Department of Pure and Applied Chemistry , Faculty of Science and Technology , 2641 Yamazaki , Noda-shi , Chiba 278-8510 , Japan.
ACS Comb Sci. 2020 Jan 13;22(1):35-41. doi: 10.1021/acscombsci.9b00158. Epub 2019 Dec 16.
Automated structural analysis techniques are required to accelerate materials research. In this study, we developed an algorithm to automate Rietveld analysis, which is a method for crystal structure refinement using powder diffraction patterns. This algorithm features the repeated generation of a set of initial values, followed by one-shot refinement. Accurate results were obtained without any strategy for the sequence of refinement, as is often used in manual analysis. Implementation and testing of the automated algorithm provided fitting results that were comparable to those of manual analysis, even when inaccurate initial values for structural parameters were input. Moreover, the much shorter time was required for the developed automatic analysis method than for manual analysis. The developed method will likely facilitate the analysis of large amounts of diffraction data, allowing the accumulation of structural data that can enhance the efficacy of materials research.
自动化的结构分析技术是加速材料研究的必要手段。在本研究中,我们开发了一种自动化 Rietveld 分析的算法,该方法可用于使用粉末衍射图谱进行晶体结构精修。该算法的特点是重复生成一组初始值,然后进行单次精修。无需使用通常在手动分析中使用的精修顺序策略,即可获得准确的结果。自动化算法的实现和测试提供了与手动分析相当的拟合结果,即使输入了结构参数的不准确初始值也是如此。此外,与手动分析相比,开发的自动分析方法所需的时间要短得多。所开发的方法可能会促进大量衍射数据的分析,从而积累可以增强材料研究效果的结构数据。