Inverse Free Electron Laser
accelerators
for
5th generation light sources
P. Musumeci
UCLA Department of Physics and
Astronomy
Catalina Island, Oct 2nd, 2010
Outline
• Inverse Free Electron Lasers
• Outlook of Past and Future IFEL experiments
• Design of a compact laser accelerator suitable
as injector for an advanced light source
• Control of beam longitudinal phase space at
optical scale: The Linear pre-buncher.
• An example of IFEL-driven FEL
• Conclusions
IFEL history
• Palmer. Journal of Applied Physics. 43, 3014, 1972. Interaction of
relativistic particles and free electromagnetic waves in presence of a
static helical magnet.
• Courant, Pellegrini, Zakowicz. Physical Review A, 32, 2813, 1985
High energy Inverse Free Electron laser Accelerator.
• IFEL for many years has been considered a form of pay-back to the
High Energy Physics community for the fundamental contribution
from HEP-driven accelerator research to the development of FreeElectron Lasers.
IFEL Interaction
In an IFEL the electron beam absorbs
energy from a radiation field.
In an FEL energy in the e-beam is
transferred to a radiation field
High power laser
u 
K 2 
1

r,n 


2
2
2 n 

Undulator magnetic field to couple high power radiation
with relativistic electrons
 r2
w  K 2 

 1 

2 
2 
eE 0
Kl 
mc 2 k
K
eB
mck w
Significant energy exchange between the
particles and the wave happens when the
resonance condition is satisfied.
Why you don’t want to hear
anymore about IFELs
•
Complicate experiment. Difficult requirements on laser and
magnet technology.
•
Synchrotron losses at high energy. NOT feasible for HEP multi-
TeV machines.
•
Gradient is energy dependent. Ion linac-like dynamic.
•
Dwarfed by successes of laser/plasma and beam/plasma
schemes.
Why IFELs (again…)?
•
IFEL scales ideally well for mid-high energy range (50 MeV – up to few GeV) due to
– high power laser wavelengths available (10 um, 1 um, 800 nm)
– permanent magnet undulator technology (cm periods)
•
Simulations show high energy/ high quality beams with gradients >500 MeV/m achievable
with current technology!
– 70 MeV/m gradient already demonstrated at UCLA
– 70 % trapping already demonstrated at BNL.
– Preservation of injected e-beam quality/emittance. (Essentially 1D acceleration)
•
Microbunching: still the preferred interaction for longitudinal phase space manipulation at
optical scale.
•
Efficient mechanism to transfer energy from laser to electrons
•
Anybody interested in a compact 1-2 GeV injector?
– Laser-plasma accelerators. Main competitors. But….
o
o
o
o
Need > 40-50 TW laser power to accelerate beams to 1 GeV.
Strongly non-linear injection mechanism.
Controlled injection ?
Beam quality ??
– Injector + (phase-locking) microbuncher for other kinds of advanced accelerators
– Injector for advanced light sources (ICS or FELs)
STELLA2 experiment
80 % of electrons accelerated,
energy spread less than 0.5 % FWHM
~30 GW@  = 10.6 mm,
gain up to 17 % of initial beam energy
W. Kimura et al. First demonstration of high
trapping efficiency and narrow energy spread
in a laser accelerator,
PRL, 92, 154801 (2004)
Diffraction dominated IFEL @ UCLA
•
–
–
–
–
15 MeV beam accelerated to over
35 MeV in 25 cm
Relative energy gain 150 %
Accelerating gradient ~70 MeV/m !
Observation of higher harmonic
IFEL interaction
2000
1600
200
150
1200
a.u.
•
•
IFEL Advanced Accelerator
at the Neptune Laboratory
0.5 TW 10.6 mm laser
Strongly tapered Kurchatov
undulator
Highest recorded IFEL
acceleration
a.u.
•
800
100
50
0
24
26
P. Musumeci et al.,High energy gain of
trapped electrons in a tapered diffractiondominated IFEL PRL, 94, 154801 (2005)
28
30
Energy (MeV)
400
0
10
15
20
25
30
Electron energy (MeV)
35
32
34
Inverse Free Electron Laser:
lessons learned
• Even though radiation guiding would help, significant gain can be
obtained controlling the diffraction effects
• Strong tapering of both period and field is possible.
• Prebunching helps beam quality.
• There is no laser wavelength preference intrinsic in the IFEL equations
– NIR lasers advantages
• Commercial high power sources available
• Table-top-sized laser systems.
• Mitigated diffraction effects
Inverse Free Electron Laser experiments
Energy gain (MeV)
NEPTU
10
STELL
MIFEL
A2
5
10
A1
BNL-IF
1
STELL
EL
A
IFELA
0.1
1
10
2
10
3
10
4
10
10
Radiation power (MW)
6
NE
Current IFEL projects
Most of them UCLA-centric
Microbunching experiment at Neptune (7th
harmonic)
Helical bunching experiment at Neptune (again
harmonic coupling, interesting beam modes)
Permanent magnet helical undulator development.
Praseodymium based cryogenic undulator.
Prebunching at 800 nm at SLAC
High repetition rate IFEL experiment
at LLNL
High gradient helical IFEL
experiment at BNL
Proposal for experiment at
SPARC-LIFE(Italy)
LLNL -IFEL
Short laser pulse IFEL
100 fs
• Gradient profile of undulator + 800
nm light requires > 3TW laser (4-5
TW preferred)
• Laser system is CPA, flashlamp
pumped, Ti:Sapphire
– 100 fs fiber oscillator
– >500 mJ, <120 fs, 10 Hz
– 100 mJ UV arm for photocathode
• Undulator has 19 periods; requires
~50 fs slippage of on-resonance
particles
– Significant laser intensity
variation over interaction length!
Laser Electric
Field
• 3D simulation of
IFEL.
• Captured bunch
is ~ 100 fsec.
• Short laser pulse
results in tail in
energy
distribution.
Current Status: Laser and
experimental layout are under
construction
50 cm UCLA undulator
Chicane couples in IFEL
drive laser and allows
compression of blow-out
mode electron bunch.
Spectrometer and
diagnostic beamline
Quad triplets match
into undulator
Laser entrance port;
not shown is vacuum
transport line from
compressor
50 MeV beam from
LLNL photo-gun/linac
Radiabeam Ucla BNL-IFEL
COllaboratioN: RUBICON
The experiment main goal is to achieve energy gain and gradient significantly
larger than what possible with conventional RF accelerators to propose IFEL
as a viable technology for mid-high energy range accelerators.
This can be achieved
using the existing ATF ebeam and high power
CO2 laser system
TOGETHER WITH
Helical geometry.
Permanent magnet
double tapered undulator.
Parameter
Initial e-beam energy
Laser wavelength
Laser peak power
Nominal length of wiggler, Lw
Rayleigh range
Laser focal spot size (w)
Location of laser waist inside wiggler
Undulator length
Fixed Value
50 MeV
10 um
0.5 TW
60 cm
9 cm
550 um
30 cm
60 cm
Table 1. Parameters for BNL high gradient high energy gain IFEL experiment
Interaction with
circularly polarized laser
is always ON
Factor ~2.3 extra gradient for
same electric field.
Planar undulator
Helical undulator
4
2
0
0.000
50.0µm
x
Electron transverse velocity is
never zero.
Energy gain
Helical interaction
0.005
0.010
0.015
0.020
0.025
0.030
0.005
0.010
0.015
0.020
0.025
0.030
0.0m
-50.0µm
0.000
Distance along the undulator (m)
Planar

JJ ( K ) K
 kK l
 sin( )
z
2

vs.
Helical

K
 kK l  sin( )
z

Optimized undulator tapering design
o Use regular NdFeB magnets. Br = 1.22 T
o Take into account not ideal laser transverse profile M2 = 1.5
o Provide large enough gap (15 mm) to minimize laser losses
o >98 % transmission to allow for recirculating schemes.
Mechanical design finalized
Particle trajectory
- Magnets ordered
- Machining started
RUBICON to demonstrate IFEL Recirculation
• IFEL does not need to wait for any plasma recombination time-scale.
– Laser power can be recirculated to increase average power and wall-plug
efficiency !!!
• A 22-m reamplification loop will carry 6 pulses (12 ns apart), to
achieve RUBICON goal of pulse train IFEL acceleration.
IFEL undulator
F1
F3
F3
IP
F2
2*F3
Beam loading and phase front evolution
Next step in IFEL simulations !!!
Amplifier:
100 cm x 2 passes
ZnSe
window
IFEL efficiency
• Beam loading or pump depletion effects for high accelerated beam
charge ( 1 nC @ 1GeV = 1 J of energy ).
• Modified Genesis version + script to take into account varying period.
• Simulate radiation
(and particle) IFEL
dynamics with
GENESIS 1.3
Power along the undulator
Power profile along the bunch for max current
• Energy extraction very
efficient (> 80%)
adjusting tapering to
compensate for peak
power variation along
the undulator.
1 GeV IFEL design:
• If successful, these experiments (LLNL+ BNL) will pave
the way for
Application of IFEL scheme
as 5th generation light source driver
• Compact-size accelerator
• ESASE (Zholents, PRL 92, 224801, 2004) benefits intrinsic
– Exponential gain length reduction due to peak current increase.
– Absolute timing synchronization with external laser optical phase
at attosecond level.
– Control of FEL radiation pulse envelope.
• Need control of output energy spread !!!
• Valid competitor for first Advanced Accelerator driven/ 5th
generation light radiation source. See talk @ FEL 2006,
Berlin.
Useful scalings for IFEL accelerator
Assuming no guiding and a single stage helical undulator
The ideal relationship between the Rayleigh range and the total undulator length is
Lu  6 zr
A tight focus increases the intensity, but only in one spot.
A large zr maximizes the gradient over the entire undulator length
The final energy (assuming a constant K and a constant resonant phase)
will be given by
d r2 kKKl e0 Z 0
  12 2sin( r 2)
dz

mc
zr

4
P K  6 10
zr

P(W ) K
In order to have the final energy 1 GeV (f2 = 106) with a 1 um laser, zr = 20 cm
and K ~ 4
The laser power P needs to be 10 TW or higher
Praesodymium based cryogenic
undulator
Cryogenic undulator + 10 TW laser power “green-field” design
Helical undulator to maximize energy exchange
(interaction always ON).
Initial energy
50 MeV
Final energy
1000 MeV
Fully permanent magnet design (no iron poles)
Avg gradient
640 MV/m
Final energy spread
<0.1 %
Laser wavelength
800 nm
Laser power
10 TW
Laser spot size (w0)
0.3 mm
Keep magnetic field amplitude well under the Halbach
limit for 6 mm gap to ensure technical feasibility.
1 GeV goal with minimum laser power to get to the softx-ray region.
3
20
110
Energy (MeV)
16.667
13.333
10
6.667
3.333
800
600
400
200
0
0
0.25
0.5
0.75
1
1.25
Distance along the undulator
Period (cm)
Field (kGaus s)
1.5
0
0
0.25
0.5
0.75
1
1.25
Distance along the undulator (m)
1.5
IFEL longitudinal phase space
Tapering optimization
•
The undulator period and magnetic field amplitude are changed trying to control
the resonant phase of acceleration and the longitudinal phase space parameters.
From KMR, IEEE. J. Quantum Electronics, 1981
•
Compromise between stability (low resonant phase) and gradient (high resonant
pahse). Varying phase along the undulator.
Improvements in gradient, energy spread and peak current !
•
For r -> p/2
-> 0
E vs. phase spa ce
Longi tudina l pha sespa ce
Relative energy spread
0.1
3
2 10
z2ave  1 .48
.50 01
04 84
0.05
max ( pz2)  1 66
95 3.4
2.3 02
06
p z2
i
0
3
1 10
 0.05
 0.1
0
0
2
4
6
6
 8 10
6
66
 6 10  410
4 10
z2 z2 ave
i
6
 2 10 0
0
6
24 10
Hamiltonian of IFEL interaction
• In the longitudinal phase space (for small variation around the
design energy), the Hamiltonian of the system looks like a
physical pendulum
K 2 K l2
k 1

 2 KKl cos( )
kw 2
2
2
H ( , )   

2
(
)
This phase space flow explain
why Inverse Free Electron Laser
is a strong longitudinal lens.
From IFEL thesis, 2004
Lasers 2001
Longitudinal bunching and
aberrations
• Harmonic potential: limited by initial energy spread
• Cos-like potential: limited by non linerarities
Potential function
Phase Space
4
5
Deltap/p
2
0
2
0
a
2
0
2
asd
5
cos potential
harmonic potential
0
2
4
p hase
cos potential
harmonic potential
Lasers 2001
6
Higher Harmonic IFEL
• Higher harmonic interaction has been first observed in UCLA
experiment, and then studied in SLAC experiment.
• More recently the efficiency of the interaction has also been shown in
the Neptune 7th harmonic IFEL experiment.
• Even harmonic interaction is also strong when there is an angle
between electron and laser beams.
• New concept: Combine first harmonics to “linearize” the IFEL buncher.
6
4
2
0
First 4 harmonics
P arabolic pote ntial
cos-like potential
2
0
2
The >90 % bunching factor-buncher
• Seed with harmonics of Ti:Sa laser
• Need to control relative phase and amplitude (phase retardation
plates)
10
kKl (  0)
e0 E0
IR laser pulse
Laser energy (in 100 fs)
to bunch 120 MeV beam
800 nm
100 uJ
400 nm
130 uJ
2
m 0 c
Non linear harmonic
generation crystals
5
0
5
 10
6
7
 1.610
 810
7
810
0
1.610
6

Electric field waveform
245
266 nm
85 uJ
200 nm
50 uJ
Energy
240
Beam Trajectory (mm)
0.4
Laser propagation direction
0.2
235
0.0
-0.2
-0.4
Angle for even
harmonic-coupling
0.0
0.1
0.2
230
6
 110
0.3
0.4
Distance along the undulator (m)
7
 510
0
z (m )
7
510
6
110
Linear “perfect” IFEL pre-Buncher
•
•
•
•
•
By using a multiple-harmonic buncher one could
approximate harmonic oscillator and linearize the potential.
47
S. Pottorf and X.J. Wang, “Harmonic Inverse Free Electron
46
Laser Micro -buncher”, BNL –68013 (2000).
Not worth for “coherent radiation production” since  Vpbipart  0.51145
0
bunching factor is already 0.5-0.6.
44
Significant improvements for injection into advanced
accelerator.
43
Particle tracking simulations show >99 % capture and
below 0.1 % energy spread !!!
5
0
 Vpbipart 1
Longitudinal phasespace and energy spectrum
Energy
110
3
110
3
800
800
600
600
400
400
200
200
0
2
4
Phase
6
0
z  1 .5m
Captured fraction
99.5 %
Energy spread
0.04 %
100
energy distribution
200
300
400
5
FEL radiation from IFEL accelerator
• Sending the IFEL beam into an undulator
 FEL radiation @ = 3 nm (water window)
• Slippage dominated regime.
• Start-to-end simulations
Current peak FWHM 80 nm or 250 as
1.7 GeV energy
2
25
spike distance 800 nm
20
1.5
15
1
10
0.5
5
0
-0.2
Energy (GeV)
Current(kA)
E-spread (10^-3)
0
0.8
1.8
2.8
Distance along the bunch (mkm)
3.8 From 20TW IFEL design
Slippage
• Slippage in the undulator
Ls  Lwu   N 
• Slippage in a gain length
Lc 
Lg
w

• Different FEL dynamics (weak superradiance) when Lb~ Lc
2006 proposed solution:
Insert slippage sections between undulators
• Larger energy (because of longer period SPARC-like undulators).
• Smaller gain.
• Between undulator sections we insert a magnetic delay section for
the electron beam to realign current and radiation spikes.
• The slippage section effectively is a positive R56 region that helps the
conversion between energy modulation and bunching. Optical
Klystron
• Need to seed for longitudinal coherence
Coordinate along the undulator (m)
14
Radiation
12
e-
10
8
6
4
2
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Coordinate along the bunch (mm)
Undulator sections
Use low charge@injection approach
• Use 0.1 mm-mrad from low-charge operation of RF photoinjector.
Assuming emittance is preserved through IFEL.
• Usually get a factor of 10 enhancement from ESASE mechanism
• p/2 resonant phase + perfect linear pre-bunching give an extra
improvement in compression.
• We obtain 5 kA – 0.1 mm-mrad at 1 GeV.
IFEL-driven soft-x-ray FEL
• Efficiency can be increased by
laser recirculation.
• Option to HHG seed FEL
• GW-level peak power @ 3 nm.
• Intrinsic synchronization of
microbunch structure with optical
phase.
Strawman design
10 TW laser system
5J -500 fs
Linear
Prebuncher
RF photogun
Strongly tapered undulator
47
46.363
RF Linac
Energy
46
 ip a rtFR AME 0.511 45
44
43.607
43
0
 2 p
45 MeV
4.5 m
0.5 m
10
 ip a rtFR AME
P h ase
6 p
1 GeV
z  0
1.5 m
3.5 m
Cryogenic short-period
FEL undulator
Conclusions
• Laser accelerators have made tremendous
progress and will soon be competitive with more
conventional machines.
• IFEL accelerator among these offers control of
the beam properties.
– Radiabeam-UCLA-BNL will show high gradient helical
IFEL acceleration.
– UCLA-LLNL IFEL will show high rep-rate, good beam
quality.
• If successful, these experiments will pave the
way towards IFEL-based compact soft-x ray
radiation source.
• Ultrashort probe beams will come from a synergy
between laser and accelerator worlds.
Acknowledgements
Collaborators:
S. Anderson, LLNL
I. Pogorelsky, V. Yakimenko, BNL
A. Murokh, A. Tremaine, Radiabeam Technologies
F. O’Shea, E. Hemsink, G. Andonian, R. Li, M. Westfall, J. B.
Rosenzweig, UCLA
Funding agencies:
DTRA
DOE-HEP / DOE-BES
University of California Office of the
President