We examine the lattice thermal conductivities $\left({\kappa }_l\right)$ of $\mathrm{L}{\mathrm{i}}_{2}X(X=\mathrm{O},\mathrm{S},\mathrm{Se},\mathrm{Te})$ using a first-principles Peierls-Boltzmann transport methodology. We find low ${\kappa }_l$ values ranging between 12 and 30 $\mathrm{W}{\mathrm{m}}^{-1}{\mathrm{K}}^{-1}$ despite light Li atoms, a large mass difference between constituent atoms, and tightly bunched acoustic branches, all features that give high ${\kappa }_l$ in other materials including BeSe (630 $\mathrm{W}{\mathrm{m}}^{-1}{\mathrm{K}}^{-1}$), BeTe (370 $\mathrm{W}{\mathrm{m}}^{-1}{\mathrm{K}}^{-1}$), and cubic BAs (3170 $\mathrm{W}{\mathrm{m}}^{-1}{\mathrm{K}}^{-1}$). Together these results suggest a missing ingredient in the basic guidelines commonly used to understand and predict ${\kappa }_l$. Unlike typical simple systems (e.g., Si, GaAs, SiC), the dominant resistance to heat-carrying acoustic phonons in $\mathrm{L}{\mathrm{i}}_2\mathrm{Se}$ and $\mathrm{L}{\mathrm{i}}_2\mathrm{Te}$ comes from interactions of these modes with two optic phonons. These interactions require significant bandwidth and dispersion of the optic branches, both present in $\mathrm{L}{\mathrm{i}}_{2}X$ materials. These considerations are important for the discovery and design of new materials for thermal management applications and give a more comprehensive understanding of thermal transport in crystalline solids.

• Received 23 December 2015
• Revised 10 May 2016

DOI:https://doi.org/10.1103/PhysRevB.93.224301