New insights on photocatalytic H2 liberation from water using transition-metal oxides: lessons from cluster models of molybdenum and tungsten oxides.

Molecular hydrogen (H2) is an excellent alternative fuel. It can be produced from the abundantly present water on earth. Transition-metal oxides are widely used in the environmentally benign photocatalytic generation of H2 from water, thus actively driving scientific research on the mechanisms for this process. In this study, we investigate the chemical reactions of W3O5(-) and Mo3O5(-) clusters with water that shed light on a variety of key factors central to H2 generation. Our computational results explain why experimentally Mo3O5(-) forms a unique kinetic trap in its reaction while W3O5(-) undergoes a facile oxidation to form the lowest-energy isomer of W3O6(-) and liberates H2. Mechanistic insights on the reaction pathways that occur, as well as the reaction pathways that do not occur, are found to be of immense assistance to comprehend the hitherto poorly understood pivotal roles of (a) differing metal-oxygen and metal-hydrogen bond strengths, (b) the initial electrostatic complex formed, (c) the loss of entropy when these TMO clusters react with water, and (d) the geometric factors involved in the liberation of H2.

Study of MoNbO(y) (y = 2-5) anion and neutral clusters using photoelectron spectroscopy and density functional theory calculations: impact of spin contamination on single point calculations.

Results of a study combining anion photoelectron spectroscopy and density functional theory calculations on the heteronuclear MoNbO(y)(-) (y = 2-5) transition metal suboxide cluster series are reported and analyzed. The photoelectron spectra, which exhibit broad electronic bands with partially resolved vibrational structure, were compared to spectral simulations generated from calculated spectroscopic parameters for all computationally determined energetically competitive structures. Although computational results on the less oxidized clusters could not be satisfactorily reconciled with experimental spectra, possibly because of heavy spin contamination found in a large portion of the computational results, the results suggest that (1) neutral cluster electron affinity is a strong indicator of whether O-atoms are bound in M-O-M bridge positions or M═O terminal positions, (2) MoNbO(y) anions and neutrals have structures that can be described as intermediate with respect to the unary (homonuclear) Mo(2)O(y) and Nb(2)O(y) clusters, and (3) structures in which O-atoms preferentially bind to the Nb center are slightly more stable than alternative structures. Several challenges associated with the calculations are considered, including spin contamination, which appears to cause spurious single point calculations used to determine vertical detachment energies.

Structures of trimetallic molybdenum and tungsten suboxide cluster anions.

Anion photoelectron spectra of Mo(3)O(y)(-) and W(3)O(y)(-) (y = 3-6) are reported and analyzed using density functional theory results in an attempt to determine whether electronic and structural trends in the less oxidized clusters (y = 3, 4) could elucidate the disparate chemical properties of the M(3)O(y)(-) (M = Mo, W, y = 5, 6) species. In general, cyclic structures are calculated to be more stable by at least 1 eV than extended structures, and the lowest energy structures calculated for the most reduced species favor M = O terminal bonds. While the numerous low-energy structures found for Mo(3)O(y)(-)/Mo(3)O(y) and W(3)O(y)(-)/W(3)O(y) were, in general, similar, various structures of W(3)O(y)(-)/W(3)O(y) were found to be energetically closer lying than analogous structures of Mo(3)O(y)(-)/Mo(3)O(y). Additionally, the Mo-O-Mo bridge bond was found to be a more stabilizing structural motif than the W-O-W bridge bond, with the oxygen center in the former having the highest negative charge. Based on this, the observation of trapped intermediates in reactions between Mo(3)O(y)(-) and water or CO(2) that are not observed in analogous W(3)O(y)(-) reactivity studies may be partially attributed to the role of bridge bond fluxionality.

Study of Nb2O(y) (y = 2-5) anion and neutral clusters using anion photoelectron spectroscopy and density functional theory calculations.

A study combining anion photoelectron spectroscopy and density functional theory calculations on the transition metal suboxide series, Nb(2)O(y)(-) (y = 2-5), is described. Photoelectron spectra of the clusters are obtained, and Franck-Condon simulations using calculated anion and neutral structures and frequencies are used to evaluate the calculations and assign transitions observed in the spectra. The spectra, several of which exhibit partially resolved vibrational structure, show an increase in electron affinity with increasing cluster oxidation state. Hole-burning experiments suggest that the photoelectron spectra of both Nb(2)O(2)(-) and Nb(2)O(3)(-) have contributions from more than one structural isomer. Reasonable agreement between experiment and computational results is found among all oxides.

H2 production from reactions between water and small molybdenum suboxide cluster anions.

Reactions between molybdenum suboxide cluster anions, Mo(x)O(y)(-) (x=1-4; y < or = 3x), and water (H(2)O and D(2)O) have been studied using mass spectrometric analysis of products formed in a high-pressure, fast-flow reactor. Product distributions vary with the number of metal atoms in the cluster. Within the MoO(y)(-) oxide series, product masses correspond to the addition of one water molecule, as well as a H/D exchange with MoO(4)H(-). Within the Mo(2)O(y)(-) oxide series, product evolution and distribution suggest sequential oxidation via Mo(2)O(y)(-)+H(2)O/D(2)O-->Mo(2)O(y+1)(-)+H(2)/D(2) reactions for y<5, while for Mo(2)O(5)(-), Mo(2)O(6)H(2)/D(2)(-) is produced. Mo(2)O(6)(-) does not appear to be reactive toward water. For the Mo(3)O(y)(-) oxide series, sequential oxidation similarly is suggested for y<5, while Mo(3)O(5)(-) reactions result in Mo(3)O(6)H(2)/D(2)(-) formation. Mo(3)O(6)(-) appears uniquely unreactive. Mo(3)O(7)(-) and Mo(3)O(8)(-) react to form Mo(3)O(8)H(2)/D(2)(-) and Mo(3)O(9)H(2)/D(2)(-), respectively. Lower mass resolution in the Mo(4)O(y)(-) mass range prevents unambiguous mass analysis, but intensity changes in the mass spectra do suggest that sequential oxidation with H(2)/D(2) evolution occurs for y<6, while Mo(4)O(y+1)H(2)/D(2)(-) addition products are formed in Mo(4)O(6)(-) and Mo(4)O(7)(-) reactions with water. The relative rate constants for sequential oxidation and H(2)O/D(2)O addition for the x=2 series were determined. There is no evidence of a kinetic isotope effect when comparing reaction rates of H(2)O with D(2)O, suggesting that the H(2) and D(2) losses from the lower-oxide/hydroxide intermediates are very fast relative to initial reaction complex formation with H(2)O or D(2)O. The rate constants determined here are two times higher than those determined in identical reactions between W(2)O(y)(-)+H(2)O/D(2)O.

Study of MoVO(y) (y = 2-5) anion and neutral clusters using anion photoelectron spectroscopy and density functional theory calculations.

The vibrationally resolved anion photoelectron (PE) spectra of MoVO(y)(-) (y = 2 - 5) metal suboxide clusters are presented and analyzed in the context of density functional theory (DFT) calculations. The electronically congested spectra reflect an increase in cluster electron affinity with increasing oxidation state. Ion beam hole-burning results reveal the features in the PE spectra of MoVO(2)(-) and MoVO(4)(-) are a result of only one anion isomer, while at least two isomers contribute to electronic structure observed in the PE spectrum of MoVO(3)(-). Spectral features of the binary systems are compared to their pure analogs, Mo(2)O(y) and V(2)O(y). An attempt to characterize the anion and neutral electronic and molecular structures is made by comparison with results from DFT calculations. However, reconciliation between the cluster spectra and the calculated spectroscopic parameters is not as straightforward as in previous studies on similar systems (Yoder, B. L.; Maze, J. T.; Raghavachari, K.; Jarrold, C. C. J. Chem. Phys. 2005, 122, 094313 and Mayhall, N. J.; Rothgeb, D. W.; Hossain, E.; Raghavachari, K.; Jarrold, C. C. J. Chem. Phys. 2009, 130, 124313).

CO2 reduction by group 6 transition metal suboxide cluster anions.

Reactions between small group 6 transition metal suboxide clusters, M(x)O(y)(-) (M = (98)Mo or (186)W; x = 1-4; y < or = 3x) and both CO(2) and CO were studied in gas phase using mass spectrometric analysis of high-pressure, fast flow reaction products. Both Mo(2)O(y)(-) and W(2)O(y)(-) show evidence of sequential oxidation by CO(2) of the form, M(2)O(y)(-)+CO(2)-->M(2)O(y+1)(-)+CO for the more reduced species. Similar evidence is observed for the trimetallic clusters, although Mo(3)O(6)(-) appears uniquely unreactive. Lower mass resolution in the M(4)O(y)(-) range precludes definitive product mass assignments, but intensity patterns suggest the continued trend of sequential oxidation of the more reduced end of the M(4)O(y)(-) oxide series. Based on thermodynamic arguments, cluster oxidation by CO(2) is possible if D(0)(O-Mo(x)O(y)(-)) > 5.45 eV. Although simple bond energy analysis suggests that tungsten oxides may be more reactive toward CO(2) compared to molybdenum oxides, this is not born out experimentally, suggesting that the activation barrier for the reduction of CO(2) by tungsten suboxide clusters is very high compared to analogous molybdenum suboxide clusters. In reactions with CO, suboxides of both metal-based oxides show CO addition, with the product distribution being more diverse for Mo(x)O(y)(-) than for W(x)O(y)(-). No evidence of cluster reduction by CO is observed.

Disparate product distributions observed in Mo((3-x))W(x)O(y) (-) (x=0-3; y=3-9) reactions with D(2)O and CO(2).

Results of gas phase reactivity studies on group six transition metal suboxide clusters, Mo(3)O(y) (-), Mo(2)WO(y) (-), MoW(2)O(y) (-), and W(3)O(y) (-) (Mo((3-x))W(x)O(y) (-), x=0-3; y=ca. 3-9) with both D(2)O and CO(2) are reported. Sequential oxidation for the more reduced species, Mo((3-x))W(x)O(y) (-)+D(2)O/CO(2)-->Mo((3-x))W(x)O(y+1) (-)+D(2)/CO, and dissociative addition for certain species, Mo((3-x))W(x)O(y) (-)+D(2)O/CO(2)-->Mo((3-x))W(x)O(y+1)D(2) (-)/Mo((3-x))W(x)O(y+1)CO(-), is evident in the product distributions observed in mass spectrometric measurements. Reactions with D(2)O proceed at a rate that is on the order of 10(2) higher than for CO(2). The pattern of reaction products reveals composition-dependent chemical properties of these group six unary and binary clusters. At the core of this variation is the difference in Mo-O and W-O bond energies, the latter of which is significantly higher. This results in a larger thermodynamic drive to higher oxidation states in clusters with more tungsten atoms. However, addition products for more oxidized W-rich clusters are not observed, while they are observed for the more Mo-rich clusters. This is attributed to the following: In the higher oxides (e.g., y=8), addition reactions require distortion of local metal-oxygen bonding, and will necessarily have higher activation barriers for W-O bonds, since the vibrational potentials will be narrower. The binary (x=1,2) clusters generally show sequential oxidation to higher values of y. This again is attributed to higher W-O bond energy, the result being that stable binary structures have W atoms in higher oxidation states, and Mo centers both in more reduced states and sterically unhindered. The reduced Mo center provides a locus of higher reactivity. An unusual result that is not readily explained is the chemically inert behavior of Mo(3)O(6) (-).

Water reactivity with tungsten oxides: H(2) production and kinetic traps.

In a recent mass spectrometry/photoelectron spectroscopy study on the reactions between W(2)O(y) (-) (y=2-6) and water, Jarrold and co-workers [J. Chem. Phys. 130, 124314 (2009)] observed interesting differences in the reactivity of the different cluster ions. Particularly noteworthy is the observation that the only product with the incorporation of hydrogens is a single peak corresponding to W(2)O(6)H(2) (-). As reactions between metal oxide clusters and small molecules such as water have high potential for catalytic applications, we carried out a careful study to obtain a mechanistic understanding of this observed reactivity. Using electronic structure calculations, we identified and characterized multiple modes of reactivity between unsaturated tungsten oxide clusters [W(2)O(y) (-) (y=4-6)] and water. By calculating the free energy corrected reaction profiles, our results provide an explanation for the formation of W(2)O(6)H(2) (-). We propose a mechanism in which water reacts with a metal oxide cluster and eliminates H(2). The results from our calculations show that this is nearly a barrierless process for all suboxide clusters with the exception of W(2)O(5) (-).

Termination of the W(2)O(y) (-)+H(2)O/D(2)O-->W(2)O(y+1) (-)+H(2)/D(2) sequential oxidation reaction: An exploration of kinetic versus thermodynamic effects.

Several mechanisms proposed and calculated for the sequential oxidation of tungsten suboxide clusters by H(2)O/D(2)O [Mayhall et al., J. Chem. Phys. 131, 144302 (2009)] are evaluated using anion photoelectron spectroscopy of an apparent intermediate, W(2)O(6)D(2) (-). The spectrum of W(2)O(6)D(2) (-) is consistent with the W(2)O(5) (-)+D(2)O-->W(2)O(6) (-)+D(2) intermediate in which the initial water addition involves the interaction of the oxygen from D(2)O with a tungsten atom, approaching from a direction with the least repulsion from the W(2)O(5) (-) oxygen atoms, coupled with the interaction between a deuterium with a tungsten-tungsten bridging oxygen on the cluster. The presence of W(2)O(6)H(2) (-) and W(2)O(6)D(2) (-) suggests that there is insufficient internal energy in the complex to surmount the barrier for rearrangement required for tungsten hydride and hydroxide formation necessary for H(2) or D(2) evolution, which was calculated to be energetically favorable. The quality of the calculations is verified by direct comparison between experimental photoelectron spectra of W(2)O(5) (-) and W(2)O(6) (-) and spectral simulations generated from the lowest energy structures calculated for W(2)O(5) (-), W(2)O(6) (-) and their corresponding neutrals. The results shed light on the importance of repulsion on the pathway a reaction follows under room temperature conditions.

Structures of Mo(x)W(3-x)O(6) (x = 0-3) anion and neutral clusters determined by anion photoelectron spectroscopy and density functional theory calculations.

The structures of Mo(3)O(6), Mo(2)WO(6), MoW(2)O(6), and W(3)O(6) and their associated anions were studied using a combination of anion photoelectron (PE) spectroscopy and density functional theory calculations. The 3.49 eV photon energy anion PE spectra of all four species showed broad electronic bands with origins near 2.8 eV. Calculations predict that low-spin, cyclic structures are the lowest energy isomers for both the anion and neutral species. The lowest energy neutral structures for all four species are analogous, C(3v) (Mo(3)O(6) and W(3)O(6)) or C(s) (mixed clusters) symmetry structures in which all three metal atoms are in formally equivalent oxidation states, with singlet ground electronic states. The lowest energy isomers predicted for Mo(3)O(6)(-) and W(3)O(6)(-) are the same with doublet electronic states. The lowest energy structures calculated for the mixed anions are lower symmetry, with the tungsten centers in higher oxidation states than the molybdenum centers. However, C(s) symmetry structures are competitive, and appear to be the primary contributors to the observed spectra. Spectral simulations based on calculated spectroscopic parameters validate the assignments. This series of clusters is strikingly different from the Mo(2)O(4)/MoWO(4)/W(2)O(4) anion and neutral series described recently [Mayhall et al., J. Chem. Phys. 130, 124313 (2009)]. While the average oxidation state is the same for both series, the structures determined for the Mo(2)O(4)/MoWO(4)/W(2)O(4) anions and neutrals were dissimilar and lower symmetry, and high spin states were energetically favored. This difference is attributed to the large stabilizing effect of electronic delocalization in the more symmetric trimetallic cyclic structures that is not available in the bimetallic species.

Electronic structures of MoWO(y)- and MoWO(y) determined by anion photoelectron spectroscopy and DFT calculations.

The anion photoelectron spectra of MoWO(y)(-) (y=2-5) and density functional theory (DFT) calculations on MoWO(y)(-) and MoWO(y) are reported and compared to previous comparable studies on Mo(2)O(y)(-)/Mo(2)O(y) and W(2)O(y)(-)/W(2)O(y). The property governing the structure of the lowest energy MoWO(y) anion and neutral clusters is the stronger W-O bond relative to the Mo-O bond, which results in the stabilization of structures in which the Mo center is in a much lower oxidation state than the W center. Anion PE spectra show a much larger change in structure between anion and neutral states than what was observed in the pure Mo(2)O(y)(-) and W(2)O(y)(-) spectra. DFT calculations show increased single-metal localization of spin with respect to the pure metal oxide clusters.

Unusual products observed in gas-phase W(x)O(y)- + H2O and D2O reactions.

Addition of H(2)O and D(2)O to small tungsten suboxide cluster anions W(x)O(y)(-) (x = 1-4; y < or = 3x) was studied using mass spectrometric measurements from a high-pressure fast flow reactor. Within the WO(y)(-) mass manifold, which also includes WO(4)H(-), product masses correspond to the addition of one to three H(2)O or D(2)O molecules. Within the W(2)O(y)(-) cluster series, product distributions suggest that sequential oxidation W(2)O(y)(-) + H(2)O/D(2)O --> W(2)O(y+1)(-) + H(2)/D(2) occurs for y < 5, while for W(2)O(5)(-), W(2)O(6)H(2)(-)/W(2)O(6)D(2)(-) is primarily produced. W(2)O(6)(-) does not appear reactive. For the W(3)O(y)(-) cluster series, sequential oxidation with H(2) and D(2) production occurs for y < 6, while W(3)O(6)(-) and W(3)O(7)(-) produce W(3)O(7)H(2)(-)/W(3)O(7)D(2)(-) and W(3)O(8)H(2)(-)/W(3)O(8)D(2)(-), respectively. Lower mass resolution in the W(4)O(y)(-) mass range prevents definitive product assignments, but intensity patterns suggest that sequential oxidation with H(2)/D(2) evolution occurs for y < 6, while W(4)O(y+1)H(2)(-)/W(4)O(y+1)D(2)(-) products result from addition to W(4)O(6)(-) and W(4)O(7)(-). Based on bond energy arguments, the H(2)/D(2) loss reaction is energetically favored if the new O-W(x)O(y)(-) bond energy is greater than 5.1 eV. The relative magnitude of the rate constants for sequential oxidation and H(2)O/D(2)O addition for the x = 2 series was determined. There are no discernable differences in rate constants for reactions with H(2)O or D(2)O, suggesting that the H(2) and D(2) loss from the lower-oxide/hydroxide intermediates is very fast relative to the addition of H(2)O or D(2)O.

Tungsten carbide revisited: New anion photoelectron spectrum and density functional theory calculations.

A new anion photoelectron (PE) spectrum of WC(-) is presented and analyzed using existing and new calculations. This spectrum is different from the PE spectrum previously published [X. Li et al., J. Chem. Phys. 111, 2464 (1999)], which we suggest was actually the PE spectrum of W(-). The ground anion state is determined by calculations and comparison with spectral features to be the (2)Delta(32) state. The (3)Delta(1)-(2)Delta(32) transition is observed at an electron binding energy of 2.155 eV, which corresponds to the adiabatic electron affinity of WC. The bond length of the anion is determined to be 1.771(5) A. A number of spectral assignments corresponding to both excited anion and neutral states are made based on previously obtained fluorescence spectra [S. M. Sickafoose et al., J. Chem. Phys. 116, 993 (2002)] and density functional theory calculations.

Solvation of O(2)(-) and O(4)(-) by p-difluorobenzene and p-xylene studied by photoelectron spectroscopy.

Anion photoelectron spectra of the O(2)(-) . arene and O(4)(-) . arene complexes with p-xylene and p-difluorobenzene are presented and analyzed with the aid of calculations on the anions and corresponding neutrals. Relative to the adiabatic electron affinity of O(2), the O(2)(-) . arene spectra are blueshifted by 0.75-1 eV. Solvation energy alone does not account for this shift, and it is proposed that a repulsive portion of the neutral potential energy surface is accessed in the detachment, resulting in dissociative photodetachment. O(2)(-) is found to interact more strongly with the p-difluorobenzene than the p-xylene. The binding motif involves the O(2)(-) in plane with the arene, interacting via electron donation along nearby C-H bonds. A peak found at 4.36(2) eV in the photoelectron spectrum of O(2)(-) . p-difluorobenzene (p-DFB) is tentatively attributed to the charge transfer state, O(2)(-) . p-DFB(+). Spectra of O(4)(-) . arene complexes show less blueshift in electron binding energy relative to the spectrum of bare O(4)(-), which itself undergoes dissociative photodetachment. The striking similarity between the profiles of the O(4)(-) . arene complexes with the O(4)(-) spectrum suggests that the O(4)(-) molecule remains intact upon complex formation, and delocalization of the charge across the O(4)(-) molecule results in similar structures for the anion and neutral complexes.

Identification and characterization of novel sodium/potassium-ATPase inhibitors by virtual screening of a compound database.

The medicinal value of cardiac glycoside inhibitors for the treatment of congestive heart failure symptoms stems from their ability to specifically inhibit the ion transport activity of the transmembrane enzyme sodium/potassium-ATPase (Na/K-ATPase) in myocardial cells. In this study, we used the inhibitory potencies of 39 cardiac glycoside analogues for the development of a quantitative structure-activity relationship (QSAR) model for Na/K-ATPase inhibition. In conjunction with a substructure and similarity search, the QSAR model was used to select ten potential inhibitors from a commercial compound database. The inhibitory potencies of these compounds were measured and four were found to be more active than the commonly used inhibitor ouabain. The results of the bioassays were incorporated into a second QSAR model, whose physical interpretation suggested that the nature of substituents in positions 10, 12, and 17 at the cyclopentanoperhydrophenanthrene core of the inhibitors was critical for enzyme inhibition. All descriptors of the QSAR models were conformation-independent, making the search protocol a suitable tool for the rapid virtual screening of large compound databases for novel inhibitors.