Accelerator Issues and Design Paul Emma, SLAC Dec' 12, 2003

1
Accelerator Issues and DesignPaul Emma,
SLACDec. 12, 2003

• Design of Compression and Acceleration Systems
• Technical Challenges
• Full System Simulations

2
Slice versus Projected Emittance

For a collider
collision integrates over bunch length
projected emittance is important
e- slips back w.r.t. photons by lr ( 1.5 Å) per
period
FEL integrates over slippage length slice
emittance is important
3
SASE X-ray FEL is very sensitive to electron
slice emittance
lr 1.5 Å
courtesy S. Reiche
Instead of mild luminosity loss, power nearly
switches OFF. However, longer wavelength, such
as 15 Å (4.5 GeV), is much easier (eN ? 6 mm).
4
Nominal System Design

• 1.5-Å SASE FEL Linac
• Requirements
• Acceleration to 14.1 GeV (3 GeV min.)
• Bunch compression to 3.4 kA
• Emittance preservation (lt20 slice of
  1-mm-mrad)
• Final energy spread (0.01 slice, lt0.1
  projected)
• Minimal sensitivity to system jitter (charge,
  phase, voltage, ...)
• Diagnostics integrated into optics
• Flexible operations (1.5 Å ?15 Å, low-charge,
  chirp, etc.)

use 2 compressors, 3 linacs
5
Nominal System Design

• Constraints
• Use existing SLAC linac compatible with PEP-II
  operation
• Undulator located beyond research yard

6
LCLS versus SLC

• LCLS Advantages
• Shorter linac (1 km lt 3 km)
• Shorter bunch in linac (1 mm ? 0.2 mm ? 0.02 mm)
• Lower charge (1 nC lt 7 nC)
• Slice emittance important, not projected
• No positrons, no sextupoles, no rolls, no DRs,
  no RTLs, no arcs
• Round beams (no x-y coupling issues)
• Disadvantages
• Lower initial linac energy (135 MeV lt 1.2 GeV)
• Smaller emittance (1/1 mm lt 4/40 mm)
• Emittance more critical (gt2 mm kills FEL power)
• Tighter RF, charge, timing jitter tols (0.1
  deg)
• CSR is new issue
• RF gun less stable platform than damping ring

7
Design Strategy

• Design longitudinal optics first
• Set proper compression in two stages
• Minimize final energy spread
• Minimize Ipk and Ef sensitivity to gun charge and
  timing jitter
• Design transverse optics second
• Minimize transverse wakefields, CSR, and
  chromatic effects
• Build in emittance, energy spread, bunch-length
  diagnostics
• Track entire system
• Iterate design

8
Nominal LCLS Linac Parameters for 1.5-Å FEL
Single bunch, 1-nC charge, 1.2-mm slice
emittance, 120-Hz repetition rate
6 MeV ?z ? 0.83 mm ?? ? 0.05
250 MeV ?z ? 0.19 mm ?? ? 1.6
4.54 GeV ?z ? 0.022 mm ?? ? 0.71
14.1 GeV ?z ? 0.022 mm ?? ? 0.01
135 MeV ?z ? 0.83 mm ?? ? 0.10
Linac-X L 0.6 m ?rf -160?
rf gun
Linac-1 L ?9 m ?rf ? -25
Linac-2 L ?330 m ?rf ? -41
Linac-3 L ?550 m ?rf ? -10
new
Linac-0 L 6 m
undulator L 125 m
21-1b 21-1d
X
21-3b 24-6d
25-1a 30-8c
...existing linac
BC-1 L ?6 m R56? -39 mm
BC-2 L ?22 m R56? -25 mm
DL-1 L ?12 m R56 ?0
LTU L 275 m R56 ? 0
SLAC linac tunnel
research yard
(RF phase frf 0 at accelerating crest)
9
RMS Bunch Length and Energy Spread
sector-21
sector-25
sector-30
FFTB
sd
sz
10
time profile
energy profile
phase space
sz 830 mm
sz 190 mm
sz 830 mm
sz 23 mm
sz 830 mm
sz 23 mm
FINAL
sz 190 mm
sz 23 mm
11
X-band RF used to Linearize Compression (f
11.424 GHz)
S-band RF curvature and 2nd-order momentum
compaction cause sharp peak current spike
12
Transverse Wakefields and Component Misalignments
Choose b-phase adv/cell for each linac to
minimize emittance dilution
L2 phase adv/cell optimized
L3 phase adv/cell optimized
sz 22 mm
sz 195 mm
x
also misaligned quads/BPMs generate dispersion ?
De
wakes on
wakes off
wakes on
wakes off
13
Transverse Optics from Cathode to e- Dump
LCLS MAD Deck ? Cathode to e- Dump (2200
elements)
Dyx,y ? 30º
Dyx,y ? 75º
Thanks to M. Woodley
14
RMS Transverse Beam Sizes from Cathode to e- Dump
1 mm
100 mm
undulator
10 mm
15
Alignment and Roll Tolerances (most gt 1 mm, gt 1
deg)
16
Linac RF Section Modifications
If modulators on 20-6, -7, and -8 used for
injector, lose another 670 MeV (1.56 GeV total)
17
Injector to Linac Interface
courtesy L. Bentson
Linac Responsibility Starts Here (21-1b)
18
Linac-1 Through BC1
21-1b
21-1c
21-1d
21-3b
19
BC2 Area
24-6d
25-1a
20
Moveable Chicanes (BC1 shown)
BPM critical for energy feedback (20 mm
resolution)
offset 17 to 30 cm (24 cm nominal)
collimator
BPM
quadrupole
screen
21
Field quality requirement too tight with fixed
chicane...

• Also needed
• BPM res. 20 mm
• BPM linearity
• profile monitor
• collimator

• requires b2/b0 lt 0.002 _at_ r 2 cm (moveable
  chicane requires 0.070)

SPPS dipoles b2/b0 lt 0.010 _at_ 2 cm (just
barely met)
22
Future Multiple Undulators
4º
2º
N
S
-2º
23
Linac-To-Undulator (LTU)
vertical bends
energy centroid spread meas. (3?10-5 10-4)
collimation

• vertical bend 4.7 mrad
• horizontal jog 1.25 m
• energy diagnostics
• emittance diagnostics
• collimators
• CSR cancellation
• branch points for future undulators
• spontaneous undulator possible

4 e-wires, 6 collimators
24
Collimation for Undulator Protection
?2.5 mm
well shadowed in x, y, and E
25
Electron Dump
x-rays ?
quads
soft bend
e- ?
permanent vert. bends
powered vert. bends
e- ?
quad
screen (sE/E 10-5 ? 5 mm)
dump
26
Specification Sheets on Every New Magnet

• BX01 DL1 dipole
• z-location
• field
• current
• trim info.
• alignment tol.s
• length
• max/min strength
• etc...

27
Technical Challenges

• Coherent Synchrotron Radiation in Bends
• projected emittance growth
• micro-bunching instability ( LSC see Z. Huang
  talk)

• Emittance Preservation in Linacs
• transverse wakefields
• misalignments chromaticity

• Machine Stability
• gun and rf system jitter
• energy and bunch length feedback

28
Coherent Synchrotron Radiation
coherent power
N ? 6?109
l-1/3
incoherent power
vacuum chamber cutoff
sz
29
Coherent Synchrotron Radiation (CSR)

• Powerful radiation generates energy spread in
  bends
• Induced energy spread breaks achromatic system
• Causes bend-plane emittance growth (short bunch
  is worse)

coherent radiation for l gt sz
bend-plane emittance growth
sz
l
L0
DE/E 0
s
Dx
e
R
DE/E lt 0
?
overtaking length L0 ? (24szR2)1/3
Dx R16(s)DE/E
30
Coherent Synchrotron Radiation (CSR) in SPPS
Chicane ON
Chicane OFF
gex 34.2 ? 0.7 mm
gex 27.6 ? 0.6 mm
31
Coherent Synchrotron Radiation (CSR) in SPPS
Bend-plane emittance is consistent with
calculations and sets upper limit on CSR effect
32
CSR Micro-bunching
S. Heifets, S. Krinsky, G. Stupakov,
SLAC-PUB-9165, March 2002
CSR amplifies small modulations on bunch current
? Successive bend-systems cause micro-bunching ?
Growth of slice-energy spread emittance.
without heater
sd ? 3?10-6
avoid!
First observed by M. Borland (ANL) in LCLS
Elegant tracking
33
Misalignments, Steering, and Emittance Correction
trajectory after steering
BPM, quad, and RF misalignments (each at 300 mm
rms)... then steered in Elegant
gex ? 5 mm gey ? 2 mm
34
Emittance Correction with Trajectory Bumps
100 seeds
steering coils
De/e ? 15
gex ? 1.02 mm gey ? 1.09 mm
Thanks to M. Borland (ANL/APS)
35
Jitter Budget (lt1 minute time-scale)
measured RF performance
klystron phase rms ? 0.07 (20 sec)
klystron ampl. rms ? 0.06 (60 sec)
36
Start-to-End Tracking Simulations

• Track entire machine to evaluate beam brightness
  FEL
• Track machine many times with jitter to test
  stability budget (M. Borland, ANL)

Parmela
Elegant
Genesis
space-charge
compression, wakes, CSR,
SASE FEL with wakes
37
Sliced e- Beam to Evaluate FEL (Dz ? 0.7 mm)
After full system tracking (also studied by S.
Reiche)
gex
gey
zx
zy
Lg lt 4 m
38
Machine Stability Simulations

• Track LCLS 230 times with Parmela? Elegant?
  Genesis
• Include wakes, CSR, etc. jitter budget (M.
  Borland, ANL)

39
Emittance and Energy Spread Diagnostics

• 5 energy spread meas. stations (optimized for
  small bx)

• 5 emittance meas. stations designed into optics
  (Dyx,y)

• slice measurements possible with transverse RF
  (L0 L3)

sE
sE
sE
sE
sE
see also P. Krejcik talk
40
Transverse RF deflector as diagnostic
41
Summary

• Linac design optimized for nominal 1.5-Å
  operation
• Design is flexible to accommodate 15-Å,
  low-charge, chirp
• CSR growth of projected emittance not slice
• Much experience on SLAC linac with wakefield
  control
• Beam diagnostics built into design
• Full system tracking to
• Evaluate e- brightness preservation,
• Calculate SASE gain,
• Simulate pulse-to-pulse stability.

Full tracking with errors shows FEL saturation at
1.5 Å, but a very challenging machine!