Flow-based model generates crystal structures for materials design
Topology-Stratified Materials Discovery with A Flow-Based Generative Model
Artificial Intelligence
Summary
Designing new materials involves creating crystal structures with specific properties, which is complicated by their diverse shapes and chemical components. The authors developed a computer model called UFO-MGen that learns the detailed geometric features of crystals to generate accurate and stable structures. This model outperforms previous ones in producing unique, stable, and novel crystals, and it can be fine-tuned to target specific properties. This approach could speed up discovering new materials for tough environments like aerospace or energy systems.
What this means in practice
- •For materials engineers: Generate stable and novel crystal structures to design advanced materials for extreme environments such as aerospace and fusion reactors.
- •For additive manufacturing teams: Produce new crystal configurations with tailored properties to optimize materials used in 3D printing of high-performance parts.
Authors
Jingyi Zhou, Oyshee Chowdhury, Noah Oyeniran, Chongze Hu
Abstract
Accurate generation of crystal structures is the foundation to the discovery of high-performance materials for extreme-environment applications, such as aerospace, additive manufacturing, and fusion energy systems. Although generative modeling has emerged as a promising approach for crystal design, its performance remains limited by the complex crystal structures and diverse chemical compositions. In this work, we develop UFO-MGen, a universal flow-based generative model that learns topological features of Wyckoff representations and leverages this information to accurately generate crystals across vast structural and chemical spaces. Compared with state-of-the-art generative models, UFO-MGen achieves the highest crystal generation success rate under a rigorous multi-stability evaluation framework, the highest SUN (stable, unique, novel) rate, and a remarkable extrapolation capability that has not been reported by previous models. Furthermore, a fine-tuning module is implemented to UFO-MGen for property-constrained crystal generation, enabling the inverse materials design toward target properties. The UFO-MGen opens a new avenue for accelerated materials discovery and providing a foundation for universal materials intelligence.