Self-amplified spontaneous emission for a single pass free-electron laser

Article (PDF Available)inPhysical Review Special Topics - Accelerators and Beams 14(6):060712 · June 2011with33 Reads
DOI: 10.1103/PhysRevSTAB.14.060712
Abstract
SPARC (acronym of ''Sorgente Pulsata ed Amplificata di Radiazione Coerente'', i.e. Pulsed and Amplified Source of Coherent Radiation) is a single pass free-electron laser designed to obtain high gain amplification at a radiation wavelength of 500 nm. Self-amplified spontaneous emission has been observed driving the amplifier with the high-brightness beam of the SPARC linac. We report measure-ments of energy, spectra, and exponential gain. Experimental results are compared with simulations from several numerical codes.
Self-amplified spontaneous emission for a single pass free-electron laser
L. Giannessi,
1,
*
D. Alesini,
2
P. Antici,
2
A. Bacci,
2,4
M. Bellaveglia,
2
R. Boni,
2
M. Boscolo,
2
F. Briquez,
10
M. Castellano,
2
L. Catani,
8
E. Chiadroni,
2
A. Cianchi,
8
F. Ciocci,
1
A. Clozza,
2
M. E. Couprie,
10
L. Cultrera,
2
G. Dattoli,
1
M. Del Franco,
1
A. Dipace,
1
G. Di Pirro,
2
A. Doria,
1
A. Drago,
2
W. M. Fawley,
11
M. Ferrario,
2
L. Ficcadenti,
2
D. Filippetto,
2
F. Frassetto,
6
H. P. Freund,
12
V. Fusco,
1,2
G. Gallerano,
1
A. Gallo,
2
G. Gatti,
2
A. Ghigo,
2
E. Giovenale,
1
A. Marinelli,
9,2
M. Labat,
10
B. Marchetti,
8
G. Marcus,
9
C. Marrelli,
2
M. Mattioli,
2
M. Migliorati,
2,5
M. Moreno,
5
A. Mostacci,
5
G. Orlandi,
13
E. Pace,
2
L. Palumbo,
2,5
A. Petralia,
1
M. Petrarca,
2
V. Petrillo,
3,4
L. Poletto,
6
M. Quattromini,
1
J. V. Rau,
7
S. Reiche,
13
C. Ronsivalle,
1
J. Rosenzweig,
9
A. R. Rossi,
2,4
V. Rossi Albertini,
7
E. Sabia,
1
L. Serafini,
4
M. Serluca,
5
I. Spassovsky,
1
B. Spataro,
2
V. Surrenti,
1
C. Vaccarezza,
2
M. Vescovi,
2
and C. Vicario
13
1
ENEA C.R. Frascati, Via E. Fermi 45, 00044 Frascati (RM), Italy
2
INFN-LNF, Via E. Fermi 40, 00044 Frascati, RM, Italy
3
Universita
`
degli Studi di Milano, Via Celoria 16, 20133 Milano, Italy
4
INFN-Mi,Via Celoria 16, 20133 Milano, Italy
5
Universita
`
La Sapienza, Piazzale Aldo Moro 1, 00185 Roma, Italy
6
CNR-IFN, Via Trasea 7, 35131 Padova, Italy
7
ISM-CNR Via del Fosso del Cavaliere 100, 00133 Roma, Italy
8
INFN-Roma II, Via della Ricerca Scientifica 1, 00133 RM, Italy
9
UCLA, 405 Hilgard Avenue, Los Angeles, California 90095-1547, USA
10
SOLEIL, L’Orme des Merisiers Saint-Aubin, BP 48 91192 GIF-sur-Yvette Cedex, France
11
Sincrotrone Trieste S.C.p.A., Area Science Park, S.S. 14 Km 163.5, I-34149 Trieste, Italy
12
Science Applications International Corporation, McLean, Virginia 22102, USA
13
Paul Scherrer Institute, 5232 Villigen, PSI, Switzerland
(Received 24 February 2011; published 29 June 2011)
SPARC (acronym of ‘Sorgente Pulsata ed Amplificata di Radiazione Coerente’’, i.e. Pulsed and
Amplified Source of Coherent Radiation) is a single pass free-electron laser designed to obtain high gain
amplification at a radiation wavelength of 500 nm. Self-amplified spontaneous emission has been
observed driving the amplifier with the high-brightness beam of the SPARC linac. We report measure-
ments of energy, spectra, and exponential gain. Experimental results are compared with simulations from
several numerical codes.
DOI: 10.1103/PhysRevSTAB.14.060712 PACS numbers: 41.60.Cr, 41.50.+h, 42.55.Vc
I. INTRODUCTION
Recent progress in accelerator technology has led to
wide-range tunable laser radiation in the vacuum-
ultraviolet and x-ray spectral regions by means of the
development of free-electron lasers (FEL) based on the
principle of self-amplified spontaneous emission (SASE)
[15]. In a SASE FEL, lasing occurs in a single pass of a
relativistic, high-brightness electron beam driven through a
long undulator magnetic structure. SASE emission and its
properties in both frequency and time domains have been
observed and studied at various wavelengths in several
experiments [610]. More recently, radiation in the hard
x-ray range, at wavelength of 1.5 A
˚
, has been obtained at
LCLS [11].
SPARC is a single pass FEL amplifier test facility de-
signed to study the amplification process under various
operating conditions. A layout of the SPARC FEL is shown
in Fig. 1.
The electron beam is generated in a high-brightness rf
gun [Fig. 1 (A)], which was implemented in the first
operating phase for examining the beam evolution in the
drift between the gun and the main linac [12]. The
electron beam then enters in a linac composed of three
traveling wave-type accelerating sections [Fig. 1 (B) and
Fig. 2].
The undulator [Fig. 1 (D) and Fig. 3] consists of six
permanent magnet sections with period
u
¼ 2:8cm. The
intersections between the modules host quadrupoles for
horizontal focusing and radiation diagnostic stations.
Each station is equipped with actuators allowing the in-
sertion of alumina screens to monitor the electron beam
size and position, and aluminum mirrors to extract the
radiation. At the end of the undulator sequence an in-
vacuum spectrometer built by the LUXOR Laboratory
(Padova) [13,14] is the main radiation diagnostic.
*
luca.giannessi@enea.it
Published by the American Physical Society under the terms of
the Creative Commons Attribution 3.0 License. Further distri-
bution of this work must maintain attribution to the author(s) and
the published article’s title, journal citation, and DOI.
PHYSICAL REVIEW SPECIAL TOPICS - ACCELERATORS AND BEAMS 14, 060712 (2011)
1098-4402=11=14(6)=060712(8) 060712-1 Published by the American Physical Society
Commissioning of the SPARC FEL began in au-
tumn 2008 with the following main goals: (1) transport
the beam through the vacuum chamber up to the beam
dump consistently with the matching condition in the un-
dulator; (2) characterization of the spontaneous and stimu-
lated radiation; and (3) demonstration of ‘velocity
bunching’ techniques [15,16] according to the invariant
envelope condition [17]. These steps were completed dur-
ing winter 2009, and the first SASE FEL spectra were
obtained in February of the same year. A few months later,
a substantial increase of the brilliance of the radiation
extracted from the FEL source was obtained with a longi-
tudinally flattop e-beam current, by increasing the bunch
charge and by anticipating the phase in the gun to reduce
the debunching in the first stage of acceleration, permitting
a higher beam current. In this paper we report on SPARC
lasing performance obtained so far in SASE mode. In the
next section we discuss the accelerator performance, the
diagnostics, and the characteristics of the electron beam
available for the SASE FEL amplification experiment. We
then describe in detail the energy calibration of the spec-
trometer used as the main diagnostic. Finally, we report on
observation and analysis of SASE emission at 500 nm.
II. ELECTRON BEAM CHARACTERIZATION
The SPARC beam diagnostic system has been designed
to ensure the possibility of optimizing the beam accelera-
tion and matching through all the linac structures and the
undulator, and to measure the main beam parameters prior
to injection in the FEL. The beam envelope is recon-
structed by measuring the rms beam size on four screens
along the linac: at the entrance of each rf structure and at
the exit of the accelerator, in the seven meter transfer line
(TL) leading to the undulator. The TL [Fig. 1 (C)] hosts six
quadrupoles arranged in two triplets and is designed to
match the beam to undulator optics. Along the transfer line
a dipole-based magnetic spectrometer deflects the beam on
a Ce:Yag screen 100 m thick, allowing the measurement
of beam energy and energy spread. The first triplet in the
TL is used to measure the projected emittance by quadru-
pole scan [18]. An S-band, 5 cell, standing wave rf deflect-
ing cavity (RFD) [19,20], placed before the dipole, in
combination with the first quadrupole triplet and the dipole
itself, is used to measure the slice longitudinal (bunch
length, slice energy spread, longitudinal trace space) and
transverse (slice emittance) beam parameters. During
FIG. 2. SPARC linac accelerator provides a final maximum
beam energy of about 178 MeV. The first two linac accelerating
structures are surrounded by two long solenoids (blue structures)
providing the additional focusing (with a maximum field of
0.18 T) required to match the beam envelope to the linac,
according to the invariant envelope conditions, when the linac
operates in velocity bunching compression mode.
FIG. 3. SPARC undulator consisting of six sections made by
75 periods of 2.8 cm each (77 periods including two termination
periods), with gaps variable in the range from 25 to 8.4 mm
(limited by the vacuum chamber) corresponding to a maximum
undulator parameter K
u
2:2 (see definition below) at the
minimum gap.
FIG. 1. General SPARC layout. (A) High-brightness rf gun
(BNL/UCLA/LCLS design). (B) Linac, composed by three
traveling wave-type accelerating sections. (C) Transfer line.
(D) Undulator, made by six permanent magnet modules, realized
by ACCEL Gmbh. (E) Seeding equipment.
L. GIANNESSI et al. Phys. Rev. ST Accel. Beams 14, 060712 (2011)
060712-2
commissioning, the deflecting voltage in the RFD was
limited to a maximum of 1.5 MV; this limited the resolu-
tion to 50 fs=pixel at 150 MeV. The overall measurement
resolution is affected by the beam dimensions on the screen
with the RFD switched off. For typical SPARC beam sizes
at the screen (70 m), the calculated resolution is about
90 fs at 150 MeV. Via a similar argument, the resolution of
the magnetic spectrometer can be estimated to be
5 keV=pixel at 150 MeV, and the overall measurement
resolution on the relative energy spread to be about 0.01%.
The beam used in the SASE FEL experiment was ob-
tained with a longitudinally flattop laser pulse, with a
length of 6–8 ps (FWHM). Whereas the cathode can
nominally deliver up to 1 nC, the charge extracted in the
FEL experiments was about 400 pC. This choice resulted
from an optimization of the achievable beam brightness,
which was mainly limited by the field gradient at the gun.
A higher charge, with the gun operated at a maximum
achievable gradient of about 105 MV=m, was accompa-
nied by an increase of the beam emittance without a
corresponding increase of the peak current. Magnetic com-
pression has not been implemented at SPARC, while ve-
locity bunching was not considered as an option because of
the consistent drop in the beam energy and of the large
energy spread induced at the end of the acceleration phase,
when operating the first linac section off crest. A typical
reconstructed longitudinal phase space is shown in Fig. 4.
We measured an energy spread of the order of 0.1% and
an energy stability better than 0.1%. In Fig. 5 the beam
current along the electron bunch, deduced by the phase
space of Fig. 4, is shown. In this condition the maximum
current is about 53 A and the rms bunch length is 2.6 ps.
Figure 6 presents the beam energy (blue) and the incoher-
ent energy spread (red) as a function of the position along
the bunch. The main parameters of the beam during the
experiment are summarized in Table I. The transverse
emittance measured with the quadrupole scan at the end
of the linac is 2.9(2.5) mm-mrad in the vertical (horizontal)
plane.
FIG. 4. Phase space of the e-beam measured with time/energy
dispersion induced by rf-deflector/spectrometer dipole.
FIG. 5. Beam current as a function of time along the micro-
bunch. Error bars represent 1 standard deviation evaluated by
combining quadratically the relative local uncertainties on
charge and bunch length ( 5%).
FIG. 6. Beam energy spread (A) and beam energy (B) as a
function of time along the bunch. Error bars represent 1
standard deviation on the local energy spread measure. The
energy error bars are calculated as 3 standard deviations of
the energy measured at the specific longitudinal position. This
parameter is related to the instrumental accuracy of the single
shot measurement and does not reflect shot to shot energy
fluctuations or drifts as indicated in Table I.
TABLE I. Main beam parameters measured during the experi-
ment. The error figures represent 1 standard deviation over 20
acquisitions (except for the charge which is obtained with 100
acquisitions). The slice energy spread uncertainty is instead
calculated as the standard deviation of the energy spread aver-
aged over the bunch length, in a single shot measurement.
Beam energy (MeV) 152:08149:8
Relative energy spread
(projected, rms)
9:0 10
4
0:8 10
4
Relative energy spread
(slice, rms)
2:0 10
4
0:8 10
4
Bunch length (ps, rms) 2:60 0:05
Electron charge (pC) 400 20
Beam current (A) 53 4
Repetition rate (Hz) 10
Vertical emittance
(projected, mm mrad)
2:9 0:3
Horizontal emittance
(projected, mm mrad)
2:5 0:3
SELF-AMPLIFIED SPONTANEOUS EMISSION FOR A ... Phys. Rev. ST Accel. Beams 14, 060712 (2011)
060712-3
During a typical eight-hours shift, we observed slow
drifts, mainly in the beam energy, in the range listed in
the table.
III. RADIATION DIAGNOSTICS
The main diagnostic used to measure the output radia-
tion is an in-vacuum spectrometer [13,14] operating in the
spectral range 35–370 nm. The instrument is a 1 m normal
incidence spectrometer with a Princeton UV grade CCD
camera, allowing the detection of spectra both in single
shot and in integrated mode. The CCD camera and the
necessary upstream optics were calibrated in energy. The
spectral image permits the reconstruction of the total pulse
energy together with the spectral parameters as central
wavelength, linewidth, and, in the vertical direction, of
the spot size and position of the radiation beam.
A layout of the present setup is shown in Fig. 7.
After exiting from the last undulator section, a metallic
aluminum mirror deflects the light beam towards the spec-
trometer. The FEL radiation is P polarized and the theo-
retical P reflectivity has been used in deriving a calibration
factor. Different filters are available along the radiation
beam line, mounted on two filter wheels (FW1, FW2).
Neutral density filters allow the attenuation of the radiation
intensity at the CCD for wavelengths longer than 350 nm.
Bandpass filters may be used to remove intense signals
from long wavelengths while observing the higher order
harmonics of the FEL. All the filters were characterized by
a wavelength transmission curve which is used for the
energy calibration. At the entrance of the spectrometer,
the light passes through a rectangular slit of variable hori-
zontal width (302000 m). Assuming a circular beam at
the slit entrance, the energy lost at the slit was recovered as
a function of its width by measuring the spot profile in the
vertical direction. Calibration curves of the slit attenuation
were derived by reconstructing the spot of a fixed source
while varying the slit aperture. This procedure was re-
peated using both the SASE FEL radiation (at 500 nm)
and the radiation from the seed laser (at 400 nm) available
at SPARC, obtaining similar results. The spectrometer
spectral range covers the interval 35–560 nm with three
different gratings (600, 1200, 2400 grooves=mm) operat-
ing at different wavelengths. The first two gratings are
Al=MgF2 coated, while the last is Pt coated. Each grating
was characterized by its own efficiency curve. The spec-
trometer is equipped with a CCD detector (Versarray,
1300B-Princeton Instruments) with a gain of 1:14 e
=
ADU (analog to digital unit). The CCD efficiency
(e
=incident photon) is given by the product of the quan-
tum efficiency (interacting photon=incident photon) and
the quantum yield (e
=interacting photon). The spec-
trometer efficiency is given by the product of the grating
efficiency and the CCD efficiency (see Fig. 8). We have
accounted for the device magnification which zooms on the
CCD the image of the entrance slit with a magnification
factor of 1.374. The global calibration is the product of the
calibration curves of all the elements crossed by the light at
a given wavelength: last mirror, filters, grating, CCD.
The calibration procedure was tested at 500 nm by
comparing the energy reconstructed from the SASE spec-
tra with the direct measurement from a calibrated high
sensitivity pyroelectric detector (see Fig. 9). The data set
FIG. 8. Energy calibration including the contribution of the
CCD detector, the reflectivity of the different gratings. A plati-
num coated grating (2400 grooves) covers the short wavelength
range (black continuous) and two aluminum coated gratings,
1200 grooves (blue dotted) and 600 grooves (red dashed) cover
the middle and long wavelength ranges, respectively.
FIG. 7. Layout of the spectrometer and radiation diagnostics at
the end of the undulator sequence.
FIG. 9. Comparison of energy measured with pyroelectric
detector (blue) and integrating the spectral images with the
appropriate calibration factor (red). Error bars (red) correspond
to 1 standard deviation over 100 shots. Numbers in parentheses
indicate, for each data set, the undulator sections set at resonance.
L. GIANNESSI et al. Phys. Rev. ST Accel. Beams 14, 060712 (2011)
060712-4
has been obtained by varying the number of undulator
sections participating in amplification.
IV. SASE EXPERIMENTS
When the beam with the parameters listed in Table I was
injected in the SPARC undulator, FEL saturation was
nearly reached at a resonant wavelength of
0
¼
u
ð1 þ
K
2
u
=2Þ=ð2
2
Þ’500 nm, where is the Lorentz factor. The
normalized undulator strength K
u
¼ e
0
B
u
u
=ð2m
0
cÞ
was 2.07. We determined the overall undulator focusing
by imposing a periodic condition to the transverse beam
Twiss parameters over the focusing defocusing (FODO)
lattice composed by an undulator section and the drift
between two adjacent undulators hosting the quadrupole,
and by equalizing the transverse average Twiss
x;y
coef-
ficients in the two directions (see Fig. 10). With this
condition, we have h
x
i¼h
y
i’1:5m and the rms
beam radius averaged over the length of the undulator
section is about 120 m.
The evolution of the radiation pulse energy as a function
of the position in the undulator sequence was obtained by
turning off the FEL interaction via progressively opening
the gap of the sections. The pulse energy measured at the
fundamental and third harmonic wavelength is shown in
Fig. 11. The sequence is represented by black circles with
each point corresponding to an average over 20 events. The
error bars correspond to 2 standard deviations of the shot
to shot fluctuations.
After closing all the undulator gaps to the nominal value,
we optimized the output energy by varying the relative
phase between the cathode drive laser and the rf accelera-
tion fields, leading to an efficiency improvement of about 1
order of magnitude (see Fig. 11, squares). We observed an
overall amplification factor close to 10
7
, with an estimated
gain length L
g
0:7m. The effective Pierce coefficient
[21], derived from the gain length according to the relation
¼
u
=ð4
ffiffi
3
p
L
g
Þ,is 1:8 10
3
, and may be
compared with the homogeneous one-dimensional Pierce
coefficient estimated from the SPARC parameters, which
is 3 10
3
. According to numerical simulations,
saturation is expected at a pulse energy 0.1 mJ. The maxi-
mum energy collected was about 0.01 mJ, corresponding to
about two gain lengths below full saturation. The radiation
parameters after optimization are summarized in the histo-
grams in Fig. 12, where the statistics of radiation energy,
relative spectral linewidth, central wavelength, and rms
spot size are shown.
The pulse energy on the fundamental and third harmonic
166 nm has been compared with the predictions of
four simulation codes:
PERSEO [22] (red), GENESIS 1.3 [ 23]
FIG. 10. Calculated evolution of the
x
(continuous red) and
y
(dotted blue) Twiss coefficients in the periodic FODO of
length L
p
¼ 256 cm composed by an undulator section of length
L
u
¼ 215:6cmand a quadrupole (Q) in the drift between two
adjacent modules.
FIG. 11. Pulse energy vs longitudinal position in the undulator.
Black symbols represent the measured data on the fundamental.
Set (A) circles (average over 20 shots after the first undulator,
z 5m), before optimization, set (B) squares (average over 50
shots), after optimizing laser to injector phase. Third harmonic
(3h) triangles, were acquired with all the undulator gaps at
resonance. Error bars represent 2 standard deviations of the
shot to shot fluctuations (error bar is missing after the first
section where the spectrum is the result of two integrations
over 100 s). The lines represent simulations with
PERSEO (red),
GENESIS 1.3 (brown), MEDUSA (blue), and GINGER (green). The
third harmonic simulation data are represented by dash-dotted
lines.
SELF-AMPLIFIED SPONTANEOUS EMISSION FOR A ... Phys. Rev. ST Accel. Beams 14, 060712 (2011)
060712-5
(brown), MEDUSA [24,25] (blue), and GINGER [26] (green)
(see Fig. 11). All the calculations have been performed by
assuming a beam with the longitudinal phase space corre-
sponding to the measured data in Figs. 5 and 6, and with
transverse emittance given by the quadrupole scan mea-
surement (2:5=2:9mm-mrad). The code
PERSEO, which is
a time dependent code including three-dimensional effects
via a correction of the FEL coupling coefficients, has been
run assuming a rms beam radius
r
¼ 120 m. In the
cases simulated with all the other codes, the beam has
been supposed matched to the undulator, with the Twiss
parameter h
x
i¼h
y
1:5m. GENESIS 1.3 and GINGER
integrate three-dimensional particle equations of motion
averaged over the undulator period, while
MEDUSA is based
on nonaveraged equations. All codes employ the slowly
varying envelope approximation to Maxwell’s equations.
Notwithstanding the differences, all codes predict a similar
gain length, and the energy values during the growth are
comparable downstream from the third section. The final
energy is close to the experimental value measured in the
optimized case. The reason for the discrepancy after
the first undulator sections may be explained considering
the differences between the codes in the transverse mode
representation of the field. In this regard,
GENESIS 1.3 uses a
Cartesian grid in ðx; yÞ, in which charge is mapped to the
nearest-neighbor grid point.
GINGER uses an axisymmetric
grid in which charge is mapped to the nearest radial grid
cells on a proportional basis and
MEDUSA uses a modal
decomposition for the field rather than a grid-based repre-
sentation. Finally,
PERSEO simulates a single transverse
mode. As pointed out in a previous comparison between
GINGER and GENESIS 1.3 [27], a Cartesian grid supports a
much larger number high-order modes (HOMs) than is true
for the other representations. These HOMs are over-
whelmed by the gain in the low-order modes and become
a progressively smaller component of the radiation as the
interaction proceeds into the exponential regime and then
to saturation. A similar argument may justify the differ-
ences on the third harmonics, where simulations made with
the
GENESIS 1.3 and GINGER codes show higher initial shot
noise. In this regard the specific shot noise algorithm used
may also introduce a difference [26,28]. Analytical esti-
mates based on the formulas reported in Ref. [29], ex-
ploited to make the comparison with the experimental
results, reproduce the full power curve vs z of either
fundamental and third harmonics with a fairly satisfactory
agreement, as reported in Ref. [30].
The optimization consisted in a fine-tuning of the laser to
rf phase that probably compensated a phase drift which
occurred after the initial beam characterization. Such a
phase drift may have affected the matching condition as
well as the beam emittances and correlated energy spread.
We studied this effect by running
PERSEO with different
conditions of matching and emittances. The simulation (A)
in Fig. 13 represents the matched nominal case as in
Fig. 11. Simulation (B) has been obtained by increasing
the beam rms radius by 30%, corresponding to
x
¼
150 m. This simulation fits quite accurately the experi-
mental data relevant to the not optimized sequence along
the undulator. The simulation (C) corresponds to a case
with nominal matching and with a reduced emittance of
2.0 mm-mrad in both planes. The slice emittance, which is
the parameter affecting the FEL gain, could be lower than
the value measured with the quadrupole scan and listed
in Table I. In this case the effective beam size would be
FIG. 13. Comparison between the experimental data on the
fundamental harmonic (as in Fig. 11) and simulations done with
PERSEO in different conditions of matching and emittances: (A)
emittances (x=y)(2:5=2:9mm-mrad) and
r
¼ 120 m; (B)
emittances (x=y)(2:5=2:9mm-mrad) and
r
¼ 150 m;
(C) emittances (x=y)(2:0=2:0mm-mrad) and
r
¼ 100 m;
(D) emittances (x=y)(2:0=2:0mm-mrad) and
r
¼ 150 m.
FIG. 12. Histograms of radiation pulse energy, wavelength,
linewidth, and spot size at the spectrometer slit for a set of 50
events corresponding to the highest energy measured, set (B) in
Fig. 11.
L. GIANNESSI et al. Phys. Rev. ST Accel. Beams 14, 060712 (2011)
060712-6
x
¼ 100 m and the corresponding gain length is shorter.
In this regime the gain length reduction is however mainly
due to the growth of current density, rather than to the
reduction of inhomogeneous broadening associated to the
emittances. Simulation (D) in Fig. 13 has been obtained
with the same emittance as in case (C), but with the same
cross section as in (B) and fits with a similar accuracy the
experimental data. We also simulated a mismatch of the
beam through the undulator line in
MEDUSA by making
small changes in the Twiss parameters. A change in the
Twiss parameters that results in an increase in the average
beam envelope equal to 150 m, yields with
MEDUSA
simulation results that are quite close to the experimental
data.
Figure 14 shows the behavior of the radiation linewidth
as a function of the longitudinal position in the undulator.
The spectra are measured about two meters after the last
undulator. The geometry of the vacuum chamber and the
transport line to the spectrometer selects the low diver-
gence portion of the radiation field, affecting both the
measured energy and linewidth.
The lines represent predictions by
PERSEO, GENESIS 1.3,
MEDUSA, and GINGER. The GENESIS 1.3 calculation of the
spectrum shown in the figure is given by the field at the
coordinate z, propagated in the far field.
A view of the evolution of the spectrum during the
exponential growth is given in Fig. 15. The spectra were
acquired during a different shift than that of the previous
figures, but with very similar beam parameters to those
listed in Table I. The picture represents a set of six spectra
obtained by progressively suppressing the amplification
process in the first part of the undulator, by detuning the
resonance in selected sections. This procedure allowed the
measure of the spectrum generated in the remaining
modules, with an increasing number of undulators partic-
ipating in amplification, while keeping unchanged the
geometry of the radiation detection. The vertical axis in
each picture indicates the position on the vertical entrance
slit of the spectrometer. The number on the upper left
corner represents the number of active undulator sections.
The intensities of the different images are normalized to
the peak value. The upper spectrum in Fig. 15, obtained
with only the last section being resonant, is the result of an
integration over 100 shots. The energy from (the last) 2–6
undulator sections allowed acquisition in single shot mode.
The spiky nature of the SASE radiation is already apparent
in the spectrum obtained with only two undulator sections.
V. CONCLUSIONS
In this paper we report for the first time the lasing
performance obtained in SASE mode at SPARC. For a
peak current of about 53 A at 500 nm wavelength, we
observed an overall amplification factor close to 10
7
, with
an estimated gain length of 0.7 m. The maximum energy
collected was about 0.01 mJ. Detailed spectra measure-
ments were done by progressively turning off the FEL
interaction in selected undulators. The evolution of the
spectrum in exponential gain regime exhibits the spiky
5
FIG. 14. Experimental linewidth of the FEL radiation (black
circles) as a function of the longitudinal position in the undu-
lator. The lines represent simulation data obtained with
PERSEO
(red), GENESIS 1.3 (brown), MEDUSA (blue), GINGER (green). The
error bars represent 2 standard deviation of the measured
linewidth evaluated over 50 samples.
FIG. 15. Evolution of the spectrum during the exponential gain
growth. The number of sections closed is reported on the upper
left corner. The vertical axis in each picture indicates the position
on the vertical entrance slit of the spectrometer.
SELF-AMPLIFIED SPONTANEOUS EMISSION FOR A ... Phys. Rev. ST Accel. Beams 14, 060712 (2011)
060712-7
nature of SASE, with about 20=40 longitudinal modes
detected in the frequency spectrum after six undulator
sections. Simulations made with the
PERSEO, GENESIS 1.3,
MEDUSA, and GINGER codes have been compared to the
experimental data both at the fundamental and third har-
monic frequency. While the agreement between predicted
and observed data at the fundamental is remarkable, a
fairly reasonable agreement has been also observed at the
third harmonics measured at the end of the undulator.
ACKNOWLEDGMENTS
We are grateful to P. Musumeci for his contribution
in the project design and in the commissioning of the
SPARC photoinjector. This work has been supported by
Ministero dell’Istruzione, dell’Universita
`
e della Ricerca
(MIUR - DM1834 RIC.4-12-2002).
[1] A. M. Kondratenko and E. L. Saldin, Part. Accel. 10, 207
(1980).
[2] H. Haus, IEEE J. Quantum Electron. 17, 1427 (1981).
[3] G. Dattoli, A. Marino, A. Renieri, and F. Romanelli, IEEE
J. Quantum Electron. 17, 1371 (1981).
[4] R. Bonifacio, C. Pellegrini, and L. M. Narducci, Opt.
Commun. 50, 373 (1984).
[5] K. J. Kim, Nucl. Instrum. Methods Phys. Res., Sect. A
250, 396 (1986).
[6] S. Milton et al., Science 292, 2037 (2001).
[7] V. Ayvazian et al., Eur. Phys. J. D 20, 149 (2002).
[8] A. Murokh et al., Phys. Rev. E 67, 066501 (2003).
[9] V. Ayvazyan et al., Phys. Rev. Lett. 88, 104802
(2002).
[10] T. Shintake et al., Phys. Rev. ST Accel. Beams 12, 070701
(2009).
[11] P. Emma et al., Nat. Photon. 176, 1038 (2010).
[12] M. Ferrario et al., Phys. Rev. Lett. 99, 234801 (2007).
[13] L. Poletto et al., Rev. Sci. Instrum. 72, 2868
(2001).
[14] L. Poletto et al., Rev. Sci. Instrum. 75, 4413 (2004).
[15] L. Serafini and M. Ferrario, AIP Conf. Proc. 581,87
(2001).
[16] M. Ferrario et al., Phys. Rev. Lett. 104
, 054801
(2010).
[17] L. Serafini and J. B. Rosenzweig, Phys. Rev. E 55, 7565
(1997).
[18] P. Piot, L. Carr, W. S. Graves, and H. Loos, Phys. Rev. ST
Accel. Beams 6, 033503 (2003).
[19] D. Alesini et al., in Proceedings of DIPAC 2009,
TUOA01, p. 146.
[20] L. Ficcadenti et al.,inProceedings of the 16th IEEE
International Pulsed Power Conference, Albuquerque,
NM (IEEE, Piscataway, NJ, 2007), p. FRPMN030.
[21] J. R. Pierce, Traveling-Wave Tubes (Van Nostrand
Company, New York, 1950).
[22] L. Giannessi, in Proceeding of the 2006 FEL Conference
(2006), p. MOPPH026 [http://www.jacow.org].
[23] S. Reiche, Nucl. Instrum. Methods Phys. Res., Sect. A
429, 243 (1999).
[24] H. P. Freund, Phys. Rev. ST Accel. Beams 8, 110701
(2005).
[25] H. P. Freund, S. G. Biedron, and S. V. Milton, IEEE J.
Quantum Electron. 36, 275 (2000).
[26] W. Fawley, Phys. Rev. ST Accel. Beams 5, 070701
(2002).
[27] Z. Huang and K. J. Kim, Phys. Rev. ST Accel. Beams 10,
034801 (2007).
[28] H. P. Freund, L. Giannessi, and W. H. Miner, Jr., J. Appl.
Phys. 104, 123114 (2008).
[29] G. Dattoli, P. L. Ottaviani, and S. Pagnutti, Technical
Report Booklet for FEL Design (ENEA Edizioni
Scientifiche Frascati, 2007) (unpublished).
[30] G. Dattoli, P. Ottaviani, and S. Pagnutti, Technical Report
No. SPARC FEL 09 006, 2009.
L. GIANNESSI et al. Phys. Rev. ST Accel. Beams 14, 060712 (2011)
060712-8
    • "Despite the FERMI peak current and brightness at the first glance not appearing to be highly challenging, the simultaneous requirement at the undulators of low longitudinal and transverse emittance, and a nearly constant ∼ kA current over a few hundreds of fs is in fact quite demanding. In table 1 we compare the present beam operational parameters of – 2 – the FERMI photoinjector with those of other FEL photoinjectors, including LCLS at SLAC [14], PSI [15], FLASH [16] [17], SPARC [18] [19], and the X-FEL Photo Injector Test facility at DESY, Zeuthen site (PITZ) [20] [21]. At PITZ they have studied several bunch charge configurations, including the 500 pC regime of interest to FERMI. "
    [Show abstract] [Hide abstract] ABSTRACT: The FERMI@Elettra project is a seeded free electron laser (FEL) source, based on the High Gain Harmonic Generation (HGHG) scheme. It is designed to supply photons in a spectral range from 65 to 20 nm with the first undulator line (FEL-1) and from 20 nm to 4 nm with the second undulator line (FEL-2). After a first period of commissioning of the electron beam up to 100 MeV at low charge, started in August 2009, several phases of installations and beam commissioning periods have been alternated through the 2010. On December 2010 the first FEL light in the FEL-1 line was obtained at 65 nm and 43 nm by adopting as an initial seed the third harmonic of a TiSa laser. The complete optimization and commissioning of the photo-injector has been carried on in parallel with the Linac and FEL commissioning, from a conservative set-up to the final designed configuration. This paper reports the electron beam characterization in the injector area, the comparison with the theoretical expectations and the experimental process which resulted in a high brightness electron beam. This beam was optimized to be compressed and then transported through the undulators of FEL-1 where intense photons ranging from 65 nm to 20 nm were generated [1,2].
    Full-text · Article · May 2013
  • [Show abstract] [Hide abstract] ABSTRACT: The injection of a seed in a free-electron laser (FEL) amplifier reduces the saturation length and improves the longitudinal coherence. A cascaded FEL, operating in the high-gain harmonic-generation regime, allows us to extend the beneficial effects of the seed to shorter wavelengths. We report on the first operation of a high-gain harmonic-generation free-electron laser, seeded with harmonics generated in gas. The third harmonics of a Ti:sapphire laser, generated in a gas cell, has been amplified and up-converted to its second harmonic (λ(rad)=133 nm) in a FEL cascaded configuration based on a variable number of modulators and radiators. We studied the transition between coherent harmonic generation and superradiant regime, optimizing the laser performances with respect to the number of modulators and radiators.
    Article · Nov 2011
  • [Show abstract] [Hide abstract] ABSTRACT: The effect of partial prebunching of an electron beam on the saturation length, efficiency, and longitudinal coherence is studied for a nominal x-ray free-electron laser design. Results indicate that partial prebunching of as little as 1% results in substantial (1) shortening of the saturation length, (2) increases in the efficiency, and (3) improvement in longitudinal coherence.
    Full-text · Article · Mar 2012
Show more

    Recommended publications

    Discover more