Superconducting Accelerator for ERL based FEL EUV light

Superconducting Accelerator for ERL based FEL EUV light source at KEK/Japan Hiroshi Sakai, Kensei Umemori, Taro Konomi, Takayuki Kubo, Eiji Kako (KEK), Masaru Sawamura(QST), Tomoko Ota(Toshiba) 1. Introduction for high-power EUV Light source 2. Learn from c. ERL cryomodule 3. EUV main linac cavity & cryomodule design 4. Development for EUV main linac cavity 5. Summary 6 th IHEP-KEK meeting. (2017. Jul. 15) 1

Introduction • 10 -k. W class EUV sources are required in the future for Next Generation Lithography In order to realize 10 -k. W class EUV light source, ERL-FEL is the most promising light source (High repetition rate (≦ 1. 3 GHz) and high current linac system). Breakthrough for EUV light by using FEL (with ERL) X-ray pulse duration ~ 50 ps Micro-bunching -> SASE lasing high peak power X-ray pulse duration ~ 10 fs LPP of 13. 5 nm 250 W level now Need breakthrough for higher EUV light (>1 k. W) G. Dattoli et al. , NIM-A (2001) In case of normal conducting accelerator, The repetition rate of FEL is less than 100 Hz High repetition with SC cavity is needed for k. W laser

EUV/X-ray FELs Low power LCLS SACLA Type of linac Normal conducting Operation mode Pulse FLASH Euro-XFEL High power LCLSII EUV-FEL Super conducting Long pulse CW US Japan Germany US ------- ERL scheme No No No Yes Repetition rate 120 30～ 60 <5000 <27000 1 M 162. 5 M Beam energy （Me. V) 14300 6000～ 8000 1250 17500 4000 800 Wavelength(nm) 0. 15 0. 08 4. 2 -52 0. 05 ~0. 3 13. 5 Pulse energy(m. J) ~10 <0. 5 ~10 ~1 ~0. 1 Average Power （W) ~1 ~1 <0. 6 ~1000 >10000 Beam dump power （W) ~1. 5 k ~0. 5 k ~6 k ~0. 5 M ~1 M ~0. 1 M Status Operation 2009 Operation 2011 Operation 2004 Construction 2017 Construction 2020 Planning Country ERL helps to make high power CW FEL and reduce the beam dump power (important in future)

Design Concept for high repetition high current EUV-FEL • Target : 10 k. W power @ 13. 5 nm, (800 Me. V, 10 m. A) EUV Source (ERL) • Use available technology (based on SASE-FEL) without too much development • Make ERL scheme by c. ERL designs, technologies and operational experiences p p 2 nd turn t RF field p t t Injector Energy recovery by SRF cavity Beam Dump R 56 < 0, T 566 < 0 1 st turn 2 nd turn bunch t t Bunch compre ssor 1 st turn Arc + Chicane Main Superconducting Linac SASE-FEL is an available technology now Undulator（for FEL) Bunch decompr essor 2 nd Arc R 56 > 0, T 566 > 0 FEL light p p t t Energy recovery is needed for accelerating more than 10 m. A to reduce beam dump and save RF power. This operational experience with high current is studied in Compact ERL (c. ERL) at KEK

Learn from Comact ERL (c. ERL) in KEK (c. ERL cryomodules) Main linac module See details by [K. Umemori ; ”Long operational experience with beam in Compact-ERL cryomodules”. SRF 2017] HOM damped (for 100 m. A circulation to suppress HOM-BBU in design) 9 -cell cavity (ERL-model 2)× 2 Compact ERL Layout RF frequency: 1. 3 GHz 20 Me. V Recirculation (return) loop o agn i rd cto e j In line Input power : 20 k. W CW (SW) stic Eacc: Unloaded-Q: Dump Main-linac Cryostat Merger 8. 6 MV/m（operational） Q 0 > 1 1010 Two 9 -cell SC cavities HOM absorber Tuner Injector-linac Photocathode DC gun (Not SRF Gun) 3 Me. V 2 -cell Cavity Tuner Injector module 2 -cell cavity × 3 Double coupler Input couplers E. Kako MOPB 097 SRF 2017 e. Requirement was satisfied at V. T. Heavy F. E was met @9 -10 MV/m after string assembly. Cryostat RF frequency: 1. 3 GHz Input power : 10 k. W/coupler (10 m. A, 5 Me. V) 180 k. W/coupler (100 m. A, 10 Me. V) Eacc: Input Coupler 7. 6 MV/m（5 Me. V） 15 MV/m (10 Me. V) Unloaded-Q: Q 0 > 1 1010 e- HOM Coupler & RF Feedthrough Requirement was satisfied at V. T and for initial 10 m. A requirement. Success 1 m. A operation 100% Energy recovery without HOM-BBU. 5 Operated from 2013. Current increased.

Parameter Specification Wavelength 13. 5 nm Output power 10 k. W Bunch chare 60 p. C Beam energy 800 Me. V Accelerating gradient 12. 5 MV/m (main linac) Number of SRF cavity 9 -cell cavity× 64 Beam repetition 162. 5 MHz Beam current 9. 75 m. A Design strategy (main linac) Epeak/Eacc is 1. 5 times reduced from c. ERL cavity to overcome field emission. 8. 6 MV/m 12. 5 MV/m 10 k. W FEL output 2 nd Arc ~20 m decompressed EUV-FEL Design Presented by Norio NAKAMURA ERL 2015（https: //www. bnl. gov/erl 2015/ ） Bunch compressed 1 st Arc Chicane ~200 m Beam Dump 10 Me. V, 10 m. A A 0 m 1 , Undulators(FEL) V e M 800 Main Linac ation er l e W ec V D Injector Diagnostic Line M / e 8 ion 10 M of ～ t a A r / y e V r cel 0 Me cove x 10 m c A 80 re Energy recovery V M y o Merger t erg 90 Is needed. 7 Injector Linac En Gun 10 Me. V, 10 m. A

FEL Performance by simulation Electron beam parameters: E=800 Me. V, Qb=60 p. C, fb=162. 5/325 MHz Helical undulator parameters: K=1. 652, lu=28 mm, Lu=4. 9 m(18 units), Lg=1. 12 m Bunch compression scheme: 1 st Arc(DBA), R 56=0. 3115 m 88. 5 m. J 79. 5 m. J FEL power without tapering: 12. 9/25. 8 k. W @ 9. 75/19. 5 m. A FEL power with 2% tapering: 14. 4/28. 8 k. W @ 9. 75/19. 5 m. A 10 k. W class high power EUV light source is NOT just a dream!

Injector & Merger Design ref. E. Kako “MOPB 097” poster in SRF 2017 Cryomodule design Vacuum Vessel Segmented ceramic duct 5 K Duct 5 K Panel 80 K Shield 2 K He Jacket Electron beam 2 K Gas Return Pipe Anode (0 V) Ti chamber NEG pump Cathode (-500 k. V) c. ERL Injector cryomodule Photocathode DC gun e- Input coupler HOM coupler cavity cells 5 K Support 80 K Base-plate Input Coupler 2 -cell Cavity HOM Coupler Solenoid magnets Buncher cavity 2 -cell SC cavity B : Bending magnets（θ=15°, ρ=1 m） Q : Quadrupole magnet injection point recirculation loop electron beam Einj~10. 5 Me. V Injector section (side view) matching section Merger section (top view) Injector part of c. ERL will be used for EUV-FEL light source.

Design of Main Linac Cavity How to overcome field emission EUV cavity – TESLA-type 9 -cell cavity + Large beam pipes(100 f & 110 f) Only end cell was modified to match the impedance to beam pipe. EUV cavity f 100 HOM f 70 HOM damper f 110 HOM damper c. ERL cavity (Model 2) – HOM damped cavity for 100 m. A operation f 80 HOM damper Parameters for acceleration mode ERL Model 2 EUV 1300 MHz Rsh/Q 897 Ω Ep/Eacc 3. 0 Frequency HOM damper ERL Model 2 EUV Iris diameter 80 mm 70 mm ~1000 Ω Qo×Rs 289 Ω ～ 270 Ω ~ 2. 0 Hp/Eacc 42. 5 Oe/(MV/m) ～ 42. 0 Oe/(MV/m) From c. ERL stable beam operation of 8. 6 MV/m in 3 years with less trip ratio. Stable operation at 12. 5 MV/m seems achievable due to reduced Ep/Eacc.

Detailed calculated parameters of EUV main linac cavity • • Acc. Ep/Eacc is 2. 0 because the center cell is TESLA shape. EUV monopole HOM is lower than c. ERL because the c. ERL was optimized for dipole HOMs Cavity Parameters Frequency (MHz) Iris diameter (mm) R/Q (Ω) G (Ω) Ep/Eacc Hp/Eacc (m. T/(MV/m)) BBU limit KEK-EUV 1300 70 1009 269 2. 0 4. 23 >190 m. A Monopole HOM KEK-c. ERL 1300 80 897 289 3. 0 4. 25 ~600 m. A TESLA 1300 70 1036 270 2. 0 4. 26 ~10 m. A Calc by Superfish Norio NAKAMURA ERL 2015（ https: //www. bnl. go v/erl 2015/ ） Dipole HOM Calc by MW-Stdio (c. ERL case calc by MAFIA) EUV cavity satisfied our requirement of 10 m. A beam operation by keeping Ep/Eacc ~ 2.

Concepts of EUV main linac cryomodule Cold Box Support Room Temp. 80 K Shield 80 K 5 K Shield 80 K 80 K 2 K Return Pipe & Cavity Support 80 K 80 K HOM Damper Input Coupler Damper 2 K gas out HOM antenna &BPM 2 K 2 -Phase Tuner 5 K out 80 K out 2 K in 5 K in Cable W/C line 80 K in 2 K 2 -Phase 5 K 80 K 300 K Tuning Shaft • • Input Coupler Tuning Shaft EUV module consists 4 cavities and the design based on STF and ERL module. 11 Coupler position is opposite direction each cavity, because the beam pipe sizes are different. Input coupler and tuner are same type of ERL and STF. HOM damper needs new development.

Detailed and Modified from c. ERL main linac cryomodule Input coupler • c. ERL main linac coupler is working well up to CW 15 k. W power (double windows) • Qext=2 x 107 require 4~5 k. W input power for Eacc=12. 5 MV/m • We could apply for EUV main linac cavity by using c. ERL main linac coupler. Frequency tuner • Rough tuning by Slide-Jack tuner controlled by motor • Full stroke 3 mm (~1 MHz) • Fine tuning by piezo tuner • Precision <nm • Working very well at c. ERL and STF • Apply for EUV c. ERL Input coupler 12 c. ERL Tuner

Test of HOM damper for EUV cavity • • • c. ERL HOM absorber (ferrite) has cracks during cool down and not bakable ⇒ not good for SRF usage. Al. N (Sienna Tec. : STL-150 D) is a candidate for absorber. ⇒ Al. N is already tested at Jlab and DESY We started the measurement of RF parameter @ low temperature, outgassing and testing the brazing. c. ERL HOM damper HIP ferrite change Al. N ring RF parameter measurement @ 80 K Al. N sample Cool to 80 K Little difference of epsilon of Al. N between R. T and 80 K We have enough absorption by using Al. N material @80 K. 13

Outgassing rate of Al. N ring was measured after 48 h x 150 o. C baking. After 1000 h, outgassing rate is lower than 10 -8 Pa*m 3/s/m 2 Surface area 0. 020724 Al. N ring 1 day c. ERL damper is not bakeable. So water rinse is not allowed. Al. N (STL-150 D) damper is bakeable and easy for cleaning. It is enough for installation inside the cryomodule ~ Measured by Shinji Terui (KEK) 14

Brazing test of Al. N based HOM damper prototype • • • Two Al. N cylinder were brazed in the copper cylinder which has the comb pattern. Brazed by Silver at 750 degree under Hydrogen Furnace. We tried thermal test. Unfortunately, Al. N cylinder is broken after first 80 K thermal cycle. Ultrasonic testing in the water bath was done after brazing. Al. N ring of 2 nd prototype much tighter connected than that of 1 st one. But not all area of Al. N ring was connected. We need to search more tighter brazing condition including thermal cycle. 1 st : Test piece of brazing Same brazing condition but shape was optimized to fit Al. N at center position. 2 nd : HOM damper prototype Ultra sonic test Crack to 80 K 2 nd prototype is tighter connected. 60% area connected but not perfect. Connection was not tight and crack occurred under cooling to 80 K (but very fast (1 hour) ). T. Ota et al. , “Development of HOM absorbers for CW Superconducting cavities in Energy Recovery LINAC” (SRF 2017)) 15

For reliable operation after assembly work Toward the reliable operation, • We made the horizontal test stand for testing the performance after cryomodule assembly including HOM damper, input coupler, tuner and magnetic shield. • Furthermore, now we try to establish the local clean boose and slow pumping system in this horizontal test standto overcome field emission after string assembly. For making very clean local boose Horizontal test stand was build and already carried out high power test in KEK. Slow pumping system Open clean bench KOACH (ISO class 1) 16

EUV-FEL Light Source Study Group for Industrialization since 2015 By all Japan association to realize EUV-FEL light source

Summary • EUV-FEL-ERL is promising to open the era to the highest EUV light source and can make more than 10 k. W@ 13. 5 nm EUV light in design. • We learned that most severe problem is field emission on long beam operation. • EUV cavity has been designing by including c. ERL beam operation experience for EUV-ERL/FEL accelerator. • Cavity based on KEK-c. ERL +TESLA cavity to reduce Epk/Eacc. ⇒TESLA center cell + beam line damper with modified end-cell & beam pipe • This designed cavity could make 10 m. A beam operated without HOM BBU instability and large HOM heat load with 12. 5 MV/m accelerating field by extrapolated from c. ERL operation. • Cryomodule has been designing based on STF+ERL cryomodule. • HOM damper was newly developed for EUV cavity. We found Al. N material is suitable for our cryomodule HOM damper on RF parameters and outgassing. Unfortunately, we did not find good brazing condition. • For a reliable cryomodule operation, we made horizontal test stand for study of more reliable string assembly work to prevent the dust coming into the cavity. • We also do the nitrogen doping/infusion work for future high-Q operation. (See poster; T. Konomi, et, al. , “Trial of Nitrogen Infusion and doping by using J-PARC Furnace”, THPB 021 SRF 2017 poster ) 18 • Finally, all Japan association with KEK promote to build EUV-FEL light source.

Thank you for your attention!

backup 20

Successful ~ 1 m. A CW beam operation with energy recovery at c. ERL by using KEK ERL-model-2 cavity beam current : CW 0. 9 m. A at 8 th/Mar/2016. Beam current (CW) Energy Recovery is almost 100. 0% (error +-0. 03%) Δptotal=Δ(Pin 1+Pin 2 -Pref 1 -Pref 2) No HOM-BBU was observed. 17. 14 MV x 900 u. A = 15. 4 k. W : KEK ERL-model-2 cavity 1) Large beampipe + HOM absorbers (not HOM coupler) 2) Optimize cell shape for ERL operation from TESLA HOM absorber Frequency Q 0 (): TESLA cavity Gradient 28 MV/m measured VT 1 e+10 Coupling 3. 8 % (1. 9%) 897 Ω (1007Ω) Ep/Eacc 3. 0 (2. 0) （lower）ML 2 8. 57 MV （upper） ML 1 8. 57 MV HOM absorber 1300 MHz Rsh/Q Main linac HOM BBU 600 m. A (10 m. A) Hp/Eacc 42. 5 Oe/(MV/m) Pref Pin e-