Taking materials dynamics to new extremes using machine learning interatomic potentials

Abstract

Understanding materials dynamics under extreme conditions of pressure, temperature, and strain rate is a scientific quest that spans nearly a century. Atomic simulations have had a considerable impact on this endeavor because of their ability to uncover materials’ microstructure evolution and properties at the scale of the relevant physical phenomena. However, this is still a challenge for most materials as it requires modeling large atomic systems (up to millions of particles) with improved accuracy. In many cases, the availability of sufficiently accurate but efficient interatomic potentials has become a serious bottleneck for performing these simulations as traditional potentials fail to represent the multitude of bonding. A new class of potentials has emerged recently, based on a different paradigm from the traditional approach. The new potentials are constructed by machine-learning with a high degree of fidelity from quantum-mechanical calculations. In this review, a brief introduction to the central ideas underlying machine learning interatomic potentials is given. In particular, the coupling of machine learning models with domain knowledge to improve accuracy, computational efficiency, and interpretability is highlighted. Subsequently, we demonstrate the effectiveness of the domain knowledge-based approach in certain select problems related to the kinetic response of warm dense materials. It is hoped that this review will inspire further advances in the understanding of matter under extreme conditions.

Keywords

Machine learning / interatomic potential / extreme condition / domain knowledge / materials science

References

[1] Asay JR.High-pressure shock compression of solids.2012;New YorkSpringer Science & Business Media

[2] Drickamer HG.Electronic transitions and the high pressure chemistry and physics of solids.2013;New YorkSpringer Science & Business Media

[3] Graham RA.Solids under high-pressure shock compression: mechanics, physics, and chemistry.2012;New YorkSpringer Science & Business Media

[4] Woollam JA.High-pressure and low-temperature physics.2012;New YorkSpringer Science & Business Media

[5] Konôpková Z,Gómez-Pérez N.Direct measurement of thermal conductivity in solid iron at planetary core conditions..Nature2016;534:99-101

[6] Monnereau M,Margerin L.Lopsided growth of Earth’s inner core..Science2010;328:1014-7

[7] Raleigh CB.Mechanisms of plastic deformation of olivine..J Geophys Res1968;73:5391-406

[8] Wenk HR,Hemley RJ,Shu J.The plastic deformation of iron at pressures of the Earth’s inner core..Nature2000;405:1044-7

[9] Valiev R.Nanostructuring of metals by severe plastic deformation for advanced properties..Nat Mater2004;3:511-6

[10] Yu Q,Li J.Strong crystal size effect on deformation twinning..Nature2010;463:335-8

[11] Zhou X,Zhu L.High-pressure strengthening in ultrafine-grained metals..Nature2020;579:67-72

[12] Aoyama T,Iyama A,Shimizu K.Giant spin-driven ferroelectric polarization in TbMnO3 under high pressure..Nat Commun2014;5:4927

[13] Cao Y,Demir A.Correlated insulator behaviour at half-filling in magic-angle graphene superlattices..Nature2018;556:80-4

[14] Cao Y,Fang S.Unconventional superconductivity in magic-angle graphene superlattices..Nature2018;556:43-50

[15] Zhao J,Li W.A combinatory ferroelectric compound bridging simple ABO₃ and A-site-ordered quadruple perovskite..Nat Commun2021;12:747 PMCID:PMC7854592

[16] Tan C,Li Y.A high performance wearable strain sensor with advanced thermal management for motion monitoring..Nat Commun2020;11:3530 PMCID:PMC7363829

[17] Tsiklis DS.Handbook of techniques in high-pressure research and engineering.2012;New YorkSpringer Science & Business Media

[18] Andreoni W.Handbook of materials modeling: methods: theory and modeling.2019;Springer International Publishing

[19] Andreoni W.Handbook of materials modeling: applications: current and emerging materials.2020;Springer

[20] Collins LA,Ding YH.A review on ab initio studies of static, transport, and optical properties of polystyrene under extreme conditions for inertial confinement fusion applications..Physics of Plasmas2018;25:056306

[21] Lees AW.The computer study of transport processes under extreme conditions..J Phys C: Solid State Phys1972;5:1921-8

[22] Nisoli C,Niezgoda SR,Lookman T.Long-time behavior of the ω→’α transition in shocked zirconium: Interplay of nucleation and plastic deformation..Acta Materialia2016;108:138-42

[23] Zong H,Lookman T.Collective nature of plasticity in mediating phase transformation under shock compression..Phys Rev B2014;89

[24] Zong H,Lookman T.Twin boundary activated α→’ω phase transformation in titanium under shock compression..Acta Materialia2016;115:1-9

[25] Zong H,Ding X.Nucleation mechanism for hcp→’bcc phase transformation in shock-compressed Zr..Phys Rev B2020;101:144105

[26] Zong H,Ding X.The kinetics of the ω to α phase transformation in Zr, Ti: analysis of data from shock-recovered samples and atomistic simulations..Acta Materialia2014;77:191-9

[27] Zong H,Ding X,Ackland GJ.hcp→’ω phase transition mechanisms in shocked zirconium: A machine learning based atomic simulation study..Acta Materialia2019;162:126-35

[28] Hohenberg P.Inhomogeneous Electron Gas..Phys Rev1964;136:B864-71

[29] Kohn W.Self-consistent equations including exchange and correlation effects..Phys Rev1965;140:A1133-8

[30] Botu V,Chapman J.Machine learning force fields: construction, validation, and outlook..J Phys Chem C2017;121:511-22

[31] Daw MS.Embedded-atom method: derivation and application to impurities, surfaces, and other defects in metals..Phys Rev B1984;29:6443-53

[32] Finnis MW.A simple empirical N-body potential for transition metals..Philosophical Magazine A2006;50:45-55

[33] Stillinger FH.Computer simulation of local order in condensed phases of silicon..Phys Rev B Condens Matter1985;31:5262-71

[34] Tersoff J.New empirical approach for the structure and energy of covalent systems..Phys Rev B Condens Matter1988;37:6991-7000

[35] Tersoff J.Empirical interatomic potential for silicon with improved elastic properties..Phys Rev B Condens Matter1988;38:9902-5

[36] Blank TB,Calhoun AW.Neural network models of potential energy surfaces..J Chem Phys1995;103:4129-37

[37] de Pablo JJ,Webb MA.New frontiers for the materials genome initiative..npj Comput Mater2019;5:1-23

[38] Deringer VL,Csányi G.Machine learning interatomic potentials as emerging tools for materials science..Adv Mater2019;31:e1902765

[39] Ong SP.Accelerating materials science with high-throughput computations and machine learning..Comput Mater Sci2019;161:143-50

[40] Rosenbrock CW,Shapeev AV.Machine-learned interatomic potentials for alloys and alloy phase diagrams..npj Comput Mater2021;7:1-9

[41] Friederich P,Proppe J.Machine-learned potentials for next-generation matter simulations..Nat Mater2021;20:750-61

[42] Mishin Y.Machine-learning interatomic potentials for materials science..Acta Materialia2021;214:116980

[43] Chen C,Tran R,Chu I.Accurate force field for molybdenum by machine learning large materials data..Phys Rev Materials2017;1:043603

[44] Byggmästar J,Nordlund K.Machine-learning interatomic potential for radiation damage and defects in tungsten..Phys Rev B2019;100

[45] Zeni C,Glielmo A.On machine learning force fields for metallic nanoparticles..Adv Phys X2019;4:1654919

[46] Marchand D,Glensk A.Machine learning for metallurgy I. A neural-network potential for Al-Cu..Phys Rev Materials2020;4:103601

[47] Jinnouchi R,Kresse G.On-the-fly machine learning force field generation: application to melting points..Phys Rev B2019;100:014105

[48] Thomas JC,Natarajan AR.Machine learning the density functional theory potential energy surface for the inorganic halide perovskite CsPbBr₃..Phys Rev B2019;100:134101

[49] Podryabinkin EV.Active learning of linearly parametrized interatomic potentials..Comput Mater Sci2017;140:171-80

[50] Gubaev K,Hart GL.Accelerating high-throughput searches for new alloys with active learning of interatomic potentials..Comput Mater Sci2019;156:148-56

[51] Zhang L,Wang H,Weinan E.Active learning of uniformly accurate interatomic potentials for materials simulation..Phys Rev Materials2019;3:023804

[52] Bisbo MK.Efficient global structure optimization with a machine-learned surrogate model..Phys Rev Lett2020;124:086102

[53] Onat B,Kermode JR.Sensitivity and dimensionality of atomic environment representations used for machine learning interatomic potentials..J Chem Phys2020;153:144106

[54] Botu V.Adaptive machine learning framework to accelerate ab initio molecular dynamics..Int J Quantum Chem2015;115:1074-83

[55] Botu V.Learning scheme to predict atomic forces and accelerate materials simulations..Phys Rev B2015;92:094306

[56] Pozdnyakov SN,Bartók AP,Csányi G.Incompleteness of atomic structure representations..Phys Rev Lett2020;125:166001

[57] Bartók AP,Csányi G.On representing chemical environments..Phys Rev B2013;87

[58] Batra R,Kim C.General atomic neighborhood fingerprint for machine learning-based methods..J Phys Chem C2019;123:15859-66

[59] Drautz R.Atomic cluster expansion for accurate and transferable interatomic potentials..Phys Rev B2019;99:014104

[60] Willatt MJ,Ceriotti M.Atom-density representations for machine learning..J Chem Phys2019;150:154110

[61] Himanen L,Morooka EV.DScribe: library of descriptors for machine learning in materials science..Comput Phys Commun2020;247:106949

[62] Kocer E,Erturk H.Continuous and optimally complete description of chemical environments using Spherical Bessel descriptors..AIP Advances2020;10:015021

[63] Zuo Y,Li X.Performance and cost assessment of machine learning interatomic potentials..J Phys Chem A2020;124:731-45

[64] Novotni M.Shape retrieval using 3D Zernike descriptors..Computer-Aided Design2004;36:1047-62

[65] Shapeev AV.Moment tensor potentials: a class of systematically improvable interatomic potentials..Multiscale Model Simul2016;14:1153-73

[66] Behler J.Generalized neural-network representation of high-dimensional potential-energy surfaces..Phys Rev Lett2007;98:146401

[67] Zong H,Ding X,Lookman T.Developing an interatomic potential for martensitic phase transformations in zirconium by machine learning..npj Comput Mater2018;4:1-8

[68] Pun GPP,Ramprasad R.Physically informed artificial neural networks for atomistic modeling of materials..Nat Commun2019;10:2339

[69] Helfrecht BA,Pireddu G,Ceriotti M.A new kind of atlas of zeolite building blocks..J Chem Phys2019;151:154112

[70] Thompson A,Trott C,Tucker G.Spectral neighbor analysis method for automated generation of quantum-accurate interatomic potentials..J Comput Phys2015;285:316-30

[71] Zhang L,Wang H,E W.Deep potential molecular dynamics: a scalable model with the accuracy of quantum mechanics..Phys Rev Lett2018;120:143001

[72] Tang L,Wen T,Kramer M.Short- and medium-range orders in Al90Tb10 glass and their relation to the structures of competing crystalline phases..Acta Materialia2021;204:116513

[73] Wu J,Zhang L.Deep learning of accurate force field of ferroelectric HfO₂..Phys Rev B2021;103:024108

[74] Yang M,Parrinello M.Liquid-liquid critical point in phosphorus..Phys Rev Lett2021;127:080603

[75] Zhang L,Car R.Phase diagram of a deep potential water model..Phys Rev Lett2021;126:236001

[76] Zhang L,Wang H.End-to-end symmetry preserving inter-atomic potential energy model for finite and extended systems..arXiv preprint arXiv2018;1805:09003

[77] Naden Robinson V,Ackland GJ,Hermann A.On the chain-melted phase of matter..Proc Natl Acad Sci U S A2019;116:10297-302 PMCID:PMC6535020

[78] Zong H,Hermann A.Free electron to electride transition in dense liquid potassium..Nat Phys2021;17:955-60

[79] Zong H,Ackland GJ.Understanding high pressure molecular hydrogen with a hierarchical machine-learned potential..Nat Commun2020;11:5014 PMCID:PMC7538439

[80] Novikov IS,Podryabinkin EV.The MLIP package: moment tensor potentials with MPI and active learning..Mach Learn Sci Technol2021;2:025002

[81] Skinner AJ.Neural networks in computational materials science: training algorithms..Modelling Simul Mater Sci Eng1995;3:371-90

[82] Bholoa A,Smith R.A new approach to potential fitting using neural networks..Nucl Instrum Methods Phys Res B2007;255:1-7

[83] Behler J,Donadio D.Metadynamics simulations of the high-pressure phases of silicon employing a high-dimensional neural network potential..Phys Rev Lett2008;100:185501

[84] Malshe M,Raff LM,Bukkapatnam S.Parametrization of analytic interatomic potential functions using neural networks..J Chem Phys2008;129:044111

[85] Sanville E,Smith R.Silicon potentials investigated using density functional theory fitted neural networks..J Phys Condens Matter2008;20:285219

[86] Eshet H,Kühne TD,Parrinello M.Ab initio quality neural-network potential for sodium..Phys Rev B2010;81:184107

[87] Handley CM.Potential energy surfaces fitted by artificial neural networks..J Phys Chem A2010;114:3371-83

[88] Behler J.Neural network potential-energy surfaces in chemistry: a tool for large-scale simulations..Phys Chem Chem Phys2011;13:17930-55

[89] Behler J.Atom-centered symmetry functions for constructing high-dimensional neural network potentials..J Chem Phys2011;134:074106

[90] Artrith N,Behler J.Erratum: high-dimensional neural-network potentials for multicomponent systems: applications to zinc oxide..Phys Rev B2012;86:153101

[91] Raff L,Hagan M.Neural networks in chemical reaction dynamics.2012;OUP USA

[92] Sosso GC,Caravati S,Bernasconi M.Neural network interatomic potential for the phase change material GeTe..Phys Rev B2012;85

[93] Behler J.Representing potential energy surfaces by high-dimensional neural network potentials..J Phys Condens Matter2014;26:183001

[94] Behler J.Constructing high-dimensional neural network potentials: a tutorial review..Int J Quantum Chem2015;115:1032-50

[95] Artrith N.An implementation of artificial neural-network potentials for atomistic materials simulations: performance for TiO₂..Comput Mater Sci2016;114:135-50

[96] Hajinazar S,Kolmogorov AN.Stratified construction of neural network based interatomic models for multicomponent materials..Phys Rev B2017;95:014114

[97] Kobayashi R,Junge T,Curtin WA.Neural network potential for Al-Mg-Si alloys..Phys Rev Materials2017;1:053604

[98] Bochkarev AS,Mossa S.Anharmonic thermodynamics of vacancies using a neural network potential..Phys Rev Materials2019;3

[99] Patra TK,Chan H,Narayanan B.A coarse-grained deep neural network model for liquid water..Appl Phys Lett2019;115:193101

[100] Unke OT.PhysNet: A neural network for predicting energies, forces, dipole moments, and partial charges..J Chem Theory Comput2019;15:3678-93

[101] Pun GPP,Hickman J,Mishin Y.Development of a general-purpose machine-learning interatomic potential for aluminum by the physically informed neural network method..Phys Rev Materials2020;4:113807

[102] Stricker M,Mak E.Machine learning for metallurgy II. A neural-network potential for magnesium..Phys Rev Materials2020;4

[103] Hajinazar S,Sandoval ED,Kolmogorov AN.MAISE: construction of neural network interatomic models and evolutionary structure optimization..Comput Phys Commun2021;259:107679

[104] Ko TW,Goedecker S.A fourth-generation high-dimensional neural network potential with accurate electrostatics including non-local charge transfer..Nat Commun2021;12:398

[105] Lin YS,Mishin Y.Development of a physically-informed neural network interatomic potential for tantalum..arXiv preprint arXiv2021;2101:06540

[106] Schütt KT,Kindermans PJ,Müller KR.SchNet - a deep learning architecture for molecules and materials..J Chem Phys2018;148:241722

[107] Andolina CM,Saidi WA.Optimization and validation of a deep learning CuZr atomistic potential: Robust applications for crystalline and amorphous phases with near-DFT accuracy..J Chem Phys2020;152:154701

[108] Hofmann T,Smola AJ.Kernel methods in machine learning..Ann Statist2008;36:1171-220

[109] Müller KR,Rätsch G,Schölkopf B.An introduction to kernel-based learning algorithms..IEEE Trans Neural Netw2001;12:181-201

[110] Glielmo A,De Vita A.Accurate interatomic force fields via machine learning with covariant kernels..Phys Rev B2017;95

[111] Bartók AP,Kondor R.Gaussian approximation potentials: the accuracy of quantum mechanics, without the electrons..Phys Rev Lett2010;104:136403

[112] Babaei H,Hashemi A.Machine-learning-based interatomic potential for phonon transport in perfect crystalline Si and crystalline Si with vacancies..Phys Rev Materials2019;3:074603

[113] Wang Q,Zhang L,Shapeev A.Predicting the propensity for thermally activated β events in metallic glasses via interpretable machine learning..npj Comput Mater2020;6:1-12

[114] Seko A,Tsuda K.Machine learning with systematic density-functional theory calculations: application to melting temperatures of single- and binary-component solids..Phys Rev B2014;89:054303

[115] Zhao L,Ding X,Ackland GJ.Commensurate-incommensurate phase transition of dense potassium simulated by machine-learned interatomic potential..Phys Rev B2019;100:220101

[116] Szlachta WJ,Csányi G.Accuracy and transferability of Gaussian approximation potential models for tungsten..Phys Rev B2014;90:104108

[117] Artrith N.High-dimensional neural network potentials for metal surfaces: A prototype study for copper..Phys Rev B2012;85:045439

[118] Kruglov I,Yanilkin A.Energy-free machine learning force field for aluminum..Sci Rep2017;7:8512

[119] Dragoni D,Csányi G.Achieving DFT accuracy with a machine-learning interatomic potential: thermomechanics and defects in bcc ferromagnetic iron..Phys Rev Materials2018;2:013808

[120] Maillet JB,Csányi G.Machine-learning based potential for iron: plasticity and phase transition. AIP Conference Proceedings..AIP Publishing LLC2018;1979:050011

[121] Ibarra-Hernández W,Avendaño-Franco G,Kolmogorov AN.Structural search for stable Mg-Ca alloys accelerated with a neural network interatomic model..Phys Chem Chem Phys2018;20:27545-57

[122] Hajinazar S,Cullo AJ.Multitribe evolutionary search for stable Cu-Pd-Ag nanoparticles using neural network models..Phys Chem Chem Phys2019;21:8729-42

[123] Sivaraman G,Baur M.Machine-learned interatomic potentials by active learning: amorphous and liquid hafnium dioxide..npj Comput Mater2020;6:1-8

[124] Artrith N,Behler J.Neural network potentials for metals and oxides - First applications to copper clusters at zinc oxide..Phys Status Solidi B2013;250:1191-203

[125] Sivaraman G,Krishnamoorthy AN.Experimentally driven automated machine-learned interatomic potential for a refractory oxide..Phys Rev Lett2021;126:156002

[126] Jain ACP,Glensk A,Curtin WA.Machine learning for metallurgy III: a neural network potential for Al-Mg-Si..Phys Rev Materials2021;5

[127] Zhao L,Ding X.Anomalous dislocation core structure in shock compressed bcc high-entropy alloys..Acta Materialia2021;209:116801

[128] Li QJ,Lam S.Development of robust neural-network interatomic potential for molten salt..Cell Rep Phys Sci2021;2:100359

[129] Lam ST,Ballinger R,Li J.Modeling LiF and FLiBe molten salts with robust neural network interatomic potential..ACS Appl Mater Interfaces2021;13:24582-92

[130] van Duin ACT,Lorant F.ReaxFF: a reactive force field for hydrocarbons..J Phys Chem A2001;105:9396-409

[131] Cheng B,Pickard CJ.Evidence for supercritical behaviour of high-pressure liquid hydrogen..Nature2020;585:217-20

[132] Musil F,Yang J,Day GM.Machine learning for the structure-energy-property landscapes of molecular crystals..Chem Sci2018;9:1289-300

[133] Muhli H,Bartók AP.Machine learning force fields based on local parametrization of dispersion interactions: Application to the phase diagram of C₆₀..arXiv preprint arXiv2021;2105:02525

[134] Caro MA,Laurila T.Machine learning driven simulated deposition of carbon films: from low-density to diamondlike amorphous carbon..Phys Rev B2020;102:174201

[135] Rowe P,Gasparotto P,Michaelides A.An accurate and transferable machine learning potential for carbon..J Chem Phys2020;153:034702

[136] Caro MA,Koskinen J,Csányi G.Growth mechanism and origin of high sp³ content in tetrahedral amorphous carbon..Phys Rev Lett2018;120:166101

[137] Bartók AP,Bernstein N.Machine learning a general-purpose interatomic potential for silicon..Phys Rev X2018;8

[138] Bonati L.Silicon Liquid structure and crystal nucleation from ab initio deep metadynamics..Phys Rev Lett2018;121:265701

[139] Deringer VL,Csányi G.Origins of structural and electronic transitions in disordered silicon..Nature2021;589:59-64

[140] Niu H,Invernizzi M.Molecular dynamics simulations of liquid silica crystallization..Proc Natl Acad Sci U S A2018;115:5348-52

[141] Unruh D,Goodnick SM.Training a machine-learning driven gaussian approximation potential for Si-H interactions..arXiv preprint arXiv2021;2106:02946

[142] Deringer VL,Csányi G.A general-purpose machine-learning force field for bulk and nanostructured phosphorus..Nat Commun2020;11:5461

[143] Niu H,Piaggi PM.Ab initio phase diagram and nucleation of gallium..Nat Commun2020;11:2654

[144] Mocanu FC,Lee TH.Modeling the phase-change memory material, Ge₂Sb₂Te₅, with a machine-learned interatomic potential..J Phys Chem B2018;122:8998-9006

[145] Yang Y,Sun J.Rippling ferroic phase transition and domain switching in 2D materials..Adv Mater2021;e2103469

[146] Hajibabaei A.Universal machine learning interatomic potentials: surveying solid electrolytes..J Phys Chem Lett2021;12:8115-20

[147] Hajibabaei A,Kim KS.Sparse Gaussian process potentials: application to lithium diffusivity in superionic conducting solid electrolytes..Phys Rev B2021;103:214102

[148] Kostiuchenko T,Neugebauer J.Impact of lattice relaxations on phase transitions in a high-entropy alloy studied by machine-learning potentials..npj Comput Mater2019;5:1-7

[149] Lahnsteiner J,Bokdam M.Long-range order imposed by short-range interactions in methylammonium lead iodide: comparing point-dipole models to machine-learning force fields..Phys Rev B2019;100:094106

[150] Lapointe C,Thiry L.Machine learning surrogate models for prediction of point defect vibrational entropy..Phys Rev Materials2020;4:063802

[151] Scarselli F,Tsoi AC,Monfardini G.The graph neural network model..IEEE Trans Neural Netw2009;20:61-80