Abstract

Semiconductors are promising in photoelectric and thermoelectric devices, for which the thermal transport properties are of particular interest. However, they have not been fully understood, especially when crystalline imperfections are present. Here, using cadmium telluride (CdTe) as an example, we illustrate how grain boundaries (GBs) affect the thermal transport properties of semiconductors. We develop a machine-learning force field from density functional theory calculations for predicting the lattice thermal conductivity (LTC) via equilibrium molecular dynamics simulations. The LTC of crystalline CdTe decreases with the relationship of κL∼1/T in the simulation temperature range of 300 – 900 K, in which the isotropic LTC decreases from 3.34 to 0.23 W/(m⋅K) due to the enhanced anharmonicity. More important, after introducing GBs, the LTC is suppressed in all directions, especially in the direction normal to the GB planes. More severe LTC suppression occurs in CdTe with Σ9 GB than that with Σ3 GB at 300 K, decreasing by 92.8% and 61.4% along the direction normal to the GB planes compared to the isotropic LTC of the crystalline CdTe, respectively. The decreased LTC is consistent with the weaker bonding near GB planes and lower shear modulus of the defective material. The analyses of the phonon dispersion curves, vibrational density of states, and phonon participation ratio indicate that the decreased LTC mainly arises from phonon scattering at GBs. Overall, our work highlights that GBs can greatly influence the LTC of semiconductors, thus providing a promising approach for thermal property design.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call