An Injector-prebuncher for a Plasma Electron Accelerator

M. Lampel, C. Pellegrini, R. Zhang, UCLA, Dept of Physics, C. Joshi, UCLA, Dept of EE,

Los Angeles, CA 90095, and W. M. Fawley, LBL, Berkeley, CA 94720

ABSTRACT the design is based.

TABLE I
Optimum operation of a plasma beat-wave or wakefield
Parameter Design Value
accelerator requires an injected beam consisting of a train of electron
bunches separated by the plasma wavelength, with each bunch in the Emittance (unnormalised, 0.5 π mm-mrad
train having a length much shorter than the plasma wavelength, and rms.)
the capability of controlling the relative phase of the electron bunches
and plasma wave. The typical plasma wavelength is about 0.1 mm, Energy spread (rms.) 2%
requiring a bunch length of about 10 to 20 µm, which is difficult to Injector beam energy >2 MeV
achieve with conventional RF based injectors. In this paper we
describe an electron accelerator-buncher system based on a Average charge in bunch 1 nC
photoinjector and an FEL, which can satisfy the plasma accelerator (bunchlet) (108 e-)
requirements.
Bunchlet length 30 - 60 fs
INTRODUCTION Bunchlet diameter (2σ) <200 µm
Phase Control
An injector of up to 25 MeV beam energy is proposed to supply the
absolute +\- 1 ps
Plasma Beat Wave Accelerator (PBWA) [1] with a high quality (ε =
relative <30 fs
0.5 π mm-mrad at full energy) electron beam of sufficient intensity (~
1 nC per pulse) to allow capture and acceleration by the PBWA of a
Injection phase control of the bunchlets facilitates
significant number of electrons to the GeV energy range. The injector
optimization of the PBWA for acceleration, reducing the energy
will be based on proven technology from the Saturnus laboratory [2],
spread and emittance of the PBWA accelerated beam. Absolute
with upgrades of both the BNL RF gun [3], and the Plane Wave
timing is relaxed because the electron pulse can be much longer than
Transformer (PWT) accelerator [4]. The system will provide 1 nC
average micropulse charge focused to a 100 µm spot diameter at the the laser pulse. However relative timing of the plasma beat waves
with the electron bunchlets becomes necessary. A nonlinear optics
injection point. A FIR FEL is proposed as an optical pre-buncher.
Phase locked pre-bunching of the electron pulse into 10 - 20 µm long control method of phase locking bunchlet formation to the beat wave
produced by the lasers is necessary. Several possibilities exist to
bunchlets will allow control of the relative phase of the electron
achieve this and detailed studies will be done to identify the most
bunchlet and the plasma wave. Such control will aid in minimizing
likely candidates so that the issue will be resolved before the optical
PBWA electron beam final energy spread and emittance. The optical
bunching experiment is begun. We note that the options include: 1)
pre-buncher/ final focus system may need to be an achromatic and
Use of a beatwave-plasma interaction to produce 100 µm phase
isochronous system to second or perhaps third order to maintain the
locked radiation to seed an FEL amplifier and; 2) use of nonlinear
bunchlets shape after exiting the optical resonator.
index of refraction changes in a plasma, due to a plasma beat-wave
interaction, to control the optical path length, thus the relative phase,
PBWA REQUIREMENTS
of 100 µm light.
A PBWA operates by mixing two lines from a drive The baseline injector system uses the technology of the
laser in a plasma whose parameters are chosen so that the Saturnus experiment. The Saturnus hardware represents all the
beat between the laser lines matches the plasma frequency. hardware needed for the injector, except for the experiment specific
We assume a multi-terawatt 1 µm laser system as the drive final focus into the PBWA, and phase locking of the electron pulse to
laser system for the PBWA. The PBWA requirements dictate the PBWA laser system.
the design of the injector so that efficient matching can
occur. Table 1 shows the top level parameters upon which
 strength.
INJECTOR DESCRIPTION GINGER calculations for a master oscillator-power
amplifier (MOPA) FIR FEL configuration have been done to begin
A standard RF power system consisting of a modulator and examining the performance of a pre-buncher. Table II gives the
klystron runs a photo-cathode RF gun (0-5 MeV) and an PWT (pulse GINGER parameters of the electron and optical beams studied for the
wave transformer) accelerator section (5-25 MeV). The electron oscillator and the amplifier.
beam generated in the gun is focused by bucking solenoids so that it Table II
has proper size entering the accelerator. X-Y steering magnets correct Electron Beam Parameters
for any small mechanical misalignments. After exiting the Oscillator and Amplifier
accelerator section the beam travels through a beamline guided and
focused by dipole and quadrupole magnets. These magnets match the Electron beam energy 21 MeV
electron pulse into the FIRFEL for prebunching. A final transport Micropulse current
section matches the beam into the PBWA for acceleration by the 150 A
plasma wave. Micropulse length (fwhm) 12 ps
The modulator and klystron provide up to 35 MW,
Micropulse spacing 10 ns
approximately 7 MW to the electron gun and the rest
available to the PWT. The design of the modulator is taken Normalized emittance (edge) 10 π mm-mrad
from the present modulator for the Saturnus experiment,
capable of running at 5 pps. The klystron will be a SLAC Resonator Parameters, osc.
XK-5 of which UCLA currently has two available for this
Single pass gain 2.4
project.
Fourth harmonic light of the PBWA laser drives the Power in spontaneous 15 W
photocathode. The gun itself is the second generation of the emission after 1 pass
successful Brookhaven National Laboratory rf gun. A slight Saturated power (peak) 180 MW
modification of the length of the first cavity, to 0.6 λ, results in an
Resonator length 1.5 m
increased electron beam brightness (a decrease in beam emittance)
over the original "half cavity" design. Round trip time 10 ns
The PWT is a high gradient standing wave accelerator. Undulator Parameters
Several engineering and fabrication improvements are proposed for Oscillator and Amplifier
the new injector: A simplification of the internal cooling of the Undulator period 4 cm
structure; use of all metal sealing surfaces for greater radiation Undulator length, osc. 1m
resistance and better vacuum properties; redesign of the "disks" to
Undulator length, amp. 0.8 m
reduce the maximum surface gradient. This improved PWT has been
fabricated and tested with low level RF power and has been installed On axis magnetic field 1.04 T
into the Saturnus beamline. In order to reach 25 MeV two PWT path length correction factor 11.3
structures will be used. K2/2
The GINGER code test case is based on a
THE FEL AS AN OPTICAL BUNCHER parabolic electron bunch in z of total length 16 ps. The undulator is
linear with external quadrupoles providing focusing in the wiggle
The requirements on the pre-buncher are that ~10 µm plane.
bunches spaced 100 µm apart are injected into the plasma beat wave
The oscillator is assumed to have 60% net cavity
accelerator with emittance and energy spread within the acceptance of reflectance, with a net cavity detuning that is 0% of slippage.
the PBWA. An FEL naturally bunches electrons onto a fraction of the Figure 1 displays results from GINGER showing the
optical wave being amplified, one wavelength apart. From this is it is bunching of the electrons at the end of the amplifier undulator. The
quite clear why an FEL can be used as the PBWA pre-buncher. first pulse displayed, at a time 1.33 ps into the bunch, demonstrates
Matching operation of the pre-buncher to produce the beam the excellent bunching possible through the FEL mechanism. The
desired for injection into the PBWA leads to setting the operating majority of electrons are clearly bunched very sharply in phase over
point of the FEL in terms of energy, undulator length, and field
 Figure 1:

Electron Beam bunching at six slices in time

at the amplifier undulator exit.

~60°, corresponding to less than 20 µm. The forward part of the REFERENCES
electron pulse is well bunched because the optical fields propagate
from the rear to the front due to the slippage of the electron beam. 1) Measurements of the Beatwave Dynamics in Space and
The bunching parameter is about 0.7. To increase the number of well Time, A. Lal, K. Wharton, D. Gordon, C.E. Clayton, M. Everett, C.
Joshi, PAC 95, May 1-5, Dallas, TX
bunched slices, we will consider in future calculations two potentially
beneficial effects: 1) tapering the undulator, and; 2) use of a
2) The UCLA Compact High Brightness Accelerator, P.
waveguide in the amplifier to reduce slippage.
Davis, G. Hairapetian, M. Hogan, C. Joshi, S. Park, C. Pellegrini, J.B.
The amplifier needs a source of 100 µm light to start it.
Rosenzweig, G. Travish, R. Zhang
Two possibilities exist: The FEL oscillator could be built as the
source. The MOPA configuration clearly provides a complete 3) Multipole Field Measurements in the SLAC/BNL #3 RF
solution to the problem of getting 100 µm light, but is expensive and Gun, D. Palmer, R. H. Miller, H. Winick, X.J. Wang, K. Batchelor, M.
does not solve the phasing problem between the bunchlets and the Woodle, I. Ben-Zvi, PAC 95, May 1-5, Dallas, TX
PBWA beatwave. A second possibility would be to use the beatwave
to produce a 100 µm seed for the FEL amplifier directly. This gives 4) Initial Operation of the Plane Wave Transformer (PWT)
automatic phasing between the bunchlets and the wave. An initial Linac R. Zhang, P. Davis, G. Hairapetian, M. Hogan, C. Joshi, S. Park,
calculation indicates that injecting the two lines from the drive laser C. Pellegrini, J.B. Rosenzweig, G. Travish, PAC 95, May 1-5, Dallas,
into a plasma will produce at least 100 kW of coherent 100 µm light TX
in a f/1 cone angle. As can be seen from Table II this power is
several orders of magnitude greater than the spontaneous emission at
the end of the first pass.