Neural Network Helps Find Unknown Patterns in Halogen Crystals
Russian researchers have developed an approach for analyzing crystal structure using neural networks. It has revealed previously unknown patterns in compounds of chlorine, bromine, and iodine.
The Russian Science Foundation's press service reported the development. The approach helps identify connections that are difficult to detect through standard search. Chemists from St. Petersburg State University and Samara State Technical University combined topological assessment of crystals with AI tools. The program determines the proximity of atoms based on the geometry of the substance, and the neural network identifies statistical patterns in combinations of contacts. The approach was tested on around five hundred crystals containing ions of chlorine, bromine, iodine, and other halogens. The analysis helped uncover previously unknown variants of atomic interactions, whose nature remains to be studied using quantum chemical calculations.
