Exploring the Impact of Active Site Structure on the Conversion of Methane to Methanol in Cu-Exchanged Zeolites.

In the past, Cu-oxo or -hydroxy clusters hosted in zeolites have been suggested to enable the selective conversion of methane to methanol, but the impact of the active site's stoichiometry and structure on methanol production is still poorly understood. Herein, we apply theoretical modeling in conjunction with experiments to study the impact of these two factors on partial methane oxidation in the Cu-exchanged zeolite SSZ-13. Phase diagrams developed from first-principles suggest that Cu-hydroxy or Cu-oxo dimers are stabilized when O2 or N2O are used to activate the catalyst, respectively. We confirm these predictions experimentally and determine that in a stepwise conversion process, Cu-oxo dimers can convert twice as much methane to methanol compared to Cu-hydroxyl dimers. Our theoretical models rationalize how Cu-di-oxo dimers can convert up to two methane molecules to methanol, while Cu-di-hydroxyl dimers can convert only one methane molecule to methanol per catalytic cycle. These findings imply that in Cu clusters, at least one oxo group or two hydroxyl groups are needed to convert one methane molecule to methanol per cycle. This simple structure-activity relationship allows to intuitively understand the potential of small oxygenated or hydroxylated transition metal clusters to convert methane to methanol.

Machine Learning-Based Prediction of Activation Energies for Chemical Reactions on Metal Surfaces.

In computational surface catalysis, the calculation of activation energies of chemical reactions is expensive, which, in many cases, limits our ability to understand complex reaction networks. Here, we present a universal, machine learning-based approach for the prediction of activation energies for reactions of C-, O-, and H-containing molecules on transition metal surfaces. We rely on generalized Bronsted-Evans-Polanyi relationships in combination with machine learning-based multiparameter regression techniques to train our model for reactions included in the University of Arizona Reaction database. In our best approach, we find a mean absolute error for activation energies within our test set of 0.14 eV if the reaction energy is known and 0.19 eV if the reaction energy is unknown. We expect that this methodology will often replace the explicit calculation of activation energies within surface catalysis when exploring large reaction networks or screening catalysts for desirable properties in the future.

Free and internal energies for the adsorption of short alkanes into the zeolite SSZ-13 from ab initio molecular dynamics simulations.

Electronic structure calculations have become a valuable tool in understanding chemical reactions of hydrocarbons in zeolite pores. However, commonly applied approaches to calculate free energies based on static electronic structure calculations significantly overestimate the entropic penalty for molecular adsorption into zeolite pores. Here, we use ab initio molecular dynamics (AIMD) simulations to model the adsorption of methane, ethane, and propane to purely siliceous and protonated SSZ-13. In our analyses we focus on the internal and Helmholtz free energies of adsorption of each molecule and compare our results to various approaches for the calculation of free energies based on static calculations. We find that only an approach that retains two thirds of the translational entropy of the adsorbate upon adsorption compares favorably with AIMD simulations. However, comparison to experimental measurements of Gibbs free energies of adsorption reported in the literature implies that we might not have captured the full complexity of alkane adsorption in our model. We expect that results in this work will help to develop a better understanding of alkane adsorption in zeolites, and that the provided data will serve as a benchmark for free energy calculations of alkane adsorption in zeolites in the future.

Modeling Electrochemical Processes with Grand Canonical Treatment of Many-Body Perturbation Theory.

Electrocatalysis plays a key role in sustainable energy conversion and storage. It is critical to model the grand canonical treatment of electrons, which accounts for the electrochemical potential explicitly, at the atomic scale and understand these reactions at electrified interfaces. However, such a grand canonical treatment for electrocatalytic modeling is in practice restricted to a treatment of electronic structure with density functional theory, and more accurate methods are in many cases desirable. Here, we develop an original workflow combining the grand canonical treatment of electrons with many-body perturbation theory in its random phase approximation (RPA). Using the potential dependent adsorption of carbon monoxide on the copper (100) facet, we show that the grand canonical RPA energetics provide the correct on-top Cu geometry for CO at reducing potential. The match with experimental results is significantly improved compared to the functionals at the generalized gradient approximation level, which is the most commonly used approximation for electrochemical applications. We expect this development to pave the way to further electrochemical applications using RPA.

Diffusion Barriers for Carbon Monoxide on the Cu(001) Surface Using Many-Body Perturbation Theory and Various Density Functionals.

First-principles calculations play a key role in understanding the interactions of molecules with transition-metal surfaces and the energy profiles for catalytic reactions. However, many of the commonly used density functionals are not able to correctly predict the surface energy as well as the adsorption site preference for a key molecule such as CO, and it is not clear to what extent this shortcoming influences the prediction of reaction or diffusion pathways. Here, we report calculations of carbon monoxide diffusion on the Cu(001) surface along the [100] and [110] pathways, as well as the surface energy of Cu(001), and CO-adsorption energy and compare the performance of the Perdew-Burke-Ernzerhof (PBE), PBE + D2, PBE + D3, RPBE, Bayesian error estimation functional with van der Waals correlation (BEEF-vdW), HSE06 density functionals, and the random phase approximation (RPA), a post-Hartree-Fock method based on many-body perturbation theory. We critically evaluate the performance of these methods and find that RPA appears to be the only method giving correct site preference, overall barrier, adsorption enthalpy, and surface energy. For all of the other methods, at least one of these properties is not correctly captured. These results imply that many density functional theory (DFT)-based methods lead to qualitative and quantitative errors in describing CO interaction with transition-metal surfaces, which significantly impacts the description of diffusion pathways. It is well conceivable that similar effects exist when surface reactions of CO-related species are considered. We expect that the methodology presented here will be used to get more detailed insights into reaction pathways for CO conversion on transition-metal surfaces in general and Cu in particular, which will allow us to better understand the catalytic and electrocatalytic reactions involving CO-related species.

Single-atom gold oxo-clusters prepared in alkaline solutions catalyse the heterogeneous methanol self-coupling reactions.

In an effort to obtain the maximum atom efficiency, research on heterogeneous single-atom catalysts has intensified recently. Anchoring organometallic homogeneous catalysts to surfaces creates issues with retaining mononuclearity and activity, while the several techniques developed to prepare atomically dispersed precious metals on oxide supports are usually complex. Here we report a facile one-pot synthesis of inorganometallic mononuclear gold complexes formed in alkaline solutions as robust and versatile single-atom gold catalysts. The complexes remain intact on impregnation onto supports or after drying in air to give a crystalline powder. They can be used to interrogate the nuclearity of the catalytically active gold site for reactions known to be catalysed by oxidized gold species. We show that the [Au1-Ox]- cluster directs the heterogeneous coupling of two methanol molecules to methyl formate and hydrogen with a 100% selectivity below 180 °C. The reaction is industrially important as well as the key step in methanol steam reforming on gold catalysts.

Computational description of key spectroscopic features of zeolite SSZ-13.

The catalytic properties of zeolites are intimately linked to the distribution and relative positions of Al atoms and defects in the pore network. However, characterizing this distribution is challenging, in particular when different local Al arrangements are considered. In this contribution we use a combination of first principles calculations and experimental measurements to develop a model for the Al-distribution in protonated SSZ-13. We furthermore apply this model to understand trends in OH-IR, 27Al-NMR and 29Si-NMR spectra. We use a Boltzmann distribution to predict the proton position for a given local Al configuration and show that for each configuration several H positions are occupied. Therefore a multi-peak spectrum in OH-IR vibrational spectroscopy is observed for all Al configurations, which is in line with experimentally measured spectra for zeolites at different Si/Al ratios. From NMR spectroscopy we find that the proton position leads to significant shifts in 27Al-NMR and 29Si-NMR spectra due to the modification of the local strain, which is lost when a uniform background charge is introduced. These findings are supported by experimental measurements. Finally we discuss the shortcomings of the presented model in terms of unit cell size and the impact of adjacent unit cells.

Anisotropic Synthesis of Armchair Graphene Nanoribbon Arrays from Sub-5 nm Seeds at Variable Pitches on Germanium.

At widths below 10 nm, armchair graphene nanoribbons become semiconductors. One promising route to synthesize nanoribbons is chemical vapor deposition (CVD) of hydrocarbons on Ge(001), and synthesis from seeds reduces nanoribbon polydispersity. In this contribution, we advance the seed-initiated synthesis of nanoribbons and explore the impact of seed size and nanoribbon spacing on growth kinetics. Periodic arrays of graphene seeds are lithographically patterned and etched to reduce their diameter. The viability of initiating synthesis from sub-5 nm seeds is demonstrated, and the pitch between nanoribbons is reduced from 500 to 50 nm to show that crowding effects do not perturb nanoribbon growth kinetics. The invariance of kinetics with pitch in combination with density functional theory (DFT) calculations indicate that (1) the growth species for synthesis has a diffusion length of ≪50 nm and/or (2) the kinetics are strongly attachment-limited. These results demonstrate that seed-initiated synthesis on Ge(001) is a promising route for creating dense arrays of armchair graphene nanoribbons for semiconductor electronics applications.

Alignment of semiconducting graphene nanoribbons on vicinal Ge(001).

Chemical vapor deposition of CH4 on Ge(001) can enable anisotropic growth of narrow, semiconducting graphene nanoribbons with predominately smooth armchair edges and high-performance charge transport properties. However, such nanoribbons are not aligned in one direction but instead grow perpendicularly, which is not optimal for integration into high-performance electronics. Here, it is demonstrated that vicinal Ge(001) substrates can be used to synthesize armchair nanoribbons, of which ∼90% are aligned within ±1.5° perpendicular to the miscut. When the growth rate is slow, graphene crystals evolve as nanoribbons. However, as the growth rate increases, the uphill and downhill crystal edges evolve asymmetrically. This asymmetry is consistent with stronger binding between the downhill edge and the Ge surface, for example due to different edge termination as shown by density functional theory calculations. By tailoring growth rate and time, nanoribbons with sub-10 nm widths that exhibit excellent charge transport characteristics, including simultaneous high on-state conductance of 8.0 µS and a high on/off conductance ratio of 570 in field-effect transistors, are achieved. Large-area alignment of semiconducting ribbons with promising charge transport properties is an important step towards understanding the anisotropic nanoribbon growth and integrating these materials into scalable, future semiconductor technologies.

Consequences of exchange-site heterogeneity and dynamics on the UV-visible spectrum of Cu-exchanged SSZ-13.

The speciation and structure of Cu ions and complexes in chabazite (SSZ-13) zeolites, which are relevant catalysts for nitrogen oxide reduction and partial methane oxidation, depend on material composition and reaction environment. Ultraviolet-visible (UV-Vis) spectra of Cu-SSZ-13 zeolites synthesized to contain specific Cu site motifs, together with ab initio molecular dynamics and time-dependent density functional theory calculations, were used to test the ability to relate specific spectroscopic signatures to specific site motifs. Geometrically distinct arrangements of two framework Al atoms in six-membered rings are found to exchange Cu2+ ions that become spectroscopically indistinguishable after accounting for the finite-temperature fluctuations of the Cu coordination environment. Nominally homogeneous single Al exchange sites are found to exchange a heterogeneous mixture of [CuOH]+ monomers, O- and OH-bridged Cu dimers, and larger polynuclear complexes. The UV-Vis spectra of the latter are sensitive to framework Al proximity, to precise ligand environment, and to finite-temperature structural fluctuations, precluding the precise assignment of spectroscopic features to specific Cu structures. In all Cu-SSZ-13 samples, these dimers and larger complexes are reduced by CO to Cu+ sites at 523 K, leaving behind isolated [CuOH]+ sites with a characteristic spectroscopic identity. The various mononuclear and polynuclear Cu2+ species are distinguishable by their different responses to reducing environments, with implications for their relevance to catalytic redox reactions.

Can Dynamics Be Responsible for the Complex Multipeak Infrared Spectra of NO Adsorbed to Copper(II) Sites in Zeolites?

Copper-exchanged SSZ-13 is a very efficient material in the selective catalytic reduction of NO(x) using ammonia (deNO(x)-SCR) and characterizing the underlying distribution of copper sites in the material is of prime importance to understand its activity. The IR spectrum of NO adsorbed to divalent copper sites are modeled using ab initio molecular dynamics simulations. For most sites, complex multi-peak spectra induced by the thermal motion of the cation as well as the adsorbate are found. A finite temperature spectrum for a specific catalyst was constructed, which shows excellent agreement with previously reported data. Additionally these findings allow active and inactive species in deNO(x)-SCR to be identified. To the best of our knowledge, this is the first time such complex spectra for single molecules adsorbed to single active centers have been reported in heterogeneous catalysis, and we expect similar effects to be important in a large number of systems with mobile active centers.

Modelling the adsorption of short alkanes in protonated chabazite: The impact of dispersion forces and temperature.

The adsorption of alkanes in a protonated zeolite has been investigated at different levels of theory. At the lowest level we use density-functional theory (DFT) based on semi-local (gradient-corrected) functionals which account only for the interaction of the molecule with the acid site. To describe the van der Waals (vdW) interactions between the saturated molecule and the inner wall of the zeolite we use (i) semi-empirical pair interactions, (ii) calculations using a non-local correlation functional designed to include vdW interactions, and (iii) an approach based on calculations of the dynamical response function within the random-phase approximation (RPA). The effect of finite temperature on the adsorption properties has been studied by performing molecular dynamics (MD) simulations based on forces derived from DFT plus semi-empirical vdW corrections. The simulations demonstrate that even at room temperature the binding of the molecule to the acid site is frequently broken such that only the vdW interaction between the alkane and the zeolite remains. The finite temperature adsorption energy is calculated as the ensemble average over a sufficiently long molecular dynamics run, it is significantly reduced compared to the T = 0 K limit. At a higher level of theory where MD simulations would be prohibitively expensive we propose a simple scheme based on the averaging over the adsorption energies in the acid and in the purely siliceous zeolite to account for temperature effects. With these corrections we find an excellent agreement between the RPA predictions and experiment.

Early stages of water/hydroxyl phase generation at transition metal surfaces--synergetic adsorption and O-H bond dissociation assistance.

The dissociation of water is a key elementary step in many processes. From density functional theory, we show on several transition metal surfaces (Ru, Co, Rh, Ir, Ni, Pd and Pt) that water prefers to chemisorb as a H-bonded dimer, one molecule being chemisorbed by the O atom, but the second one developing only a weak interaction with the surface. Counterintuitively, the molecule in the dimer that shows the smallest activation energy for O-H dissociation is the one interacting weakly with the surface. The H-bonded dimer provides a clear synergy for its chemisorption and assists the dissociation of the H-bond acceptor water molecule. Two different classes of O-H activation pathways are clearly identified with a linear activation energy-reaction energy relationship, of Brønstedt-Evans-Polanyi type.

Van der Waals interactions between hydrocarbon molecules and zeolites: periodic calculations at different levels of theory, from density functional theory to the random phase approximation and Møller-Plesset perturbation theory.

The adsorption of small alkane molecules in purely siliceous and protonated chabazite has been investigated at different levels of theory: (i) density-functional (DFT) calculations with a gradient-corrected exchange-correlation functional; DFT calculations using the Perdew-Burke-Ernzerhof (PBE) functional with corrections for the missing dispersion forces in the form of C(6)∕R(6) pair potentials with (ii) C(6) parameters and vdW radii determined by fitting accurate energies for a large molecular data base (PBE-d) or (iii) derived from "atoms in a solid" calculations; (iv) DFT calculations using a non-local correlation functional constructed such as to account for dispersion forces (vdW-DF); (v) calculations based on the random phase approximation (RPA) combined with the adiabatic-coupling fluctuation-dissipation theorem; and (vi) using Hartree-Fock (HF) calculations together with correlation energies calculated using second-order Møller-Plesset (MP2) perturbation theory. All calculations have been performed for periodic models of the zeolite and using a plane-wave basis and the projector-augmented wave method. The simpler and computationally less demanding approaches (i)-(iv) permit a calculation of the forces acting on the atoms using the Hellmann-Feynman theorem and further a structural optimization of the adsorbate-zeolite complex, while RPA and MP2 calculations can be performed only for a fixed geometry optimized at a lower level of theory. The influence of elevated temperature has been taken into account by averaging the adsorption energies calculated for purely siliceous and protonated chabazite, with weighting factors determined by molecular dynamics calculations with dispersion-corrected forces from DFT. Compared to experiment, the RPA underestimates the adsorption energies by about 5 kJ/mol while MP2 leads to an overestimation by about 6 kJ/Mol (averaged over methane, ethane, and propane). The most accurate results have been found for the "hybrid" RPA-HF method with an average error of less than 2 kJ/mol only, while RPA underestimates the adsorption energies by about 8 kJ/mol on average. MP2 overestimates the adsorption energies slightly, with an average error of 5 kJ/mol. The more approximate and computationally less demanding methods such as the vdW-DF density functional or the C(6)∕R(6) pair potentials with C(6) parameters from "atoms in a solid" calculations overestimate the adsorption energies quite strongly. Relatively good agreement with experiment is achieved with the empirical PBE+d method with an average error of about 5 kJ/mol.

Structure and properties of metal-exchanged zeolites studied using gradient-corrected and hybrid functionals. I. Structure and energetics.

The structural and energetic properties of purely siliceous, proton-, and Cu- and Co-exchanged chabazite have been studied using periodic density-functional (DFT) calculations with both conventional gradient-corrected exchange-correlation functionals and hybrid functionals mixing exact (i.e., Hartree-Fock) and DFT exchange. Spin-polarized and fixed-moment calculations have been performed to determine the equilibrium and excited spin-configurations of the metal-exchanged chabazites. For the purely siliceous chabazite, hybrid functionals predict a slightly more accurate cell volume and lattice geometry. For isolated Al/Si substitution sites, gradient-corrected functionals predict that the lattice distortion induced by the substitution preserves the local tetrahedral symmetry, whereas hybrid functionals lead to a distorted Al coordination with two short and two long Al-O bonds. Hybrid functionals yield a stronger cation-framework binding that conventional functionals in metal-exchanged zeolites, they favor shorter cation-oxygen bonds and eventually also a higher coordination of the cation. Both types of functionals predict the same spin in the ground-state. The structural optimization of the excited spin-states shows that the formation of a high-spin configuration leads to a strong lattice relaxation and a weaker cation-framework bonding. For both Cu- and Co-exchanged chabazite, the prediction of a preferred location of the cation in a six-membered ring of the zeolite agrees with experiment, but the energy differences between possible cation locations and the lattice distortion induced by the Al/Si substitution and the bonding of the cation depends quite significantly on the choice of the functional. All functionals predict similar energy differences for excited spin states. Spin-excitations are shown to be accompanied by significant changes in the cation coordination, which are more pronounced with hybrid functionals. The consequences of electronic spectra and chemical reactivity are analyzed in the following papers.

Structure and properties of metal-exchanged zeolites studied using gradient-corrected and hybrid functionals. II. Electronic structure and photoluminescence spectra.

The influence of the choice of the exchange-correlation functional (semilocal gradient corrected or hybrid functionals) on the electronic properties of metal-exchanged zeolites has been investigated for Cu- and Co-exchanged chabazite. The admixture of exact exchange in hybrid functionals increases the fundamental gap of purely siliceous chabazite, leading to better agreement with experiment and many-body perturbation theory for close-packed SiO(2) polymorphs where detailed experimental information is available. For the metal-exchanged chabazite the increased exchange splitting strongly influences the position of the cation states relative to the framework bands-in general, gradient-corrected functionals locate the occupied cation states close to the valence-band maximum of the framework, while hybrid functionals shift the occupied cation states to larger binding energies and the empty states to higher energies within the fundamental gap. The photoluminescence spectra have been analyzed using fixed-moment total-energy calculations for excited spin states in structurally relaxed and frozen geometries. The geometrical relaxation of the excited states leads to large differences in excitation and emission energies which are more pronounced in calculations using hybrid functionals. Due to the stronger relaxation effects calculated with hybrid functionals, the large differences in the electronic spectra calculated with both types of functionals are not fully reflected in the photoluminescence spectra.

Structure and properties of metal-exchanged zeolites studied using gradient-corrected and hybrid functionals. III. Energetics and vibrational spectroscopy of adsorbates.

The influence of the exchange-correlation functional (semilocal gradient corrected or hybrid functional) on density-functional studies of the adsorption of CO and NO in Cu- and Co-exchanged chabazite has been investigated, extending the studies of the structural and electronic properties of these materials [F. Göltl and J. Hafner, J. Chem. Phys. 136, 064501 (2012); 136, 064502 (2012)] and including for comparison carbonyls and nitrosyls of Cu and Co. Hybrid functionals predict much lower adsorption energies than conventional semilocal functionals, in better agreement with experiment as far as data are available for comparison. The calculated adsorption energies show a strong linear correlation with the stability of the cation sites. For Cu(I)-chabazite the calculated adsorption energies span almost the interval between the adsorption energies calculated for pure neutral and positively charged Cu-carbonyls and nitrosyls. For divalent Cu(II) and Co(II) the adsorption energies at cations in chabazite are much lower than the metal-molecule binding energies in the free carbonyls or nitrosyls, especially for the most stable cation location in a six-membered ring of the chabazite structure. For the stretching modes of adsorbed CO only hybrid functionals reproduce the blueshift of the frequency reported for all Cu(I)- and Co(II)-zeolites. For Cu(II)-chabazite both types of functionals predict a blueshift, the larger value calculated with hybrid functionals being in better agreement with observation. For NO adsorbed on Cu(I)-chabazite all functionals produce a redshift, the smaller value derived with hybrid functionals being in better agreement with experiment. For NO adsorbed in Cu(II)- and Co(II)-chabazite gradient-corrected functionals produce the best agreement with experiment for cations located in a six-membered ring. Semilocal functionals tend to underestimate the frequencies, while hybrid functionals tend to overestimate. The decisive factors determining the influence of the functionals are the larger HOMO-LUMO gap and the larger bandgap of the zeolite host, as well as the larger exchange-splitting of the cation eigenstates predicted with hybrid functionals. For Co(II)-chabazite the tendency to overestimate the exchange-splitting and to stabilize a high-spin state lead to better results with semilocal functionals. Finally, a comprehensive discussion of the influence of the exchange-correlation functional on the physico-chemical properties of these complex systems, based all three papers of this series is presented.

Alkane adsorption in Na-exchanged chabazite: the influence of dispersion forces.

The importance of dispersion forces for the correct description of the adsorption of short alkanes in Na-exchanged and purely siliceous chabazite has been investigated at different levels of theory: (i) standard density-functional (DFT) calculations using the Perdew, Burke, and Ernzerhof (PBE) exchange-correlation functional in the generalized gradient approximation, (ii) dispersion corrections based on empirical force fields according to Grimme [J. Computat. Chem. 134, 1463 (2004)- PBE-d], (iii) calculations based on the van der Waals density functional (vdW-DF) proposed by Dion et al. [Phys. Rev. Lett. 92, 246401 (2004)], and (iv) using the random phase approximation (RPA) in combination with the adiabatic-connection fluctuation-dissipation theorem (RPA-ACFDT), using wave-functions calculated at the DFT and Hartree-Fock (HF) levels. A full relaxation of the adsorbate-zeolite complex was performed at the PBE, PBE-d, and vdW-DF levels. RPA and RPA-HF energies were calculated for the optimized configurations. A critical analysis of the results shows that the most accurate description is achieved at the RPA level with HF exchange energies, while both PBE-d and vdW-DF overestimate the strength of the interaction with the acid site.