Simulation of Phase-Change-Memory and Thermoelectric Materials using Machine-Learned Interatomic Potentials: Sb2Te3

Konstantinos Konstantinou; Juraj Mavračić; Felix C. Mocanu; Stephen R. Elliott

doi:10.1002/pssb.202000416

Simulation of Phase-Change-Memory and Thermoelectric Materials using Machine-Learned Interatomic Potentials: Sb₂Te₃

Konstantinos Konstantinou
, Juraj Mavračić
, Felix C. Mocanu
, Stephen R. Elliott^*

^*Corresponding author for this work

Physics

Research output: Contribution to journal › Article › Scientific › peer-review

22 Citations (Scopus)

Abstract

Density-functional-theory (DFT)-based, ab initio molecular dynamics (AIMD) simulations of amorphous materials generally suffer from three computer-resource-related limitations due to their O(N³) cubic scaling with model system size, N. They are limited to a maximum model size of N ≈500 atoms; they are limited to time scales <1 ns; and, usually, only a single model can be simulated in any one investigation. This article discusses a machine-learned, linear-scaling (O(N)), DFT-accurate interatomic potential (a Gaussian approximation potential, GAP), originally developed by Mocanu et al. [J. Phys. Chem. B 2018, 122, 8998] using a Gaussian process regression method for the ternary phase-change-memory material Ge₂Sb₂Te₅ (GST). The chemical transferability of this GAP potential is explored in an application to the case of simulating amorphous models of the phase-change-memory and thermoelectric material Sb₂Te₃, an end-member of the GST compositional tie-line GeTe–Sb₂Te₃. The GAP-model results are compared with those obtained from conventional DFT-based AIMD simulations.

Original language	English
Article number	2000416
Number of pages	8
Journal	Physica Status Solidi (B): Basic Research
Volume	258
Issue number	9
Early online date	2020
DOIs	https://doi.org/10.1002/pssb.202000416
Publication status	Published - 2021
Publication type	A1 Journal article-refereed

Funding

This paper is dedicated to Professor David Drabold on the occasion of his 60 birthday. David is a great Anglophile, and S.R.E. recalls with much pleasure the many visits that he has made to Cambridge over the years to discuss matters amorphous. Via our membership of the UK's HEC Materials Chemistry Consortium, which was funded by EPSRC (EP/L000202, EP/R029431), this work used the ARCHER UK National Supercomputing Service ( http://www.archer.ac.uk ). th

Keywords

ab initio molecular dynamics
amorphous SbTe
density functional theory
Gaussian approximation potential
machine-learned potentials
phase-change-memory materials
thermoelectric materials

Publication forum classification

Publication forum level 1

ASJC Scopus subject areas

Electronic, Optical and Magnetic Materials
Condensed Matter Physics

Access to Document

10.1002/pssb.202000416

Cite this

@article{ea17381f1ca54498b9850611f06cc4e4,

title = "Simulation of Phase-Change-Memory and Thermoelectric Materials using Machine-Learned Interatomic Potentials: Sb2Te3",

abstract = "Density-functional-theory (DFT)-based, ab initio molecular dynamics (AIMD) simulations of amorphous materials generally suffer from three computer-resource-related limitations due to their O(N3) cubic scaling with model system size, N. They are limited to a maximum model size of N ≈500 atoms; they are limited to time scales <1 ns; and, usually, only a single model can be simulated in any one investigation. This article discusses a machine-learned, linear-scaling (O(N)), DFT-accurate interatomic potential (a Gaussian approximation potential, GAP), originally developed by Mocanu et al. [J. Phys. Chem. B 2018, 122, 8998] using a Gaussian process regression method for the ternary phase-change-memory material Ge2Sb2Te5 (GST). The chemical transferability of this GAP potential is explored in an application to the case of simulating amorphous models of the phase-change-memory and thermoelectric material Sb2Te3, an end-member of the GST compositional tie-line GeTe–Sb2Te3. The GAP-model results are compared with those obtained from conventional DFT-based AIMD simulations.",

keywords = "ab initio molecular dynamics, amorphous SbTe, density functional theory, Gaussian approximation potential, machine-learned potentials, phase-change-memory materials, thermoelectric materials",

author = "Konstantinos Konstantinou and Juraj Mavra{\v c}i{\'c} and Mocanu, \{Felix C.\} and Elliott, \{Stephen R.\}",

note = "Funding Information: This paper is dedicated to Professor David Drabold on the occasion of his 60 birthday. David is a great Anglophile, and S.R.E. recalls with much pleasure the many visits that he has made to Cambridge over the years to discuss matters amorphous. Via our membership of the UK's HEC Materials Chemistry Consortium, which was funded by EPSRC (EP/L000202, EP/R029431), this work used the ARCHER UK National Supercomputing Service ( http://www.archer.ac.uk ). th Publisher Copyright: {\textcopyright} 2020 Wiley-VCH GmbH Copyright: Copyright 2020 Elsevier B.V., All rights reserved.",

year = "2021",

doi = "10.1002/pssb.202000416",

language = "English",

volume = "258",

journal = "Physica Status Solidi (B): Basic Research",

issn = "0370-1972",

publisher = "John Wiley and Sons Inc",

number = "9",

}

TY - JOUR

T1 - Simulation of Phase-Change-Memory and Thermoelectric Materials using Machine-Learned Interatomic Potentials

T2 - Sb2Te3

AU - Konstantinou, Konstantinos

AU - Mavračić, Juraj

AU - Mocanu, Felix C.

AU - Elliott, Stephen R.

N1 - Funding Information: This paper is dedicated to Professor David Drabold on the occasion of his 60 birthday. David is a great Anglophile, and S.R.E. recalls with much pleasure the many visits that he has made to Cambridge over the years to discuss matters amorphous. Via our membership of the UK's HEC Materials Chemistry Consortium, which was funded by EPSRC (EP/L000202, EP/R029431), this work used the ARCHER UK National Supercomputing Service ( http://www.archer.ac.uk ). th Publisher Copyright: © 2020 Wiley-VCH GmbH Copyright: Copyright 2020 Elsevier B.V., All rights reserved.

PY - 2021

Y1 - 2021

N2 - Density-functional-theory (DFT)-based, ab initio molecular dynamics (AIMD) simulations of amorphous materials generally suffer from three computer-resource-related limitations due to their O(N3) cubic scaling with model system size, N. They are limited to a maximum model size of N ≈500 atoms; they are limited to time scales <1 ns; and, usually, only a single model can be simulated in any one investigation. This article discusses a machine-learned, linear-scaling (O(N)), DFT-accurate interatomic potential (a Gaussian approximation potential, GAP), originally developed by Mocanu et al. [J. Phys. Chem. B 2018, 122, 8998] using a Gaussian process regression method for the ternary phase-change-memory material Ge2Sb2Te5 (GST). The chemical transferability of this GAP potential is explored in an application to the case of simulating amorphous models of the phase-change-memory and thermoelectric material Sb2Te3, an end-member of the GST compositional tie-line GeTe–Sb2Te3. The GAP-model results are compared with those obtained from conventional DFT-based AIMD simulations.

AB - Density-functional-theory (DFT)-based, ab initio molecular dynamics (AIMD) simulations of amorphous materials generally suffer from three computer-resource-related limitations due to their O(N3) cubic scaling with model system size, N. They are limited to a maximum model size of N ≈500 atoms; they are limited to time scales <1 ns; and, usually, only a single model can be simulated in any one investigation. This article discusses a machine-learned, linear-scaling (O(N)), DFT-accurate interatomic potential (a Gaussian approximation potential, GAP), originally developed by Mocanu et al. [J. Phys. Chem. B 2018, 122, 8998] using a Gaussian process regression method for the ternary phase-change-memory material Ge2Sb2Te5 (GST). The chemical transferability of this GAP potential is explored in an application to the case of simulating amorphous models of the phase-change-memory and thermoelectric material Sb2Te3, an end-member of the GST compositional tie-line GeTe–Sb2Te3. The GAP-model results are compared with those obtained from conventional DFT-based AIMD simulations.

KW - ab initio molecular dynamics

KW - amorphous SbTe

KW - density functional theory

KW - Gaussian approximation potential

KW - machine-learned potentials

KW - phase-change-memory materials

KW - thermoelectric materials

U2 - 10.1002/pssb.202000416

DO - 10.1002/pssb.202000416

M3 - Article

AN - SCOPUS:85096827858

SN - 0370-1972

VL - 258

JO - Physica Status Solidi (B): Basic Research

JF - Physica Status Solidi (B): Basic Research

IS - 9

M1 - 2000416

ER -