Technique and Performance of Powder Diffraction in Crystal Structure Studies

Abstract

A common use of powder diffraction data is for crystal structure studies. Since the pioneering papers of Rietveld (1967, 1969), powder diffraction has been improved in many ways. Some advances in powder diffraction techniques since an earlier review (Cheetham and Taylor 1977) are described, and an indication is given of how the Rietveld method is performing with X-ray and neutron diffraction. This method has been more popular with crystallographers than the integrated-intensity method since it attacks the superposition problem directly and allows more complex structures to be refined. It has been asserted (Sakata and Cooper 1969) that calculated e.s.d. values in the Rietveld method are low by a factor of about two, although the derived positional parameters have never been faulted. This does not negate the value of the method as corrections to the e.s.d. values can be computed (Cooper et aZ. 1981; Scott 1983). The problem of precision versus accuracy is universal; in fact most e.s.d. values published in single-crystal studies are probably low by a similar amount because of the widespread practice of omitting large amounts of 'weak' data in order to artificially lower the residuals and e.s.d. values. It is shown that powder methods, especially the Rietveld, have performed well in a variety of applications.