Alexander Bagger

Ammonia plays a central role in global fertilizer production and is increasingly recognized as a versatile energy carrier within emerging sustainable energy systems. The La/Co system is an example of an ammonia production catalyst with efficiency greater than that of current commercial catalysts. A spin-mediated promotion mechanism has been proposed as the mechanism behind the catalytic activity. While the resulting catalytic promotion is clearly observed, the change in spin has not yet been directly demonstrated experimentally. In the present study, scanning tunneling microscopy (STM) and scanning tunneling spectroscopy (STS) measurements were performed on a Co(101̅15) crystal and compared with density functional theory (DFT) calculations. The purpose was to obtain an understanding of the properties of the Co(101̅15) crystal surface as a first step to verify the electronic structure changes behind spin promotion. The STM results revealed a stepped surface as expected. A comparison between the STS measurements and DFT calculations indicated that the spin of the Co atoms changes across the surface. This could potentially be caused by adsorbates on the crystal surface or by an interaction between the STM tip and the sample surface.

In aqueous electrocatalytic environments, the protonation of stable *CO and *N2 to *COH/*CHO and *NNH is fundamentally challenged by the competing hydrogen evolution reaction (HER). In this context, direct dissociation of these molecules offers an alternative pathway, forming reactive *C or *N fragments that can be reduced further if surface bonding is not too strong. However, the strong C≡O and N≡N bonds make dissociation challenging under ambient electrochemical conditions, and the extent to which solvent, local field, and electrolyte ions can promote bond dissociation remains unclear. In this work, we quantify these dissociation pathways at a Co(B5) step site using density functional theory, implicit-solvent calculations, and explicit-solvent calculations using enhanced sampling with machine-learned interatomic potential (MLIP). Implicit-solvent calculations reveal only weak sensitivity to dielectric constant, dipole moment, and solvent radius, with barrier variations typically ≤0.1 eV. In contrast, the local H-bonding network in explicit-solvent simulations lowers the dissociation barrier by ∼0.3 eV relative to vacuum by stabilizing the dissociating atoms. Across different solvents and aqueous solutes, the barrier correlates qualitatively with hydrogen-bond donor strength, pKa, and the local H-coordination number near the transition-state region. Interfacial cations (Li+ and Cs+) modify the work function but do not significantly change the dissociation barrier compared with pure water, suggesting that solvent-mediated interactions dominate under the studied conditions.

Electrocatalysis is central to electrified energy conversion technologies, including the production of fuels and chemicals from renewable electricity. In recent years, atomistic modeling based on density functional theory (DFT) has become an indispensable tool for catalysis researchers because it can provide a molecular-level perspective of catalytic reactivity that is otherwise inaccessible through experimental measurements. Computational electrocatalysis model inputs are intended to capture fundamental understanding of the complex interfacial environment of the catalytic active site, but they rely on practical approximations that can affect the reliability of insights gained from DFT. This Perspective outlines a framework for rigorous and reproducible integration of theory and experiment in heterogeneous electrocatalysis research. We discuss state-of-the-art computational methods and sources of error in model development and interpretation. We describe opportunities for validation of theoretical models and experimental interpretations as they pertain to the structure of active sites, reaction energetics, and microkinetic modeling. We provide examples to help computational researchers develop models that move beyond obtaining qualitative agreement with reactivity trends, and toward the development of rigorous mechanistic insights that can be leveraged to predict promising new catalysts or operating conditions. Rigor and reproducibility are enhanced when theory and experiment have multiple nodes of connection, thereby enhancing the longevity and robustness of conclusions and insights about electrocatalytic reactivity.

Electrochemical ethylene oxidation can produce ethylene oxide and ethylene glycol, but selectivity is limited by competition with the oxygen evolution reaction and overoxidation. At the same time, the mechanistic literature is fragmented across oxo-, peroxo-, and hydroxyl-mediated pathways, creating ambiguity and hindering systematic improvement. Using density functional theory (DFT) and scaling relations, we build a unified descriptor-based framework comparing the three pathways across fcc metals, metal(II) oxides, and rutile-structured oxides. The oxygen binding strength is found to control ethylene activation via oxo and peroxo species, while a carbon-based descriptor describes the hydroxyl pathway, and product binding defines an overoxidation limit. The resulting selectivity maps identify the catalyst properties required for each route, rationalize the high performance of individual catalysts such as oxidized Pd, and provide transferable guidelines for designing selective alkene oxidation catalysts.

Nitric acid is essential in fertilizer production, yet its current industrial synthesis via the Haber-Bosch and Ostwald process accounts for 1 % of the chemical sector’s energy consumption and produces significant greenhouse gas emissions. Despite considerable efforts, previous attempts at electrochemical nitrogen oxidation have proven unsuccessful due to competition with the oxygen evolution reaction (OER) and the inherent difficulty of activating the nitrogen (N2) bond under ambient conditions. In this work, we propose a phase transition-based approach inspired by electrochemical ammonia synthesis to circumvent these challenges. The strategy is centered around the formation of a metal nitride by direct nitridation of a suitable metal, followed by an oxidation step employing molecular oxygen (O2). Li and Ca are identified as particularly promising candidates to mediate such a pathway based on reaction thermodynamics, and we show experimentally that Li nitride can indeed be oxidized to yield nitrate in the presence of O2. The proposed concept demonstrates considerable potential for direct N2 oxidation, encouraging further computational and experimental investigation.

Metal–support interactions (MSI) play a critical role in heterogeneous catalysis by influencing catalyst stability and reactivity. A long-term objective is to model catalytic systems in simulated experiments, to obtain atomistic insights into MSI under reaction conditions. Achieving this goal requires extensive modelling of large atomistic systems. Machine-learned force fields (MLFF’s) enable simulations at these scales by providing orders of magnitude speedups compared to density functional theory (DFT). However, effective fine-tuning of MLFFs requires smart sampling of the configurational space of supported nanoparticle systems. Notably, the chemically relevant interfacial region constitutes only a small fraction of the total atoms, highlighting the need for cost-effective training strategies. Here, we demonstrate two such strategies using Co/TiO2 as a model system: cross-polymorph training and reducing support slab layers. The latter lowers the number of electrons in each training structure, significantly reducing computational cost, which scales roughly cubically with electrons. Combining both strategies produced an accurate MLFF capable of revealing a novel minimum-energy configuration for a Co nanoparticle on TiO2 anatase. The presented strategies are applicable to all metal oxide-supported metal nanoparticle systems and enable the development of MLFFs capable of probing MSI at the atomistic scale.

The catalytic performance of the electrochemical CO2 reduction reaction (CO2RR) is highly sensitive to the electrolyte microenvironment, including ion identity, concentration, and buffering chemistry. While such electrolyte effects are well established for metallic catalysts, their mechanistic roles in porous Ni–N–C systems remain less explored. Here, we systematically investigate the influence of cation (Li+, K+, Cs+, and tetramethylammonium, TMA+) and anion (phosphate, sulfate, and bicarbonate) identity and concentration on CO2-to-CO conversion. Contrary to expectations, Cs+ does not outperform K+, whereas TMAHCO3 electrolytes exhibit a pronounced enhancement in CO2RR activity, reaching a CO partial current density of 25 mA cm–2 at –0.7 VRHE in 0.5 M solution. Electrochemical impedance spectroscopy (EIS) and distribution of relaxation times (DRT) analysis reveal that this activity enhancement does not directly correlate with interfacial ion-/charge-transfer or mass transport behaviors. Instead, Fourier-transform infrared (FTIR) spectroscopy shows a significant increase in bulk CO2 availability in TMAHCO3 electrolytes. Overall, these results suggest that electrolyte composition affects both interfacial properties and CO2 availability in the bulk, influencing CO2RR activity within the porous Ni–N–C catalyst system.

Asymmetric catalysis underpins pharmaceutical synthesis but is dominated by homogeneousmolecular catalysts that complicate product separation and recycling. Heterogeneousalternatives, including hybrid systems in which molecular catalysts are immobilised on solidsupports, often suffer from diminished activity and poor stability. Here, we use densityfunctional theory and microkinetic modelling to evaluate intrinsically chiral platinum surfacesas a heterogeneous alternative, taking the hydrogenation of hydroxyacetone to 1,2-propanediol (1,2-PD) as a model reaction. On the mirror-image Pt(643)S and Pt(643)ᴿ surfaces,hydroxyacetone chemisorbs at the undercoordinated chiral kink, and the effective transition-state barriers for the two reaction pathways differ by 0.21 eV, favouring S-1,2-PD on Pt(643)S and R-1,2-PD on Pt(643)ᴿ. By contrast, hydroxyacetone is only physisorbed on the achiral Pt(111) terrace and is hydrogenated to a racemic product with a lower effective barrier of 0.75 eV. Microkinetic modelling that accounts for the realistic coexistence of kink and terrace sites identifies an operational window near 500 K and 10⁻² bar in which strong substrate binding at the chiral kink outcompetes the racemic terrace pathway, yielding enantiomeric excesses approaching 99 %. These results establish chiral Pt(643) surfaces as a viable platform for fully heterogeneous asymmetric hydrogenation and motivate experimental verification.

CO electrolysis with cobalt yielding Fischer–Tropsch type products that are strongly temperature dependent, but weakly cation dependent.

The electrochemical reduction of nitrate (NO₃^-) to ammonia (NH₃) offers a sustainable route for nitrogen cycle remediation and decentralized NH₃ production. In this work, we systematically investigated the impact of electronic structure and wettability in regulating the catalytic performance of molecular catalysts using functionalized iron phthalocyanines (FePc-R, R = NH₂, COOH, CN, and t-Bu) supported on carbon nanotubes. The strongly hydrophilic FePc-NH₂/CNT (electron-donating functional group-containing) catalyst achieved a maximum Faradaic efficiency of 94.1% at -0.6 V_RHE and a partial current density of 83.9 mA cm^-2 toward NH₃ at -0.9 V_RHE. In contrast, strongly hydrophilic FePc-COOH/CNT and weakly hydrophilic FePc-CN/CNT, containing electron-withdrawing functional groups, delivered a lower performance across all potentials. Density functional theory (DFT) calculations revealed that electron-donating functional groups elevate the Fe-center HOMO level, facilitating hydrogenation of NH_x intermediates and enhancing turnover frequency. In situ X-ray absorption spectroscopy (XAS) confirmed that Fe-N₄ coordination in FePc-NH₂/CNT remains stable across all tested potentials, while electron-withdrawing functional group-containing catalysts (FePc-COOH/CNT and FePc-CN/CNT) exhibited Fe-Fe cluster formation at -0.8 and -0.7 V_RHE, respectively. Furthermore, coupled mass transport and reaction modeling indicated that more hydrophilic surfaces reduce the diffusion layer thickness, promoting NO₃^- accessibility and NH₃ formation. Together, these findings decoupled the synergistic role of electronic tuning and wettability control in governing both activity and stability, providing mechanistic design principles for molecular and heterogeneous catalysts in the reduction of electrochemical NO₃^- to NH₃.