skip to main content
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network

Abstract

Machine learning (ML) has emerged as a powerful complement to simulation for materials discovery by reducing time for evaluation of energies and properties at accuracy competitive with first-principles methods. We use genetic algorithm (GA) optimization to discover unconventional spin-crossover complexes in combination with efficient scoring from an artificial neural network (ANN) that predicts spin-state splitting of inorganic complexes. We explore a compound space of over 5600 candidate materials derived from eight metal/oxidation state combinations and a 32-ligand pool. We introduce a strategy for error-aware ML-driven discovery by limiting how far the GA travels away from the nearest ANN training points while maximizing property (i.e., spin-splitting) fitness, leading to discovery of 80% of the leads from full chemical space enumeration. Over a 51-complex subset, average unsigned errors (4.5 kcal/mol) are close to the ANN’s baseline 3 kcal/mol error. By obtaining leads from the trained ANN within seconds rather than days from a DFT-driven GA, this strategy demonstrates the power of ML for accelerating inorganic material discovery.

Authors:
ORCiD logo [1];  [1]; ORCiD logo [1]
  1. Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States
Publication Date:
Research Org.:
Massachusetts Inst. of Technology (MIT), Cambridge, MA (United States)
Sponsoring Org.:
USDOE Office of Science (SC), Basic Energy Sciences (BES) (SC-22)
OSTI Identifier:
1421993
Alternate Identifier(s):
OSTI ID: 1508752
Grant/Contract Number:  
SC0018096; N00014-17-1-2956; CBET-1704266; ACI-1548562
Resource Type:
Published Article
Journal Name:
Journal of Physical Chemistry Letters
Additional Journal Information:
Journal Name: Journal of Physical Chemistry Letters Journal Volume: 9 Journal Issue: 5; Journal ID: ISSN 1948-7185
Publisher:
American Chemical Society
Country of Publication:
United States
Language:
English
Subject:
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY; 36 MATERIALS SCIENCE

Citation Formats

Janet, Jon Paul, Chan, Lydia, and Kulik, Heather J. Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network. United States: N. p., 2018. Web. doi:10.1021/acs.jpclett.8b00170.
Janet, Jon Paul, Chan, Lydia, & Kulik, Heather J. Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network. United States. doi:10.1021/acs.jpclett.8b00170.
Janet, Jon Paul, Chan, Lydia, and Kulik, Heather J. Thu . "Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network". United States. doi:10.1021/acs.jpclett.8b00170.
@article{osti_1421993,
title = {Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network},
author = {Janet, Jon Paul and Chan, Lydia and Kulik, Heather J.},
abstractNote = {Machine learning (ML) has emerged as a powerful complement to simulation for materials discovery by reducing time for evaluation of energies and properties at accuracy competitive with first-principles methods. We use genetic algorithm (GA) optimization to discover unconventional spin-crossover complexes in combination with efficient scoring from an artificial neural network (ANN) that predicts spin-state splitting of inorganic complexes. We explore a compound space of over 5600 candidate materials derived from eight metal/oxidation state combinations and a 32-ligand pool. We introduce a strategy for error-aware ML-driven discovery by limiting how far the GA travels away from the nearest ANN training points while maximizing property (i.e., spin-splitting) fitness, leading to discovery of 80% of the leads from full chemical space enumeration. Over a 51-complex subset, average unsigned errors (4.5 kcal/mol) are close to the ANN’s baseline 3 kcal/mol error. By obtaining leads from the trained ANN within seconds rather than days from a DFT-driven GA, this strategy demonstrates the power of ML for accelerating inorganic material discovery.},
doi = {10.1021/acs.jpclett.8b00170},
journal = {Journal of Physical Chemistry Letters},
number = 5,
volume = 9,
place = {United States},
year = {2018},
month = {2}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record
DOI: 10.1021/acs.jpclett.8b00170

Citation Metrics:
Cited by: 20 works
Citation information provided by
Web of Science

Figures / Tables:

Figure 1. Figure 1.: (top) Representative descriptors in MCDL-25: metal properties, metal-adjacent (i.e., local ligand properties), and global ligand properties. (bottom) Representative complexes including Fe(II)(H2O)6 in training data and increasingly distant complexes from the training data (left to right): Fe(II)(bpy)3, Fe(II)(H2O)2(furan)4, and Fe(II)(bpy)2(furan)2. The closest training point and its distance is indicatedmore » below each complex.« less

Save / Share:

Works referencing / citing this record:

Enumeration of de novo inorganic complexes for chemical discovery and machine learning
journal, January 2020

  • Gugler, Stefan; Janet, Jon Paul; Kulik, Heather J.
  • Molecular Systems Design & Engineering, Vol. 5, Issue 1
  • DOI: 10.1039/c9me00069k

Enumeration of de novo inorganic complexes for chemical discovery and machine learning
journal, January 2020

  • Gugler, Stefan; Janet, Jon Paul; Kulik, Heather J.
  • Molecular Systems Design & Engineering, Vol. 5, Issue 1
  • DOI: 10.1039/c9me00069k

    Figures/Tables have been extracted from DOE-funded journal article accepted manuscripts.