CLIC and High Gradient FEL Design

1. CLIC and High Gradient FEL Design D. Schulte for the CLIC collaboration and FEL friends D. Schulte, LSUM, Ankara, October 2013

2. What is the Connection to FELs? • CERN does not do light sources • It is not part of CERN’s mandate • But use of X-band in FELs in other labs would help CLIC for a number of tasks • Further technical developments with industry • Will create the industrial basis • Performance studies of accelerator parts and systems • From components up to large scale main linac system test • We think that FELs can profit from X-band technology • For you to judge based on further studies • Need to find one/several laboratories to build an FEL and help them as needed (including RF, instrumentation, alignment, beam dynamics, test stands, industrial contacts …) • This is whywe do this D. Schulte, LSUM, Ankara, October 2013

3. FEL Overview A. Aksoy Looked a bit into a linac design for a typical Angstrøm FEL We do not know the real user needs Swiss FEL (C-band, approved): E=5.8GeV Q=200pC σz=7μm ε≈200nm-500nm Proposal of Ch. Adolphsen et al. shows concept for X-band E=6GeV Q=250pC σz=8μm ε≈400nm-500nm As example we did chose E=6GeV, Q=250pC, σz=8μm, ε≈400nm Do not study injector (use the one from PSI for now) or undulator D. Schulte, LSUM, Ankara, October 2013

4. Longitudinal Dynamics A. Latina E [MeV] E [MeV] E [GeV] • Example structure: a/λ=0.14 and G=67.5MV/m • σz = 7.96 μm, σE = 0.0071%, σE,slice = 0.0027% • (Swiss FEL: σz=7μm, σE,slice = 0.006%) • Looks promising but detailed studies needed • realistic figure of merit for final beam distribution • radiation in compressors • operational margins • … D. Schulte, LSUM, Ankara, October 2013

5. Transverse Dynamics (Strong) CLIC lattice and simplified wakefield Stability of beam with initial jitter requires to stay above red line Emittance growth with 100um tolerances: We need dispersion free steering or CLIC-style alignment for FEL D. Schulte, LSUM, Ankara, October 2013

6. ExampleSASE FEL Calculation A. Aksoy

7. Cost Minimisation Preliminary Based on simple cost model Uses CLIC structure database (K. Sjobak, A. Grudiev) Beam dynamics constraints included Many solutions at almost the same cost Can chose most reasonable parameter set Need to refine cost model design constraints D. Schulte, LSUM, Ankara, October 2013

8. Electron linac RF unit layout based on the existing (industrialized) RF sources (klystron and modulator) 2x ScandiNova solid state modulators 2x CPI klystrons 410 kV, 1.6 s flat top 50 MW 1.5 s (Operated @45MW) I. Syratchev, modified by me X 5.2 100 (90) MW 1.5 s ~11 m, 16.3 cm TE01 transfer line (RF=0.9) Inline RF distribution network TE01 900 bend Common vacuum network Preliminary 468 MW (418 MW) 150 ns x 10 accelerating structures @68.8MV/m (65MV/m) 46.8MV (41.8MW) input power 10 m, 7.5 active This unit should provide ~516 (488) MeV acceleration beam loading. Need 12 (12) RF units. Cost 51.7 a.u., 4% more than optimum D. Schulte, LSUM, Ankara, October 2013

9. Examples for Basic Parameters Preliminary D. Schulte, LSUM, Ankara, October 2013

10. FEL Interest • Had a meeting of several institutes that are interested in use of X-band for FELs • TAC, Australian Light Source, Elettra, Shanghai, Uppsala , (CERN) • Agreed to join forces to coordinated R&D and exploit synergy • Between CLIC and FELs • Between different FEL projects • Started common discussion groups to report in February • TAC and Australia want to go for new FEL • Similar timescales but TAC more advanced • Elettra and Shanghai want to use X-band for upgrades • Uppsala wants to explore • Other institutes expressed interest in a longer timescale • E.g. PSI, Norway D. Schulte, LSUM, Ankara, October 2013

11. Timescale • Prepare a CDR (O(12months)) • To establish a project with an attractive scope and good, robust design and reasonable funding prospects • To propose and justify R&D phase toward a TDR and project proposal • Mainly theoretical work based on existing hardware experience and simulations • This work will profit from close collaboration between different FEL proponents and CLIC • One can imagine a “modular CDR”, where parts are shared • Prepare a project proposal/TDR (O(4 years)) • This will require hardware developments • E.g. an RF unit • There may be high potential for synergy between different FEL projects as well as CLIC in this phase • FEL construction • Also at this stage collaboration appears beneficial • The level of mutual benefits will evolve with the designs D. Schulte, LSUM, Ankara, October 2013

12. CDR Activities • Defining the goals and main parameters of the FELs • User needs • Beam time structure • Other beam parameters • Putting together an integrated model to evidence beam performance • Sources • Linacs • Photon lines • Defining the RF design • Pulse compressors • Structures • Klystrons • … • Other components could maybe treated in less detail • Cost estimate and site study • Definition of the R&D for the TDR • Coordinated with other projects D. Schulte, LSUM, Ankara, October 2013

13. Options • Baseline: Single bunch with 50Hz • Option 1a: Single bunch with 500Hz • Option 1b: Few bunches with 50Hz • With space to separate them into different photon lines • Option 2: Few bunches with 500Hz • As above • Option 3: Full CLIC bunch train at 50Hz • Can the experiments survive this? D. Schulte, LSUM, Ankara, October 2013

14. Conclusion • X-band seems a good technology for an X-FEL • Simplistic example study with CLIC structure and RF design and soon available commercial klystrons already promises good performance and cost • Your FEL project might profit from X-band • CLIC would profit from fostering the use of X-band technology • We are looking for collaborations on X-band FELs • Would need to define the scope that users like • Number of bunches • Repetition rate • Other beam parameters • Based on this one can define a CDR design • Also cost will need to be considered D. Schulte, LSUM, Ankara, October 2013

15. Reserve D. Schulte, LSUM, Ankara, October 2013

16. 80fs Options 20ms Baseline • Single bunches at 50Hz • Cheapest option • Can either have a single user line or several • But need to distribute pulses 2ms • Single bunches at up to 500Hz • Can either have a single user line or several • More costly klystrons • 10 times higher power consumption • Heating issues in the structures • Injector issues D. Schulte, LSUM, Ankara, October 2013

17. Options II 50ns 20ms • A few bunches in a train spaced by 20-50ns (or so), trains 50Hz • Can use a kicker to distribute bunches into different beam lines • Requires several photon lines • Linac not significantly more costly than baseline • Power consumption similar to baseline 50ns 2ms • A few bunches in a train spaced by 0.5ns, trains 50Hz • Can use a kicker to distribute bunches into different beam lines • Requires several photon lines • More costly klystrons • 10 times higher power consumption than baseline • Heating issues in the structures • Injector issues D. Schulte, LSUM, Ankara, October 2013

18. Options III 0.5ns 20ms • Many bunches (e.g. 300) in a train spaced by 0.5ns, trains 50Hz • Will all go down the same photon line • Linac not significantly more costly than baseline • Power consumption similar to baseline • Is this good for anybody? • At which spacing could the bunches be used in the same line? D. Schulte, LSUM, Ankara, October 2013