Interpretable physics-informed machine learning approaches to accelerate electrocatalyst development

Abstract

Identifying exceptional electrocatalysts from the vast materials space remains a formidable challenge. Machine learning (ML) has emerged as a powerful tool to address this challenge, offering high efficiency while maintaining good accuracy in predictions. From this perspective, we provide a brief overview of recent advancements in ML for electrocatalyst discoveries. We emphasize the applications of physics-informed ML (PIML) models and explainable artificial intelligence (XAI) to electrocatalyst development, through which valuable physical and chemical insights can be distilled. Additionally, we delve into the challenges faced by PIML approaches, explore future directions, and discuss potential breakthroughs that could revolutionize the field of electrocatalyst development.

Keywords

Electrocatalysts / machine learning / physics-informed machine learning / explainable artificial intelligence

References

[1] Seh ZW,Dickens CF,Nørskov JK.Combining theory and experiment inelectrocatalysis: insights into materials design.Science2017;355:eaad4998

[2] Ding K,Leung MT.Recent advances in the data-driven development of emerging electrocatalysts.Curr Opin Electrochem2023;42:101404

[3] Banoth P,Kollu P.Introduction to electrocatalysts. In Noble metal-free electrocatalysts: new trends in electrocatalysts for energy applications. Vol. 2; Washington: American Chemical Society, 2022; pp. 1-37.

[4] Santos DMF.Advanced materials for electrochemical energy conversion and storage devices.Materials2021;14:7711 PMCID:PMC8706487

[5] Chen L,Chen A,Hu X.Targeted design of advanced electrocatalysts by machine learning.Chin J Catal2022;43:11-32

[6] Liao X,Xia L.Density functional theory for electrocatalysis.Energy Environ Mater2022;5:157-85

[7] Chen ZW,Ou P.Unusual Sabatier principle on high entropy alloy catalysts for hydrogen evolution reactions.Nat Commun2024;15:359 PMCID:PMC10774414

[8] Peng J,Akkiraju K.Human- and machine-centred designs of molecules and materials for sustainability and decarbonization.Nat Rev Mater2022;7:991-1009

[9] Yang T,Zhou J.High-throughput screening of transition metal single atom catalysts anchored on molybdenum disulfide for nitrogen fixation.Nano Energy2020;68:104304

[10] Yang T,Song TT,Feng YP.High-throughput identification of exfoliable two-dimensional materials with active basal planes for hydrogen evolution.ACS Energy Lett2020;5:2313-21

[11] Shen L,Yang T,Feng YP.High-throughput computational discovery and intelligent design of two-dimensional functional materials for various applications.Acc Mater Res2022;3:572-83

[12] Zhou J,Yang M,Kong W.Discovery of hidden classes of layered electrides by extensive high-throughput material screening.Chem Mater2019;31:1860-8

[13] Steinmann SN,Seh ZW.How machine learning can accelerate electrocatalysis discovery and optimization.Mater Horiz2023;10:393-406

[14] Liu C.Finding physical insights in catalysis with machine learning.Curr Opin Chem Eng2022;37:100832

[15] Xin H,Pillai HS,Huang Y.Interpretable machine learning for catalytic materials design toward sustainability.Acc Mater Res2024;5:22-34

[16] Kayode GO.Latent variable machine learning framework for catalysis: general models, transfer learning, and interpretability.JACS Au2024;4:80-91 PMCID:PMC10807004

[17] Rangarajan S.Chapter 6 - Artificial intelligence in catalysis. In Artificial intelligence in manufacturing, Elsevier, 2024; pp. 167-204.

[18] Zhang Y,Reddy GK.Descriptor-free design of multicomponent catalysts.ACS Catal2022;12:10562-71

[19] Jäger MOJ,Canova FF,Foster AS.Efficient machine-learning-aided screening of hydrogen adsorption on bimetallic nanoclusters.ACS Comb Sci2020;22:768-81 PMCID:PMC7739401

[20] Fan X,Huang D.From single metals to high-entropy alloys: how machine learning accelerates the development of metal electrocatalysts.Adv Funct Mater2024;34:2401887

[21] Park Y,Bang K.Machine learning filters out efficient electrocatalysts in the massive ternary alloy space for fuel cells.Appl Catal B Environ2023;339:123128

[22] Esterhuizen JA,Linic S.Interpretable machine learning for knowledge generation in heterogeneous catalysis.Nat Catal2022;5:175-84

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

[24] Tran R,Shuaibi M.The Open Catalyst 2022 (OC22) Dataset and challenges for oxide electrocatalysts.ACS Catal2023;13:3066-84

[25] Jain A,Hautier G.Commentary: the materials project: a materials genome approach to accelerating materials innovation.APL Mater2013;1:011002

[26] Saal JE,Aykol M,Wolverton C.Materials design and discovery with high-throughput density functional theory: the Open Quantum Materials Database (OQMD).JOM2013;65:1501-9

[27] Zhou J,Costa MD.2DMatPedia, an open computational database of two-dimensional materials from top-down and bottom-up approaches.Sci Data2019;6:86 PMCID:PMC6561947

[28] Montavon G,Fazli S.Learning invariant representations of molecules for atomization energy prediction. In Proceedings of the 26th International Conference on Neural Information Processing Systems, Curran Associates Inc.: Red Hook, USA, 2012; Vol 1, pp 440-8.

[29] Rupp M,Müller KR.Fast and accurate modeling of molecular atomization energies with machine learning.Phys Rev Lett2012;108:058301

[30] Bartók AP,Csányi G.On representing chemical environments.Phys Rev B2013;87:184115

[31] Willatt MJ,Ceriotti M.Feature optimization for atomistic machine learning yields a data-driven construction of the periodic table of the elements.Phys Chem Chem Phys2018;20:29661-8

[32] Jäger MOJ,Federici Canova F,Foster AS.Machine learning hydrogen adsorption on nanoclusters through structural descriptors.npj Comput Mater2018;4:96

[33] De S,Csányi G.Comparing molecules and solids across structural and alchemical space.Phys Chem Chem Phys2016;18:13754-69

[34] Mai H,Chen D,Caruso RA.Machine learning for electrocatalyst and photocatalyst design and discovery.Chem Rev2022;122:13478-515

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

[36] Motagamwala AH,Dumesic JA.Microkinetic analysis and scaling relations for catalyst design.Annu Rev Chem Biomol Eng2018;9:413-50

[37] Pérez-Ramírez J.Strategies to break linear scaling relationships.Nat Catal2019;2:971-6

[38] Batchelor TA,Winther SH,Jacobsen KW.High-entropy alloys as a discovery platform for electrocatalysis.Joule2019;3:834-45

[39] Artyushkova K,Olson TS,Atanassov P.Predictive modeling of electrocatalyst structure based on structure-to-property correlations of x-ray photoelectron spectroscopic and electrochemical measurements.Langmuir2008;24:9082-8

[40] Hearst M,Osuna E,Scholkopf B.Support vector machines.IEEE Intell Syst Their Appl1998;13:18-28

[41] Sun H,Gao L.High throughput screening of single atomic catalysts with optimized local structures for the electrochemical oxygen reduction by machine learning.J Energy Chem2023;81:349-57

[42] Arjmandi M,Motevassel M.Evaluating algorithms of decision tree, support vector machine and regression for anode side catalyst data in proton exchange membrane water electrolysis.Sci Rep2023;13:20309 PMCID:PMC10662483

[43] Hossain SS,Rushd S,Cheng CK.Interaction effect of process parameters and Pd-electrocatalyst in formic acid electro-oxidation for fuel cell applications: implementing supervised machine learning algorithms.Int J Energy Res2022;46:21583-97

[44] Tamtaji M,Hu Z,Chen G.A surrogate machine learning model for the design of single-atom catalyst on carbon and porphyrin supports towards electrochemistry.J Phys Chem C2023;127:9992-10000

[45] Anbari E,Iranshahi D.Experimental investigation and development of a SVM model for hydrogenation reaction of carbon monoxide in presence of Co–Mo/Al₂O₃ catalyst.Chem Eng J2015;276:213-21

[46] Sun J,Guan J.Interpretable machine learning-assisted high-throughput screening for understanding NRR electrocatalyst performance modulation between active center and C-N coordination.Energy Environ Mater2024;7:e12693

[47] Tan S,Song G.Machine learning and Shapley Additive Explanation-based interpretable prediction of the electrocatalytic performance of N-doped carbon materials.Fuel2024;355:129469

[48] Zhang Y,Ma N.Directly predicting N₂ electroreduction reaction free energy using interpretable machine learning with non-DFT calculated features.J Energy Chem2024;97:139-48

[49] Wei C,Yang Z.Data-driven design of double-atom catalysts with high H₂ evolution activity/CO₂ reduction selectivity based on simple features.J Mater Chem A2023;11:18168-78

[50] Ying Y,Luo X,Huang H.Unravelling the origin of bifunctional OER/ORR activity for single-atom catalysts supported on C₂N by DFT and machine learning.J Mater Chem A2021;9:16860-7

[51] Lin S,Wang Y,Chen Z.Directly predicting limiting potentials from easily obtainable physical properties of graphene-supported single-atom electrocatalysts by machine learning.J Mater Chem A2020;8:5663-70

[52] Lu S,Jia Z.Symbolic transform optimized convolutional neural network model for high-performance prediction and analysis of MXenes hydrogen evolution reaction catalysts.Int J Hydrogen Energy2024;85:200-9

[53] Roy D,Das A.Unravelling CO₂ reduction reaction intermediates on high entropy alloy catalysts: an interpretable machine learning approach to establish scaling relations.Chemistry2024;30:e202302679

[54] Wang Y,Ma N.Machine learning accelerated catalysts design for CO reduction: an interpretability and transferability analysis.J Mater Sci Technol2025;213:14-23

[55] Jia X.Machine learning enabled exploration of multicomponent metal oxides for catalyzing oxygen reduction in alkaline media.J Mater Chem A2024;12:12487-500

[56] Yang H,Wang Q. Convolutional neural networks and volcano plots: screening and prediction of two-dimensional single-atom catalystsar. arXiv 2024, arXiv:2402.03876. Available online: https://doi.org/10.48550/arXiv.2402.03876. (accessed 15 Jan 2025)

[57] Gilmer J,Riley PF,Dahl GE. Neural message passing for quantum chemistry. arXiv 2017, arXiv:1704.01212. Available online: https://doi.org/10.48550/arXiv.1704.01212. (accessed 15 Jan 2025)

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

[59] Park Y,Hwang S.Scalable parallel algorithm for graph neural network interatomic potentials in molecular dynamics simulations.J Chem Theory Comput2024;20:4857-68

[60] Merchant A,Schoenholz SS,Cheon G.Scaling deep learning for materials discovery.Nature2023;624:80-5 PMCID:PMC10700131

[61] Deng B,Jun K.CHGNet as a pretrained universal neural network potential for charge-informed atomistic modelling.Nat Mach Intell2023;5:1031-41

[62] Chen C.A universal graph deep learning interatomic potential for the periodic table.Nat Comput Sci2022;2:718-28

[63] Tang D,Luber S.Machine learning interatomic potentials for heterogeneous catalysis.Chemistry2024;30:e202401148

[64] Batatia I,Simm GNC,Csányi G. MACE: higher order equivariant message passing neural networks for fast and accurate force fields. arXiv 2022, arXiv:2206.07697. Available online: https://doi.org/10.48550/arXiv.2206.07697 (accessed 15 Jan 2025)

[65] Batatia I,Kovács DP. The design space of e (3)-equivariant atom-centered interatomic potentials. arXiv 2022, arXiv:2205.06643. Available online: https://doi.org/10.48550/arXiv.2205.06643 (accessed 15 Jan 2025)

[66] Riebesell J,Benner P. Matbench discovery - an evaluation framework for machine learning crystal stability prediction. arXiv 2023, arXiv:2308.14920. Available online: https://doi.org/10.48550/arXiv.2308.14920 (accessed 15 Jan 2025)

[67] Batatia I,Chiang Y. A foundation model for atomistic materials chemistry. arXiv 2023, arXiv:2401.00096. Available online: https://doi.org/10.48550/arXiv.2401.00096 (accessed 15 Jan 2025)

[68] Open AI; Achiam J, Adler S, Agarwal S, et al. Gpt-4 technical report. arXiv 2023, arXiv:2303.08774. Available online: https://doi.org/10.48550/arXiv.2303.08774 (accessed 15 Jan 2025)

[69] Yao Y,Xu K,Sun Z.A survey on large language model (LLM) security and privacy: the good, the bad, and the ugly.High Confid Comput2024;4:100211

[70] Chang Y,Wang J.A survey on evaluation of large language models.ACM Trans Intell Syst Technol2024;15:1-45

[71] Augenstein I,Cha M.Factuality challenges in the era of large language models and opportunities for fact-checking.Nat Mach Intell2024;6:852-63

[72] Patil R.A review of current trends, techniques, and challenges in large language models (LLMs).Appl Sci2024;14:2074

[73] Beltagy I,Cohan A. SciBERT: a pretrained language model for scientific text. arXiv 2019, arXiv:1903.10676. Available online: https://doi.org/10.48550/arXiv.1903.10676 (accessed 15 Jan 2025)

[74] Wang L,Du Y,Gao Y. CataLM: empowering catalyst design through large language models. arXiv 2024, arXiv:2405.17440. Available online: https://doi.org/10.48550/arXiv.2405.17440 (accessed 15 Jan 2025)

[75] Ding R,Tan A,Liu J.Unlocking new insights for electrocatalyst design: a unique data science workflow leveraging internet-sourced big data.ACS Catal2023;13:13267-81

[76] Minh D,Li YF.Explainable artificial intelligence: a comprehensive review.Artif Intell Rev2022;55:3503-68

[77] Wang SH,Wang S,Xin H.Infusing theory into deep learning for interpretable reactivity prediction.Nat Commun2021;12:5288 PMCID:PMC8421337

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

[79] Noh J,Kim S.Uncertainty-quantified hybrid machine learning/density functional theory high throughput screening method for crystals.J Chem Inf Model2020;60:1996-2003

[80] Abed J,Sanspeur RY.Pourbaix machine learning framework identifies acidic water oxidation catalysts exhibiting suppressed ruthenium dissolution.J Am Chem Soc2024;146:15740-50

[81] Zhang J,Huang S.Design high-entropy electrocatalyst via interpretable deep graph attention learning.Joule2023;7:1832-51

[82] Deringer VL,Bernstein N,Ceriotti M.Gaussian process regression for materials and molecules.Chem Rev2021;121:10073-141 PMCID:PMC8391963

[83] Ulissi ZW,Tsai C.Automated discovery and construction of surface phase diagrams using machine learning.J Phys Chem Lett2016;7:3931-5

[84] Christensen AS,Faber FA.FCHL revisited: faster and more accurate quantum machine learning.J Chem Phys2020;152:044107

[85] 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

[86] Togninalli M,Llinares-López F,Borgwardt K. Wasserstein weisfeiler-lehman graph kernels. arXiv 2019, arXiv:1906.01277. Available online: https://doi.org/10.48550/arXiv.1906.01277 (accessed 15 Jan 2025)

[87] Grisafi A,Salanne M.Predicting the charge density response in metal electrodes.Phys Rev Mater2023;7:125403

[88] Satorras VG,Welling M. E(n) equivariant graph neural networks. arXiv 2021, arXiv:2102.09844. Available online: https://doi.org/10.48550/arXiv.2102.09844 (accessed 15 Jan 2025)

[89] Zhang X,Helwig J. Artificial intelligence for science in quantum, atomistic, and continuum systems. arXiv 2023, arXiv:2307.08423. Available online: https://doi.org/10.48550/arXiv.2307.08423 (accessed 15 Jan 2025)

[90] Zitnick CL,Kolluru A. Spherical channels for modeling atomic interactions. arXiv 2022, arXiv:2206.14331. Available online: https://doi.org/10.48550/arXiv.2206.14331 (accessed 15 Jan 2025)

[91] Passaro S. Reducing SO₃ convolutions to SO₂ for efficient equivariant GNNs. arXiv 2023, arXiv:2302.03655. Available online: https://doi.org/10.48550/arXiv.2302.03655 (accessed 15 Jan 2025)

[92] Liao YL,Das A. EquiformerV2: improved equivariant transformer for scaling to higher-degree representations. arXiv 2023, arXiv:2306.12059. Available online: https://doi.org/10.48550/arXiv.2306.12059 (accessed 15 Jan 2025)

[93] Hastie TJ.Generalized additive models. In Statistical models in S, 1st ed.; Routledge, 2017; pp 249-307.

[94] Ouyang R,Ahmetcik E,Ghiringhelli LM.SISSO: a compressed-sensing method for identifying the best low-dimensional descriptor in an immensity of offered candidates.Phys Rev Mater2018;2:083802

[95] Lin X,Chang X,Zhao Z.High-throughput screening of electrocatalysts for nitrogen reduction reactions accelerated by interpretable intrinsic descriptor.Angew Chem Int Ed2023;135:e202300122

[96] Ding Z,Ma A.Single-atom catalysts based on two-dimensional metalloporphyrin monolayers for electrochemical nitrate reduction to ammonia by first-principles calculations and interpretable machine learning.Int J Hydrogen Energy2024;80:586-98

[97] Shu W,Liu JX.Structure sensitivity of metal catalysts revealed by interpretable machine learning and first-principles calculations.J Am Chem Soc2024;146:8737-45

[98] Su Y,Ye Y.Automation and machine learning augmented by large language models in a catalysis study.Chem Sci2024;15:12200-33 PMCID:PMC11304797

[99] Liu X.Toward next-generation heterogeneous catalysts: empowering surface reactivity prediction with machine learning.Engineering2024;39:25-44

[100] Yang Z.Applications of machine learning in alloy catalysts: rational selection and future development of descriptors.Adv Sci2022;9:e2106043 PMCID:PMC9036033

[101] Ribeiro MT,Guestrin C. “Why should I trust you?”: Explaining the predictions of any classifier. arXiv 2016, arXiv:1602.04938. Available online: https://doi.org/10.48550/arXiv.1602.04938 (accessed 15 Jan 2025)

[102] Van der Maaten L, Hinton G. Visualizing data using t-SNE.J Mach Learn Res2008;9:2579-605https://www.jmlr.org/papers/volume9/vandermaaten08a/vandermaaten08a.pdf. (accessed 2025-01-15).

[103] Lundberg S. A unified approach to interpreting model predictions. arXiv 2017, arXiv:1705.07874. Available online: https://doi.org/10.48550/arXiv.1705.07874 (accessed 15 Jan 2025)

[104] Omidvar N,Huang Y.Explainable AI for optimizing oxygen reduction on Pt monolayer core–shell catalysts.Electrochem Sci Adv2024;4:e202300028

[105] Li Y,Li T,Liu Y.Accelerating materials discovery for electrocatalytic water oxidation via center-environment deep learning in spinel oxides.J Mater Chem A2024;12:19362-77

[106] Roh J,Kwon H.Interpretable machine learning framework for catalyst performance prediction and validation with dry reforming of methane.Appl Catal B Environ2024;343:123454

[107] Ding R,Chen P.Machine learning-guided discovery of underlying decisive factors and new mechanisms for the design of nonprecious metal electrocatalysts.ACS Catal2021;11:9798-808

[108] Pillai HS,Wang SH.Interpretable design of Ir-free trimetallic electrocatalysts for ammonia oxidation with graph neural networks.Nat Commun2023;14:792 PMCID:PMC9922329

[109] Zhong M,Min Y.Accelerated discovery of CO₂ electrocatalysts using active machine learning.Nature2020;581:178-83

[110] Pablo-García S,Vargas-Hernández RA.Fast evaluation of the adsorption energy of organic molecules on metals via graph neural networks.Nat Comput Sci2023;3:433-42 PMCID:PMC10766545

[111] Schütt KT,Gastegger M. Equivariant message passing for the prediction of tensorial properties and molecular spectra. arXiv 2021, arXiv:2102.03150. Available online: https://doi.org/10.48550/arXiv.2102.03150 (accessed 15 Jan 2025)

[112] Szymanski NJ,Fei Y.An autonomous laboratory for the accelerated synthesis of novel materials.Nature2023;624:86-91 PMCID:PMC10700133