Resolution of a discrepancy of x-ray attenuation measurements of silicon in the energy range 25-50 keV

Precise measurements of the x-ray attenuation coefficient of crystalline silicon were made in the energy range 24 to 50 keV in 1985. As we show in this short paper, these measurements are of the highest precision currently available (1%-1.5%) for these energies. However, comparisons with theory were unable to resolve a residual discrepancy which reached seven standard deviations of the experimental precision over this range of energies. The most likely cause of the discrepancy was thought to lie in the estimation of the thermal diffuse scattering cross-section. We show that the dominant factor was, instead, the accuracy of the theory for the photoelectric component of the attenuation. Comparison with theory based on Chantler's work shows agreement with experiment to within one standard deviation.