Short Bunch Beam Diagnostics
Patrick Krejcik
Stanford Linear Accelerator Center
2575 Sand Hill Rd, Menlo Park CA 94025
Abstract. With the emergence of 4th generation PEL based light sources there is now
considerable interest in both producing and characterizing ultra-short (<100 fs) electron bunches.
Knowledge of the extremely high peak current in a short bunch is required to diagnose the SASE
(self amplified stimulated emission) process. Measuring the femtosecond duration of the pulse is
inherently interesting, particularly for experimenters using the beam to measure fast phenomena
(e.g. femto-chemistry). Diagnostic techniques that have the necessary femtosecond resolution
will be reviewed: These include high-power RF transverse deflecting structures that "streak" the
beam in the accelerator allowing the bunch length to be recorded on a profile monitor. Electro
optic crystal diagnostics use the electric field of the electron bunch to modulate light thereby
exploiting the femtosecond technology of high bandwidth visible lasers. Coherent synchrotron
radiation (CSR) from dipole magnets and optical diffraction radiation (ODR) both result in
radiation with wavelengths of the order of the bunch length and hence in the terahertz band
which can be detected by a variety of techniques. The role of each of these techniques is
discussed in terms of its application at the Linac Coherent Light Source (LCLS) and the Short
Pulse Photon Source (SPPS) currently under construction at SLAC.
INTRODUCTION
The Linac Coherent Light Source (LCLS) [1] to be built at SLAC utilizes electron
bunches as short as 80 femtoseconds rms to generate self-amplified stimulated
emission (SASE) X-ray radiation in a PEL. The production and tuning of these ultrashort bunches is critical to the performance of the LCLS but the measurement of such
short bunch lengths is a considerable challenge that cannot be met, for example, with
conventional streak camera technology. The technologies and limitations to various
bunch length measurement techniques are reviewed in this paper together with plans
for developing and testing these techniques at SLAC.
The new Short Pulse Particle Source (SPPS) [2] at SLAC offers a nearer term
opportunity to test and compare these different diagnostic techniques with bunches as
short as 30 fs rms, far shorter than anything so far produced in a high energy electron
accelerator. The locations of the SPPS and LCLS installations are shown in figure 1 in
relation to the other accelerators at SLAC. SPPS will only produce spontaneous x-ray
radiation from its undulator but the peak spectral brightness from this source, at 1024
photons s"1 mm"2 mr"2 (0.1% bandwidth), still exceeds that of any existing x-ray
source.
CP648, Beam Instrumentation Workshop 2002: Tenth Workshop, edited by G. A. Smith and T. Russo
2002 American Institute of Physics 0-7354-0103-9
162
North Damping Ring
SPPS Bunch
SPPS
SPPS Bunch
Bunch
Compressor
Chicane
Compressor
Chicane
Compressor Chicane
New LCLS
New LCLS
Injector
Injector
SPPS X-Rays
SPPS X-Rays
LCLS
LCLS
X-Ray FEL
X-Ray FEL
FIGURE 1. The locations
bunch
compressor
chicane
isis shown
superimposed
on
the
locations of
of the
the SPPS
SPPS linac
linac bunch
bunch compressor
compressorchicane
chicaneis
shownsuperimposed
superimposedon
onthe
the
FIGURE 1. The locations
of
the
SPPS
linac
shown
SLAC site together
with
the
LCLS
facility
which
utilizes
the
last
one
third
ofofthe
linac
totogenerate
together
with
the
LCLS
facility
which
utilizes
the
last
one
third
the
linac
generate
SLAC site together with the LCLS facility which utilizes the last one third of the linac to generate
coherent
x-rays from
a PEL
coherent
FEL located
locateddownstream
downstreamof
ofthe
thebeam
beamswitch
switchyard.
yard.
coherent x-rays
x-rays from
from aa FEL
located
downstream
of
the
beam
switch
yard.
To
the application
application of
of the
the various
various short
short bunch
bunch diagnostic
diagnostictechniques
techniquesthe
the
To illustrate
illustrate the
the
application
of
the
various
short
bunch
diagnostic
techniques
the
accelerator
layout
of
the
LCLS
is
shown
in
figure
2
with
indicators
for
the
locations
of
of
the
LCLS
is
shown
in
figure
2
with
indicators
for
the
locations
accelerator layout of the LCLS is shown in figure 2 with indicators for the locations ofof
the
ininthe
devices.
Bunches of
of approximately
approximately 333ps
psrms
rmslength
lengthare
areproduced
producedin
theRF
RF
the various
various devices.
devices. Bunches
Bunches
of
approximately
ps
rms
length
are
produced
the
RF
photoinjector
and
accelerated
to
150
MeV
where
a
magnetic
chicane
compresses
the
photoinjector
accelerated
to
MeV
where
a
magnetic
chicane
compresses
the
photoinjector and accelerated to 150 MeV where a magnetic chicane compresses the
bunches
to
an
rms
length
of
630
fs.
A
second
bunch
compressor
is
located
at
the
length
of
630
fs.
A
second
bunch
compressor
is
located
at
the
bunches to an rms length of 630 fs. A second bunch compressor is located at the
4.5GeV
its
final
bunch
length
of
75
fs.
the bunch
bunch attains
attains its
its final
final bunch
bunchlength
lengthof
of75
75fs.
fs.
4.5GeV point
point where
where the
the
bunch
attains
At
the
injector,
confirmation
of
the
critical
initial
longitudinal
the
injector,
confirmation
of
the
critical
initial
longitudinal
distributionof
the
At the injector, confirmation of the critical initial longitudinal distribution
distribution
ofofthe
the
bunch
is
independently
diagnosed
with
both
electro-optic
profiling
of
the
bunch
and
diagnosed
with
both
electro-optic
profiling
of
the
bunch
and
bunch is independently diagnosed with both electro-optic profiling of the bunch and
with
system
long S-band
S-band transverse
transverse deflecting
deflectingcavity.
cavity.The
Theelectro-optic
electro-opticsystem
system
with aa short,
short, 0.55
0.55 m
m long
long
S-band
transverse
deflecting
cavity.
The
electro-optic
can
make
use
of
the
already
available
high-bandwidth
TiiSapphire
laser
system
for
already
available
high-bandwidth
Ti:Sapphire
laser
system
forthe
the
can make use of the already available high-bandwidth Ti:Sapphire laser system for
the
injector
to
easily
achieve
the
necessary
sub-picosecond
resolution
to
characterize
the
injector
to
easily
achieve
the
necessary
sub-picosecond
resolution
to
characterize
the
injector to easily achieve the necessary sub-picosecond resolution to characterize the
gun pulse.
150
energy
of
the
injector
means
that
only
a
short
gun
pulse. The
The relatively
relatively low
low 150
150 MeV
MeV
energy
of
the
injector
means
that
only
a
short
MeV energy of the injector means that only a short
section of
structure
isis required
to
streak
section
of transverse
transverse deflecting
deflecting structure
structure is
required to
to streak
streak the
the beam
beam on
on aaa
deflecting
required
the
beam
on
downstream
profile
monitor
using
about
1
MW
of
power
split
off
from
one
of
downstream profile monitor using
using about
about 11 MW
MW of
of power
power split
split off
off from
from one
oneofofthe
the
the
injector
klystrons.
injector klystrons.
A
second transverse
in
the
high-energy
linac
with
A second
transverse deflecting
deflecting cavity
cavity is
is also
also located
locatedin
inthe
thehigh-energy
high-energylinac
linacwith
withaaa
deflecting
cavity
is
also
located
length
of
2.44
m
and
is
powered
with
25
MW
from
a
separate
klystron.
length of 2.44 m and is powered
powered with
with 25
25 MW
MW from
from aa separate
separateklystron.
klystron.
iEO
ll
φ
φφ
Tz
Tz
Tz
TC
∆∆E
E
150 MeV
MeV
150
2.8
2.8 ps
ps
φφ
TC
TC
∆E
CSR
250
250 MeV
MeV
630
630 fs
fs
∆∆EE
CSR
CSR
ZZφφ
EO
EO
∆∆EE
4.5
GeV
4.5
4.5GeV
GeV
75
fs
75
75fsfs
FIGURE
schematic layout
layout of
of the
the LCLS accelerator
accelerator and bunch
bunch compressor system
system showing the
the
FIGURE 2.
2. A
A schematic
of the LCLS
LCLS accelerator and
and bunch compressor
compressor systemshowing
showing the
types and locations of
the
various
diagnostics
to
measure
bunch
length
and
characterize
the
longitudinal
types and locations of the various
various diagnostics
diagnostics to
to measure
measurebunch
bunchlength
lengthand
andcharacterize
characterizethe
thelongitudinal
longitudinal
phase
(EO), Transverse
Transverse Cavity (TC),
(TC), Terahertz power
power monitors (Tz),
(Tz),
phase space
space of
of the
the beam:
beam: Electo-Optics
Electo-Optics
Electo-Optics (EO),
(EO), Transverse Cavity
Cavity (TC),Terahertz
Terahertz powermonitors
monitors (Tz),
Coherent Synchrotron Radiation
monitors
(CSR),
Energy
spread
monitors
(AE),
Beam
Phase
monitors
Coherent Synchrotron Radiation monitors
monitors (CSR),
(CSR), Energy
Energy spread
spreadmonitors
monitors(∆E),
(∆E),Beam
BeamPhase
Phasemonitors
monitors
((|)),and
(Z <|>).
(φ),and Zero-phase
Zero-phase measurement
measurement locations
locations
locations (Z
(Z φ).
φ).
163
The energy and energy spread of the beam are measured on beam position monitors
and The
profile
monitors
at allspread
the dispersive
locations
where on
thebeam
beamposition
is bent.monitors
The very
energy
and energy
of the beam
are measured
short
bunches
generate
a
coherent
component
to
the
synchrotron
light
in
the
and profile monitors at all the dispersive locations where the beam is bent. The bunch
very
compressor
chicane
bends
so the light
will also
to light
monitor
thebunch
bunch
short bunches
generate
a coherent
component
to be
thediagnosed
synchrotron
in the
length
at these
locations.
and
control ofto the
chicane
bunch
compressor
chicane
bends soTuning
the light
willfeedback
also be diagnosed
monitor
the bunch
compressors
is facilitated
by Tuning
rapid, pulse-to-pulse
CSR power
length at these
locations.
and feedbackmeasurement
control of of
thethechicane
bunchby
narrow
band terahertz
detectors.
Beam
phase monitors
are located
at each
compressors
is facilitated
by rapid,
pulse-to-pulse
measurement
of the
CSRaccelerating
power by
sector.
accelerating
section downstream
of the second
bunchatcompressor
will be
narrowThe
band
terahertz detectors.
Beam phase monitors
are located
each accelerating
used
forThe
zero-phase
crossing
measurement
length
in conjunction
the
sector.
accelerating
section
downstreamof
of bunch
the second
bunch
compressor with
will be
profile
monitor
locatedcrossing
in the final
bend before
the undulator.
used for
zero-phase
measurement
of bunch
length in conjunction with the
The SPPS
linac
bunch
compressor
is similarly equipped with short
profile
monitor
located
in the
final bend chicane
before the[3]
undulator.
bunch
to bunch
tune and
measure its
performance.
Thediagnostics
SPPS linac
compressor
chicane
[3] is similarly equipped with short
bunch
to tune and measure
performance.
Eachdiagnostics
of these techniques
will be its
described
in the following sections and their
Each of
techniques
will be described
following
sections
andwill
theirbe
potential
forthese
reaching
the necessary
resolutionin inthethe
LCLS and
SPPS
potential for
the necessary
resolution
in electron
the LCLS
and also
SPPS
will bean
discussed.
The reaching
full temporal
characterization
of the
bunch
involves
discussed. of
The
temporal
of measured
the electron
also involves
an
assessment
thefull
timing
jitter characterization
and how it can be
on bunch
a pulse-to-pulse
basis.
assessment of the timing jitter and how it can be measured on a pulse-to-pulse basis.
TRANSVERSE DEFLECTING CAVITIES
TRANSVERSE DEFLECTING CAVITIES
The principal of the transverse cavities is shown in figure 3 where it can be seen
principal
of the transverse cavities
is shownorincrabbing,
figure 3 can
where
it can be on
seen
that The
a strong
longitudinal-to-transverse
correlation,
be imposed
the
that
a
strong
longitudinal-to-transverse
correlation,
or
crabbing,
can
be
imposed
on
bunch when the cavity is operated close to zero phase crossing of the appliedthe
RF
bunch when
the cavity of
is measuring
operated close
to length
zero phase
crossing
of the
applied
RF
voltage.
This technique
bunch
was in
fact used
during
the early
voltage.
This
technique
of
measuring
bunch
length
was
in
fact
used
during
the
early
commissioning years of the injector at SLAC[4] and in more recent times has been
commissioning
years offor
the measuring
injector at very
SLAC[4]
in more recent
timestohas
proposed
as a method
shortand
bunches[5].
In order
testbeen
this
proposed as a method for measuring very short bunches[5]. In order to test this
technique we have just completed recommissioning one of the original SLAC
technique we have just completed recommissioning one of the original SLAC
deflecting structures [6] by installing it in the SLAC linac where it can be used with a
deflecting structures[6] by installing it in the SLAC linac where it can be used with a
variety of beams during normal accelerator operations[7].
variety of beams during normal accelerator operations[7].
V(/)t
V(t)
\y\7
e−
σz
〈∆y〉
σy
2π
TMu
TM11
T
2,44
2.44 m
∆ψ ≈ 660°
0°
t
βc
t
βp
FIGURE3.3.The
The transverse
transverse deflecting
deflecting cavity
cavity can
can be
be used
FIGURE
used to
to crab
crab the
the bunch
bunch so
so that
that itsitsprojected
projected
transversebeam
beamsize
sizemeasured
measuredon
onaadownstream
downstream profile
profile monitor
monitor is
transverse
is proportional
proportionaltotobunch
bunchlength.
length.
164
250
200
150
I100
50
vertical distance [mm]
horizontal distance [mm]
0.5 vertical distance [mm] 1.5
Projected Vertical Beam Profile: o = 0.11728 mm
0
0.05
0.1
0.15
0.2
horizontal distance [mm]
0.25
0.5
vertical distance [mm]
horizontal distance [mm]
1
1.5
vertical distance [mm]
0.5 vertical distance [mm] 1.5
Projected Vertical Beam Profile: o = 0.53865 mm
... ...i...
V
0
0.05
0.1
0.15
0.2
0.25
0.5
horizontal distance [mm]
1
1.5
2
vertical distance [mm]
FIGURE 4. Profile monitor images and projected vertical image profile with the RF transverse cavity
FIGURE
4. with
Profile
imagesshowing
and projected
profile with the RF transverse cavity
off
(left) and
themonitor
RF on (right)
how thevertical
beam isimage
streaked.
off (left) and with the RF on (right) showing how the beam is streaked.
165
A profile monitor screen (p) is positioned at a suitable transverse betatron phase
advance from the deflecting cavity (c). The beam size measured on the screen is the
quadrature sum of the initial vertical beam size, <7yo2=/^£yo and the RF induced
crabbing.
(1)
The advantage of this technique is that a large deflecting voltage, Vo can be
achieved at short RF wavelength, /I by applying high peak RF power supplied by the
same type of klystron as is used to power the normal accelerating sections.
A recent test of this technique with a 2.44 m long 2856 MHz device at SLAC[7]
showed that 10 jim bunch length resolution could be achieved with a RF deflecting
amplitude of 17.5 MV on a bunch with a normalized vertical emittance of 3*10-5 m
rad. Images of the streaked beam are shown in figure 4.
In addition to measuring the rms bunch length, information can also be gained on
the bunch length distribution. The measured profile is a convolution of the actual
longitudinal distribution in the bunch with the vertical beam size at the screen as
determined by the transverse emittance. When the longitudinal distribution is nonGaussian, due to the bunch being over compressed for example, the measured profile
is also non-Gaussian. Particle tracking[8] of the LCLS beam show that quite complex
longitudinal bunch distributions are to be expected as a result of the compression
dynamics together with the strong wakefields experienced in the linac that are excited
by the short bunch. Simulation[9] of the transverse beam profile from the RF
deflecting cavity shows that adequate resolution can be achieved to reveal most of the
details of the complex longitudinal distribution of the LCLS bunch.
The RF deflecting cavities offer such a powerful resolution of the beam properties
along the longitudinal z-coordinate of the bunch that it is reasonable to consider
measuring other properties as a function of the "longitudinal slice" position along the
bunch. Consider, for example, that the horizontal width of the beam can be measured
at different vertical positions on the screen and hence at different positions along the
bunch, as shown in figure 5 a. If this is combined with a standard emittance measuring
technique such as varying the strength of an appropriate upstream quadrupole then the
horizontal emittance of each of the slices along the bunch can be determined.
Another technique is to use a profile monitor screen in a region of large horizontal
dispersion, after a bending magnet, so that the horizontal beam size becomes a
measure of energy spread in the beam. Recording the horizontal beam size along the
vertical axis of the profile monitor gives a measurement of what is referred to as the
slice energy spread of the bunch. The slice emittance and slice energy spread together
with the bunch length and charge are key parameters in the performance of an PEL.
Although this bunch length measurement technique disrupts the beam it is possible
to power the device on and off on a pulse-by-pulse basis so that it can operate with a
1% or less duty cycle and minimize the impact to the downstream users.
166
ED=6.2 GeV
0,6
0,6 .
0.4
0.4 .
-0.4
-0,4 .
-0,6
-0.6 .
-0.6 -0,4 -0,2 0,0 0.2
0
0.4 0.6 0.6
x (mm)
-0.6 -0,4 -0,2 0,0 0.2
2000 4000 6000 8000
N
0.4 0.6 0.8
x (mm)
E Q =1 4.3
0,20
GeV
0.1 5
0,10
0,05
0,00
-0.05
-0.1 0
-0,1 5
-0,20
-0,15
-0,10
-0,05
0,00
0,05
0,10
0,15
0,ZO
x (mm)
watch-point phase space—Input Icls29nov01 .ele
lattice: Icls29nov01,lte
LCLS beam
beam (from
(from P.
P. Emma,
Emma, ref.
ref. 1
and 9)
9)
FIGURE
FIGURE 5.
5. a.
a. Expected
Expected longitudinal
longitudinal slice
slice profile
profile for
for the
the LCLS
land
from
from which
which the
the horizontal
horizontal slice
slice emittance
emittance could
could be
be reconstructed,
reconstructed, and
and b.
b. the
the slice
slice energy
energy spread
spread
obtained
beam at
dispersion profile
profile monitor
monitor location
location (ibid).
(ibid).
obtained by
by viewing
viewing the
the streaked
streaked beam
at aa high
high dispersion
ZERO-PHASE
ZERO-PHASE CROSSING
CROSSING MEASUREMENTS
MEASUREMENTS
A
RF
A more
more invasive
invasive measurement
measurement of
of the
the bunch
bunch length
length distribution
distribution uses
uses the
the RF
accelerating
accelerating cavities
cavities that
that are
are part
part of
of the
the linac.
linac. Rather
Rather than
than accelerate
accelerate the
the beam
beam on
on crest
crest
the
by
the bunch
bunch is
is moved
moved to
to the
the RF
RF zero-phase
zero-phase crossing
crossing where
where the
the accelerating
accelerating field
field seen
seen by
each
each particle
particle changes
changes rapidly
rapidly along
along the
the bunch
bunch length
length thereby
thereby inducing
inducing aa correlated
correlated
energy
The bunch
bunch length
energy spread,
spread, or
or chirp,
chirp, along
along the
the bunch,
bunch, as
as shown
shown in
in figure
figure 6.
6. The
length
which
was
first
rotated
onto
the
energy
axis
by
the
RF
is
now
transformed
into the
the
which was first rotated onto the energy axis by the RF is now transformed into
167
spatial coordinate by observing the transverse beam profile on a down-stream highspatial coordinate
by observing the transverse beam profile on a down-stream highdispersion
profile monitor.
dispersion
profilehorizontal
monitor. beam size <JX at a profile monitor location where the
The measured
The measured
horizontal
beamsum
sizeof σthe
a profile
x atbeam
dispersion
is r/x is the
quadrature
size monitor
<7xo, withlocation
the RF where
off, andthethe
η
is
the
quadrature
sum
of
the
beam
size
σ
,
with
the
RF
off,
and the
dispersion
is
x0
contribution fromx the energy spread:
contribution from the energy spread:
σx = σ
2
x0
  dE  2πeVrf
+σ η 
+
cos ϕ rf
  dz  0 λ E
rf 0

2
z
2
x




2
(2)(2)
where
(dE/dz)
the initial energy spread in the beam.
0 is
where
(dE/dz)
0 is the initial energy spread in the beam.
^^^^^^^lii∆E
X
Beamline
ηx
RF
RF at zerocrossing phase
Resultant energy
chirp on bunch
Profile monitor
intensity
FIGURE 6. A rotation in longitudinal phase space is induced by moving the bunch off crest to the zeroFIGURE
6. A rotation in longitudinal phase space is induced by moving the bunch off crest to the zerophase crossing where the resultant correlated energy spread is measured on a high-dispersion location
phase crossing where the resultant correlated energy spread is measured on a high-dispersion location
profile monitor.
profile monitor.
Good bunch length resolution can be obtained if the induced correlated energy
Good is
bunch
can be obtained
if the
induced
energy
spread
muchlength
greater resolution
than the incoherent
energy spread
in the
bunch.correlated
For this reason
spread
is
much
greater
than
the
incoherent
energy
spread
in
the
bunch.
For
this
reason
it is usually necessary to use a large portion of the linac in this off-crest mode.
it is usually
necessary
to
use
a
large
portion
of
the
linac
in
this
off-crest
mode.
This technique has been used extensively in short bunch applications such as FEL
This technique
has the
been
usedlength[10,11].
extensively in
short
applications
asfor
PEL
injectors
to measure
bunch
The
samebunch
technique
describedsuch
above
injectors
to
measure
the
bunch
length[10,ll].
The
same
technique
described
above
measuring the slice emittance with the transverse cavity can also be used infor
measuring
thewith
slice
emittance crossing
with the
transverse cavity can also be used in
conjunction
the zero-phase
measurement.
conjunction
with therelies
zero-phase
crossing
measurement.
This technique
on a rotation
in longitudinal
phase space of the bunch, induced
This
relies
a rotationthis
in from
longitudinal
phase spaceenergy
of thecorrelations
bunch, induced
by
thetechnique
RF. In order
to on
distinguish
wakefield-induced
in
is possible
to rotate the
opposite directions
by performing
the in
bythe
thebunch
RF. Init order
to distinguish
thisbunch
from in
wakefield-induced
energy
correlations
on the two
zero-phase
crossings
of the directions
RF in turn.byThe
wakefieldthe
themeasurements
bunch it is possible
to rotate
the bunch
in opposite
performing
contribution
will
be
seen
to
either
aid
or
oppose
the
RF
induced
energy
spread
and can
measurements on the two zero-phase crossings of the RF in turn. The wakefield
thus
subtracted
out
of
the
bunch
length
reconstruction.
contribution will be seen to either aid or oppose the RF induced energy spread and can
The more invasive
zero-phase
crossing measurements means they are
thus subtracted
out of thenature
bunchoflength
reconstruction.
usually
onlyinvasive
performednature
duringofinitial
tune up of
the accelerator
or to cross
checkthey
otherare
The more
zero-phase
crossing
measurements
means
bunch
length
measuring
techniques.
usually only performed during initial tune up of the accelerator or to cross check other
bunch length measuring techniques.
168
ELECTRO OPTIC TECHNIQUES
At
techniques begin
begin to
to give
give out
out
At extremely
extremely short
short bunch
bunch lengths
lengths where
where the
the previous
previous techniques
through
to other
other technologies.
technologies. It
It is
is possible
possible to
to take
take
through lack
lack of
of RF
RF power
power we
we can
can move
move to
advantage of the advances in laser science where femtosecond technology has
has been
developed.
have very
very high
electric fields
fields associated
associated with
with
developed. Very
Very short
short electron
electron bunches
bunches have
high electric
them
to modulate
laser pulse.
pulse. The
The problem
problem of
of measuring
measuring the
the
them which
which can
can be
be used
used to
modulate aa laser
electron bunch length then moves to one of
of measuring
measuring the
the laser
laser pulse
pulse modulation,
modulation, for
for
which several techniques exist.
As an illustration, we consider the setup shown schematically in figure 7, where an
electro-optic (EO) crystal responds to the electric field from the electron
electron bunch.
bunch.
Beam pipe
pipe
Beam
Electron bunch
Co-propagating
Laser pulse
Polarizer
ECnCrystal
EO
Crystal
Spectrometer
Spectrometer
Analyzer
Analyzer
I
ωl
t
Initial laser chirp
ωs
t
Bunch charge
charge
Bunch
Gated spectral
spectral signal
signal
Gated
FIGURE 7. The electric field from a bunch modulates the polarization in an electro optic
optic crystal
crystal and
and
gates the transmission of a chirped laser pulse.
The electric field of the bunch alters the optical retardation along one axis of the
birefringent
birefringent crystal according to
2
3
+X
E +...~]
XiE+X

P=
= εe00[x
2E
1E + X
2 E + 3X3 E + K
2
3
(3)
(3)
The optical transmission through a polarizer-analyzer pair is modulated
modulated as
as the
polarization vector in the crystal rotates under the influence of the electric field of the
bunch. By superimposing a chirp on the incoming laser pulse the time
time structure
structure of
of the
the
electron beam induced modulation can be measured in the frequency domain with a
spectrometer.
169
rr
.L.
2N
N eee
9 2
9
E
9
10
=
×
E = 9 × 10 r 2πe σ
r 2π σ zz
m..kk..ss units
units
m.k.s
m
1/γ
1/γ
FIGURE
Point
charge
approximation
for
the
electric
field at
at a distance
distance rr from
from a short,
short, relativistic
relativistic
FIGURE8.8.8.Point
Pointcharge
chargeapproximation
approximationfor
forthe
theelectric
electric field
field
FIGURE
at aa distance
r from aa short,
relativistic
bunch.
bunch.
bunch.
10 µm
10 µm
33fs rms
33fs rms
Frequency [THz]
FIGURE9.9.The
The bandwidth
bandwidth of
of aa Gaussian
Gaussian bunch extends into the THz
FIGURE
THz band
band for
for short
short (lOum)
(10µm)bunches.
bunches.
FIGURE 9. The bandwidth of a Gaussian bunch extends into the THz band for short (10µm) bunches.
The electric
electric field
field of
of an
an ultra-relativistic
ultra-relativistic bunch of Nee electrons at a distance
distance r,r, from
from
The
The
electric
field of
an
ultra-relativistic
bunch of Ne electrons
at a distance r, from
the
crystal
shown
in figure
figure
for aa Gaussian
<JzZ[12].
[12].
the
crystal
isis shown
in
88 for
bunch of length σ
the crystal
shown inthe
figure
8 forangle,
a Gaussian
bunchsmall
of lengththe
σ [12]. can be placed
At high
highisenergies
energies
opening
l/%
At
the opening
1/
γ, is very
very small so
so thez crystal
crystal can be placed
At
high
energies
the
opening
angle,
1/
γ
,
is
very
small
so
the
crystal
can
be is
placed
somedistance
distance from
from the
the beam
beam without
without loss of bunch length resolution.
The
field
some
resolution.
The
field
is also
also
some
distance
from
the
beam
without
loss
of
bunch
length
resolution.
The
field
is MV
also
quite
strong,
for
example
a
10
|Lim
long
bunch
with
1
nC
charge
will
generate
a
72
quite strong, for example a 10 µm long
will generate a 72 MV
-11 field
quite
strong,
exampleofa r10= µm
longAtbunch
with 1 nC
charge
will
generate
a 72 MV
m"
at for
a distance
1 mm.
low energies
around
the
gun
and
the
around
the
gun
and injector
injector
the
m
-1 field at a distance of r = 1
fieldneeds
at a distance
of
r
=
1
mm.
At
low
energies
around
the
gun
and
injector
the
mcrystal
crystal
needs
to
be
brought
into
closer
proximity
with
the
beam
to
maintain
good
to be brought
with the beam to maintain good
crystal
needs
to
be brought into closer proximity with the beam to maintain good
bunchlength
length resolution.
resolution.
bunch
bunch length resolution.
170
In a beam pipe the high frequency component of the wakefields contain the bunch
length distribution information. The bandwidth of radiation produced from a Gaussian
bunch, given by its Fourier transform,
F(co) = e
2
(4)
extends into the terahertz region for the ultra-short bunches considered here, as shown
in figure 9.
COHERENT RADIATION FROM A SHORT BUNCH
Terahertz radiation is generated by the bunch through a variety of means. Special
cases of wakefields for short bunches are coherent transition radiation and coherent
diffraction radiation. In addition, the synchrotron radiation from an ultra-short bunch
in a bending field has a pronounced coherent component at wavelengths comparable
to the bunch length.
Coherent synchrotron radiation
The power spectrum of the incoherent synchrotron radiation is proportional to the
number of electrons, N, in the bunch whereas the coherent term is enhanced by N2, as
indicated by equation (5) from reference[13]
(5)
The power spectrum, taken from Wang [13], shown in figure 10, peaks in the
terahertz band. It is strongly dependant on bunch length so that a power meter tuned
on a narrow wavelength band, for example 0.2 mm in the case shown in figure 10,
would produce a clear signal inversely proportional to bunch length. A number of
detectors are available for THz radiation, including Schottky devices and germanium
photodetectors whose band can be extended into this range. Bandpass filters can also
be fabricated at these wavelengths, which should be centered at a wavelength
corresponding to the nominal bunch length for an application.
171
1 THz
incoherent
10
Bandpass
Filter
Detector
FIGURE
FIGURE10.
10.Power
Powerspectrum
spectrumofofcoherent
coherentsynchrotron
synchrotronradiation
radiationcalculated
calculatedfor
fortwo
twodifferent
different bunch
bunch
lengths
lengths(after
(afterWang
Wang[13]).
[13]).
Measurements
Measurements ofof the
the power
power inin the
the THz
THz band
band with
with detectors
detectors such
such as
as these,
these, or
or
bolometers,
do
not
yield
any
phase,
or
bunch
shape,
information.
The
measurement
bolometers, do not yield any phase, or bunch shape, information. The measurement
still
still performs
performs the
the useful
useful function
function ofof providing
providing an
an rms
rms measurement
measurement ofof the
the bunch
bunch
length,
which
is
useful
in
tuning
applications
along
the
length
of
the
accelerator.
length, which is useful in tuning applications along the length of the accelerator.
The
Thephase
phaseand
andarrival
arrivaltime
timeofofthe
thebunch
bunchcan,
can,however,
however,be
bemeasured
measuredby
byaafast
fastgating
gating
technique
using
femtosecond
laser
technology.
The
output
from
a
Ti:Sapphire
technique using femtosecond laser technology. The output from a TiiSapphire laser,
laser,
for
forexample,
example,isisdirected
directedonto
ontoaasemiconductor
semiconductorsubstrate
substrateof
ofGaAs
GaAsor
orradiation
radiation damaged
damaged
Si-on-sapphire,
Si-on-sapphire, generating
generating photo
photo electrons.
electrons. Metallized
Metallized electrodes
electrodes on
on the
the surface
surface
forming
formingaaHertzian
Hertziandipole
dipoleatatTHz
THzwavelengths
wavelengthsallow
allowcurrent
current toto flow
flow when
when the
the laser
laser
pulse
pulseand
andTHz
THzradiation
radiationare
arecoincident
coincidentinintime.
time.Such
Suchdevices,
devices,shown
shownininfigure
figure 11,
11,are
are
referred
referredtotoasasAuston
Austonswitches
switches[14]
[14]and
andare
areused
usedininTHz
THzpump
pumpprobe
probeapplications.
applications.
Coherent
Coherentdiffraction
diffraction radiation
radiation
The
TheTHz
THzdetection
detectiontechniques
techniquesdescribed
describedabove
abovefor
forCSR
CSRcan
can also
also be
be applied
applied to
to the
the
detection
of
coherent
diffraction
radiation
in
the
beam
pipe.
It
is
anticipated
that
detection of coherent diffraction radiation in the beam pipe. It is anticipated that aa
number
numberofofthese
theserelatively
relativelysimple
simpledevices
deviceswould
wouldbe
beused
usedalong
alongan
anFEL
PELaccelerator
accelerator to
to
aid
aidininbunch
bunchlength
lengthtuning
tuningand
andfeedback
feedback control.
control.CSR
CSRpower
power would
would be
be monitored
monitored atat
the
the bunch
bunch compressor
compressor bends
bends and
and CDR
CDR atat the
the diagnostic
diagnostic station
station preceding
preceding the
the
undulator.
undulator. Pump
Pump probe
probe timing
timing measurements
measurements with
with THz
THz radiation
radiation could
could be
be
conveniently
conveniently made
made where
where high
high bandwidth
bandwidth laser
laser light
light isis already
already available,
available, atat the
the
injector
injectorfor
forexample,
example,ororatatthe
theundulator
undulatorentrance
entrancewhere
whereFEL
PELexperiments
experiments also
also employ
employ
femtosecond
femtosecondlasers
lasersfor
forpump
pumpprobe
probecharacterization
characterizationofofthe
theradiation.
radiation.
172
FIGURE 11. An Auston switch comprises of metallized electrodes forming a dipole on a semiconductor substrate.
ACKNOWLEDGMENTS
The contributions of Paul Emma to the beam dynamics of short bunch production,
tuning and diagnostics are gratefully acknowledged.
This work is supported under the United States of America Department of Energy
contract DE-AC03-76SF00515.
REFERENCES
1. "LCLS CDR", SLAC-R-593, April 2002.
2. M. Cornacchia et al., "A Subpicosecond Photon Pulse Facility for SLAC", SLAC-PUB-8950, LCLSTN-01-7,Aug 2001. 28pp.
3. L. Bentson, P. Emma, P. Krejcik, "A New Bunch Compressor Chicane for the SLAC Linac to Produce 30-fsec,
30-kA, 30-GeV Electron Bunches", EPAC-2002, Paris, France, June 3-7, 2002. p.682-685
4. R.H. Miller, R.F. Koontz, D.D. Tsang, "The SLAC Injector", IEEE Trans. Nucl. Sci., June 1965, p.
804-8.
5. X.-J. Wang, "Producing and Measuring Small Electron Bunches", Proc. of the 1999 Part. Ace. Conf.,
New York, NY, March 1999.
6. O. H. Altenmueller, et. al., "Investigations of Traveling-Wave Separators for the Stanford Two-Mile
Linear Accelerator", The Review of Scientific Instruments, Vol. 35, Number 4, April 1964.
7. R. Akre, L. Bentson, P. Emma, P. Krejcik, "Bunch Length Measurements Using A Transverse RF
Deflecting Structure In The Slac Linac", EPAC-2002, Paris, France, June 3-7, 2002. p. 1882-1884.
8. M. Borland et al, "Start-To-End Jitter Simulations of the Linac Coherent Light Source", Proc. of the 2001 Part.
Ace. Conf., Chicago, H, June 2001, p. 2707-2709.
9. P. Emma, J. Frisch, P. Krejcik, LCLS-TN-00-12, Aug. 2000.
10. D. X. Wang, and G. A. Krafft, "Measuring Longitudinal Distribution And Bunch Length Of Femtosecond
Bunches With RF Zero-Phasing Method", Proc. of the 1997 Part. Ace. Conf., Vancouver, May 1997, p. 2020.
11. W.S. Graves et al.," Ultrashort Electron Bunch Length Measurements at DUVFEL", Proc. of the 2001 Part.
Ace. Conf., Chicago, H., June 2001, p. 2224-2226.
12. J.D. Jackson, "Classical Electrodynamics", 2nd Ed., John Wiley & Sons, New York, 1975.
13. D. X. Wang, "Electron Beam Instrumentation Techniques Using Coherent Radiation", Proc. of the
1997 Part. Ace. Conf., Vancouver, May 1997, p. 1976-1980.
14. D. H. Auston, K. P. Cheung, and P. R. Smith, "Picosecond photoconducting Hertzian dipoles", Appl. Phys.
Lett. 45 (3), p. 284-286 (1984).
173