Aug 2026· ACS Applied Energy Materials· 0 citations· 33 references
TL;DR
This study demonstrates an example of how machine-learning-driven materials discovery can accelerate catalyst design while simultaneously offering an experimental solution to the persistent challenge of high ORR overpotential.
Abstract
Machine learning has significantly reduced the computational power necessary to estimate the free energy of adsorption of key reaction intermediates on a diverse range of catalytic surfaces. Nevertheless, translating this computational capability into the discovery and experimental validation of catalysts necessitates targeting specific questions where this faster computation can yield the most significant impact. Using the electrochemical oxygen reduction reaction (ORR) as a model reaction due to its known linear scaling relationships between key catalytic intermediates, we demonstrate that the Open Catalyst Project’s machine-learning-based calculations of adsorption energies can inform experimental catalytic research. The primary challenge we addressed was the structural effect of the ORR, wherein the higher Miller index facets of Pt exhibit diminished ORR activity in comparison to Pt(111). The Open Catalyst Project’s rapid relaxation energy calculations enabled us to screen a large number of bimetallic materials for each key intermediate of the ORR over a wide range of crystal facets. The Open Catalyst Project was able to identify that PdAu3 can overcome the negative structural effects observed on the higher Miller index facets of Pt for the ORR. Electrochemical experimentation via rotating disk electrode linear sweep voltammetry and Tafel slope analysis revealed that polycrystalline PdAu3 nanoparticles exhibit an improved onset potential for the ORR compared to commercial polycrystalline Pt/C. Thus, this study demonstrates an example of how machine-learning-driven materials discovery can accelerate catalyst design while simultaneously offering an experimental solution to the persistent challenge of high ORR overpotential.
Molecular design of an active electrocatalyst can determine the primary product. Precise selection of the active metal center influences the adsorption of reactants, intermediates, and product selectivity and ultimately affects the efficiency of the process. Here, we predict a potent catalyst, namely, Fe@C6N6, with enhanced activity and low overpotential for CO2 reduction reactions (CO2RRs) utilizing the density functional theory (DFT) approach, coupled with machine-learning-based model building. A series of 24 transition-metal single-atom catalysts (SACs) anchored on the C6N6 monolayer were screened, among which 19 exhibit effective CO2 activation. Employing density functional theory (DFT), a systematic study on reaction pathways demonstrates that 13 catalysts favor CH4 formation, while the remaining 6 exhibit selectivity toward CH3OH production. Additionally, 9 among the 19 catalysts are further screened out as they preferentially promote the competing hydrogen evolution reaction (HER). The active SACs demonstrate strong CO2RR activity, with Fe@C6N6 emerging as the most promising catalyst with the lowest limiting potential (0.39 V). To further establish atomic–property relationships, a machine-learning (ML) regression model was developed to predict adsorption energies using statistically significant descriptors and identify key factors governing CO2 adsorption. Furthermore, a classification model developed using the elemental characteristics of the metal atoms shows excellent correlation with the DFT results across the dataset. This combined DFT and machine-learning approach provides mechanistic insights for the rational development of efficient SACs for CO2 electroreduction.
Unknown authors· ACS Applied Energy Materials· 0 citations
Catalytic N2O decomposition in the presence of O2 is a key process for addressing environmental challenges, such as greenhouse gas emissions and ozone layer depletion. However, the identification of efficient catalysts for this reaction remains challenging owing to the limitations of conventional methods. In this study, we employ a machine learning approach designed to accelerate the discovery of effective direct N2O decomposition catalysts. Starting with 51 catalysts and conducting 37 cycles of a closed-loop discovery system (machine-learning prediction + experiment), 633 catalysts are experimentally tested. Over 10 multi-elemental catalysts exhibiting superior activity are identified, surpassing the performance of the originally identified best catalyst. Among them, Rh(1)–Pd(2)/ZrO2_EP exhibits the highest catalytic performance for N2O decomposition. Through control experiments and a combination of ex situ and in situ characterizations, we identify the essential role of each component within the catalyst system. Catalytic N2O decomposition in the presence of O2 is a key process for addressing environmental challenges, yet identifying efficient catalysts for this reaction remains challenging. Here, the authors employ a machine learning approach to accelerate the discovery of effective catalysts for direct N₂O decomposition.
Chenxi He, S. Mine, Yuan Jing et al.· Nature Communications· 0 citations
The discovery and development of high-performance catalysts, which is crucial across all catalysis areas, requires advanced technologies and innovative approaches. Recently, machine learning (ML) has shown promise in accelerating this process, but its capability and examples of discovery of truly novel catalysts have remained limited. In this study, we describe an ML approach that goes beyond the traditional element pool, incorporating elements that have not been previously studied, to develop highly efficient catalysts for ethanol synthesis via CO2 hydrogenation. Starting with an initial data set of 58 catalysts (274 data points obtained at reaction temperatures ranging from 240-400 °C), we conducted 24 iterations of a closed-loop discovery system (ML predictions + experimental validation), testing a total of 555 catalysts (2477 data points), and building a large experimental data set. More than 50 catalysts with superior activity were discovered through this data-driven approach. The multielemental Pd(0.8)-Au(0.3)/K(2.5)-Sr(1)-Fe(20)-Zn(4)-Cd(2)-Yb(1)-Re(1)/CeO2(25%)-ZrO2 catalyst, where the numbers in parentheses represent weight percent (wt %), was identified as the most effective catalyst for ethanol synthesis (ethanol space-time yield: 8.2 mmol gcat-1 h-1 with a CO2 conversion of 57.6% and an ethanol selectivity of 23.2% under reaction conditions of 360 °C, 4 MPa, 12 L gcat-1 h-1, H2/CO2 = 3/1). Comprehensive characterizations, including in situ/operando techniques such as X-ray absorption spectroscopy (XAS), ambient-pressure X-ray photoelectron spectroscopy (AP-XPS), and diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), enable us to highlight the critical roles of each constituting element in improving ethanol synthesis efficiency.
Pengfei Du, Abdellah Ait El Fakir, S. Mine et al.· Journal of the American Chem...· 0 citations
High-entropy alloy (HEA) catalysts provide a vast compositional design space for optimizing the oxygen reduction reaction (ORR), but their complexity presents significant challenges for computational screening. In this work, we develop a machine-learning-assisted framework to screen Pt-skin ORR electrocatalysts supported on IrPdPtRhRu HEA subsurface using *O and *OH binding energies as descriptors of catalytic activity. The model employs a simple multiple linear regression approach combined with zone-based representations of the local atomic environment to capture adsorption trends in Pt-skin architectures. Compositional screening identifies Pt- and Rh-rich subsurface as optimal, with minimal Pd and moderate Ru content which is consistent with prior noble metal HEA screening studies. The predicted adsorption trends follow expectations from d-band theory, providing additional physical validation of the model predictions.
J. M. Aliman, J. L. P. Ebaya, K. J. Bongabong et al.· Journal of Physics, Conferen...· 0 citations
Emerging opportunities in physics-informed machine learning, graph neural networks, generative artificial intelligence, active learning, and autonomous closed-loop DFT-ML-MKM workflows are discussed as promising directions for accelerating the discovery of next-generation electrocatalysts with enhanced activity, selectivity, and long-term stability.
Swetarekha Ram, Shalini Tomar, S. Bhattacharjee· Chemical Communications· 0 citations
A comprehensive account of ML applications in NRR, covering curated experimental databases, feature engineering based on atomic, structural, and DFT‐derived descriptors, and ML‐guided insights into single‐atom, dual‐atom, alloy, oxide, nitride, and defect‐engineered catalysts are presented.
