Review of rf structure test results
Steffen Döbert, ACE, 16.01.2008
• High gradient results
• 30 GHz
• 11.4 GHz
For more detailed high gradient results please see:
The first ACE, June 2007
http://indico.cern.ch/conferenceDisplay.py?confId=15452
The x-band structure design and testing workshop: June 2007
http://indico.cern.ch/conferenceDisplay.py?confId=15112
The High-Gradient workshop, October 2006
http://hg2006.web.cern.ch/HG2006
Copper
Tungsten
Damage in high field areas
Clamped-Iris Structure Tests in CTF II
0 0.5 1 1.5 2 2.5 3
x 106
0
50
100
150
200
No. of shots
Pe
ak
Ac
ce
lera
tin
g f
ield
(M
V/m
)
3.5 mm tungsten iris3.5 mm tungsten iris after ventilation3.5 mm copper structure3.5 mm molybdenum structureCLIC goal loadedCLIC goal unloaded
Accelerating Structure Tests in CTF II
Short, 16 ns rf pulses
New Record for classical accelerating structures !
Reached nominal 30 GHz CLIC values :
150 MV/m 70 ns
Overview of 30 GHz results
Molybdenum shows higher gradient but different slope
HDS performs worse than round brazed structure
Hybrid Damped Structure (HDS)
CLIC damped and detuned accelerating structure:30 GHz, 150 MV/m, 70 ns, < 10-6 trip probability
Accelerating Structures made out of milled quadrants
HDS60
New Materials for High-Gradient
Copper has still the best performance at low break down rate
Damage vs aperture or group velocity
HDS60 Large HDS60 Small
Evidence for correlation between damage and power flow (a,vg,P):
Criteria for optimizing rf designs (P/C):
Px(1/3)//2a < threshold
Parameters for 30 GHz test strcutures
HDS4vg2.6_thick_150degC40vg8_pi/2
Recent 30 GHz results
NDS4_vg2.5_thick result
Summary of 30 GHz results in 2007
Probing phase advance and P/C theory
Structure 2a(mm)
P (MW)
E (MV/m
)
PT1/3/C (wue)
C30vg4.7 3.5 20.2 92 7.5
HDS60vg8.0 3.8 16.1 61 5.6
HDS60vg5.1 3.2 13.3 75 5.5
C40vg7.4_pi/2 4.0 19.2 65 6.2
HDS4vg2.6_thick 3.5 7.5 67 2.8
NDS4vg2.5_thick 3.5 8.6 75 3.2
All measured data at 70 ns pulse length and 10-3 breakdown rate
Summary of 30 GHz results
First CLIC x-band structure
One tested at KEK and one tested at SLAC
150 MV/m peak, 125 MV/m avg
150 ns pulse length
No breakdown monitoring
Typical NLC/GLC prototype structures
Length: 60 cm
Phase advance:120 deg
Group velocity: 4 %
a/: 0.17
Es/Eacc: 2.2
Pin (65 MV/m): 59 MW
Coupler: mode luncher
Preparation: H-brazing,diamond turning
Performance of NLC/GLC structures
58 60 62 64 66 68 70 7210
-2
10-1
100
Bre
akd
ow
n r
ate
per
ho
ur
Average gradient
FXC3FXB-006FXB-007H60vg4R17H60vg4S17-1H60vg4S17-3FXD1FXC5Average trip rate goal
Hybrid damped structures (HDX) at x-band
Frequency scaling
Scaled structures show very similar performance
HDS-type structures show consistently limited performance
A reference structure for CLIC from NLC
Length: 53 cm
Phase advance: 120 deg
Group velocity: 3 %
a/: 0.13
Es/Eacc: 2.2
Pin (65 MV/m):41 MW
Coupler: mode luncher
Preparation: H-brazing,diamond
turning
Tests of old NLC structures at short pulses
T53vg3MC can be used as a first reference for the new CLIC parameters
CLIC goal
Pulse Length Dependence
Structure P (MW)
E (MV/m)
PT1/3/C (wue)
T53vg3MC (50ns) 118 110 18
T53vg3MC (100ns)
107 105 20
H75vg3 (150 ns) 155 97 27
HDX11vg5 (70 ns)
59 60 9
All data around ~10-6 breakdown rate
Summary of 11 GHz results
First result on power ramping during filling:
100 ns ramp (50%-100%) + 100 ns flat top: 97 MV/m at 10-6 BDR
Tapered Damped Structure Test in ASSET
Test results
Successful experimental verification of strong cell damping and benchmarking of codes
3D model of single cell SW structure
David Martin
Assembly of a three cell SW structure made by KEK
SLAC/KEK results on short SW accelerating structures
Yasuo Higashi, KEK
Breakdown rate vs. accelerating gradient, all breakdowns, flat pulse, a/l~0.21
Single Cell SW1
0.1
1
10
100
1000
90 100 110 120 130 140
50 ns70 ns90 ns
Bre
akdo
wn
rate
[#
/ hou
r]
Gradient [MV/m]
0.1
1
10
100
1000
90 100 110 120 130 140
75 ns150 ns300 ns
Gradient [MV/m]
Single Cell SW2
Time of flat pulse after filling time
Time of flat pulse after filling time
V. Dolgashev, S. Tantawi
Conclusions on recent structure tests
(some of them preliminary)
Current CLIC design within experimentally demonstrated region
27 wue have been measured (Design used 18)
120 MW input Power for 100 ns into first cell of T53 (the structures showing a promising gradient are not damped)
Hybrid Damped Structures show performance deficit (short phase advance, slots, quadrants and milling)
Copper is still the best material to make accelerating structures (Molybdenum still has some potential, shallow slope seen in previous experiments could be due to iris clamping, slow processing as usual)
Exactly scaled structures seem to perform independent of frequency (therefore 30 GHz test are still meaningful)
Some doubts on P/C theory used to optimize this years structures
Quadrant technology appears not mature
Short phase advance seems not beneficial
The end, reserve slides following
Scattered Dark Spots
Areas of DiscolorationPatchy breakdown areas along sides of irises
High Current Region
Input Coupler
Iris
Post mortem inspection of HDX11cu
