Stochastic fitting of specular X-ray reflectivity data using StochFit

• Published in: Journal of applied crystallography 2008-12, Vol.41 (6), p.1187-1193
• Main Authors: Danauskas, Stephen M., Li, Dongxu, Meron, Mati, Lin, Binhua, Lee, Ka Yee C.
• Format: Article
• Language: English
• Subjects: ALGORITHMS; COMPUTER CALCULATIONS; computer programs; COMPUTERIZED SIMULATION; Crystallography; Density; ELECTRON DENSITY; electron density distribution; Electrons; INORGANIC, ORGANIC, PHYSICAL AND ANALYTICAL CHEMISTRY; INTERFACES; MATHEMATICAL METHODS AND COMPUTING; REFLECTIVITY; S CODES; specular X-ray reflectivity; Stochastic models; StochFit; THIN FILMS; X-RAY DIFFRACTION; X-rays
• ISSN: 1600-5767; 0021-8898; 1600-5767
• DOI: 10.1107/S0021889808032445

Specular X‐ray reflectivity data provide detailed information on the electron density distribution at an interface. Typical modeling methods involve choosing a generic electron density distribution based on an initial speculation of the electron density profile from the physical parameters of the experimental system. This can lead to a biased set of solutions. StochFit provides stochastic model‐independent and model‐dependent methods for analyzing X‐ray reflectivities of thin films at an interface. StochFit divides an electron density profile into many small boxes and stochastically varies the electron density of these boxes to locate the best fit to a measured reflectivity. Additionally, it provides the ability to perform model‐dependent fitting with a stochastic search of the parameter space in order to locate the best possible fit. While model‐independent profile search algorithms have been described previously, they are difficult to implement because of the heavy computational requirements, and there is a dearth of software available to the general scientific public utilizing these techniques. Several cases that illustrate the usefulness of these techniques are presented, with a demonstration of how they can be used in tandem.