|
|
Tutorial - X-ray streak cameras
|
 |
A streak camera is a time-resolved
detector, which works similar to a TV tube. An (optical) signal
is converted into an electron pulse on a photocathode, accelerated
by a strong electric field and deflected by a fast voltage ramp
on two parallel deflector plates. Electrons arriving early see
a low electric field and are deflected less than electrons arriving
later, which are deflected more. Time information is thereby
encoded by a spatial coordinate. A temporally extended signal
is converted into a streak on the electron detector. The magnetic
lens focuses the emitted electrons. |
|
| A GaAs photoconductive switch produces the voltage
ramp on the deflection plates. The short laser pulse that triggers
the switch is produced by a titanium sapphire femtosecond laser.
The infrared (IR) pulse generates carriers in the semiinsulating
substrate and closes the switch between two Gold electrodes,
launching a ~500 V pulse. The pulse has a rise time on the order
of ~100 ps, broadened by dispersion of the cables, and a slower
trailing edge. |
  |
|
|
|
|
| Streak camera are used for time-resolved experiment
in order surpass the time resolution given by the length of the probe pulse.
In x-ray experiments at synchrotron electron storage rings the time-resolution
of stroboscopic experiments is given by the length of the x-ray pulse, which
is typically between 30 ps and 100 ps long. A streak camera can resolve
the dynamics of the sample within the x-ray pulse. In a pump-probe style
experiment a first laser pulse from a Ti:sapphire amplified laser generates
a dynamic response of the sample. A symchronized x-ray pulse probes the
state of the sample, here by being partially absorbed in the sample, and
is detected by the streak camera. The deflectors of the streak camera are
ramped by a pulse that is derived from the same laser pulse that was used
for pumping. Two more laser pulses are split off and converted to UV to
monitor the temporal resolution of the streak camera. Using x-rays a resolution
of 1-2 picoseconds has been achieved. For UV pulses a much higher time resolution
of a few 100 femtoseconds is achievable because the energy bandwidth of
emitted electrons is much reduced compared to that of x-ray generated electrons. |
 |
Synchronization to the storage ring requires
a phase-locked laser ocillator which fires laser pulses, here at 62.5
MHz, synchronuously with the 500 MHz repetition rate of the storage
ring. The laser amplifier picks and amplifies oscillator pulses at
a much lower frequency of 5 KHz, synchronously with the round trip
bunch marker of the storage ring to ascertain that always the same
electron bunch is chosen, usually a cam-shaft or single/dual bunch
pulse. The low repetition rate of the laser is necessary to provide
enough laser power to trigger the photoconductive switch and to start
the dyanmics in a millimeter size area of the sample. This scheme
has the downside that the majority of the x-ray pulses have to be
discarded and cannot be used to sample the dynamic response of the
sample. Suppression of these pulse is achieved by a DC bias on the
sweep plates and by using a gated channelplate detector.Laser pump
- x ray probe experiments are extremely challenging because of the
very low usable x-ray flux and the complex timing requirements. |
|
| A streaked x-ray pulse of about 70 ps
length is shown together with two UV fiducials. The time resolution
of this measurement was about 4 ps. To obtain good statistics, several
10000 streaks were averaged on the streak camera CCD detector at a
5 KHz repetition rate. |
 |
|
|
