DFT Variants for Mixed-Metal Oxides. Benchmarks Using Multi-Center Cluster Models.

Mixed-metal oxides, e.g., V-Mo and Bi-Mo, are promising selective oxidation catalysts. Yet, their intricate chemical composition and electronic structure often confound DFT methods. This study addresses problems arising from the simultaneous presence of two kinds of transition metals, by probing eight functionals-five hybrid functionals (MN15, M06, PBE0-D3, B3LYP-D3, and TPSSh-D3), the meta-GGA functional M06-L-D3, the range-separated functional ωB97XD, and the GGA functional PBE-D3. We examine the ability of these functionals to localize reducing electrons, and to reproduce reaction energies from CCSD(T) calculations. Accordingly, hybrid functionals containing 20% or more exact exchange perform considerably better in both tests. The B3LYP-D3 approach exhibits the lowest overall mean absolute deviation of reaction energies (OMAD), 21 kJ mol-1, and gave electron distributions as expected from the local lattice structure according to the pseudo-Jahn-Teller effect. MN15 and PBE0-D3 reproduced the electron distributions, but bore slightly higher OMAD values, at 31 and 32 kJ mol-1. Despite acceptable OMAD values, M06 (28 kJ mol-1) and TPSSh (23 kJ mol-1) in some cases did not yield the expected electron distributions. The range-separated functional ωB97XD experienced the opposite problem, yielding correct electron distributions but a poor OMAD of 41 kJ mol-1. M06-L-D3 and PBE-D3 performed relatively poorly, regarding the electron distribution and the OMAD values, 39 and 65 kJ mol-1, respectively.

Transformations of Organic Molecules over Metal Surfaces: Insights from Computational Catalysis.

Much-needed progress in catalytic science, in particular regarding heterogeneous catalysis, is associated with the transition from largely empirical research to rational design of new and improved catalysts and catalytic processes. To achieve this goal, fundamental atomic-scale understanding of catalytic processes is required, which can be achieved with the help of theoretical modeling, in particular, using methods based on quantum chemical calculations. In this review we illustrate the current progress by discussing examples from the authors' work in which complex reaction networks involving organic molecules on transition-metal surfaces have been studied using density functional theory. We review some of the success stories where theory helped to interpret experimental observations and provided atomistic insights into the mechanisms, which were not definitively known before. In other cases, partial disagreement between theoretical results and existing experimental evidence calls for further reconciliation studies.

C-O cleavage of aromatic oxygenates over ruthenium catalysts. A computational study of reactions at step sites.

We studied the C-O cleavage of phenolate and catecholate at step sites of a Ru catalyst using periodic DFT methods at the GGA level. Both C-O scission steps are associated with activation barriers of about 75 kJ mol(-1), hence are significantly more facile than the analogous reactions on Ru terraces. With these computational results, we offer an interpretation of recent experiments on the hydrodeoxygenation of guaiacol (2-methoxyphenol) over Ru/C. We hypothesize that the experimentally observed dependency of the product selectivity on the H2 pressure is related to the availability of step sites on a Ru catalyst.

Assessment of hybrid density functionals for the adsorption of carbon monoxide on platinum model clusters.

We examined computationally the adsorption of CO on various sites of (111) facets of the model clusters Pt79 and Pt225 with the semilocal exchange-correlation functionals PBE, TPSS, and M06L as well as their corresponding hybrid DFT variants PBE0, TPSSh, and M06. The adsorption of CO molecules on Pt(111) is a well-known challenge for the Kohn-Sham DFT approach because one has to treat adequately the electronic structure of the metallic moiety and simultaneously control the self-interaction in the adsorbate. Indeed, in the context of the so-called CO puzzle, hybrid DFT methods do not appear to be beneficial.

Predicting adsorption enthalpies on silicalite and HZSM-5: A benchmark study on DFT strategies addressing dispersion interactions.

We evaluated the accuracy of periodic density functional calculations for adsorption enthalpies of water, alkanes, and alcohols in silicalite and HZSM-5 zeolites using a gradient-corrected density functional with empirical dispersion corrections (PBE-D) as well as a nonlocal correlation functional (vdW-DF2). Results of both approaches agree in acceptable fashion with experimental adsorption energies of alcohols in silicalite, but the adsorption energies for n-alkanes in both zeolite models are overestimated, by 21-46 kJ mol(-1). For PBE-D calculations, the adsorption of alkanes is exclusively determined by the empirical dispersion term, while the generalized gradient approximation-DFT part is purely repulsive, preventing the molecule to come too close to the zeolite walls. The vdW-DF2 results are comparable to those of PBE-D calculations, but the latter values are slightly closer to the experiment in most cases. Thus, both computational approaches are unable to reproduce available experimental adsorption energies of alkanes in silicalite and HZSM-5 zeolite with chemical accuracy.

Does the preferred mechanism of a catalytic transformation depend on the density functional? Ethylene hydrosilylation by a metal complex as a case study.

For an extended set of density functionals (BP86, BLYP, B3LYP, B3PW91, PBE, PBE0, mPWPW, MPW1K, M06-L, M06, MPW3LYP, TPSS) we explored the key steps of four mechanisms of ethylene hydrosilylation (Glaser-Tilley, Chalk-Harrod, modified Chalk-Harrod, and σ-bond metathesis) by a Rh(I) catalyst, previously studied at the B3LYP level. The Chalk-Harrod and the σ-bond metathesis mechanisms were determined to be preferred for all these functionals. The preference among these two mechanisms and the corresponding highest relative barriers (6.6-11.8 kcal·mol(-1)) depend on the functional. To a certain extent, the differences in the description of the reaction can be traced back to the correlation part of the functionals. For the most notable functional-dependent barrier, similar values were calculated when the LYP correlation functional and the functionals M06-L and M06 were employed, but distinctively different values resulted from the functionals PBE, PW91, and TPSS.

Metal-free polymerization of phenylsilane: tris(pentafluorophenyl)borane-catalyzed synthesis of branched polysilanes at elevated temperatures.

The strong organoborane Lewis acid B(C6F5)3 catalyzes the polymerization of phenylsilane at elevated temperatures forming benzene and SiH4 as side-products. The resulting polymer is a branched polysilane with an irregular substitution pattern, as revealed by 2D NMR spectroscopy. Having explored the mechanism of this novel metal-free polymerization by computational chemistry methods at the DFT level, we have suggested that unusual cationic active species, namely monomer-stabilized silyl cations, propagate the polymerization. Hydride abstraction of SiH3 moiety by the catalyst in the initiation step was found to be kinetically preferred by around 9 kcal mol(-1) over activation by coordination of the monomer at the aromatic ring. The formation of linear Si-Si bonds during propagation was calculated to be less favorable than branching and ligand scrambling, which accounts for the branched and highly substituted form of the polymer that was obtained. This novel type of polymerization bears the potential for further optimization with respect to degree of polymerization and structure control for both primary as well as secondary silanes, which can be polymerized by sterically less hindered boranes.

Activation of hydrogen peroxide by ionic liquids: mechanistic studies and application in the epoxidation of olefins.

Imidazolium-based ionic liquids that contain perrhenate anions are very efficient reaction media for the epoxidation of olefins with H2O2 as an oxidant, thus affording cyclooctene in almost quantitative yields. The mechanism of this reaction does not follow the usual pathway through peroxo complexes, as is the case with long-known molecular transition-metal catalysts. By using in situ Raman, FTIR, and NMR spectroscopy and DFT calculations, we have shown that the formation of hydrogen bonds between the oxidant and perrhenate activates the oxidant, thereby leading to the transfer of an oxygen atom onto the olefin demonstrating the special features of an ionic liquid as a reaction environment. The influence of the imidazolium cation and the oxidant (aqueous H2O2, urea hydrogen peroxide, and tert-butyl hydrogen peroxide) on the efficiency of the epoxidation of cis-cyclooctene were examined. Other olefinic substrates were also used in this study and they exhibited good yields of the corresponding epoxides. This report shows the potential of using simple complexes or salts for the activation of hydrogen peroxide, owing to the interactions between the solvent medium and the active complex.

Size dependence of the adsorption energy of CO on metal nanoparticles: a DFT search for the minimum value.

With a density functional theory method, we studied computationally the size dependence of adsorption properties of metal nanoparticles for CO as a probe on Pd(n) clusters with n = 13-116 atoms. For large particles, the values slowly decrease with cluster size from the asymptotic value for an (ideal) infinite surface. For clusters of 13-25 atoms, starting well above the asymptotic value, the adsorption energies drop quite steeply with increasing cluster size. These opposite trends meet in an intermediate size range, for clusters of 30-50 atoms, yielding the lowest adsorption energies. These computational results help to resolve a controversy on the size-dependent behavior of adsorption energies of metal nanoparticles.

Lanthanum modulated reaction pacemakers on a single catalytic nanoparticle.

Promoters are important in catalysis, but the atomistic details of their function and particularly their role in reaction instabilities such as kinetic phase transitions and oscillations are often unknown. Employing hydrogen oxidation as probe reaction, a Rh nanotip for mimicking a single Rh nanoparticle and field electron microscopy for in situ monitoring, we demonstrate a La-mediated local catalytic effect. The oscillatory mode of the reaction provides a tool for studying the interplay between different types of reaction pacemakers, i.e., specific local surface atomic configurations that initiate kinetic transitions. The presence of La shifts the bistable reaction states, changes the oscillation pattern and deactivates one of two pacemaker types for the La-free surface. The observed effects originate from the La-enhanced oxygen activation on the catalyst. The experimental observations are corroborated by micro-kinetic model simulations comprising a system of 25 coupled oscillators.

Kinetic and Computational Studies of CO Oxidation and PROX on Cu/CeO2 Nanospheres.

As supported CuO is well-known for low temperature activity, CuO/CeO2 nanosphere catalysts were synthesized and tested for CO oxidation and preferential oxidation of CO (PROX) in excess H2. For the first reaction, ignition was observed at 95 °C, whereas selective PROX occurred in a temperature window from 50 to 100 °C. The catalytic performance was independent of the initial oxidation state of the catalyst (CuO vs. Cu0), suggesting that the same active phase is formed under reaction conditions. Density functional modeling was applied to elucidate the intermediate steps of CO oxidation, as well as those of the comparably less feasible H2 transformation. In the simulations, various Cu and vacancy sites were probed as reactive centers enabling specific pathways. Supplementary Information: The online version contains supplementary material available at 10.1007/s11244-023-01848-x.

Thermal decomposition of branched silanes: a computational study on mechanisms.

The initial steps of the thermal decomposition of silanes in the gas phase were examined by DFT-B3LYP calculations, with particular attention being paid to the way in which the reactivity pattern changes with the degree of branching of the silane. Besides the established pathways-1,2-hydrogen shift, H(2) elimination, and homolytic dissociation-1,3-hydrogen shift was also explored as an initial reaction step which leads to disilene structures. Subsequent silylene insertion and initial steps of radical chain reactions were also studied. To estimate the energetic changes with temperature, various reaction free energies and the corresponding activation free energies up to 650 °C were calculated. Accordingly, the leading reaction channel at room temperature is 1,2-hydrogen shift with subsequent silylene insertion; for higher degrees of branching, competing pathways (homolytic dissociation, 1,3-hydrogen shift, and radical polymerization) gain in relative importance. At high temperatures, the rate-determining step changes to homolytic dissociation, and thereby the apparent rates of decomposition become dependent on the degree of branching.

Scalable properties of metal clusters: a comparative study of modern exchange-correlation functionals.

The performance of eight generalized gradient approximation exchange-correlation (xc) functionals is assessed by a series of scalar relativistic all-electron calculations on octahedral palladium model clusters Pd(n) with n = 13, 19, 38, 55, 79, 147 and the analogous clusters Au(n) (for n up through 79). For these model systems, we determined the cohesive energies and average bond lengths of the optimized octahedral structures. We extrapolate these values to the bulk limits and compare with the corresponding experimental values. While the well-established functionals BP, PBE, and PW91 are the most accurate at predicting energies, the more recent forms PBEsol, VMTsol, and VT{84}sol significantly improve the accuracy of geometries. The observed trends are largely similar for both Pd and Au. In the same spirit, we also studied the scalability of the ionization potentials and electron affinities of the Pd clusters, and extrapolated those quantities to estimates of the work function. Overall, the xc functionals can be classified into four distinct groups according to the accuracy of the computed parameters. These results allow a judicious selection of xc approximations for treating transition metal clusters.

Active sites and deactivation of room temperature CO oxidation on Co3O4catalysts: combined experimental and computational investigations.

Co3O4is a well-known low temperature CO oxidation catalyst, but it often suffers from deactivation. We have thus examined room temperature (RT) CO oxidation on Co3O4catalysts by operando DSC, TGA and MS measurements, as well as by pulsed chemisorption to differentiate the contributions of CO adsorption and reaction to CO2. Catalysts pretreated in oxygen at 400 °C are most active, with the initial interaction of CO and Co3O4being strongly exothermic and with maximum amounts of CO adsorption and reaction. The initially high RT activity then levels-off, suggesting that the oxidative pretreatment creates an oxygen-rich reactive Co3O4surface that upon reaction onset loses its most active oxygen. This specific active oxygen is not reestablished by gas phase O2during the RT reaction. When the reaction temperature is increased to 150 °C, full conversion can be maintained for 100 h, and even after cooling back to RT. Apparently, deactivating species are avoided this way, whereas exposing the active surface even briefly to pure CO leads to immediate deactivation. Computational modeling using DFT helped to identify the CO adsorption sites, determine oxygen vacancy formation energies and the origin of deactivation. A new species of CO bonded to oxygen vacancies at RT was identified, which may block a vacancy site from further reaction unless CO is removed at higher temperature. The interaction between oxygen vacancies was found to be small, so that in the active state several lattice oxygen species are available for reaction in parallel.

Periodic structural changes in Pd nanoparticles during oscillatory CO oxidation reaction.

Nanoparticle (NP) catalysts are ubiquitous in energy systems, chemical production, and reducing the environmental impact of many industrial processes. Under reactive environments, the availability of catalytically active sites on the NP surface is determined by its dynamic structure. However, atomic-scale insights into how a NP surface reconstructs under reaction conditions and the impact of the reconstruction on catalytic activity are still lacking. Using operando transmission electron microscopy, we show that Pd NPs exhibit periodic round-to-flat transitions altering their facets during CO oxidation reaction at atmospheric pressure and elevated temperatures. This restructuring causes spontaneous oscillations in the conversion of CO to CO2 under constant reaction conditions. Our study reveals that the oscillatory behavior stems from the CO-adsorption-mediated periodic restructuring of the nanocatalysts between high-index-faceted round and low-index-faceted flat shapes. These atomic-scale insights into the dynamic surface properties of NPs under reactive conditions play an important role in the design of high-performance catalysts.

Mo2TiC2 MXene-Supported Ru Clusters for Efficient Photothermal Reverse Water-Gas Shift.

Driving metal-cluster-catalyzed high-temperature chemical reactions by sunlight holds promise for the development of negative-carbon-footprint industrial catalysis, which has yet often been hindered by the poor ability of metal clusters to harvest and utilize the full spectrum of solar energy. Here, we report the preparation of Mo2TiC2 MXene-supported Ru clusters (Ru/Mo2TiC2) with pronounced broadband sunlight absorption ability and high sintering resistance. Under illumination of focused sunlight, Ru/Mo2TiC2 can catalyze the reverse water-gas shift (RWGS) reaction to produce carbon monoxide from the greenhouse gas carbon dioxide and renewable hydrogen with enhanced activity, selectivity, and stability compared to their nanoparticle counterparts. Notably, the CO production rate of MXene-supported Ru clusters reached 4.0 mol·gRu-1·h-1, which is among the best reported so far for photothermal RWGS catalysts. Detailed studies suggest that the production of methane is kinetically inhibited by the rapid desorption of CO from the surface of the Ru clusters.

The origin of the particle-size-dependent selectivity in 1-butene isomerization and hydrogenation on Pd/Al2O3 catalysts.

The selectivity of 1-butene hydrogenation/isomerization on Pd catalysts is known to be particle size dependent. Here we show that combining well-defined model catalysts, atmospheric pressure reaction kinetics, DFT calculations and microkinetic modeling enables to rationalize the particle size effect based on the abundance and the specific properties of the contributing surface facets.