Mössbauer spectroscopy of minerals

We compare two different approaches to fitting the Fe-57 Mössbauer spectra of paramagnetic state minerals: the widespread practice of using Lorentzian line doublets with adjustable Γ widths and a new method (Rancourt and Ping, 1991) based on quadrupole splitting distributions (QSDs). We argue that there is no physical justification for the former and that the latter is a theoretically correct approach. With the same number of free parameters, the QSD method performs better. The Lorentzian doublet method 1) significantly overestimates the background, 2) puts overly large wings or tails on the main absorption peaks, and 3) gives unphysically large values of Γ. Whereas Lorentzian doublets are often arbitrarily (and incorrectly) assigned to model sites in the average unit cell, QSDs represent the most information that can be distilled from the spectra and are caused by the various local and long-range structural and chemical environments acting concomitantly.