Lattice Thermal Conductivity: An Accelerated Discovery Guided by Machine Learning

Abstract

In the present work, we used machine learning (ML) techniques to build a crystal-based model that can predict the lattice thermal conductivity (LTC) of crystalline materials. To achieve this, first, LTCs of 119 compounds at various temperatures (100− 1000 K) were obtained based on density functional theory (DFT) and phonon calculations, and then, these data were employed in the next learning process to build a predictive model using various ML algorithms. The ML results showed that the model built based on the random forest (RF) algorithm with an R 2 score of 0.957 was the most accurate compared with the models built using other algorithms. Additionally, the accuracy of this model was validated using new cases of four compounds, which was not seen for the model before, where a good matching between calculated and predicted LTCs of the new compounds was found. To find candidates with ultralow LTCs (<1 W m −1 K −1 ) at room temperature, the model was used to screen compounds (32116) in the Inorganic Crystal Structure Database. From the screened compounds, Cs 2 SnI 6 and SrS were selected to validate the ML prediction using the counterpart theoretical calculations (DFT and phonon), and it was found that the outcome behaviors by both methods (ML prediction and DFT/phonon calculations) are fairly consistent. Considering the type of employed feature, the prime novelty in this work is that the built model can credibly predict the LTC−temperature behaviors of new compounds that are constructed based on prototype structures and chemical compositions, without the use of any DFT-relaxed structure parameters. Accordingly, using the periodic table, prototype structures, and the RF-based model, the LTC−temperature behavior of a huge number of compounds can be predicated.

Machine Learning Accelerated Discovery of Promising Thermal Energy Storage Materials with High Heat Capacity

“…The RF is created by the random feature selections and bagging as introduced by Breiman, which helps to reduce overfitting. The performance of the RF algorithm on regression problems has been demonstrated in several application areas such as predicting atomic local environment and lattice thermal conductivity of crystalline material . See ref for more information on random forest.…”

“…The performance of the RF algorithm on regression problems has been demonstrated in several application areas such as predicting atomic local environment 40 and lattice thermal conductivity of crystalline material. 41 See ref 38 for more information on random forest.…”

Recent advances in designing thermoelectric materials

“…CBH or bonding chemistry can be potentially manipulated by many elemental descriptors such as electronegativity, coordination number, and boiling/melting points as learned from physics, chemistry, and more recently machine learning models. 399,402,403,[435][436][437][438] These descriptors also significantly affect the electronic dispersion, especially the effective masses, which largely determine the electronic transport properties. Generally, a higher electronegativity difference between the atoms results in more ionic bonds, whereas a smaller difference leads to covalent bonding.…”

Recent progress, mechanisms, and perspectives for crystal and interface chemistry applying to the Zn metal anodes in aqueous zinc‐ion batteries

“…Therefore, the balance should be found for the rational proportion of salt to solvent, contributing to the high ionic conductivity in the aqueous electrolyte and SEI 89,118,119 The advantages of theoretical simulation should be made full use of, especially artificial intelligence, data mining, and machine learning, which have gained great attention recently 120 . − 122 These tools are increasingly becoming a treasure for researchers in a variety of science and technology facilitating interdisciplinary collaboration, and advanced theoretical simulation is helping to improve the efficiency of joint work.…”