Rotation of X-ray polarization in the glitches of a silicon crystal monochromator

EXAFS studies on dilute samples are usually carried out by collecting the fluorescence yield using a large-area multi-element detector. This method is susceptible to the 'glitches' produced by all single-crystal monochromators. Glitches are sharp dips or spikes in the diffracted intensity at specific crystal orientations. If incorrectly compensated, they degrade the spectroscopic data. Normalization of the fluorescence signal by the incident flux alone is sometimes insufficient to compensate for the glitches. Measurements performed at the state-of-the-art wiggler beamline I20-scanning at Diamond Light Source have shown that the glitches alter the spatial distribution of the sample's quasi-elastic X-ray scattering. Because glitches result from additional Bragg reflections, multiple-beam dynamical diffraction theory is necessary to understand their effects. Here, the glitches of the Si(111) four-bounce monochromator of I20-scanning just above the Ni K edge are associated with their Bragg reflections. A fitting procedure that treats coherent and Compton scattering is developed and applied to a sample of an extremely dilute (100 micromolal) aqueous solution of Ni(NO₃)₂. The depolarization of the wiggler X-ray beam out of the electron orbit is modeled. The fits achieve good agreement with the sample's quasi-elastic scattering with just a few parameters. The X-ray polarization is rotated up to ±4.3° within the glitches, as predicted by dynamical diffraction. These results will help users normalize EXAFS data at glitches.Rotation of X-ray polarization at the glitches of a monochromator composed of single crystals of silicon is observed. This effect can be explained by a model taking full account of the X-ray source, the effects of multiple-beam dynamical diffraction, and the coherent and Compton scattering from the sample.