Strategies for achieving femtosecond
synchronization in Ultrafast Electron
Diffraction
John Byrd
R. B. Wilcox, G. Huang, L. R. Doolittle
Lawrence Berkeley National Laboratory
Workshop On Ultrafast Electron Sources For Diffraction And
Microscopy Applications
UCLA, December 14-16 2012
1
When in the Course of human events, it becomes necessary for one people to dissolve the political bands which have connected them
with another, and to assume among the powers of the earth, the separate and equal station to which the Laws of Nature and of Nature's
God entitle them, a decent respect to the opinions of mankind requires that they should declare the causes which impel them to the
separation.
We hold these truths to be self-evident, that all men are created equal, that they are endowed by their Creator with certain unalienable
Rights, that among these are Life, Liberty and the pursuit of Happiness.
That to secure these rights, Governments are instituted among Men, deriving their just powers from the consent of the governed, That
whenever any Form of Government becomes destructive of these ends, it is the Right of the People to alter or to abolish it, and to
institute new Government, laying its foundation on such principles and organizing its powers in such form, as to them shall seem most
likely to effect their Safety and Happiness. Prudence, indeed, will dictate that Governments long established should not be changed for
light and transient causes; and accordingly all experience hath shewn, that mankind are more disposed to suffer, while evils are
sufferable, than to right themselves by abolishing the forms to which they are accustomed. But when a long train of abuses and
usurpations, pursuing invariably the same Object evinces a design to reduce them under absolute Despotism, it is their right, it is their
duty, to throw off such Government, and to provide new Guards for their future security.
Check here if you agree
• We have been focused on synchronization
issues at FELs where one of the main issues is
stable timing distribution and synchronization of
remote lasers.
• I’ll try to concentrate on issues relevant to labscale experiments for UED.
2
<10fs pump/probe experiments
drive timing system design
• ≤10fs X-ray pulses already on LCLS, FLASH
• Want timing uncertainty ≤ pulse width
– Otherwise pulse is statistically widened
– Or, timing range is statistically sampled (then “binned” if measured)
– And/or shots are wasted, reducing effective reprate
pump
probe
jitter
statistics
detect timing,
“bin” data by time
wasted
shots
valid data range
3
Sources of jitter in a UED system
• Assume RF gun-based to achieve <50 fsec bunches for UED
RF
Control
Master
Clock
Laser
control
Laser
HV
Modulator
Buncher
Sample
Gun
Electron beam:
Gun voltage Amp+phase
Buncher Amp+phase
PC laser arrival time
Dispersive drift
Timing distribution:
Master clock jitter
Link jitter
Beam
diags
Laser:
Oscillator phase noise
Amplifier
4
Jitter from electron bunch compression
d  DE/E
d
s zi
‘space charge chirp’
z
late
sdi
Dtrf-laser
early
d
Dtsample
Path-Length EnergyDependent Beamline
z
Dtrf-laser
Dtrf-laser
Dtrf-laser
V = V0sin(kz)
• Relative phase jitter of the electron bunch and RF
is converted to energy jitter.
z
• The time jitter is compressed by the compression
factor
• Early and late bunches have different
compression
• Overfocused beams begin to increase time jitter.
RF field stability: low-level RF
control
RF
Control
HV
Modulator
Master
Clock
Forward, Reverse and
Cavity power probes
Buncher
Sample
Gun
Beam
diags
• Use modern digital RF controller to measure and stabilize the cavity
field.
– Feedback within RF pulse can only occur for long RF pulses >20 microseconds
– Feedback cannot control shot-to-shot variable noise from the RF source
• Modern RF controllers can achieve <10-4 amplitude and 0.01 deg
phase stability.
RF source stability
• For pulsed RF sources:
– Variable charging of the PFN delivers variation of the
high voltage to the klystron
– Variable firing of the thyratron switch
– Klystron is often run near saturation so HV variation
usually results in a phase shift.
– “Breakdown” in any part of the RF path (klystron,
SLED, waveguide, cavity, load) can cause plasma
induced reflections, phase shifts. These “breakdowns”
can be well below the limit for an RF trip and may be
already a part of “normal” operations.
7
Example: LCLS Linac (F.J. Decker)
– 0.35 deg to 0.03 deg
Un-SLEDed, HV=340kV
?
BC1: E =250 MeV
LCLS Jitter Status in 2012
HV=300kV
8
8
RF source stability
• For CW or quasi-CW RF sources:
– Klystron must be operated with some overhead to
provide feedback control
– AM/PM conversion from variable cavity tuning
– HV PS harmonics
– RF clock phase noise
9
How good does the clock have to
be?
clock
•
•
•
•
experiment
Determined by delay difference tD = tA – tB
High frequency: differential noise, frequency >1/(2tD)
Low frequency: phase delay change Dt = tD x (Df/f)
Example: 200m fiber
– tD is 1mS
– High frequency noise above 500kHz < 1fs
– Long term frequency drift < 10-9
10
Optical clocks are good enough
<0.1fs jitter above 500KHZ
Song, et al, Opt. Expr. 19, 14518 (2011)
Kubina et al, Opt. Expr. 13, 904 (2005)
RF and optical frequencies, at exact integer multiples
amplitude
•
~10-15 freq. stability
2 3 4 5...
reprate
RF
100MHZ
2e6, 2e6+1...
optical
frequency
• Commercially available
200THz
Menlo Systems
11
Pulsed lasers are naturally quiet
Er:fiber laser:
J. A. Cox et al,
Opt. Lett. 35, 3522 (2010)
• <1fs above 100kHz
– Electro-optic modulators have ~1MHz BW
12
Stabilized optical link timing
distribution
receiver
transmitter
CW
laser
wRF
AM
FS
Rb
ref
VCO or laser
RF phase
detect,
correct
wRF
optical
delay
sensing
wRF
delay error, fs
• RF clock controls remote oscillator
• ~10fs is about the limit
– 0.01 degree phase error
– 10fs at 3GHz
• Currently used in LCLS and
Fermi@Elettra
Out-of-loop resuts:
Controlling VCXO, 200m fiber
8.4fs, 20 hours to 2kHz (loop BW)
time, hours
13
Synching mode-locked lasers
with RF
n*frep
Basic Phase-locked loop
Master
Clock
Trep
BP
slave
ML Laser
Df
H
• ML Oscillator is a sub-harmonic of the clock frequency.
• Best performance if the photo-detected harmonic of oscillator
frequency is the clock frequency. Otherwise, additional frequency
multiplication is needed, reducing resolution.
• Possible AM/PM conversion at the PD
• ML oscillator is a dynamic device. Feedback response H should be
designed to dynamic response of oscillator (piezo, piezo driver, etc.)
Laser-laser synchronization
Trep
master
n*frep
n*frep
BP
BP
ML Laser
Trep
slave
ML Laser
Df
H
Detection and
bandpass filter
carrier/envelope
offset
repetition rate
0
m*frep+fceo
n*frep
frequency
Shelton (14GHz)
Bartels (456THz)
Shelton et al, O.L. 27, 312 (2002)
Bartels et al, O.L. 28, 663 (2003)
present
work (5THz)
Optimizing RF lock for ti:sapphire
laser
• Use modern control techniques
– Determine open loop transfer function
– Add filter to prevent oscillation with high gain (30kHz LPF)
Transfer function:
laser
amplitude
DAC
39kHz
resonance
ADC
step
response
phase
16
RF locking results with tisaf
• In-loop measurement compared with difference between
two externally referenced measuements
FFT of noise
In-loop:
Outofloop:
21fs RMS
1Hz to 170kHz
Integrated
RMS jitter
26fs RMS
30Hz to 170kHz
Jitter spectral density
of laser and reference
control
bandwidth
17
Effect of amplifiers on CEP
Schultze et al,
Opt. Exp. 18, 27291 (2010)
3mJ
6fs
100kHz
88as
240as
• CEP thru example optical parametric amp, 240as long term
• Dispersion changes CEP
– Carrier and envelope velocity are different
– Dispersion controlled to minimize pulse width, thus
stable
18
Out-of-loop lock diagnostics
• Compare ML phase with measured buncher
phase
RF
Control
Master
Clock
Laser
control
Laser
HV
Modulator
Buncher
Gun
Dispersive drift
Beam
diags
19
Post-sample diagnostics
• Measure electron charge, position and angle following sample
• Use deflecting cavity to measure beam-RF jitter.
• Use magnetic spectrometer to measure energy jitter. Should be
correlated to energy jitter induced by timing jitter at buncher.
20
Noise measurement and control depends
on repetition (sample) rate
• High reprate enables high bandwidth feedback
– Control BW ≈ sample rate/10
• Integrated jitter above sample rate is “shot to shot”
100Hz
100kHz
21
A high rep-rate RF gun for UED
(Daniele Filippetto)
• APEX Phase I RF gun has been built as R&D for a
high rep-rate FEL
– CW 187 MHz gun, 750 keV, 1 MHz laser rep-rate
(could be higher), low emittance
– Because of low frequency RF gun, beam dynamics
quasi-DC. 1.3 GHz buncher.
– Expected RF stability DV/V~10-4 and Df~0.01 deg
– Deflecting cavity and spectrometer diagnostics.
– High rep-rate allows for broadband RF and beambased feedback.
– If laser pump/electron probe jitter can be reduced to
<10 fsec, diffraction images can be integrated.
– Expected operation in 2013.
Parameter
Value
Energy
750
keV
Charge
1-3x105
fC
laser spot
(rms)
50-1000
μm
repetition
rate
1-106
Hz
emittance
0.03-0.6
min. bunch
length (rms)
100
22μm
fs
NGLS@Berkeley
• The eventual goal is to provide remote synchronization between all
FEL driver systems: x-rays, lasers, and RF accelerators. Our current
focus is to synch user laser systems with timing diagnostics.
PC laser
Laser heater
Timing
diagnostics
RF control
Seed lasers
User lasers
Master
NGLS Approach: RF and BB Feedback
GUN
0.8 MeV
Heater
100 MeV
BC1
210 MeV
L0
L1
Lh
CM1
CM2,3
3.9
SPREADER
2.4 GeV
BC2
685 MeV
L2
L3
CM9
CM4
CM27
CM10
Δστ
ΔE
Δστ
ΔE
ΔE
ΔEτ
SP
SP
SP
SP
CW SCRF provides potential for highly stable beams…
Measure e- energy (4 locations), bunch length (2
locations), arrival time (end of machine)
Feedback to RF phase & amplitude, external lasers
Stabilize beam energy (~10-5 ?), peak current (few
%?), arrival time (<20 fs)
Conclusions
• UED is the ideal setup for pump-probe
– Pump and probe generated by same laser
• Laser-RF stability requires careful control of RF and
laser with out-of-loop comparisons.
– Greatest potential for improvement.
– CW RF can be stabilized to DV/V~10-4 and Df~0.01 deg
– Potential for significant improvement in laser lock
• Further improvement using beam-based feedback to
stabilize source.
– High rep-rate will help.
25