Deep neural networks for accurate predictions of crystal stability

Ye, Weike; Chen, Chi; Wang, Zhenbin; Chu, Iek-Heng; Ong, Shyue Ping

doi:10.1038/s41467-018-06322-x

Title: Deep neural networks for accurate predictions of crystal stability

Abstract

Predicting the stability of crystals is one of the central problems in materials science. Today, density functional theory (DFT) calculations remain comparatively expensive and scale poorly with system size. Here we show that deep neural networks utilizing just two descriptors-the Pauling electronegativity and ionic radii-can predict the DFT formation energies of C₃A₂D₃O₁₂ garnets and ABO₃ perovskites with low mean absolute errors (MAEs) of 7-10 meV atom^-1 and 20-34 meV atom^-1, respectively, well within the limits of DFT accuracy. Further extension to mixed garnets and perovskites with little loss in accuracy can be achieved using a binary encoding scheme, addressing a critical gap in the extension of machine-learning models from fixed stoichiometry crystals to infinite universe of mixed-species crystals. Finally, we demonstrate the potential of these models to rapidly transverse vast chemical spaces to accurately identify stable compositions, accelerating the discovery of novel materials with potentially superior properties.

Authors:

Ye, Weike; Chen, Chi; Wang, Zhenbin; Chu, Iek-Heng;

Publication Date:: Tue Sep 18 00:00:00 EDT 2018

Research Org.:: Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States). National Energy Research Scientific Computing Center (NERSC); Univ. of California, Oakland, CA (United States)

Sponsoring Org.:: USDOE; USDOE Office of Science (SC), Basic Energy Sciences (BES) (SC-22). Materials Sciences & Engineering Division

OSTI Identifier:: 1469730

Alternate Identifier(s):: OSTI ID: 1543748; OSTI ID: 1559147

Grant/Contract Number:: AC02-05CH11231

Resource Type:: Published Article

Journal Name:: Nature Communications

Additional Journal Information:: Journal Name: Nature Communications Journal Volume: 9 Journal Issue: 1; Journal ID: ISSN 2041-1723

Publisher:: Nature Publishing Group

Country of Publication:: United Kingdom

Language:: English

Subject:: 36 MATERIALS SCIENCE; Science & Technology; Other Topics

Citation Formats


                    Ye, Weike, Chen, Chi, Wang, Zhenbin, Chu, Iek-Heng, and Ong, Shyue Ping. Deep neural networks for accurate predictions of crystal stability.  United Kingdom: N. p., 2018. 
Web.  doi:10.1038/s41467-018-06322-x.

Copy to clipboard


                    Ye, Weike, Chen, Chi, Wang, Zhenbin, Chu, Iek-Heng, & Ong, Shyue Ping. Deep neural networks for accurate predictions of crystal stability.  United Kingdom.  https://doi.org/10.1038/s41467-018-06322-x

Copy to clipboard


                    Ye, Weike, Chen, Chi, Wang, Zhenbin, Chu, Iek-Heng, and Ong, Shyue Ping. Tue .  
"Deep neural networks for accurate predictions of crystal stability".  United Kingdom.  https://doi.org/10.1038/s41467-018-06322-x.

Copy to clipboard


                    
@article{osti_1469730,

  title        = {Deep neural networks for accurate predictions of crystal stability},

  author       = {Ye, Weike and Chen, Chi and Wang, Zhenbin and Chu, Iek-Heng and Ong, Shyue Ping},

  abstractNote = {Predicting the stability of crystals is one of the central problems in materials science. Today, density functional theory (DFT) calculations remain comparatively expensive and scale poorly with system size. Here we show that deep neural networks utilizing just two descriptors-the Pauling electronegativity and ionic radii-can predict the DFT formation energies of C3A2D3O12 garnets and ABO3 perovskites with low mean absolute errors (MAEs) of 7-10 meV atom-1 and 20-34 meV atom-1, respectively, well within the limits of DFT accuracy. Further extension to mixed garnets and perovskites with little loss in accuracy can be achieved using a binary encoding scheme, addressing a critical gap in the extension of machine-learning models from fixed stoichiometry crystals to infinite universe of mixed-species crystals. Finally, we demonstrate the potential of these models to rapidly transverse vast chemical spaces to accurately identify stable compositions, accelerating the discovery of novel materials with potentially superior properties.},

  doi          = {10.1038/s41467-018-06322-x},

  journal      = {Nature Communications},

  number       = 1,

  volume       = 9,

  place        = {United Kingdom},

  year         = {Tue Sep 18 00:00:00 EDT 2018},

  month        = {Tue Sep 18 00:00:00 EDT 2018}

}

Copy to clipboard

Journal Article:

Free Publicly Available Full Text

Publisher's Version of Record
https://doi.org/10.1038/s41467-018-06322-x

Other availability

Search WorldCat to find libraries that may hold this journal

Citation Metrics:

Cited by: 162 works

Citation information provided by
Web of Science

Save / Share:

Export Metadata

Save to My Library

Works referenced in this record:

Generalized Gradient Approximation Made Simple
journal, October 1996

Perdew, John P.; Burke, Kieron; Ernzerhof, Matthias
Physical Review Letters, Vol. 77, Issue 18, p. 3865-3868
DOI: 10.1103/PhysRevLett.77.3865

Generating derivative structures at a fixed concentration
journal, June 2012

Hart, Gus L. W.; Nelson, Lance J.; Forcade, Rodney W.
Computational Materials Science, Vol. 59
DOI: 10.1016/j.commatsci.2012.02.015

Accelerating materials property predictions using machine learning
journal, September 2013

Pilania, Ghanshyam; Wang, Chenchen; Jiang, Xun
Scientific Reports, Vol. 3, Issue 1
DOI: 10.1038/srep02810

Deep learning
journal, May 2015

LeCun, Yann; Bengio, Yoshua; Hinton, Geoffrey
Nature, Vol. 521, Issue 7553
DOI: 10.1038/nature14539

Room-temperature magnetoresistance in an oxide material with an ordered double-perovskite structure
journal, October 1998

Kobayashi, K. -I.; Kimura, T.; Sawada, H.
Nature, Vol. 395, Issue 6703
DOI: 10.1038/27167

Projector augmented-wave method
journal, December 1994

Blöchl, P. E.
Physical Review B, Vol. 50, Issue 24, p. 17953-17979
DOI: 10.1103/PhysRevB.50.17953

The Principles Determining the Structure of Complex Ionic Crystals
journal, April 1929

Pauling, Linus
Journal of the American Chemical Society, Vol. 51, Issue 4
DOI: 10.1021/ja01379a006

Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides
journal, September 1976

Shannon, R. D.
Acta Crystallographica Section A, Vol. 32, Issue 5, p. 751-767
DOI: 10.1107/S0567739476001551

Machine learning bandgaps of double perovskites
journal, January 2016

Pilania, G.; Mannodi-Kanakkithodi, A.; Uberuaga, B. P.
Scientific Reports, Vol. 6, Issue 1
DOI: 10.1038/srep19375

The emergence of perovskite solar cells
journal, July 2014

Green, Martin A.; Ho-Baillie, Anita; Snaith, Henry J.
Nature Photonics, Vol. 8, Issue 7, p. 506-514
DOI: 10.1038/nphoton.2014.134

Including crystal structure attributes in machine learning models of formation energies via Voronoi tessellations
journal, July 2017

Ward, Logan; Liu, Ruoqian; Krishna, Amar
Physical Review B, Vol. 96, Issue 2
DOI: 10.1103/PhysRevB.96.024104

Approximation capabilities of multilayer feedforward networks
journal, January 1991

Hornik, Kurt
Neural Networks, Vol. 4, Issue 2
DOI: 10.1016/0893-6080(91)90009-T

Die Gesetze der Krystallochemie
journal, May 1926

Goldschmidt, V. M.
Die Naturwissenschaften, Vol. 14, Issue 21
DOI: 10.1007/BF01507527

Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set
journal, October 1996

Kresse, G.; Furthmüller, J.
Physical Review B, Vol. 54, Issue 16, p. 11169-11186
DOI: 10.1103/PhysRevB.54.11169

Commentary: The Materials Project: A materials genome approach to accelerating materials innovation
journal, July 2013

Jain, Anubhav; Ong, Shyue Ping; Hautier, Geoffroy
APL Materials, Vol. 1, Issue 1
DOI: 10.1063/1.4812323

Inhomogeneous Electron Gas
journal, November 1964

Hohenberg, P.; Kohn, W.
Physical Review, Vol. 136, Issue 3B, p. B864-B871
DOI: 10.1103/PhysRev.136.B864

Low temperature synthesis of Li5La3Nb2O12 with cubic garnet-type structure by sol–gel process
journal, February 2013

Peng, Hongjian; Wu, Qing; Xiao, Lihong
Journal of Sol-Gel Science and Technology, Vol. 66, Issue 1
DOI: 10.1007/s10971-013-2984-y

Li−Fe−P−O ₂ Phase Diagram from First Principles Calculations
journal, February 2008

Ong, Shyue Ping; Wang, Lei; Kang, Byoungwoo
Chemistry of Materials, Vol. 20, Issue 5
DOI: 10.1021/cm702327g

The Nature of the Chemical Bond. iv. the Energy of Single Bonds and the Relative Electronegativity of Atoms
journal, September 1932

Pauling, Linus
Journal of the American Chemical Society, Vol. 54, Issue 9
DOI: 10.1021/ja01348a011

Origin of ferroelectricity in perovskite oxides
journal, July 1992

Cohen, Ronald E.
Nature, Vol. 358, Issue 6382
DOI: 10.1038/358136a0

Perovskite oxides for visible-light-absorbing ferroelectric and photovoltaic materials
journal, November 2013

Grinberg, Ilya; West, D. Vincent; Torres, Maria
Nature, Vol. 503, Issue 7477, p. 509-512
DOI: 10.1038/nature12622

Bulk superconductivity at 91 K in single-phase oxygen-deficient perovskite ${Ba}_{2}$ ${YCu}_{3}$ $O_{9 - δ}$
journal, April 1987

Cava, R. J.; Batlogg, B.; van Dover, R. B.
Physical Review Letters, Vol. 58, Issue 16
DOI: 10.1103/PhysRevLett.58.1676

Machine Learning Energies of 2 Million Elpasolite $(A B C_{2} D_{6})$ Crystals
journal, September 2016

Faber, Felix A.; Lindmaa, Alexander; von Lilienfeld, O. Anatole
Physical Review Letters, Vol. 117, Issue 13
DOI: 10.1103/PhysRevLett.117.135502

Oxidation energies of transition metal oxides within the $GGA + U$ framework
journal, May 2006

Wang, Lei; Maxisch, Thomas; Ceder, Gerbrand
Physical Review B, Vol. 73, Issue 19
DOI: 10.1103/PhysRevB.73.195107

Structure and Ionic-Transport Properties of Lithium-Containing Garnets Li ₃ Ln ₃ Te ₂ O ₁₂ (Ln = Y, Pr, Nd, Sm−Lu)
journal, September 2006

O'Callaghan, Michael P.; Lynham, Danny R.; Cussen, Edmund J.
Chemistry of Materials, Vol. 18, Issue 19
DOI: 10.1021/cm060992t

Universal fragment descriptors for predicting properties of inorganic crystals
journal, June 2017

Isayev, Olexandr; Oses, Corey; Toher, Cormac
Nature Communications, Vol. 8, Issue 1
DOI: 10.1038/ncomms15679

A general-purpose machine learning framework for predicting properties of inorganic materials
journal, August 2016

Ward, Logan; Agrawal, Ankit; Choudhary, Alok
npj Computational Materials, Vol. 2, Issue 1
DOI: 10.1038/npjcompumats.2016.28

Predicting the Thermodynamic Stability of Solids Combining Density Functional Theory and Machine Learning
journal, May 2017

Schmidt, Jonathan; Shi, Jingming; Borlido, Pedro
Chemistry of Materials, Vol. 29, Issue 12
DOI: 10.1021/acs.chemmater.7b00156

The thermodynamic scale of inorganic crystalline metastability
journal, November 2016

Sun, Wenhao; Dacek, Stephen T.; Ong, Shyue Ping
Science Advances, Vol. 2, Issue 11
DOI: 10.1126/sciadv.1600225

Predicting the thermodynamic stability of perovskite oxides using machine learning models
journal, July 2018

Li, Wei; Jacobs, Ryan; Morgan, Dane
Computational Materials Science, Vol. 150
DOI: 10.1016/j.commatsci.2018.04.033

Combinatorial screening for new materials in unconstrained composition space with machine learning
journal, March 2014

Meredig, B.; Agrawal, A.; Kirklin, S.
Physical Review B, Vol. 89, Issue 9
DOI: 10.1103/PhysRevB.89.094104

Accuracy of density functional theory in predicting formation energies of ternary oxides from binary oxides and its implication on phase stability
journal, April 2012

Hautier, Geoffroy; Ong, Shyue Ping; Jain, Anubhav
Physical Review B, Vol. 85, Issue 15
DOI: 10.1103/PhysRevB.85.155208

Prediction model of band gap for inorganic compounds by combination of density functional theory calculations and machine learning techniques
journal, March 2016

Lee, Joohwi; Seko, Atsuto; Shitara, Kazuki
Physical Review B, Vol. 93, Issue 11
DOI: 10.1103/PhysRevB.93.115104

Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysis
journal, February 2013

Ong, Shyue Ping; Richards, William Davidson; Jain, Anubhav
Computational Materials Science, Vol. 68
DOI: 10.1016/j.commatsci.2012.10.028

Similar Records in DOE PAGES and OSTI.GOV collections:

Charge-density based convolutional neural networks for stacking fault energy prediction in concentrated alloys

Journal Article Arora, Gaurav ; Kamrava, Serveh ; Tahmasebi, Pejman ; ... - Materialia

A descriptor-less machine learning (ML) model based only on charge density images extracted from density functional theory (DFT) is developed to predict stacking fault energies (SFE) in concentrated alloys. The model is based on convolutional neural networks (CNNs) as one of the promising ML techniques for dealing with complex images and data. Identification of correct descriptors is a key bottleneck to develop ML models for predicting materials properties. Often, in most ML models, textbook physical descriptors such as atomic radius, valence charge and electronegativity are used as descriptors which have limitations because these properties change in concentrated alloys when multiplemore »« less
https://doi.org/10.1016/j.mtla.2022.101620

Full Text Available
Accurate Fe–He machine learning potential for studying He effects in BCC-Fe

Journal Article Pitike, Krishna Chaitanya ; Setyawan, Wahyu - Journal of Nuclear Materials

Nanostructured Ferritic Alloys (NFAs) containing oxide nanoparticles are prospective advanced structural materials in future fusion reactors. Lacking fusion neutron sources with an adequate flux, modeling is critical to understand radiation damage in the NFAs, including helium effects. Machine learning interatomic potentials (MLPs) have been demonstrated to be capable of describing atomic interactions in chemically complex systems with comparable accuracy to density functional theory (DFT) but with much less computational costs. Hence, MLPs are chosen to study He bubble accumulation in chemically complex NFAs. As a preliminary step, we develop a Fe–He potential, based on ~10,000 atomic configurations. The developed MLPmore »« less
https://doi.org/10.1016/j.jnucmat.2022.154183

Full Text Available
Electronegativity is the average one-electron energy of the valence-shell electrons in ground-state free atoms

Journal Article Allen, L C - Journal of the American Chemical Society; (USA)

It is argued that electronegativity is the third dimension of the Periodic Table, and that x{sub spec} = (m{epsilon}{sub p} + n{epsilon}{sub s})/(m + n), for representative elements where {epsilon}{sub p}, {epsilon}{sub s} are the p, s ionization energies and m, n the number of p, s electrons. Values of spectroscopic x are obtained to high accuracy from the National Bureau of Standards atomic energy level tables and closely match the widely accepted Pauling and Allred Rochow scales. x{sub spec} rationalizes the diagonal separation between metals and non-metals in the Periodic Table, the formation of noble gas molecules, metallization ofmore »« less
https://doi.org/10.1021/ja00207a003
Exploring Ca–Ce–M–O (M = 3d Transition Metal) Oxide Perovskites for Solar Thermochemical Applications

Journal Article Gautam, Gopalakrishnan Sai ; Stechel, Ellen B. ; Carter, Emily A. - Chemistry of Materials

Solar thermochemical (STC) processes hold promise as efficient ways to generate renewable fuels, fuel precursors, or chemical feedstocks using concentrated sunlight. Specifically, one actively researched approach is the two-step STC cycle, which uses a redox-active, off-stoichiometric, transition-metal oxide material to split water and/or CO₂, generating H₂ and/or CO, respectively, or syngas (a combination of H₂ and CO). Identifying novel metal oxides that yield larger reduction extents (practically achievable off-stoichiometries) than the state-of-the-art CeO₂ is critical. Here, we explore the chemical space of Ca–Ce–M–O (M = 3d transition metal, except Cu and Zn) metal oxide perovskites, with Ca and/or Ce occupyingmore »« less
https://doi.org/10.1021/acs.chemmater.0c02912

Full Text Available
Accuracy of DFT computed oxygen-vacancy formation energies and high-throughput search of solar thermochemical water-splitting compounds

Journal Article Baldassarri, Bianca ; He, Jiangang ; Qian, Xin ; ... - Physical Review Materials

The enthalpy change involved in metal oxide reduction is a key quantity in various processes related to energy conversion and storage, and is of particular interest for computational prediction. Often this prediction involves the simulation of a high temperature reduction process with a 0K methodology like density functional theory (DFT), and it is not infrequent for the high temperature and 0K stable crystal structures to differ. This introduces a conundrum with regards to the choice of crystal structure to utilize in the computation, with approaches in the literature varying and experimental validation remaining scarce. In this work we address bothmore »« less
https://doi.org/10.1103/physrevmaterials.7.065403

Similar Records

Title: Deep neural networks for accurate predictions of crystal stability

Abstract

Citation Formats

Generalized Gradient Approximation Made Simple journal, October 1996

Generating derivative structures at a fixed concentration journal, June 2012

Accelerating materials property predictions using machine learning journal, September 2013

Deep learning journal, May 2015

Room-temperature magnetoresistance in an oxide material with an ordered double-perovskite structure journal, October 1998

Projector augmented-wave method journal, December 1994

The Principles Determining the Structure of Complex Ionic Crystals journal, April 1929

Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides journal, September 1976

Machine learning bandgaps of double perovskites journal, January 2016

The emergence of perovskite solar cells journal, July 2014

Including crystal structure attributes in machine learning models of formation energies via Voronoi tessellations journal, July 2017

Approximation capabilities of multilayer feedforward networks journal, January 1991

Die Gesetze der Krystallochemie journal, May 1926

Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set journal, October 1996

Commentary: The Materials Project: A materials genome approach to accelerating materials innovation journal, July 2013

Inhomogeneous Electron Gas journal, November 1964

Low temperature synthesis of Li5La3Nb2O12 with cubic garnet-type structure by sol–gel process journal, February 2013

Li−Fe−P−O 2 Phase Diagram from First Principles Calculations journal, February 2008

The Nature of the Chemical Bond. iv. the Energy of Single Bonds and the Relative Electronegativity of Atoms journal, September 1932

Origin of ferroelectricity in perovskite oxides journal, July 1992

Perovskite oxides for visible-light-absorbing ferroelectric and photovoltaic materials journal, November 2013

Bulk superconductivity at 91 K in single-phase oxygen-deficient perovskite Ba 2 YCu 3 O 9 − δ journal, April 1987

Machine Learning Energies of 2 Million Elpasolite ( A B C 2 D 6 ) Crystals journal, September 2016

Oxidation energies of transition metal oxides within the GGA + U framework journal, May 2006

Structure and Ionic-Transport Properties of Lithium-Containing Garnets Li 3 Ln 3 Te 2 O 12 (Ln = Y, Pr, Nd, Sm−Lu) journal, September 2006

Universal fragment descriptors for predicting properties of inorganic crystals journal, June 2017

A general-purpose machine learning framework for predicting properties of inorganic materials journal, August 2016

Predicting the Thermodynamic Stability of Solids Combining Density Functional Theory and Machine Learning journal, May 2017

The thermodynamic scale of inorganic crystalline metastability journal, November 2016

Predicting the thermodynamic stability of perovskite oxides using machine learning models journal, July 2018

Combinatorial screening for new materials in unconstrained composition space with machine learning journal, March 2014

Accuracy of density functional theory in predicting formation energies of ternary oxides from binary oxides and its implication on phase stability journal, April 2012

Prediction model of band gap for inorganic compounds by combination of density functional theory calculations and machine learning techniques journal, March 2016

Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysis journal, February 2013

Generalized Gradient Approximation Made Simple
journal, October 1996

Generating derivative structures at a fixed concentration
journal, June 2012

Accelerating materials property predictions using machine learning
journal, September 2013

Deep learning
journal, May 2015

Room-temperature magnetoresistance in an oxide material with an ordered double-perovskite structure
journal, October 1998

Projector augmented-wave method
journal, December 1994

The Principles Determining the Structure of Complex Ionic Crystals
journal, April 1929

Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides
journal, September 1976

Machine learning bandgaps of double perovskites
journal, January 2016

The emergence of perovskite solar cells
journal, July 2014

Including crystal structure attributes in machine learning models of formation energies via Voronoi tessellations
journal, July 2017

Approximation capabilities of multilayer feedforward networks
journal, January 1991

Die Gesetze der Krystallochemie
journal, May 1926

Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set
journal, October 1996

Commentary: The Materials Project: A materials genome approach to accelerating materials innovation
journal, July 2013

Inhomogeneous Electron Gas
journal, November 1964

Low temperature synthesis of Li5La3Nb2O12 with cubic garnet-type structure by sol–gel process
journal, February 2013

Li−Fe−P−O ₂ Phase Diagram from First Principles Calculations
journal, February 2008

The Nature of the Chemical Bond. iv. the Energy of Single Bonds and the Relative Electronegativity of Atoms
journal, September 1932

Origin of ferroelectricity in perovskite oxides
journal, July 1992

Perovskite oxides for visible-light-absorbing ferroelectric and photovoltaic materials
journal, November 2013

Bulk superconductivity at 91 K in single-phase oxygen-deficient perovskite ${Ba}_{2}$ ${YCu}_{3}$ $O_{9 - δ}$
journal, April 1987

Machine Learning Energies of 2 Million Elpasolite $(A B C_{2} D_{6})$ Crystals
journal, September 2016

Oxidation energies of transition metal oxides within the $GGA + U$ framework
journal, May 2006

Structure and Ionic-Transport Properties of Lithium-Containing Garnets Li ₃ Ln ₃ Te ₂ O ₁₂ (Ln = Y, Pr, Nd, Sm−Lu)
journal, September 2006

Universal fragment descriptors for predicting properties of inorganic crystals
journal, June 2017

A general-purpose machine learning framework for predicting properties of inorganic materials
journal, August 2016

Predicting the Thermodynamic Stability of Solids Combining Density Functional Theory and Machine Learning
journal, May 2017

The thermodynamic scale of inorganic crystalline metastability
journal, November 2016

Predicting the thermodynamic stability of perovskite oxides using machine learning models
journal, July 2018

Combinatorial screening for new materials in unconstrained composition space with machine learning
journal, March 2014

Accuracy of density functional theory in predicting formation energies of ternary oxides from binary oxides and its implication on phase stability
journal, April 2012

Prediction model of band gap for inorganic compounds by combination of density functional theory calculations and machine learning techniques
journal, March 2016

Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysis
journal, February 2013