Evaluation of X-ray Camera As a Tool for Automated Beam Characterization

Abstract: Several methods for analysing materials and proteins use highly concentrated beams of X-rays, e.g. SAXS and X-ray crystallography. To evaluate the outgoing beam, it is of high importance to know the light distribution of the incoming beam. Previously, a method for this has been to focus the X-ray beam onto a pinhole in front of a photodiode, a so called pinhole measurement. Although this method gives information about the radial distribution of the beam, it is very time-consuming. In this report a faster alternative has been developed and evaluated. In this new method an image is taken with an X-ray camera in the focus of the beam. Algorithms are then used to replicate a pinhole measurement by applying virtual pinholes. Different pixels in an image act differently, referred to as spatial noise. This must be compensated for before information about the beam may be extracted. To do this, the camera noise was characterized and a calibration procedure developed for its minimization. It was shown that the spatial noise was greatly reduced, making the temporal shot noise the new largest noise source. Although the noise was successfully reduced, the calibration procedure failed to accurately remove all signal not originating from registered photons. Measurements done with low photon intensities, large exposure times or at high temperatures are therefore less accurate. The measured camera signal was transformed into incident photon intensity using a responsivity proportionality constant. This constant was estimated by comparing the results from real and virtual pinhole measurements for several photon intensities and pinholes. The results gave a responsivity proportionality constant of 0,03 DN/X-ph. Further measurements were done concerning the temperature dependence of the camera responsivity and to investigate possible bleaching. The results indicated that the responsivity was held constant under changing temperatures and that the camera remained unbleached during the 114h long measurement. Finally, real and virtual pinhole measurements were done for a series of pinholes and compared using the responsivity proportionality constant. A maximum relative deviation of 6% was measured between the two, indicating that virtual pinhole measurements give accurate results. The largest deviations of the measurement seem to occur when using small or large pinholes. These errors, however, have a high potential of being further minimized, resulting in higher accuracy.