Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Saturnus: The UCLA High-Gain Infrared FEL Project

Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

Nuclear Instruments and Methods m Physics Research A304 (1991) 155-158 155

North-Holland

Saturnus : the UCLA high-gain infrared FEL project


F. Aghamir, W.A. Barletta, D.B . Cline, J.W. Dodd, S .C. Hartman, T.C. Katsouleas, J. Kolonko,
S. Park, C. Pellegrini and J.-C. Terrien
Center for Advanced Accelerators, Department of Physics, UCLA, 405 Hilgard Ave., Los Angeles, CA 90024-1547, USA

J.G. Davis, C.J. Joshi, N.C. Luhmann and D .B . McDermott


Department of Electrical Engineering UCLA, 405 Hilgard Ave . Los Angeles, CA 90024-1594, USA

S.N. Ivanchenkov, Yu.Yu. Lachin and A.A . Varfolomeev


1 V. Kurchatov Institute of Atomic Energy, Moscow 123182, USSR

We present the status of Saturnus : an infrared FEL operating in the 10 [Lm wavelength region, driven by a compact 20 MeV lmac
with a photomlector, under construction at UCLA . The 1 .5 cm period, 0.5 T peak-field undulator is being built at the Kurchatov
IAE. The FEL is being designed to operate primanly in the self-amplified spontaneous emission mode . We plan to study the startup
from noise, optical guiding, saturation, sidebands and superradiance, with emphasis on the effects important for future short-wave-
length operation of FELs . The photomjector follows closely the Brookhaven design . Electrons are injected into an accelerating
section based on the plane-wave transformer design developed by Swenson at SAIC. Simulation of the linac and FEL show a gain
length of 10 cm, and saturation power of 50 MW

1. Introduction Table 1
A compact 20 MeV linac with an rf laser-driven Saturnus design parameters

electron gun is being built at UCLA . This linac will be Electron beam
used to study the production of high-brightness electron Energy, nominal 16 .5 MeV
beams, and to drive a high-gain, 10 .6 win wavelength Energy spread, rms 0.2%
FEL amplifier, capable of operating in the SASE mode. Peak current 200 A
Saturnus will use a 1.5 cm period, 5.5 kG peak-field Klystron frequency 2.865 GHz
undulator being developed and built at the Kurchatov Pulse repetition rate 5 Hz
Macropulse duration 3.5 Ws
Institute of Atomic Energy . A list of the main FEL,
Micropulse duration, rms - 1 .6 ps
undulator and lmac design parameters are given in table
Charge per bunch -1 nC
1, and a layout of the system is shown in fig. 1.
Normalized emittance, rms 15 mm mrad
Saturnus will mainly study FEL physics in the high-
gain regime, including startup from noise, optical guid- Undulator
ing, sidebands, saturation, and superradiance, with em- Drift tube i .d. within undulator 4 mm
Period 1.5 cm
phasis on the effects important for future short wave-
Length 60 cm
length operation of FELs [1]. For the electron beam we
Fixed gap between pole pieces 5 mm
will study ways to improve the beam brightness and
Field on axis 5.5 kG
peak current. We will have two beam lines, one leading
straight into the undulator, the other designed for beam FEL output radiation
diagnostic and longitudinal bunch compression. This Wavelength 10 .6 win
Gain length 10 cm
second beam line will also be used for other particle
Saturation length 130 cm
beam physics experiments, including plasma wakefield
Peak saturated power 50 MW
acceleration [2].

0168-9002/91/$03 .50 1 1991 - Elsevier Science Publishers B.V. (North-Holland) IV . PROPOSALS/STATUS REPORTS
156 F Aghamtr et al / Saturnus the UCLA high-gain IR FEL project

2. Photoemitting rf electron gun

The electron injector is based on the photocathode,


rf gun designed at Brookhaven National Laboratory [3].
The cathode will be illuminated at 70' or 2'30' off
axis . The idea to use photocathodes in rf electron guns
originated at Los Alamos National Laboratory [4]. The
gun consists of one and one-half cavity cells operating
in the Tr-mode and resonating at 2.856 GHz. Powdered
by 5 MW extracted from a 25 MW SLAC klystron, the
electron gun can produce a beam energy approaching
4.5 MeV . The gun is followed by a solenoid to focus the
beam into the accelerating section.
Photoemtsston is initiated by irradiating the cathode
with a frequency quadrupled light pulse from a 1o 20 30
Nd : YAG/glass laser. Two sapphire viewports are ori- DISTANCE FROM CATHODE (cm)
ented 70° from the electron beam axis and situated Fig. 2. PARMELA simulation of the electron beam radius
symmetrically about the line perpendicular to the planar propagating through the gun and focusing solenoid . The exit of
cathode's surface. Initially a copper cathode will be the gun is 8 cm downstream from the cathode. The end of the
used (quantum efficiency >_ 10 -5 ). For the linearly focusing solenoid is 26 cm from the cathode surface
polarized laser beam. off-axis incidence preferentially
enhances the photoemission by providing an electric electric fields within the 1 2 cells and the PARMELA
field component parallel to the axis . For a LaB6 crystal- code, as extended by McDonald [7] to include photo-
line photocathode, twice the amount of photoemission cathodes, to model the electron beam dynamics . The
at 70 ° has been observed compared to axial illumina rms beam size along the x- or y-axis versus the distance
tion at 0 ° [5]. For long-pulse studies ( >_ 30 ps) the downstream from the photocathode surface is shown to
second port will be used to align the infected pulse from fig. 2, with the focusing solenoid field "on" (focusing
the first port . field strength -- 800 G) .
For short-pulse studies (< 3 ps), off-axis illumina- Without the focusing coil, the beam out of the gun is
tion causes a time delay across the breadth of the laser strongly divergent and only a small part of it would
pulse when it strikes the cathode surface, thereby pro- propagate through the accelerating section. For the
ducing a much longer electron pulse . To preserve an parameters in table 1, PARMELA yields a normalized
ultrafast electron pulse, a mirror placed between the emittance along the x- or y-axis of 15 mm mrad at the
gun and the ltnac, near the electron beam axis, will be exit of the focusing solenoid.
used to illuminate the cathode. The time delay di- Fig. 3 shows the longitudinal phase space, with an
minishes to about 1 ps . energy spread of about ±0 .5% and a phase acceptance
corresponding to a pulse duration of about 4 ps . The
2 1. Rf gun simulations incident laser pulse shall be on the order of 2-4 ps .

Computer simulations of the electron gun have been


done using the SUPERFISH [6] code to model the

4 .62

U
a
w 4 60
A- COILS Z
B- R F GUN LU
C-
D-
VACUUM PORT
P W T LINAC w
E- QUAD TRIPLET E K G
F- DIPOLE 4 .58
H
G- BEAM DUMP
H- UNDULATOR Q
J- OPTICAL DIAG
K- WAKEFIELD EXP
Fig. 1 . Layout of the UCLA free electron laser Two beam 1038 1040 1042
lines are shown The straight path transits the undulator. The
bent path allows beam diagnosis and an auxiliary means to PARTICLE PHASE (DEGREES)

magnetically compress the electron bunches. Fig. 3. Longitudinal phase space at the exit of the solenoid .
F Aghamir et al. / Saturnus the UCLA high-gam IR FEL project 15 7

Table 2
Solid-state laser performance values
Stage Wave- Pulse Pulse
length duration energy
YAG oscillator 1 .064 ~Lm 80 ps 0 .3 WJ
(76 .16 MHz PRF)
Regen. amplifier 1 .064 pin 70 ps 15 mJ
(5 Hz PRF)
Twice doubled 266 rim 50 ps 2 mJ
Add fiber and grating pair 266 rim >_ 2 ps 0 .25 mJ Fig. 4 . Electric field pattern for a half cell of the prototype
plane-wave transformer linac

2 .2 . Nd : YAG/glass laser
4. Hybrid undulator
The ultraviolet fourth harmonic of a high-powdered
Nd : YAG laser system illuminates the photocathode . The hybrid undulator has been designed and is un-
The solid-state laser system contains : (a) a Coherent", der construction at the Kurchatov Institute of Atomic
AntareSTM Nd : YAG laser oscillator, 22 W, mode- Energy in the USSR . The primary magnetic flux is
locked at 38 .08 MHz, (2) a Continuumg, model RGA- provided by C-shaped iron (possible vanadium-per-
44-5, Nd : silicate glass regenerative amplifier, operating mendur) yokes where between the poles thin blocks of
at 5 Hz, (3) an optical fiber and diffraction grating pair neodymium-iron-boron magnets are placed to provide
to provide pulse chirping and compression, respectively, additional magnetic flux along the undulator axis . The
and (4) two frequency doubling KD * P crystals . Ad- field strength is also adjusted by moving the thin Nd-
ditional optics, power isolators, mirrors, etc., complete Fe-B blocks towards or away from the axis on a
the system . Table 2 specifies performance values for the set-screw mount. Fig. 5 is a schematic diagram of the
solid-state laser system . To produce a train of electron Kurchatov undulator. The initial assembly will have 40
bunches sufficient for FEL oscillation (100-200 pulses periods with each period being 1 .5 cm . The gap distance
separated by 13 .1 ns) two more amplification stages between the "yoke" pole pieces is fixed at 5 mm . A
after the regenerative amplifier are required . preliminary mock-up of the undulator having five peri-
ods was measured and yielded an on-axis peak field of
5.5 kG . A computer simulation using the TOSCA [9]
3. Plane-wave transformer linac code gives similar results of approximately 5.5 kG on
axis .
The linac is the prototype plane-wave transformer
structure designed by Swenson [8]. Capable of achieving
20 MeV electron energies within a distance of 0.5 m, the
prototype consists of a cylindrical tank with eight
washers placed along the axis. The distance between
each washer is evenly spaced . Half this distance is the
spacing given between the end washer and the respec-
tive end wall of the tank. The space between the outer
diameter of the tank and the washer array supports a
TEM-style "coaxial" standing wave . The washer array
is equivalent to the "center conductor" . The plane wave
traveling back and forth along the structure couples
power into the "cells" producing a TM oz -like mode
between the washers. This is equivalent to a periodically
loaded coax-cable that transforms the transverse field of
the TEM standing plane waves into a longitudinal elec-
tric field along the axis ready to accelerate particles. A
SUPERFISH simulation of the PWT linae's electric
field shows in fig. 4 the longitudinal orientation on axis
as well as the transition from the outer-lying transverse
fields . Fig. 5 . The Kurchatov hybrid undulator (g = 5 mm).

IV . PROPOSALS/STATUS REPORTS
15 8 F Aghamir et al / Saturnus the UCLA high-gain IR FEL project

E+09

E+08
3 E+07

w E+06
3 E+05
O
a
E+04

0. E+03
E+02
CD
w E+01
w
E+00
E-01
2 .4 .6 .8 1 .0 1 .2 1 .4

UNDULATOR LENGHT (in) UNDULATORLENGHT(in)


Fig 6. TDA simulations of SASE using the parameters m table 1 - (a) power growth through the undulator and (b) radius of the
optical beam within the undulator. Optical guiding is evident since the optical beam remains confined near the electron beam, once
power growth predominates

5. FEL output: diagnosis and simulations Acknowledgements

Our group appreciates the help received from H.


The FEL's output wavelength will closely match the
Kirk, K. Batchelor, J. Xie, J. Sheehan, G. Bennett, J.
10 .6 Win radiation which CO, lasers produce, simplify-
Wurtele, M. Allen, G. Loew, and H . Hoag . This work is
ing the diagnosis of the output radiation. For cryogenic
supported by DOE Grant no . DE-FGO-90ER40565 .
infrared detectors sensitive to 10 I.Lm wavelengths, the
expected power output from spontaneous emission in
the FEL is marginally detectable. For 1 nC charge per References
bunch, 1 .2 X 10 7 photons per micropulse are expected [I] C. Pellegrim, Nucl . Instr. and Meth . A272 (1988) 364.
from spontaneous emission within an angle 0 - 1(yt/N ) [2] J Smolm, T. Katsouleas, C. Joshi, P. Davis and C. Pel-
and a linewidth 1/(2N) . This amount can be detected legrmi, Bull . Am . Phys . Soc 35 (1990) 2103
directly . For smaller charges one can integrate the sig- cf . : T. Katsouleas, J.J Su, W B. Mon, C. Joshi and J.M
nal over many pulses . Dawson, m Microwave and Particle Beam Sources and
The FEL performance in SASE is evaluated using Directed Energy Concepts, ed . H.E . Brandt, Proc. SPIE
the TDA code developed by Tran and Wurtele at MIT 1601 (1989) 428.
K Batchelor, H Kirk, J. Sheehan, M. Woodle and K.
(101 . Using the parameters of table 1, TDA predicts a
McDonald, Proc. Europ. Particle Accelerator Conf., Rome,
gain length of 10 cm, with a saturation power of 50
Italy, 1988 ;
MW . This assumes that the full 200 A peak current is
also : K. Batchelor et al ., Proc . 1989 Particle Accelerator
achievable. TDA also assumes free-space propagation, Conf ., Chicago, IL, p 273
whereas our 0.2 mm wide electron traverses the undula- [4] J.S . Fraser and R.L . Sheffield, IEEE J. Quantum Electron.
tor inside a - 4 mm inner diameter drift tube. QE-23 (1987) 1489;
A calculation employing complete waveguide modes J.S Fraser, et al, Proc . 1987 IEEE Particle Accelerator
may modify these results . Fig. 6a shows a graph of the Conf ., eds. E.R . Lindstrom and L.S . Taylor, Washington,
power vs distance along the undulator as projected by DC, p. 1705 ;
TDA. Fig. 6a uses an input power of 40 mW, corre- M.E . Jones and W Peter, Theory and simulation of
high-brightness electron beam production from laser-
sponding to the spontaneous radiation emitted in one
irradiated photocathodes in the presence of DC and RE
gain length, within the angle and linewidth given above.
electric fields, 6th Int. Conf. on High-Power Particle
This radiation is assumed to be focused to a spot radius
Beams, Japan, 1986 .
(XLundul)i/2/4m at the centre of the first gain length . [5] G.T. Bennett, private communication
Fig. 6b plots the calculated rYris optical beam radius [6] K. Halbach and R.F . Holsinger, Part . Accel. 7 (1976) 213 .
vs the distance along the undulator. Notice that once [71 K.T . McDonald, IEEE Trans. Electron . Devices ED-35
growth occurs the optical beam collapses and remains (1988) 2052 .
in the vicinity of the electron beam as it propagates [8] D.A . Swenson, Proc . Europ. Particle Accelerator Conf .,
down the entire undulator. This is indicative of optical vol. 2, 1988, Rome, Italy, ed . S. Tazzari, p 1418 .
guiding. IEEE Proc . vol. B127 (6) (1980) .
[10] T.M . Tran and J S Wurtele, Comput . Phys Commun . 54
(1989) 263

You might also like