STELLA-II: Staged Monoenergetic Laser
Acceleration - Experiment Update
W. D. Kimura*, M. Babzien1", I. Ben-Zvi1", L. C. Campbell*, D. B. Cline11,
C. E. Dilley*, J. C. Gallardo1, S. C. Gottschalk*, K. P. Kusche*1,
R. H. Pantell*, I. V. Pogorelsky1, D. C. Quimby*, J. Skaritka1,
L. C. Steinhauer§, V. Yakimenko1, and F. Zhou11
*STI Optronics, Inc., 2755 Northup Way, Bellevue, WA 98004
^Brookhaven National Laboratory, Upton, NY 11973
^University of California, Los Angeles, Los Angeles, CA 90095
^Stanford University, Stanford, CA 94305
^University of Washington, Redmond Plasma Physics Laboratory, Redmond, WA 98052
Abstract. The goal of STELLA-II is to demonstrate staged monoenergetic laser acceleration of
microbunches using an inverse free electron laser (IFEL) buncher and IFEL accelerator. A key
feature of this experiment is the usage of a single high-intensity laser beam to simultaneously
drive both the buncher and accelerator. A chicane between the buncher and accelerator causes
microbunches to form at the entrance to the accelerator. All hardware has been installed at the
Accelerator Test Facility (ATF) located at Brookhaven National Laboratory, including a new
laser beam transport system to permit delivering higher laser power. Preliminary test results
indicate that modulation and acceleration of the microbunches are occurring with the new
system. Energy gains >13% have been observed. Current experiments are being conducted with
the ATF CO2 laser operating at a pulse length of -180 ps. In late autumn 2002, the ATF CO2
laser will be upgraded to produce pulse lengths of <10 ps at approximately the same output pulse
energy. This higher peak power will enable higher acceleration and more complete trapping of
the laser-generated microbunches in the accelerator. This higher acceleration and trapping will
also result in a clean separation of the accelerated microbunch electrons from the unaccelerated
ones while at the same time maintaining a narrow energy spread.
INTRODUCTION
The development of practical linacs based upon laser acceleration mechanisms will
require staging the process multiple times in order to obtain high net energy gain [1].
Moreover, it is critical during the staging process that the accelerated electrons remain
grouped tightly together as a microbunch(es) with narrow energy and phase spread.
The former attribute we refer to as being monoenergetic and the latter represents
maintaining a short bunch length in longitudinal space. Thus, useful staging requires
more than resynchronizing the microbunches with the accelerating wave in each stage;
it must also be done in a manner that does not degrade the microbunch qualities.
CP647, Advanced Accelerator Concepts: Tenth Workshop, edited by C. E. Clayton and P. Muggli
© 2002 American Institute of Physics 0-7354-0102-0/02/$19.00
269
CO2
LASER BEAM
BEAM!
CO
2 LASER
ADJUSTABLE
OPTICAL
DELAY
STAGE
DIPOLE
MAGNET
FOCUSING LENSES
Accelerator
(IFEL2)
Buncher
(IFEL1)
E-BEAM
FOCUSING
LENSES
VACUUM
PIPE
mnmm
E
-BEAM
E-BEAM
UNDULATOR
UNDULATOR
UNDULATOR
FOCUSING
MAGNET
MAGNET
MAGNET
LENSES
ARRAY
ARRAY
SPECTROMETER
VIDEO CAMERA
MIRROR WITH
CENTRAL HOLE
HOLE
MIRROR
MIRROR WITH
WITH
CENTRAL
CENTRAL HOLE
HOLE
EE-BEAM
-BEAM
D-
== QUADRUPOLE
QUADRUPOLE MAGNET
MAGNET
FIGURE 1. Schematic layout for the first STELLA experiment where staging
staging was first demonstrated.
demonstrated.
The Staged Electron Laser Acceleration (STELLA) experiment demonstrated the
basic staging process using two inverse free
free electron lasers (IFEL) [2]. A schematic
layout of this experiment is shown in Fig. 1.
1. The output from the ATF CO
CO2
2 laser is
split into two beams -– the first
first beam is focused into the first undulator (IFEL1), which
serves as a buncher, and the second beam is sent to a delay stage
stage and
and then
then focused
focused into
into
the second undulator (IFEL2), which acts as the accelerator. The purpose of the
buncher is to modulate the e-beam energy. This leads to the formation of ~3-fs long
microbunches at the accelerator located 2 m downstream of the buncher. By adjusting
the phase delay we demonstrated the ability to resynchronize the microbunches with
the laser light driving the accelerator.
STELLA-II builds upon the success of these first experiments. The primary goal of
STELLA-II is to demonstrate monoenergetic acceleration of the microbunches. To do
this requires separating the microbunches in energy from the unaccelerated
background electrons and trapping the microbunches in the laser beam ponderomotive
potential well. This separation implies the need to impart significant energy gain on
the microbunches. Thus, key differences
differences between the first STELLA experiment and
and
STELLA-II is utilizing higher laser power from an upgraded ATF CO
CO2
laser and
and using
using
2 laser
a tapered undulator for IFEL2.
Another key feature of STELLA-II is using aa single
single laser
laser beam
beam to
to drive
drive both
both the
the
buncher and accelerator. This greatly reduces phase jitter between the two devices
and allows minimizing the separation distance between the buncher and accelerator by
using a chicane rather than a drift space. The laser beam transport system
system was
was also
modified to withstand the much higher laser pulse energy from the upgraded laser.
modified
Figure 2 gives a schematic layout for the STELLA-II experiment.
All the major hardware components have been delivered to the ATF. This includes
different bunchers [an electromagnet (EM) and a fixed-gap permanent-magnet
two different
permanent-magnet
(PM) device], a hybrid PM/EM chicane, and two undulators (untapered and tapered).
These devices are described below and preliminary results are presented.
270
CONVEX
CONVEX MIRROR
MIRROR-
CO2 LASER BEAM
BUNCHER
(IFEL1)
DIPOLE
MAGNET
SPECTROMETER
SPECTROMETER
VIDEO
VIDEO CAMERA
CAMERA
VACUUM
PIPE
E
-BEAM
E-BEAM
FOCUSING
FOCUSING
LENSES
LENSES
WINDOW
LENS
ACCELERATOR
(IFEL2)
TAPERED
UNDULATOR
ARRAY
ARRAY
E-BEAM
VACUUM
VACUUM
CHAMBER
CHAMBER
CHICANE
PARABOLIC
E
BEAM
PARABOLIC
E.-BEAM
MIRRORWITH
WITH
FOCUSING
FOCUSING MIRROR
CENTRAL HOLE
HOLE
LENSES
CENTRAL
LENSES
FIGURE
experiment.
FIGURE 2. Schematic
Schematic layout for the STELLA-II experiment.
DESCRIPTION
DESCRIPTION OF HARDWARE AND SYSTEMS
Laser Beam Transport System
The
The ATF
ATF laser
laser presently delivers approximately 5-J laser pulses with 180-ps pulse
length. Once
Once upgraded the laser will deliver about the same amount
length.
amount of
of pulse
pulse energy,
energy,
but the
the pulse length will be <10 ps. To transport this amount
but
amount of
of pulse
pulse energy,
energy, the
the laser
laser
beam diameter
diameter must
must be large enough to keep the fluence
beam
fluence on
on the
the optics
optics below
below their
their
damage threshold,
threshold, in
in particular on any transmissive optics, which tend to have much
damage
much
lower damage
damage limits
limits than metal mirrors. Consequently, metal mirrors are used
lower
used
wherever possible;
possible; however, a window is still needed on the e-beam vacuum pipe to
wherever
to
permit transmission
transmission of the laser beam. Thus, one requirement of
permit
of the
the laser
laser beam
beam
transport design
design is
is to
to position this window where the laser beam has a large size.
transport
Another requirement
requirement is
is to focus the laser beam in the center of the accelerator
Another
accelerator
(IFEL2)
as
tightly
as
possible
to maximize the laser intensity. This
(IFEL2) as tightly as
This implies
implies the
the need
need
for
a
short
Rayleigh
range,
which
means
the
vacuum
pipe
upstream
of
IFEL2
for a short
upstream of IFEL2 must
must
increase in
in diameter
diameter to accommodate the rapidly expanding laser
increase
laser beam.
beam. A
A triplet
triplet
located immediately
immediately upstream of the buncher (see Fig. 2) limits the
located
the maximum
maximum size
size of
of
the laser
laser beam; however, it is large enough to provide the short Rayleigh range
the
range desired
desired
for the experiment. Nonetheless, at this point in the
for
the laser
laser beam
beam transport
transport the
the beam
beam isis
still too
too small
small for the beamline window. Hence, there was
still
was aa need
need to
to further
further expand
expand the
the
laser beam
beam size.
size.
laser
To solve
solve this
this problem, we use a NaCl lens positioned just
To
just before
before the
the beamline
beamline
window and
and aa 90°
90° off-axis
off-axis parabolic mirror as depicted in Fig. 2. The combination of
window
the lens
lens and
and parabolic mirror provides both the short Rayleigh range
the
range and large
large beam
beam
size at
at the
the window.
window. Although this scheme has an internal focus,
size
focus, this focus
focus occurs
occurs
within the
the beamline
beamline vacuum.
vacuum.
within
271
Cradle for
Tilt
FIGURE3.
Photograph of
of 90°
90° off-axis
off-axis parabolic
FIGURE
on
remote-controlled
cradle.
FIGURE
3.3. Photograph
Photograph
of
90°
off-axis
parabolic mirror
mirror on
on remote-controlled
remote-controlledcradle.
cradle.
Figure333is
photograph of
of the
the parabolic
parabolic copper
copper mirror,
Figure
mirror,
which
has
4-mm
dia.
hole
Figure
isisaaaphotograph
photograph
of
the
parabolic
copper
mirror, which
which has
has aaa 4-mm
4-mmdia.
dia.hole
hole
drilledthrough
throughits
itscenter
center for
for transmission
transmission of
of the
the e-beam.
e-beam. The
The mirror
drilled
the
mirror is
mirror
is supported
supported on
on aa
vacuum-compatible, remotely
remotely adjustable
adjustable cradle
cradle that
vacuum-compatible,
that
both
vertical
and
vacuum-compatible,
adjustable
cradle
that provides
provides both
both vertical
vertical and
and
horizontal
tilt
control.
horizontal
horizontal tilt
tilt control.
control.
Bunchers
Bunchers
photographof
ofthe
the EM
EM buncher
buncher is
is shown
shown in
in Fig.
Fig. 4.
AAphotograph
4. It
It is
is aa 3-pole
3-pole device
device with
withfield
field
clamps
on
its
ends
to
control
the
magnet
field
distribution.
It
is
also
designed
clamps on its ends to control the magnet field
field distribution. It is also designed to
to be
be
slightlyoff
off resonance.
resonance. These
These attributes
attributes enable
enable it
it to
a
slightly
to modulate
modulate the
the e-beam
e-beam by
by only
only
only a
small
amount
(~±0.4%)
despite
being
driven
by
very
high
laser
peak
power.
Due
to
small amount (~±0.4%) despite being driven by very high laser peak power. Due to
theshort
shortRayleigh
Rayleighrange,
range, the
the laser
laser intensity
intensity inside
inside the
the buncher
is
the
buncher is
is also
also small.
small.
the buncher
Pole
Field clamp
coil
FIGURE 4. Photograph of EM buncher.
FIGURE
FIGURE 4.
4. Photograph
Photograph of
of EM
EM buncher.
buncher.
272
FIGURE
FIGURE5.5.
5.Photograph
Photographof
ofPM
PMbuncher
buncherlying
FIGURE
Photograph
of
PM
buncher
lyingon
onits
itsside
sideto
to show
show the
the gap.
gap.
The
ThePM
PMbuncher,
buncher,shown
showninin
inFig.
Fig.5,5, isisaa5-period
to be
on-resonance
The
PM
buncher,
shown
Fig.
5-period device
device designed
designed to
be on-resonance
for
foraaa45.6
45.6MeV
MeVe-beam.
e-beam. This
Thispermits
permitsititit to
to operate
operate at
at lower
lower laser
for
45.6
MeV
e-beam.
This
permits
to
operate
at
lower
laser intensities
intensities despite
despite
having
havingaaalarge
largegap.
gap. Field
Fieldclamps
clampsare
arelocated
located inside
the C-frame,
C-frame, which
having
large
gap.
Field
clamps
are
located
inside the
the
C-frame,
which is
is the
the same
same
basic
basicdesign
designasas
asfor
forthe
theundulators
undulatorsused
usedin
inIFEL2.
IFEL2.
basic
design
for
the
undulators
used
in
IFEL2.
Chicane
Chicane
The
Thechicane
chicane(see
(seeFig.
Fig. 6)6)
6) uses
uses aa 3-pole
the energy
The
chicane
(see
Fig.
uses
3-pole PM
PM configuration
configuration to
to convert
convert the
energy
modulation
to
density
modulation.
It
has
been
pretuned
assuming
±0.4%
modulation
modulation
to
density
modulation.
It
has
been
pretuned
assuming
±0.4%
modulation
modulation to density modulation. It has been pretuned
±0.4%
bythe
thebuncher.
buncher. Using
Usingthe
themain
maincoil
coilto
tochange
change the
the magnetic
magnetic field
field about
this nominal
nominal
by
by
the
buncher.
Using
the
main
coil
to
change
the
magnetic
field
about this
pointcontrols
controlswhen
whenthe
themicrobunches
microbunchesarrive
arrive in
in phase
phase relative
relative to
to the
the laser
laser light
light in the
point
point
controls
when
the
microbunches
arrive
in
phase
relative
to
the
laser
accelerator. Energizing
Energizingthis
thismain
maincoil
coilalso
alsocauses
causesdeflection
deflection of
of the
the e-beam,
e-beam, which
which can
accelerator.
accelerator.
Energizing
this
main
coil
also
causes
deflection
of
the
e-beam,
becompensated
compensatedusing
usingtrim
trimcoils
coilson
onthe
theends
endsof
ofthe
the chicane.
chicane. The
The magnetic
magnetic field
field of the
the
be
be
compensated
using
trim
coils
on
the
ends
of
the
chicane.
chicaneisisisoriented
orientedorthogonal
orthogonalto
tothe
thebuncher
buncherand
andthe
thetapered
tapered undulator
undulator to
to minimize
minimize echicane
chicane
oriented
orthogonal
to
the
buncher
and
the
tapered
undulator
to
beaminteraction
interactionwith
withthe
thelaser
laserbeam
beaminside
insidethe
thechicane.
chicane.
beam
beam
interaction
with
the
laser
beam
inside
the
chicane.
mil
FIGURE 6.
6. Photograph
Photograph of
of hybrid
hybrid PM/EM
PM/EM chicane.
FIGURE
chicane.
FIGURE
6. Photograph
of hybrid
PM/EM chicane.
273
FIGURE7.7. Photograph
Photograph of
oftapered
tapered undulator.
undulator.
FIGURE
Tapered Undulator
Undulator
Tapered
Figure7 7isisa aphotograph
photographofofthe
the tapered
tapered undulator.
undulator. ItIt is
is the
the same
Figure
same undulator
undulator used
used
during
the
first
STELLA
experiment
[3]
except
with
one
end
of
the
during the first STELLA experiment [3] except with one end of the magnet
magnet array
array
taperedtotosmaller
smaller gap.
gap. Presently
Presently the
the gap
gap taper
taper is
is set
set at
tapered
at 8%;
8%; itit isis capable
capable ofof aa
maximum
taper
of-19%.
maximum taper of ≈19%.
PRELIMINARY RESULTS
PRELIMINARY RESULTS
Initial tests indicate the PM buncher is undermodulating the e-beam by producing a
Initial tests indicate the PM buncher is undermodulating the e-beam by producing a
modulation
of only roughly ±0.2% instead of the needed ±0.4%. This implies the
modulation
of only
roughly
±0.2% isinstead
needed ±0.4%.
Thistoimplies
the
laser intensity
within
the buncher
lower of
thantheexpected.
Recall due
the short
laser
intensity
within
the
buncher
is
lower
than
expected.
Recall
due
to
the
short
Rayleigh range, the laser beam diameter at the buncher is large (>1 cm).
Rayleigh
range, in
thethelaser
beamdistribution
diameter can
at lead
the to
buncher
large (>1 Such
cm).
Nonuniformities
intensity
weaker ismodulation.
Nonuniformities
in
the
intensity
distribution
can
lead
to
weaker
modulation.
Such
nonuniformities might be caused by diffraction effects due to, say, the central hole in
nonuniformities
might be
causedtests
by diffraction
to, say, the
central holethein
the parabolic mirror.
Further
and analysiseffects
will bedueconducted
to understand
the
parabolic
Further tests and analysis will be conducted to understand the
cause
for the mirror.
smaller modulation.
causeThe
for STELLA-II
the smaller modulation.
experiment can still be performed since the weaker modulation
The results
STELLA-II
can still
be performed
thepreliminary
weaker modulation
only
in less experiment
tightly bunched
electrons.
Figure 8since
shows
raw data
only
in less tightly
electrons.
Figure 8ofshows
preliminary
data
fromresults
the electron
energy bunched
spectrometer
as a function
the chicane
phaseraw
delay.
from
the 8(a),
electron
as a function
of the images
chicanewhere
phaseenergy
delay.
Figures
(c), energy
and (e) spectrometer
are the spectrometer
video camera
Figures
8(a),
and (e)Figure
are the
energy
increases
to (c),
the right.
8(b),spectrometer
(d), and (f) video
are thecamera
energyimages
profileswhere
through
the
increases
the right.
8(b), arbitrarily
(d), and (f)
are the0°energy
through
the
center oftothese
images.Figure
We have
assigned
phase profiles
to Fig. 8(c),
which
showed
maximum
this particular
set 0°
of data.
maximum
center
of the
these
images.acceleration
We have for
arbitrarily
assigned
phase Indeed,
to Fig. a8(c),
which
acceleration
of >13% acceleration
was measured,
to our knowledge
is theIndeed,
largest aamount
of
showed
the maximum
forwhich
this particular
set of data.
maximum
accelerationofobserved
frommeasured,
an IFEL thus
far.to our knowledge is the largest amount of
acceleration
>13% was
which
As the phase
delayfrom
is adjusted
from Fig. 8(c), we see evidence that a group
acceleration
observed
an IFEL±100°
thus far.
ofAselectrons
shifting
in energy.
energy
peaksweare
broad,
is
the phaseis delay
is adjusted
±100°The
from
Fig. 8(c),
seequite
evidence
thatwhich
a group
nonoptimal
bunching The
of theenergy
electrons
due are
to thequite
undermodulation
by is
ofconsistent
electrons with
is shifting
in energy.
peaks
broad, which
consistent with nonoptimal bunching of the electrons due to the undermodulation by
274
Spectrometer Output (arb. units)
EnergyShift
Shift(%)
(%)
Energy
140
140
-6 -4
-4 -2
-2 00 22 44 66 88 10
1012
1214
1416
16
-6
120
120
100
80
60
40
20
0
-3
- 3-2 - -12 0- 11 02 1 32 43 45 56 6 7 7 88
EnergyShift
Shift(MeV)
(MeV)
Energy
(a) Phase
Phase delay
delay ==-100°
(a)
-100°
(b)
(b)
Spectrometer Output (arb. units)
Energy
EnergyShift
Shift(%)
(%)
-6
-6 -4-4 -2-2 00 22 44 66 88 10
1012
1214
1416
16
100
80
60
40
20
20
0
-3
- 3-2 - -12 0- 11 02 1 32 43 45 56 6 7 7 88
Energy
EnergyShift
Shift(MeV)
(MeV)
(c)
(c) Phase
Phase delay
delay == 0°
0°
(d)
(d)
Energy
EnergyShift
Shift(%)
(%)
Spectrometer Output (arb. units)
100
^100
-6
-6 -4-4 -2-2 00 22 44 66 88 10
1012
1214
1416
16
80
I80
60
I 60
O
® 40
40
I
<fl
(e)
+100°
(e) Phase
Phase delay
delay ==+100°
20
2
°
00
-3
- 3-2 - -12 0- 11 02 1 32 43 45 56 6 7 7 88
Energy
EnergyShift
Shift(MeV)
(MeV)
(f)
(f)
FIGURE
FIGURE 8.
8. Preliminary
Preliminary experimental
experimental results
resultsfor
forSTELLA-II.
STELLA-II. (a),
(a),(c),
(c),and
and(e)
(e)are
areraw
rawvideo
videooutput
output
from
from the
the spectrometer
spectrometer camera
camera with
with energy
energydispersion
dispersion increasing
increasingto
tothe
theright.
right, (b),
(b),(d),
(d),and
and(f)
(f)are
areline
line
profiles
profiles through
through the
the center
center of
of (a),
(a), (c),
(c), and
and (e),
(e), respectively.
respectively.
275
the buncher. Hence, this preliminary data seems to indicate that the chicane is
functioning properly.
Even with 13% energy gain, this preliminary data shows that the microbunch
electrons have not gained enough energy to separate from the background electrons.
As shown below, our model predicts at least 20% energy gain will be necessary for
this separation to occur. This requires setting the accelerator undulator to 19% gap
taper since the amount of energy gain is directly related to the amount of taper. A
larger taper also requires higher laser intensity to drive it. All this points to the need
for the upgraded CO2 laser, which should provide more than enough peak power to
drive a 19% gap tapered undulator and the EM buncher rather than the PM buncher.
MODEL PREDICTIONS FOR UPGRADED LASER
Assuming a 19% gap tapered undulator and the upgraded CO2 laser with 1 TW/cm2
at the center of the undulator, Fig. 9 gives the model predictions for the STELLA-II
experiment. The chicane phase has been adjusted for minimum energy spread of the
microbunch and a high resolution spectrometer is assumed.
Energy Shift (%)
-8-4
12
10
> 8
f
6
!c
4
0
4
8
12
16 20 24
1 2
CD
£ 0
-2
-3
-2
-
1
0
1
-4
Output Phase (rad)
-2
0
2
4
6
8
10
12
Energy Shift (MeV)
(a)
(b)
FIGURE 9. Model predictions for STELLA-II using upgraded ATF CO2 laser, (a) Energy-phase
diagram, (b) Energy histogram.
This shows in Fig. 9(a) the microbunch electrons trapped in a fairly small group
(see upper left-hand corner of phase diagram). These electrons have a narrow energy
spread as seen in Fig. 9(b) and are well separated from the background electrons.
Note, the energy gain is 20%.
CONCLUSIONS
The STELLA-II experiment has begun obtaining its first data. An energy gain
>13% has already been observed. Complete energy separation of the trapped
276
microbunches from the background electrons requires an energy gain of at least 20%.
To achieve this requires utilizing the higher laser peak power that will be available
from the upgraded ATF CO2 laser and a 19% gap-tapered undulator. This upgrade
should be completed by the end of 2002 at which point the STELLA-II primary goal
of demonstrating staged monoenergetic laser acceleration can be achieved. In the
meantime, the experiment will be operated at lower laser power in order to further
characterize and optimize the equipment.
ACKNOWLEDGMENTS
The authors wish to acknowledge Dr. Xijie Wang and the staff at the ATF for thensupport of this experiment. This work was sponsored by the U. S. Department of
Energy, Grants Nos. DE-FG03-98ER41061, DE-AC02-98CH10886, and DE-FG0392ER40695.
REFERENCES
1. P. Sprangle, "Laser Driven Plasma Accelerators," in these Proceedings.
2. W. D. Kimura, A. van Steenbergen, M. Babzien, I. Ben-Zvi, L. P. Campbell, C. E. Dilley, D. B.
Cline, J. C. Gallardo, S. C. Gottschalk, P. He, K. P. Kusche, Y. Liu, R. H. Pantell, I. V. Pogorelsky,
D. C. Quimby, J. Skaritka, L.C. Steinhauer, and V. Yakimenko, Phys. Rev. Lett. 86, 4041-4043
(2001).
3. W. D. Kimura, L. P. Campbell, C. E. Dilley, S. C. Gottschalk, D. C. Quimby, A. van Steenbergen,
M. Babzian, I. Ben-Zvi, J. C. Gallardo, K. P. Kusche, I. V. Pogorelsky, J. Skaritka, V. Yakimenko,
D. B. Cline, P. He, Y. Liu, L. C. Steinhauer, and R. H. Pantell, Phys. Rev. ST Accel. Beams, 4,
101301 (2001).
277