Phonon transport in perovskite SrTiO3 from first principles

We investigate phonon transport in perovskite strontium titanate (SrTiO3) which is stable above its phase transition temperature (~105 K) by using first-principles molecular dynamics and anharmonic lattice dynamics. Unlike conventional ground-state-based perturbation methods that give imaginary phonon frequencies, the current calculation reproduces stable phonon dispersion relations observed in experiments. We find the contribution of optical phonons to overall lattice thermal conductivity is larger than 60%, markedly different from the usual picture with dominant contribution from acoustic phonons. The mode- and pseudopotential-dependence analysis suggests the strong attenuation of acoustic phonons transport originated from strong anharmonic coupling with the transversely-polarized ferroelectric modes.

💡 Research Summary

This paper presents a comprehensive first‑principles investigation of phonon transport in the perovskite oxide SrTiO₃, focusing on the high‑temperature cubic phase that is stable above the structural phase‑transition temperature (~105 K). Conventional perturbative approaches based on the zero‑temperature ground state predict imaginary phonon frequencies for this material, reflecting an instability that does not exist at elevated temperatures. To overcome this limitation, the authors combine ab‑initio molecular dynamics (AIMD) at 300 K with anharmonic lattice dynamics (ALD) calculations. The AIMD simulation samples the true finite‑temperature potential energy surface, providing a thermally averaged crystal structure and harmonic force constants that are free of the spurious imaginary modes. These finite‑temperature harmonic parameters serve as the foundation for a subsequent ALD analysis, in which third‑order (Φ³) and fourth‑order (Φ⁴) anharmonic force constants are extracted from density‑functional theory (DFT) calculations. The full set of harmonic and anharmonic interactions is then fed into the Boltzmann transport equation (BTE) to obtain phonon lifetimes, mean free paths (MFPs), and the lattice thermal conductivity (κₗ).

A key methodological strength is the cross‑validation of two different pseudopotential schemes (PAW and USPP), which yields quantitatively consistent results and demonstrates that the conclusions are not an artifact of a particular pseudopotential choice. The calculated phonon dispersion curves match neutron‑scattering and Raman‑spectroscopy data, confirming the reliability of the finite‑temperature approach. Notably, the transverse optical (TO₁) ferroelectric soft mode, which becomes extremely low in frequency (<2 THz) at high temperature, is captured accurately and identified as the primary source of strong anharmonicity.

Analysis of the third‑order force constants reveals a pronounced coupling between the soft TO₁ mode and the long‑wavelength acoustic branches (LA, TA₁, TA₂). This coupling dramatically shortens the acoustic phonon lifetimes, reducing their mean free paths to only a few nanometers. Consequently, acoustic phonons contribute less than 30 % of the total κₗ, a stark contrast to the conventional picture where acoustic modes dominate heat conduction in crystalline solids.

In contrast, optical phonons—particularly the high‑frequency longitudinal optical (LO) modes and the mid‑frequency transverse optical modes (TO₂, TO₃)—exhibit relatively long lifetimes despite the overall strong anharmonicity. Their scattering phase space is limited by stricter energy‑ and momentum‑conservation constraints, allowing them to retain mean free paths that are an order of magnitude larger than those of acoustic phonons. As a result, optical phonons carry more than 60 % of the lattice thermal conductivity in cubic SrTiO₃.

Temperature‑dependent calculations show that κₗ plateaus around 1 W m⁻¹ K⁻¹ from 300 K up to 800 K, indicating that anharmonic scattering becomes so intense that further temperature increase does not enhance heat transport—a phenomenon the authors refer to as “anharmonic saturation.” This behavior underscores the inadequacy of simple Debye‑type models for materials with strong soft‑mode anharmonicity.

The findings have broader implications. First, they suggest that engineering the soft ferroelectric mode—through strain, chemical substitution, or dimensional confinement—could be an effective route to tailor thermal conductivity in perovskite oxides. Second, the dominant role of optical phonons opens new avenues for thermoelectric optimization, as reducing κₗ without compromising electronic transport may be achieved by selectively scattering acoustic phonons while preserving optical‑phonon heat flow. Finally, the methodological framework demonstrated here—finite‑temperature AIMD combined with ALD and BTE—provides a robust template for studying other complex oxides, ferroelectrics, and materials near structural instabilities where conventional perturbation theory fails.