The average absolute error on an individual concentration was 2.5 wt%. The
errors are, however, not random. The conentrations of the most highly
absorbing phases are systematically underestimated, while those of the
least absorbing phases are overestimated.
I believe that this is a microabsorption effect, but have not seen a
completely satisfactory treatment in the literature. I can develop
empirical correlations of the error with (mu-muaverage) or mu/muaverage, and
use these to "adjust" the calculated concentrations. This adjustment
reduces the average absolute error on an individual conentration to 1.2 wt%.
Has anyone addressed microabsorption effects in Rietveld quantitative
analysis analytically? Can corrections be made without any a priori
knowledge of the sample, so that they can be made on a routine basis?
_______________________________________
James A. Kaduk
Amoco Corporation
P.O. Box 3011, MC F-9
150 W. Warrenville Rd.
Naperville IL 60566
kaduk@amoco.com
708-420-4547 (O), 708-420-5252 (FAX)
_______________________________________
From: SMTP%"Nicholas Calos <N.Calos@mailbox.uq.oz.au>" 18-NOV-1994 07:22:57.17
Date: Fri, 18 Nov 1994 16:09:33 +1000 (GMT+1000)
From: Nicholas Calos <N.Calos@mailbox.uq.oz.au>
Subject: Re: Quantitative Phase Analysis
Dear Jim,
The issues of microabsorption contrast have been addressed by John Taylor
-Taylor and Matulis, J.Appl.Cryst. (1991) v.24 p.14- (and references
therein) in light of earlier work by Brindley (1945).
To properly correct for absorption contrast, the grain sizes of the
individual phases must be known (e.g. using SEM).
However, once known, we have found better than 1% absolute accuracy may be
achieved in good systems and typically, about 5% with no especial care.
Regards,
Nick Calos