January 17, 2002 J. Rosenzweig Experimental Results and Computational Modeling of Pulse Compression and High Gain at the VISA SASE FEL (and related topics) CSR Workshop - Zeuthen January, 17 2002 James Rosenzweig UCLA Department of Physics and Astronomy
Experimental Results and Computational Modeling of Pulse Compression and High Gain at the VISA SASE FEL (and related topics). James Rosenzweig UCLA Department of Physics and Astronomy. CSR Workshop - Zeuthen January, 17 2002. Acknowledgments. - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
January 17, 2002 J. Rosenzweig
Experimental Results and Computational Modeling of Pulse Compression and High Gain
at the VISA SASE FEL (and related topics)
CSR Workshop - Zeuthen January, 17 2002
James RosenzweigUCLA Department of Physics and Astronomy
January 17, 2002 J. Rosenzweig
AcknowledgmentsAcknowledgments
• VISA is a large collaboration (BNL, LLNL, SLAC, UCLA). C. Pellegrini (UCLA) is spokesman
• UCLA was lead on experimental data-taking and analysis� Aaron Tremaine (post-doc, ex-UCLA student). Experiment.� Alex Murokh (student). Experiment. � Ron Agusstson (student). ELEGANT simulation (originally for ATF
compressor expt.)� Sven Reiche (post-doc). GENESIS� JBR (Expt. diagnosis; simulations), CP (theory)� Stealth collaborators in expt/data analysis: P. Emma, H-D. Nuhn
• Extremely difficult experiment to perform and understand.
January 17, 2002 J. Rosenzweig
VISA BeamlineVISA Beamline
• Gun and Linac Section (1.6 cell photo-emission gun and 2 SLAC type linac structures operating at S-Band, generate 71 MeV beam)
• 20° double-bend dispersive transport section
• Beamline III, with VISA matching optics and 4-m strong focusing undulator (K=1.26)
Linac Sections
Gun
Transport
20° Dispersive SectionMatching Line
Undulator
January 17, 2002 J. Rosenzweig
Measurements on Electron Beam at Linac ExitMeasurements on Electron Beam at Linac Exit
• Emittance was measured with the quad scan after the linac. For a typical charge of 200-500 pC emittance was optimized at
n 1.3 2.7 mn 1.3 2.7 m
• The beam current in the linac was measured by applying a linear chirp to the beam and measuring its profile after 20° bend. With wake-field correction the current value is found:
Ip 55 AmpIp 55 Amp
January 17, 2002 J. Rosenzweig
Beam Profile Monitors
8 diagnostic ports
e-beam
vacuum chamber
actuator
OTR to CCD
FEL light out
Undulator DiagnosticsUndulator Diagnostics
FEL Optical Diagnostics
Beam Profile Monitors
January 17, 2002 J. Rosenzweig
Tune “Optimization”Tune “Optimization”
• Initially an FEL radiation pulse energy was measured ~ 1-10 nJ, in accord with the measured beam brightness.
-300.0000
-200.0000
-100.0000
0.0000
100.0000
200.0000
300.0000
400.0000
0 5 10 15 20 25
x-beta [m]
dispersion [cm]
y-beta [m]
old tune-300.0000
-200.0000
-100.0000
0.0000
100.0000
200.0000
300.0000
400.0000
0 5 10 15 20 25
x-beta [m]
dispersion [cm]
y-beta [m]
new tune
• With the new tune the FEL radiation intensity went up to ~ 10 µJ. Why?
• In the attempt to compensate for the dispersion, a new tune was developed:
January 17, 2002 J. Rosenzweig
Saturation and Physical ModelSaturation and Physical Model
• With the high gain an FEL saturation in 3.6 m was observed:
• How does the gain length measurement agree with the high gain SASE-FEL theory? Not that well if we believe beam parameters at linac exit…
M. Xie numerical model: 18.7 cm gain length at 140 pC charge and 50 Amp peak current corresponds to the sliced emittance of <0.35 mm-mrad
Lg = 18.7 cm
January 17, 2002 J. Rosenzweig
(old tune) (new tune)
More inconsistencies in the data More inconsistencies in the data • Highest gain observed after changing rf phase of linac• Change of the tune significantly altered all SASE radiation properties, indicating changes
of basic electron beam properties:
800 820 840 860 880 900
G ~ 103
many spikesspike width ~ 0.1% centered at 830 nm
800 820 840 860 880 900
G ~ 107
single spikespike width ~ 1% centered at 845 nm
low gain stable condition high gainvery unstable, 100% fluctuations
January 17, 2002 J. Rosenzweig
4/30/01: Bunch compression hypothesis4/30/01: Bunch compression hypothesis• High gain observed for running ~4-5 degrees forward of crest; horizontal beam size expands inside of
undulator• Strong bunch compression in the dispersive section was suggested, due to mistuning of linac energy from the
nominal value. Effective R56 can change sign, order of magnitude due to T566, off energy operation.
• Increase in peak current reduces FEL gain length, explains the observed spectral behavior (watch for growth due to dispersion mismatch…)
• Longitudinal transformation highly nonlinear• Measure compression in final VISA runs!
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
-0.02 -0.01 0 0.01 0.02
R56
dp/p
z f zi R56p
p
1
2T566
p
p
2
z f zi R56p
p
1
2T566
p
p
2
T566 R56
p p 7 m rad2T566 R56
p p 7 m rad2
R56 p0 0.15 m
January 17, 2002 J. Rosenzweig
Single Golay Cell Experimental Set-upSingle Golay Cell Experimental Set-up
System allows following measurements:1. Scanning linac RF phase and observe CTR signal (test for a possible
bunch compression).2. Inserting a remote controlled low-pass filter for a quantitative
measure of a bunch length when compared to PARMELA/ELEGANT model.
e-beam
Golay cell
90° polarizer
parabolic mirrors
FEL light
Wavenumber [cm-1]
Tra
nsm
issi
on
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 20 40 60
40 µm
detector window cut-off
100 µm
400 µm
removable low-pass filter
filter transmission
January 17, 2002 J. Rosenzweig
Using Collimator to Map Linac RF PhaseUsing Collimator to Map Linac RF Phase
• To understand the nature of the compression, one has to keep a track of the linac RF phase jitter.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
0 5 10 15 20 25
x [mm]
• It was found that the bending dipole to ATF Beamline 1 acts as a scraper with the 1.5 cm aperture.
• Charge loss at the scraper depends on the beam energy and is very sensitive to changes in the RF phase.
Measuring the charge loss at the collimator allows to1. Calibrate the linac RF phase shot-by-shot.2. Use the same system operating point for FEL measurements.
1.5 cm aperture
January 17, 2002 J. Rosenzweig
Results of the Measurements Results of the Measurements
0
0.005
0.01
0.015
0.02
0.05 0.1 0.15 0.2 0.25 0.3
Linac RF phase span of 2°
Charge [nC]
Goa
ly S
igna
l [m
V]
• Initial test indicated strong CTR signal dependence on linac RF phase.
• CTR measurement of subpicosecond bunches� Emittance Diagnostic
• Current increase at what cost?• Beam physics in the compressor: phase space monitoringBeam physics in the compressor: phase space monitoring
January 17, 2002 J. Rosenzweig
The Neptune CompressorThe Neptune Compressor
2'-4"
Edge provideshorizontal focusing
(and steering)22.5º bend angles
• Horizontally focusing edge angles fore and aft• Mitigate vertical focusing, no cross-over in chicane
January 17, 2002 J. Rosenzweig
CTR interferometry for pulse lengthCTR interferometry for pulse length
• Data gives filtered filtered autocorrelation of the temporal beam profile. Need to take into account “missing” long wavelengths
• Short beam, some ancillary structures
• Near resolution limit0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
10 12 14 16 18 20 22
Autcorrelation Data
Nor
mal
ized
Sig
nal
Delay [psec]
t = 0.63 ps
Interferogram for shortest pulse length
January 17, 2002 J. Rosenzweig
Emittance Growth in the CompressorEmittance Growth in the Compressor
• The compressor, pulse length, and emittance diagnostics allow us to examine the issue of emittance growth in bends.
• In particular, the slit based measurement permits us to view the the evolution of the transverse phase space as the emittance increases.
• Experimental Procedure:� Set bend angle to design value of 22.5°, keep R56 constant� Measure linac phase and pulse length, map compression� Vary phase and measure emittance
January 17, 2002 J. Rosenzweig
Emittance Versus Linac PhaseEmittance Versus Linac Phase
5
10
15
20
25
55 60 65 70 75 80 85 90
Norm
aliz
ed
Em
itta
nce [
mm
mra
d]
Linac Phase [deg]
Sharp increase is a consistent feature in
data
Maximum compression
January 17, 2002 J. Rosenzweig
Phase space reconstruction shows bifurcationPhase space reconstruction shows bifurcation
January 17, 2002 J. Rosenzweig
Simulation of experimentSimulation of experiment
• Different codes model different processes (acceleration fields versus velocity fields.)
• Codes employed:� TREDI: Full story, but noisy..� PARMELA: Provides input distributions for TREDI. Point-to-point
space charge for comparison, no acceleration fields. Noisy.� ELEGANT: only acceleration fields, approximate.� Heuristic calculation of space-charge between longitudinal slices.
• Initial simulations indicate that for this experiment, acceleration fields do not contribute much emittance growth, the space charge fields are the dominant effect.
January 17, 2002 J. Rosenzweig
Simulation resultsSimulation results
0
5
10
15
20
25
55 60 65 70 75 80 85 90
DataPARMELATREDI
Em
itta
nce
[mm
mra
d]
Linac Phase [deg.]
• Simulation is difficult. Number of macro-particles is low because of time-intensive space-charge calculations.
• Sharp emittance increase when “fold over” begins is missing in simulations.
• Improve existing tools, use heuristic model
January 17, 2002 J. Rosenzweig
Heuristic analysisHeuristic analysis
• To analyze the effect of space-charge in the compressor, we model the beam as a series of longitudinal slices.
• Since the beam energy spread is heavily correlated to slice position, we assume that there is no energy spread within a single slice
• Space-charge forces push a slice based on the fields at its centroid due to the other slices.
• Use standard envelope equations to evolve the sizes of single slices.
January 17, 2002 J. Rosenzweig
Evolution Without Space-chargeEvolution Without Space-charge
Configuration Space Long. Phase Space
Beam “folds over” in configuration space.
January 17, 2002 J. Rosenzweig
Effect of space-charge in the modelEffect of space-charge in the model
0
5
10
15
20
25
260 280 300 320 340 360 380 400 420
Parmela Simulation
No
rma
liz
ed
Em
itta
nc
e [
mm
mra
d]
Z [cm]
0
5 10 -6
1 10 -5
1.5 10 -5
2 10 -5
2.5 10 -5
3 10 -5
-0.2 0 0.2 0.4 0.6 0.8 1 1.2
Slice Model
No
rma
liz
ed
Em
itta
nc
e [
m r
ad
]
Z [m]
• Slices repel strongly in (and after) the last magnet
• This destroys the dispersion cancellation at the compressor exit ( & ’ 0)
• Space-charge + dispersion grows emittance after the compressor as well
January 17, 2002 J. Rosenzweig
Slice Model SimulationSlice Model Simulation
• With space-charge beam “fold over” is not perfect as seen in configuration space.
• In phase space, this shows up as a bifurcation• We see evidence for a two-peak initial longitudinal profile. Presently adding to
this to all simulations, expect enhanced bifurcation
-5
-4
-3
-2
-1
0
1
2
3
-1 -0.5 0 0.5
Trace Space
X' [
mra
d]
X [mm]
-1
-0.5
0
0.5
-1.5 -1 -0.5 0 0.5
Configuration Space
X [m
m]
Z [mm]
January 17, 2002 J. Rosenzweig
Summary and conclusionsSummary and conclusions
• Proper understanding of compression and beam performance requires large effort in diagnosis and simulation — in tandem
• At 70 MeV, ELEGANT/GENESIS combination very robust� “Pathological” running conditions at VISA explained
• Some verification of CSR importance at VISA• Computational tools are developing to meet experimental
demands• The more details of beam 6D phase space revealed, the
better� FEL is excellent “phase space diagnostic”� Phase space tomography is high energy analogue of slits
• CSR spectrum should also be very useful• High brightness beams have a wealth of applications, equal