Trond Ramsvik

1

DC Spark Test System for CLIC

• Trond Ramsvik
• TS / MME

2
Outline

• Experimental Setup
• Breakdown characteristics of pure metals
• Esat, Energy, Power
• Mechanical surface treatments
• Molybdenum EDM ? Rolled
• Copper / GlidCop / CuZr EDM ? Milled
• Heat treatments
• Results from annealing with e-beam and oven
• Field Stability
• Mass Spectroscopy Studies
• Knowledge obtained from DC spark tests
• Further plans

3
Experimental Setup
Sphere / Plane geometry
4
Typical Conditioning Curves - Scaled
5
Saturated Breakdown Fields and Energies
1Valid for non-heated Molybdenum
6
Saturated Breakdown Fields and Energies
7
Power Flow (Scaled)
DC 1 mm in diameter gt 0.79 mm2
RF HDS Ti Pmax 50 MW _at_ 40 ns
RF circumference x width 16 mm2
P PDC x(16/0.79)
RF circular Mo Pmax 65 MW _at_ 70 ns
RF circular Cu Pmax 40 MW _at_ 40 ns
Lower power flow available in the discharge
compared to RF
8
Chromium
Esat (491 ? 11) MV/m
dgap 16.7 µm ? 14.4 µm 14 decrease

• Intensive breakdown conditioning of chromium
  shows
• equal and higher breakdown fields than
  Molybdenum
• less erosion than Titanium.

9
Breakdown Conditions
In addition to the type of electrode materials,
the breakdown characteristics in vacuum for a
given field depend on several other important
parameters

• Electrode Geometry and Gap Distance

• Electrode Surface Finishing Treatment
• standard metallurgical polishing techniques
• mechanical
• chemical
• electrochemical
• heat treatment
• Conditioning Processes
• removal of contamination and surface smoothing
• field emission
• repeated breakdown events
• Residual gas pressure

10
Molybdenum EDM ?Cold Rolled
EDM
Rolled / Chem. cleaned
Mo
Mo
11
CuZr EDM ?Milling
12
Glidcop EDM ?Milling
13
Comparison Cu CuZr - GlidCop
120
240-280
393
180
190
340
Solidus Liquidus
14
Comparison Cu CuZr - GlidCop
15
Possible implications for CLIC

• the breakdown characteristics are similar for
  all three Cu materials

• a final decision of cavity materials should be
  based on other parameters such as from the
  on-going fatigue measurements.

• the choice of mechanical surface finishing
  techniques are important to shorten the
  breakdown conditioning time.
• EDM 50 and 200 breakdown events for CuZr and
  GlidCop, respectively.
• Milling Immediate conditioning for both Cu
  materials
• more extreme differences between EDM treated and
  rolled Mo electrodes

16
Mo - heated with e-beam
4 hours in air between heating and mounting in
spark system
Conditioning almost immediately to 450 MV/m
17
Mo heated in oven
Faster conditioning
Immediate conditioning not observed
No clear improvement in the conditioning speed
with increasing temperatures
18
Field Stability of Conditioned Mo
1
4
60
1
29
1000oC for 2 h
19
Field Stability of Conditioned Mo
2900
480
453
20
Mass Spectroscopy

• Goal
• To provide quantitative information about gas
  releases during breakdown events.

21
Gas Releases - Mo
Example Release of Hydrogen Gas
Pumping Speed 0.3 Litre/sec
22
Gas Releases - Mo
Release of gas due to breakdown events
Correlation Pressure Rise lt-gt
Number of Molecules
Number of Molecules per unit pressure rise
Releases of H2 and CO gas dominate
Releases of H2 and CO gas dominate
23
Gas releases - Mo
Less energy needed to release H2
24
Gas releases heat treated Mo
Current limiting resistor removed -gt Egap
1/2Cdis U2
An increase in the energy over the gap causes
more gas releases
25
Gas Experiments Air/Mo
Laboratory Air
26
X-ray Photo Emission Spectroscopy
Reference Mo
27
Summary - Gas Experiments
28
Knowledge obtained from DC spark tests

• the ranking of breakdown fields in RF and DC
  experiments is similar for high breakdown rates
• the saturated breakdown fields vary with up to
  one order of magnitude among the studies
  electrode materials
• the effort in finding the optimum material must
  continue
• alloys?
• the breakdown conditioning speed can be
  drastically improved by correct choice of
  pre-treatments
• surface finishing technigues (milling, EDM, ...)
• heat treatments (e-beam ? oven)
• breakdown rate experiments seem to give similar
  results in RF and DC
• Should be given more importance in future studies
• for molybdenum and tungsten the vacuum quality
  influences the ultimate breakdown fields

29
Future Plans

• Finish the construction of the new spark system
• Two systems running in parallel -gt facilitate
  higher throughput of materials and preparation
  techniques
• Improvements in the experimental setup
• XYZ movements
• E-beam heating ? 1000oC
• In-situ treatments
• several samples
• variation of energy over gap more convenient
• Upgrade of maximum voltage to 30 kV
• Dedicate the old spark system to breakdown rate
  experiments
• New setup to increase the repetition rate
• Goal 0.5 Hz -gt 500000 runs corresponds to 12
  days

Antoine Descoeudres
30
Contributors

• Sergio Calatroni
• Ahmed Cherif
• Antoine Descoeudres
• Gonzalo Arnau Izquierdo
• Samuli Heikkinen
• Holger Neupert
• Alessandra Reginelli
• Mauro Taborelli
• Ivo Wevers
• Walter Wuensch
• CLIC Study Team

31
Automatic Spark Conditioning
Molybdenum (Mo) Tip and Sample
Histogram
Spark Scan
32
Depth Profile - Mo
33
Comparison OFE Cu CuZr - GlidCop
1 M. Taborelli, M. Kildemo, S. Calatroni, Phys.
Rev. ST-AB, 7, 092003 (2004) 2 A. Hassanein, Z.
Insepov, J. Norem, A. Moretti, Z. Qian, A. Bross,
Y. Torun, R. Rimmer, D. Li, M. Zisman, D. N.
Seidman, and K. E. Yoon, Phys. Rev. ST-AB, 9,
062001 (2006) Milled Electro Discharge
Machined
34
Mass Spectroscopy
FIG. 2. Calibration to determine the relation
between the recorded current from the RGA and the
corresponding CO pressure. (A) CO pressure as
function of RGA ion current. The open blue
squares represent the experimental values. (B)
Evolution of the pressure during the first few
seconds of a pumpdown of CO. The start pressure
was 510-7 mbar. The measured data is
represented by blue open circles. The red line
shows the resulting linear fit through the
measured points in both figure A and B.
35
Mass Spectroscopy
36
Field Stability of Conditioned Mo