A bunch compressor design and several X-band FELs Yipeng Sun, ARD/SLAC 2011-04-13, LCLS-II meeting

A bunch compressor designand several X-band FELs

Yipeng Sun, ARD/SLAC

2011-04-13, LCLS-II meeting

Presentation TitlePage 2

Design of two bunch compressors

Bunch compressionPage 3

Magnetic bunch compression

e- source

Energy modulation (correlation):RF structure, laser, wake field etc.

RF3-Dip. Chicane

Dispersive region:chicane, wigglerarc, dogleg etc.

Bunch L phase space

Different bunch compressors

3(4) dipole chicane, R56 <0, T566 >0achromatic to any order

Wiggler, R56 <0, T566 >0achromatic to any order?

Arc, R56 >0, T566 >0

2-Dip. Dogleg w/ quad+sextupole, R56 >0, T566 tunable

Chicane w/ quadrupole+sextupole, R56 tunable, T566 tunable

NLCTA chicane shape


Dispersion relations


Bunch compressor with dipoles and drifts


General chicane (1)


General chicane (1)


General chicane (2)


General chicane (2)


An FEL withLCLS injector (S-band+X-band

harmonic)Plus X-band Linac2 and Linac3


Scaling

Total length of acceleratorAssume 70% RF in linac

250-10pC

Final bunch length versus bunch charge


Longitudinal wake potential

'long' range


Linac3 length needed for de-chirp after BC2


Accelerator shape (LCLS injector + X-band)


LiTrack, LCLS, 250pC, 3kA


LiTrack, LCLS injector+X-band, 250pC, 3kA


Optics

LCLS

LCLS-Injector + X-band


Elegant simulation, 250 pC, 3 kA (w/ and w/o CSR)

LCLS

LCLS w/o CSR


Elegant simulation, 250 pC, 3 kA

LCLS

LCLS-Injector + X-band (½ R56 in BC2, 0.7 bending angle)



LCLS (L3, 30degree)



Elegant simulation, 250 pC, 5 kA, Projected emittance

LCLS (L3, 30degree)



Elegant simulation, 250 pC, 5 kA, Trajectory

LCLS (L3, 30degree)



LCLS-Injector + X-band (0.5 R56 in BC2, 0.7 bending angle), 250 pC, 5 kA

BC1 end BC2 end Linac3 endBC2 entrance


Potential X-band advantage over S-band

• Maintain a flat energy profile when pushing for shorter bunch length and higher peak current (i.e. 6kA at 250pC), due to stronger X-band longitudinal wake in Linac3, to remove energy correlation (chirp); plus possible cancellation of nonlinear chirp between RF, wake and CSR effects.

• Similar or smaller CSR emittance growth in BC2, benefiting from a weaker dipole and a larger energy correlation generated in Linac2 (previous argument)

• Compact (300m vs 1000m, at 14GeV)

• For LCLS, increasing current from 3kA to 6kA requires a smaller L1 phase to generate a longer bunch in ~400m Linac2, so that the L wake chirp is much smaller, and the bunch is compressed more in BC2 with same L2 phase; if keeping similar L1 phase and increasing L2 phase (i.e. from 36d to 37.5), the final energy profile will be very nonlinear.



LCLS (L1, 19degree; L2, 36degree; L3, 30degree)

LCLS (L1, 22degree; L2, 37.5degree; L3, 0degree)


An X-band RF based FEL with optics linearization

250 pC


Bunch length after compression

Final coordinate (square)

Minimum lengthMinimum length

Neglect small initial un-correlated energy spread

1st order optimal compression:

2nd order optimal compression:

3rd order optimal compression:


Full compression using optics linearization

1st order dispersion

2nd order dispersion

3rd order dispersion


Minimize CSR (1) short interaction time

New design BC1 (1) first order

B10.2m7 degree

B20.2m3 degree

B30.2m-3 degree

B40.2m-7 degreeQD

QF

R56 = 17 mm

New design BC1 (2) second order

SF1&2 SD1&2 symmetricK3(SF1) = -K3(SD2)K3(SF2) = -K3(SD1)

T166 = T266 = 0; T566 = 170 mm


Minimize CSR (2) phase space matching

general

x

X’

CSR

Large βx

X’

CSR

Small β x

X’

CSR

Optimalβ and α

specific

Optimized to minimize CSR impact on emittance


X-band based 2 stage FEL (1) 250pc, 300micron


Final profile at 7GeV (collimation in middle of BC1)


Slice emittance evolution, 250 pC, 6 kA

BC1 entrance BC2 entrance Linac3 endBC1 end


An X-band RF based FEL with normal chicane BC

10 pC


Max bunch length w/o harmonic RF


Bunch compressor and linac design

BC1 BC2 Linac cell

X-band based 2 stage FEL (3) 10pc, 40micron

54 MeV (C. Limborg) 6 GeV

FEL simulation Setup

• FEL at 2 keV , 6 Å (FEL at 8 keV, 1.5 Å)• Electron Charge 10 pC, Centroid Energy 6

GeV, peak current 3 kA with profile as shown in previous slides– S2E file down to undulator entrance

• LCLS Undulator with larger gap lw = 3 cm (1.5 cm); beta-function ~ 15 m

Juhao Wu


FEL performance

Juhao Wu

1.5 angstrom6 angstrom


BC parameters summary


Possible test at NLCTA


Motivation and simulation condition

Motivationo Demonstrate effective bunch compression (5 to 10 times) with x-band RF

Scheme 1: use normal chicane + positive RF chirp (current NLCTA) Scheme 2: use optics w/ higher order dispersion + positive/negative RF

chirp (need to install 4/6 sextupoles in the big chicane)o Investigate tolerances on timing jitter, misalignment etc.; emittance growth

Simulation condition: In Elegant, including transverse and longitudinal wake, coherent synchrotron

radiation (CSR), longitudinal space charge (LSC) and velocity bunching0.5 million macro-particlesFor scheme 1, current operating opticsFor scheme 2, new optics20 pC beam at 5MeV, 0.5ps RMS bunch length, 5e-3 RMS energy spread, 1

m.mrad transverse emittanceBeam energy: 60 MeV at BC1, 120 MeV at BC2


NLCTA optics (current operation)

R56 =-73mm T566 = 111mm R56 =-10mm

T566 = 15mm


Scheme 1 (1) L phase, current and bunch length

Initial

Linac1

BC1

Linac2

BC2


Scheme 1 (2) no compression, on crest

Initial

Linac1

BC1

Linac2

BC2


Scheme 1 (3) 2 stage compress 20 times, end


Scheme 1 (4) effect of timing jitter, near full compression

Timing jitter between laser and RF (assumed same for two RF sections)

On phase + 115 fs (0.5 degree) - 115 fs


Scheme 1 (5) effect of timing jitter, under compression

Timing jitter between laser and RF (assumed same for two RF sections)

On phase + 115 fs (0.5 degree) - 115 fs


Scheme 2 (1) optics

Chicane w/ quadrupole+sextupole, R56 tunable, T566 tunable

6 meters long

Install 4/6 sextupoles in the big chicane


Scheme 2 (2) L phase and current


Scheme 2 (3) 1 stage compress 10 times, end


Scheme 2 (4) Sensitivity to timing jitter

Deviation between analytical formulae and simulation due to:Small difference of beam(RF) parameters being employedCollective effects in simulation


Thank you for your patience!

I would like to thank the following people for their great help and useful discussions:

C. Adolphsen , K. Bane, A. Chao, Y. Cai, Y. Ding, J. England, P. Emma, Z. Huang, C. Limborg, Y. Jiao, Y. Nosochkov, T. Raubenheimer, M. Woodley, W. Wan, J. Wu


Current less sensitive to RF phase jitter

20pC, 80 micron

L =6 m L =9 mrf = -25°

L =330 mrf = -41°

L =550 mrf = -10°

BC-1L =6 m

R56= -36 mm

BC-2L =22 m

R56= -25 mm DL-2L =66 mR56 = 0

DL-1L =12 mR56 0

undulatorL =120 m

6 MeVz 0.83 mm 0.1 %

150 MeVz 0.83 mm 0.10 %

250 MeVz 0.19 mm 1.8 %

4.3 GeVz 0.022 mm 0.76 %

13.6 GeVz 0.022 mm 0.01 %

...existing linac

L0

rfgun

L3L1 X

LhL =0.6 mrf=-160

L2

LCLS

L = 16 mrf = -40°

L = 72 mrf = -40°

L 850 mrf = 0°

BC-2L 14 m

R56= -36 mm

BC-3L 18 m

R56= -11 mm

undulatorL =? m

6 MeVz 2.0 mm 0.1 %

120 MeVz 0.5 mm 2.0 %

375 MeVz 0.1 mm 1.4 %

1.64 GeVz 0.020 mm

0.5 %

20.5 GeVz 0.020 mm 0.01 %

L3L0

LhL =1.4 mrf= -191

rfgun C L1

BC-1L 4 m

R56= -76 mm

L = 8 mrf -22°

L2

TESLA-XFEL

(2003 parameters)

Paul Emma

Energy change + optics (dispersion) (2)Emittance & trajectory (slice)

For sufficiently large slice number, one can assume same energy change in one slice

Change slice emittanceChange slice trajectory

Other terms

CSR energy change + phase rotation (smear)Emittance & trajectory (slice)

For over-compress, CSR-process can be treated as an integral process, with continuing bunch compression (lengthening).

Change slice trajectory & emittance

Negligible

Documents

A bunch compressor design and several X-band FELs Yipeng Sun, ARD/SLAC 2011-04-13, LCLS-II meeting