A two-oscillator echo enabled tunable soft x-rays

Transcription

1 A two-oscillator echo enabled tunable soft x-rays FLS 2010 Workshop SLAC J.S. Wurtele Co workers: P. Gandhi, X.-W. Gu, G. Penn, A. Zholents R. R. Lindberg, K.-J. Kim 1. Overview of scheme 2. Walkthrough of system 3. Preliminary performance 4. Conclusions

2 Echo Enabled FEL Stupakov, PRL 2009 Radiator Compact method for lasing at high harmonics without cascade Stability of bunching determined by ebeam and seed lasers Longitudinal coherence and small bandwidth is important for many users FEL oscillators provide longitudinal coherence but are not tunable over a wide band

3 Hard X-Ray FEL Oscillator Store an X-ray pulse in a Bragg cavity multi-pass gain & spectral cleaning Provide mev bandwidth (Δω/ω ~ 10-7 ) MHz pulse repetition rate high average brightness Originally proposed in 1984 by Collela and Luccio and resurrected in 2008 (KJK, S. Reiche, Y. Shvyd ko, PRL 100, (2008) KJK FLS 2010 March 1 5, 2010 KJK But mirrors are worse in the soft x-ray regime 3

4 A two-oscillator echo enabled tunable soft x-rays Advantages: Eliminates seed lasers (hard to make at high rep rate) Longitudinal coherence from oscillators Tunable by changing R56 (echo scheme) Radiator Disadvantages: Requires optics in soft xray regime This study is preliminary and not optimized; there may be surprises!

5 FEL Model We use the model of Lindberg and Kim (PRST-AB, 2009). Transverse physics is included using a modal analysis and integrated transverse orbits (finite emittance effects) The particle motion is one-dimensional. The code is run in MATLAB and will be freely available. θ j = 2η j k k u J j η j = K[JJ] 2γ 2 1 k u E µ eiθ + C.C. w = w K[JJ] I π Re < e iθ > H γ k u Eµ * + 2 α kk u w α = K[JJ] I 2π Im (1 + iα) < e iθ > H γ k u Eµ * α 2 kk u w 2 E = K[JJ] I 2π < e iθ > ie K[JJ] γ k u µ * γ H = 4(1+ z n 2 ) w 2 + (1 + iα x )(1+ z n 2 ) 2 I π Im 1 + iα k u 1 iα < e iθ > Eµ * H z n = (z z 0 )σ px σ x = 2(s s 0 )σ px k u 1 µ = w e itan 1 (α ) π (1 iα)(1+ z 2 n ) + w 2

6 The tunable two-oscillator echo-enhance FEL Oscillator 1 Oscillator 2 Radiator Chicane 1 Chicane 2

7 The tunable two-oscillator echo-enhance FEL Tunability is realized by changing R 56 (as in echo scheme). Oscillators are held at fixed frequencies and there is no need for seed lasers. Oscillator 1 Oscillator 2 Radiator Chicane 1 Chicane 2 Undulator a w =7.2 λ w =7 cm L w =21 Radiation λ = 43nm w = 120µm Power = 2 MW Optics Cavity Size= 37 m Mirror Radius= 36.5 m Power Loss = 75% Undulator a w =13.7 λ w =10 cm L w =10 m Radiation λ = 215nm w =1500 µm Cavity Power = 380 MW Optics Cavity Size= 37 m Mirror Radius= 77 m Cavity Power Loss = 0.5%

8 Power build up in 200 passes

9

10

11 Performance -FEL stability requires nearly identical bunching shot-to-shot -After start-up, run simulation. -Plot phase space at critical locations every 25 passes - Simulated with a single slice model with no time-dependence -The laser power output exhibits some slow variation with pass number, which needs to be understood.

12 Pass 125

13 Pass 150

14 Pass 175

15 Pass 200

16 Phase space at last 9 of 200 passes for the first oscillator

17 Power build-up in the first oscillator Final Pass Initial pass

18 Tunability -The echo enabled harmonic bunching peak can be shifted by adjusting the chicanes. -This is a consequence of the echo scheme, not the two-oscillator configuration -The oscillator scheme (within the single-slice model) allows for tuning.

19

20

21

22

23

24

25

26 Issues for future study 1. Further optimization---detailed studies for longer numbers of passes 2. Time dependent, 2d simulations in the oscillator (Ginger) 3. Slippage and cavity detuning 4. Sensitivity to errors 5. Mirror damage limits 6. Low current limitations (need enough power to bunch) 7. Limits to performance (beam brightness, harmonics) 8. Size minimization (also set by rep rate) 9. Parameters for proof-of-principle experiment at JLAB

27 laser field along wiggler ) matlab code W Genesis code ( 10 8 r e w o P r e s a L Comparison for single pass high gain FEL ) 10m 0 ( e z 10i -2 s z (m) m a 10 x Laser Beam Waist Vs laser beam size e 10-4 b matlab code for laser Waist 9 r Genesis code for laser beam size e 8 s a l 7, ) 6 m ( 5 t s i a W z (m) Benefits of a Matlab simulation tool based on the reduced model: -Requires fewer macro particles (10x 100x) and less computation time compared to a 3D code such as GENESIS. -Matlab interface is easy to modify and highly portable. Further work: -Investigate the regimes in which the model fails to capture 3D effects properly and understand why this happens. In particular, there is an as yet not explored discrepancy in the radiation waist between the reduced model and GENESIS after saturation -Implement time-dependence in the Matlab code.

28 laser field along wiggler ) matlab code W Genesis code ( 10 8 r e w o P r e s a L Comparison for single pass high gain FEL z (m) Benefits of a Matlab simulation tool based on the reduced model: -Requires fewer macro particles (10x 100x) and less computation time compared to a 3D code such as GENESIS. -Matlab interface is easy to modify and highly portable. Further work: -Investigate the regimes in which the model fails to capture 3D effects properly and understand why this happens. In particular, there is an as yet not explored discrepancy in the radiation waist between the reduced model and GENESIS after saturation -Implement time-dependence in the Matlab code.