RESEARCH ON OUR SITE
 
 

Towards femtosecond diffraction with the XPAD pixel detector

SOLEIL Company Contents > All the news > News 2010 > Towards femtosecond diffraction with the XPAD pixel detector

SOLEIL recently launched an ambitious project which aims at producing synchrotron radiation pulses of 100 femtoseconds width: the femto-slicing project. The first beamlines concerned will be TEMPO (photo-emission) and CRISTAL (diffraction). CCPM (Marseille), the CRG D2AM beamline at ESRF and the Detectors group at SOLEIL have designed and produced one of the detectors needed for the future experiments carried out on CRISTAL beamline.

The femto-slicing project is based on the interaction between a 30 fs infrared laser pulse (1 fs = 10-15 s) and one of the electron bunches that rotate in the storage ring of the synchrotron, whose spatial distribution corresponds to a duration of several tens of ps (1 ps = 10-12 s). When it takes place under suitable conditions, this interaction leads to a change in energy of the electrons in a fs-slice of the bunch, which in turn causes a slight change in their trajectory. The ‘sliced’ electrons thus radiate fs x-ray pulses which can be spatially separated from the regular ps-pulses in a beamline.
Obtaining short (ps) or ultra-short (fs) pulses enables to probe the dynamics of atomic or molecular states, collective movements in solids, or the course of chemical reactions, as they occur after a fs laser excitation. In the so-called ‘pump-probe’ experiment scheme, the system studied is excited (pumped) by a fs laser pulse and probed after a chosen delay by a pulse of a different ‘colour’, such as x-rays. This procedure must be repeated in a stroboscopic way, in order to generate the necessary statistics and reliability of measurement. Photo-induced metastable states are generally investigated through this type of experiments, since they quickly relax to the ground state (i.e., in less than 1 ms). The temporal resolution of the experiment is mainly determined by the length of the probe pulses (ps or fs).

Pump-probe experiments are already carried out at synchrotron radiation centres, where the beam delivered is naturally pulsed due to the ‘bunch’ structure of the electron beam in the storage ring. At SOLEIL in the ‘8-bunch’ filling mode, eight electron bunches of 70 ps duration (full-width at half-maximum, FWHM) go around the ring within 1.18 µs, each of them being separated from the others by a time interval of 147 ns. In the framework of the femto-slicing project, these pulses will reach a width of 100 fs FWHM. The first beamlines concerned will be the photo-emission beamline TEMPO and the diffraction beamline CRISTAL.

On CRISTAL, a diffraction beamline with an undulator source, the pump-probe experiments will consist in (1) exciting a solid using a laser pulse, and (2) measuring after a chosen delay the diffraction pattern produced by a short or an ultra-short x-ray pulse. Modern diffraction methods take in use two-dimensional (2D) detectors, which allow several diffraction spots to be measured at once. To perform time-resolved experiments, only the x-ray pulse that immediately follows the laser pulse should be counted (and therefore detected). Since the pump laser has a frequency of 1 to 10 kHz, the detector must be activated on an x-ray pulse every 0.1 to 1 ms, whilst avoiding measuring the neighbouring x-ray pulses. In the ‘8-bunch’ mode at SOLEIL for example, this means that the counting of diffracted photons must be enabled for no longer than 2x147~300 ns. The 2D detectors normally used in diffraction, CCD cameras or image plates, cannot switch that quickly. This is the reason why the selection of the x-ray pulse to be measured is usually performed by a mechanical selector, or ‘chopper’. Unfortunately choppers are limited to a maximum frequency of 1 kHz, and cannot be flexibly tuned in frequency, Moreover, they involve a very complex mechanical design (rotation at supersonic speed, requiring operation in a vacuum and with very severe constraints concerning stability).

In collaboration with the CCPM (Marseille) and the CRG D2AM beamline at ESRF, the Detectors group at SOLEIL currently develops a hybrid pixel 2D detector. One of the imagers available for the experiments at SOLEIL consists of eight modules of seven hybrid integrated circuits (chips) on a single silicon sensor (see Figure 1). Each of these chips consists of 80 x 120 pixels measuring 130 µm along each side, and the complete imager has a footprint of 7.28 cm x 12.48 cm.

Imageur basé sur le chip XPAD3

Figure 1 : Imager based on the XPAD3 chip. The detection surface, consisting of eight modules with seven chips each, has a footprint of 7.28 cm x 12.48 cm.

The 3.2 version of this detector has been optimised to be able to count during a period specified by the width of a logic gate, at any frequency. The rest of the time, the counting of photons is disabled. A chip from this detector was tested on the CRISTAL beamline in the machine’s 8-bunch mode. A piece of Teflon placed in the incident x-ray beam was used as a scatterer. Figure 2 represents the scattered intensity measured by the XPAD3.2 chip and averaged over its 9600 pixels as a function of the length of the gate. The measurement was performed at the frequency of revolution of the electrons in the ring, i.e. 847 kHz, for two energies commonly used in diffraction: 12 keV and 24 keV. The staircase graph shows that it is possible to enable and disable the detector quickly enough to detect in a controlled way the scattering that originates from only one x-ray pulse, or from a small number of these pulses. The rise time from one step to the following one is only 80 ns, which shows that the synchronisation of the electronics of each pixel, on the one hand, and the other phenomena causing a fluctuation of the time response of the detector, on the other hand, are controlled on this timescale. The XPAD3.2 detector can thus easily be used to select one x-ray pulse in the 8-bunch or hybrid modes of SOLEIL, among the x-ray pulses spaced at 147 ns intervals.

Figure 2 : Total intensity measured as a function of the width of the detection gate at 12 keV and 24 keV. The lower diagram shows (in orange) the eight x-ray pulses at intervals of 147 ns, and the gate widths at the times indicated by the pink arrows.

Figure 3 shows a part of the diffraction pattern for Teflon obtained on a chip at 12 keV with an 80 ns gate opened at a 10 kHz frequency, for a total exposure time of 85 s. These are the conditions under which picosecond pump-probe experiments will be carried out at SOLEIL. It should be noted that the intensity diffracted by Teflon is comparable to the weak diffuse scattering signals that are observed for example in crystals close to phase transitions. The XPAD3.2 detector will therefore allow for detailed studies of photo-induced phase transition mechanisms. In addition the same gated detection mode can be used in the context of the femto-slicing project, in order to reduce the background noise due to bunches that did not interact with the laser. These results are therefore very encouraging for the pico- and femto-second diffraction experiments at SOLEIL.

Figure 3 : Part of the Teflon diffraction diagram measured at 12 keV with the XPAD3.2 chip (1 cm x 1.5 cm) placed at 50 cm away from the sample. The measurement frequency is 10 kHz and the gate opening period is 80 ns. The colour scale is in number of photons.

Accueil