Higher-order thermal transport theory for phonon thermal transport in semiconductors using lattice dynamics calculations and the Boltzmann transport equation
Ankit Jain, Yagyank Srivastava, Amey G. Gokhale, Nidheesh Virakante, Hardik L. Kagdada · 2025
Read Higher-order thermal transport theory for phonon thermal transport in semiconductors using lattice dynamics calculations and the Boltzmann transport equation by Ankit Jain, Yagyank Srivastava, Amey G. Gokhale, Nidheesh Virakante, Hardik L. Kagdada (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-23307.txt.
Introduction
Ankit Jain and four co-authors wrote this 2025 review of newer theory for how heat moves through semiconductor crystals. Heat in such solids rides on waves of atom motion called phonons. The standard method pairs atom motion math with a flow equation and has matched tests well with no fitted values.
Over the past decade, many materials broke that simple picture. The basic method samples the energy surface only near rest positions and counts only three wave clashes. The review gathers the fixes built since then: richer clash counts and fuller treatments of the flow equation. It maps which method fits which class of material.
Catalog record
| CREATOR | Ankit Jain, Yagyank Srivastava, Amey G. Gokhale, Nidheesh Virakante, Hardik L. Kagdada |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 65 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-23307
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Higher-order thermal transport theory for phonon thermal transport in semiconductors using lattice dynamics calculations and the Boltzmann transport equation Ankit Jain, Yagyank Srivastava, Amey G. Gokhale, Nidheesh Virakante, Hardik L. Kagdada Abstract The phonon thermal conductivity of semiconducting periodic solids can be obtained using the lattice dynamics calculations along with the Boltzmann transport equation and with input from density functional theory calculations. These calculations have resulted in an excellent agreement with experiments without requiring any fitting parameters. However, over the last decade, many material systems have been identified where the lowest level lattice dynamics theory, which is based on the relaxation time approximation solution of the Boltzmann transport equation and considers potential energy surface sampling around the static equilibrium positions of atoms with only three-phonon scatterings, is proved insufficient in describing the thermal transport physics. In this article, we review these higher-order developments in the lattice dynamics theory to describe thermal transport in periodic semiconducting solids. We start with a brief discussion of the lowest-order theory and discuss its limitations along with proposed developments to address these limitations. We discuss prominent success cases of these higher-order developments and present our recommendations on their use for various material systems. Considering that many of these higher-order developments are computationally more demanding compared to the lowest-order theory, we also discussed data-driven approaches to accelerate these calculations. This review article is intended to serve as a reference for both novice and experienced researchers in this field. I Introduction The understanding of phonon thermal transport physics in semiconductors is crucial for applications such as thermoelectric energy conversion, heat dissipation, and memory storage devices [ 1 , 2 , 3 ] . The thermal transport in non-metallic crystalline materials (periodic semiconducting and insulating solids) is predominantly through atomic vibrations, i.e., phonons [ 4 , 5 ] . Traditionally, the thermal design of semiconducting solids is led by resource-consuming experiments based on a trial-and-error approach guided by simple physics principles such as the presence of heavy atomic species and complex material unitcell [ 6 ] . However, with recent advances in computational resources, it is now possible to predict the phonon thermal conductivity of materials without requiring any experimental fitting parameters [ 7 , 8 , 3 , 9 , 10 ] ; thus paving the way for computational discovery of materials with targeted thermal properties. Lattice dynamics calculations and molecular dynamics simulations are two commonly employed tools for investigating thermal transport in solids [ 11 , 12 ] . In lattice dynamics calculations, the atoms are treated as a coupled spring-mass system, and the phonon modes are obtained as eigenmodes of the coupled system with linear/harmonic springs. [ 13 ] The non-linearity/anharmonicity is then introduced as a perturbation to harmonic modes [ 5 ] . This approach is ideal for perfectly periodic systems, and any deviation from this is handled via the perturbative treatment. The lattice dynamics approach provides a bottom-up mode-level understanding of the thermal transport physics, and the thermal conductivity is obtained by summing over the contributions of all modes. In contrast, in molecular dynamics simulations, the atoms are allowed to move in real-time, and the time-evolution of atomic trajectories is obtained by using Newton’s laws. These atomic trajectories are then used to obtain the thermal conductivity. Unlike lattice dynamics, the molecular dynamics simulations can account for interaction anharmonicity to the highest order, and the inclusion of lattice imperfections is trivial in these simulations. While these simulations treat atoms as classical particles (occupation governed by the Boltzmann distribution), methodologies to account for quantum corrections are also explored in the literature [ 14 ] . The thermal conductivity obtained from lattice dynamics calculations and molecular dynamics simulations depends on interatomic interactions [ 7 , 8 , 9 ] . Consequently, the applicability of these approaches was originally limited to qualitative/trend predictions as the interactions were obtained from empirical forcefields [ 15 , 16 , 17 ] , fitted to reproduce one or more experimentally observed properties of the given material [ 18 , 19 , 20 ] . With the availability of ab-initio-based density functional theory (DFT) driven interatomic interactions, it is now possible to obtain the actual thermal conductivity of materials from these approaches, directly comparable with experiments [ 21 , 7 , 8 , 3 , 9 ] . The lattice dynamics calculations have an advantage in this regard as they require interatomic force constant evaluation only once, while molecular dynamics simulations need interatomic interactions at each of the millions of simulation timesteps. Despite this advantage, the computational cost of interatomic force constants evaluation from DFT is high at several thousand to tens of thousands cpu-hours [ 7 ] and approximate data-driven methodologies, such as based on machine learning, are being developed to reduce the computational cost [ 22 , 23 , 24 , 25 , 26 ] . On top of interatomic interactions, the accuracy of results obtained from the lattice dynamics calculations also depend on several other factors such as (not limited to) (i) methodology of interatomic force constant extraction (finite-difference, compressive-sensing, etc) [ 27 , 28 , 29 , 30 ] , (ii) phonon-phonon interactions (three-phonon, four-phonon scatterings, etc) [ 31 , 32 , 33 ] , (iii) phonon eigenmode renormalization (to account for anharmonicity) [ 34 ] , and (iv) employed thermal transport theory [Boltzmann transport equation (BTE) based particle-like thermal transport or coherent-based wave-like transport] [ 35 , 36 , 37 ] . This tutorial-cum-review is focused on lattice dynamics based approach for phonon thermal conductivity prediction of periodic solids. We will begin with a brief discussion of basic lattice dynamics theory and its success stories. Next, we will discuss limitations of this basic theory and the developments presented in the literature to address these limitations. We will next present computational challenges and associated data-driven approaches to address computational challenges. Lastly, we will discuss our recommendations on the use of lattice dynamics approach for different material systems. We note that, while for completeness, we present a brief discussion of basic lattice dynamics theory and its success cases, the readers are referred to our previous tutorial for an in-depth discussion of basic lattice dynamics theory [ 9 ] . II Phonon Thermal Conductivity The phonon thermal transport theory is discussed in detail in Refs. [ 4 , 5 , 38 , 39 ] . The phonons at wavevector \bm{q} are defined as eigenmodes (3 N modes, enumerated by mode-index \nu , where N is the number of atoms in the unitcell) of the Dynamical matrix {\bm{D}}_{{\bm{q}}} whose elements, D_{\bm{q}}^{3(b-1)+\alpha,3(b^{{}^{\prime}}-1)+\beta} , are given by: \begin{split}D_{\bm{q}}^{3(b-1)+\alpha,3(b^{{}^{\prime}}-1)+\beta}=\\ \frac{1}{\sqrt{m_{b}m_{b^{{}^{\prime}}}}}\sum_{l^{{}^{\prime}}}\Phi_{b0;b^{{}^{\prime}}l^{{}^{\prime}}}^{\alpha\beta}\exp{\{i[{\bm{q}}.({\bm{r}}_{b^{{}^{\prime}}l^{{}^{\prime}}}-{\bm{r}}_{b0})]\}},\end{split} (1) where the summation is over all unit-cells in the lattice, m_{b} is the mass of atom b in the unit-cell, \bm{r}_{bl} is the position vector of atom b in the l^{th} unit-cell, \alpha , \beta \in(x,y,z) denotes cartesian directions, and \Phi_{ij}^{\alpha\beta} is the real-space ( ij,\alpha\beta )-element of the harmonic force constant matrix \bm{\Phi} . The phonon frequencies, \omega_{{\bm{q}}\nu} , are obtained from the diagonalization of the dynamical matrix as: \omega_{{\bm{q}}\nu}^{2}{\bm{e}}_{{\bm{q}}\nu}={\bm{D}}_{{\bm{q}}}\cdot{\bm{e}}_{{\bm{q}}\nu}, (2) where {\bm{e}}_{{\bm{q}}\nu} is the phonon eigenvector. In the lattice dynamics approach, the phonon thermal conductivity is obtained from the steady-state analysis of the BTE: {\bm{v}_{{\bm{q}}\nu}}\cdot\nabla{n_{{\bm{q}}\nu}}=\left(\frac{\partial n}{\partial t}\right)_{coll}, (3) where the left-hand