The FERMI FEL project at Trieste

Abstract

The main features of the proposed new IR/FIR FEL user facility, FERMI, are described.

Full text

Page 1

THE FERMI FEL PROJECT AT TRIESTE
D. Bulfone, F. Cargnello, G. D'Auria, F. Daclon, M. Ferianis, M. Giannini,
G. Margaritondo§, A. Massarotti, A. Rindi, R. Rosei, C. Rubbia¥, R. Visintini,
R.P. Walker†, A. Wrulich, D. Zangrando, Sincrotrone Trieste S.c.p.A.
F. Ciocci, G. Dattoli, A. De Angelis¶, A. Dipace, A. Doria, G. P. Gallerano, F. Garosi, L. Giannessi,
E. Giovenale, L. Mezi, P. L. Ottaviani, A. Renieri, E. Sabia, A. Segreto‡, A. Torre, ENEA-Frascati
M. Castellano, P. Patteri, S. Tazzari, F. Tazzioli, INFN-Frascati
F. Cevenini, INFN-Naples and Univ. Naples, Dept. of Physics
A. Cutolo, Univ. Naples, Dept. of Electronic Engineering
† corresponding author; ¶ guest; ‡ study grant holder
§ also at EPFL, Lausanne, Switzerland; ¥ also at CERN, Geneva, Switzerland
The main features of the FERMI project are described,
including beam transport design and FEL performance
calculations.
I. INTRODUCTION
The FERMI (Free Electron Radiation and Matching
Instrumentation) project aims to construct an Infra-Red FEL
user facility covering a broad spectral range (2-250 µm) to
complement the high brightness radiation from the ELETTRA
synchrotron radiation facility at Trieste [1]. A unique feature
of the project will be the possibility of carrying out "pump-
probe" experiments using synchronized radiation beams from
FERMI and ELETTRA on the same sample.
The project was launched at a meeting of Italian FEL
experts held in Trieste on the 18th November 1994, chaired by
C. Rubbia, as a collaboration between Sincrotrone Trieste,
ENEA (Frascati), INFN (Frascati) and the University of
Naples (Department of Electronic Engineering).
The facility will make use of an existing linac, that forms
part of the ELETTRA injection system, and a hall into which
the beam can be extracted. In addition, for the first phase of
the project equipment will be used from the suspended
INFN/ENEA "SURF" FEL experiment [2], including the
undulator, beam transport magnets and optical cavity.
Development of the facility is foreseen in three phases.
Phase 1 will consist of setting up the first FEL (FEL-1) to
cover the range 5-20 µm. Initially, the existing optical cavity
will be used to demonstrate laser operation at a fixed
wavelength of 16 µm. Later, new mirrors will be installed
allowing tuneability over the range 5-20 µm with a higher
output coupling, permitting a first user experiment to be
carried out. In Phase 2 a second FEL (FEL-2) will be set up
for the 15-250 µm region. To allow a full support for user
experiments a new building will be constructed (Phase 3)
hosting several experimental stations. As part of this phase a
new linac gun will also be implemented to reduce the
emittance, which is necessary to reach the shortest
wavelengths on FEL-1 (2-5 µm).
II. LAYOUT
Figure 1 shows the layout of the facility adjacent to the
ELETTRA storage ring. The beam from the first 100 MeV
part of the 1.5 GeV injector linac will be transported into an
existing hall adjacent to the linac tunnel, sufficiently large
(32 m x 10 m) to accommodate both FELs as well as space for
power supplies etc. and an area for optical diagnostics and
pilot experiments. In the final phase the FEL beams will be
Fig. 1 Layout of the FERMI FEL facility.

Page 2

transported along the klystron gallery (above the linac tunnel)
and into a new FEL experimental area (~ 21 m x 17 m) which
will host a number of experimental hutches. The location of
the new building will allow the possibility of easily
transporting synchrotron radiation beams into the FEL
experimental area, or FEL radiation into the ELETTRA
experimental hall.
III. LINAC
The low energy part of the linac has been designed to
operate in a FEL mode in the 20-75 MeV range with a 10 µs
pulse length. First tests in this mode of operation are reported
elsewhere in these Proceedings [3]. Table 1 summarises the
expected linac parameters, on which the FEL performance
calculations were based. It will be seen that the linac
performance is not sufficient to allow saturation to be reached
below about 5 µm, and therefore possible schemes for
upgrading the gun are being evaluated.
Table 1. Main linac parameters in the FEL-mode
Energy, MeV
Macropulse repetition rate, Hz
Macropulse length, µs
Charge per micropulse, nC
Micropulse length (FWHM), ps
Peak current, A
Micropulse repetition rate, MHz
Normalized emittance (rms), mm mrad
Energy spread at 75 MeV, %
20-75
10
10
0.4
10
37.5
20.8-31.3
62.5
± 0.3
Fig. 2 Layout of the transport line and two FELs
IV. BEAM TRANSPORT
A particular requirement of the transport line is to provide
a variable path length dependence with energy, in order to
maintain high peak currents or to allow bunch length
manipulations in conjunction with linac phase adjustments.
Fig. 2 shows a suitably flexible design using four bends for
each FEL line. Fig. 3 shows the optical functions for FEL-1 in
the isochronous case. The design is compatible with operation
over the full 20-75 MeV range using the existing magnets.
Fig. 3 FEL-1 transport line optics in the isochronous case.
V. UNDULATORS AND OPTICAL CAVITIES
Table 2 summarises the undulator parameters. The
undulator for FEL-1 was built by Ansaldo in collaboration
with ENEA, Frascati [4]. The minimum gap of 20 mm allows
a vacuum aperture of 15 mm which gives minimal diffraction
losses at 20 µm. The FEL-2 undulator is designed to have the
shortest period allowing 250 µm to be reached at the lowest
envisaged energy of 20 MeV, compatible with a vacuum
aperture of 40 mm for minimal diffraction losses.
Table 2. Main parameters of the FEL undulators.
FEL-1
2.20
0.044
50
20
0.35
1.44
FEL-2
2.28
0.095
24
45
0.43
3.79
Length, m
Period, m
Number of periods
Minimum gap, mm
Peak field, T
K parameter, max.
The optical cavity length has been chosen to be 5.4 m,
which allows the same bunch separation to be obtained in
ELETTRA with a regular filling pattern with 24 bunches. For
FEL-1 the mirror adjustment system for the SURF project will
be used which has combined PZT and motorized adjustment
of mirror tilt and longitudinal position [5]. Mirrors are
available for the first demonstration experiment at 16 µm.
Thereafter other mirrors will be substituted to cover the
5-20 µm range. No decision has yet been taken about the
FEL-2 optical cavity that could employ metal mesh mirrors
and a waveguide cavity.

Page 3

VI. PERFORMANCE CALCULATIONS
Fig. 4. Small signal gain for FEL-1 and FEL-2.
The estimated small signal gain (Gmax.) , including the
effects of emittance, energy spread and slippage is shown in
fig. 4 as a function of undulator K value for various electron
beam energies. Table 3 presents a summary of the optimized
parameters at various wavelengths, assuming 5 % output
coupling and 1 % losses. It can be seen that the gain is
sufficient above about 5 µm to reach saturation in a
sufficiently small time (τr) with respect to the macropulse
duration. The results indicate that high micropulse energies
(Emicro.) and power (Pmicro.) can be produced, with low
average power (Pav.) due to the low duty cycle.
Table 3. Results of FEL performance calculations.
FEL-1
10
34
FEL-2
50
30
λ, µm
E, MeV
Gmax. (%)
τr, µs
Emicro., µJ
Pmicro., MW 2.2
Pmacro., kW 0.80 0.67 0.56 0.49 2.7
Pav., mW3042
5
48
18.3 27.3 31.2 31.4 24.0 33.1 33.6 25.0
6.3 3.8 3.2 3.2
27 24 2018
2.0 1.6 1.4
16
27
20
24
20
46
100 250
2320
4.3
96
9.0
3.0
62
5.8
1.7
2.9
40
3.8
1.1
4.1
23
2.2
0.64
3838 34152 120 79
Fig. 5 Results of pulse propagation analysis (see text).

effects has been made, using a code developed by the ENEA
group based on the solution of the FEL integral equation
including saturation effects. Figure 5 shows the output power
averaged over the macropulse and the optical pulse length,
near the beginning of the macropulse (small signal value) and
at the end (saturated value), as a function of cavity length
detuning (δL) for 2 cases : FEL-1 at 16 µm (upper) and FEL-2
at 250 µm (lower). In general the results confirm that
saturation is reached in the 5-250 µm range, and that the pulse
energy and power is in agreement with the previous
calculations. It can also be seen that the pulse length at small
detuning is smaller than the electron pulse length (rms = 4.2
ps), but increases with detuning and due to saturation,
particularly in the case of large slippage (e.g. at 250 µm).
A more accurate numerical analysis of pulse propagation
VII. PRESENT STATUS
The work presented here is a summary of a more detailed
Conceptual Design Report that has recently been completed.
The present phase of activity concerns the identification of
potential users and possible collaborators, as well as sources
of funding. A formal collaborative agreement between the
Italian partners is also being prepared.
VIII. REFERENCES
[1] C.J. Bocchetta et al., this Conference.
[2] M. Castellano et al., Proc. 3rd European Particle
Accelerator Conference, Editions Frontières (1992) p. 611
[3] G. D'Auria et al., this Conference.
[4] F. Rosatelli et al., Proc. 1991 US Particle Accelerator
Conference, p. 2760.
[5] M. Castellano et al., Nucl. Instr. Meth. A304 (1991) 204.