Transcript TW FEL “Death

TW FEL “Death-Ray“ Studies
Josef Frisch, Zhirong Huang, Yi Jiao
FEL Optimized for Imaging




Many imaging experiments are similar
 Hard X-rays ~1 Angstrom, minimum focus size,
maximum flux.
 Liquid or gas target injection
 Pump laser
 CCD arrays surrounding sample
Use a special purpose imaging chamber designed for rapid
change of samples
FEL optimized for maximum flux at 1 Angstrom, ~1030fsec
Conversations with experimenters suggest that 2TW in
10fs is required for imaging of single bio-molecules
Self Seeding Improves FEL Efficiency
Seeding
undulator
Diamond crystal
band-stop filter
Gain
Undulator
Energy
extraction
taper
Narrow band seed power
Band stop filter produces
narrow band signal in tail
Minimal mirrors
(maybe just KB set)
Chicane delays beam so that it
interacts with tail of filtered pulse
A “notch” filter produces a narrow
band peak in the transmitted
spectrum
Idea invented by
E. Saldin at DESY
High Energy Operation - Estimates


Vary electron voltage while keeping operating wavelength fixed at 1
Angstrom.
Assume that peak current increases as the 4th root of beam energy.




Undulator wavelength changes, keep same peak B field as LCLS undulator so
"K" changes
Roughly matches what we can use in LCLS_I, is probably conservative
Assume 0.6um slice emittance and 1MeV energy spread (before
spontaneous energy spread is added).
Run Ming Xie formula with spontaneous energy spread (Saldin et al,
“Design Formulas for VUV and X-RAY FELs”) calculated for 20 gain
lengths (iterate).

Allow Beta function to optimize from 10 to 100 Meters
No Taper
~60GW at 28 GeV
0.4 emittance, higher
peak current, more
agressive parameters
90GW at 28 GeV
Usable to 45 GeV
500eV FEL can use high
beam energy as well !
~25 GeV beam with
existing LCLS_I
undulator should work
at 25 KeV
TW FEL Genesis Simulations by Zhirong
Huang
27 GeV, 0.6um Emittance, 5KA peak current, 1.4 MeV espread,
4.5cm Undulator, K=4.95, 30M beta function
4TW (!!)
700M undulator(!!)
NOT OPTIMAL!
Need Simulations
• The simulation above is based on a guess at beam
parameters from Ming-Xi calculations. It may be VERY far
from optimal
• 700 M undulator is huge – but not completely insane
– Undulators are $200K/M, this is $140M for an undulator, LCLS_II
is $400M, XFEL is >$1B
• E-beam parameters are reasonable, but R&D might improve
emittance or peak current.
• Energies up to 45 GeV in principal available from the SLAC
linac. (probably can’t use that much energy)
• High K helical undulators (Superconducting) may be possible.
• Want to explore the design phase space.
Simulations
•
Use Genesis in CW beam mode to evaluate the performance of various
accelerator and undulator configurations
– Slice simulation takes ~15 seconds on a PC – can afford to run a LOT of
simulations.
•
•
•
•
•
Use various electron beam energies, emittance, peak current.
Use linear and helical undulators, set for 1 angstrom output wavelength
with maximum practical K at each undulator wavelength.
Optimize output power at various undulator lengths (each 100M) for beta
function and K (2 parameters).
Note: need to apply physics to constrain parameters and to aid
optimization.
Create a large “library” of simulations to explore the phase space
– >106 simulation possible – probably 104 optimized conditions.
•
Generate an empirical model or software tool to predict the performance
of an FEL give input e-beam parameters (energy, emittance, peak current),
and undulator parameters (length, max B-field for PM, or SC).
Automatically optimizing
 Varying one input parameter while keep
others constant
Generate the corresponding input file for
Genesis steady-state simulation
Read the Genesis output, record interested
quantities, e.g. power, transverse radiation
size, etc.
Find out the optimal input parameter
All these jobs are integrated into a MATLAB
code.
Scan sx, sy, at the undulator entrance and quadrupole gradient
for a higher radiation power
2.7
2.8
2.6
2.5
2.2
2.4
Pmax (TW)
2.4
2
1.8
2.3
2.2
1.6
2.1
1.4
2
0
10
20
30
40
50
60
 (m)
x
1.9
0
10
20
30
40
50
60
 y (m)
3
Before scanning : sx,= 11.6 mm, sy
= 9.1 mm, K = 26.7 T/m, power =
2.63 TW
After scanning : sx,= 9.3 mm, sy =
9.3 mm, K = 22.4 T/m, power =
2.83 TW
2.5
2
Pmax (TW)
Pmax (TW)
2.6
1.5
1
0.5
0
0
5
10
15
20
25
30
Quad gradient (T/m)
35
40
45
50
Other Issues
•
Transverse mode quality may degrade for long tapered undulators
– Juhao is working on this?
•
Full 3-d effects may be important
– Will need full 3-d genesis runs for some interesting cases
•
Do we trust Genesis in this regime?
– Need to cross-compare against other FEL simulations
•
Experiments
– Want to compare prediction with the HXRSS seeding experiment at LCLS late this
year.
– Need results soon enough to predict, not post-dict!
•
Need to do this quickly!
– SLAC is designing LCLS_II, working on Sector 0 test facility.
– Will soon determine what parts of the LINAC are available in the future
– Knowing possible future uses of the LINAC and undulator hall may affect decisions
now!
– Around the world $5B of FEL projects proposed or under construction – these
results might affect their decisions.
Benefit to SLAC



Imaging experiments are an important application of
FELs.
Provides a unique capability at SLAC: No other lab will
build a 30 GeV accelerator for an FEL.
Not made obsolete by XFEL or any other proposed FEL
project
LCLS Publicity image shows single
molecule imaging, but this is NOT
POSSIBLE with present day FELs!