声子散射计算：新采样技术降低其复杂性

热导率（κ）与辐射特性（ε）是评估材料在热电和光学应用性能的核心指标，它们与声子间相互作用的散射过程紧密相关。近期研究揭示，除了常规的三声子散射，四声子散射在高温或特殊材料中也极其重要。

Fig. 1 Maximum likelihood estimation of τ-1λ3ph and τ-1λ4ph.

然而，精确计算这些散射过程需要巨大的计算资源，挑战了材料科学的研究边界。因此，开发新的计算方法以降低计算成本，对于高性能材料的发展至关重要。

Fig. 2 Calculating confidence interval of τ̂-1λ3ph and τ̂-1λ4ph.

美国普渡大学机械工程系的Guang Lin（林光）教授和Xiulin Ruan（阮修林）教授领导的团队，提出了一种基于采样技术与最大似然估计（MLE）的方法，显著降低了声子散射计算的复杂性，并加速了热导率（κ）与辐射特性（ε）的预测速度。

Fig. 3 Average computational cost of estimating scattering rate with respect to sample sizes for Si. a τ-1λ3ph, b τ-1λ4ph.

在弛豫时间近似（RTA）条件下，该采样技术旨在通过所有声子散射过程中的部分样本来估计每种声子模式λ的三声子（3ph）和四声子（4ph）散射率。

Fig. 4 Estimation of κ with the sampling method.

这个概念很简单，但实施起来效果却出奇的好。在获得散射率数据后，研究团队利用所有声子模式的散射率和仅使用红外活跃声子模式的散射率，分别确定了热导率κ和辐射特性ε。

Fig. 5 Estimation of the anisotropic κ of LiCoO2 with the sampling method.

作者的方法不仅计算成本低，而且能在大幅度缩减计算时间的同时，保持高精度的预测结果。这使得作者能够以前所未有密集的32×32×32的精细倒空间网格划分来精确计算硅的四声子散射，并获得了一个收敛的热导率数值，与实验数据的吻合度大幅提升。

Fig. 6 Time-saving of predicting lattice thermal conductivity.

该方法的准确性和高效性使其成为热学和光学应用材料高通量筛选的理想选择。该文近期发表于npj Computational Materials 10: 31 (2024)。

Fig. 7 Estimation of dielectric function for MgO at 300 K.

Editorial Summary

Sampling technique: Effectively reduces the complexity of phonon scattering calculations

Thermal conductivity (κ) and radiative properties (ε) are essential indicators for assessing material performance in thermoelectric and optical applications, closely related to the scattering processes of phonon interactions. Recent studies have revealed that, in addition to conventional three-phonon scattering, four-phonon scattering plays a vital role in materials at high temperatures or with special properties. However, accurately calculating these scattering processes requires substantial computational resources, challenging the frontiers of material science research. Therefore, developing new computational methods to reduce the cost of calculations is crucial for the advancement of high-performance materials. 

A team lead by Prof. Guang Lin and Prof. Xiulin Ruan from Department of Mechanical Engineering, Purdue University, USA, proposed an approach based on sampling and maximum likelihood estimation (MLE) to reduce the computational cost of phonon scattering calculations and accelerate the predictions of κ and ε. Under the relaxation time approximation (RTA), the sampling method aims to estimate the scattering rates of 3ph and 4ph scattering for each phonon mode λ (denoted as  and , respectively) by sampling a subset from all phonon scattering processes. The concept is simple but works surprisingly well. After that, κ and ε are determined by using the scattering rate of all phonon modes and the IR-active phonon mode only, respectively. The authors demonstrate that the sampling method can significantly reduce the computational cost of the predictions of thermal conductivity and radiative properties. This allows the authors to revisit the thermal conductivity prediction of Si with an unprecedented q-mesh of 32×32 ×32, resulting in a converged thermal conductivity value that closely aligns with experimental data. The accuracy and efficiency of the approach make it ideal for high-throughput screenings of materials for thermal and optical applications. This article was recently published in npj Computational Materials 10: 31 (2024).

原文Abstract及其翻译

Sampling-accelerated prediction of phonon scattering rates for converged thermal conductivity and radiative properties （采样加速预测声子散射率以获得收敛的热导率和辐射特性）

Ziqi Guo, Zherui Han, Dudong Feng, Guang Lin & Xiulin Ruan

Abstract The prediction of thermal conductivity and radiative properties is crucial. However, computing phonon scattering, especially for four-phonon scattering, could be prohibitively expensive, and the thermal conductivity for silicon after considering four-phonon scattering is significantly under-predicted and not converged in the literature. Here we propose a method to estimate scattering rates from a small sample of scattering processes using maximum likelihood estimation. The calculation of scattering rates and associated thermal conductivity and radiative properties are dramatically accelerated by three to four orders of magnitude. This allows us to use an unprecedented q-mesh (discretized grid in the reciprocal space) of 32 × 32 × 32 for calculating four-phonon scattering of silicon and achieve a converged thermal conductivity value that agrees much better with experiments. The accuracy and efficiency of our approach make it ideal for the high-throughput screening of materials for thermal and optical applications.

摘要 

热导率和辐射特性的准确预测对科学研究至关重要。然而，计算声子散射，尤其是涉及复杂的四声子散射，不仅成本高昂，而且现有文献中对硅的热导率进行的预测，考虑到四声子散射后，结果往往低于实际值且数值并未达到理论上的收敛状态。在这里，我们提出了一种方法，该方法只需对少量散射过程样本进行分析，就能够利用最大似然估计技术来估算散射率。通过这种策略，我们显著加快了散射率及其相关的热导率和辐射特性的计算速度，达到了比传统方法快上三到四个数量级的效率。这使得我们能够使用前所未有密集的32×32×32的精细倒空间网格划分来精确计算硅的四声子散射，并获得了一个收敛的热导率数值，与实验数据的吻合度大幅提升。我们方法的高准确性和高效率使其成为热学和光学应用材料筛选领域的理想工具，尤其适合进行大规模的高通量材料筛查。