Inverse Free Electron Laser accelerators

for5th 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 Free-Electron Lasers.

IFEL Interaction

Undulator magnetic field to couple high power radiation with relativistic electrons

wmckeB

K kmc

eEK l 2

02

2 12 2w

r

K

Significant energy exchange between the particles and the wave happens when the resonance condition is satisfied.

In an FEL energy in the e-beam is transferred to a radiation field

In an IFEL the electron beam absorbs energy from a radiation field.

r,n u

22n1K2

2

High power laser

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 Need > 40-50 TW laser power to accelerate beams to 1 GeV. o Strongly non-linear injection mechanism. o Controlled injection ?o Beam quality ??

– Injector + (phase-locking) microbuncher for other kinds of advanced accelerators

– Injector for advanced light sources (ICS or FELs)

STELLA2 experiment

W. Kimura et al. First demonstration of high trapping efficiency and narrow energy spread in a laser accelerator,PRL, 92, 154801 (2004)

80 % of electrons accelerated, energy spread less than 0.5 % FWHM~30 GW@ l = 10.6 mm, gain up to 17 % of initial beam energy

Diffraction dominated IFEL @ UCLA

• IFEL Advanced Accelerator at the Neptune Laboratory

• 0.5 TW 10.6 m laser• Strongly tapered Kurchatov

undulator• Highest recorded IFEL

acceleration– 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

• IFEL Advanced Accelerator at the Neptune Laboratory

• 0.5 TW 10.6 m laser• Strongly tapered Kurchatov

undulator• Highest recorded IFEL

acceleration– 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

10 15 20 25 30 350

400

800

1200

1600

2000

a.u.

Electron energy (MeV)

24 26 28 30 32 340

50

100

150

200

a.u.

Energy (MeV)

P. Musumeci et al.,High energy gain of trapped electrons in a tapered diffraction-dominated IFEL PRL, 94, 154801 (2005)

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

101 102 103 104 105 1060.1

1

10

En

erg

y g

ain

(M

eV

)

Radiation power (MW)

Inverse Free Electron Laser experiments

Current IFEL projectsMost 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• 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 photo-

cathode• Undulator has 19 periods; requires

~50 fs slippage of on-resonance particles– Significant laser intensity

variation over interaction length!

Laser Electric Field

100 fs

• 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 MeV beam from LLNL photo-gun/linac

Chicane couples in IFEL drive laser and allows compression of blow-out mode electron bunch.

Quad triplets match into undulator

50 cm UCLA undulator

Spectrometer and diagnostic beamline

Laser entrance port; not shown is vacuum transport line from compressor

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 e-beam and high power CO2 laser system

TOGETHER WITH

Helical geometry.

Permanent magnet double tapered undulator.

Table 1. Parameters for BNL high gradient high energy gain IFEL experiment

Parameter Fixed Value

Initial e-beam energy 50 MeV

Laser wavelength 10 um

Laser peak power 0.5 TW

Nominal length of wiggler, Lw 60 cm

Rayleigh range 9 cm

Laser focal spot size (w) 550 um

Location of laser waist inside wiggler 30 cm

Undulator length 60 cm

Helical interactionElectron transverse velocity is never zero.

Interaction with circularly polarized laser is always ON

Factor ~2.3 extra gradient for same electric field.

)sin(2

)(

KKJJkK

z l sin( )l

KkK

z

vs.Planar Helical

0.000 0.005 0.010 0.015 0.020 0.025 0.030-50.0µm

0.0m

50.0µm

0.000 0.005 0.010 0.015 0.020 0.025 0.0300

2

4

xDistance along the undulator (m)

En

erg

y g

ain

Planar undulator Helical undulator

Optimized undulator tapering designo Use regular NdFeB magnets. Br = 1.22 T

o Take into account not ideal laser transverse profile M2 = 1.5o Provide large enough gap (15 mm) to minimize laser losses

o >98 % transmission to allow for recirculating schemes.

Particle trajectoryMechanical design finalized- Magnets ordered- Machining started

RUBICON to demonstrate IFEL Recirculation

IP

F1

F2

F3 F3

2*F3 Amplifier:100 cm x 2 passes

ZnSewindow

• 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

Beam loading and phase front evolution Next step in IFEL simulations !!!

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.

• Energy extraction very efficient (> 80%) adjusting tapering to compensate for peak power variation along the undulator.

• Simulate radiation (and particle) IFEL dynamics with GENESIS 1.3

Power profile along the bunch for max current

Power 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

ru zL 6

In order to have the final energy 1 GeV (gf2 = 106) with a 1 um laser, zr = 20 cm

and K ~ 4

KWPz

KPz

mc

Zerr )(106212 4

2002

The final energy (assuming a constant K and a constant resonant phase)will be given by

The laser power P needs to be 10 TW or higher

A tight focus increases the intensity, but only in one spot.A large zr maximizes the gradient over the entire undulator length

)sin( rlr kKK

dz

d

Praesodymium based cryogenic undulator

Cryogenic undulator + 10 TW laser power “green-field” design

Helical undulator to maximize energy exchange (interaction always ON).

Fully permanent magnet design (no iron poles)

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 soft-x-ray region.

Initial energy 50 MeV

Final energy 1000 MeV

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

0 0.25 0.5 0.75 1 1.25 1.50

3.333

6.667

10

13.333

16.667

20

Period (cm)Field (kGauss)

Distance along the undulator

0 0.25 0.5 0.75 1 1.25 1.50

200

400

600

800

1 103

Distance along the undulator (m)

Ene

rgy

(MeV

)

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.

• 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 yr -> p/2-> 0

E vs. phase space Longitudinal phasespace

8 106 6 10

6 4 106 2 10

6 0 2 106

0

1 103

2 103

pz2i

z2i

z2ave

0 2 4 60.1

0.05

0

0.05

0.1

Rel

ativ

e en

ergy

spr

ead z2ave 1.48018

max pz2( ) 1663.402

E vs. phase space Longitudinal phasespace

8 106 4 10

6 0 4 106

0

1 103

2 103

pz2i

z2i

z2ave

0 2 4 60.1

0.05

0

0.05

0.1

Rel

ativ

e en

ergy

spr

ead z2ave 1.50044

max pz2( ) 1952.306

From KMR, IEEE. J. Quantum Electronics, 1981

From IFEL thesis, 2004Lasers 2001

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

2

)cos(222

1),(

)( 22

2l

lw

KKKK

kkH

This phase space flow explain why Inverse Free Electron Laser is a strong longitudinal lens.

Lasers 2001

Longitudinal bunching and aberrations

• Harmonic potential: limited by initial energy spread

• Cos-like potential: limited by non linerarities

0 2 4 65

0

5

cos potentialharmonic potential

Phase Space

phase

Del

tap/

pa

2 0 22

0

2

4

cos potentialharmonic potential

Potential function

asd

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.

2 0 2

0

2

4

6

First 4 harmonicsParabolic potentialcos-like potential

• Seed with harmonics of Ti:Sa laser• Need to control relative phase and amplitude (phase retardation

plates)

Laser energy (in 100 fs) to bunch 120 MeV beam

800 nm 100 uJ

400 nm 130 uJ

266 nm 85 uJ

200 nm 50 uJ

Electric field waveform

0.0 0.1 0.2 0.3 0.4

-0.4

-0.2

0.0

0.2

0.4

Bea

m T

raje

ctor

y (m

m)

Distance along the undulator (m)

Laser propagation direction

Non linear harmonic generation crystals

IR laser pulse

The >90 % bunching factor-buncher

Angle for even harmonic-coupling

1.6 106 8 10

7 0 8 107 1.6 10

610

5

0

5

10

kKl 0( )

e0 E0

m0 c2

1 106 5 10

7 0 5 107 1 10

6230

235

240

245

z (m)

Ene

rgy

• By using a multiple-harmonic buncher one could approximate harmonic oscillator and linearize the potential.

• S. Pottorf and X.J. Wang, “Harmonic Inverse Free Electron Laser Micro -buncher”, BNL –68013 (2000).

• Not worth for “coherent radiation production” since bunching factor is already 0.5-0.6.

• Significant improvements for injection into advanced accelerator.

• Particle tracking simulations show >99 % capture and below 0.1 % energy spread !!!

Linear “perfect” IFEL pre-Buncher

Captured fraction 99.5 %

Energy spread 0.04 %

Longitudinal phasespace and energy spectrum

0 2 4 6

200

400

600

800

1 103

Phase

Ene

rgy

0 100 200 300 400

200

400

600

800

1 103

energy distribution z 1.5m

5 0 543

44

45

46

47

Vpbipart 0

0.511

Vpbipart 1

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

0

0.5

1

1.5

2

-0.2 0.8 1.8 2.8 3.8

Distance along the bunch (mkm)

0

5

10

15

20

25

Energy (GeV)

Current(kA)

E-spread (10^-3)

spike distance 800 nm

Current peak FWHM 80 nm or 250 as1.7 GeV energy

From 20TW IFEL design

Slippage • Slippage in the undulator

• Slippage in a gain length

• Different FEL dynamics (weak superradiance) when

Lb~ Lc

g

w

L

cL

u

w

LsL N

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

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.50

2

4

6

8

10

12

14

Coordinate along the bunch (m)

Coord

inat

e al

ong

the

un

dula

tor

(m)

Undulator sections

e-

Radiation

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

Strawman design

• 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.

10 TW laser system5J -500 fs

RF Linac

4.5 m 1.5 m 3.5 m

45 MeV

0.5 m

RF photogunLinear Prebuncher

Strongly tapered undulator

0 1043

44

45

46

47

Phase

Energ

y

46.363

43.607

ipart FRAME 0.511

6 2 ipart FRAME

z 0

< 10 meters !!!Cryogenic short-period FEL undulator

FEL

1 GeV

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.

AcknowledgementsCollaborators: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