Limits to scaling relations between adsorption energies?

Insights into Electrochemical CO2 Reduction on Metallic and Oxidized Tin Using Grand-Canonical DFT and In Situ ATR-SEIRA Spectroscopy

Electrochemical CO2 reduction (CO2R) to formate is an attractive carbon emissions mitigation strategy due to the existing market and attractive price for formic acid. Tin is an effective electrocatalyst for CO2R to formate, but the underlying reaction mechanism and whether the active phase of tin is metallic or oxidized during reduction is openly debated. In this report, we used grand-canonical density functional theory and attenuated total reflection surface-enhanced infrared absorption spectroscopy to identify differences in the vibrational signatures of surface species during CO2R on fully metallic and oxidized tin surfaces. Our results show that CO2R is feasible on both metallic and oxidized tin. We propose that the key difference between each surface termination is that CO2R catalyzed by metallic tin surfaces is limited by the electrochemical activation of CO2, whereas CO2R catalyzed by oxidized tin surfaces is limited by the slow reductive desorption of formate. While the exact degree of oxidation of tin surfaces during CO2R is unlikely to be either fully metallic or fully oxidized, this study highlights the limiting behavior of these two surfaces and lays out the key features of each that our results predict will promote rapid CO2R catalysis. Additionally, we highlight the power of integrating high-fidelity quantum mechanical modeling and spectroscopic measurements to elucidate intricate electrocatalytic reaction pathways.

Accessing the usefulness of atomic adsorption configurations in predicting the adsorption properties of molecules with machine learning

We present a systematic study into the effect of adding atomic adsorption configurations into the training and validation dataset for a neural network's predictions of the adsorption energies of small molecules on single metal and bimetallic, single crystal surfaces. Specifically, we examine the efficacy of models trained with and without H and X atomic adsorption configurations, where X is C, N, or O, to predict XHn adsorption energies. In addition, we compare our machine learning models to traditional simple scaling relationships. We find that models trained with the atomic adsorption configurations outperform models trained with only molecular adsorption configurations, with as much as a 0.37 eV decrease in the MAE. We find that models trained with the atomic adsorption configurations slightly outperform traditional scaling relationships. In general, these results suggest it may be possible to vastly reduce the number of adsorption configurations one needs for training and validation datasets by supplementing said data with the adsorption configurations of composite atoms or smaller molecular fragments.

Interpretable design of Ir-free trimetallic electrocatalysts for ammonia oxidation with graph neural networks

The electrochemical ammonia oxidation to dinitrogen as a means for energy and environmental applications is a key technology toward the realization of a sustainable nitrogen cycle. The state-of-the-art metal catalysts including Pt and its bimetallics with Ir show promising activity, albeit suffering from high overpotentials for appreciable current densities and the soaring price of precious metals. Herein, the immense design space of ternary Pt alloy nanostructures is explored by graph neural networks trained on ab initio data for concurrently predicting site reactivity, surface stability, and catalyst synthesizability descriptors. Among a few Ir-free candidates that emerge from the active learning workflow, Pt₃Ru-M (M: Fe, Co, or Ni) alloys were successfully synthesized and experimentally verified to be more active toward ammonia oxidation than Pt, Pt₃Ir, and Pt₃Ru. More importantly, feature attribution analyses using the machine-learned representation of site motifs provide fundamental insights into chemical bonding at metal surfaces and shed light on design strategies for high-performance catalytic systems beyond the d-band center metric of binding sites.

Free-atom-like d states beyond the dilute limit of single-atom alloys

Through a data-mining and high-throughput density functional theory approach, we identify a diverse range of metallic compounds that are predicted to have transition metals with "free-atom-like" d states that are highly localized in terms of their energetic distribution. Design principles that favor the formation of localized d states are uncovered, among which we note that site isolation is often necessary but that the dilute limit, as in most single-atom alloys, is not a pre-requisite. Additionally, the majority of localized d state transition metals identified from the computational screening study exhibit partial anionic character due to charge transfer from neighboring metal species. Using CO as a representative probe molecule, we show that localized d states for Rh, Ir, Pd, and Pt tend to reduce the binding strength of CO compared to their pure elemental analogues, whereas this does not occur as consistently for the Cu binding sites. These trends are rationalized through the d-band model, which suggests that the significantly reduced d-band width results in an increased orthogonalization energy penalty upon CO chemisorption. With the multitude of inorganic solids that are predicted to have highly localized d states, the results of the screening study are likely to result in new avenues for heterogeneous catalyst design from an electronic structure perspective.

Chemical reaction networks and opportunities for machine learning

Chemical reaction networks (CRNs), defined by sets of species and possible reactions between them, are widely used to interrogate chemical systems. To capture increasingly complex phenomena, CRNs can be leveraged alongside data-driven methods and machine learning (ML). In this Perspective, we assess the diverse strategies available for CRN construction and analysis in pursuit of a wide range of scientific goals, discuss ML techniques currently being applied to CRNs and outline future CRN-ML approaches, presenting scientific and technical challenges to overcome.

Stick or Spill? Scaling Relationships for the Binding Energies of Adsorbates on Single-Atom Alloy Catalysts

Single-atom alloy catalysts combine catalytically active metal atoms, present as dopants, with the selectivity of coinage metal hosts. Determining whether adsorbates stick at the dopant or spill over onto the host is key to understanding catalytic mechanisms on these materials. Despite a growing body of work, simple descriptors for the prediction of spillover energies (SOEs), i.e., the relative stability of an adsorbate on the dopant versus the host site, are not yet available. Using Density Functional Theory (DFT) calculations on a large set of adsorbates, we identify the dopant charge and the SOE of carbon as suitable descriptors. Combining them into a linear surrogate model, we can reproduce DFT-computed SOEs within 0.06 eV mean absolute error. More importantly, our work provides an intuitive theoretical framework, based on the concepts of electrostatic interactions and covalency, that explains SOE trends and can guide the rational design of future single-atom alloy catalysts.