A Generative Diffusion Model for Amorphous Materials

2025-07-07

AI Summary

To address the limited performance of generative models in inverse design of amorphous materials, this work proposes— for the first time—the DiffAmorph framework, a diffusion-based generative model jointly trained on computational simulations and experimental characterization data. It incorporates information-theoretic validation and conditional generation to ensure physical fidelity and controllability. The model efficiently generates diverse amorphous structures spanning broad preparation conditions, compositional spaces, and heterogeneous data sources. It enables large-scale sampling at low cooling rates and accurately reproduces short- and medium-range order as well as macroscopic properties—including ductile–brittle transitions and mesoporous features—in both silica and metallic glasses. Generation speed exceeds conventional molecular dynamics by up to three orders of magnitude, and the model supports experimental-data-driven synthesis of target structures. This work establishes the first efficient, interpretable, and cross-modal generative paradigm for rational design of amorphous materials.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Web Mining and Content Analysis: Web data generation and simulation
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Generative models show great promise for the inverse design of molecules and inorganic crystals, but remain largely ineffective within more complex structures such as amorphous materials. Here, we present a diffusion model that reliably generates amorphous structures up to 1000 times faster than conventional simulations across processing conditions, compositions, and data sources. Generated structures recovered the short- and medium-range order, sampling diversity, and macroscopic properties of silica glass, as validated by simulations and an information-theoretical strategy. Conditional generation allowed sampling large structures at low cooling rates of 10$^{-2}$ K/ps to uncover a ductile-to-brittle transition and mesoporous silica structures. Extension to metallic glassy systems accurately reproduced local structures and properties from both computational and experimental datasets, demonstrating how synthetic data can be generated from characterization results. Our methods provide a roadmap for the design and simulation of amorphous materials previously inaccessible to computational methods.

Problem

• Generative model for amorphous materials design
• Fast generation of diverse amorphous structures
• Accurate reproduction of local structures and properties

Innovation

• Diffusion model generates amorphous structures faster
• Validated short- and medium-range order recovery
• Conditional generation samples large structures efficiently