Velocity bunching from Sband photoinjectors Julian Mc Kenzie

Velocity bunching from S-band photoinjectors Julian Mc. Kenzie 1 st July 2011 Ultra Bright Electron Sources Workshop Cockcroft Institute STFC Daresbury Laboratory, UK

Introduction • Normal conducting S-band RF guns are often the gun of choice for modern FELs • Have provided very low emittance beams • However, FELs typically require multiple stages of magnetic compression • Velocity bunching schemes have been proposed for low bunch charge applications such as electron diffraction • Can we apply the same techniques to 100 p. C bunches to serve as an FEL driver?

S-band RF gun • • ALPHA-X / Strathclyde (TU/e + LAL) 2. 5 cell, 2998. 5 MHz Cu photocathode 266 nm laser Courtesy Bas van der Geer, Marieke de Loos, Pulsar Physics

ASTRA simulations • Take on-axis field-maps, feed into ASTRA • Assume thermal emittance as per LCLS measurements: 0. 9 mm mrad per mm (rms) of laser spot* • Assume gun can achieve peak on-axis field of 100 MV/m • Beam energy on exit ~6 Me. V Red = Ez field from gun Blue = Bz field of combined bucking and focussing solenoids • Start with low charge (10 p. C) and scale up to high charge. . . * D. Dowell, “Unresolved emittance issues of the LCLS gun”, 5/08/2010, LBNL workshop on “Compact X-Ray FELs using High-Brightness Beams”

Shortest bunch from gun • At the multiple-picosecond level, it is safe to assume that the bunch length from the gun is equivalent to that of the drive laser • This assumption breaks down sub-ps due to space charge limitations Assumed a 0. 5 mm diameter beam, Gaussian temporal profile, and scanned laser pulse length. Minimum of ~250 fs electron bunch. Similar figure to results from Osaka University, K. Kan et al at IPAC’ 10/Linac’ 10

Shortest bunch from gun • Can reduce bunch length by increasing laser diameter • However, there is a trade-off with emittance • Previously assumed linear correlation between laser spot size and emittance • This emittance cannot be improved but bunch length can • Therefore initially use 0. 5 mm spot, assume best case laser, 10 fs rms

Add bunching cavity • • • 2 m long S-band cavity Operating at the bunching zero-cross phase 7. 5 MV/m Gun Buncher Bunch length continues to increase after gun Operate gun at -15°to help mitigate this effect Place bunching cavity as close to gun as possible Bunch length comes to a focus ~6 m from cathode Minimum of 27 fs rms NB// using different gun/solenoid fieldmaps here

Buncher cavity length RED = 1 m GREEN = 2 m BLUE = 3 m Don’t gain anything by increasing to 3 m, therefore utilise 2 m long buncher

Capture cavity Gun • Add linac cavity at waist to capture the short bunch length • For simulations used 2 m long S -band cavity operating at 20 MV/m • Beam energy on exit ~ 50 Me. V Buncher Linac

Transverse focussing schemes Gun Buncher Linac RED = no solenoids BLUE = solenoid around buncher and before linac GREEN = small solenoid at end of buncher

Optimisation • Utilise genetic/evolutionary optimisation algorithm • Multi-objective shows trade-off between transverse emittance and bunch length • Uses non-dominated sorting technique, based off NSGA-II* • 100 generations of 60 runs each, takes overnight *Kalyanmoy Deb et al. , IEEE Transactions on Evolutionary Computation 6 (2) 2002, pp 183 -197

Optimisation parameters • • Laser spot size (flat-top) Laser pulse duration (Gaussian) Gun field strength Gun phase Gun solenoid strength Buncher field strength Buncher solenoid strength

Optimisation front @ 10 p. C RED = small solenoid at buncher exit Gun Buncher BLUE = solenoids all around buncher Gun Buncher

Manual versus genetic optimisation Gun Buncher Linac

Manual optimisation Genetic optimisation 10 p. C 50 Me. V NB// head of bunch to the right

100 p. C optimisation RED = small solenoid at buncher exit Gun Buncher BLUE = solenoids all around buncher Gun Buncher

Optimisation fronts at various bunch charges RED = 10 p. C GREEN = 100 p. C BLUE = 250 p. C

Selected bunches (100 p. C) Bunch A Bunch B Parameter A B units Emittance 0. 71 1. 10 mm mrad Bunch length 153 23 fs Peak current 331 3340 A Energy spread 58 187 ke. V Energy 48 50 Me. V

Bunch A: 300 A 100 p. C 50 Me. V NB// head of bunch to the right Bunch B: 3 k. A

Optimised parameters • 100 p. C, 50 Me. V Parameter Bunch A Bunch B units Laser radius rms 0. 40 0. 45 mm Laser length rms 140 50 fs Gun peak field 97 71. 5 MV/m Gun phase -2 -10 ° Gun solenoid peak field 0. 248 0. 246 T Buncher peak field 15. 6 14. 6 MV/m Buncher solenoid peak field 0. 038 0. 039 T Linac operated on-crest, 20 MV/m

Bunch B to 240 Me. V Gun Buncher 6 Me. V Linac 0 Linac 1 50 Me. V Linac 2 240 Me. V

Jitter • 500 runs • 10, 000 macroparticles per run • Random jitter based on the following sigmas (cut off at 3 sigma) Parameter(s) Sigma Unit Bunch charge 1 p. C Laser position (x, y) 0. 1 mm Laser timing 100 fs RF gradients 0. 1 % RF phases 0. 1 ° Solenoid strengths 0. 1 % NB// all RF jitter applied individually to each cavity, similarly with solenoids (except bucking and gun solenoid locked)

Jitter ~ 0. 2 mm mrad ~ 15 fs ~ 600 fs ~ 0. 6 Me. V

Tolerances: Arrival time

Summary • A velocity bunching scheme was presented based around an S-band gun and followed by a 2 m long S-band buncher and a further S-band capture cavity • This scheme can provide 100 p. C bunches to the sub-ps level, k. A peak current and 1 mm mrad emittance • Simulated beam parameters are capable of delivering beam to an FEL • However, jitter remains a big issue