Specialising neural network potentials for accurate properties and application to the mechanical response of titanium

Large scale atomistic simulations provide direct access to important materials phenomena not easily accessible to experiments or quantum mechanics-based calculation approaches. Accurate and efficient interatomic potentials are the key enabler, but their development remains a challenge for complex materials and/or complex phenomena. Machine learning potentials, such as the Deep Potential (DP) approach, provide robust means to produce general purpose interatomic potentials. Here, we provide a methodology for specialising machine learning potentials for high fidelity simulations of complex phenomena, where general potentials do not suffice. As an example, we specialise a general purpose DP method to describe the mechanical response of two allotropes of titanium (in addition to other defect, thermodynamic and structural properties). The resulting DP correctly captures the structures, energies, elastic constants and γ-lines of Ti in both the HCP and BCC structures, as well as properties such as dislocation core structures, vacancy formation energies, phase transition temperatures, and thermal expansion. The DP thus enables direct atomistic modelling of plastic and fracture behaviour of Ti. The approach to specialising DP interatomic potential, DPspecX, for accurate reproduction of properties of interest “X”, is general and extensible to other systems and properties.

Pushing the boundaries of lithium battery research with atomistic modelling on different scales

Computational modelling is a vital tool in the research of batteries and their component materials. Atomistic models are key to building truly physics-based models of batteries and form the foundation of the multiscale modelling chain, leading to more robust and predictive models. These models can be applied to fundamental research questions with high predictive accuracy. For example, they can be used to predict new behaviour not currently accessible by experiment, for reasons of cost, safety, or throughput. Atomistic models are useful for quantifying and evaluating trends in experimental data, explaining structure-property relationships, and informing materials design strategies and libraries. In this review, we showcase the most prominent atomistic modelling methods and their application to electrode materials, liquid and solid electrolyte materials, and their interfaces, highlighting the diverse range of battery properties that can be investigated. Furthermore, we link atomistic modelling to experimental data and higher scale models such as continuum and control models. We also provide a critical discussion on the outlook of these materials and the main challenges for future battery research.

First-Principle and Atomistic Modelling in Materials Science

In the last two decades, the importance of Computational Materials Science has continuously increased due to the steadily growing availability of computer power [...]