Seeing Is Believing: Experimental Spin States from Machine Learning Model Structure Predictions DOI Creative Commons
Michael G. Taylor, Tzuhsiung Yang,

Sean Lin

et al.

The Journal of Physical Chemistry A, Journal Year: 2020, Volume and Issue: 124(16), P. 3286 - 3299

Published: March 28, 2020

Determination of ground-state spins open-shell transition-metal complexes is critical to understanding catalytic and materials properties but also challenging with approximate electronic structure methods. As an alternative approach, we demonstrate how alone can be used guide assignment spin from experimentally determined crystal structures complexes. We first identify the limits distance-based heuristics distributions metal-ligand bond lengths over 2000 unique mononuclear Fe(II)/Fe(III) To overcome these limits, employ artificial neural networks (ANNs) predict spin-state-dependent classify experimental based on agreement ANN predictions. Although trained hybrid density functional theory data, exploit method-insensitivity geometric enable ground states for majority (ca. 80-90%) structures. utility by data-mining literature spin-crossover (SCO) complexes, which have observed temperature-dependent changes, correctly assigning almost all (>95%) in 46 Fe(II) SCO complex set. This approach represents a promising complement more conventional energy-based spin-state at low cost machine learning model.

Language: Английский

Machine Learning for Electrocatalyst and Photocatalyst Design and Discovery DOI
Haoxin Mai, Tu C. Le, Dehong Chen

et al.

Chemical Reviews, Journal Year: 2022, Volume and Issue: 122(16), P. 13478 - 13515

Published: July 21, 2022

Electrocatalysts and photocatalysts are key to a sustainable future, generating clean fuels, reducing the impact of global warming, providing solutions environmental pollution. Improved processes for catalyst design better understanding electro/photocatalytic essential improving effectiveness. Recent advances in data science artificial intelligence have great potential accelerate electrocatalysis photocatalysis research, particularly rapid exploration large materials chemistry spaces through machine learning. Here comprehensive introduction to, critical review of, learning techniques used research provided. Sources electro/photocatalyst current approaches representing these by mathematical features described, most commonly methods summarized, quality utility models evaluated. Illustrations how applied novel discovery elucidate electrocatalytic or photocatalytic reaction mechanisms The offers guide scientists on selection research. application catalysis represents paradigm shift way advanced, next-generation catalysts will be designed synthesized.

Language: Английский

Citations

277
A Comprehensive Discovery Platform for Organophosphorus Ligands for Catalysis DOI
Tobias Gensch, Gabriel dos Passos Gomes, Pascal Friederich

et al.

Journal of the American Chemical Society, Journal Year: 2022, Volume and Issue: 144(3), P. 1205 - 1217

Published: Jan. 12, 2022

The design of molecular catalysts typically involves reconciling multiple conflicting property requirements, largely relying on human intuition and local structural searches. However, the vast number potential requires pruning candidate space by efficient prediction with quantitative structure–property relationships. Data-driven workflows embedded in a library can be used to build predictive models for catalyst performance serve as blueprint novel designs. Herein we introduce kraken, discovery platform covering monodentate organophosphorus(III) ligands providing comprehensive physicochemical descriptors based representative conformer ensembles. Using quantum-mechanical methods, calculated 1558 ligands, including commercially available examples, trained machine learning predict properties over 300000 new ligands. We demonstrate application kraken systematically explore organophosphorus how existing data sets catalysis accelerate ligand selection during reaction optimization.

Language: Английский

Citations

222
Computational Discovery of Transition-metal Complexes: From High-throughput Screening to Machine Learning DOI Creative Commons
Aditya Nandy, Chenru Duan, Michael G. Taylor

et al.

Chemical Reviews, Journal Year: 2021, Volume and Issue: 121(16), P. 9927 - 10000

Published: July 14, 2021

Transition-metal complexes are attractive targets for the design of catalysts and functional materials. The behavior metal-organic bond, while very tunable achieving target properties, is challenging to predict necessitates searching a wide complex space identify needles in haystacks applications. This review will focus on techniques that make high-throughput search transition-metal chemical feasible discovery with desirable properties. cover development, promise, limitations "traditional" computational chemistry (i.e., force field, semiempirical, density theory methods) as it pertains data generation inorganic molecular discovery. also discuss opportunities leveraging experimental sources. We how advances statistical modeling, artificial intelligence, multiobjective optimization, automation accelerate lead compounds rules. overall objective this showcase bringing together from diverse areas computer science have enabled rapid uncovering structure-property relationships chemistry. aim highlight unique considerations motifs bonding (e.g., variable spin oxidation state, strength/nature) set them their apart more commonly considered organic molecules. uncertainty relative scarcity motivate specific developments machine learning representations, model training, Finally, we conclude an outlook opportunity accelerated complexes.

Language: Английский

Citations

216
Metal-based electrocatalytic conversion of CO2 to formic acid/formate DOI
Peng Ding, Haitao Zhao, Tingshuai Li

et al.

Journal of Materials Chemistry A, Journal Year: 2020, Volume and Issue: 8(42), P. 21947 - 21960

Published: Jan. 1, 2020

This review summarizes recent progress in the development of metal-based electrocatalysts for reduction CO2 to formic acid/formate. The current challenges and future research directions materials are also proposed.

Language: Английский

Citations

161
Accelerated dinuclear palladium catalyst identification through unsupervised machine learning DOI

Julian A. Hueffel,

Theresa Sperger, Ignacio Funes‐Ardoiz

et al.

Science, Journal Year: 2021, Volume and Issue: 374(6571), P. 1134 - 1140

Published: Nov. 25, 2021

Although machine learning bears enormous potential to accelerate developments in homogeneous catalysis, the frequent need for extensive experimental data can be a bottleneck implementation. Here, we report an unsupervised workflow that uses only five points. It makes use of generalized parameter databases are complemented with problem-specific silico acquisition and clustering. We showcase power this strategy challenging problem speciation palladium (Pd) catalysts, which mechanistic rationale is currently lacking. From total space 348 ligands, algorithm predicted, experimentally verified, number phosphine ligands (including previously never synthesized ones) give dinuclear Pd(I) complexes over more common Pd(0) Pd(II) species.

Language: Английский

Citations

115
Reaction-based machine learning representations for predicting the enantioselectivity of organocatalysts DOI Creative Commons
Simone Gallarati, Raimón Fabregat, Rubén Laplaza

et al.

Chemical Science, Journal Year: 2021, Volume and Issue: 12(20), P. 6879 - 6889

Published: Jan. 1, 2021

A machine learning model for enantioselectivity prediction using reaction-based molecular representations.

Language: Английский

Citations

89
tmQM Dataset—Quantum Geometries and Properties of 86k Transition Metal Complexes DOI Creative Commons
David Balcells, Bastian Bjerkem Skjelstad

Journal of Chemical Information and Modeling, Journal Year: 2020, Volume and Issue: 60(12), P. 6135 - 6146

Published: Nov. 9, 2020

We report the transition metal quantum mechanics (tmQM) data set, which contains geometries and properties of a large metal–organic compound space. tmQM comprises 86,665 mononuclear complexes extracted from Cambridge Structural Database, including Werner, bioinorganic, organometallic based on variety organic ligands 30 metals (the 3d, 4d, 5d groups 3 to 12). All are closed-shell, with formal charge in range {+1, 0, −1}e. The set provides Cartesian coordinates all optimized at GFN2-xTB level, their molecular size, stoichiometry, node degree. were computed DFT(TPSSh-D3BJ/def2-SVP) level include electronic dispersion energies, highest occupied orbital (HOMO) lowest unoccupied (LUMO) HOMO/LUMO gap, dipole moment, natural center; polarizabilities also provided. Pairwise representations showed low correlation between these properties, providing nearly continuous maps unusual regions chemical space, for example, combining wide gaps low-energy HOMO orbitals electron-rich centers. can be exploited data-driven discovery new complexes, predictive models machine learning. These may have strong impact fields chemistry plays key role, catalysis, synthesis, materials science. is an open that downloaded free https://github.com/bbskjelstad/tmqm.

Language: Английский

Citations

80
Navigating through the Maze of Homogeneous Catalyst Design with Machine Learning DOI
Gabriel dos Passos Gomes, Robert Pollice, Alán Aspuru‐Guzik

et al.

Trends in Chemistry, Journal Year: 2021, Volume and Issue: 3(2), P. 96 - 110

Published: Jan. 14, 2021

Language: Английский

Citations

79
autodE: Automated Calculation of Reaction Energy Profiles— Application to Organic and Organometallic Reactions DOI
Tom A. Young,

Joseph J. Silcock,

Alistair J. Sterling

et al.

Angewandte Chemie International Edition, Journal Year: 2020, Volume and Issue: 60(8), P. 4266 - 4274

Published: Oct. 27, 2020

Abstract Calculating reaction energy profiles to aid in mechanistic elucidation has long been the domain of expert computational chemist. Here, we introduce autodE ( https://github.com/duartegroup/autodE ), an open‐source Python package capable locating transition states (TSs) and minima delivering a full profile from 1D or 2D chemical representations. is broadly applicable study organic organometallic classes, including addition, substitution, elimination, migratory insertion, oxidative reductive elimination; it accounts for conformational sampling both TSs compatible with many electronic structure packages. The general applicability demonstrated complex multi‐step reactions, cobalt‐ rhodium‐catalyzed hydroformylation Ireland–Claisen rearrangement.

Language: Английский

Citations

77
Beyond hydrogen bonding: recent trends of outer sphere interactions in transition metal catalysis DOI
Jonathan Trouvé,

Rafael Gramage‐Doria

Chemical Society Reviews, Journal Year: 2021, Volume and Issue: 50(5), P. 3565 - 3584

Published: Jan. 1, 2021

The implementation of interactions beyond hydrogen bonding in the 2ndcoordination sphere transition metal catalysts is rare. However, it has already shown great promise last 5 years, providing new tools to control activity and selectivity as here reviewed.

Language: Английский

Citations

77