1. Vertical Test Results of 9-Cell Cavities for LCLS-II
Dan Gonnella
For the LCLS-II Project

2. Outline
Introduction to LCLS-II
Challenges in Production
Cavity Vertical Test Results
Summary & Outlook for ILC

3. Outline
Introduction to LCLS-II
Challenges in Production
Cavity Vertical Test Results
Summary & Outlook for ILC

4. Introduction to LCLS-II
The SRF linac is closely based on the XFEL/ILC/TESLA Design
It consists of 35 cryomodules each with 8 cavities – total of 280 cavities
The GHz 9-cell cavities have a very ambitious Q specification:
Additionally, cavities must reach 19 MV/m
2.5 x 1010 at 16 MV/m and 2 K
2.7 x 1010 at 16 MV/m and 2 K
Q specification lowered in VT due to addition of stainless steel blank on short side of cavity – adds ~0.8 nΩ of Rres

5. Cavity Preparation
In order to achieve the ambitious Q specification, all 9-cell cavities for LCLS-II are prepared with nitrogen-doping
Original recipe:
140 μm bulk EP
800oC degas for 3 hours in vacuum
2 minutes at 800oC in mTorr of N2
6 minutes at 800oC in vacuum
5-7 μm light EP

6. Cavity Preparation
In order to achieve the ambitious Q specification, all 9-cell cavities for LCLS-II are prepared with nitrogen-doping
Original recipe:
140 μm bulk EP
800oC degas for 3 hours in vacuum
2 minutes at 800oC in mTorr of N2
6 minutes at 800oC in vacuum
5-7 μm light EP

7. Cavity Preparation
In order to achieve the ambitious Q specification, all 9-cell cavities for LCLS-II are prepared with nitrogen-doping
Original recipe:
140 μm bulk EP
800oC degas for 3 hours in vacuum
2 minutes at 800oC in mTorr of N2
6 minutes at 800oC in vacuum
5-7 μm light EP
Cavities are being produced by Research Instruments GmbH (RI) and Ettore Zanon S.p.a (EZ)
Niobium was procured from Tokyo Denkai and OTIC Ningxia

8. Outline
Introduction to LCLS-II
Challenges in Production
Cavity Vertical Test Results
Summary & Outlook for ILC

9. Challenges in Production
Early LCLS-II cavities were plagued by two main issues that limited performance below the required cavity spec:
Poor flux expulsion in the raw niobium sheet material
Improper fabrication techniques at one of the cavity vendors

10. Q vs E for First Article Cavities
Some Great Cavities
Some Good Cavities
All these cavities tested in very low (0-1 mG) magnetic fields

11. Residual Resistance in First Articles
Rres higher than in the prototyping phase TD
Expected 2-3 nΩ, found 5-9 nΩ
Spread is a result of differences in ambient magnetic fields

12. Issues with Original Recipe
Material from both vendors showed worse flux expulsion when treated at 800oC in single-cell cavity tests than material used during the prototyping and R&D stage
All material meets specifications
This means that magnetic field specifications would need to be tighter in order to minimize the need for efficient flux expulsion
Cryomodule results thus far have shown ambient magnetic fields less than the LCLS-II CM Spec of 5 mG – further improvement on this would be difficult
140 μm bulk EP was insufficient to remove the damage layer, adding additional Rres
Worse flux expulsion is an independent phenomena from N-Doping – It impacts all cavity preparation methods

13. Cavity Preparation
In order to achieve the ambitious Q specification, all 9-cell cavities for LCLS-II are prepared with nitrogen-doping
Updated recipe:
200 μm bulk EP
900oC degas for 3 hours in vacuum for TD cavities
950/975oC degas for 3 hours in vacuum for NX cavities
2 minutes at 800oC in mTorr of N2
6 minutes at 800oC in vacuum
5-7 μm light EP

14. Residual Resistance After Recipe Change
Spread lower due to near 100% FE TD
After Increasing Bulk EP and Degas Temperature,
Rres is consistently ~2 nΩ

15. Q vs E for First Article Cavities

16. Q vs E for First Article Cavities
Highest Q0 Ever Obtained in Production SRF Cavities
Cavities now consistently meet Q0 specification

17. Improper Cavity Fabrication
Q Slope Onset MV/m
Cavities produced at Vendor A showed significantly lower quench fields than those produced at Vendor B
Additionally a strong Q slope was present in many of the cavities, suggesting normal conducting inclusions were present
LCLS-II project staff were placed on site to evaluate and fix the underlying problems
Subset of cavities, behavior observed in most

18. Vendor A – Surface Roughness
This Issue was Fixed
Data from Y. Trenikhina
EP at Vendor A was producing a very rough surface
Roughness was as high as 15 μm
It was found that EP temperature was leading to the EP no longer being in the “polishing regime” but in the “etching regime”

19. Vendor A – Grinding Technique
Visual inspection of the interior of Vendor A cavities after fabrication showed the presence of many defects
Some pits had visible normal conducting media embedded
Theorized that aggressive grinding on the dumbbell surface led to the “burying” of media
This normal conducting media would then be revealed after bulk EP
All procedures were reworked with LCLS-II staff to improve the fabrication procedures
Additionally, it was found that a dirty weld stackup procedure was leading to weld defects which further limited cavity gradient

20. Changes to Vendor A Procedures
After evaluation and rewriting of Vendor A cavity fabrication procedures, performance for Vendor A cavities significantly improved
Performance is now in line with cavities produced at Vendor B and what we would expect
Cavities that were fabricated with improper procedures will be reworked

21. Outline
Introduction to LCLS-II
Challenges in Production
Cavity Vertical Test Results
Summary & Outlook for ILC

22. Quench Field Results - Vendor A
Maximum Eacc
Category
<Emax> [MV/m]
Original Grinding
17.6
Restart
New Weld-Stackup
Rework

23. Quench Field Results - Vendor A
Maximum Eacc
Category
<Emax> [MV/m]
Original Grinding
17.6
Restart
22.4
New Weld-Stackup
Rework

24. Quench Field Results - Vendor A
Maximum Eacc
Category
<Emax> [MV/m]
Original Grinding
17.6
Restart
22.4
New Weld-Stackup
24.0
Rework

25. Quench Field Results - Vendor A
Maximum Eacc
Lesson Learned:
Great care and vendor oversight is necessary to ensure good cavity fabrication
Category
<Emax> [MV/m]
Original Grinding
17.6
Restart
22.4
New Weld-Stackup
24.0
Rework
20.0

26. Quench Field Results - Vendor B
Maximum Eacc
Vendor B cavities show an average quench field of 22.8 MV/m

27. Quench Field Results By Material
Maximum Eacc
Vendor B cavities show an average quench field of 22.8 MV/m
No significant difference in quench fields for cavities made from different materials

28. Q0 Results - Vendor A
Q0 at 16 MV/m
Cavities made by Vendor A with the original grinding procedures showed low Q0 primarily due to strong Q slope
Improved procedures has resulted in an average Q0 of 3.2x1010

29. Q0 Results - Vendor B Q0 at 16 MV/m
Q0 performance strongly depends on material and heat treatment due to differences in flux expulsion

30. Q0 Results - Vendor B Q0 at 16 MV/m
Q0 performance strongly depends on material and heat treatment due to differences in flux expulsion
Material/Heat Treatment
<Q0>
TD/800
2.5x1010
TD/900
NX/900
NX/950/975

31. Q0 Results - Vendor B Q0 at 16 MV/m
Q0 performance strongly depends on material and heat treatment due to differences in flux expulsion
Material/Heat Treatment
<Q0>
TD/800
2.5x1010
TD/900
3.3x1010
NX/900
NX/950/975

32. Q0 Results - Vendor B Q0 at 16 MV/m
Q0 performance strongly depends on material and heat treatment due to differences in flux expulsion
Material/Heat Treatment
<Q0>
TD/800
2.5x1010
TD/900
3.3x1010
NX/900
2.6x1010
NX/950/975

33. Q0 Results - Vendor B
Q0 at 16 MV/m
Q0 performance strongly depends on material and heat treatment due to differences in flux expulsion
Lesson Learned:
Flux expulsion must be verified on a material batch level to ensure Q0 meets specification
Material/Heat Treatment
<Q0>
TD/800
2.5x1010
TD/900
3.3x1010
NX/900
2.6x1010
NX/950/975
3.5x1010

34. Outline
Introduction to LCLS-II
Challenges in Production
Cavity Vertical Test Results
Summary & Outlook for ILC

35. Summary & Outlook for ILC
Early LCLS-II cavity production faced two main challenges to meeting the required specifications
In order to meet gradient spec, strict vendor oversight and reworking of procedures was necessary
This will become even more important for ILC which requires higher gradients than LCLS-II
In order to meet the Q0 spec, a recipe change was required twice mid-production
All future niobium should first be checked for flux expulsion on single-cells
More long term: identify a material spec that governs flux expulsion

36. Summary & Outlook for ILC
Both cavity vendors are now consistently producing cavities that exceed the very ambitious LCLS-II specifications
Average Q0 for the project has been demonstrated to be well above 3x1010
Cavities show an average quench field of ~23 MV/m, consistent with expectations on nitrogen-doped cavities produced with the LCLS-II recipe