Advanced Intraundulator Electron Beam
Diagnostics Using COTR Techniques*
A. H. Lumpkin, W. J. Berg, S. Biedron, M. Borland, Y. C. Chae,
R. Dejus, M. Erdmann, Z. Huang, K.-J. Kim, J. Lewellen, Y. Li,
S. V. Milton, E. Moog, D. W. Rule1", V. Sajaev, and B. X. Yang
Advanced Photon Source, Argonne National Laboratory, Argonne, IL 60439
Abstract. A significant advance in intraundulator electron-beam diagnostics has recently been
demonstrated based on coherent optical transition radiation (COTR) imaging. We find signal
strengths from a microbunched beam in a UV-visible free-electron laser to be several orders of
magnitude higher than that of incoherent optical transition radiation. In addition we report that
the far-field images of COTR interferograms carry information about beam size and asymmetry,
divergence, and pointing.
INTRODUCTION
One of the standard means of imaging electron beams is the use of optical transition
radiation (OTR) as the conversion mechanism [1-4]. Its inherently good spatial
resolution and ultrafast time response must be balanced with its lower conversion
efficiency in some applications with bright beams. However, in the case of
microbunched electron beams, we report a significant advance in intraundulator
diagnostics based on the signal enhancements and structure observed in coherent
optical transition radiation (COTR). Longitudinal microbunching of the electron beam
occurs as it copropagates with the emitted synchrotron radiation within the undulator.
The density modulation develops at the fundamental wavelength of the light, which
leads to a growing fraction of electrons emitting in phase, or coherently. A favorable
instability results in an exponential growth of the light generation in a self-amplified
spontaneous emission (SASE) free-electron laser (PEL) [5-8]. The signal strength
now goes as the square of the number of particles involved (bnN)2, where bn is the
amplitude of the Fourier component of the electron distribution with spatial frequency
kn, and N is the total number of particles. We have routinely had to use neutral density
(ND) filters providing attenuation factors of 104 to 105 (!) for COTR from 200-pC
electron beams that have been microbunched. This would be an unheard of scenario
with OTR imaging, and the light yield exceeds most scintillators.
In the case of the Advanced Photon Source (APS) PEL operating in the UV-visible
regime, we have used standard CCD imaging cameras to obtain near-field, far-field,
and spectral information on the COTR [9-12]. In addition to the unprecedented signal
* Work supported by U.S. Department of Energy, Office of Basic Energy Sciences under Contract No. W-31-109-ENG-38.
f
Carderock Division, NSWC, West Bethesda, MD 20817.
CP648, Beam Instrumentation Workshop 2002: Tenth Workshop, edited by G. A. Smith and T. Russo
2002 American Institute of Physics 0-7354-0103-9
212
strength in the near-field images (used to measure beam size), we have found
strength instructure
the near-field
images
(used to measure beam size), we have found
noticeable
and 0x-6
y asymmetry in the far-field images plus narrow-band
noticeable
structure In
andour
θx-θtwo-foil
in thewe
far-field
images
plus interferometer
narrow-band
y asymmetry
spectral
emission.
geometry,
see clear
COTR
spectralthat
emission.
In ourbytwo-foil
geometry,
we see
COTR
interferometer
images
are explained
the product
of a bunch
formclear
factor
(developed
from the
imagessize)
that and
are explained
the product
of a bunch
form factor
from
beam
the singlebyelectron
interference
pattern
[13]. (developed
The spacings
ofthe
the
beam
size)
and
the
single
electron
interference
pattern
[13].
The
spacings
of
interference fringe peaks act like an internal "measurement grid" for beam sizethe
(as
interference
fringe
peaks act via
likefringe
an internal
“measurement
grid”
for beam size
well
as carrying
information
visibilities
about beam
divergence).
For(as
our
well aswecarrying
fringe
Forfringe
our
optics
actuallyinformation
infer bettervia
beam
sizevisibilities
sensitivityabout
(a~30beam
jim) divergence).
from the COTR
optics
we
actually
infer
better
beam
size
sensitivity
(σ~30
µm)
from
the
COTR
fringe
visibility than from the beam size image directly! A brief description of our
visibility than
beam
size will
image
directly! A brief description of our
experiments
andfrom
somethe
results
to date
be presented.
experiments and some results to date will be presented.
EXPERIMENTAL BACKGROUND
EXPERIMENTAL BACKGROUND
These experiments were performed at the APS using beam accelerated to 211 MeV
These experiments were performed at the APS using beam accelerated to 217 MeV
by the S-band linac normally used as part of the injector system for the 7-GeV storage
by the S-band linac normally used as part of the injector system for the 7-GeV storage
ring. We have obtained data with both the thermionic rf gun, with an 8-ns-long
ring. We have obtained data with both the thermionic rf gun, with an 8-ns-long
macropulse,
and the photocathode (PC) rf gun, which generates a single micropulse at
macropulse, and the photocathode (PC) rf gun, which generates a single micropulse at
66 Hz.
The
guns
and linac
linac are
are described
described elsewhere.
elsewhere. The
Thebeam
beamisistransported
transportedtoto
Hz. The guns [14,15]
[14,15] and
the
set
of
undulators
(maximum
of
nine)
in
the
low-energy
undulator
testline
line
the set of undulators (maximum of nine) in the low-energy undulator test
(LEUTL)
tunnel
[16].
A
schematic
of
the
experiment
is
given
in
Figure
1.
It
is
not
(LEUTL) tunnel [16]. A schematic of the experiment is given in Figure 1. It is not toto
scale,
40-m transport
transport line
line between
between the
thethree-screen
three-screen
scale, and
and there
there is
is an
an approximately
approximately 40-m
emittance
station
and
the
entrance
of
the
undulators.
emittance station and the entrance of the undulators.
When
are installed,
installed, the
the magnetic
magneticstructure
structurelength
lengthisis21.6
21.6m.
m.The
The
When all
all nine
nine undulators
undulators are
properties
of
the
undulators
and
diagnostic
stations
have
been
previously
described
properties of the undulators and diagnostic stations have been previously described
[17].
cells have
have aa period
period of
of 3.3
3.3 cm
cmand
andlength
lengthL=2.4
L=2.4m.
m.
[17]. Very
Very briefly,
briefly, the
the undulator
undulator cells
Visible Light Diagnostics (VLD)
——— for UR and CTR
UV DriveLaser
Electron
Spectrometer
Dipole
ORIEL
UV-Visible
Spectrometer
Beam
Dump
Spectrometer
3-Screen
Emittance
Measurement
-Streak
Camera
Measurement
CDR
Monitor
FIGURE 1.
1. A
A schematic
schematic of
FIGURE
of the
the APS
APS SASE
SASE FEL
PEL and
and the
the ten
ten intraundulator
intraundulatordiagnostic
diagnosticlocations.
locations.
213
They
They have
have aa fixed
fixed gap
gap with
with aa field
field parameter
parameter K=3.1.
K=3.1. There
There isis approximately
approximately aa 0.38-m
0.38-m
space
space after
after each
each undulator.
undulator. This
This space
space isis used
used for
for diagnostics
diagnostics and
and focusing
focusing and
and
steering
steering elements
elements before
before the
the first
first undulator
undulator and
and after
after each
each of
of the
the installed
installed undulator
undulator
sections.
sections. A
A schematic
schematic of
of these
these stations
stations isis shown
shown in
in Figure
Figure 2.
2. The
The screens
screens on
on the
the first
first
actuator
actuator include
include positions
positions for
for aa YAG:Ce/mirror,
YAGiCe/mirror, aa mirror
mirror at
at 45°,
45°, and
and aa thin
thin (6
(6 µm)
jim) Al
Al
foil
foil mounted
mounted with
with its
its surface
surface normal
normal to
to the
the beam
beam direction.
direction. This
This thin
thin foil
foil serves
serves two
two
functions:
functions: (1)
(1) to
to block
block the
the stronger,
stronger, visible
visible undulator
undulator radiation
radiation (UR)
(UR) and
and (2)
(2) to
to
generate
generate OTR
OTR or
or COTR
COTR as
as the
the e-beam
e-beam transits
transits the
the foil/vacuum
foil/vacuum interfaces.
interfaces. A
A digital
digital
camera
camera views
views the
the e-beam
e-beam images
images from
from the
the YAG
YAG and
and the
the reflected
reflected undulator
undulator radiation
radiation
from
from the
the mirror.
mirror. The
The second
second actuator,
actuator, located
located 63
63 mm
mm downstream,
downstream, involves
involves aa
retractable
retractable mirror
mirror at
at 45°
45° to
to the
the beam
beam direction.
direction. Another
Another digital
digital camera
camera and
and moveable
moveable
lens
provide
both
near-field
and
far-field
(focus
at
infinity)
imaging.
This
lens provide both near-field and far-field (focus at infinity) imaging. This visible
visible light
light
detector
detector (VLD)
(VLD) system
system thus
thus provides
provides both
both beam
beam size
size and
and angular
angular distribution
distribution data.
data.
Both
Both neutral
neutral density
density (ND)
(ND) filters
filters and
and bandpass
bandpass (BP)
(BP) filters
filters are
are selectable
selectable by
by up
up to
to
three
three filter
filter wheels
wheels in
in front
front of
of the
the cameras.
cameras.
As
As indicated
indicated in
in Figure
Figure 2,
2, aa remotely
remotely controlled
controlled pick-off
pick-off mirror
mirror and
and lens
lens system
system can
can
be
be used
used to
to redirect
redirect the
the UR
UR or
or COTR
COTR to
to an
an Oriel
Oriel UV-visible
UV-visible spectrometer.
spectrometer. Spectral
Spectral
effects
effects were
were observed
observed including
including the
the onset
onset of
of sideband
sideband production
production after
after FEL
PEL saturation.
saturation.
For
For the
the purposes
purposes here,
here, itit was
was important
important to
to verify
verify the
the COTR
COTR narrow-band
narrow-band spectrum
spectrum
centered
centered around
around the
the fundamental
fundamental SASE
SASE wavelength.
wavelength. Chromatic
Chromatic effects
effects in
in spatial
spatial
focus
focus or
or the
the time-domain
time-domain are
areavoided
avoidedwith
withCOTR.
COTR.
CCD Imager
BP Filter
ND Filter
To UV-Visible
Spectrometer
Diagnostic Chamber
"Button" BPM—
Mirror
Retractable Mirror
CCD Imager
Lens
BP Filter
ND Filter
YAG Crystal/Mirror/OTR Foil
Secondary Emission Monitor
Quadrupole x/y Corrector.
Undulator-^
Retractable Mirror
Undulator
FIGURE
FIGURE 2.
2. A
A detailed
detailed schematic
schematic of
of the
the intraundulator
intraundulator diagnostic
diagnostic stations
stations showing
showing the
the YAG
YAG
actuator/cameras
actuator/cameras and
and the
the downstream
downstream 45°
45° pick-off
pick-off mirror
mirrorwith
withthe
thevisible
visiblelight
light detector
detector (VLD)
(VLD) cameras.
cameras.
214
EXPERIMENTAL RESULTS AND DISCUSSION
EXPERIMENTAL RESULTS AND DISCUSSION
We now present examples of results from the near-field imaging, far-field imaging,
We now present examples of results from the near-field imaging, far-field imaging,
and the imaging spectrometer.
and the imaging spectrometer.
Near-Field
Near-Field Imaging
Imaging
The
beam size
size and
and profile
profile
The near-field
near-field focus
focus position
position of
of the
the lens
lens provides
provides beam
information.
In
our
geometry
we
can
completely
eliminate
the
competition
from
the
information. In our geometry we can completely eliminate the competition from the
dominant
SASE
light
by
using
the
blocking
foil
at
the
upstream
position
from
the
45°
dominant SASE light by using the blocking foil at the upstream position from the 45°
pick-off
beam sizes
are shown
shown in
in
pick-off mirrors
mirrorsatateach
eachstation.
station. Examples
Examples of
of the
the z-dependent
z-dependent beam
sizes are
Figure
3.
In
this
case
the
FEL
fundamental
was
at
530
nm.
We
note
that
the
needed
Figure 3. In this case the PEL fundamental was at 530 nm. We note that the needed
ND
variation in
in the
the
NDvalue
valuechanged
changed from
from 00 atat VLD1
VLD1 to
to 4.5
4.5 at
at VLD5,
VLD5, 6.
6. A
A significant
significant variation
observed
beam
sizes
is
observed,
in
contrast
to
the
expected
constant
beam
sizes
at
the
observed beam sizes is observed, in contrast to the expected constant beam sizes at the
sampling
points
for
a
well-matched
beam.
Possible
contributions
to
this
effect
are
sampling points for a well-matched beam. Possible contributions to this effect are
beam
due to
to COTR,
COTR, aa
beam match
match into
into the
the undulators,
undulators, beam
beam image
image size
size reductions
reductions due
transverse
dependence
of
the
bunching
fraction
coupled
with
the
COTR
effect,
transverse dependence of the bunching fraction coupled with the COTR effect,
chromatic
effects
on
lens
focus
for
broadband
OTR
versus
narrow-band
COTR,
and aa
chromatic effects on lens focus for broadband OTR versus narrow-band COTR, and
camera
we believe
believe the
the major
major
camera focus
focus error
error in
in the
the early
early stations.
stations. It
It is
is noted
noted that
that we
contributions
the VLDO
VLD0 and
and
contributionsare
arethe
the first
first three.
three. Since
Since we
we see
see small
small beam
beam structures
structures in
in the
VLD1
believe focus
focus effects
effects are
are
VLD1 cameras
cameras on
on some
some shots
shots out
out of
of the
the 100
100 images,
images, we
we believe
minimal.
minimal.
1400
1400
x-size
y-size
Beam Size FWHM (µm)
1200
1200 1000
1000 800
800 600
600 400
400 200 200
0
VLDO VLD1
VLD1 VLD2
VLD2 VLD3
VLD3 VLD4
VLD4 VLD5
VLD5 VLD6 VLD7 VLD8 VLD9
VLD0
Location
ZZ Location
FIGURE3.3. An
Anexample
exampleof
ofz-dependent,
z-dependent, x–
x- and
andy–beam
y-beam sizes
sizes observed
observed in
in the undulator areas. These
FIGURE
These
dataare
arefrom
fromMarch
March30,
30,2001.
2001.
data
215
We
the beam
beam spot
spot narrowing
narrowing that
that occurs
occurs when
when
We have
have performed
performed aa calculation
calculation of
of the
COTR
is
used
as
the
conversion
mechanism.
If
the
bunching
fraction
is
assumed
to be
COTR is used as the conversion mechanism. If the bunching fraction
2 is assumed to be
2 term for the peak
uniform
across
the
Gaussian
transverse
dimension
then
the
(b
N)
n
uniform across the Gaussian transverse dimension then the (bn N) term for the peak
intensity
of the
the “observed”
"observed" peak.
peak. We
We have
have
intensity will
will approximately
approximately give
give aa V2
√2 narrowing
narrowing of
experimentally
tested
this
simple
model
by
taking
data
without
(COTR)
and
with
experimentally tested this simple model by taking data without (COTR) and with
(OTR)
a
500-nm
short
pass
filter,
which
would
attenuate
by
100
all
wavelengths
(OTR) a 500-nm short pass filter, which would attenuate by 100 all wavelengths
greater
this effect
effect is
is shown
shown in
in Figure
Figure 4.4. The
The
greater than
than about
about 520
520 nm.
nm. An
An example
example of
of this
horizontal
gain has
has occurred
occurred starting
starting atat VLD5
VLD5 isis
horizontalbeam
beam size
size observed
observed when
when significant
significant gain
narrower
The incoherent
incoherent OTR
OTR source
source
narrower when
when the
the bandpass
bandpass filter
filter is
is not
not used
used (COTR).
(COTR). The
provides
at VLD8
VLD8 the
the single
single filter
filter is
is insufficient
insufficient
providesthe
theactual
actual e-beam
e-beam size.
size. We
We note
note that
that at
totoblock
so clean
clean OTR
OTR imaging
imaging isis not
notachieved
achieved
blockthe
theentire
entire strong
strong microbunching
microbunching signal,
signal, so
for
this
point.
for this point.
1000
1000 •
OTR- X
COTR-X
Observed FWHM (µm)
800
800 -
600
600 -
400
400 -
200 200
0
0
22
4
4
6I
8
10
10
Location (VLD#)
ZZ Location
(VLD#)
FIGURE4.4. An
Anexample
example of
of z-dependent
z-dependent x-beam
x–beam sizes
sizes observed
FIGURE
observed with
with (OTR)
(OTR) and
and without
without (COTR)
(COTR)the
the
500-nmshort
shortpass
passfilter
filter at
at VLD5
VLD5 -– VLD8.
VLD8. Data
Data are
500-nm
are from
from December
December 20,
20,2001.
2001.
Far-Field Imaging
Imaging
Far-Field
The camera-to-lens
camera-to-lens zz separation
separation can
can be
be remotely
remotely controlled
The
controlled by
by the
the stepper
stepper motor.
motor.
For far-field
far-field imaging
imaging angular
angular distribution
distribution patterns
patterns are
For
are obtained.
obtained. Based
Based on
on the
the
analytical model
model described
described in
in Ref.
Ref. 13,
the single-electron
analytical
13, the
single-electron OTR
OTR interference
interference (OTRI)
(OTRI)
pattern for
for two
two sources
sources 63
63 mm
mm apart
apart is
is first
first calculated.
calculated. This
This includes
pattern
includes effects
effects due
due to
to
beamdivergence.
divergence. For
For example,
example, we
we estimate
estimate the
the beam
beam scattering
beam
scattering from
from the
the first
first foil
foil as
as
about 0.1
0.1 to
to 0.2
0.2 mrad
mrad at
at 220
220 MeV.
MeV. The
The calculated
calculated OTRI
about
OTRI pattern
pattern exhibits
exhibits fringes
fringes atat
± 1.9, ±± 4.6,
4.6, ±± 6.3,
6.3, and
and ±± 7.5
7.5 mrad.
mrad. The
The bunch
bunch form
±1.9,
form factor
factor isis multiplied
multiplied times
timesthis
this
pattern.
Large
beam
(>
100
µm)
sizes
have
very
narrow
form
factor
functions
pattern. Large beam (> 100 jim) sizes have very narrow form factor functions ininθ0
space. Figure
Figure 55 shows
shows an
an example
example of
of the
the COTR
COTR fringe
space.
fringe visibility
visibility variation
variation with
with beam
beam
size.
The
COTR
second
fringe
peaks
are
visible
for
beam
sizes
less
than
50
size. The COTR second fringe peaks are visible for beam sizes less than 50 µm.
jim. The
The
observed first
first fringe
fringe peak
peak position
position varies
varies rapidly
rapidly for
> 100
observed
for aσ >
100 µm.
jim. An
An example
example of
of an
an
216
image is given in Figure 6. In this case the initial beam size asymmetry in σx and σy
imageinisthe
given
in Figure
6. θInand
thismultiple
case thepeaks
initialinbeam
sizelooking
asymmetry
<jx and
ay
θy. By
at theinfringe
peak
results
single
peaks in
x
results
in
the
single
peaks
in
0
and
multiple
peaks
in
0
.
By
looking
at
the
fringe
peak
X
relative intensities for σy = 15,
20, and 30 µm, they θy pattern was found to be
relative
intensities
for
<j
=
15,
20,and
andtotal
30 divergences
jim, the 0yofpattern
found The
to be
y
consistent with a beam size of 30 µm
about was
0.2 mrad.
σy
consistent
with
a
beam
size
of
30
jim
and
total
divergences
of
about
0.2
mrad.
The
ay
= 30 µm result is smaller than the limiting resolution in the optical system of about
= 30 jim result is smaller than the limiting resolution in the optical system of about
one pixel at 80 µm/pixel for near-field focus. Further studies are needed in this area to
one
pixel
at 80 jim/pixel
near-field
Furtherwe
studies
needed elsewhere
in this areathe
to
develop
accuracy
on the for
beam
size. focus.
In addition,
haveare
reported
develop accuracy on the beam size. In addition, we have reported elsewhere the
sensitivity of the symmetry of the ± θy peaks to e-beam steering with the correctors
sensitivity of the symmetry of the ± 0y peaks to e-beam steering with the correctors
[18].
[18].
1 .0
Relative Intensity (arb. units)
25 µ m
C
0 .8
50 µ m
0 .6
0 .4
0.4
100 µ m
incoherent
_ case
incoheren
case
0 .2
0 .0
-0.01 0
-0.00 5
0 .0 00
0 .0 05
0 .0 10
AAngle
ngle (radians)
(radians)
FIGURE
beam sizes,
sizes, aσyy == 25,
25, 50,
50,
FIGURE5.5.AAcomparison
comparisonofofCOTR
COTRinterference
interference fringe
fringe visibility
visibility for
for different
different beam
and
and100
100µm.
urn.
0
5
6X (mrad)
FIGURE6.6.An
Anexample
exampleofofaaCOTR
COTRinterference
interference image
image taken
taken after
after undulator
undulator 88 in
FIGURE
in aa case
case when
when this
this was
was
post-FELsaturation.
saturation. The
Theclear
clearθ0x-θ
asymmetry
and
the
fringe
visibility
support
corresponding
beam
x-6yasymmetry
post-FEL
and
the
fringe
visibility
support
corresponding
beam
y
sizesofofσax≈~100
100µm,
um,and
andσay≈~30
30µm,
um,respectively.
respectively.
sizes
x
y
217
Imaging Spectrometer
The expected microbunching at the fundamental wavelength is clearly illustrated in
a spectrum from one of our early studies [19]. As shown in Figure 7, the SASE PEL
peak and the COTR peak overlap in wavelength at 530 nm. The COTR is a little
broader in this example taken after undulator 5. The electron beam energy jitter can
be observed directly in the wavelength jitter of the images. As a reference the GreNe
calibration line is shown at 543.5 nm, and its width indicates operational resolution of
about 0.8 nm (a) for these conditions.
1 CTR
520 530 540 550 560 570 580
Lambda (nm)
FIGURE 7. A comparison of the SASE PEL spectrum, the COTR spectrum, and the calibration laser
signal obtained after undulator 5. The discrete COTR line is quite different from the broadband OTR
spectrum.
SUMMARY
In summary, the COTR imaging techniques provide all the advantages of OTR
imaging with the additional features of signal enhancements up to 105 and
monochromatic spectral output. These features provide a near-ideal, intraundulator
electron-beam characterization capability. We hope to extend the techniques to the
VUV in the coming year as the APS PEL project pushes to operate at 130 nm.
218
ACKNOWLEDGEMENTS
The authors acknowledge the support of Om Singh, Antanas Rauchas, Efim
Gluskin, and Rodney Gerig of the APS.
REFERENCES
1. Wartski, L. et. al., J. Appl. Phys. 46, 3644 (1975).
2. Rule, D. W., Nucl. Instrum. Methods in Phys. Res. B24/25 901 (1987).
3. Lumpkin, A. H. et al., NIM A296, 150 (1990), D. W. Rule et al., Idem, p. 739 (1990).
4. Lumpkin, A. H. et al., Nucl. Instrum. Methods in Phys. Res. A429, 336 (1999).
5. Derbenev, Y. S., Kondratenko, A. M., and Saldin, E. L., Nucl. Instrum. Methods A193, 452 (1982).
6. Orzechowski, T. J. et al., Nucl. Instrum. Methods A250 144 (1986).
7. Hogan, M. et al., Phys. Rev. Lett. 81, 4897 (1998).
8. Milton, S. V. et al., Phys. Rev. Lett., 85 988 (2000).
9. Tremaine, A. et al., Phys. Rev. Lett. 81, 5816 (1998).
10. Lumpkin, A. H. et al., Phys. Rev. Lett. 86, 79 (2001).
11. Lumpkin, A. H. et al., "Evidence for Microbunching "Sidebands" in a Saturated PEL Using Coherent Optical
Transition Radiation," accepted in Phys. Rev. Lett. March 2002.
12. Lumpkin, A. H. et al., "Comprehensive z-Dependent Measurements of Electron-beam Microbunching using
COTR in a Saturated SASE PEL," Nucl. Instrum. Methods in Phys. Res. (in press).
13. Rule, D. W., and Lumpkin A. H., Proc. of the IEEE 2001 Particle Accelerator Conference, Chicago, IL, pp.
1288-1290 (2001).
14. Lewellen, J. W. et al., Proc. of the 1998 Linac Conference, Chicago, ANL-98/28, Vol. 2, pp. 863-865 (1999).
15. Biedron, S. et al., Proc. of the IEEE 1999 Particle Accelerator Conference, New York, NY, pp. 2024-2026
(1999).
16. Milton, S. V. et al., Nucl. Instrum. Methods in Phys. Res., A407, 210 (1998).
17. Gluskin, E. et al., Nucl. Instrum. Methods in Phys. Res., A429, 358 (1999).
18. Lumpkin, A. H. et al., Proc. of the 2001 Particle Accelerator Conference, Chicago, IL, pp. 550-552 (2001).
19. Lumpkin, A. H. et al., "Utilization of CTR to Measure the Evolution of Electron-Beam Microbunching in a
SASE PEL," Nucl. Instrum. Methods in Phys. Res. (in press).
219