Thermal conductivity of WC: Microstructural design driven by first-principles simulations

Abstract

The relationships between the microstructure and the thermal conductivity of binderless WC have been quantified, considering crystal orientation, isotopic abundance, porosity, and grain size. A significantly higher conductivity is predicted in the out-of-plane (c-axis) direction vs. the in-plane (a-axis) direction, using first principles simulations. Isotopic enrichment of the tungsten sublattice is predicted to increase conductivity, e.g., by a factor of 4–5 in the absence of boundary scattering. The results suggest that for an isotopically pure single crystal a thermal conductivity exceeding 1000 W m−1 K−1 may be achievable normal to the basal plane. The conductivity of samples with various porosities could be well fit by a minimum surface area (exponential) model, with a porosity exponent of b = 4.4. Experiment and simulation show a strong grain size dependence to conductivity below 1 µm, with a saturation beyond ∼10 µm. The experimental plateau values for κ were ∼45 % lower than those of the simulations due to deviations from perfect stoichiometry. We also find a higher scattering coefficient in the experiments, likely due to effects of grain size distribution and elongation. Our study clarifies the physical origin of disagreeing literature reports as being predominantly due to grain boundary scattering and enables microstructural design for thermally demanding environments.