Technology Category
Machine Learning: Hardware-aware MLKnowledge Representation and Reasoning: Other Foundations of Knowledge Representation & ReasoningData Mining & Knowledge Management: Other Foundations of Data Mining & Knowledge Management
Benchmarking Universal Interatomic Potentials on Zeolite Structures
🤖 AI Summary
The applicability of general-purpose interatomic potentials (IPs) to zeolitic porous materials—ranging from pure silica and aluminosilicates to complex Cu/K/organic-cation-containing systems—remains inadequately benchmarked. Method: This work systematically evaluates both analytical potentials (GFN-FF, UFF, Dreiding) and state-of-the-art machine learning (ML) potentials (CHGNet, ORB-v3, MatterSim, eSEN-30M-OAM, PFP-v7, EquiformerV2) against DFT (with dispersion correction) and experimental data, assessing accuracy in geometry optimization and energy prediction. Contribution/Results: GFN-FF outperforms other analytical potentials but exhibits significant errors under high strain. All ML potentials achieve DFT-level accuracy; eSEN-30M-OAM demonstrates the most robust and consistent performance across diverse compositions and structural motifs. This study establishes ML potentials as practically viable for high-throughput zeolite discovery and provides a rigorous, composition-aware guidance framework for selecting reliable IPs in multiscale modeling of multicomponent porous materials.
Technology Category
Application Category
📝 Abstract
Interatomic potentials (IPs) with wide elemental coverage and high accuracy are powerful tools for high-throughput materials discovery. While the past few years witnessed the development of multiple new universal IPs that cover wide ranges of the periodic table, their applicability to target chemical systems should be carefully investigated. We benchmark several universal IPs using equilibrium zeolite structures as testbeds. We select a diverse set of universal IPs encompassing two major categories: (i) universal analytic IPs, including GFN-FF, UFF, and Dreiding; (ii) pretrained universal machine learning IPs (MLIPs), comprising CHGNet, ORB-v3, MatterSim, eSEN-30M-OAM, PFP-v7, and EquiformerV2-lE4-lF100-S2EFS-OC22. We compare them with established tailor-made IPs, SLC, ClayFF, and BSFF using experimental data and density functional theory (DFT) calculations with dispersion correction as the reference. The tested zeolite structures comprise pure silica frameworks and aluminosilicates containing copper species, potassium, and organic cations. We found that GFN-FF is the best among the tested universal analytic IPs, but it does not achieve satisfactory accuracy for highly strained silica rings and aluminosilicate systems. All MLIPs can well reproduce experimental or DFT-level geometries and energetics. Among the universal MLIPs, the eSEN-30M-OAM model shows the most consistent performance across all zeolite structures studied. These findings show that the modern pretrained universal MLIPs are practical tools in zeolite screening workflows involving various compositions.
Problem
Research questions and friction points this paper is trying to address.
Benchmarking universal interatomic potentials on zeolite structures
Evaluating accuracy of universal potentials versus tailored ones
Assessing performance for silica and aluminosilicate systems
Innovation
Methods, ideas, or system contributions that make the work stand out.
Benchmarking universal interatomic potentials on zeolite structures
Comparing analytic and machine learning IPs with DFT references
eSEN-30M-OAM shows best performance across all zeolites
🔎 Similar Papers
