Convolutional Neural Network of Atomic Surface Structures to Predict Binding Energies for High-Throughput Screening of Catalysts

Back, Seoin; Yoon, Junwoong; Tian, Nianhan; Zhong, Wen; Tran, Kevin; Ulissi, Zachary W.

doi:10.26434/chemrxiv.8150666.v1

Cited by 125 publications

(89 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The rapid rise of machine learning and deep learning in recent years, along with advances in computational capability, has led to a number of projects focused on applications in materials science. 15 , 16 In particular, various techniques have been employed toward data-driven materials discovery, such as high-throughput computation, 17 , 18 natural language processing, 19 − 21 “design-to-device” pipelines 22 , 23 and deep learning. 24 , 25 …”

Section: Introductionmentioning

confidence: 99%

3-D Inorganic Crystal Structure Generation and Property Prediction via Representation Learning

Court

Yildirim

Jain

et al. 2020

J. Chem. Inf. Model.

View full text Add to dashboard Cite

Generative models have been successfully used to synthesize completely novel images, text, music, and speech. As such, they present an exciting opportunity for the design of new materials for functional applications. So far, generative deep-learning methods applied to molecular and drug discovery have yet to produce stable and novel 3-D crystal structures across multiple material classes. To that end, we, herein, present an autoencoder-based generative deep-representation learning pipeline for geometrically optimized 3-D crystal structures that simultaneously predicts the values of eight target properties. The system is highly general, as demonstrated through creation of novel materials from three separate material classes: binary alloys, ternary perovskites, and Heusler compounds. Comparison of these generated structures to those optimized via electronic-structure calculations shows that our generated materials are valid and geometrically optimized.

show abstract

Section: Introductionmentioning

confidence: 99%

3-D Inorganic Crystal Structure Generation and Property Prediction via Representation Learning

Court

Yildirim

Jain

et al. 2020

J. Chem. Inf. Model.

View full text Add to dashboard Cite

show abstract

“…Compared with a full quantum-mechanics treatment of many-body systems, the simplicity of physics-inspired descriptors comes at a cost of limited generalization, particularly for high-throughput materials screening. Incorporation of multifidelity site features into reactivity models with machine learning (ML) algorithms has shown early promise for the prediction of adsorption energies, with an accuracy comparable to the typical error (~0.1−0.2 eV) of density functional theory (DFT) calculations [8][9][10][11][12][13][14][15][16] . However, the approach is largely black-box in nature, prohibiting its physical interpretation.…”

mentioning

confidence: 99%

Bayesian learning of chemisorption for bridging the complexity of electronic descriptors

2020

View full text Add to dashboard Cite

Building upon the d-band reactivity theory in surface chemistry and catalysis, we develop a Bayesian learning approach to probing chemisorption processes at atomically tailored metal sites. With representative species, e.g., *O and *OH, Bayesian models trained with ab initio adsorption properties of transition metals predict site reactivity at a diverse range of intermetallics and near-surface alloys while naturally providing uncertainty quantification from posterior sampling. More importantly, this conceptual framework sheds light on the orbitalwise nature of chemical bonding at adsorption sites with d-states characteristics ranging from bulk-like semi-elliptic bands to free-atom-like discrete energy levels, bridging the complexity of electronic descriptors for the prediction of novel catalytic materials.

show abstract

“…25,39 Graph-based learning, wherein small molecules or crystals are presented as undirected graphs with atoms described as nodes and edges representing the connections between the atoms, has been used to accurately account for the underlying structural and chemical properties for a diverse class of materials including small molecules, 39 periodic materials, 25,40 metal-organic frameworks, 8 and selected surfaces. 6 However, a successful implementation of such graph-based representations, or any surrogate model framework, for complex surface models incorporating a combination of multiple adsorbates, high-coverage ensembles, and complex surface geometries (steps, kinks, and other defects), remains highly challenging. The ACE-GCN model constitutes a simple strategy for treating these sources of complexity.…”

Section: Adsorbate Chemical Environment-based Graph Neural Networkmentioning

confidence: 99%

Adsorbate chemical environment-based machine learning framework for heterogeneous catalysis

Ghanekar

Deshpande

Greeley

2021

Preprint

View full text Add to dashboard Cite

Heterogeneous catalytic reactions are influenced by a subtle interplay of atomic-scale factors, ranging from the catalysts' local morphology to the presence of high adsorbate coverages.Describing such phenomena via computational models requires generation and analysis of a large space of surface atomic configurations. To address this challenge, we present the Adsorbate Chemical Environment-based Graph Convolution Neural Network (ACE-GCN), a screening workflow that can account for atomistic configurations comprising diverse adsorbates, binding locations, coordination environments, and substrate morphologies. Using this workflow, we develop catalyst surface models for two illustrative systems: (i) NO adsorbed on a Pt3Sn(111) alloy surface, of interest for nitrate electroreduction processes, where high adsorbate coverages combine with the low symmetry of the alloy substrate to produce a large configurational space, and (ii) OH* adsorbed on a stepped Pt(221) facet, of relevance to the Oxygen Reduction Reaction, wherein the presence of irregular crystal surfaces, high adsorbate coverages, and directionally-dependent adsorbate-adsorbate interactions result in the configurational complexity. In both cases, the ACE-GCN model, having trained on a fraction (~10%) of the total DFT-relaxed configurations, successfully ranks the relative stabilities of unrelaxed atomic configurations sampled from a large configurational space. This approach is expected to accelerate development of rigorous descriptions of catalyst surfaces under in-situ conditions.

show abstract

Convolutional Neural Network of Atomic Surface Structures to Predict Binding Energies for High-Throughput Screening of Catalysts

Abstract: We present an application of deep-learning convolutional neural network of atomic surface structures using atomic and Voronoi polyhedra-based neighbor information to predict adsorbate binding energies for the application in catalysis.

Cited by 125 publications

References 0 publications

3-D Inorganic Crystal Structure Generation and Property Prediction via Representation Learning

3-D Inorganic Crystal Structure Generation and Property Prediction via Representation Learning

Bayesian learning of chemisorption for bridging the complexity of electronic descriptors

Adsorbate chemical environment-based machine learning framework for heterogeneous catalysis

Contact Info

Product

Resources

About