Classification and statistical analysis of structural disorder in crystalline materials

Approximately 50% of entries in the Inorganic Crystal Structure Database (ICSD; https://www.fiz-karlsruhe.de/en) exhibit some form of structural disorder. This work aims to provide a thorough analysis of structurally disordered materials within the ICSD, using data extracted from crystallographic information files. To achieve this, we derive a classification of structurally disordered crystalline materials described by their spatially averaged structures and introduce a range of quantitative measures of structural disorder. The overarching aim of this classification and analysis is to facilitate high-throughput and machine learning studies of disordered materials. To demonstrate the application of our approach, we perform statistical analysis of the disordered compounds reported in the ICSD to identify general trends in the distribution of disorder across different chemical elements, structures and classes of materials.