Machine-learning prediction of facet-dependent CO coverage on Cu electrocatalysts

Abstract

Copper-based electrocatalysts, which hold great promise in selectively reducing CO₂ into multicarbon products, have attracted significant recent interest, both experimentally and theoretically. While many studies have suggested a strong dependence of catalytic selectivity on the concentration of the ^*CO reaction intermediate on the Cu surface, it remains challenging for a direct experimental probe of the CO coverage. This necessitates a reliable computational method that can accurately establish the theoretical coverage-dependent phase diagram of CO adsorbates on the catalyst. Here we propose a scheme composed of density functional theory calculations, machine-learning force fields and graph neural networks as a solution. This method enables a fast screening of 7 million adsorption configurations based on a small set of density functional theory data, with a balance between accuracy and efficiency tuned by the combinatorial use of machine-learning force field and graph neural network models. We have investigated eight different Cu facets and discovered that the high-index facets such as (310), (210) and (322) exhibit a much higher CO coverage than the low-index counterparts such as (111), leading to an increased opportunity for C–C coupling for the former. Our results can provide a new perspective for the understanding of the fundamental role of CO coverage on the Cu surface for electrochemical CO₂ reduction.

Keywords

Machine-learning force fields / density functional theory / graph neural networks / coverage effect / electrochemical CO₂ reduction

References

[1] Beck A,van Bokhoven JA.Identifying and avoiding dead ends in the characterization of heterogeneous catalysts at the gas-solid interface.Nat Catal2023;6:873-84

[2] Li X,Fang Y,Perez-Ramirez J.Advances in heterogeneous single-cluster catalysis.Nat Rev Chem2023;7:754-67

[3] Bruix A,Andersen M.First-principles-based multiscale modelling of heterogeneous catalysis.Nat Catal2019;2:659-70

[4] Handoko AD,Jenndy ,Seh ZW.Understanding heterogeneous electrocatalytic carbon dioxide reduction through operando techniques.Nat Catal2018;1:922-34

[5] Tran K.Active learning across intermetallics to guide discovery of electrocatalysts for CO₂ reduction and H₂ evolution.Nat Catal2018;1:696-703

[6] Greeley J,Bondarenko AS.Alloys of platinum and early transition metals as oxygen reduction electrocatalysts.Nat Chem2009;1:552-6

[7] Wu Z,Li Y,Li J.Improving the DFT computational accuracy for CO activation on Fe surfaces by Bayesian error estimation functional with van der Waals correlation.Comput Theor Chem2023;1219:113968

[8] Araujo RB,Dos Santos EC.Adsorption energies on transition metal surfaces: towards an accurate and balanced description.Nat Commun2022;13:6853 PMCID:PMC9652424

[9] Nørskov JK,Logadottir A.Trends in the exchange current for hydrogen evolution.J Electrochem Soc2005;152:J23

[10] Nørskov JK,Logadottir A.Origin of the overpotential for oxygen reduction at a fuel-cell cathode.J Phys Chem B2004;108:17886-92

[11] Greeley J,Bonde J,Nørskov JK.Computational high-throughput screening of electrocatalytic materials for hydrogen evolution.Nat Mater2006;5:909-13

[12] Liu X,Peng H,Chan K.Understanding trends in electrochemical carbon dioxide reduction rates.Nat Commun2017;8:15438 PMCID:PMC5458145

[13] Zhan C,Rettenmaier C.Revealing the CO coverage-driven C–C coupling mechanism for electrochemical CO₂ reduction on Cu₂O nanocubes via operando raman spectroscopy.ACS Catal2021;11:7694-701 PMCID:PMC8256421

[14] Cave ER,Kuhl KP.Trends in the catalytic activity of hydrogen evolution during CO₂ electroreduction on transition metals.ACS Catal2018;8:3035-40

[15] Ding H,Yang X.Role of surface hydrogen coverage in C–C coupling process for CO₂ electroreduction on Ni-based catalysts.ACS Catal2024;14:14330-8

[16] Zheng S,Pan J,Li S.Multi-center cooperativity enables facile C–C coupling in electrochemical CO₂ reduction on a Ni₂P catalyst.ACS Catal2023;13:2847-56

[17] Zheng, S.; Ding, H.; Yang, X.; Li, S.; Pan, F. Automating discovery of electrochemical urea synthesis reaction paths via active learning and graph theory. https://www.chinesechemsoc.org/doi/full/10.31635/ccschem.024.202404955 (accessed 2024-02-07).

[18] Ghanekar PG,Greeley J.Adsorbate chemical environment-based machine learning framework for heterogeneous catalysis.Nat Commun2022;13:5788 PMCID:PMC9527237

[19] Xiong Y,Zhou J,Hao F.Electrochemical nitrate reduction: ammonia synthesis and the beyond.Adv Mater2024;36:e2304021

[20] Zheng S,Shi ZZ,Pan F.The loss of interfacial water-adsorbate hydrogen bond connectivity position surface-active hydrogen as a crucial intermediate to enhance nitrate reduction reaction.J Am Chem Soc2024;146:26965-74

[21] Enger BC.Nickel and Fischer-Tropsch synthesis.Catal Rev2012;54:437-88

[22] Chen Y,Duyar MS,Khodakov AY.Carbon-based catalysts for Fischer-Tropsch synthesis.Chem Soc Rev2021;50:2337-66

[23] Weststrate CJK,Garcia Rodriguez D,Fredriksson HOA.Mechanistic insight into carbon–carbon bond formation on cobalt under simulated Fischer-Tropsch synthesis conditions.Nat Commun2020;11:750 PMCID:PMC7005166

[24] Ai C,Tygesen AS,Hansen HA.Impact of hydrogen concentration for CO2 reduction on PdHx : a combination study of cluster expansion and kinetics analysis.J Catal2023;428:115188

[25] Deshpande S,Greeley J.Graph theory approach to determine configurations of multidentate and high coverage adsorbates for heterogeneous catalysis.npj Comput Mater2020;6:345

[26] Bang K,Park Y,Han SS.Machine learning-enabled exploration of the electrochemical stability of real-scale metallic nanoparticles.Nat Commun2023;14:3004 PMCID:PMC10213026

[27] Liu P,Avargues N.Combining machine learning and many-body calculations: coverage-dependent adsorption of CO on Rh(111).Phys Rev Lett.2023;130:078001

[28] Jenner B,Mckenzie P.Completeness results for graph isomorphism.J Comput Syst Sci2003;66:549-66

[29] Sumaria V.CO organization at ambient pressure on stepped Pt surfaces: first principles modeling accelerated by neural networks.Chem Sci2021;12:15543-55 PMCID:PMC8654054

[30] Yang Y,Kitchin JR.Machine-learning accelerated geometry optimization in molecular simulation.J Chem Phys2021;154:234704

[31] Jha D,Ward L.Enabling deeper learning on big data for materials informatics applications.Sci Rep2021;11:4244 PMCID:PMC7895970

[32] Zhang L,Wang H,E W.Deep potential molecular dynamics: a scalable model with the accuracy of quantum mechanics.Phys Rev Lett2018;120:143001

[33] Ulissi ZW,Xiao J.Machine-learning methods enable exhaustive searches for active bimetallic facets and reveal active site motifs for CO₂ reduction.ACS Catal2017;7:6600-8

[34] Sumaria V,Tao FF.Atomic-scale mechanism of platinum catalyst restructuring under a pressure of reactant gas.J Am Chem Soc2023;145:392-401

[35] Xu W,Andersen M.Predicting binding motifs of complex adsorbates using machine learning with a physics-inspired graph representation.Nat Comput Sci2022;2:443-50

[36] Boes JR,Winther K.Graph theory approach to high-throughput surface adsorption structure generation.J Phys Chem A2019;123:2281-5

[37] Eren B,Patera LL.Activation of Cu(111) surface by decomposition into nanoclusters driven by CO adsorption.Science2016;351:475-8

[38] Pérez-Gallent E,Calle-Vallejo F.Spectroscopic observation of a hydrogenated CO dimer intermediate during CO reduction on Cu(100) electrodes.Angew Chem Int Ed Engl2017;56:3621-4

[39] Nakano H,Fujitani T.Structure-dependent kinetics for synthesis and decomposition of formate species over Cu(111) and Cu(110) model catalysts.J Phys Chem B2001;105:1355-65

[40] Hori Y,Koga O.Electrochemical reduction of carbon dioxide at various series of copper single crystal electrodes.J Mol Catal A Chem2003;199:39-47

[41] Wang S,Su H.First-Principles microkinetic study of methanol synthesis on Cu(221) and ZnCu(221) surfaces.Chin J Chem Phys2018;31:284-90

[42] Schouten KJP,Koper MTM.Structure sensitivity of the electrochemical reduction of carbon monoxide on copper single crystals.ACS Catal2013;3:1292-5

[43] Guo W.Effect of local metal microstructure on adsorption on bimetallic surfaces: atomic nitrogen on Ni/Pt(111).J Chem Phys2013;138:174702

[44] Deshpande S.First-Principles analysis of coverage, ensemble, and solvation effects on selectivity trends in NO electroreduction on Pt₃Sn alloys.ACS Catal2020;10:9320-7

[45] Ojeda M,Nilekar AU,Mavrikakis M.CO activation pathways and the mechanism of Fischer-Tropsch synthesis.J Catal2010;272:287-97

[46] Stróż K.Plane groups - from basic to advanced crystallographic concepts.Z Kristallogr Cryst Mater2003;218:642-9

[47] Hoffmann, F. Symmetry in the plane: about wallpaper patterns, islamic mosaics, drawings from escher, and heterogeneous catalysts. In Introduction to Crystallography; Hoffmann, F., Eds.; Springer International Publishing: Cham, 2020; pp 127-50.

[48] Hahn, T. The 17 plane groups (two-dimensional space groups). In International Tables for Crystallography Volume A: Space-group symmetry; Hahn, T., Ed. Springer Netherlands: Dordrecht, 2002; pp 92-109.

[49] Chanussot L,Goyal S.Open catalyst 2020 (OC20) dataset and community challenges.ACS Catal2021;11:6059-72

[50] Lu D,Chen M.86 PFLOPS deep potential molecular dynamics simulation of 100 million atoms with ab initio accuracy.Comput Phys Commun2021;259:107624

[51] Weng M,Qian G.Identify crystal structures by a new paradigm based on graph theory for building materials big data.Sci China Chem2019;62:982-6

[52] Li S,Wang Z.Graph-based discovery and analysis of atomic-scale one-dimensional materials.Natl Sci Rev2022;9:nwac028 PMCID:PMC9170357

[53] Li S,Chen D,Nie Z.Encoding the atomic structure for machine learning in materials science.WIREs Comput Mol Sci2022;12:e1558

[54] Xie T.Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties.Phys Rev Lett2018;120:145301

[55] Steiner M.Autonomous reaction network exploration in homogeneous and heterogeneous catalysis.Top Catal2022;65:6-39 PMCID:PMC8816766

[56] Gu GH,Jung Y.Automated exploitation of the big configuration space of large adsorbates on transition metals reveals chemistry feasibility.Nat Commun2022;13:2087 PMCID:PMC9043206

[57] Qiao Z,Vahdat A,Anandkumar A.State-specific protein-ligand complex structure prediction with a multiscale deep generative model.Nat Mach Intell2024;6:195-208

[58] Grabow LC,Nørskov JK.Understanding trends in catalytic activity: the effect of adsorbate-adsorbate interactions for CO oxidation over transition metals.Top Catal2010;53:298-310

[59] Lausche AC,Khan TS.On the effect of coverage-dependent adsorbate-adsorbate interactions for CO methanation on transition metal surfaces.J Catal2013;307:275-82

[60] Huang Y,Hirunsit P.Electrochemical reduction of CO₂ using copper single-crystal surfaces: effects of CO^* coverage on the selective formation of ethylene.ACS Catal2017;7:1749-56

[61] Hou J,Li J,Lu Q.Correlating CO coverage and CO electroreduction on Cu via high-pressure in situ spectroscopic and reactivity investigations.J Am Chem Soc2022;144:22202-11

[62] Jin J,Min Q.Constrained C₂ adsorbate orientation enables CO-to-acetate electroreduction.Nature2023;617:724-9

[63] Zheng Y,Ma Z.Seeded growth of gold-copper janus nanostructures as a tandem catalyst for efficient electroreduction of CO₂ to C₂₊ products.Small2022;18:e2201695

[64] Sun W,Jiang Y.V-doped Cu₂ Se hierarchical nanotubes enabling flow-cell CO₂ electroreduction to ethanol with high efficiency and selectivity.Adv Mater2022;34:e2207691

[65] Yan X,Wu Y.Boosting CO₂ electroreduction to C₂₊ products on fluorine-doped copper.Green Chem2022;24:1989-94

[66] Reller C,Volkova E.Selective electroreduction of CO₂ toward ethylene on nano dendritic copper catalysts at high current density.Adv Energy Mater2017;7:1602114

[67] Xiang K,Fu Y.Boosting CO₂ electroreduction towards C₂₊ products via CO^* intermediate manipulation on copper-based catalysts.Environ Sci Nano2022;9:911-53