Constructing a HighGain FEL:

1
Constructing a High-Gain FEL How to Make it Work
P. Emma SLAC August 25, 2009
assigned title
2
Many Accelerator/Undulator Challenges

• RF Photocathode Gun
• Need ?1-µm normalized emittance all day long
• ?10-ps initial bunch length with up to 1 nC
  charge
• Emittance Preservation in LINAC
• Transverse wakefields (misalignments)
• Dispersion and chromatic effects
• Bunch Compression
• Coherent synchrotron radiation (CSR)
• Micro-bunching instability
• Machine Stability
• RF phase, voltage, gun-timing, charge
• Pointing stability lt10 of beam size
• Long Undulator
• Need few micron straight trajectory
• Field tolerances, absolute alignment
• Longitudinal wakefields (small, optically smooth
  chamber)

3
Focus of this Presentation

• Will not talk about physics or FEL design.
• Many here know the physics better.
• Will talk about the practical issues of starting
  up such a machine, and preparing for this.
• Many others here have similar or more experience
  (DESY, Spring-8, Frascati, ANL, BNL, etc).
• This is from one point of view (mine at LCLS),
  and does not cover all issues (e.g., x-ray
  diagnostics and transport).

4
7 Key Points

• 0. Good machine design (big issue - not addressed
  here)
• Detailed start-to-end tracking simulations from
  cathode through undulator, with as much physics
  as possible.
• Tolerances studied and tuning strategies well
  developed before first beam.
• Diagnostics well integrated into machine design.
• High-level applications ready on day 1 (fast,
  precise, automated measurement and tuning e, sz,
  BBA, ).
• Pre-Beam measurement and checkout of all
  components, including controls with local
  (tunnel) verification.
• Beam-based feedback loops (hold upstream systems
  steady while commissioning downstream systems).
• Adequate budget and time to do it right.

5
1. Detailed start-to-end tracking simulations
from cathode through undulator, with as much
physics as possible
6
Start with 2D Longitudinal Phase Space
1. Start-to-End Simulations
after L2
energy profile
phase space
time profile
after DL1
sz 830 mm
sz 190 mm
after L1
after BC2
sz 830 mm
sz 23 mm
after X-RF
after L3
sz 830 mm
sz 23 mm
after BC1
at und.
sz 190 mm
sz 23 mm
7
Micro-bunching Identified During LCLS Simulations
1. Start-to-End Simulations
heater damps bunching
Effect discovered by M. Borland (ANL) using his
Elegant tracking code
8
1. Start-to-End Simulations
Resistive-Wall Wake in Undulator
0.2 nC
1 nC
Cylindrical, Copper, r 2.5 mm Bane/Stupakov
AC-wake model
9
Simulations Suggest Lower Bunch Charge
1. Start-to-End Simulations

• 1 nC ? 0.25 nC
• Smaller gun emittance (1.1 µm ? 0.5 µm)
• Reduced wakefields in undulator
• Better control of transverse wakes in linac
• Mitigate CSR in compressors
• Less micro-bunching
• FEL still produces up to 4 mJ/pulse (50 GW)

10
Longitudinal Jitter Simulations (Generate full
list of stability requirements for drive laser
and RF)
1. Start-to-End Simulations
Lg
0.09
0.004
energy
energy spread
96 fs
bunch length
timing
RF phase jitter lt 0.1º
Pout
10
peak current
pulse length
11
2. Tolerances and tuning strategies developed
before first beam
12
Transverse Wakefields May Destroy Beam Brightness
2. Tolerances and Tuning
HEAD
sz 0.02 mm
x'/sx'
sz 0.20 mm
sz 2.00 mm
Wakes Reduced with Shorter Bunch
N 3?1010, b 30 m, L 3 m, eN 1 mm, E0 1
GeV, Dx 1 mm
e-
misaligned RF accelerating structure
13
Steering in RF Structures (X-band here)
2. Tolerances and Tuning
11.4 GHz, 3.7-mm iris
X-emittance after X-band Cavity (µm)
Beam position in X-band Cavity (mm)
14
Simulations with Random Component Misalignments
2. Tolerances and Tuning
trajectory after steering
trajectory corrected
large emittance growth
Elegant by M. Borland
15
Emittance Correction with Trajectory Variations
2. Tolerances and Tuning
steering coils
optimize x and y emittances with two x- and two
y-steering coils
e meas.
16
Tweaker quads allow critical chicane dispersion
correction
2. Tolerances and Tuning
and same at LCLS
with two quads correct h and (ah bh?),
orthogonally
17
Bipolar Tuning Quad in BC1 Chicane
2. Tolerances and Tuning
X emittance after BC1 Chicane (µm)
Quadrupole Integrated Gradient (kG)
18
2. Tolerances and Tuning
Sources of Transverse Beam Jitter
19
2. Tolerances and Tuning
Undulator Alignment Strategy

• Beam-based alignment using large energy changes
  (4 to 14 GeV)
• Quadrupole magnets are well aligned to dnstr. end
  of undulator (CMM)
• Beam-finder wires used to align upstream end to
  beam

calculate
Beam Finder Wire
Undulator
14 GeV
7 GeV
4 GeV
move
Quads and BPMs are not aligned well enough
initially
Vary electron energy and record BPM readings (BBA)
Undulator ends already precisely aligned to
quadrupoles
Quadrupole and BPM position corrections now
applied
Beam-finder wire needed to align upstream end to
beam
Full system now aligned after 3 iterations
H.-D. Nuhn, et al
20
3. Diagnostics built into machine design
21
Electron Beam Diagnostics
3. Diagnostics
gun
4 wire scanners 8 colls
4 wire scanners 6 colls
TCAV0
3 wires 2 OTR
3 OTR
vert. dump
heater
stopper
m wall
sz1
sz2
3 wires 3 OTR
BC2 4.3 GeV
14 GeV
TCAV3 5.0 GeV
BC1 250 MeV
135 MeV
14 GeV
undulator 14 GeV
stopper

• 2 Transverse RF cavities (135 MeV 5 GeV)
• 7 YAG screens (at E ?135 MeV)
• 12 OTR screens at E ? 135 MeV
• 15 wire scanners (each with x y wires)
• 4 beam phase monitors (2805 MHz)
• CSR/CER pyroelectric bunch length monitors at BC1
  BC2
• Gun spectrometer line, injector spectrometer
  line, BC1 stopper
• 200 BPMs and toroids

22
Many Emittance Measurement Options
3. Diagnostics

• OTR screens and wire-scanners both available
• Multi-screen or quad-scan with beam waist at
  screen
• Quad scan (invasive), multi-wire is non-invasive
• Transverse RF for time-resolved measurements
• Measurement stations in each critical section

OTR screen
23
3. Diagnostics
Wire Scanner Emittance Measurements Near End of
Linac
sy 49 mm
sy 41 mm
sy 57 mm
sy 35 mm
Individual wire-scanner profiles
x (b0g - 2a0a g0b)/2
24
Wire Scanner Vibration
3. Diagnostics
sy 97 mm
sy 45 mm
Scan before addition of 10? reducer gear
Scan after addition of 10? reducer gear
J. Frisch
25
OTR Screens Compromised by Coherent
Radiation(especially after bunch compression)
3. Diagnostics
Bright beam becomes a bigger problem?
Coherent Transition, Diffraction, Synchrotron,
and Edge Radiation corrupt image
Image is unusable
26
Transverse RF Deflectors for Time-Resolved
Measurements
3. Diagnostics
off-axis screen
single-shot, absolute bunch length measurement

• Deflector used to measure
• absolute bunch length,
• time-sliced emittance,
• time-sliced energy spread,
• electron arrival time jitter

27
Measuring Bunch Arrival Time Jitter
3. Diagnostics
e-
S-band (2856 MHz)
BPM
V(t)
Q 0.25 nC
slope -2.34 mm/deg
Now measure BPM jitter both with transverse RF
OFF, and then ON (at constant phase)
Timing Jitter (w.r.t. RF) (110 mm)/(2.34
mm/deg) 0.047 deg ? 46 fsec rms
28
Estimating X-ray Pulse Energy with Well Placed
BPMs
3. Diagnostics
10 MeV/e- (2.4 mJ)
initial DEi
? 100 meters ?
Vertical Bend
DE DEf - DEi
final DEf
vary FEL power with oscillations record e-
energy loss
29
4. High-level physics applications ready on day
1 (fast, precise, automated measurement and
tuning)
30
4. High-Level Applications
Correlation Plot GUI in Matlab
H. Loos
31
4. High-Level Applications
Laser Heater Spatial Alignment
IR
Calculate and re-align laser
poor heating?
e-
good heating
H. Loos
32
4. High-Level Applications
Fast, Precise RF Phasing with Beam
e.g., calibrate gun voltage
vary RF phase read BPMs
33
5. Pre-Beam measurement and checkout of all
components, including controls with local
(tunnel) verification
34
5. Measurement and Checkout
Component Measurements (e.g., RF Gun)
D. Dowell, et al
Race-track cavity design
quadrupole component of RF
35
Undulator Measurements
5. Measurement and Checkout

• Undulator K measured and recorded over 6 mm x
  range

Z. Wolf, Y. Levashov, H.-D. Nuhn, et al
36
5. Measurement and Checkout
All Beamline Components Get Local Checkout
100 page checkout lists
Check all component locations, and
magnet polarities
Check using real controls as well
collimator travel
wire scanners
BPM cables orientations
etc.
37
6. Beam-based feedback loops (hold upstream
systems steady while commissioning downstream
systems)
38
6. Feedback
The Importance of Feedback Loops(beam-based and
other)

• The main function of feedback is usually thought
  of as jitter reduction
• But there are more important functions
• De-coupling of machine sections, minimizing (for
  example) steering in the undulator during a
  quad-scan in the injector
• De-coupling of effects, holding (for example) the
  beam position constant at a wire-scanner while
  running a quad-scan
• Machine recovery after a maintenance day is much
  more automatic
• Long term drift control (termperature and
  day/night especially)

39
Feedback Loop Examples in the LCLS

• 9 electron trajectory loops (BPMs and steering
  coils)
• Drive laser pointing
• Gun launch angle
• Injector trajectory
• X-band RF structure X Y position
• Post-BC1 launch
• Post-BC2 launch
• End-of-linac launch
• Linac-to-undulator trajectory
• Undulator launch
• Bunch charge loop (toroid and laser waveplate)
• 6x6 longitudinal loop (energy at 4 locations
  peak current at each compressor very important)
• Many RF-based phase and amplitude loops (set
  points are then adjusted with beam-based loops)

Feedbacks run 24/7 enabling other work by holding
FEL steady
40
6. Feedback
Feedback Systems - Bunch Length Energy (66)
J. Frisch
CSR-based bunch length monitor
41
Transverse Stability of LCLS Injector
6. Feedback
one of 50 jittering trajectories
50 shots at 10 Hz (250 MeV, after BC1)
1-s beam size
Q 0.25 nC
Ax 3.9 rms
Ay 3.4 rms
42
Transverse Stability of LCLS Linac
6. Feedback
one of 50 jittering x-trajectories
one of 50 jittering y-trajectories
50 shots at 10 Hz (14 GeV, near end of linac)
Q 0.25 nC
1-s beam size
Ax 14.2 rms (needs work)
Ay 9.5 rms
goal lt10
goal lt10
43
7. Budget and Time to Do It Right
44
7. Budget and Time
Adequate Budget and Time

• LCLS was a 420 M project
• 17 years of design, simulations, and optimization
  (not full time)
• 4 years of construction
• Hundreds of people involved
• 2.5 years of commissioning
• and with an existing linac

45
A Few More Points to Make

• A good electron gun solves many problems
• Soon after planning how to complete some machine
  setup task, consider how to automate that process
• Be prepared to spend 80 of commissioning time on
  controls (amazing but true)!
• Then get ready for the next 15... safety system
  interlock testing (5 allowed for FEL physics).

46
Finally, this slide was shown many times over the
years
Finally, these worries have been relieved
SASE FEL is not forgiving instead of mild
luminosity loss, power nearly switches OFF
electron beam must meet brightness requirements
47
Thanks for your attention