DFT‐based Machine Learning for Ensemble Effect of Pd@Au Electrocatalysts on CO₂ Reduction Reaction

Abstract

The ensemble effect due to variation of Pd content in Pd−Au alloys have been widely investigated for several important reactions, including CO2 reduction reaction (CO2RR), however, identifying the stable Pd arrangements on the alloyed surface and picking out the active sites are still challenging. Here we use a density functional theory (DFT) based machine‐learning (ML) approach to efficiently find the low‐energy configurations of Pd−Au(111) surface alloys and the potentially active sites for CO2RR, fully covering the Pd content from 0 to 100 %. The ML model is actively learning process to improve the predicting accuracy for the configuration formation energy and to find the stable Pd−Au(111) alloyed surfaces, respectively. The local surface properties of adsorption sites are classified into two classes by the K‐means clustering approach, which are closely related to the Pd content on Au surface. The classification is reflected in the variation of adsorption energy of CO and H: In the low Pd content range (0–60 %) the adsorption energies over the surface alloys can be tuned significantly, and in the medium Pd content (37‐68 %), the catalytic activity of surface alloys for CO2RR can be increased by increase the Pd content and attributed to the meta‐stable active site over the surface. Thus, the active site‐dependent reaction mechanism is elucidated based on the ensemble effect, which provides new physical insights to understand the surface‐related properties of catalysts.

Local descriptors-based machine learning model refined by cluster analysis for accurately predicting adsorption energies on bimetallic alloys

“…[44][45][46] Among the most relevant local chemical descriptors proposed to date, those focused on correlating the nearest neighboring surface atoms to the adsorption site have gained signicant attention. These descriptors include the average of elemental properties, 47 atom-specic ngerprints derived from elemental properties, 48,49 and use ML models based on neural networks, [50][51][52] or graph neural networks. [53][54][55][56] However, these descriptors have difficulties adapting to other adsorbates or ML-based model architectures, hindering their transferability to other systems.…”

A prediction model for CO2/CO adsorption performance on binary alloys based on machine learning

“…28 Liu investigated the adsorption energies on binary alloy surfaces of Pd n Au 16−n alloyed surface with different Pd content (n = 1-16) by ML prediction and concluded that the isolated Pd top sites surrounded by Au atoms are stable adsorption sites. 29 Nayak et al predicted adsorption energies of H, O, N, OH, NO, and CO on fcc(111) surface top sites of 25 different transition metals including Ir, Pt, and Au with an average root-mean-square error (RMSE) of about 0.4 eV by random forest regression. 30 Prediction of adsorption energies on metal and alloy surfaces was also reported using the XGBoost regression, 31 articial neural network algorithm, 32,33 random forest, 34 and other methods.…”

“…the entire topmost layer of the metal was replaced by another metal element, while the actual conguration of alloys in real catalysts could be much more complex. 29 In this study, we directly exploit the adsorption energies of CO 2 and CO on surfaces of a wide range of binary alloys using ML methods without any assumptions of linearity, i.e. we do not assume the adsorption energy on alloys to be a linear combination of adsorption affinity on its two component metal surfaces.…”

“…the entire topmost layer of the metal was replaced by another metal element, while the actual configuration of alloys in real catalysts could be much more complex. 29 …”

Unravelling the Alloy Effect of Porous Binary Pd-Based Nanostructures on Electrocatalytic Oxidation of Carbon Monoxide in Different Electrolytes

“…Pd-based electrocatalysts are well-known for their outstanding CO Oxid activity, due to their ability to enhance the activation/dissociation of reactants (CO + O 2 ) under low applied potential and weaken the adsorption of carbonaceous intermediates; however, the earth-scarcity and high-cost of Pd conspired the commercial utilization. ,− This issue could be vanquished via reducing amount of Pd, by alloying Pd with inexpensive and earth-abundant metals (i.e., Ni, Mn, Cu, Fe, and Co) and/or using supports (i.e., carbon and metal oxide), − which reduce the loading of Pd usage and enhance the CO Oxid activity . Mainly, alloying effect generates a kind of electronic interaction (i.e., synergism and strain), − which endows the activation/dissociation of reactants, eases H 2 O dissociation to release oxygenated species for accelerating the CO Oxid kinetics, and facilitates the desorption of product (i.e., CO 2 ) under low potential . For instance, Pd nanocrystals supported on Co–ZIF-derived hierarchical porous carbon nanosheets (Pd/ZIF-67/C) exhibited electrocatalytic CO Oxid that outperformed commercial Pd/C by 4.0 times in HClO 4 medium, due to the interaction of Pd with ZIF-67/C, porous structure of ZIF-67/C and abundance active sites of Co–N x .…”