Materials generation improves by fixing crystal framework shapes
Framework-Conditional Open Materials Generation
Machine Learning
Summary
Predicting new materials that are stable and unique is tough because many methods either rely on known structures or try to guess everything from scratch. The authors created a new method that generates materials by keeping the framework shape fixed and only changing the chemical parts inside, which makes the problem easier and more flexible. Their approach finds new material structures more often without needing to learn complex atomic arrangements. They also examined existing data sets to understand how common duplicate structures are.
What this means in practice
- •For materials scientists: Generate novel stable inorganic materials by fixing known crystal frameworks and varying compositions to find new stable structures efficiently.
- •For chemical engineers: Design new material candidates faster by using framework-conditional generation to reduce complexity while exploring stable compositions.
Authors
Maya M. Martirossyan, Hillary Pan, Philipp Höllmer, Stefano Martiniani
Abstract
Theoretical predictions of novel materials rely on a variety of techniques, from data-mining-based ion substitution methods to fully de novo generative models. Each approach imposes specific constraints and encounters unique challenges when extrapolating beyond the distribution of known materials to discover new ones. The foremost goal of materials generation is to produce stable, unique, and novel (S.U.N.) structures. Predicted structures that meet these criteria, however, typically result from novel chemical compositions mapped onto known structural prototypes or frameworks. We leverage this heuristic insight to present a new generative paradigm for the de novo prediction of inorganic crystalline materials constrained by their structural framework---i.e., the unit cell shapes and respective particle positions. Our method, termed OMatG-FC (Open Materials Framework-Conditional Generation), learns via a combination of discrete flow matching on the atomic species and stochastic interpolants for the volume scaling of the unit cell lattice, while fixing in-place the fractional coordinates and the unit-volume lattice vectors, dramatically reducing the dimensionality of the problem compared to unconstrained (framework-free) generation. Our method is more flexible than existing prototype-constrained methods, demonstrating capability for discovery of unseen, stable prototypes derived from known structural frameworks. We also investigate the number of unique frameworks in the MP-20 dataset, demonstrating the propensity of duplicate frameworks within the dataset and providing a framework-split version. We benchmark the performance of our model using the LeMat-GenBench suite using three ML interatomic potentials and achieve state-of-the-art S.U.N. rate below the convex hull, eschewing the need to learn atomic packings for materials discovery.