Search Menu
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information
Search Scholarly Publications

Title: Active Learning A Neural Network Model For Gold Clusters & Bulk From Sparse First Principles Training Data

Abstract

Abstract Small metal clusters are of fundamental scientific interest and of tremendous significance in catalysis. These nanoscale clusters display diverse geometries and structural motifs depending on the cluster size; a knowledge of this size‐dependent structural motifs and their dynamical evolution has been of longstanding interest. Given the high computational cost of first‐principles calculations, molecular modeling and atomistic simulations such as molecular dynamics (MD) has proven to be an important complementary tool to aid this understanding. Classical MD typically employ predefined functional forms which limits their ability to capture such complex size‐dependent structural and dynamical transformation. Neural Network (NN) based potentials represent flexible alternatives and in principle, well‐trained NN potentials can provide high level of flexibility, transferability and accuracy on‐par with the reference model used for training. A major challenge, however, is that NN models are interpolative and requires large quantities ( or greater) of training data to ensure that the model adequately samples the energy landscape both near and far‐from‐equilibrium. A highly desirable goal is minimize the number of training data, especially if the underlying reference model is first‐principles based and hence expensive. Here, we introduce an active learning (AL) scheme that trains a NN model on‐the‐fly with minimal amountmore » of first‐principles based training data. Our AL workflow is initiated with a sparse training dataset ( 1 to 5 data points) and is updated on‐the‐fly via a Nested Ensemble Monte Carlo scheme that iteratively queries the energy landscape in regions of failure and updates the training pool to improve the network performance. Using a representative system of gold clusters, we demonstrate that our AL workflow can train a NN with 500 total reference calculations. Using an extensive DFT test set of ∼1100 configurations, we show that our AL‐NN is able to accurately predict both the DFT energies and the forces for clusters of a myriad of different sizes. Our NN predictions are within 30 meV/atom and 40 meV/Å of the reference DFT calculations. Moreover, our AL‐NN model also adequately captures the various size‐dependent structural and dynamical properties of gold clusters in excellent agreement with DFT calculations and available experiments. We finally show that our AL‐NN model also captures bulk properties reasonably well, even though they were not included in the training data.« less

Authors:
 [1];  [2];  [3];  [2];  [4];  [2]
+ Show Author Affiliations
  1. Argonne National Lab. (ANL), Argonne, IL (United States)
  2. Argonne National Lab. (ANL), Argonne, IL (United States); Univ. of Illinois, Chicago, IL (United States)
  3. Indian Inst. of Technology Madras, Chennai (India)
  4. Univ. of Louisville, KY (United States)
Publication Date:
Research Org.:
Argonne National Laboratory (ANL), Argonne, IL (United States)
Sponsoring Org.:
USDOE Office of Science (SC), Basic Energy Sciences (BES)
OSTI Identifier:
1756631
Alternate Identifier(s):
OSTI ID: 1786050
Grant/Contract Number:  
AC02-06CH11357; AC02-05CH11231
Resource Type:
Accepted Manuscript
Journal Name:
ChemCatChem
Additional Journal Information:
Journal Volume: 12; Journal Issue: 19; Journal ID: ISSN 1867-3880
Publisher:
ChemPubSoc Europe
Country of Publication:
United States
Language:
English
Subject:
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY

Citation Formats

Loeffler, Troy D., Manna, Sukriti, Patra, Tarak K., Chan, Henry, Narayanan, Badri, and Sankaranarayanan, Subramanian. Active Learning A Neural Network Model For Gold Clusters & Bulk From Sparse First Principles Training Data. United States: N. p., 2020. Web. doi:10.1002/cctc.202000774.
Copy to clipboard
Loeffler, Troy D., Manna, Sukriti, Patra, Tarak K., Chan, Henry, Narayanan, Badri, & Sankaranarayanan, Subramanian. Active Learning A Neural Network Model For Gold Clusters & Bulk From Sparse First Principles Training Data. United States. https://doi.org/10.1002/cctc.202000774
Copy to clipboard
Loeffler, Troy D., Manna, Sukriti, Patra, Tarak K., Chan, Henry, Narayanan, Badri, and Sankaranarayanan, Subramanian. Wed . "Active Learning A Neural Network Model For Gold Clusters & Bulk From Sparse First Principles Training Data". United States. https://doi.org/10.1002/cctc.202000774. https://www.osti.gov/servlets/purl/1756631.
Copy to clipboard
@article{osti_1756631,
title = {Active Learning A Neural Network Model For Gold Clusters & Bulk From Sparse First Principles Training Data},
author = {Loeffler, Troy D. and Manna, Sukriti and Patra, Tarak K. and Chan, Henry and Narayanan, Badri and Sankaranarayanan, Subramanian},
abstractNote = {Abstract Small metal clusters are of fundamental scientific interest and of tremendous significance in catalysis. These nanoscale clusters display diverse geometries and structural motifs depending on the cluster size; a knowledge of this size‐dependent structural motifs and their dynamical evolution has been of longstanding interest. Given the high computational cost of first‐principles calculations, molecular modeling and atomistic simulations such as molecular dynamics (MD) has proven to be an important complementary tool to aid this understanding. Classical MD typically employ predefined functional forms which limits their ability to capture such complex size‐dependent structural and dynamical transformation. Neural Network (NN) based potentials represent flexible alternatives and in principle, well‐trained NN potentials can provide high level of flexibility, transferability and accuracy on‐par with the reference model used for training. A major challenge, however, is that NN models are interpolative and requires large quantities ( or greater) of training data to ensure that the model adequately samples the energy landscape both near and far‐from‐equilibrium. A highly desirable goal is minimize the number of training data, especially if the underlying reference model is first‐principles based and hence expensive. Here, we introduce an active learning (AL) scheme that trains a NN model on‐the‐fly with minimal amount of first‐principles based training data. Our AL workflow is initiated with a sparse training dataset ( 1 to 5 data points) and is updated on‐the‐fly via a Nested Ensemble Monte Carlo scheme that iteratively queries the energy landscape in regions of failure and updates the training pool to improve the network performance. Using a representative system of gold clusters, we demonstrate that our AL workflow can train a NN with 500 total reference calculations. Using an extensive DFT test set of ∼1100 configurations, we show that our AL‐NN is able to accurately predict both the DFT energies and the forces for clusters of a myriad of different sizes. Our NN predictions are within 30 meV/atom and 40 meV/Å of the reference DFT calculations. Moreover, our AL‐NN model also adequately captures the various size‐dependent structural and dynamical properties of gold clusters in excellent agreement with DFT calculations and available experiments. We finally show that our AL‐NN model also captures bulk properties reasonably well, even though they were not included in the training data.},
doi = {10.1002/cctc.202000774},
journal = {ChemCatChem},
number = 19,
volume = 12,
place = {United States},
year = {Wed Jun 10 00:00:00 EDT 2020},
month = {Wed Jun 10 00:00:00 EDT 2020}
}
Copy to clipboard
Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record
https://doi.org/10.1002/cctc.202000774
Copyright Statement
Other availability
Search WorldCat Search WorldCat to find libraries that may hold this journal
Save / Share:
Export Metadata
Save to My Library
You must Sign In or Create an Account in order to save documents to your library.

Works referenced in this record:

Oxide‐Supported Gold Clusters and Nanoparticles in Catalysis: A Computational Chemistry Perspective
journal, September 2018

Computational high-throughput screening of electrocatalytic materials for hydrogen evolution
journal, October 2006

  • Greeley, Jeff; Jaramillo, Thomas F.; Bonde, Jacob
  • Nature Materials, Vol. 5, Issue 11, p. 909-913
  • DOI: 10.1038/nmat1752

Nonscalable Oxidation Catalysis of Gold Clusters
journal, December 2013

  • Yamazoe, Seiji; Koyasu, Kiichirou; Tsukuda, Tatsuya
  • Accounts of Chemical Research, Vol. 47, Issue 3
  • DOI: 10.1021/ar400209a

Structural Optimization of Silver Clusters up to 80 Atoms with Gupta and Sutton-Chen Potentials
journal, May 2005

  • Shao, Xueguang; Liu, Xiaomeng; Cai, Wensheng
  • Journal of Chemical Theory and Computation, Vol. 1, Issue 4
  • DOI: 10.1021/ct049865j

Is coupled cluster singles and doubles (CCSD) more computationally intensive than quadratic configuration interaction (QCISD)?
journal, April 1989

  • Scuseria, Gustavo E.; Schaefer, Henry F.
  • The Journal of Chemical Physics, Vol. 90, Issue 7
  • DOI: 10.1063/1.455827

Catalysis by clusters with precise numbers of atoms
journal, July 2015

Gold Catalysis in (Supra)Molecular Cages to Control Reactivity and Selectivity
journal, October 2018

  • Jans, Anne C. H.; Caumes, Xavier; Reek, Joost N. H.
  • ChemCatChem, Vol. 11, Issue 1
  • DOI: 10.1002/cctc.201801399

Optical absorption spectra of Au 7 , Au 9 , Au 11 , and Au 13 , and their cations: Gold clusters with 6, 7, 8, 9, 10, 11, 12, and 13 s ‐electrons
journal, September 1994

  • Collings, B. A.; Athanassenas, K.; Lacombe, D.
  • The Journal of Chemical Physics, Vol. 101, Issue 5
  • DOI: 10.1063/1.467535

The shape of Au8: gold leaf or gold nugget?
journal, January 2013

  • Serapian, Stefano A.; Bearpark, Michael J.; Bresme, Fernando
  • Nanoscale, Vol. 5, Issue 14
  • DOI: 10.1039/c3nr01500a

First-principles study of the geometric and electronic structure of Au 13 clusters: Importance of the prism motif
journal, April 2008

Structural study of gold clusters
journal, March 2006

  • Xiao, Li; Tollberg, Bethany; Hu, Xiankui
  • The Journal of Chemical Physics, Vol. 124, Issue 11
  • DOI: 10.1063/1.2179419

Less is more: Sampling chemical space with active learning
journal, June 2018

  • Smith, Justin S.; Nebgen, Ben; Lubbers, Nicholas
  • The Journal of Chemical Physics, Vol. 148, Issue 24
  • DOI: 10.1063/1.5023802

Oxidative Dehydrogenation of Cyclohexane on Cobalt Oxide (Co 3 O 4 ) Nanoparticles: The Effect of Particle Size on Activity and Selectivity
journal, September 2012

  • Tyo, Eric C.; Yin, Chunrong; Di Vece, Marcel
  • ACS Catalysis, Vol. 2, Issue 11
  • DOI: 10.1021/cs300479a

Density-functional study of Au n ( n = 2 – 2 0 ) clusters: Lowest-energy structures and electronic properties
journal, July 2002

High Throughput Experimental and Theoretical Predictive Screening of Materials − A Comparative Study of Search Strategies for New Fuel Cell Anode Catalysts
journal, October 2003

  • Strasser, Peter; Fan, Qun; Devenney, Martin
  • The Journal of Physical Chemistry B, Vol. 107, Issue 40
  • DOI: 10.1021/jp030508z

Large piezoelectric response of van der Waals layered solids
journal, January 2018

  • Manna, Sukriti; Gorai, Prashun; Brennecka, Geoff L.
  • Journal of Materials Chemistry C, Vol. 6, Issue 41
  • DOI: 10.1039/C8TC02560F

Enhanced Piezoelectric Response of AlN via CrN Alloying
journal, March 2018

Tuning the piezoelectric and mechanical properties of the AlN system via alloying with YN and BN
journal, September 2017

  • Manna, Sukriti; Brennecka, Geoff L.; Stevanović, Vladan
  • Journal of Applied Physics, Vol. 122, Issue 10
  • DOI: 10.1063/1.4993254

Mechanical stability of possible structures of PtN investigated using first-principles calculations
journal, March 2006

Spherical harmonics based descriptor for neural network potentials: Structure and dynamics of Au 147 nanocluster
journal, May 2017

  • Jindal, Shweta; Chiriki, Siva; Bulusu, Satya S.
  • The Journal of Chemical Physics, Vol. 146, Issue 20
  • DOI: 10.1063/1.4983392

Sustainable Gold Catalysis in Water Using Cyclodextrin‐tagged NHC‐Gold Complexes
journal, August 2019

Density functional theory in surface chemistry and catalysis
journal, January 2011

  • Norskov, J. K.; Abild-Pedersen, F.; Studt, F.
  • Proceedings of the National Academy of Sciences, Vol. 108, Issue 3
  • DOI: 10.1073/pnas.1006652108

Unraveling the Planar-Globular Transition in Gold Nanoclusters through Evolutionary Search
journal, November 2016

  • Kinaci, Alper; Narayanan, Badri; Sen, Fatih G.
  • Scientific Reports, Vol. 6, Issue 1
  • DOI: 10.1038/srep34974

ANN-based estimator for distillation using Levenberg–Marquardt approach
journal, March 2007

  • Singh, Vijander; Gupta, Indra; Gupta, H. O.
  • Engineering Applications of Artificial Intelligence, Vol. 20, Issue 2
  • DOI: 10.1016/j.engappai.2006.06.017

Atom-centered symmetry functions for constructing high-dimensional neural network potentials
journal, February 2011

  • Behler, Jörg
  • The Journal of Chemical Physics, Vol. 134, Issue 7
  • DOI: 10.1063/1.3553717

Embedded-atom-method functions for the fcc metals Cu, Ag, Au, Ni, Pd, Pt, and their alloys
journal, June 1986

Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set
journal, July 1996

Nested sampling in the canonical ensemble: Direct calculation of the partition function from NVT trajectories
journal, September 2013

  • Nielsen, Steven O.
  • The Journal of Chemical Physics, Vol. 139, Issue 12
  • DOI: 10.1063/1.4821761

Two-to-three dimensional transition in neutral gold clusters: The crucial role of van der Waals interactions and temperature
journal, January 2019

Atomically Precise Gold Nanoclusters as New Model Catalysts
journal, March 2013

  • Li, Gao; Jin, Rongchao
  • Accounts of Chemical Research, Vol. 46, Issue 8
  • DOI: 10.1021/ar300213z

Quantum Monte Carlo calculations of the optical gaps of Ge nanoclusters using core-polarization potentials
journal, January 2007

Global minimization of gold clusters by combining neural network potentials and the basin-hopping method
journal, January 2015

  • Ouyang, Runhai; Xie, Yu; Jiang, De-en
  • Nanoscale, Vol. 7, Issue 36
  • DOI: 10.1039/C5NR03903G

Characterization of Elastic Modulus Across the (Al 1–x Sc x )N System Using DFT and Substrate-Effect-Corrected Nanoindentation
journal, November 2018

  • Wu, Dong; Chen, Yachao; Manna, Sukriti
  • IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Vol. 65, Issue 11
  • DOI: 10.1109/TUFFC.2018.2862240

New algorithm in the basin hopping Monte Carlo to find the global minimum structure of unary and binary metallic nanoclusters
journal, April 2008

  • Kim, Hyoung Gyu; Choi, Si Kyung; Lee, Hyuck Mo
  • The Journal of Chemical Physics, Vol. 128, Issue 14
  • DOI: 10.1063/1.2900644

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

Embedded-atom method: Derivation and application to impurities, surfaces, and other defects in metals
journal, June 1984

Pt Nanoclusters Confined within Metal–Organic Framework Cavities for Chemoselective Cinnamaldehyde Hydrogenation
journal, April 2014

  • Guo, Zhiyong; Xiao, Chaoxian; Maligal-Ganesh, Raghu V.
  • ACS Catalysis, Vol. 4, Issue 5, p. 1340-1348
  • DOI: 10.1021/cs400982n

Multi-fidelity machine learning models for accurate bandgap predictions of solids
journal, March 2017

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

Ionic vs. van der Waals layered materials: identification and comparison of elastic anisotropy
journal, January 2018

  • McKinney, Robert W.; Gorai, Prashun; Manna, Sukriti
  • Journal of Materials Chemistry A, Vol. 6, Issue 32
  • DOI: 10.1039/C8TA04933E

Bond order potential for gold
journal, September 2012

A high-throughput infrastructure for density functional theory calculations
journal, June 2011

Long-range Finnis–Sinclair potentials
journal, March 1990

Reactive forcefield for simulating gold surfaces and nanoparticles
journal, June 2010

Neural network potential-energy surfaces in chemistry: a tool for large-scale simulations
journal, January 2011

  • Behler, Jörg
  • Physical Chemistry Chemical Physics, Vol. 13, Issue 40
  • DOI: 10.1039/c1cp21668f

An implementation of artificial neural-network potentials for atomistic materials simulations: Performance for TiO2
journal, March 2016

Metal Catalysts for Heterogeneous Catalysis: From Single Atoms to Nanoclusters and Nanoparticles
journal, April 2018

A coarse-grained deep neural network model for liquid water
journal, November 2019

  • Patra, Tarak K.; Loeffler, Troy D.; Chan, Henry
  • Applied Physics Letters, Vol. 115, Issue 19
  • DOI: 10.1063/1.5116591

Asynchronous multicanonical basin hopping method and its application to cobalt nanoclusters
journal, June 2005

  • Zhan, Lixin; Chen, Jeff Z. Y.; Liu, Wing-Ki
  • The Journal of Chemical Physics, Vol. 122, Issue 24
  • DOI: 10.1063/1.1940028

Dynamics of Place-Exchange Reactions on Monolayer-Protected Gold Cluster Molecules
journal, May 1999

  • Hostetler, Michael J.; Templeton, Allen C.; Murray, Royce W.
  • Langmuir, Vol. 15, Issue 11
  • DOI: 10.1021/la981598f

Neural network potentials for dynamics and thermodynamics of gold nanoparticles
journal, February 2017

  • Chiriki, Siva; Jindal, Shweta; Bulusu, Satya S.
  • The Journal of Chemical Physics, Vol. 146, Issue 8
  • DOI: 10.1063/1.4977050

Global minima for transition metal clusters described by Sutton–Chen potentials
journal, January 1998

  • Doye, Jonathan P. K.; Wales, David J.
  • New Journal of Chemistry, Vol. 22, Issue 7
  • DOI: 10.1039/a709249k

Describing the Diverse Geometries of Gold from Nanoclusters to Bulk—A First-Principles-Based Hybrid Bond-Order Potential
journal, June 2016

  • Narayanan, Badri; Kinaci, Alper; Sen, Fatih G.
  • The Journal of Physical Chemistry C, Vol. 120, Issue 25
  • DOI: 10.1021/acs.jpcc.6b02934

Recent Advances in Asymmetric Gold Catalysis
journal, June 2010

Active Learning the Potential Energy Landscape for Water Clusters from Sparse Training Data
journal, January 2020

  • Loeffler, Troy D.; Patra, Tarak K.; Chan, Henry
  • The Journal of Physical Chemistry C, Vol. 124, Issue 8
  • DOI: 10.1021/acs.jpcc.0c00047

Active learning of uniformly accurate interatomic potentials for materials simulation
journal, February 2019

Gold Nanoclusters as Electrocatalysts for Energy Conversion
journal, January 2020

The solution of nonlinear inverse problems and the Levenberg-Marquardt method
journal, July 2007

Subnanometre platinum clusters as highly active and selective catalysts for the oxidative dehydrogenation of propane
journal, February 2009

  • Vajda, Stefan; Pellin, Michael J.; Greeley, Jeffrey P.
  • Nature Materials, Vol. 8, Issue 3
  • DOI: 10.1038/nmat2384

Comprehensive genetic algorithm for ab initio global optimisation of clusters
journal, March 2016

Catalysis by Small Metal Clusters
journal, September 1979

Metal Catalysts for Heterogeneous Catalysis: From Single Atoms to Nanoclusters and Nanoparticles
journal, April 2018

Structural study of gold clusters
journal, March 2006

  • Xiao, Li; Tollberg, Bethany; Hu, Xiankui
  • The Journal of Chemical Physics, Vol. 124, Issue 11
  • DOI: 10.1063/1.2179419

Neural network potentials for dynamics and thermodynamics of gold nanoparticles
journal, February 2017

  • Chiriki, Siva; Jindal, Shweta; Bulusu, Satya S.
  • The Journal of Chemical Physics, Vol. 146, Issue 8
  • DOI: 10.1063/1.4977050
    Sort by:
    [ × clear filter / sort ]

    Similar Records in DOE PAGES and OSTI.GOV collections: