1. H 2, partial gas, vacuum furnace for diffusion bonding against brazing 25 Nov. 2014, CERN Toshi Higo (KEK)

2. Contents Experiences and comments Assembly methods GLC/NLC developments Cavity inner surface Junction between mating surfaces Vacuum baking Intentional or associated heat treatment in Hydrogen environment Conclusion 2014/11/25CLIC Structure Review2

3. Assembly methods Stacking – Clamping, electroforming, ….. Welding – EBW, TiG, ….. Diffusion bonding – H 2 furnace – Vacuum furnace with/without partial gas Brazing – H 2 furnace – Vacuum furnace We should study the related change of characteristics in copper bulk or surface 2014/11/25CLIC Structure Review3

4. High gradient performance depending on the assembly method Bulk material and surface – Processed material, HIP, ….. – As machined or heat treated – Hydrogen solution Current flow across junction – Bonded  mutual diffusion  shape preserved – Brazed  flow, meniscus  smooth – Compressed  pressure  need stress 2014/11/25CLIC Structure Review4

5. Technology developed for GLC/NLC 2014/11/25CLIC Structure Review5 Stacking Diffusion bonding Brazing Final brazing and Vacuum baking Machining

6. Diffusion bonding Brazing 2014/11/25CLIC Structure Review6 Well established in 1atm H2 in 100mTorr gas in vacuum Liquid braze material float, slip, flow out Solid process at high temperature for long period with a little pressure small cell slippage small deformation Surface modification surface smoothing absorbed gas, Bulk material change freeing absorbed gas less crystal defect diminishing voids

7. Extended to present-day CERN/SLAC/KEK test flow Design for CLIC (CERN) Fabrication of parts (KEK) Bonding (SLAC) CP (SLAC) VAC bake (SLAC) High power test (NLCTA- SLAC) High power test (Nextef- KEK) 2014/11/257CLIC Structure Review

8. Things to be considered Cavity surface – Conductivity – Surface smoothness Junction – Pressure – Filling material – Gap Geometry preservation – Frequency – Alignment 2014/11/25CLIC Structure Review8

9. Cavity inner surface Modified depending on filled gas or residual gas Crystal orientation and faceting  field enhancement Attacked from heater and others existing in vacuum furnace wall and bottom Gas purity; especially H 2 dew point on water content 2014/11/25CLIC Structure Review9

10. Hydrogen furnace for T24_THU_#1 2014/11/25CLIC Structure Review10 Hydrogen quality should be kept, such as dew point. Stainless steel greening happened from Au brazing! Possible? Copper coated surface may subject oxydization? What about copper inside structure? Sent KEK  Tsinghua

11. Surface change in vacuum furnace due to hour-long high-temperature process Faceting – Crystal cut pattern on surface Sublimation – 1 micron through 900 degC, 1 hour in vacuum Reduced mean free path – Partial filled with 10 Pa gas – With H 2 or Ar, etc. Oxygen on surface – Diffusion into body to disappear 2014/11/25CLIC Structure Review11

12. Long-period process in vacuum furance 2014/11/25CLIC Structure Review12 Other materials Mn, Fe, etc. Back to normal Natural crystal growth, boundary, facet: OK or may not OK on HG Other material should be avoided: evacuation system & or purging cycle, furnace base, heater, etc.  such care should be in mind

13. Other example in hydrogen / vacuum furnace Under metal cap in the same furnace as above In bad vacuum furnace Smoothing in hydrogen furnace Need good vacuum and take profit from gas furnace 2014/11/2513CLIC Structure Review Vacuum furnace 1040 ℃, 1hr Migration into Ti Box Migrated material other than Ti

14. Junction between mating surfaces Welding – Mechanical contact – Contact pressure  stress exists Brazing – Liquid, flow over, meniscus, ….. – Other material than body Diffusion bonding – Mutual diffusion – May suffer from imperfect junction 2014/11/25CLIC Structure Review14

15. Joint made applying liquid-phase process Liquid flow into cavity – Ag-based  flow and fill gap – Au-based  Less flow – Sheet  fill rough gap, but mutual slip Stopping liquid – By dike as CERN-made cavity for KEK in early 1992 – Non-continuous (cut ) groove for wires 2014/11/25CLIC Structure Review15

16. Brazed cavity test through the recent high -gradient view. 2014/11/25CLIC Structure Review16 The same design as CERN 1992. Silver brazed in vacuum furnace brazed as MHI does for many S- and C-band structures. Single-cell test planned at Nextef-B.

17. Junction made by solid-phase process Mutual diffusion 2014/11/25CLIC Structure Review17 Au  Cu, Cu  Cu 800 degC for 1hr? Smooth, a few 10 microns size diffusion makes good bonding to bond 1 micron-class flat surfaces.

18. Bad experience: Large gap created in DB of  80, 900C, 1hr, 0.1MPa 2014/11/25CLIC Structure Review18 Start with <1 micron flatness. Leak tightness confirmed. Huge gap created at the outer side. Process not understood.

19. 2014/11/25CLIC Structure Review19

20. KEK specification on flatness 2014/11/25CLIC Structure Review20 In drawing as shown above But actually the final judgment done by KEK.

21. Flatness typical patterns 2014/11/25CLIC Structure Review21 Specified to be less than 0.5micron Both accepted Convex, Potato-chip Convex Potato-chip

22. Need to study the effect in high gradient feature Imperfect bonding  finite gap 2014/11/25CLIC Structure Review22 Unless the mutual diffusion process starts, no bonding happens. Remaining gap between cells Large current flow across the gap may introduce small sparc Micro gas burst from deep inside, from remaining voids Markus Aicherer

23. Vacuum baking For klystron – Toshiba less than 500 degC for a week or so – SLAC 550 degC, 2 weeks? For structure – SLAC routine for LC: 650 C for 2 weeks – FNAL: 1000C firing in vacuum as parts stage – KEK experience: 60cm structure, 490 C for a week What is the difference among these – The above structures gave reasonable HG performance – Need to be studied 2014/11/25CLIC Structure Review23

24. GLC/NLC standard at SLAC & klystron-like at KEK 2014/11/25CLIC Structure Review24 SLAC 650 ℃～ 2 week VAC baking set in A VAC can H2H2 H2OH2O N 2 O/CO CO 2 Cx HyCx Hy KEK : Double evacuation 500 ℃, ~1week Problem in gasket H 2 evacuation an order of magnitude

25. Partial hydrogen brazing followed by vacuum baking in consideration 1.VAC 900 degC 2.Partial H 2 1hr? 3.Vacuum 1hr? 4.Temp down 5.Extraction and seal in clean environment Ti 、 SUS etc. 2014/11/25CLIC Structure Review25

26. 2014/11/25CLIC Structure Review26 FNAL process for NLC Vacuum furnace + cleanroom From Tug Arkan as of ISG-XI Vacuum furnace connected to cleanroom RF tuning in class 1000 cleanroom No 1atm hydrogen furnace, no baking but 1000C processing High gradient meeting GLC/NLC spec in undamped structure What is the inevitable process? Cleanness may improve the initial processing stage.

27. Intentional or associated heat treatment in Hydrogen environment Surface smoothness – Preferred Cleaning – Carbon contents from surface – Better in high gradient Question: – Hydrogen process is MUST or replaced by others or ….. 2014/11/25CLIC Structure Review27

28. 28 Pulsed DC-HV breakdown experiment at Saitama univ. As diamond turned  + In-situ He sputtering  + Hydrogen brazing process Side BD 1 st inter-electrode BD Carbon on surface plays an important role 2014/11/25CLIC Structure Review

29. 29 After H 2 furnace 1040C, 5min As diamond turned with Kerosene Carbon and oxygen depth profile by AES Carbon decreases through high-temperature process. Cu C C O O 2014/11/25CLIC Structure Review

30. Hydrogen diffusion DiffusionConstant[ Ｔ ]= 8.34*10^-3*Exp[-137320/8.3/ Ｔ ]; Diffusion length 800 ℃, 1hr  40μm Temperature Length [m] 2014/11/25CLIC Structure Review30 Hydrogen cleans surface diffuses into surface and further diffuses into body Effect of hydrogen effect located under thin surface and in body should be studied from high-gradient view point.

31. Hydrogen diffusion into body during diffusion bonding process 31 CERN_EDMS_104124 & 104125 ： Chiara Pasquino, Politecnico di Milano 2014/11/25CLIC Structure Review

32. Many high gradient cavities are made by hydrogen brazing Valery single-cell test cavities at SLAC – Are not always (or never?) vacuum baked at 650C ATF-based S-band RF gun cavities – Made by hydrogen-furnace at KEK – Au-brased at higher than 1000 degC by a few times – High gradient performance Gradual application of power  Gradient of >100 MV/m at cathode is normally realized X-band RF gun tested at ASTA – Not vacuum baked? (Check Juwen) – Operated > 150MV/m at cathode!? These are all SW, but essentially the same in TW!? 2014/11/25CLIC Structure Review32

33. ATF-base AF Gun High precision machining + brazing 交通大学（中国）用 RFGun 早稲田大学用 ECC-RFGun KEK(ATF) 、放医研、産総研、早稲田、 京大、阪大、交通大学（中国）など 合計１３台製作 2014/11/25CLIC Structure Review33

34. Conclusion What to be studied – Hydrogen solution in body – Crystal surface, boundary, facet, … – Surface cleanness, dust free, foreign material free For stable high gradient > 50 MV/m – Which characteristics is needed? – How the above to be realized? To understand the mechanism – Firstly focus at 50 – 100 MV/m range – Should be addressed experimentally – Will be done in a simple setup – Then extend to the actual TW structure 2014/11/25CLIC Structure Review34