[论文解读] Interatomic potentials: Achievements and challenges

Martin H. Müser, Sergey V. Sukhomlinov+1 more|arXiv (Cornell University)|Apr 20, 2022

Machine Learning in Materials Science被引用 21

一句话总结

A pedagogical review of empirical interatomic potentials, their classifications (two-body, many-body, open/closed shell, charge transfer, dispersion, machine-learned), and the consequences of many-body effects on properties like elasticity and phase behavior.

ABSTRACT

Interatomic potentials approximate the potential energy of atoms as a function of their coordinates. Their main application is the effective simulation of many-atom systems. Here, we review empirical interatomic potentials designed to reproduce elastic properties, defect energies, bond breaking, bond formation, and even redox reactions. We discuss popular two-body potentials, embedded-atom models for metals, bond-order potentials for covalently bonded systems, polarizable potentials including charge-transfer approaches for ionic systems and quantum-Drude oscillator models mimicking higher-order and many-body dispersion. Particular emphasis is laid on the question what constraints ensue from the functional form of a potential, e.g., in what way Cauchy relations for elastic tensor elements can be violated and what this entails for the ratio of defect and cohesive energies, or why the ratio of boiling to melting temperature tends to be large for potentials describing metals but small for short-ranged pair potentials. The review is meant to be pedagogical rather than encyclopedic. This is why we highlight potentials with functional forms sufficiently simple to remain amenable to analytical treatments. Our main objective is to provide a stimulus for how existing approaches can be advanced or meaningfully combined to extent the scope of simulations based on empirical potentials.

研究动机与目标

• Explain why interatomic potentials are essential for simulating many-atom systems.
• Survey different classes of potentials and their typical applications and limitations.
• Discuss how the functional form constrains physical properties and transferability.
• Highlight pathways to combine or extend potential forms to cover multi-bonding situations.

提出的方法

• Classify interaction potentials by bonding type, topology, and theory basis.
• Discuss construction approaches including bottom-up (quantum/DFT input) and top-down (virial expansion) designs.
• Explain the role of two-body versus many-body terms and how they relate to elastic and defect properties.
• Describe explicit versus implicit many-body potentials and their transferability implications.
• Survey representative potentials across covalent, metallic, ionic, and van der Waals systems.
• Incorporate discussion of machine-learned potentials as data-driven extensions.

实验结果

研究问题

• RQ1 What are the main classes of interatomic potentials and their characteristic features?
• RQ2 How do many-body interactions influence elastic properties, defect energies, and phase behavior?
• RQ3 What constraints does the functional form impose on transferability and physical accuracy?
• RQ4 How can different potential classes be combined or extended to handle mixed bonding environments?

主要发现

• Two-body potentials can capture general trends but often fail to describe metals and systems with significant many-body effects.
• Many-body terms are essential for correctly predicting elastic properties, defect energies, and bond order in open-shell and covalent systems.
• Cauchy relations are violated in real materials, illustrating the impact of many-body interactions on elastic tensors.
• Effective pair potentials may reproduce some bulk properties but typically lack transferability across structures or state points.
• Charge-transfer and dispersion models (including explicit many-body dispersion and polarizable approaches) improve accuracy for ionic and van der Waals systems.
• Machine-learned potentials offer a data-driven route to capture complex interactions, while preserving consistency with underlying physics.