engleski

Improving the Accuracy of Structures Solved by Powder Diffraction

Crystal structure models used as starting models for Rietveld refinement, may not be close enough to the global minimum to allow for convergence. More likely, the refinement will be trapped in a nearby local minimum. We suspect that this is often the case with moderately sized crystal structures of molecular compounds, thus hindering precise structure determination. Herein we describe how performing a series of optimizations from slightly different starting models may help finding the global minimum (model which gives the best fit to the measured data). The program Topas [Version 4.1, Bruker-AXS, Karlsruhe, Germany] allows the user to define new starting values of parameters (by e.g. adding a random small shift) before a new refinement cycle is initiated. Since molecular geometry is distorted after the shifts are applied, setting the proper values of geometry restraints seems to be crucial. By applying this method the system explores the space around the initial model which allows for the determination of the global minimum with better reliability (Fig. 1). We will present refinements on two moderately sized compounds, for both of which lower minima were found after this approach was applied.

powder diffractiom ; accuracy ; simulated annealing ; Rietveld refinement

nije evidentirano