Rapid Prediction of Anisotropic Lattice Thermal Conductivity: Application to Layered Materials

Thermal conductivity plays a crucial role in many applications; the use of single-crystal and textured polycrystalline materials in such applications necessitates understanding the anisotropy in thermal transport. The measurement of anisotropic lattice thermal conductivity κL(θ,ϕ) is quite challenging. To address this need through computations, we build upon our previously developed isotropic model for κL and incorporate the directional (angular) dependence by using the elastic tensor obtained from ab initio calculations and the Christoffel equations for the speed of sound. With the anisotropic speed of sound and intrinsic material properties as input parameters, we can predict the direction-dependent κL(θ,ϕ). We validate this new model by comparing with the experimental data from the literature—the predicted κL is within an average factor difference of 1.8 of experimental measurements, spanning 5 orders of magnitude in κL. To demonstrate the utility and computational tractability of this model, we calculate κL(θ,ϕ) of ∼2200 layered materials that are expected to exhibit anisotropic thermal transport. We consider both van der Waals and ionic layered structures with binary and ternary chemistries and analyze the anisotropy in their κL. The large-scale study has revealed many layered structures with interesting anisotropy in κL.