Machine Learning Elastic Strain Engineering

PI: Ju Li, Department of Nuclear Science and Engineering, MIT
PI: Alexander Shapeev, Center for Computational and Data-Intensive Science and Engineering, Skoltech 

Nanostructured materials can withstand much higher elastic strain without mechanical failure than their conventional counterparts, opening up a huge parameter space for rational engineering of material properties. This project aims to systematically study the effect of high strain and strain gradient in inducing changes of technologically relevant properties in semiconductor crystals, including band gap, band structure and nonlinear optical susceptibilities. We will use high-throughput ab initio calculations to compute such properties in strain space, up to the ideal elastic strain limit of a crystal, and develop advanced machine learning modules to efficiently represent the dependence of material properties on strain and strain gradient. The planned machine learning modules will allow us to detect critical phenomena as high strain is imposed on a material, including direct-indirect bandgap transition, metal-insulator transition, as well as band inversion and topological phase transition. These critical phenomena are promising in forming the basis of next-generation electronic and optoelectronic devices. The machine learning modules will also allow us to elucidate the fundamental correlation between the strain gradient and enhanced higher-order nonlinear optical response in centrosymmetric semiconductors, leading to design principles of new device architectures that result in optimal nonlinear optical responses in strained semiconductors.

Back to the list >>