High-throughput Computations in Materials Science 

Mingli Yang*

Research Center for Materials Genome Engineering, Sichuan University, Chengdu 610065, China

ABSTRACT: High-throughput computation (HTC) is one of the key techniques in materials genome engineering (MGE). Utilizing high-performance computers and integrating multi-scale computational tools, HTC develops various automatic workflows to realize the design, computation and screening of target materials, provides information on their microstructures, preparing and processing conditions, physiochemical properties, in-service performances and lifetime, and directs the preparation, characterization, manufacture and utilization of novel materials.

Computations on materials usually start with well-designed computational models extracted from complex materials, predicting their structures and properties by numerical computations, which is considered as one kind of high-performance computations. HTC for materials (MHTC) is however a combination of conventional materials computations and high-throughput computations. The multi-component configurations, multi-scale hierarchical structures, multi-step preparing and processing, and complex service environments generate a huge number of computational models, a typical compute-intensive task particularly fit for HTC.

MHTC completes virtually “trials and errors” that are conventionally completed by experiments, saving cost and time by dramatically reducing experimental effort. In this sense, MHTC can be regarded as an engineering application of computational materials. MHTC is characterized by: (1) Multi-factor design and screening. (2) Automatic work flows. (3) Efficient computations. (4) Data-integrated pre- and post-computations.

Some established MHTC platforms have made great contribution in materials discovery. Supported by National Key Research and Development Program, MHTC techniques and platforms developed fast in China in the past four years. MHTC will proceed in the direction of computational platforms for novel materials design for screening, computational tools and workflows for specific materials, and integration with big data and artificial intelligence in the coming years.

Acknowledgement: National Key Research and Development Program “Key Techniques and Supporting Platforms of Material Genome Engineering” and all researchers in this program.