Monitoring and Manipulation of

Sub-picosecond Beams

J.B. Rosenzweig
UCLA Department of Physics and Astronomy
PAC 2001, Chicago, June 18, 2001
 Outline

• Applications of sub-picosecond beams

• Sub-picosecond beam generation
• Sub-picosecond beam manipulation
• Sub-picosecond beam diagnosis
• The future...
Applications of Sub-picosecond Beams:
Advanced Accelerators
Laser-based Accelerators Wake-field Accelerators
Very high field implies short wavelength, use of Beam used as power source for
lasers, plasmas high frequency, high impedance, accelerator
Generally produced in pulse trains Need high harmonic content in current:
short bunches/modulated/ramped beam
Relatively small charge/bunch (~ ), emittance
High charge (>nC), relaxed emittance
Few micron length, fsec time scales.. Sub-mm length, 100 fsec time scales..
LEAP expt.
at Stanford 1

0.5

W (arb. units)
0

-0.5

W I
-1

0 2 4 6 8 10 12
kz/2

Ramped beam, gives transformer

ratio R>>2 in wakefield accelerator
Applications of Sub-picosecond Beams:
Sub-picosecond Radiation Sources

X-ray SASE Free-Electron Lasers Thomson Scattering Sources

A fast growing beam-radiation instability; Ultra-short from collision of short pulse
needs cold, dense “lasing media” lasers and electron beams; needs “luminosity”
High energy (>10 GeV), low energy spread Low energy! (~20-200 MeV)
Very high current (>kA) Very high current (>kA)
Excellent emittance (~ mm-mrad) Focusability, good emittance

Short pulse (<100 fsec) beams Short pulse (<100 fsec) beams, sub-psec timing

laser beam

Needs highest brightness 2I

Be = 2
beams for gain... n electron beam
scattered x-rays
 Production of Beams from High-field
RF Photoinjectors
• Prompt photo-emission with ultra-short
lasers allows very short e-beam production
• Injection of laser pulse to ~30 fsec level
– Is cathode prompt enough?
– Longitudinal beam dynamics?
• Emittance “compensation” depends on
acceleration rate γ’/γ being similar to 1.6 cell high gradient BNL/SLAC/UCLA RF gun

transverse plasma frequency k p = 2 e 3b

4 rn Space-charge defocusing

– Implies longitudinal plasma motion (debunching)

– Partially controlled with rf focusing (BNL)
RF focusing
z
– Limit on achievable beam density, brightness
– Need also to compress!
Forward RF wave
 Example Advanced Accelerator
Laboratory: Neptune at UCLA

• Photoinjector: 1.6 cell rf gun, post-acceleration (PWT) linac

• Inline chicane compressor
• S-bend compression beamline
• Experimental program in Plasma/Beam/Laser-base acceleration concepts
– Very short beam time structures, and external synchronization

Beam experimental area

Chicane Negative R56 S-bend

PWT linac compressor compression line

to RF gun
Laser experimental area
 Electron Beam Pulse Shaping using
Photocathode Drive laser
• Optimized shapes for beam
applications can be “programmed”
with laser

1.5
• Witness beams (AWA), ramps?

1
Wakefield (MV/m)
• Pulse trains

0.5
0
– Below rf wavelength (microbunching)

-0.5
– Integer multiples of rf wavelength

-1
• Photoelectron dynamics can distort

-1.5

20

40

60

80
the pulse profiles

z (cm)
Simulated bunch train from AWA photoinjector,
with excited wake-fields (courtesy [Link], ANL)
– Cathode delay
– RF focusing
– Space charge, wake-fields
Pulse Structure Progamming
Example: Beatwave injection
Scheme proposal by
CO2 Laser Filip, et al. (UCLA EE)
Electro-Optical
Switch

Drive Laser

Plasma Cell Accelerator

Electron beam and beatwave are phase locked at cathode.

Are they phase-locked at the Pasma BWA experiment?
Pulse Structure Preservation ( );

∂ϕ ∂
f
=
1
ϕ
Defeating RF Focusing

o
Pulse longitudinal phase space at Phase space after linac (note Phase space after chicane,
cathode; current projection positive chirp, back of crest) pulse structure restored
 Chicane Compression

• Negative chirp applied to beam

(forward of crest in linac). Positive
R56 then compresses beam
• Mapping (to first order) all initial
phases to the same final phase, ,

∂ϕ ∂
f
=
0
(like parallel-to-point optics)

ϕ
o
• Final current profile limited by rf UCLA Compressor chicane for BNL ATF expts.

curvature, or dynamical nonlinearities

• Final current has sharp rise, long tail
• Space charge and wakes degrade
compression
 Phase and configuration space
pictures of chicane compression

Pulse longitudinal phase space at Phase space after linac Phase space after chicane,
cathode; current projection (negative chirp) linear correlation removed

Ramped beam current,

but wrong orientation for
wake excitation
 Time domain measurement of
compressed pulses: Streak camera
• Resolution limited to 0.5 psec (M. Uesaka, U. Tokyo)
• Compressed beam at A0 has 1.6 kA in 2 psec rms, beam stopped due to
wake-fields in 8 cm in plasma! (N. Barov talk)
• Non-streak dimension useful for time dependent imaging (AWA, E-157)
5 1.4 100
x1 / 2 exp. intensity exp.

Integrated intensity (arb. units)

1.2 x1 / 2 sim. Intensity sim.
4 80
Experiment
Parmela 1
3 60
(mm) 0.8
[mm]
z

x1 / 2
σ

2 0.6
Maximum Energy 40

0.4
1
20
0.2

0
0 0
-110 -100 -90 -80 -70 -60 -50 -40 0 10 20 30 40 50 60 70 80
Phase [Degree] Time (ps)
Compression of 8nC bunches at Results from time-dependent imaging in
FNAL A0 Photoinjector (OTR streak data) plasma wakefield experiment at AWA
 Higher resolution measurements:
Time domain autocorrelation of CTR
• Coherent transition radiation (CTR)
pulse follows beam current
– Long wavelengths not produced efficiently
(see Reiche poster)
• Autocorrelate in Martin-Puplett
interferometer
– Multipulse scanning measurement
– Time-domain fitting
• Neptune chicane experiments:
– Bunch compressed to 0.5 psec
• Compression accompanied by
transverse emittance growth…
 Emittance growth in bends
• Transverse phase space can be
strongly affected by compression
• Critical in X-ray SASE FEL design
• Self-fields cause dispersion errors,
other phase space correlations
• “Velocity” fields
• “Acceleration fields
– Centrifugal space-charge
– Coherent synchrotron radiation (CSR)
• Brightness limitations implied;
trade-off between and I Comparison of CSR-induced energy
spectrum in experiments at CLIC
• Experiments at CLIC, Neptune, etc. (from Corsini, et al, Linac2K)
Transverse phase space post chicane
compression: Neptune experiments
Increased compression

Images after slits Phase space bifurcation

Phase space reconstructions See talk by S. Anderson later

 Velocity Bunching:
a Cure for “The Bends”?

• Proposed by Serafini, Ferrario (see

PAC paper); tool for SASE FEL
injector, avoids magnetic compression
• Inject emittance-compensated beam
at 5-7 MeV into slow-wave linac
• Perform one-quarter of synchrotron
oscillation to compress beam
• Similar to manipulations in thermionic
injector bunchers, but with high
phase space density Longitudinal phase space
schematic for velocity bunching
 Velocity bunching optimization
and implementation
• Long slow-wave structure
from tuning ω (new source)
or k (structure)?
• Alternative: use only short
bunching section to split
functions of bunching and
Simulation of INFN test facility case
acceleration 0.2

• Studies: 0.15

– INFN FEL injector test facility

(mm)

(slow-wave system)
0.1
z
σ

– ORION (S-band buncher with 0.05

S-band RF gun

X-band linacs) 0
0 1 2 3 4 5 6 7
S-band PWT Buncher z (m) X-band travelling wave linacs

Simulation of ORION system

 Alternative to Chicane:
Negative R56 Compression

Accelerator Test Facility (ATF) at

Brookhaven National Laboratory

• Opposite chirp, final phase space orientation from

chicane case. Cprrect ramped beam possible
• Space-charge and wakefields enhance compression
• Most common configuration: S-bend (Neptune)
• Much more sensitive to errors than chicane
• Impact on BNL compressor, VISA experiment
• ORION design study to mitigate this problem(?)
Experience with the VISA Experiment
at BNL ATF
• 19 cm measured gain length (see PAC contributions by A. Tremaine and A. Murokh)
• Current is low at end of linac - brightness needs to be enhanced to explain results
• Spectrum also indicates shorter FEL pulse (single spike…)
• Compression inadequate with design R56

Energy G ~ 1000
[nJ] March 15, 2001
10 6

10 4

100 800 820 840 860 880

0.01 G ~ 106
Saturation at 3.6 m
0.0001
0 1 2 3 G = 2.24 x 107
z [m] Lg = 18.7 cm!
Q = 165 pC 800 820 840 860 880
 ELEGANT simulation of beamline 3 (to VISA)

Large projected x
emittance

Beam core has

emittances 0.5 x 6
mm-mrad, 300 A

Strong compression,
with positive R56! 100
fsec beam core… Dispersion not correctly
cancelled, R56 large,
w/ wrong sign!
 Frequency domain CTR
measurements of single pulses
• Simpler, single shot tool needed for experiments
• Single shot spectra of incoherent radiation can be
used (Catravas, et al.)
• Fourier content of coherent pulse ~ follows beam
• Reconstruction? Kramers-Kronig…
• Multichannel polychromator developed by
Tohoku/Tokyo (Uesaka, et al.)
• Single channel + low pass filter at VISA
1000
1

U (microJ)
U (microJ) with Filter 0.8

100
0.6
U (µ J)

Ratio

0.4
10

0.2 Tohoku/Tokyo polychromator

measurements, benchmarked to 0.5
1 0
0 50 100 150 200 250 0 50 100 150 200 250 psec res. streak camera
σ (µm) σ (µm)
 IR CTR Diagnostics for Modulated
Beam Pulse: SASE FEL Case
• UCLA/LANL SASE FEL experiment 250

CTR (normalized) ,SASE (mV)

• Ultra-high gain (3x105) - a few gain- 200

lengths from saturation… CTR

SASE
150

• Coherent transition radiation (CTR)

from microbunching (1%) gives narrow 100

band spectrum, similabr to SASE 50

spectrum
0

• Multiple shot monochromator (signal 10 11 12 13

Wavelength ( m)
14

level, detector considerations) SASE FEL and CTR spectra (Tremaine, et al.)

•Also investigated, in use at STELLA IFEL experiment (next talk), full bunching
•Intensive use at LEUTL SASE FEL experiment (saturation = full bunching)
•Harmonics give important information on highly bunched systems
Far IR CTR Diagnostics for Modulated
Beam Injection: Neptune Case
• How do we know we have recovered the initial beat-wave derived
current profile? Interferometry not possible due to low repetition rate.

• Single shot spectra can be obtained using a 10 channel polychromator,

collaboration with Uesaka (U of Tokyo) and Shibata (Tohoku Univ.).

Fourier Transform of Modulated Beam Longitudinal Enlargement of Beat-Wave Peak with Simulated 10ch
Structure Polychromator Signal
1

Beat-Wave 1/λ

0.3
= 33.3 cm-1

Intensity (a.u.)
Intensity (a.u.)
0.8 0.6 0.4 0.2

0.25 2 0.15 1 0.05

0. 0.

25

30

35

40

45
20

40

60

80

Macropulse peak 1/λ (cm-1) 1/λ (cm-1)

Microbunch peak
 Creation of compressed, ramped
beams for high XFMR ratio PWFA
6

5
(a)
4
Optimum profile

I (kA)
3
p / p (a.u.)
2

0
0.8 1 1.2 1.4 1.6 1.8
z (mm)
z (a.u.)

Phase space Current profile Wakes in 1E16/cc plasma,

from PIC simulation
• ORION study shown
• Negative R56 gives correct ramp
• How to measure time domain features? Sub-100 fsec
fall time, etc.
 RF Sweeping: A Beam-based
Streak Camera

• Longitudinal position dispersed by

RF deflection near zero crossing Detection
Drift

• Resolution limited by emittance and RF deflector

space-charge at low energy

(Carlsten, Russell, LANL) • Alternatives in detection
• Proposed at high energy (X. Wang), system
– Spot imaging (ramp beam…)
also for LCLS (100 fsec beam) slice
– Phase space (quad scan)
diagnostics (10 fsec!)
– Spectrometer
Future directions: high brightness
particles sources from plasmas
• Ultra-short beams from plasma background
• Phase-locked to plasma accelerating wave
• Laser and electron beam-excited systems
• Simulations show good phase-space control
• Extension to ions?

Self-modulated laser wake-field data courtesy D. Umstadter, UMich

1 deg cone
Injected beam
From Suk, et al,., PRL 2001,
also see contributions from Suk
and Thomspon et al at PAC.
 Conclusions

• Ultra-short pulse generation, manipulation and

measurement are critical to advanced accelerator
and FEL development
• Concepts theoretically, computationally mature
Initial experiments giving valuable feedback on
conceptual view, and future instrumentation
• Some of what were once experiments are now tools
• Next few years should yield exciting results