High-Temperature Heat Capacity of Vibrational Modes

Ethan Meitz, Carnegie Mellon University

Photo of Ethan Meitz

The atomic-scale mechanisms responsible for the thermal properties of solids at high temperatures are poorly understood. We propose and verify an approach using a Taylor effective potential (TEP) and molecular dynamics (MD) simulations to calculate the heat capacity of phonon modes at high temperatures. This research will contribute to the development of ultrahigh thermal conductivity materials, efficient thermal storage technologies, and practical thermoelectrics. Existing methods for phonon-resolved thermal conductivity calculations rely on harmonic heat capacities (i.e., from Bose-Einstein statistics). Here, we develop a parameter-free first-principles framework to calculate individual phonon heat capacities at high temperature, where the harmonic approximation breaks. Our model unlocks the ability to study high temperature heat capacity by naturally including anharmonicity and decomposes the bulk heat capacity into contributions from each phonon mode. We study Lennard-Jones argon and Stillinger-Weber silicon and compare heat capacities from the TEP model to those obtained directly from MD simulations. When using zero-temperature phonon modes, the TEP results agree with those from MD at low temperatures, but the error grows as temperature, and thus anharmonicity, increases. This effect is accounted for by using temperature-dependent force constants resulting in less than 2% error between the TEP and MD heat capacities.