基于机器学习原子间势函数的银硫属化合物热输运性质研究

银硫属化合物Ag₂X(X=Te，Se) 是一类新型的热电材料，在室温下是一种窄带隙半导体，在400 K 附近发生一级相变，转变为超离子导体。目前，对于银硫属化合物的研究主要集中于低温的α 相或者高温的β 相，对于相变点附近银硫属化合物热输运性质的研究非常少。因此，对于银硫属化合物在相变区域的热输运性质的研究具有重要意义。在进行分子动力学计算时，相比于传统的经验势函数，机器学习原子间势函数具有更高的计算精度。矩张量势函数是一种新兴的机器学习原子间势函数，其特点在于多项式形式的基础函数和线性回归的拟合方式。

本研究对矩张量势函数的结构特性进行了分析，确定了适用于研究银硫属化合物的矩张量势函数的形式。根据自适应双重取样方法构建了合理的样本空间，并在样本空间上对矩张量势函数进行了训练和精度测试：根据训练后的矩张量势函数对银硫属化合物在相变前后的原子轨迹、径向分布函数以及原子的均方位移等性质进行了计算，计算结果与现有的研究工作相一致，表明构造的矩张量势函数是合理的。在此基础上，根据平衡态分子动力学方法，研究了银硫属化合物在相变前后和相变区内的晶格热导率的变化规律。研究结果表明，根据矩张量势函数计算的银硫属化合物的晶格热导率在相变区域存在突变。对银硫属化合物的晶格热导率各向异性的分析结果显示，在相变前，银硫属化合物的晶格热导率存在明显的各向异性。相变结束后，随着温度的升高，这种各向异性逐渐减弱。

[1] 曾辉, 曾琪. 塞贝克效应及其应用[J]. 化学物理学报, 2004, 22(3): 52-54.
[2] 柳承恩. 珀耳帖效应对无序绝缘衬底上多晶硅激光再结晶的影响[J]. Journal of Semiconductors,1985, 6(5): 481-486.
[3] 赵素琴. 伯塞活脱气体的焦耳——汤姆逊效应[J]. 青海师专学报: 教育科学, 2005, 25(4):58-61.
[4] VINING C B. An inconvenient truth about thermoelectrics[J]. Nature Materials, 2009, 8: 83-85.
[5] 厉英, 王淑兰, 张大勇, 等. 热电材料的研究现状及发展[J]. 材料导报, 2005, (4): 23-25.
[6] WU H J, ZHAO L D, ZHENG F S, et al. Broad temperature plateau for thermoelectric figure of merit ZT>2 in phase-separated PbTe0.7S0.3[J]. Nature Communications, 2014, 5.
[7] 徐亚东, 徐桂英, 葛昌纯. SiGe 系热电材料的研究动态[C]//材料导报编辑部. 2007 高技术新材料产业发展研讨会暨《材料导报》编委会年会论文集. 四川成都: 《材料导报》杂志社, 2007: 122-124+127.
[8] LI H, SONG J, XIAO J, et al. Conductive Polymers: Synergistically Improved Molecular Doping and Carrier Mobility by Copolymerization of Donor–Acceptor and Donor–Donor Building Blocks for Thermoelectric Application[J]. Advanced Functional Materials, 2020, 30(40): 2070270.
[9] 王大刚, 王雷, 王文馨, 等. 聚噻吩及其衍生物热电材料研究进展[J]. 材料导报, 2012, 26:74-78.
[10] 梁安生. 基于噻唑单元D-A 型共轭聚合物的设计合成及其热电性能研究[D]. 深圳: 深圳大学, 2017.
[11] LIU J, VAN DER ZEE B, ALESSANDRI R, et al. N-type organic thermoelectrics: demonstration of ZT>0.3[J]. Nature Communications, 2020, 11: 5694.
[12] SAKURADA S, SHUTOH N. Effect of Ti substitution on the thermoelectric properties of (Zr,Hf)NiSn half-Heusler compounds[J]. Applied Physics Letters, 2005, 86(8): 082105.
[13] 张强, 李家荡, 胡良禄, 等. Mg3Sb2 基热电材料的研究进展[J]. 太原理工大学学报, 2021,52: 683-692.
[14] TAMAKI H, SATO H, KANNO T. Isotropic Conduction Network and Defect Chemistry in Mg3+δSb2-Based Layered Zintl Compounds with High Thermoelectric Performance[J]. Advanced materials, 2016, 28.
[15] SHUAI J, MAO J, SONG S, et al. Tuning the carrier scattering mechanism to effectively improve the thermoelectric properties[J]. Energy & Environmental Science, 2017, 10: 799-807.
[16] CHEN X, WU H, CUI J, et al. Extraordinary thermoelectric performance in n-type manganese doped Mg3Sb2 Zintl: High band degeneracy, tuned carrier scattering mechanism and hierarchical microstructure[J]. Nano Energy, 2018, 52: 246-255.
[17] SONG S, MAO J, BORDELON M, et al. Joint effect of magnesium and yttrium on enhancing thermoelectric properties of n-type Zintl Mg3+δY0.02Sb1.5Bi0.5[J]. Materials Today Physics, 2019, 8: 25-33.
[18] 黄高飞. Bi2Se3 纳米热电材料的合成与物性研究[D]. 新疆: 新疆大学, 2016.
[19] 詹斌, 兰金叻, 刘耀春, 等. 氧化物热电材料研究进展[J]. 无机材料学报, 2014, 29: 237-244.
[20] ZHU J, PANDEY R. Silver tellurides: Structural, elastic, and optical properties of AgTe and Ag2Te[J]. Journal of Physics and Chemistry of Solids, 2019, 129: 41-45.
[21] KASHIDA S, WATANABE N, HASEGAWA T, et al. Electronic structure of Ag2Te, band calculation and photoelectron spectroscopy[J]. Solid State Ionics, 2002, 148: 193-201.
[22] FUJIKANE M, KUROSAKI K, MUTA H, et al. Electrical properties of α- and β-Ag2Te[J]. Journal of Alloys and Compounds, 2005, 387: 297-299.
[23] KARAKAYA I, THOMPSON W. The Ag-Te (Silver-Tellurium) System[J]. Journal of Phase Equilibria, 1991, 12: 56-63.
[24] FUJIKANE M, KUROSAKI K, MUTA H, et al. Thermoelectric properties of α- and β-Ag2Te[J]. Journal of Alloys and Compounds, 2005, 393(1): 299-301.
[25] VORONIN M V, OSADCHII E G, BRICHKINA E A. Thermochemical properties of silver tellurides including empressite (AgTe) and phase diagrams for Ag–Te and Ag–Te–O[J]. Physics and Chemistry of Minerals, 2017, 44: 639-653.
[26] KOBAYASHI M, ISHIKAWA K, TACHIBANA F, et al. Diffusion path and Haven’s ratio of mobile ions in α-Ag2Te[J]. Physical Review B, 1988, 38: 3050-3055.
[27] TACHIBANA F, KOBAYASHI M, OKAZAKI H. Residence time and density distribution of silver ions in α-Ag2Te by molecular-dynamics simulation[J]. Physical Review B, 1989, 40: 3360-3363.
[28] PEI Y, HEINZ N A, SNYDER G J. Alloying to increase the band gap for improving thermoelectric properties of Ag2Te[J]. Journal of Materials Chemistry, 2011, 21: 18256-18260.
[29] JOOD P, CHETTY R, OHTA M. Structural stability enables high thermoelectric performance in room temperature Ag2Se[J]. Journal of Materials Chemistry A, 2020, 8: 13024-13037.
[30] KARAKAYA I, THOMPSON W T. The Ag-Se (Silver-Selenium) system[J]. Bulletin of Alloy Phase Diagrams, 1990, 11: 266.
[31] DRYMIOTIS F, DAY T W, BROWN D R, et al. Enhanced thermoelectric performance in the very low thermal conductivity Ag2Se0.5Te0.5[J]. Applied Physics Letters, 2013, 103(14): 143906.
[32] WIEGERS G A. The Crystal Structure of the Low-Temperature form of Silver Selenide[J]. American Mineralogist, 1971, 56(11-12): 1882-1888.
[33] SHIMOJO F, OKAZAKI H. Phase Transition in Superionic Conductor Ag2Se:A Molecular Dynamics Study[J]. Journal of the Physical Society of Japan, 1991, 60(11): 3745-3753.
[34] RINO J P, HORNOS Y M M, ANTONIO G A, et al. Structural and dynamical correlations in Ag2Se: A molecular dynamics study of superionic and molten phases[J]. The Journal of Chemical Physics, 1988, 89(12): 7542-7555.
[35] SAKUMA T, IIDA K, HONMA K, et al. X-Ray Diffraction Study on a Superionic Conductor: α-Ag2Se[J]. Journal of the Physical Society of Japan, 1977, 43(2): 538-543.
[36] HIRATA K, MATSUNAGA T, SINGH S, et al. High-Performance Solid-State Thermal Diode Consisting of Ag2(S, Se, Te)[J]. Journal of Electronic Materials, 2020, 49: 2895-2901.
[37] BEHLER J. Atom-centered symmetry functions for constructing high-dimensional neural network potentials[J]. The Journal of Chemical Physics, 2011, 134(7): 074106.
[38] SHIMAMURA K, TAKESHITA Y, FUKUSHIMA S, et al. Computational and training requirements for interatomic potential based on artificial neural network for estimating low thermal conductivity of silver chalcogenides[J]. The Journal of Chemical Physics, 2020, 153(23): 234301.
[39] BARTÓK A P, PAYNE M C, KONDOR R, et al. Gaussian Approximation Potentials: The Accuracy of Quantum Mechanics, without the Electrons[J]. Physical Review Letters, 2010, 104: 136403.
[40] BARTOK A, CSáNYI G. Gaussian Approximation Potentials: a brief tutorial introduction[J]. International Journal of Quantum Chemistry, 2015, 115.
[41] DERINGER V L, CARO M A, CSáNYI G. Machine Learning Interatomic Potentials as Emerging Tools for Materials Science[J]. Advanced Materials, 2019, 31(46): 1902765.
[42] CSÁNYI G, WINFIELD S, KERMODE J R, et al. Expressive Programming for Computational Physics in Fortran 95+[J]. Computer Physics Communications, 2007.
[43] SHAPEEV A V. Moment Tensor Potentials: A Class of Systematically Improvable Interatomic Potentials[J]. Multiscale Modeling & Simulation, 2016, 14(3): 1153-1173.
[44] ROSENBROCK C, GUBAEV K, SHAPEEV A, et al. Machine-learned Interatomic Potentials for Alloys and Alloy Phase Diagrams[J]. npj Computational Materials, 2019, 7: 24.
[45] NOVIKOV I S, GUBAEV K, PODRYABINKIN E V, et al. The MLIP package: moment tensor potentials with MPI and active learning[J]. Machine Learning: Science and Technology, 2021, 2(2): 025002.
[46] PODRYABINKIN E V, SHAPEEV A V. Active learning of linearly parametrized interatomic potentials[J]. Computational Materials Science, 2017, 140: 171-180.
[47] GOREINOV S A, OSELEDETS I V, SAVOSTYANOV D V, et al. How to Find a Good Submatrix[M/OL]. 2010: 247-256. DOI: 10.1142/9789812836021_0015.
[48] KROGH A, VEDELSBY J. Neural Network Ensembles, Cross Validation and Active Learning[C]//NIPS’94: Proceedings of the 7th International Conference on Neural Information Processing Systems. Cambridge, MA, USA: MIT Press, 1994: 231–238.
[49] HOHENBERG P, KOHN W. Inhomogeneous Electron Gas[J]. Physical Review, 1964, 136: B864-B871.
[50] JANAK J F. Proof that ∂𝐸∂𝑛𝑖= 𝜖 in density-functional theory[J]. Physical Review B, 1978, 18: 7165-7168.
[51] KOCH W, HOLTHAUSEN M C. A Chemist’s Guide to Density Functional Theory[M]. 2nd ed. New Jersey: John Wiley and Sons, 2001: 65-91.
[52] KURTH S, PERDEW J P, BLAHA P. Molecular and solid-state tests of density functional approximations: LSD, GGAs, and meta-GGAs[J]. International Journal of Quantum Chemistry, 1999, 75(4-5): 889-909.
[53] KRYACHKO E S, LUDEñA E V. Density functional theory: Foundations reviewed[J]. Physics Reports, 2014, 544(2): 123-239.
[54] 冯真真. 几种Zintl 相化合物及MgAgSb 热电特性的理论研究[D]. 河南: 河南大学, 2016.
[55] YANG H, ZHU Y, DONG E, et al. Dual adaptive sampling and machine learning interatomic potentials for modeling materials with chemical bond hierarchy[J]. Physical Review B, 2021, 104: 094310.
[56] GUBAEV K, PODRYABINKIN E V, HART G L, et al. Accelerating high-throughput searches for new alloys with active learning of interatomic potentials[J]. Computational Materials Science, 2019, 156: 148-156.
[57] KOBAYASHI M, ISHIKAWA K, TACHIBANA F, et al. Diffusion path and Haven’s ratio of mobile ions in α-Ag2Te[J]. Physical Review B, 1988, 38: 3050-3055.
[58] SHIMOJO F, ANIYA M. Diffusion Mechanism of Ag ions in Superionic Conductor Ag2Se from Ab Initio Molecular-Dynamics Simulations[J]. Journal of the Physical Society of Japan, 2005, 74(4): 1224-1230.
[59] RAPAPORT D C. The Art of Molecular Dynamics Simulation[M]. 2nd ed. Cambridge University Press, 2004: 25-43.
[60] HAN Z, YANG X, LI W, et al. FourPhonon: An extension module to ShengBTE for computing four-phonon scattering rates and thermal conductivity[J]. Computer Physics Communications, 2022, 270: 108179.
[61] WANG Z, RUAN X. On the domain size effect of thermal conductivities from equilibrium and nonequilibrium molecular dynamics simulations[J]. Journal of Applied Physics, 2017, 121: 044301.
[62] KONG L T, BARTELS G, CAMPAñá C, et al. Implementation of Green’s function molecular dynamics: An extension to LAMMPS[J]. Computer Physics Communications, 2009, 180(6): 1004-1010.
[63] KONG L. Phonon dispersion measured directly from molecular dynamics simulations[J]. Computer Physics Communications, 2011, 182: 2201-2207.
[64] IKESHOJI T, HAFSKJOLD B. Non-equilibrium molecular dynamics calculation of heat conduction in liquid and through liquid-gas interface[J]. Molecular Physics, 1994, 81(2): 251-261.
[65] JUND P, JULLIEN R. Molecular-Dynamics Calculation of the Thermal Conductivity of Vitreous Silica[J]. Physical Review B, 1999, 59.
[66] MüLLER-PLATHE F. A simple nonequilibrium molecular dynamics method for calculating the thermal conductivity[J]. The Journal of Chemical Physics, 1997, 106(14): 6082-6085.
[67] KUBO R. Statistical-Mechanical Theory of Irreversible Processes. I. General Theory and Simple Applications to Magnetic and Conduction Problems[J]. Journal of the Physical Society of Japan, 1957, 12(6): 570-586.
[68] GREEN M S. Markoff Random Processes and the Statistical Mechanics of Time‐Dependent Phenomena. II. Irreversible Processes in Fluids[J]. The Journal of Chemical Physics, 1954, 22(3): 398-413.
[69] LI D, ZHANG B, MING H, et al. Liquid-Phase Manipulation Securing Enhanced ThermoelectricPerformance of Ag2Se[J]. ACS applied materials & interfaces, 2021, 13.