Status of x-band structure tests at SLAC

Structure working group 16.5.2007

HDX11 made out of Copper and Molybdenum by CERNH75vg3S18 and T53vg3MC from the NLC program

Status of x-band structure tests at SLAC

We are the hot topic !

ILC-News Letter

Scientific Motivation for the CLIC X-band proposal

Test HDS geometry and technology at high power{low phase advance, slotted iris, 4 quadrant design}

Test design optimization logic{constrains: surface field and Power*sqrt (pulse

length)}

Benchmark with well known NLC copper data

Learn about material dependence (Cu vs Mo)

Learn about frequency dependence{similar tests at 30 GHz in CFT3 in 2006}

Get more statistics

We are not aiming to demonstrate the CLIC structure or the CLIC gradient at X-band with these experiments !

Conditioning history at 40 ns so far

20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

70

80

90HDX11mo

Time (h)

Pe

ak

Gra

die

nt

(MV

/m)

HDX11 Mo50ns 75MV/m

0 10 20 30 40 500

20

40

60

80

100

120

140T53vg3

Time (h)

Ave

rag

e G

rad

ien

t (M

V/m

)

Conditioning history at 50 ns and 100 ns so far

Start 100 ns Run 100 MV/m; 50 ns

~2*10-5

T5350ns 100MV/m

No breakdown in 25 hours

Quick and Dirty Beta Measurement with Veetest at 100 ns down stream

4 4.2 4.4 4.6 4.8 5 5.2

x 10-9

10-21

10-20

10-19

1/E

I/E2

.5

Beta T53vg3(100 ns): = 51

190 200 210 220 230 2405

10

15

20

25

30

Surface field (MV/m)

Da

rk c

urr

en

t (a

.u)

100 ns

Calibration needed to evaluate

dark current

Vacuum behavior T53

Vacuum behavior HDX11 Mo

HDX11 conditioning statisticsOr what it takes to break a structure

Copper:

~ 20000 Break downs

~ 50 hours initial conditioning

~ 600 hours

Damaged

Molybdenum (preliminary):

~ 11500 Break downs

~ 500 h so far

“most likely damaged too”

Duration: 6 weeks with 5 weeks experiment

Scattered Dark Spots

Areas of DiscolorationPatchy breakdown areas along sides of irises

High Current Region

Input Coupler

Iris

Areas of Interest Besides the Regular Irises

HDX11 Cu post mortem inspection

Input Output

Comparing initial conditioning

Conditioning:

HDX11_Cu: 15 h, 40 ns, 80 MV/m

HDX11_Mo: 50 h, 40 ns, 75-80 MV/m

T53vg3MC: 10h, 50 ns, 120 MV/m ( was previously conditioned to 95 MV/m

2h, 50 ns, 80 MV/m

C30vg4_W: 40h, 30 ns, 80 MV/m

90h, 30 ns, 110 MV/m

C30vg4_Mo: >200h, 30 ns, 80 MV/m

Refractory metals condition slower than copper

Possible beneficial heat treatment is lost after weeks on air

Breakdown Rates

Frequency scaling experiment HDS vs HDX

scaled structures show very similar performance

More Breakdown Rates

All HDS-type structures tested so far

HDS performs consistently worse than round structuresNo significant improvement for other materials over Copper

Even More Breakdown Rates

All HDS-type structures tested so far

HDS performs consistently worse than round structuresNo significant improvement for other materials over Copper

Breakdown Rates

Frequency scaling experiment of 3.5 mm round

Again frequency scaling not inconsistentSlope most likely due to iris clamping

Breakdown Rates

HDX; 60 MV/m; 70 ns = 9 wue

T53; 110 MV/m; 50 ns = 18 wue

T53; 105 MV/m; 100 ns = 20 wue

H75; 97 MV/m; 150 ns = 27 wue

Looks pretty good to me

The Wuenschlist

Power for: 100, 105, 110 MV/m

97, 107, 118 MW

Conclusions on recent x-band tests

Typical slow processing for Moly

Slope similar to Copper

New Hypothesis: slope is not determined by material or heating but by iris clamping ?

Heating does not help for processing or is lost after a few days

Molybdenum does not show a big advantage over Copper (Therefore focus even more on copper)

Frequency scaled structures have similar performance (BD vs grad)

HDX11 Cu revealed machining and alignment issues

Structures built out of milled bars with slots did perform worse than turned and brazed structures

Structure manufacturing technology seems to play an important role

The T53 could be used as a reference structure for CLIC

What does it mean for CLIC ?

~27 Wuensches have been demonstrated (Design: 18-22 needed)

T53 based structures: 9% efficiency full length at 100 ns, 40 bunches T23vg3 should work (~14% efficiency structure demonstrated) {T23, 100 bunches at 4e9, 95 MW, T33, 80 bunches, 105 MW}

See if CLICvg1 shows a better performance

Damping has to be integrated and demonstrated

‘Old school rules’ , get engaged again with round, brazed structures made out of copper

Breakdown Rates

The CLIC proposal for High Power testing in NLCTA

Structure parameters

Scaled version of HDS11 small