

VLPC

SiPM

G-APD

Katsushi Arisaka UCLA

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 120: Vacuum based Photon Detectors

120

Summary

10282012 Katsushi Arisaka UCLA

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 121: Vacuum based Photon Detectors

121

Purpose of Photon Detector

Observe all the quantities of photons as accurate as possible The number of photons E Arrival time of photons T Position of photons X Y Z

Primary purpose of vacuum detectors Very small number of photons lt 100 photons Accurate time of photons lt 10 nsec

10282012 Katsushi Arisaka UCLA

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 122: Vacuum based Photon Detectors

12210282012

Market Price

SiPM

Silicon

HPD

Katsushi Arisaka UCLA

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 123: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 12310282012

FAQrsquos Why do we have to operate each PMT at different

HV Why is PMT response non-uniform over surface What is the cause of non-linearity How stable is PMT How often should we

calibrate Every minute Every day What external facts could change the Gain of

PMT What could damage PMTs permanently

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 124: Vacuum based Photon Detectors

Katsushi Arisaka UCLA 12410282012

More FAQrsquos What is the source of dark current and dark

pulse Are they correlated Why is PMT still the best for photon counting

application Why is APD or HPD not widely used Then who uses APD or HPD Why is the signal of PMT so fast

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 125: Vacuum based Photon Detectors

125

Closing RemarksPMTs are still used in many applications for good

reasons Intrinsically high gain Extremely low noise ndash photon counting Fast speed ( lt 1 ns) Large area ( gtgt 5 inch)

However PMTs are not perfect There are many issues to be concerned Cathode and Anode Uniformity Non Linearity Effect of Magnetic Field Long-term Stability

10282012 Katsushi Arisaka UCLA

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References
Page 126: Vacuum based Photon Detectors

126

References

Hamamatsu PMT Handbook httpsaleshamamatsucomassets

applicationsETDpmt_handbook_completepdf

Special thanks to Yuji Yoshizawa at Hamamatsu Photonics

10282012 Katsushi Arisaka UCLA

  • Vacuum based Photon Detectors
  • Outline
  • Concept of PMT
  • PMT (Photomultiplier Tube)
  • Super-Kamiokande
  • Slide 6
  • Operation of Head-On Type PMT
  • Structure of Linear-focus PMT
  • Principle of Silicon Photodiode
  • FAQ
  • Purpose of Photon Detector
  • Basic Properties
  • Outline (2)
  • Quantum Efficiency (QE)
  • Quantum Efficiency (QE) (2)
  • QE curves of 6 types
  • Typical QE
  • Transmittance of windows
  • FAQ (2)
  • FAQ (3)
  • Slide 21
  • Slide 22
  • Propagation Chain of Absolute Calibration of Photon Detectors
  • NIST High Accuracy Cryogenic Radiometer (HACR)
  • Trap Detector
  • NIST Standards Quantum efficiencies of typical Si InGaAs and
  • Sk (Cathode Sensitivity) and Skb (Cathode Blue Sensitivity)
  • Collection Efficiency (CE)
  • FAQ (4)
  • Detective Quantum Efficiency (DQE)
  • FAQ (5)
  • Dynode Structure
  • PMT Types
  • Dynode Structures ndash Side-on vs Head-on
  • Dynode Structures ndash Linear Focus vs Venetian Blind
  • Metal Channel PMT
  • Fine Mesh PMT
  • MCP (Micro Channel Plate)
  • MCP PMT
  • Principle of Image Intensifier
  • Effect of Magnetic Fields
  • Gain of PMT
  • Structure of Linear-focus PMT (2)
  • Secondary electron Emission
  • Gain (GP)
  • FAQ (6)
  • Single PE distribution
  • Gain (GI)
  • FAQ (7)
  • Gain vs Voltage Curve
  • 270 Auger-SD PMTs HV for G=2105 UCLA vs Photonis
  • FAQ (8)
  • FAQ (9)
  • Gain and Dark Current vs HV
  • Temperature Dependence of Anode Sensitivity
  • Two Types of Voltage Divider
  • Time Response
  • Time Response (2)
  • Typical TTS (Transit Time Spread)
  • Transit Time vs HV
  • Time Properties (R11410)
  • Time Resolution vs Sensitive Area
  • Imperfect Behavior of PMT
  • Uncertainties Specific to PMTs
  • Linearity
  • PMT Non Linearity
  • Pulse Linearity
  • Slide 68
  • Optimization of Anode Pulse Linearity
  • Slide 70
  • Uniformity
  • Anode Uniformity
  • Cathode Uniformity (3 inch PMT)
  • Anode Uniformity (3 inch PMT)
  • Collection Efficiency (=AnodeCathode)
  • Effect of Magnetic Field
  • Effect of Magnetic Fields (2)
  • Typical Magnetic Field Effect
  • Effect of Magnetic Field on Liner-focus 2rdquo PMT
  • Edge Effect of Magnetic Shields
  • Dark Count
  • Temperature Dependence of Dark Current
  • Dark Count Rate vs Temperature
  • After Pulse
  • After Pulse (R11410)
  • After Pulse by Helium
  • Long Term Stability
  • Typical Long-term Stability
  • Other Vacuum Devices
  • Principle of Silicon Photodiode (2)
  • APD (Avalanche Photodiode)
  • HPD (Hybrid Photodiode)
  • CMS Detector under 4 Tesla
  • Slide 94
  • LHCb experiment
  • Slide 96
  • Hamamatsu Hybrid APD
  • 1 2 3 hellip Photo-electron Distribution
  • Decay Time Measurement by HAPD
  • Leica HyD Detector for Confocal Microscope
  • 8 inch HAPD by Hamamatsu
  • Slide 102
  • QUPID (QUartz Photon Intensifying Detector)
  • Production Version QUPID
  • 1 2 and 3 PE Distribution with 2m cable
  • Intevac Electron Bombarded CMOS
  • EBAPS by Intevac
  • Energy Resolution
  • Anode Signal (E)
  • Energy Resolution (E)
  • Excess Noise Factor (ENF)
  • Single PE Distribution
  • Single PE Distribution (2)
  • ENF vs PV Ratio of 270 Auger-SD PMTs
  • FAQ (10)
  • Resolution of Hybrid Photodiode (HPD)
  • Summary Table
  • Energy Resolution vs N
  • Resolution (over Poisson Limit)
  • Summary
  • Purpose of Photon Detector (2)
  • Market Price
  • FAQrsquos
  • More FAQrsquos
  • Closing Remarks
  • References