Machine Learning Interatomic Potentials (MLIPs) promise to transform computational catalysis by delivering near-density functional theory (DFT) accuracy at a fraction of the computational cost. Here, we evaluate the Universal Machine Learning Potential for Atoms (UMA) on two data sets of transition-metal complexes. UMA enables high-throughput evaluations in seconds per structure on consumer-grade GPUs. Analysis of per-ligand Spearman rank correlations (ρ > 0.6, p < 0.05) reveals variability in ranking reliability that is not captured by aggregate metrics such as R² or RMSE. However, these inaccuracies are shown to mainly occur in the near-DFT accuracy regime where these complexes are practically indistinguishable. For square-planar Ni complexes, reliable rankings are obtained for 84% of ligands in rigid Ni–Cl₂ complexes and drop to 53% for flexible asymmetric coordination environments, particularly only when conformers differ by <2 kJ/mol. Data set 2 shows a similar trend, with 61% and 44% reliability for Ru(II) and Mn(I) complexes, respectively, and, as expected, challenges for fluxional systems with small (<5 kJ/mol) relative energy gaps. These findings highlight the promise of MLIPs for both rigid, well-defined systems and highly flexible or fluxional catalysts, while underscoring the need to combine the speed of ML with validation and domain expertise to ensure robust and meaningful chemical insights.

Traditional heterogeneous catalysis is constrained by kinetic and thermodynamic limits, such as the Sabatier principle and reaction equilibrium. Dynamic and resonant catalysts hold promise to overcome these limitations by actively oscillating a catalyst’s physical or electronic structure at the time scale of the catalytic cycle, allowing programmable control over reaction pathways, and leading to improved rate and selectivity. External stimuli such as temperature swing, mechanical strain, electric charge, and light can perturb catalyst surfaces in different ways, altering adsorbate coverage, binding energies, and transition states beyond what steady-state catalysis allows. This work surveys the current state of dynamic catalysis, introduces the concept of “stimulando” characterization for observing transient dynamics, and outlines key modeling, mechanistic, and benchmarking strategies to advance the field toward improved chemical transformation.

Capturing the dynamic behavior of active sites on complex, amorphous supports is a significant challenge in modeling single-site catalysts, particularly in surface organometallic catalysts. These systems are characterized by a well-defined chemical bonding pattern that coexists with the fluxionality of ancillary ligands and the inherent complexity of the support. Here, we present a conceptual workflow that integrates reactive molecular dynamics with advanced graph theory-based analysis to systematically explore the configurational space of supported catalysts. First, we used enhanced molecular dynamics to overcome local energy barriers and generate a diverse ensemble of structures. Then, we applied graph-based algorithms to distinguish truly distinct isomers from mere conformers and rotamers. Applying this approach to the model system of 1,1′-bis(n-butyl-cyclopentadienyl) zirconium dihydride on a dehydrated amorphous silica model, our method reveals the significant role of local silica strain in shaping the ensemble of active site configurations: catalysts grafted on silanol groups with strained confinement exhibit a diverse array of reaction pathways and significant energy stabilization, whereas less-strained environments yield a more restricted set of accessible configurations. This work demonstrates that combining molecular dynamics with graph theory provides an intuitive framework for unraveling the complex, fluxional behavior of supported catalysts.

Transesterification reactions are fundamental transformations in organic chemistry, yet performing them in aqueous media is challenging because of the competing hydrolysis reaction. In this study, we describe a mutant of alcohol oxidase from Phanerochaete chrysosporium (PcAOx-VPN) that also exhibits transesterification activity. Moreover, PcAOx-VPN displays no detectable hydrolytic activity, owing to its hydrophobic active site, which effectively excludes water. These characteristics make PcAOx-VPN a promising catalyst for transesterification reactions in aqueous media, a context that is typically compromised by competing hydrolysis.

We report here the upcycling of PET (polyethylene terephthalate) waste via semihydrogenation to make ethyl 4-(hydroxymethyl)benzoate. The reaction is catalyzed by a ruthenium pincer catalyst at 80 °C in bioderived solvents – a combination of 2-methyl THF and ethanol. A detailed mechanistic investigation through organometallic and kinetic studies, as well as chemical exchange saturation transfer (CEST) NMR spectroscopy, provides insights into the nature of active species and factors that promote and inhibit the catalytic hydrogenation of PET. Using this mechanistic knowledge, a record high turnover number of >30 000 was achieved for the hydrogenative depolymerization of end-of-life PET waste (e.g., bottles and textiles). The semihydrogenation product, ethyl 4-(hydroxymethyl)benzoate, was utilized to make precursors of various known pharmaceutical drugs, an agrochemical, as well as a new and recyclable polyester. A cradle-to-gate life cycle assessment demonstrated that using PET waste as a feedstock for EHMB production significantly reduces the environmental footprint compared to the conventional route from p-toluic acid.

Understanding how surface species evolve under reaction conditions is essential for improving catalyst design for efficient CO2 hydrogenation. This work combines systematic DFT calculations with grand canonical sampling to investigate the stability and reactivity of Ga–H species on β-Ga2O3 across a range of reaction conditions. Initial DFT studies reveal that when Ga–H species are present, they facilitate formate formation via a low-barrier pathway, largely independent of the surface termination or hydrogen site. However, grand canonical sampling shows that under a broad range of reaction conditions─especially at high oxygen chemical potentials associated with high water content─Ga–H species are thermodynamically inaccessible. Furthermore, adsorbed water molecules can block reactive sites, inhibiting CO2 activation even when hydrides are present. These findings suggest that the lack of accessible hydride species, rather than their intrinsic reactivity, could contribute to reduced catalytic performance of β-Ga2O3 under more oxidizing, high-conversion conditions.

Many catalytic reactions suffer from product inhibition, which especially hard to control in homogeneous hydrogenation due to the scaling relation between the inhibited and active states of the catalyst. We recently reported one such pathway in Mn(I) hydrogenation and demonstrated that addition of alkoxide bases could affect the thermodynamic favorability of this reaction and selectively suppress the product inhibition. Since external reaction promotors are formally not involved in reaction thermodynamics, we set to investigate the explicit molecular interactions behind these apparently environmental effects. Herein, we reveal that the thermodynamic landscape of the inhibitory process exhibits a non-monotonic dependence on the base concentration. This study related this phenomenon to the presence of two dominant mechanisms operating at different base concentrations. Specifically, the base additives can enhance the ionic strength and lower the free energy of the inhibited state at low promotor concentration. At high base concentrations this study suggested the formation of highly labile alcohol-alkoxide clusters which stabilize the free alcohol and make its addition to the catalyst unfavourable, thereby suppressing the inhibition. While relatively weak, such noncovalent interactions between reactants and reaction environment can cause substantial perturbations to the free energy of catalytic process, ultimately deciding its fate.

Chemical recycling of polyolefins represented by polyethylene (PE) and polypropylene (PP) via catalytic cracking has emerged as a promising strategy for converting waste plastics into valuable hydrocarbons. In this study, we investigated the selective hydrocracking of PP into light alkanes (C₁–C₅) using zeolite catalysts at 280 °C under 1 MPa H₂. An HMFI zeolite with high Al content exhibited the best catalytic performance among various zeolite catalysts tested. In situ DRIFTS comparing bare HMFI and externally-silylated HMFI suggested that the external surface Brønsted acid sites serve as the active sites for the cracking of PP. Combination of in situ DRIFTS and UV–vis spectroscopy analyses identified the formation and consumption of oligomeric species as a reaction intermediate during reaction. Density functional theory (DFT) calculations suggested that a route in which the carbocation and alkoxide intermediates generated by hydrocracking of PP undergo low-energy barrier transformations into gaseous products such as C₃ and C₄ hydrocarbons. This study advances the development of sustainable polyolefin recycling technologies.

Plastic waste is a major environmental issue; converting it directly into valuable chemicals by using catalysts is a promising alternative to plastic recycling. Here, we report the selective catalytic cracking of polypropylene (PP), a typical commodity plastic, to high-value light olefins (C2–C5), below pyrolytic temperature (290 °C) and without external hydrogen supply, by using zeolite catalysts. Among the H+-form zeolites with different structures, HMFI showed the highest yields of light hydrocarbons (C2–C5), of which light olefins (C2–C5) were the major products. The HMFI-catalyzed PP conversion was applicable to the upcycling of a model PP waste, resulting in a 61.9% light hydrocarbon yield. The results of catalytic and in situ IR experiments using model HMFI catalysts with a small amount of external Brønsted acid sites suggested that the Brønsted acid sites on the external surface of HMFI are indispensable for the PP conversion and are posited to be the active sites for the cracking of PP into short-chain (oligomeric) hydrocarbon species as intermediate products. Density functional theory analyses were conducted to determine plausible reaction pathways by adopting 2,4-dimethylheptene as the shortest unit of the oligomeric species. The obtained results show that the β-scission of 2,4-dimethylheptene by Brønsted acid sites yields isobutene and propylene (or a propyl alkoxide group) via carbocation intermediates with an activation energy below 118 kJ mol–1. Operando UV–vis and IR experiments under the reaction conditions, combined with ex situ 1H NMR and 13C NMR analyses of the spent catalyst, show that some of the olefins are further converted to light or heavy aromatics (coke deposit), probably via carbenium ion species.

Sulfated zirconium oxide (SZO) catalyzes the hydrogenolysis of isotactic polypropylene (iPP, Mw=13.3 kDa, Đ=2.4, <mmmm>=94 %) or high-density polyethylene (HDPE, Mn=2.5 kDa, Đ=3.6) to branched alkane products. We propose that this reactivity is driven by the pyrosulfate sites SZO, which open under mild conditions to transiently form adsorbed SO3 and sulfate groups. This adsorbed SO3 is a very strong Lewis acid that binds 15N-pyridine or triethylphosphineoxide (TEPO) (ΔEads>−39 kcal mol−1), reacts with Ph3CH to form Ph3C+, and mediates H/D exchange in dihydroanthracene-d4. DFT studies show that pyrosulfate sites open with a modest 26.1 kcal mol−1 barrier to form the adsorbed SO3 and sulfate in the presence of a tetramer of propylene. Hydride abstraction from the tertiary C−H in this model is exothermic and subsequent β-scission forms cleaved products. Analysis of the energetics provided here brackets the hydride ion affinity (HIA) of the adsorbed SO3 between 226.2 to 237.9 kcal mol−1, among largest values reported for a formally neutral Lewis acid. This study explains how SZO, a classic heterogeneous catalyst, can form carbocations by a redox neutral hydride abstraction reaction by very strong Lewis sites.

Simulation and systematic analysis of the surfaces of amorphous materials is a challenge for computational chemistry. For example, silica has found widespread industrial use as an adsorbent and catalyst support but available models for use with periodic DFT are limited in variety and representativeness of realistic materials. Herein we present a generic approach for the systematic construction of ensembles of amorphous materials surface models with varied roughness and termination characteristics. The power of the approach is shown with silica as the representative example. By combining MD simulations and Fourier-series-based randomization, bulk amorphous silica was modeled and cleaved to produce surfaces with systematically varied roughness and surface saturation. An automated saturation procedure resulted in surface models with silanol densities typical of high-temperature activation protocols in the range 0.35–2.00 OH nm−2, in excellent agreement with the experimental data on surface chemistry of dehydroxylated silica materials.

Computational exploration of chemical space is a powerful tool for designing organometallic homogeneous catalysts. While catalytic properties depend on ligand properties and spatial arrangement, the role of stereoisomerism in defining catalyst selectivity and reactivity has only been elucidated sporadically, leaving gaps in virtual screening workflows. This study investigates the necessity of exploring ligand configurations for virtual high-throughput (HT) screening of octahedral transition metal complexes. Using automated workflows, ligand configuration ensembles were generated for bisphosphine ligands with Ir(III), Ru(II), and Mn(I) metal centers. DFT calculations revealed distinct preferences for Ir(III) configurations, whereas Mn(I)- and Ru(II)-complexes displayed significant fluxionality, with multiple configurations within a 10 kJ mol−1 energy range. Linear regression analyses showed that global descriptors, such as bite angle and HOMO–LUMO gap, are transferable across configurations and metal centers, while local steric descriptors lacked such transferability. Machine learning (ML) models successfully classified ligand configurations (balanced accuracy >0.8) but struggled to predict stability across metal centers, especially for Mn(I) and Ru(II). Thus, improved descriptors of the first coordination sphere to capture fluxionality and stability more effectively can improve ML models. Overall, this study underscores the limitations of ignoring stereoisomerism in virtual HT screening, which may lead to incomplete exploration of chemical space and underrepresentation of key catalyst features. Until dynamic digital representations are developed, exhaustive stereoisomerism exploration should be implemented for screening workflows.

Transition-metal complexes serve as highly enantioselective homogeneous catalysts for various transformations, making them valuable in the pharmaceutical industry. Data-driven prediction models can accelerate high-throughput catalyst design but require computer-readable representations that account for conformational flexibility. This is typically achieved through high-level conformer searches, followed by DFT optimization of the transition-metal complexes. However, conformer selection remains reliant on human assumptions, with no cost-efficient and generalizable workflow available. To address this, we introduce an automated approach to correlate CREST(GFN2-xTB//GFN-FF)-generated conformer ensembles with their DFT-optimized counterparts for systematic conformer selection. We analyzed 24 precatalyst structures, performing CREST conformer searches, followed by full DFT optimization. Three filtering methods were evaluated: (i) geometric ligand descriptors, (ii) PCA-based selection, and (iii) DBSCAN clustering using RMSD and energy. The proposed methods were validated on Rh-based catalysts featuring bisphosphine ligands, which are widely employed in hydrogenation reactions. To assess general applicability, both the precatalyst and its corresponding acrylate-bound complex were analyzed. Our results confirm that CREST overestimates ligand flexibility, and energy-based filtering is ineffective. PCA-based selection failed to distinguish conformers by DFT energy, while RMSD-based filtering improved selection but lacked tunability. DBSCAN clustering provided the most effective approach, eliminating redundancies while preserving key configurations. This method remained robust across data sets and is computationally efficient without requiring molecular descriptor calculations. These findings highlight the limitations of energy-based filtering and the advantages of structure-based approaches for conformer selection. While DBSCAN clustering is a practical solution, its parameters remain system-dependent. For high-accuracy applications, refined energy calculations may be necessary; however, DBSCAN-based clustering offers a computationally accessible strategy for rapid catalyst representations involving conformational flexibility.

This chapter provides a brief introduction to computational chemistry in the context of zeolite research, emphasizing the capabilities and limitations of modern theoretical models for investigating their reactivity and chemical properties under operando conditions. A brief overview of the computational chemistry toolbox is given, followed by a discussion of state-of-the-art applications in zeolite chemistry and catalysis. This chapter also highlights the increasing impact of data-driven techniques, such as machine-learning potentials, in advancing computational methods in zeolite studies.

Catalysis Science & Technology, Evgeny Pidko and Núria López would like to acknowledge Weixue Li for their contributions to the Digital Catalysis themed collection as a Guest Editor.

Significant efforts have been dedicated to the direct syngas conversion into ethanol, however, achieving a high ethanol yield remains a formidable task. In this study, we present the direct syngas-to-ethanol conversion over Li-promoted RhOx/MgO catalyst (RhOx/Li2O/MgO). The ethanol space-time yield (EtOH STY) and selectivity reached 12.2 mmol gcat–1 h–1 and 20%, respectively, at a 35% CO conversion over the RhOx/Li2O/MgO catalyst. The RhOx/Li2O/MgO catalyst demonstrated superior performance in terms of both ethanol selectivity and STY compared to Rh/Li2O catalysts on other support materials and Rh/MgO catalysts promoted with other alkali metals. In situ/operando spectroscopic techniques, combined with other characterisations and theoretical calculations, have elucidated the interactions between Li2O and Rh on the MgO surface. These interactions promote the formation of new active sites and weaken CO adsorption on the Rh surface, thereby enhancing ethanol production. This work provides a promising strategy for improving ethanol yield in syngas conversion processes.

Electrocatalytic reduction of organic halides and subsequent carboxylation are promising methods for the valorization of CO2 as a C1 source in synthetic organic chemistry. The reaction mechanism underlying the selectivity and reduction mechanism of benzyl halides is highly dependent on the nature of the electrode material as well as the processes, composition, and structure of the liquid phase at the electrode–solution interface. Herein, we present a computational study on the influence of the electric double layer (EDL) on the activation of benzyl halides at different applied potentials over the Au (111) cathode. Using a multiscale modeling approach, we demonstrate that, under realistic electrocatalytic conditions, the formation of a dense EDL over the cathode hampers the diffusion of benzyl halides toward the electrode surface. A combination of classical molecular dynamics simulations and density functional theory calculations reveals the most favorable benzyl halide electro-carboxylation pathway over the EDL that does not require direct substrate adsorption to the cathode surface. The dense EDL promotes the dissociative reduction of the benzyl halides via the outer-sphere electron transfer from the cathode surface to the electrolyte. Such a reduction mechanism results in a benzyl radical intermediate, which is then converted to benzyl anions in the EDL via an additional electron transfer step.

Supported rhenium (Re) catalysts are emerging as promising candidates for hydrogenation reactions, which are crucial in industrial processes such as biomass valorization, CO2 reduction, and petroleum refining. However, despite their broad application, the structural and mechanistic understanding of these systems remains limited. The strong oxophilic nature of Re, combined with its ability to adopt multiple oxidation states, complicates the characterization of the active species even with advanced experimental techniques. In this study, we employ density functional theory calculations, alongside ab initio thermodynamic analysis, to systematically explore the structural and electronic properties of single-atom Re catalysts on a TiO2 surface, providing insights that could inform the rational catalyst design. Our calculations reveal the formation of stable Re polyhydrides on the surface under hydrogen-rich conditions. Notably, even in highly reducing environments, Re species with low formal oxidation states are thermodynamically unfavorable. The stable Re species identified on TiO2 surfaces demonstrate high reactivity toward CO2 hydrogenation. The electronic properties and computed X-ray photoelectron spectroscopy (XPS) signatures of the feasible surface species are highly influenced by the ligand environment. This work highlights the limitation of routine interpretation of experimental XPS characterization data in terms of the formal oxidation state.

Computational chemistry pipelines typically commence with geometry generation, well-established for organic compounds but presenting a considerable challenge for transition metal complexes. This paper introduces MACE, an automated computational workflow for converting chemist SMILES/MOL representations of the ligands and the metal center to 3D coordinates for all feasible stereochemical configurations for mononuclear octahedral and square planar complexes directly suitable for quantum chemical computations and implementation in high-throughput computational chemistry workflows. The workflow is validated through a structural screening of a data set of transition metal complexes extracted from the Cambridge Structural Database. To further illustrate the power and capabilities of MACE, we present the results of a model DFT study on the hemilability of pincer ligands in Ru, Fe, and Mn complexes, which highlights the utility of the workflow for both focused mechanistic studies and larger-scale high-throughput pipelines.

The premise of most studies on the homogeneous electrocatalytic CO₂ reduction reaction (CO₂RR) is a good understanding of the reaction mechanisms. Yet, analyzing the reaction intermediates formed at the working electrode is challenging and not always attainable. Here, we present a new, general approach to studying the reaction intermediates applied for CO₂RR catalyzed by a series of cobalt complexes. The cobalt complexes were based on the TPA-ligands (TPA = tris(2-pyridylmethyl)amine) modified by amino groups in the secondary coordination sphere. By combining the electrochemical experiments, electrochemistry-coupled electrospray ionization mass spectrometry, with density functional theory (DFT) calculations, we identify and spectroscopically characterize the key reaction intermediates in the CO₂RR and the competing hydrogen-evolution reaction (HER). Additionally, the experiments revealed the rarely reported in situ changes in the secondary coordination sphere of the cobalt complexes by the CO₂-initiated transformation of the amino substituents to carbamates. This launched an even faster alternative HER pathway. The interplay of three catalytic cycles, as derived from the experiments and supported by the DFT calculations, explains the trends that cobalt complexes exhibit during the CO₂RR and HER. Additionally, this study demonstrates the need for a molecular perspective in the electrocatalytic activation of small molecules efficiently obtained by the EC-ESI-MS technique.