Computational Screening of New Perovskite Materials Using Transfer Learning and Deep Learning

Abstract

As one of the most studied materials, perovskites exhibit a wealth of superior properties that lead to diverse applications. Computational prediction of novel stable perovskite structures has big potential in the discovery of new materials for solar panels, superconductors, thermal electric, and catalytic materials, etc. By addressing one of the key obstacles of machine learning based materials discovery, the lack of sufficient training data, this paper proposes a transfer learning based approach that exploits the high accuracy of the machine learning model trained with physics-informed structural and elemental descriptors. This gradient boosting regressor model (the transfer learning model) allows us to predict the formation energy with sufficient precision of a large number of materials of which only the structural information is available. The enlarged training set is then used to train a convolutional neural network model (the screening model) with the generic Magpie elemental features with high prediction power. Extensive experiments demonstrate the superior performance of our transfer learning model and screening model compared to the baseline models. We then applied the screening model to filter out promising new perovskite materials out of 21,316 hypothetical perovskite structures with a large portion of them confirmed by existing literature.

Machine learning for perovskite materials design and discovery

“…In addition to E hull , the formation energy of compounds could also be used to evaluate the formability and stability of perovskite. Li et al 63 proposed a transfer learning strategy to evaluate the stability of the ABX 3 inorganic perovskites. First, an ML transfer learning model was constructed by taking the formation energies of 570 perovskites as the target variable and the physics-informed structural and elemental parameters of perovskites as descriptors.…”

RbSnX3 (X = Cl, Br, I): promising lead-free metal halide perovskites for photovoltaics and optoelectronics

“…Temperature-dependent structural phase transitions are very common in halide perovskites. 5–7 Different perovskites have been found in multiple crystallographic phases at different temperatures and the common phases are cubic, tetragonal, and orthorhombic. 6,7 However, the cubic phase is the most common phase among the perovskite materials.…”

Is Domain Knowledge Necessary for Machine Learning Materials Properties?

“…An exemplar of the CBFV method is the Magpie [1] descriptor. This domain-derived approach (CBFV) has been successfully employed in materials informatics studies for years [2][3][4][5][6][7]. Not only has it been successful, but the information it contains is also human-readable, allowing for physically interpretable results.…”