| Abstract |
The “collect-before-destroy” approach, where the sample interacts with a femtosecond pulse from an X-ray free-electron laser (XFEL) before undergoing Coulomb explosion, has several benefits over experiments at conventional synchrotron radiation light sources. The short exposure time prevents radiation damage from accumulating, while the high intensity of the pulse permits useful information from the interaction to be recovered. Since experiments can be performed without cryocooling, reaction dynamics can be followed using e.g. pump-probe methods. Owing to the power of the XFEL pulses, weakly diffracting crystals or even non-crystalline, single particles can be examined, and with repetition rates up to 120 Hz large datasets can be collected in a short time. These advantages are particularly relevant for dilute radiation-sensitive samples, such as the Mn-cluster of photosystem II (PS II). However, experiments are currently far from routine, and several parameters necessary for successful analysis have to be derived from the data itself. Properties of the highly specialized equipment, including the detector and the sample injector, must be accounted for by the software pipeline, and extensive use of distributed multiprocessing is required to keep up with the high data rates. We describe cctbx.xfel, a processing suite for XFEL data, which relies on proven algorithms implemented in the Computational Crystallography Toolbox, cctbx. |
