t24 results and comparison to the preceding studies on clic prototype structures

T24 results and comparison to the preceding studies

on CLIC prototype structures

May 17, 2011T. Higo (KEK)

Topics 1. Processing as a whole2. Typical BD appearance pattern3. BD vs Eacc4. BDR evolution5. Comparison of BDR among structures6. BDR vs pulse width or pulse heating7. Dark current8. Remembrance of preceding pulses9. Thinking from double pulse operation10. Following pulses without stopping11. BD location

2011/5/17 2X-band collab meeting at SLAC (T. Higo)

Prototype structures at Nextef

T18 TD18

TD18_quad

T24 TD24

TD24R05


Processing/operation as a whole

Number of breakdowns

T18 = 2500 BD / 4000 hrsTD18 = 8000 BD / 4000 hrsT24 = 1200 BD / 2000 hrs


T18_VG2.4_Disk #2

0

20

40

60

80

100

120

0

500

1000

1500

2000

2500

3000

0 500 1000 1500 2000 2500 3000 3500 4000

MasterTable_Eacc_Trend till_090610

<Eacc> 51nsec Ushi<Eacc> 113ns Ushi<Eacc> 173n ushi<Eacc> 213n Ushi<Eacc> 253ns UshimotoEacc keep 253ns UshiEacc 51ns UshiEacc 113ns UshiEacc 173ns UshiEacc 213ns UshiEacc 253ns UshiEacc Usahimoto 2/21-3/23Eacc 252ns Usahi 4/1-4/7Eacc 252ns 4/7-14Eacc 312ns 4/14-16Eacc 412ns 4/23-27Eacc 331ns 4/28-5/1Eacc 252ns 5/1-11Eacc 412ns 5/11-6/10

Total BD<Eacc> MV/m Total# BD

RF-ON integrated (hr)

090610T18#2


TD18


Processed for 1744 hours.

T24#3


Typical breakdown appearance pattern

Some breakdown triggers a series of following breakdowns.After a while, it becomes stable, randomly appeared.It needs usually more time to determine the BDR in a well-stable period. Sill we need better way of BDR evaluation.


Final run at 120MV/m

1640 1645 1650 1655 1660 1665 1670 1675 16801115

1120

1125

1130

1135

1140

1145

Run 35 #ACC-BD331nsec, 120MV/m

BDR=1.1*10^-5

1560 1570 1580 1590 1600 1610 1620 1630 1640 16501000

1020

1040

1060

1080

1100

1120

1140

Run 33&34 #ACC-BD412ns, 120MV/mBDR=2.9*10^-5

Regeme not-statistically random

Well statistically developed

Time of the final point is not real.

1670 1680 1690 1700 1710 1720 1730 1740 17501140

1140.51141

1141.51142

1142.51143

1143.51144

1144.5

Run 36 #ACC-BD252nsec, 120MV/m

4BD / 67 hrsBDR=1.6*10^-6

T24#3

Not statistically stable at all for especially a low BDR period.Two categories!? Severe BD & recovery Stable regime

Ev25Ev26

Ev30Ev31 Ev32


Typical ACC BD and following recovery process

Ev26 Ev27-29 Ev30

ACC BD1st pulse BD

Recovery processNext

ACC BD

Run35T24

A few hundred pulses

Each by 1 pulse

Later in a pulse From the beginning of a pulse

Again at later in a pulse


T24#3 Run16 goal 110MV/m at 252nsec

615 620 625 630 635 640 645 650 655 660620630640650660670680690700710720

ACC-BD

RF ON (hour)0.10 1.00 10.00

1E-05

1E-04

1E-03

BDR vs time

3/1

1

timeBDR


3/1

1

timeBDR

TD18_Disk_#2 BDR~ 1.3 x 10-5 /pulse/m] during Run 51&52 (60MW, 252ns) as of the total RF-ON period of 2255 hours

TD18_#2 at 252 ns

Still decreasing in a logarithmic time scale BDR ~ t^-0.380.10 1.00 10.00 100.00 1000.00

10.0

100.0

BDR

10^-

6 /p

ulse

/m

0.00 100.00 200.00 300.00 400.00 0

20

40

60

80

100

120

# BD


BDR vs Eacc

• Exponential slope at 100 MV/m for 10MV/m

– T18 X10 – TD18 X20 – T24 X10

• Power index “n” as Eaccn

– T18 n= 26– TD18 n=25-29– T18 n=22


Breakdown rate

10-7

10-6

10-5

10-4

80 85 90 95 100 105 110 115 120

T18_VG2.4_Disk #2Breakdown rate for 252ns and 412ns

BDR(ACC) 412ns

BDR(ACC) 252ns

BDR(ACC) [1/pulse/m]

Eacc [MV/m]

0530-0610 (3700hr)

0525-0530 (3500hr)

0520-0525 (3500hr)

0515-0518(3400hr)

0511-0515 (3300hr)

0501-0507(3100hr)

0423-0427(2900hr)

0411-0414 (2800hr)

0403-0407 (2700hr)

0402-0403 (2700hr)

0401-0402 (2700hr)

0313-0323 (2500hr)

0227-0305(2300hr)

T18#2


Relevant data points of BDR vs Eacc

Steep rise as Eacc, 10 times per 10 MV/m, less steep than T18

TD18


T24#3 BDR evolution at 252ns

Lines are assuming the same exponential slope as that at 400hr

T24


BDR evolution nominal at 100MV/m 252ns

• Exponential decrease– T18 TD18 T24– 400hrs 186hrs

• Power index “n” as (elapsed time)-x

– T18 TD18 T24– - 4.3


TD18_#2 Evolution of breakdown rateTD18


T24#3BDR evolution at 252nsnormalized 100MV/m

Assuming the same exponential slope as that at 400hr

T24


Comparison of BDR among structures

Present status

Comparison of BDR vs Eacc is just the base for feasibility evaluation.

It is not easy to evaluate < 10^-7.And it is time dependent.

What is the best parameter to characterize?No more than demandnig more understanding of

BD (and BD trigger) mechanism.


Comparison of BDR in T18, TD18 and T24

TD18

T18 T24

400hrs

1677hrs

T24

T18


BDR vs pulse width or pulse heating

It is evident that BDT is a steep function of pulse heating.

But not yet well understood when comparing various structures, and even in various stages of a

structure.


T24#3 BDR vs width in various stagesT24#3


T24#3 BDR vs Pulse heating

BDR scales as exponential to pulse heating, though BDR changes as processing and the functional form is not easy to conclude.

)6.1

(18

TExpBDRTD

Final at 120MV/m

)1.4

(18

TExpBDRTD

at 105MV/m

T24#3


BDR vs DT

BDR varies exponential to DT.

The amount seems very different between T18(SLAC) and T24(KEK).

Undamped

Damped

DT

BDR

TD18 based on Faya Wang

)7.9

(18

TExpBDRTD

)7.2

(18

TExpBDRT

T24#3 at 120MV/m

)6.1

(18

TExpBDRTD


Dark current


Amount and reached level

0.01

0.1

1

10

100

50 60 70 80 90100 200

T18_#2 Dark Current evolution081128-081224-090224-090414-090515

FC-Mid [microA] (081128)

FC-Mid [microA] (081224)

FC-Mid microA (253ns, 090225)

FC-Mid microA (253ns, 090414)

FC-Mid microA 090515

FC-Mid microA

Eacc [MV/m]

T18_Disk TD18_Disk

Eacc for peak dark current of 10 mA90MV/m 70MV/m 85MV/m

T24_Disk


Field enhancement factor as of 090515

10-6

10-5

0.0001

0.001

0.005 0.015

Dark Current of T18_VG2.4_Disk #2 on 090515

I_UP/E 2.5 090515

I_Mid/E 2.5 090515

y = 0.074674 * e (-672.19x) R= 0.99772

y = 0.16415 * e (-626.47x) R= 0.99936

I/E2.5

1/Eacc [1/MV/m]

=36

=38

Following the formula and parameters of the following page

T18#2

No big change in b duringRF ON = 2300 – 2900 hrs b ~ 36—38


Evolution of dark current till early April in TD18_Disk

Dark current reduced by three order of magnitude. It followed roughly modified F-N formula. The b reduced from 70 to 40 in its initial stage of processing keeping b * Es

max ~ 5~7 GV/m but became larger to 55 at the later stage.

IPAC10TD18

3100

hr


Fitting including low current data points

b=33 (252ns)

b=26 (412ns)

T24#3

Beta seem small b = 26—33 at RF ON = 1000 hrs


Comparison of beta values

T18 TD18 T24

Meas. at RF ON(hrs) 2300 – 2900 100 – 600

3200 1000

b 36 – 38 70 40 55 26 – 33

Seems good to compare these values.Higher beta in TD18? How it changes as time?

Better to monitor systematically.2011/5/17 31X-band collab meeting at SLAC (T. Higo)

Dark current spectra

T18#2

Different widths Different acc field levels


TD18 spectrum is missing.Our fault!

TD18

Simply enough!It should be measured!


T24#3 Dark current spectra Meas. 25 Feb., 2011 @ 1400 hours after processing start

0 2 4 6 8 10 12 14 16 18

-0.016

-0.014

-0.012

-0.01

-0.008

-0.006

-0.004

-0.002

0

0.002

Eacc = 116MV/m

Eacc = 112MV/m

Eacc = 108MV/m

kE(MeV)

Ave

rage D

ark

curr

ent(

a.u

.)

T24#3


Dark current• Dark current gives a measure on HG performance– In beta– In its amount

• Dark current contains many hints– field emission site– beta as a whole and possibly local beta– Even though it is the result of the whole area with a

limited acceptance• We want to improve the monitoring– As the processing proceeds, especially at the beginning


Remembrance of the preceding pulse?

Is the breakdown trigger influenced by anything from the preceding pulse or from the general

situation of the period?


Run 98 switching among three power levels for 46.2 hours

Switching among three power levels in each 50 pulses.

90 – 95 – 100 MV/m

TD18


Comparison to the usual BDR data

3.3*10-6 / 2.6*10-5 / 1.4*10-4 [BD/pulse/m]

TD18

Data points from jumping among three levels

81

43Run 98

BDR is determined by the power level of the very pulse.


The remembrance of the preceding pulses, 50 – 100 pulses ahead, is not kept.

We need to refrain the same experiment with pulse-to-pulse switching to see

whether the remembrance is kept from the previous pulse.


Thinking from double pulse study

What about the remembrance very close to the last pulse separated by 10 – 200 nsec period?


BDR: measured and instantaneous

)()()(),,( paccpacc ThtgEfTEtBDR

g(t) = instantaneous BDR within a pulse

h(DTp) = BD potential factor determined by the integrated effect of such as pulse heating of many pulses before the BD pulse

f(Eacc) = BDR depending on Eacc

),,(),(0 pacc

T

pacc TEtBDRdtTEBDRp

We thought that the double pulse study would give us the clue.


Features investigated by double pulse operation

FG set value Period=210ns Width=200ns

200 + 10 + 200

Double (200+10+200) = Single (410ns eq.) in TD18 Run 71 (30 June – 1 July ---)

LE5ns

TE5ns

Period

Width200ns

FG set value Period=410ns Width=200ns

200 + 210 + 200

Double (200+210+200) = Single (2000 or 400ns eq.??) in TD18 Run 72, 73, 74, 89, …


TD18 Run 89Well separated double pulse at 90 MV/m

TD18


Run 25&26 T24#3 BDRdouble pulse at 110MV/m

1290 1300 1310 1320 1330 1340 1350 1360 13700

5

10

15

20

25

Former pulse BDLatter pulse BD

Guide line slope : BDR_former = 3.2*10^-6 BDR_latter = 4.0*10^-6

T24#3

1240 1250 1260 1270 1280 1290 13000

2

4

6

8

10

12

14

Former pulse BDLatter pulse BD

Guide line slope : BDR_former = 2.4*10^-6 BDR_latter = 3.7*10^-6

Run26=200+200+200Run25=200+10+200


Run28: 200+200+200

ACC-BDFirst-pulse BD

ACC-BD

First-pulse BD

T24#3


Result of double pulse study• The BDR becomes saturating after initial a few

tens of hours.• BDR of latter pulse is a little higher than that of

former pulse, but at most within a factor 2.• Former pulse BD always triggers latter pulse BD.• The breakdown timing within a pulse distributes

uniformly in the pulse.• First-pulse BD after a BD mostly starts from the

beginning of the former pulse.

T24#3


Trigger for the breakdown• Pulse heating at 100MV/m – 200ns (DT=10degC) 400ns(DT=13-14degC)

• Then from width dependence– BDR exponential rise as DTp (SLAC TD18)

– DTp=3degC h(DTp) x10

• But double pulse result– BDR of latter pulse = at most x2– It seems g(t) rises moderate even if rises

• This result supports – BDR is mostly determined by non-instantaneous

mechanism.– i.e. no memory from the preceding pulse sepaated by 10 –

200 nsec.

T24#3


Following pulses without stopping

How the structure changes after “BIG” breakdown?

The probability of breakdown is extremely high, but the breakdown does not always happen.The probability of the breakdown rapidly decreases as number of pulses.

Structure surely remembers the BD.


Run 100 at 100 MV/m level Behavior of eight pulses after BD

More than 50% of breakdowns are not accompanied by any consecutive breakdowns.Some are followed by breakdowns. Correlation between number of consecutive breakdowns and the hardness of the initial breakdown should be evaluated.

TD18


BD location


T24#3 BD locationRun 9 RF-ON=400-430 hrs

100MV/m at 252ns just after reaching

Run 16 RF-ON=620-660 hrsDuring ramping to 110MV/m at 252ns

T24#3


BD locationAnalysis program for the BD localization by RF timing has almost been developed. KEK needs careful and through analysis of BD location, back to T18 and TD18. This I was reminded by the SEM views of T18 and TD18 shown by CERN, the distribution of the arc spots and the high magnetic field spots along the structures.

We have in mind but not yet done well.


Conclusion

• Systematic understanding applicable to all of the structures is not established at all.

• We should acquire more data systematically.• We should make complementally studies with

simple setups depicting each a single change than reference to look at the relevant mechanism that appeared in the prototype structures.

• Theoretical studies are welcome in understanding and finding the strategy.


Additional materials


Deduction of the field enhancement factor

)/()/(

1053.6

5.2

5.19

mMVEmVE accs eeE

I

32180

)/(

6530 5.1

accs EE

f(Cu)=4.52eV

Assuming Es/Eacc=1.95 for T24

Fitting of modified F-N curve


Pritzkau thesis


Study with double-pulse operation

• Temperature rise was calculated based on Pritzkau thesis formula.

• For 200nsX2 double pulse, very little difference exists in the latter pulses with different off-periods from 0 to 200ns.

• Big temperature difference exists only between former pulse and latter pulse– T24: Former=10.1C vs Latter=13.3 ~14.3degC– TD18: Former=45 vs Latter=60~63C.

• Let us compare the BDR between former and latter pulse!!


Double-pulse temperature rise at unloaded gradient of 100MV/m

magnetic field= 0.23 MA/mAt end of full pulse 0.4 microsec temp= 14.2519 degC

At end of first pulse 0.2 microsec temp= 10.0776 degCAt beginniing of second pulse 0.3 microsec temp= 5.21657 degCAt end of second pulse 0.5 microsec temp= 13.6693 degCAt end of first pulse 0.2 microsec temp= 10.0776 degCAt beginniing of second pulse 0.4 microsec temp= 4.1743 degCAt end of second pulse 0.6 microsec temp= 13.2807 degC

magnetic field= 0.485 MA/mAt end of full pulse 0.4 microsec temp= 63.3727 degC

At end of first pulse 0.2 microsec temp= 44.8112 degCAt beginniing of second pulse 0.3 microsec temp= 23.196 degCAt end of second pulse 0.5 microsec temp= 60.7817 degCAt end of first pulse 0.2 microsec temp= 44.8112 degCAt beginniing of second pulse 0.4 microsec temp= 18.5614 degCAt end of second pulse 0.6 microsec temp= 59.0539 degC

Hs=0.23MA/m at the last cell of T24#2

Hs=0.485MA/m at the last cell of TD18#3

No significant difference in pulse heating temperatures among latter pulses, only 6%, but significant difference between former and latter, by 40%.


t24 results and comparison to the preceding studies on clic prototype structures

Documents

kek xband area t24 results

higocomparison of bdr

elapsed timex t18 td18

low bdr period

logarithmic time scale

hours td18

better way of bdr evaluation

nsecbdr vs time