Verification of a method for determining the degree of crystallinity using experimental and computer-generated powder diffraction patterns

Abstract

A new procedure for determining the degree of crystallinity (DOC) has been developed. It is a two-step procedure consisting of (i) the separation of the observed pattern of a partially crystallized target material into crystalline and non-crystalline parts based on the Rietveld method, and (ii) the determination of the DOC based on the Ruland method. In the latter step, a disorder factor in Gaussian form with a disorder parameter is used to redistribute the total intensity between the crystalline and non-crystalline parts. Then, the DOC and disorder parameters are optimized by minimizing the cost function under the constraint of a constant intensity distribution between the two parts irrespective of the angular interval of integration. A self-contained technique for the background/halo pattern correction has also been developed and used between the two steps. Adoption of the two-step procedure could avoid parameter correlation between 2θ-dependent functions and provide robustness. The final goal is to determine the DOC of real materials such as polyphenylene sulfide composites with crystalline and non-crystalline fillers. The present study is a first step, and its purpose is to verify the efficiency of this method. Experimental and computer-generated intensity datasets of artificially mixed patterns of known weight fractions were used for testing the procedure. The DOC could be determined routinely by the procedure presented here within an error of 1–2% for samples with DOC < 50%.