STATUS OF THE SPARX FEL PROJECT*

C.Vaccarezza, D.Alesini, M.Bellaveglia, S.Bertolucci, M.E Biagini, R.Boni, M.Boscolo,

M.Castellano, A.Clozza, L.Cultrera, G.Di Pirro, A.Drago, A.Esposito, M.Ferrario, D.Filippetto,
V.Fusco, A.Gallo, A.Ghigo, S.Guiducci, M.Migliorati, L.Palumbo, L.Pellegrino, M.Preger,
C.Sanelli, M.Serio, F.Sgamma, B.Spataro, A.Stella, F.Tazzioli, M.Vescovi, C.Vicario, INFN
Frascati, Italy
F. Ciocci, G.Dattoli, A.Doria, F.Flora, G.Gallerano, L.Giannessi, E.Giovenale, G.Messina, P.L.
Ottaviani, G.Parisi, L.Picardi, M.Quattromini, A.Renieri, C. Ronsivalle ,ENEA C.R. Frascati, Italy
S.Cialdi, C.Maroli, V.Petrillo, M. Rome, L.Serafini, INFN-Milano, Italy
L.Catani, E.Chiadroni, A. Cianchi, C.Schaerf, INFN-Roma II, Roma, Italy
P.Musumeci, INFN Roma, Italy
F.Alessandria, A.Bacci, F.Broggi, C.De Martinis, D.Giove, M.Mauri, INFN/LASA, Segrate, Italy
L.Ficcadenti, M.Mattioli, A.Mostacci, University La Sapienza, Roma, Italy
P.Emma, SLAC, Menlo Park, CA, U.S.A.
S.Reiche, J. Rosenzweig, UCLA, Los Angeles, CA, U.S.A

Abstract
The SPARX project consists in an X-ray-FEL facility Table 1: Electron beam parameters.
jointly supported by MIUR (Research Department of Beam Energy 1÷2 GeV
Italian Government), Regione Lazio, CNR, ENEA, INFN
and Rome University Tor Vergata. It is the natural Peak current 1-2.5 kA
extension of the ongoing activities of the SPARC Emittance (average) 2 mm-mrad
collaboration. The aim is the generation of electron beams
characterized by ultra-high peak brightness at the energy Emittance (slice) 1 mm-mrad
of 1 and 2 GeV, for the first and the second phase Energy spread (correlated) 0.1 %
respectively. The beam is expected to drive a single pass
FEL experiment in the range of 13.5÷6 nm and 6÷1.5 nm, Repetition Rate 50 Hz
at 1 GeV and 2 GeV respectively, both in SASE and
SEEDED FEL configurations. A hybrid scheme of RF the schematic layout of the accelerator is shown. A 150
and magnetic compression will be adopted, based on the MeV SPARC-like photoinjector [1] is meant to provide a
expertise achieved at the SPARC high brightness 300÷500 A beam, adopting the velocity bunching
photoinjector presently under commissioning at Frascati compression scheme. A first linac section L1 rises the
INFN-LNF Laboratories [1,2]. The use of beam energy up to 350 MeV, where a first magnetic
superconducting and exotic undulator sections will be chicane is foreseen mainly for comparing the overall
also exploited. In this paper we report the progress of the efficiency between the two compression methods at low
collaboration together with start to end simulation results energy. After a second linac section L2, i.e. at the energy
based on a combined scheme of RF compression around 0.5 GeV the main magnetic compressor is located
techniques. that rises the beam peak current up to Ipk~1 kA, according
to a ‘hybrid” compression scheme consisting in one RF
THE SPARX LAYOUT compression stage at low energy, plus one magnetic
A spectral range from 13 nm to 1 nm has been chicane at 0.5 GeV. A third accelerating section L3 brings
considered for the radiation. SASE-FEL’s in this the beam energy up to E~1GeV and a first extraction
wavelength range require high brightness beam at the dogleg DL1 drives the beam through a diagnostic section
undulator entrance. In Table 1 the electron beam and to the first undulator for SASE experiments in the
parameter list is reported for such a source, while in Fig. 1 radiation wavelength range of λr ~ 13.5÷6 nm. This is

Figure 1: SPARX Linac schematic layout

___________________________________________

*Work partially supported by MIUR for the realization of “Grande Infrastruttura Laser ultrabrillante per raggi X multiscopo”
what is meant for the first phase of the SPARX project. Table 2: RF compressor parameter
For the Phase II a third magnetic chicane is foreseen
downstream to compress the beam at E=1GeV and reach RF B1,B2,B3 Current Emittance
peak currents of the order of Ipk~2÷2.5 kA, and a final (μm)
phase range (gauss) (A)
energy of E=2GeV, by means of a fourth linac section L4.
A second extraction dogleg DL2 provides the beam -60±/-75± 1200,0,0 117-151 0.7
diagnostics and delivery to the second undulator for the -75±/-83± 1200,1400,0 151-249 0.8
wavelength range λr~ 5÷1.5 nm
-83±/-87± 1200,1400,0 249-458 1.3
The RF Photoinjector -87±/-91± 1200÷1800 458-1180 2.8
The injector design is based on the SPARC high
brightness photoinjector presently under commissioning
at Frascati Laboratories [1,2]. It considers a 1.1 nC bunch
10 ps long (flat top) with 1.1 mm radius, generated inside
a 1.6-cell S-band RF gun of the same type of the BNL-
SLAC-UCLA one [3] operating at 120 MV/m peak field
equipped with an emittance compensating solenoid. Three
standard SLAC type 3-m TW structures each one
embedded in a solenoid boost the beam up to 150 MeV.
According to the simulation results the beam compression
at low energy (<150 MeV), still in the space charge
dominated regime, turns out to be feasible provided that a
proper emittance compensation technique is adopted [4],
Moreover the propagation of a short bunch in the
following accelerating sections reduces the potential
emittance degradation caused by transverse wake fields,
while a proper phasing of the linac can control the Figure 2: Average current, transverse emittance and
longitudinal wake fields. A systematic study based on envelope, axial magnetic field vs. z for a final 450 A
PARMELA code simulations has been done in order to electron beam., tracked with PARMELA [7].
optimize the parameters that influence the compression BC2, while the on crest L3 section rises the beam energy
[5], the results of computations show that peak currents up to E=1 GeV and contributes to the energy spread
up to kA level are achievable at the injector exit with a reduction. For the magnetic compression in BC3, the
good control of the transverse and longitudinal emittance proper off-crest acceleration is applied also to the L3
by means of a short SW section operating at 11424 MHz section, while L4 contributes to the energy spread
[6] placed before the first accelerating section. On the cancellation and brings the beam energy up to E=2GeV.
other side the results obtained without the IV harmonic The invariant envelope matching condition will be
correction prior the RF compressor, show that is anyway applied to the lattice as proposed by Ferrario et al [8].
possible to reach good compression factors but paying for
a heavy deformation of the longitudinal bunch The Bunch Compression
distribution, a strong sensitivity to RF compressor phase, To increase the beam current up to the kA order
and a highly non linear longitudinal phase space. In Table magnitude a magnetic compression stage is foreseen in
2 a summary is reported of the RF compression studies the BC2 chicane for the beam with final energy E=1GeV,
results. A medium RF compression factor has been chosen Ipk~1 kA (Phase I), and in the BC3 chicane for the beam
and presented here as the first referring case for the with final energy E=2 GeV, Ipk~2.5 kA (Phase II). With
SPARX beam dynamics studies, it gives an average peak the previously described setting for the RF photoinjector,
current Ipk~400 A at the exit of the photoinjector, as a further compression factor of 2 in the BC2 chicane
highlighted in Table 2 and shown in Fig. 2. produces an average peak current Ipk~1 kA, able to reach
The Linac SASE saturation in the wavelength range of λr~ 10÷6 nm
at the energy of 1GeV. The parameters of the BC2
In the present configuration the SPARX accelerator compressor for this case are shortly listed in Table 3. The
(PHASE I and II) is composed of four separate S-band photoinjector incoming beam has been generated with
linac sections L1, L2, L3, L4, with E= 25 MV/m PARMELA, considering a thermal emittance value of
accelerating field, located downstream a SPARC-like εx~0.6 μm, the downstream tracking in the chicane has
photoinjector (see Fig. 1). At the entrance of each of the been performed with ELEGANT [9]. The tracking in the
three magnetic chicanes an X-band section is provided to BC2 chicane has also been checked with PARMELA and
linearize the beam longitudinal phase space prior the CSRtrack [10] codes in order to clarify the role played by
magnetic compression. For the 1 GeV channel of Phase I, the space charge effect in the transverse emittance
both L1 and L2 sections are meant to work off crest, to dilution. In Fig. 3 the obtained results for the horizontal
provide the required energy chirp to compress the beam in
 Table 3: BC2 compressor parameters.
10 10
Beam Energy E GeV 0.5 10 9
Initial rms bunch length σzi mm 210 10 8
Final rms bunch length σzf mm 90

Power (W)
10 7
Incoming energy spread σδi % .45 10 6
Momentum compaction R56 mm -27 10 5
2nd order mom. compaction T566 mm +42 10 4
10 3
and vertical slice emittance are reported, together with the 10
2
0 5 10 15 20 25 30 35
slice energy spread and current distribution along the Z (m)
bunch. The three codes outputs show that for a peak
current around Ipk~1 kA the space charge effect doesn’t
Undulator Layout
heavily contribute to dilution of the transverse emittance.
Figure 5: Genesis code simulation for the fundamental
A detailed study is undergoing for the cases with higher
harmonic wavelength λr= 12 nm.
compression factors both in BC2 and BC3 chicanes. The
results of the beam tracking up to the undulator entrance required to saturate at 12 nm, see Fig. 5. These first
in the 1 GeV channel are shown in Fig. 4. preliminary results will be the starting point for further
optimizations.

CONCLUSIONS
The SPARX project aim is the generation of electron
beams characterized by ultra-high peak brightness at the
energy of 1 and 2 GeV, (Phase I and II), for SASE and
SEEDED FEL experiments in the range of λr ~ 13.5÷1.5
nm. It is jointly supported by the Italian Government and
Regione Lazio with a five years schedule for the first
Figure 3: Slice analysis of simulation results trough the phase. The critical components such as the RF-
BC2 compressor for the horizontal emittance and energy compression scheme, magnetic chicane, etc. will be tested
spread. The 50k particles have been tracked with during the phase II of the SPARC project. A first general
ELEGANT, PARMELA and CSRtrack with “g_to_p” layout for SPARX has been proposed and first start to end
type forces. simulations for the 1 GeV channel of Phase I, have been
presented using a “hybrid” scheme of bunch compression
with RF and magnetic compression techniques involved
together.

REFERENCES
[1] L. Serafini, et al, “Status of the SPARC project”,
these proceedings.
[2] A. Cianchi, et al, “Characterization of the SPARC
Photo-injector”, these proceedings.
[3] D.T. Palmer,"The next generation photoinjector",
PhD. Thesis, Stanford University.
[4] L. Serafini and M. Ferrario,"Velocity Bunching in
PhotoInjectors" , AIP CP 581, 2001, pag. 87.
[5] C.Ronsivalle, et al, “Optimization of RF compressor
in the SPARX photoinjector”,PAC05,Knoxville(TN).
Figure 4: Energy spread, current distribution and [6] B. Spataro, et al, “Design and RF measurements of
transverse emittance along the bunch at the end of the an X-band acc. structure”, NIM A554:1-12, 2005.
SPARX 1GeV DL1 dogleg, tracked with ELEGANT. [7] Billen, “PARMELA”, LA-UR-96-1835, 1996.
[8] M. Ferrario, et al, “Beam Dynamics around
The SASE-FEL Source transition”, ICFA panel on “The Physics and
Time dependent FEL simulations, performed with the Applications of High Brightness electron Beams”,
code GENESIS [11] using the particle distributions Erice, Sicily, October 14-19 2005.
produced by the Linac simulations, presented in the [9] M. Borland, “ELEGANT” LS-287, ANL, Argonne,
previous section, are in progress showing saturation for IL 60439, USA.
90% of bunch slices after 25 m of active undulator length. [10] M.D. Dohlus, T. Limberg, “CSRtrack: Faster
We assume to use the same undulator of the SPARC Calculation of 3-D CSR Effects, FEL04, Trieste.
project [1] with six additional 2.13 m long modules [11] S. Reiche, Nucl. Instrum. & Meth. A429,243 (1999).