Ct. Chantler et Rd. Deslattes, SYSTEMATIC CORRECTIONS IN BRAGG X-RAY-DIFFRACTION OF FLAT AND CURVED CRYSTALS, Review of scientific instruments, 66(11), 1995, pp. 5123-5147
Measurements of spectral wavelengths in Bragg diffraction from crystal
s often require refractive index corrections to allow a detailed compa
rison of experiment with theory. These corrections are typically 100-3
00 ppm in the x-ray regime, and simple estimates may sometimes be accu
rate to 5% or better. The inadequacies of these estimates are discusse
d. Even with a possibly improved index of refraction estimate, this co
rrection is insufficient since additional systematics in the diffracti
on process occur at or above this level. For example, asymmetries of d
iffraction profiles with ir-polarized radiation or due to three-beam d
iffraction can approach the magnitude of refractive index corrections
for flat or curved crystals. The depth of penetration of the x-ray fie
ld inside curved crystals, the shift of the mean angle to the diffract
ing planes, and lateral shifts around the crystal surface are rarely c
onsidered but can dominate over:refractive index corrections, particul
arly for high-order diffraction or medium-energy x rays. Shifts and no
nlinearities arise when diffracting surfaces lie off the Rowland circl
e, and exhibit strong and rapidly varying angular dependencies. Johann
geometries with the source located on the Rowland circle should be av
oided to minimize profile truncation shifts from crystal ranges or min
imum grazing angles, and to avoid extreme scaling corrections. Other s
ignificant shifts are identified and illustrated, with functional rela
tions provided to allow an estimation of related magnitudes. The centr
al concerns of this paper are the effects on flat crystal diffraction
and curved crystal diffraction in the Johann geometry, with a source a
nd crystal of variable dimensions and location. Experiments often inte
rpolate or extrapolate from calibration lines, so dependencies upon th
e diffracting angle are as important as the magnitude of the correctio
ns. These dependencies are presented in formulas and graphs. (C) 1995
American Institute of Physics.