Search for content and authors |
Modeling crystallites and r.m.s. microstrain distributions with anisotropic broadening in the Rietveld method |
Luca Lutterotti |
Department of Material Engineering and Industrial Tecnology, University of Trento (DIMTI), v. Mesiano 77, Trento 38100, Italy |
Abstract |
Rietveld refinement programs even if they may include some fundamental parameter approach to model instrument broadening, they always rely on a functional description of line broadening due to microstructural characteristics. For anisotropic broadening instead they include either the description of Popa [1] or Stephens [2] to account for it. A methodology is presented here as used in the Maud Rietveld program to model line broadening using directly crystallite size and r.m.s microstrain distributions without resorting to any line profile function. The peak profile is computed directly from the convolution of the two distributions and finally convoluted also with the instrumental part. Combination of lognormal and gamma distributions are used for the crystallites and an appropriate function is used for the microstrain distribution as described by Lutterotti et al. [3]. The methodology follow closely the description in [3] but has been adapted to the Rietveld method incorporating also a general description of the anisotropic crystallite shape and microstrain in analogy to what presented by Popa and Balzar in [1]. The methodology prove to be robust for Rietveld refinement and the computation speed is sufficient for normal refinements. The method has been successfully applied to different samples and in some cases the fitting quality improves significantly respect to an equivalent approach (even using anisotropic broadening) constrained by functional peak description as used traditionally by Rietveld programs. When applied to the broadened ceria sample of the Size-Strain-Round-Robin for the ESRF BM16 beamline data the Rwp of the fitting drop from 4.6% to 3.9% using distributions instead of peak functions to model the broadened profile (see figure). The size-strain results using the distribution modeling are in agreement with the assessed results, not showing problems related to gaussian crystallite broadening as in the case of peak function fitting. [1] Popa N. and Balzar D. (2005). Acta Cryst., A61, C79. [2] Stephens P. W. (1999). J. Appl. Cryst., 32, 281-289. [3] Lutterotti L. and Scardi P. (1992). Advances in X-Ray Analysis, 35A, 577-584. |
Legal notice |
|
Related papers |
Presentation: Poster at 11th European Powder Diffraction Conference, Poster session, by Luca LutterottiSee On-line Journal of 11th European Powder Diffraction Conference Submitted: 2008-04-30 23:06 Revised: 2009-06-07 00:48 |