Nuclear Instruments and Methods in Physics Research A 429 (1999) 209}212

Measured free-electron laser microbunching using

coherent transition radiation
A. Tremaine *, J. Rosenzweig , S. Anderson , P. Frigola , M. Hogan ,
A. Murokh , C. Pellegrini , D. Nguyen
, R. She$eld

University of California, Los Angeles, 405 Higard Avenue, Los Angeles, CA 90095, USA

Los Alamos National Laboratory, Los Alamos, NM 87545, USA

Abstract

The microbunch distribution of an electron beam exiting a SASE free-electron laser has been measured using the
emitted coherent transition radiation (CTR) produced from a thin aluminum foil placed at the end of the undulator. The
wavelength of the coherent transition radiation is shown to be the same as the FEL wavelength, and thus a measure of the
beam microbunch spacing. Also, the study of the CTR linewidth and angular acceptance of the radiation captured are
shown to be derived from this coherent radiative process. Scattering e!ects on the forward emitted transition radiation
from the electron beam traversing an aluminum foil are also considered.  1999 Elsevier Science B.V. All rights
reserved.

Keywords: CTR; FEL wavelength; Microbunching; Scattering e!ects

1. Introduction shown promise in the measurement of very

Longitudinal modulation of electron beams is
now generated with many types of devices such as
the free-electron laser (FEL) [1], inverse FELs [2],
and plasma and structure acceleration [3]. The
periodic beam modulation formed from these devi-
ces is at time scales where usual diagnostic methods
like streak cameras [4] and RF sweeping [5] can-
not be used. Instead of these time-domain methods,
frequency-domain measurements using coherent
transition radiation (CTR) from metallic foils have
* Corresponding author.
short electron pulses [6}8]. The high-gain pro-
cess of FELs is directly related to beam mico-
bunching and gives an opportunity to use CTR
to measure microbunch spacing down to several
femtoseconds.
2. Theoretical background
Describing the beam distribution exiting an FEL
has been done in Ref. [9], but in the present experi-
ment, asymmetries in transverse beam dimensions
were present at the exit of the wiggler and at the
CTR foil and must be included. Thus, the electron

0168-9002/99/$ - see front matter  1999 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 8 - 9 0 0 2 ( 9 9 ) 0 0 1 0 5 - 9 VI. SASE FELs
210 A. Tremaine et al. / Nuclear Instruments and Methods in Physics Research A 429 (1999) 209}212

beam distribution now looks like 3. Experimental setup


N x y z
f (r, z)" @ exp ! ! ! The present experiment was performed at Los
(2n)p p p 2p 2p 2p Alamos National Laboratory. A 1300 MHz RF
V W X V W X
photoinjector produces a 100 bunch train of low-

 
 emittance high-current bunches. Relevant beam
; 1# b sin(nk , z) (1)
L P parameters are given in Table 1. The wiggler used is
L
the permanent magnet 2 m Kurchatov undulator
where N is the number of electrons, k is the FEL with a period of 2.01 cm and is the same one used in
@ P
radiation wavenumber and thus the modulated the high-gain SASE FEL in Ref. [10]. For the
micro-bunch beam wavenumber, and b is the
L CTR experiment, a 6 lm foil was put on an insert-
bunching factor. Analysis of the CTR follows able mount which when inserted was 2 cm away
Ref. [8] where the energy spectrum is given by from the last period of the wiggler and normal to
the electron beam. Also, the foil covered the entire


d; d; exit aperture of the wiggler so when inserted, all the

NF (u)F (u, h)s(h) . (2)
du dX @ * 2 du dX FEL radiation would be re#ected back and the
  C\
Here F (u) and F (u, h) are the square of the longi-
* 2
tudinal and transverse beam pro"le Fourier trans-
forms. s(h) is a divergence factor which is usually
ignored in analysis, but will be shown to be very
important for forward emitted CTR. Putting Eq. (1)
into Eq. (2) and integrating, the CTR photon num-
ber angular dependence is found to be
only light to propagate down the optical beamline
would be the forward emitted CTR. The close prox-
imity of the foil to the end of the wiggler is very
important. First, there would be very little electron
beam debunching [11] from space-charge e!ects
and second, the source points and the optical
beamline for the FEL (foil extracted) and CTR (foil
inserted) radiation would be the same. A calibrated
HgCdTe detector was installed 3.5 m away from


dN a(N b ) sin(h)
!20" @ , *s(h)
the source point causing the angular acceptance of
dh 4(pnk p (1!b cos(h)) the optical beamline to be about 12 mrad. Mainly
P X
coherent radiation would be collected, since the
;exp[!(nk sin(h))(p sin(
) incoherent spectrum is peaked at a much larger
P V
angle, h +1/c+30 mrad.
#p cos(
))] (3) 
W For the present experiment, the system was run
at a charge of 1.5 nC and the conditions for the
where h and
are the polar and azimuthal angles high-gain SASE FEL in Ref. [9] were reestablished.
and a is the "ne structure constant. Most of the Once the maximum SASE signal was obtained and
CTR light is found in a small annular cone with the foil inserted, it was found that a minor RF
a maximum, using an axisymmetric beam of size p ,
P phase adjustment of 23 was necessary in order to
near h+1/(2nk p . Ignoring the divergence factor maximize the CTR signal (and thus the micro-
P P
and integrating Eq. (2), the number of photons can bunching of the beam) emanating from the foil. In
be found by the straightforward relation this case, adjustments to minimize the spot size
through RF focusing would enhance the CTR radi-



a(N b ) c  p#p ation as predicted by Eq. (4) and also negligibly
N " @ L V W . (4)
!20 4(pk p nk pp change the "nal beam energy. Simulations using
P X P V W
the 3-D FEL code Ginger were done for a series of
In order to best maximize the number of photons, experimental parameters giving a bunching factor
Eq. (4) shows the electron beam must be very dense
at the foil. It should also be noted that the above
relation is for normal incidence of the electron  SASE FEL gain in excess of 10 was "rst seen in this
beam on the foil. experiment.
 A. Tremaine et al. / Nuclear Instruments and Methods in Physics Research A 429 (1999) 209}212 211

Table 1
Electron beam and undulator experimental parameters

Parameter Value

Beam energy (E) 17.5 MeV

Peak current (I) 140 A
Charge/bunch (Q) 1.5 nC
Bunch length (q) 11 ps
(FWHM)
Energy spread (*c/c) 0.5%
Wiggler period (j ) 2.01 cm
5
On axis "eld (B ) 7.4 kG

FEL wavelength (j ) 13 lm
P
FEL parameter (o) 0.008
Rms beam sizes (p , p ) 210, 160 lm
V W

between 0.008 and 0.01 for the "rst harmonic and

negligible bunching for the higher harmonics. Fig. 1. CTR/SASE through monochromator versus wavelength.

12 to give it the same scale as SASE and the

4. Results resolution for the monochrometer was found to be
0.177 lm. As seen in the "gure, the signals are
In order to compare the measured photon num- nearly centered around the same wavelength
ber with that predicted in Eq. (4), the attenuation (13 lm) as expected. This shows that the micro-
factor s(h) introduced in Ref. [7] must now be bunching of the beam is at the SASE radiation
examined. When the radial scattering of the beam wavelength agreeing with theory.
through the foil is much less than the angular
spread of the incoherent radiation, p +h 1/c,
P   5. Conclusion
the attenuation can be ignored. When this condi-
tion is violated, s(h) becomes a complicated integral
The microbunching of an electron beam due to
and must be done numerically and quickly be-
the FEL gain process has been shown to have
comes less than unity. De"ning a degradation fac-
a spacing equivalent to that of the radiation
tor, g, it is found the forward emitted CTR signal
wavelength. Also, the narrow angular spectrum
through the 6 lm foil is reduced by a factor of 0.61.
and form of the spectral linewidth were measured
It should be noted that this experiment was "rst
and shown to correspond with the theoretical pre-
attempted with a 50 lm foil in which the CTR was
dictions. When doing forward scattered CTR ex-
greatly smaller than expected (g"0.11) leading to
periments, foil thickness and scattering e!ects must
the examination of scattering issues in the foil.
be carefully considered. This method of diagnosing
Using the predictions from Ginger and numerically
a microbunched electron beam is being planned
integrating Eq. (2), the range of expected photons is
for the Visible FEL experiment at Brookehaven
N "2.8;10}4.4;10. The calibrated measured
A National Lab in which the microbunch spacing and
photon number per pulse from the HgCdTe detector
radiation wavelength will be near 800 nm.
was 3.5;10. To within experimental and simula-
tion uncertainty, the numbers agree quite well.
Next the SASE and CTR signals were sent References
through a Jerrel Ash monochrometer. The results
are shown in Fig. 1. The SASE signal is attenuated [1] R. Bonifacio, C. Pellegrini, L. Narducci, Opt. Comm. 50
by a factor of 3 and the CTR signal is multiplied by (1984) 373.

VI. SASE FELs

212 A. Tremaine et al. / Nuclear Instruments and Methods in Physics Research A 429 (1999) 209}212

[2] A. Van Steenburgen et al., Phys. Rev. Lett. 80 (1998) 289.
[3] J. Rosenzweig et al., Phys. Rev. Lett. 74 (1995) 2467.
[4] J. Gardelle et al., Phys. Rev. Lett. 79 (1996) 3905.
[5] B. Carlsten, S.J. Russell, Phys. Rev. E 53 (1996) R2072.
[6] U. Happek, A.J. Sievers, E.B. Blum, Phys. Rev. Lett. 67
(1991) 2962.
[7] A. Murokh et al., Nucl. Instr. and Meth. A, submitted for
publication.
[8] Yukio Shibata et al., Phys. Rev. E 50 (1994) 1479.
[9] J.B. Rosenzweig, G. Travish, A. Tremaine, Nucl. Instr.
Meth. A 365 (1995) 255.
[10] M. Hogan et al., Phys. Rev. Lett., submitted for publica-
tion.
[11] J. Rosenzweig et al., Nucl. Instr. Meth. A 393 (1997)
376.