Scale-invariant Machine-learning Model Accelerates the Discovery of Quaternary Chalcogenides with Ultralow Lattice Thermal Conductivity

Abstract

Intrinsically low lattice thermal conductivity (κ l ) is a desired requirement in many crystalline solids such as thermal barrier coatings and thermoelectrics. Here, we design an advanced machinelearning (ML) model based on crystal graph convolutional neural network that is insensitive to volumes (i.e., scale) of the input crystal structures to discover novel quaternary chalcogenides, AMM'Q3 (A/M/M'=alkali, alkaline-earth, post-transition metals, lanthanides, Q=chalcogens). Upon screening the thermodynamic stability of ∼ 1 million compounds using the ML model iteratively and performing density functional theory (DFT) calculations for a small fraction of compounds, we discover 99 compounds that are validated to be stable in DFT. Taking several DFT-stable compounds, we calculate their κ l using phonon-Boltzmann transport equation, which reveals ultralow-κ l (< 2 Wm −1 K −1 at room-temperature) due to their soft elasticity and strong phonon anharmonicity. Our work demonstrates the high-efficiency of scale-invariant ML model in predicting novel compounds and presents experimental research opportunities with these new compounds. K.P. and C.W.P. contributed equally.

Toward machine learning for microscopic mechanisms: A formula search for crystal structure stability based on atomic properties

“…In the last few years, materials scientists have developed different machine-learning (ML) methods to rationalize the data analysis. [23][24][25][26][27][28][29][30][31][32][33] Each method has its own specific advantages and limitations. Methods, such as random forest 34 or neural network (NN), 35 which is mainly behind the Deep Learning (DL), are very efficient 36 but not always transparent, partially blurring the comprehension of the role played by the input variables in the final results.…”