1

Mitch Soderberg Calibrations WG Meeting

Nov. 5, 2015

A few notes on ArgoNeuT and LArTPC Calibrations

What to use to Calibrate TPCs

2

•Muons - use rather straight tracks from passing muon tracks.•Laser tracks - use even straighter, and more uniform, laser tracks. •Photocathodes - create charge from a known position within TPC.•Temperature probes - necessary to account for temp/drift-velocity dependence.•Radiologicals - insert various sources near (inside?) TPC for low-energy

calibration. New idea.

A variety of approaches have been suggested to calibrate the performance of LArTPCs.

Using Muons

3

algorithm to form 3D points, that can also be applied to the 3D reconstruction of arbitrary trajectorytracks that are not necessarily straight lines.

Tracks having both the first and the last reconstructed points falling within 1.5 cm of a TPCboundary, which are also matched to a track reconstructed by the MINOS-ND (see section 6) areincluded in the through-going data sample. Figure 4 shows many examples of the reconstruction ofa few hundred overlaid through-going muons. The vast majority (>95%) of these tracks are ⇠90cm in length, consistent with entering through the upstream face of the TPC and exiting throughthe downstream face. The rest of the tracks in the sample enter or exit the TPC through one of theside walls, and can be significantly shorter than 90 cm in length. The final sample contains 14322negatively-charged through-going tracks, and 2607 positively-charged through-going tracks. Theratio of µ+/µ� in the sample (15.4%/84.6%) is higher than would be expected by comparing thebreakdown of nµ/nµ neutrinos in the NuMI flux for neutrino mode (6%/94%) [11]. This is notunexpected since the through-going sample under consideration has a higher average energy thanthat of muons from the overall flux, and the ratio of µ+/µ� increases with increasing energy.

Figure 4. 3D reconstructed tracks for a sample of overlaid through-going muon tracks crossing the Ar-goNeuT active volume.

In figure 5, the inclination of the reconstructed tracks with respect to the horizontal (left) andvertical (right) axes is shown. The mean value of the horizontal angle (0.61� for negatively-chargedtracks) suggests a slight misalignment of the ArgoNeuT TPC with respect to the neutrino beamdirection. The mean vertical angle is peaking at approximately -3�, which is consistent with theNuMI beam’s known inclination angle with respect to ArgoNeuT and the MINOS-ND as it headsdown through the decay tunnel towards Minnesota.

– 6 –

Collection Wire Number

Induction Wire Number

Sam

ple

Num

ber

Sam

ple

Num

ber

x

y

zRefs: 1.) Analysis of a Large Sample of Neutrino-Induced Muons with the ArgoNeuT Detector, C. Anderson et al, JINST 7 (2012) P10020

ArgoNeuT collected tens of thousands of “through-going”

muons originating from neutrino interactions in

surrounding NuMI environment.

Using Muons

4

Z vs. X for TG Tracks

9

Z (cm)0 10 20 30 40 50 60 70 80 90

X (c

m)

0

10

20

30

40

50ArgoNeuT Through-Going Tracks: Neutrino Mode

Cathode

Anode

TPCfieldmap TPCfieldmap

zzI need to work with Steve L. to implement the detailed E I need to work with Steve L. to implement the detailed E--field map in the wire signal simulation. field map in the wire signal simulation.

z [cm]

x [cm]

Garfield simulation

Hypothesis is that field non-uniformities produce features in

this plot

Cathode

Anode

Track enterTrack exit

Using Mouns

5

X vs. Y for TG Tracks

8

X (cm)0 10 20 30 40 50

Y (c

m)

-25

-20

-15

-10

-5

0

5

10

15

20

25ArgoNeuT Through-Going Tracks: Neutrino Mode

Cat

hode

Ano

deHypothesis is that field non-uniformities

produce features in this plot

Track enterTrack exit

Using Muons

6

The drift time from cathode to wire plane can alternatively be determined using through-goingtracks that cross just one of the two planes, either the cathode or the anode wire plane, and mea-suring their entry time or exit time.The shield plane, delimiting the TPC drift volume opposite to the cathode, is not instrumented forsignal read-out. Therefore, the successive Induction plane has to be used for drift time determi-nation. The corresponding drift distance includes the nominal drift length of the detector (`d=470mm from cathode to Shield plane at Ed) together with the gap between Shield and Induction planeat a higher field value (`g=4 mm at Eg1 = rT1 Ed, with rT1=1.45). A correction for the field changein the gap will be taken into account in the drift velocity determination, as well as a correction dueto thermal contraction of the TPC frame from room to LAr temperature in the drift direction.

Figure 25: [Left] Hit time distribution from the Induction plane for neutrino induced through-going muon tracks.The sample includes tracks crossing either the cathode or the Induction plane of the TPC. These tracks provide thehits at the edges of the distribution and thus determine the drift time measurement through Eq.5. The precision onthe determination of the r.h.s. edge of the distribution (cathode maximum crossing time detected, tmax

C = 310 ± 2µs)is the main contribution to the systematic error of the drift velocity measurement (Eq.6). [Right] The drift velocitymeasurement in ArgoNeuT, and associated systematic uncertainties, is compared with expectation values from [26]as a function of the electric field and for di↵erent LAr temperatures corresponding to the range of variation of thecurrent LAr temperature during operations.

The distribution of the reconstructed hit times in the Induction wire signals from a sampleof neutrino induced through-going muon tracks is shown in figure 25 [Left]. A fraction of thesetracks cross either the cathode or the anode plane of the TPC.As neutrino interactions are randomly distributed in time within the NuMI spill duration (�tspill =

9.7 µs), the absolute hit time corresponding to a single plane crossing is also distributed over thistime interval. The earliest neutrino interactions, at the beginning of the spill, will produce trackswith the minimum value of the absolute hit time (tmin

I ) when crossing the Induction plane, whilethe latest neutrinos in the spill produce tracks with maximum hit time (tmax

C ) when crossing thecathode.The drift time td from cathode to Induction plane can thus be determined from the hit time distri-bution of through-going tracks as

td = tmaxC � tmin

I � �tspill = 300.5 µs (5)38

The drift time from cathode to wire plane can alternatively be determined using through-goingtracks that cross just one of the two planes, either the cathode or the anode wire plane, and mea-suring their entry time or exit time.The shield plane, delimiting the TPC drift volume opposite to the cathode, is not instrumented forsignal read-out. Therefore, the successive Induction plane has to be used for drift time determi-nation. The corresponding drift distance includes the nominal drift length of the detector (`d=470mm from cathode to Shield plane at Ed) together with the gap between Shield and Induction planeat a higher field value (`g=4 mm at Eg1 = rT1 Ed, with rT1=1.45). A correction for the field changein the gap will be taken into account in the drift velocity determination, as well as a correction dueto thermal contraction of the TPC frame from room to LAr temperature in the drift direction.

Figure 25: [Left] Hit time distribution from the Induction plane for neutrino induced through-going muon tracks.The sample includes tracks crossing either the cathode or the Induction plane of the TPC. These tracks provide thehits at the edges of the distribution and thus determine the drift time measurement through Eq.5. The precision onthe determination of the r.h.s. edge of the distribution (cathode maximum crossing time detected, tmax

C = 310 ± 2µs)is the main contribution to the systematic error of the drift velocity measurement (Eq.6). [Right] The drift velocitymeasurement in ArgoNeuT, and associated systematic uncertainties, is compared with expectation values from [26]as a function of the electric field and for di↵erent LAr temperatures corresponding to the range of variation of thecurrent LAr temperature during operations.

The distribution of the reconstructed hit times in the Induction wire signals from a sampleof neutrino induced through-going muon tracks is shown in figure 25 [Left]. A fraction of thesetracks cross either the cathode or the anode plane of the TPC.As neutrino interactions are randomly distributed in time within the NuMI spill duration (�tspill =

9.7 µs), the absolute hit time corresponding to a single plane crossing is also distributed over thistime interval. The earliest neutrino interactions, at the beginning of the spill, will produce trackswith the minimum value of the absolute hit time (tmin

I ) when crossing the Induction plane, whilethe latest neutrinos in the spill produce tracks with maximum hit time (tmax

C ) when crossing thecathode.The drift time td from cathode to Induction plane can thus be determined from the hit time distri-bution of through-going tracks as

td = tmaxC � tmin

I � �tspill = 300.5 µs (5)38

where the values are obtained from figure 25 [Left].The electron drift velocity can finally be calculated as:

vd =`d + `g/rT1 � �`

td= 1.57 ± 0.02 mm/µs (6)

where the first correction to the drift length, applied to account for the field change in the gap, isbased on the linear dependance of the drift velocity on the electric field within the current rangeand the second one (�` ' 2 mm) for the thermal contraction at LAr T is inferred from NIST dataon integrated coe�cient of thermal expansion for the G10 material of the TPC structure.The result of the drift velocity measurement in (6) refers to LAr in a temperature range of ± 0.1K around the set point and under the operational electric field strength established with ' 1%precision. The estimated relative error �v/vd ' 2% includes uncertainties on the cathode andanode plane crossing time determination, related to the intrinsic precision of the hit finding algo-rithm, and uncertainties on the e↵ective drift length evaluation, mainly due to the error associatedwith the correction for the thermal contraction (not directly measured). The measured value is invery good agreement with expectations based on a common parametrization [26] of the electrondrift velocity in liquid argon as a function of electric field strength, in the actual range of LArtemperature during the run, as reported in figure 25 [Right].

6.2. Electron lifetimeAccurate determination of the electron lifetime in LAr, an indicator of the level of chemical

purity of the liquid, is a key element in the calorimetric energy reconstruction of ionization tracksand related particle identification. When measured, the electron lifetime allows to account for thefree electron loss during the drift time due to attachment to electro-negative impurities (from ma-terial outgassing and residual leaks).The impurities concentration in the liquid may vary in time, depending on the filters’ removal e�-ciency and saturation level during operation in the recirculation/recondensation loop. The electronlifetime variation in time thus needs to be periodically monitored. A ⌧e measurement per DAQ-run(about 21 hrs on average) during the beam period was considered appropriate both for the determi-nation of the charge correction factor to be used with the data collected during the correspondingDAQ-run and for filter saturation monitoring/replacement. This imposed the need for a fast, fullyautomated o↵-line procedure for the lifetime extraction during the physics run.For this measurement, the abundant “through-going track(s) event” sample (mainly muons fromupstream ⌫-interactions) is used. These impinge upon the front side of the detector and longitudi-nally cross the TPC volume, and have a narrow (and known) energy spectrum. Almost constantdirectionality and energy loss along the track make these tracks suitable for this measurement.The accelerator beam-timing signal is used as the neutrino-event trigger for ArgoNeuT and thethrough-going muon rate was found in the range of 60 (selected) events per hour.Samples of about 1500 tracks per DAQ-run are used to extract the electron lifetime through a fullyautomated o↵-line procedure derived from [27]. Selected tracks span most (at least 120) of thewires in both Collection and Induction plane and are evenly distributed along the drift distance,from cathode to anode plane (see [17] for a detailed analysis of the through-going muon sample).The peak amplitude (i.e. the detected charge Qdet) of every hit of the tracks in the sample is plotted

39

Refs: 1.) The ArgoNeuT Detector in the NuMI Low-Energy Beam Line at Fermilab, C. Anderson et al, JINST 7 (2012) P10019

MicroBooNE UV Laser System

7

Refs: 1.) A steerable UV laser system for the calibration of liquid argon time projection chambers, A. Ereditato et al, JINST 9 (2014) T11007

MicroBooNE has UV lasers on upstream/downstream end. Provide ‘map’ of drift e-field.

MicroBooNE UV Laser System

8

Laser hitting metal cathode (Photoelectric

effect)

Photocathode on Cathode

9

Field-Cage and Cathode

1/10/2008 02:59:37 f=0.20 /Users/mitch/Eagle_T962/fieldcage_longside.brd 1/10/2008 03:01:55 f=0.20 /Users/mitch/Eagle_T962/fieldcage_shortside.brd

1/10/2008 03:02:41 f=0.10 /Users/mitch/Eagle_T962/t962_cathode.brd

Field Cage - Long Field Cage - Short

Cathode

HV Connection

PhotoCathode Mounts

•Photocathode can provide charge injection from known position.•ArgoNeuT had two gold photocathodes mounted directly on TPC cathode.•Fibers carrying flash from Xenon flash lamp directed at TPC.•…complete failure because fibers pulled out of mounts during installation for

commissioning run (2008), and were subsequently abandoned.•Nonetheless…the concept is good, the execution needs to be better.•With a resistive (non-metallic) cathode in DUNE, not clear how this idea translates.

Fiber feedthroughs

10

Refs: 1.) The Photomultiplier Tube Calibration System of the MicroBooNE Experiment, J. Conrad et al, JINST 10 (2015) T06001

•MicroBooNE has a very elegant “flasher” system for calibrating PMTs (you should ask someone from that team to present).

•My bias….Fibers are a big headache…very fragile, and feedthroughs (using Stycast or other epoxies for a vacuum-tight seal) non-trivial.

•Over the long-term fibers will degrade (repeated UV light exposure), and feedthroughs have been known to gradually lose seal. Need to design for these eventualities.

Temperature Probes

11

E-field (kV/cm)0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

s)µ

(mm

/dr

iftv

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

T=87.3

T=89.3

T=91.3

ICARUS Data

as a function of E-fielddriftv

Weak dependence of drift-velocity on LAr temperature (weaker than E-field dependence).

We had thermocouples at top/bottom of liquid volume. These were inherently inaccurate, plus systematic offsets due to junction errors of feedthrough metal. Lesson learned: use RTDs; multiple reference points is

nice; calibrate these with LN or something to verify their accuracy.

Low-Energy Calibration?

12

Corey Adams, Yale University

De-Excitation Gammas in

11

0.4 MeV 1.1 MeV

0.34 MeV

1.3 MeV2 MeV

1 MeV0.9 MeV

0.5 MeV

Can have multiple gammas, each can have multiple interactions.

Further complication: there are other particles that can have similar signatures!

Preliminary

•Few MeV depositions likely only produce signal on a single wire, so extracting energy from pulse shape alone requires precise calibration of each channel.

•Ideas floating around now (ask Asaadi/Littlejohn for details) about source insertion for this purpose on SBND.

We think these blips are de-excitation gammas. Uncertainty on energy values listed not quantified, but definitely large.