An Experimental Examination of Error Functions for Bragg-Brentano Powder Diffractometry

Review of X-ray powder diffraction data of rhombohedral bismuth tri-iodide

“…But as is demonstrated below, the refined value may not represent the true aberration. Because no calibration material exists for the entire scanning range of a Bragg-Brentano diffractometer (Klug and Alexander, 1972) and no development of a single function for the elimination of errors has yet been achieved (King and Payzant, 1993), the use of an internal standard will not guarantee the accuracy of the lattice parameters refined. It has been pointed out recently (Toraya and Ochiai, 1994) that the whole powder pattern fitting technique with the absence of a structural model may be more suitable to lattice parameter refinement than the Rietveld analysis which imposes a structural constraint on each phase including the internal standard.…”

“…On the other hand, a Rietveld code not equipped with the cos 2 0 function, but only a sample displacement error function and goniometer-zero correction, will yield a value which deviates at least +0.0005 A from the true value. In situations where no singular error function will suffice dominant errors have to be deconvoluted (King and Payzant, 1993). Al powder 5 /xm in size mounted with DUCO cement into a 1.5 mm-deep sample holder cavity and surface polished in ethanol produced a cos 6 cot 6 plot which is shown in Figure 3.…”