Search

Machine learning has demonstrated great power in materials design, discovery, and property prediction. However, despite the success of machine learning in predicting discrete properties, challenges remain for continuous property prediction. The challenge is aggravated in crystalline solids due to crystallographic symmetry considerations and data scarcity. Here, the direct prediction of phonon density‐of‐states (DOS) is demonstrated using only atomic species and positions as input. Euclidean neural networks are applied, which by construction are equivariant to 3D rotations, translations, and inversion and thereby capture full crystal symmetry, and achieve high‐quality prediction using a small training set of ≈103 examples with over 64 atom types. The predictive model reproduces key features of experimental data and even generalizes to materials with unseen elements, and is naturally suited to efficiently predict alloy systems without additional computational cost. The potential of the network is demonstrated by predicting a broad number of high phononic specific heat capacity materials. The work indicates an efficient approach to explore materials' phonon structure, and can further enable rapid screening for high‐performance thermal storage materials and phonon‐mediated superconductors., A machine‐learning‐based predictive model allows a direct output of phonon density‐of‐states (DOS) by inputting atomic coordinates. The implemented Euclidean neural network preserves crystallographic symmetry and enables high‐quality prediction with small training data. The model allows for fast DOS calculation, reaching ab initio accuracy for 70–80% materials with much lowered computational cost, facilitating materials design especially in alloy systems.

Charge and thermal transport in a crystal is carried by free electrons and phonons (quantized lattice vibration), the two most fundamental quasiparticles. Above the Debye temperature of the crystal, phonon‐mediated thermal conductivity (κ L) is typically limited by mutual scattering of phonons, which results in κ L decreasing with inverse temperature, whereas free electrons play a negligible role in κ L. Here, an unusual case in charge‐density‐wave tantalum disulfide (1T‐TaS2) is reported, in which κ L is limited instead by phonon scattering with free electrons, resulting in a temperature‐independent κ L. In this system, the conventional phonon–phonon scattering is alleviated by its uniquely structured phonon dispersions, while unusually strong electron‐phonon (e‐ph) coupling arises from its Fermi surface strongly nested at wavevectors in which phonons exhibit Kohn anomalies. The unusual temperature dependence of thermal conduction is found as a consequence of these effects. The finding reveals new physics of thermal conduction, offers a unique platform to probe e‐ph interactions, and provides potential ways to control heat flow in materials with free charge carriers. The temperature‐independent thermal conductivity may also find thermal management application as a special thermal interface material between two systems when the heat conduction between them needs to be maintained at a constant level., Electron–phonon interaction that normally plays negligible role in lattice thermal conduction of crystalline materials is found to dominantly scatter phonons and limit thermal conduction in a charge‐density‐wave material, 1T‐TaS2. This work elucidates the basic transport behavior of the two most fundamental quasi‐particles in a crystal, free electrons and phonons.

Phonon density‐of‐states is a key property that governs materials thermal properties but is nontrivial to compute or measure. A neural network that carries full crystal symmetry allows a prediction of phonon density‐of‐states using a small volume of training data, approaching ab initio accuracy but with significantly increased efficiency, as demonstrated in article number 2004214, by Mingda Li, Zhantao Chen, Nina Andrejevic, Tess Smidt, and co‐workers. This work enables direct structure‐property design of materials with superior thermal properties. [Image: see text]