Gas-Phase Reactions of [VO2(OH)2]â and [V2O5 ... - ACS Publications

Article pubs.acs.org/JPCA

Gas-Phase Reactions of [VO2(OH)2]− and [V2O5(OH)]− with Methanol: Experiment and Theory Benjamin L. Harris,†,‡,§ Tom Waters,†,‡ George N. Khairallah,*,†,‡ and Richard A. J. O’Hair*,†,‡,§ †

School of Chemistry, ‡Bio21 Institute of Molecular Science and Biotechnology, and §ARC Centre of Excellence for Free Radical Chemistry and Biotechnology, The University of Melbourne, Melbourne, Victoria 3010, Australia S Supporting Information *

ABSTRACT: The gas-phase reactivity of the vanadium hydroxides [VO2(OH)2]− and [V2O5(OH)]− toward methanol was examined using a combination of ion−molecule reactions (IMRs) and collision-induced dissociation (CID) in a quadrupole ion trap mass spectrometer. Isotopelabeling experiments with CD3OH, 13CH3OH, and CH318OH were used to confirm the stoichiometry of ions and the observed sequence of reactions. The experimental data were interpreted with the aid of density functional theory calculations, carried out at the B3LYP/SDD6-311++G** level of theory. While [VO2(OH)2]− is unreactive, [V2O5(OH)]− undergoes a metathesis reaction to yield [V2O5(OCH3)]−. The DFT calculations reveal that the metathesis reaction of methanol with [VO2(OH)2]− suffers from a barrier of +0.52 eV (relative to separated reactants) but that the reaction of [V2O5(OH)]− with methanol readily proceeds via addition/elimination reactions with both transition states being below the energy of the separated reactants. CID of [V2O5(OCH3)]− (m/z 213) yields three ions arising from activation of the methoxo ligand: [V2, O6, C, H]− (m/z 211); [V2, O5, H]− (m/z 183); and [V2, O4, H]− (m/z 167). Additional experiments and DFT calculations suggest that these ions arise from losses of H2, formaldehyde and the sequential losses of H2 and CO2, respectively. The use of an 18O-labeled methoxo ligand in [V2O5(18OCH3)]− (m/z 215) showed the competing losses of H2C16O and H2C18O and [H2 and C16O18O] and [H2 and C16O2], highlighting that 16O/18O exchange between the methoxo ligand and the vanadium oxide occurs prior to the subsequent fragmentation of the ligand. DFT calculations reveal that a key step involves hydrogen atom transfer from the methoxo ligand to the oxo ligand of the same vanadium center, producing the intermediate [V2O4(OH)(OCH2)]− containing a ketyl radical ligand and a hydroxo ligand. This intermediate can either undergo CH2O loss, or the ketyl radical can couple with an oxo ligand of the adjacent vanadium center, producing [V2O3(μ2-O2CH2)]−, which is a key intermediate in the 16O/18O scrambling and in the H2 loss channel.

■

INTRODUCTION Many commodity organic chemicals are produced by heterogeneous oxidation catalysis,1 with more than a third of worldwide catalyst production based on the use of metal oxides.2 In order to improve and develop new metal oxide catalysts, attention has focused on the molecular structure and reactivity of active sites.2−9 While great strides have been made in the use of in situ and operando spectroscopic investigations of catalysts,10 detailed mechanistic studies are often hampered by factors such as the dynamic nature of the catalyst surface and its inhomogeneity. Furthermore, important reactions appear to occur at localized reaction sites (e.g., surface defects or adsorbed reactant molecules), and such sites might exhibit significantly different reactivity to that of the bulk oxide. Thus, a proper understanding of these differences is vital. Mass-spectrometry-based gas-phase studies of isolated transition-metal oxide cluster ions can provide molecular-level insights into the fundamental reactivity of clusters of known stoichiometry under well-defined experimental conditions.11,12 Many of these studies are carried out in conjunction with DFT calculations,16i and the insights that they provide can help one to understand related reactions occurring at oxide surfaces. A key aim is to identify new modes of reactivity and bonding in the gas phase that might be relevant to condensed-phase © XXXX American Chemical Society

chemistry. Numerous studies have reported on the gas-phase reactions of a vast range of cationic and anionic mononuclear and multinuclear metal oxides with organic substrates such as alkanes,14 alkenes,15 and alcohols.16,17 Relatively few complete catalytic cycles have, however, been uncovered.16g-j One widely used commodity is formaldehyde, which can be produced via oxidation of methanol over metal oxide surfaces,19 particularly molybdenum oxide surfaces20 or bulk or supported vanadium oxides.5,21 The reactions of these metal oxides likely proceed via the intermediacy of metal alkoxide species. We have previously reported two different, complete catalytic cycles involving the oxidation of methanol to formaldehyde, as shown in Scheme 1. Cycle 1 involves the dimolybdate anion, [Mo2O6(OH)]−, as catalyst and nitromethane as the terminal oxidant.16h In contrast, the mononuclear molybdate anion, [MoO3(OH)]− does not act as a catalysts because it fails to undergo the related addition/elimination reaction (cf. step a of cycle 1), highlighting the important role of the second metal center in catalysis.16h A key intermediate in cycle 1 is the Special Issue: Peter B. Armentrout Festschrift Received: May 11, 2012 Revised: July 11, 2012

A

dx.doi.org/10.1021/jp3046142 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Scheme 1. Gas-Phase Catalytic Cycles for the Oxidation of Methanol Examined in Detailed Using Mass Spectrometry Experimentsa

a

Cycle 1 is catalyzed by the dimolybdate anion, [Mo2O6(OH)]−;16h cycle 2 is catalyzed by the metavanadate anion, [VO3]−.16i,j

methoxide, [Mo2O6(OCH3)]−, which liberates formaldehyde in step b upon “heating” under conditions of collision-induced dissociation (CID). Cycle 2 utilizes the mononuclear metavanadate anion [VO3]− as the catalyst and dioxygen as the terminal oxidant.16i,j The 2e− oxidation of the alcohol was linked to the 4e− reduction of dioxygen. A key step in the process was the reaction of [VO3]− with methanol to eliminate water and form [VO2(η2-OCH2)]−, with an [η2-O,C-OCH2]2− ligand that may be regarded formally as doubly deprotonated methanol or as 2e−-reduced formaldehyde. This ligand is isoelectronic with the peroxo ligand O22− and was crucial in linking the 2e− oxidation of two equivalents of methanol to the 4e− reduction of dioxygen.16i A range of functional groups can be present on vanadium oxide surfaces, including oxo, peroxo, and hydroxide, and thus, it is of interest to examine their reactivity through the use of model systems. Cycle 2 demonstrated that models for both oxo and peroxo groups are reactive and involved in the catalytic oxidation of methanol. In order to address what role vanadium hydroxides might play in the oxidation of methanol, the present work (i) compares the reactivity of the vanadium hydroxides [VO2(OH)2]− and [V2O5(OH)]− toward methanol and (ii) examines the types of products formed upon activation of the methoxo group of [V2O5(OCH3)]− (previous studies have examined the related mononuclear species, [VO2(OH)(OCH3)]−16i,j). Where appropriate, reaction pathways were examined by kinetic and isotope labeling experiments, and the experimental data were interpreted with the aid of density functional theory (DFT) calculations.

■

NaVO3 was dissolved in 1:1 H2O/MeCN to ∼0.75 mM, and this solution was introduced into the electrospray source via a syringe pump at a flow rate of 3 μL·min−1. Typical electrospray conditions involved needle potentials of 3.8−4.2 kV and a heated capillary temperature of 200 °C. Mass selection and collision-induced dissociation were carried out using standard isolation and excitation procedures using the “advanced scan” function of the LCQ software. The instrument has been modified to permit introduction of neutral reagents into the ion trap, allowing the measurement of ion−molecule reaction rate constants. These modifications and the procedure for measuring rates have been described in detail previously,16h,22 and applications in metal ion chemistry have been highlighted.22b Gronert has shown that ions in the LCQ are thermalized by the helium bath gas and are thus essentially at room temperature when undergoing ion−molecule reactions.22c The rate constant for the reaction of [V2O5(OH)]− with methanol was measured 10 times on 4 different occasions, and the rate constants are listed in Table S1 (Supporting Information). Reaction efficiencies (ϕ) were calculated by dividing the experimentally determined rate constant (kexp) by theoretical predictions of ion−molecule collision rate constants (kado). The collision rates were calculated via the average dipole orientation (ADO) theory of Su and Bowers using the COLRATE program.23 Theoretical Calculations. Density functional theory (DFT) calculations were carried out using the Gaussian 03 program.24 The B3LYP functional was used,25 and calculations were performed using the Stuttgart effective core potential (SDD) basis set for V and the 6-311++G** basis set for C, H, and O.26 This combination of method and basis set is hereafter referred to as B3LYP/SDD6-311++G** and was chosen because it has been previously shown to be effective in calculating geometries and bond lengths close to available experimental data as it provides reasonable agreement with the results from higher-level theoretical calculations14e,f,27 to allow direct comparisons with the results of related vanadium oxide anions.16i,j Stationary points were characterized by frequency calculations, and unscaled zero-point energies are included for all species. Relevant Cartesian coordinates and energies for all species are included in the Supporting Information.

METHODS

Reagents. Sodium metavanadate, NaVO3, CH3OH (HPLC grade), CD3OH (99 atom % D), CH3CH2OH (HPLC grade), CD3CH2OH (98 atom % D), and CH3CD2OH (98 atom % D) were obtained from Aldrich. CH318OH (95 atom % 18O) was obtained from Isotec, and 13CH3OH was obtained from Cambridge Isotope Laboratories. All reagents were used without further purification. Mass Spectrometry. Mass spectrometry experiments were conducted using a modified Finnigan LCQ quadrupole ion trap mass spectrometer with an electrospray ionization source.16h B



■

Article

(3), with a five-coordinate VV center. The formation of this product is predicted to be endothermic by 0.09 eV and is thus not expected to be observed experimentally. Alternatively, the metathesis reaction proceeding via addition of methanol coupled with loss of water to yield [VO2(OH)(OCH3)]− (m/z 131) (4) is predicted to be exothermic by 0.09 eV. The fact that this reaction pathway is not observed experimentally suggests a kinetic barrier due to the presence of a high-energy transition state. Thus, the potential energy surface for the addition of methanol with the loss of neutral water was examined. The first step involves the formation of the ion− molecule complex [VO2(OH)2(CH3OH)]− (5), which is exothermic by 0.53 eV (Figure 1). The concerted transition state for transfer of a proton from methanol to a hydroxo ligand to eliminate water (TS 5−6) is high in energy, 0.52 eV above the energy of the separated reactants. This indicates that the metathesis reaction (eq 1b) should not occur, in support of the experimental results. Reaction of [V2O5(OH)]− with Methanol. The binuclear anion, [V2O5(OH)]− (m/z 199) (2), was also mass-selected in the ion trap and allowed to react with methanol (Figure 2). In

RESULTS AND DISCUSSION As noted in previous work,16i,j ESI of H2O/MeCN solutions of sodium metavanadate produced a range of vanadium oxide ions in the negative ion mode, including [VO2(OH)2]− (m/z 117) (1) and [V2O5(OH)]− (m/z 199) (2), which possesses two bridging oxo ligands.28 Numerous other, larger vanadium oxo clusters containing up to 17 vanadium atoms were also detected. For example, [V 3 O 8 ] − (m/z 281) and [V4O10(OH)]− (m/z 381) are among the most abundant ions observed below m/z 450, but their reactivity is not discussed here. Here, the gas-phase reactivity of the anions [VO 2 (OH) 2 ] − and [V 2 O 5 (OH)] − toward methanol is described. Reaction of [VO2(OH)2]− with Methanol. Mononuclear [VO2(OH)2]− (m/z 117) (1) was mass-selected and exposed to methanol vapor introduced through the helium bath gas line. It was unreactive with methanol, and products arising from either addition (eq 1a) or an addition/elimination metathesis reaction (eq 1b) were not observed over the maximum reaction time (10 s, data not shown). This suggests that reactions 1a and 1b are either endothermic or involve a high-energy transition state. [VO2 (OH)2 ]− (m/z 117) + CH3OH → [VO(OH)3 (OCH3)]− (m/z 149) → [VO2 (OH)(OCH3)]− (m/z 131) + H 2O

1a 1b

DFT Calculations on the Interaction of [VO2(OH)2]− and CH3OH. DFT calculations were carried out on the possible products of methanol addition (eq 1a) and methanol addition/water elimination (eq 1b) in order to gain insights into why these reactions did not occur, and the results are shown in Figure 1. Methanol addition via proton transfer to an oxo ligand would produce [VO(OH)3(OCH3)]− (m/z 149) Figure 2. LCQ mass spectrum showing the reaction of [V2O5(OH)]− (m/z 199) (2) and methanol to generate [V2O5(OCH3)]− (7) (m/z 213). The reaction time is 300 ms, and the concentration of MeOH is ∼2.5 × 1010 molecules·cm−3.

contrast to the mononuclear [VO2(OH)2]− (m/z 117), (2) reacted with methanol to form a product ion detected at m/z 213 of stoichiometry [V2, O6, C, H3]−. This observation is consistent with a metathesis reaction involving addition of CH3OH and loss of water (eq 2), similar to that occurring for [Mo2O6(OH)]− (Scheme 1, step a of cycle 1).16h The reaction of [V2O5(OH)]− with methanol proceeds with a rate constant of 1.9 × 10−10 cm3 molecules−1 s−1, corresponding to a reaction efficiency of ∼13% (Table S1 Supporting Information). Reactions with the isotopically labeled methanols 13CH3OH, CH318OH, and CD3OH gave ions at m/z 214, 215, and 216 (Figure S1, Supporting Information), consistent with a metathesis reaction. [V2O5(OH)2 ]− (m/z 199) + CH3OH → [V2O5(OCH3)]− (m/z 213) + H 2O Figure 1. (a) B3LYP/SDD6-311++G(d,p) calculated potential energy diagram of the reaction of [VO2(OH)2]− (1) and methanol. Energies are in eV and are relative to [VO2(OH)2]− (1) + CH3OH. (b) Calculated minimum-energy structures of the intermediates and products.

(2) −

Theoretical Investigation of [V2, O6, C, H3] (m/z 213) and the Potential Energy Surface for Interaction of [V2O5(OH)]− and CH3OH. To calculate the minimum-energy structures of the intermediate [V2, O6, C, H3]−, a number of C



Article

The reaction begins by the formation of an ion−molecule complex between [V2O5(OH)]− and methanol (8), predicted to be exothermic by 0.55 eV. The first transition state (TS 8− 9) involves a bridging oxo ligand deprotonating methanol with concomitant opening of the VO2V ring to form [(HO)2(O)V− O−V(O)2(OCH3)]− (9), which possesses only a single bridging oxo ligand. Facile intramolecular proton transfer from the hydroxo ligand of 9 to an oxo ligand via TS 9−10 yields [(HO)(O)2V−O−V(O)(OH)(OCH3)]− (10), which also possesses only a single bridging oxo ligand. The final transition state (TS 10−11) involves a hydroxo ligand on one vanadium center deprotonating a second hydroxo ligand on the adjacent vanadium center, with concomitant re-formation of the VO2V ring to form an ion−molecule complex between [V2O5(OCH3)]− and water (11). Loss of water gives the experimentally observed product [V2O5(OCH3)]− (m/z 213) (7a). Each of the reactants and transition states are below the entrance channel (Figure 4a), consistent with the reaction being observed experimentally (Figure 2). A key feature of the binuclear system [V2O5(OH)]− (2) is its bridging oxo ligands. One of these opens during methanol addition (via TS 8−9) and closes during water elimination (via TS 10−11), allowing the vanadium centers to remain fourcoordinate. This is in contrast with the high-energy fivecoordinate transition state (TS 5−6, Figure 1) predicted for the mononuclear, [VO2(OH)2]− (1) and provides an explanation for why the binuclear system is reactive while the mononuclear is not. Furthermore, this observation is in line with studies of bulk and solution vanadium catalysts, in which bridging oxo ligands are linked to reactivity.29a−d Collisional Activation and Decomposition of the Methoxo Ligand in [V2O5(OCH3)]−. Methoxo ligands related to that of [V2O5(OCH3)]− (m/z 213) (7a), are observed as

singlet and triplet isomers were considered, and some of the minima found are shown in Figure 3. The singlet isomer

Figure 3. B3LYP/SDD6-311++G(d,p) calculated isomeric forms of the anion of stoichiometry [V2, O6, C, H3]− at m/z 213. Spin values and energies relative to the lowest-energy structure, [V2O5(OCH3)]− (7a) (eV), are listed. The spin expectation values (S2) are also listed.

[V2O5(OCH3)]− (7a) was calculated to be over 0.5 eV more stable than any other isomer, and thus, this is assigned as the sole experimental product in the reaction of [V2O5(OH)]− and CH3OH (Figure 2). Calculations were also carried out to examine the formation of this ion, where all of the lowestenergy intermediates and transition states occurred on the singlet surface. The minimum-energy pathway calculated for reaction of [V 2 O 5 (OH)] − with methanol to form [V2O5(OCH3)]− is presented in Figure 4a. Structures of key intermediates are presented in Figure 4b.

Figure 4. (a) B3LYP/SDD6-311++G(d,p) calculated potential energy diagram of the reaction of [V2O5(OH)]− (2) and methanol. Energies are in eV and are relative to [V2O5(OH)]− (2) + CH3OH. (b) Calculated minimum-energy structures of the intermediates and products. D



Article

the observed ions assigned in Scheme 2 are based on the experimental isotope labeling, reactivity studies, and theoretical calculations described in detail below. The primary fragmentation products were assigned as [V2, O6, C, H]− (m/z 211) (12), [V2, O5, H]− (m/z 183) (13), and [V2, O4, H]− (m/z 167) (14). The other ions observed experimentally all originated from secondary reactions of these product ions. For example, the product at m/z 167, assigned to [V2, O4, H]−, reacted with background water to generate an ion at m/z 183, presumably from addition of water and loss of H2 (+H2O, −H2). It also underwent an equivalent reaction with background methanol to yield m/z 197 (i.e., +CH3OH, −H2). Both of these secondary products then underwent further reaction with background methanol and water (Figure 5b). Similarly, the primary fragmentation product at m/z 183 and assigned to [V2, O5, H]− reacted with methanol to produce m/z 197 via addition of methanol and loss of water. It also reacted via direct addition of water or methanol to produce m/z 201 and 215, respectively (Figure 5c). Each of these secondary product ions also underwent further reactions. Finally, the third product ion at m/z 211 was assigned to [V2, O6, C, H]−. It only underwent addition of background water and methanol to produce m/z 229 and 243, respectively, and no other reactions were observed (Figure 5d). The ion at m/z 183 was assigned to a primary fragmentation product from parent [V2O5(OCH3)]−. However, an ion at m/z 183 was also generated from a secondary reaction of m/z 167 with background water. From CID experiments conducted on [V2O5(OCD3)]− (Figure S2d, Supporting Information), we can deduce that this secondary reaction accounts for only some of the m/z 183 observed experimentally. Product [V2, O6, C, H]− at m/z 211 can only arise from direct loss of H2 from parent [V2O5(OCH3)]−. In contrast, [V2, O5, H]− at m/z 183 might arise from loss of intact H2CO or the combined loss of H2 and CO. Similarly, the product assigned to [V2, O4, H]− at m/z 167 could arise from loss of HCO2H or the combined losses of H2 and CO2 or H2O and CO. Therefore, the fragment ions at m/z 183 and 167 might arise from [V2, O6, C, H]− m/z 211 undergoing further loss of CO or CO2, respectively. To examine this further, the primary product ion at m/z 211 was mass-selected and subjected to CID, resulting in formation of m/z 167 of stoichiometry [V2, O4, H]− via the loss of CO2 (data not shown). No loss of CO was observed. Furthermore, the reactivity of product [V2, O4, H]− at m/z 167 generated from m/z 211 was equivalent to that generated directly from [V2O5(OCH3)]− (m/z 213), suggesting that they were equivalent species (data not shown). This indicated that product [V2, O4, H]− is most likely formed from [V2O5(OCH3)]− via the sequential losses of H2 (eq 3a) and CO2 (eq 4). In contrast, the fact that [V2, O6, C, H]− (m/z 211) does not undergo loss of CO indicates that [V2, O5, H]− (m/z 183) is most likely formed from parent [V2O5(OCH3)]− via direct loss of formaldehyde rather than loss of H2 followed by CO. The combined data are summarized in eqs 3 and 4.

intermediates on vanadium oxide surfaces, and are stable under moderate conditions.29a However, at high temperatures (e.g., ∼230 °C), these methoxo ligands are oxidized and dissociate via loss of formaldehyde. A related reaction occurs when the dimolybdate anion [Mo2O6(OCH3)]− is “heated” in the gas phase under conditions of CID (step b of cycle 1 in Scheme 116h). In order to establish whether a related reaction occurs for [V2O5(OCH3)]− (m/z 213), this ion was mass-selected and subjected to CID. The resultant mass spectrum is shown in Figure 5a and shows a rich set of products. Several of these

Figure 5. Mass spectra of relevance to the thermal decomposition of [V2O5(OCH3)]−: (a) CID spectrum of [V2O5(OCH3)]− (m/z 213) to generate [V2O5(OCH)]− (m/z 211), [V2O4(OH)]− (m/z183), [V2O4(H)]− (m/z 167), and other secondary products; (b) isolation of the primary product, [V2O4(H)]− (m/z 167), followed by ion− molecule reactions with background water and methanol; (c) isolation of the primary product, [V2O4(OH)]− (m/z 183), followed by ion− molecule reactions with background water and methanol; (d) isolation of the primary product, [V2O4(O2CH)]− (m/z 211), followed by ion− molecule reactions with background water and methanol. Massselected precursor ions are marked by an asterisk.

products are either at higher m/z values than the precursor ions or correspond to unlikely losses of 12 or 16 Da, suggesting that they arise from ion−molecule reactions of fragment ions with background water and methanol present in the ion trap. Attempts to establish the source of these ions by electrospraying the sodium vanadate from a pure D2O solution added further complications due to H/D exchange reactions, including those with the background, unlabeled water (data not shown). Thus, in order to confirm these assignments and thus the origin of each of the observed ions, they were examined experimentally by mass selection, and their reactivity with background water and methanol was probed (Figure 5b− d). A general scheme summarizing the formation of different ions is presented in Scheme 2. Stoichiometries and structures of

[V2O5(OCH3)]− (m/z 213) → [V2 , O6 , C, H]− (m/z 211) + H 2 → [V2 , O5 , H]− (m/z 183) + H 2CO E

3a 3b



Article

Scheme 2. Reactions Associated with the Gas-Phase Fragmentation of [V2O5(OCH3)]− and Subsequent Reactions with Background Water and Methanol

Figure 6. B3LYP/SDD6-311++G(d,p) calculated structures, spin states, and relative energies (eV) for different isomers of ions of stoichiometry (a) [V2, O6, C, H]− (m/z 211) (12); (b) [V2, O5, H]− (m/z 183) (13); and (c) [V2, O4, H]− (m/z 167) (14).

stoichiometry [V2, O6, C, H]− (12). Structures and relative energies are shown in Figure 6a. The most stable isomer was triplet [V2O4(O2CH)]− containing two V(IV) centers and a bidentate formate ligand HCO2− bridging both metal centers. Other minimum-energy structures calculated include, for example, the equivalent singlet structure that was found to be +1.25 eV higher in energy. The remaining structures presented in Figure 6 are at least 0.41 eV higher in energy. The observed reactivity of the ion at m/z 211 (Figure 5d and Scheme 2), where only neutral substrate adsorption is observed, appears to support the calculated structure with the lowest energy (i.e.; [V2O4(O2CH)]−). [V2, O5, H]− (m/z 183). The most stable isomer was mixedvalence triplet [V2O4(OH)]− (13a) with a hydroxo ligand and

[V2 , O6 , C, H]− (m/z 211) → [V2 , O4 , H]− (m/z 167) + CO2

(4)

Structures of the Product Ions from Fragmentation of [V2O5(OCH3)]−. Initial theoretical calculations were aimed at addressing the structures of the product ions of stoichiometry [V2, O6, C, H]− (m/z 211, eq 3a), [V2, O5, H]− (m/z 183, eq 3b), and [V2, O4, H]− (m/z 167, eq 4), and the results are shown in Figure 6. The validity of the structures proposed from these theoretical calculations was assessed in light of the reactivity of the various ions described above. [V2, O6, C, H]− (m/z 211). A number of isomers of both singlet and triplet spin states were considered for the ion of F



Article

methoxo ligand need to take into account these exchange processes.

one V(V) center and one V(III) center. The isomer with the hydroxo ligand at the V(V) center was favored over that with the hydroxo ligand at the V(III) center by 0.20 eV. A third isomer, singlet [V2O5(H)]− with an hydrido ligand and two V(V) centers, was predicted to be only 0.06 eV higher in energy than the most favored isomer. The similar predicted energies of these three isomers make it difficult to distinguish on theoretical grounds. However, the ion−molecule reactions of [V2, O5, H]− with background methanol (Figure 5c and Scheme 2) are consistent with a metathesis reaction involving a hydroxide ligand rather than a hydride ligand because an ion corresponding to a formula [V2O4(OCH3)]− (m/z 197) is observed rather than [V2O5(OCH3)]− (m/z 213). Thus, isomer 13c is ruled out. [V2, O4, H]− (m/z 167). The most favored isomer was mixedvalence triplet [V2O4(H)]− (14a) containing a hydrido ligand and with one V(V) center and one V(III) center. The equivalent singlet isomer was 0.70 eV higher in energy. The pentet state of an alternative isomer [V2O3(OH)]− with a hydroxo ligand was 0.48 eV above the triplet state of [V2O4(H)]−. This assignment is further supported by the observed reactivity of m/z 167 with background water and methanol (Figure 5b and Scheme 2), where the products observed at m/z 183 and 197 corresponded to the addition of the neutral substrate and loss of H2, as expected for a metathesis reaction involving a hydride ligand. Thus, isomers 14c and 14d, which contain a hydroxide ligand, are ruled out. Isotope Labeling Experiments. Isotope labeling experiments with CD3OH, 13CH3OH, and CH318OH were used to confirm the stoichiometry of ions and the observed sequence of reactions proposed in Scheme 2. Results for the three decomposition products formed in eqs 3 and 4 as well as the parent [V2O5(OCH3)]− (7a) are summarized in Table 1, and spectra are included in the Supporting Information (Figure S1−S9).

[V2O5(18OCH3)]− (m/z 215) → [V2O318O(OH)]− (m/z 185) + H 2CO → [V2O4 (OH)]− (m/z 183) + H 2C18O [V2O5(18OCH)]− (m/z 213) → [V2 , O4 , H]− (m/z 167) + OC18O → [V2 , O3 , 18O, H]− (m/z 169) + CO2

m/z [V2O5(OCH3)]− (parent ion) [V2O4(O2CH)]− (loss of H2) [V2O4(OH)]− (loss of H2CO) [V2O4(H)]− (loss of H2 + CO2)

CH3OH

13

CH3OH

CH318OH

CD3OH

213

214

215

216

211

212

213

212

183

183

183 + 185

184

167

167

167 + 169

168

5b

6a 6b

DFT Calculations on the Mechanism of Methoxo Ligand Activation. Theoretical calculations were employed to probe the mechanisms of the important experimental observations upon activation of [V2O5(OCH3)]− (7). These observations included (i) oxidation of the methoxo ligand to formaldehyde, (ii) the consecutive loss of H2 followed by CO2, and (iii) the unexpected 16O/18O scrambling observed in loss of H2CO and CO2. Formaldehyde Loss Channel (Equation 3b). Methoxo activation and loss of formaldehyde were examined first, starting from the accepted mechanisms for the mononuclear gas-phase and bulk surface catalysts. These mechanisms both involve HAT from a methoxo ligand to a terminal oxo ligand of the same vanadium center, producing a ketyl radical ligand and a hydroxo ligand.30 The product of HAT could contain two unpaired electrons, and therefore, both the singlet and triplet products and transition states were considered. In addition to the terminal oxo ligand, equivalent to the mononuclear system, the neighboring vanadium of [V2O5(OCH3)]− introduced two additional oxo ligands to which HAT might occur, that is, a bridging oxo ligand and a terminal oxo ligand of the neighboring vanadium center. Accordingly, six transition states were examined, arising from hydrogen atom transfers to each of the three possible oxo ligands for both singlet and triplet states (Supporting Information Figure S10). Calculations predicted that the lowest-energy transition state was the triplet state for HAT to a terminal oxo of the same vanadium center (TS 7a− 7h), analogous to the previous mechanisms proposed for mononuclear centers. The next possible transition state, which was 0.14 eV higher in energy, is on the singlet surface, involving HAT to a terminal oxo of the other vanadium center. Although this might also occur to produce similar products, for simplicity, the lowest transition state involving a triplet state and an oxo ligand of the same vanadium center along with the similar structure transition statge on the singlet surface were considered further. The potential energy diagram of the proposed reactions of the HAT process involving the triplet transition state is in Figure 7(a). The product of initial HAT was [V2O4(OH)(OCH2)]− (7h); this was predicted to favor a triplet spin state. As a consequence, a singlet−triplet surface crossing was required to transform a singlet reactant (7a) into a triplet product (Figure 7a and b). This is a formally forbidden process that might occur in this case due to the presence of the heavier vanadium center, as suggested for other transition-metal oxides

Table 1. The m/z of Product Anions Generated from Fragmentation of [V2O5(OCH3)]− (7) with Labeled Methoxo Ligands product ion (neutral loss)

5a

The use of an 1 8 O-labeled methoxo ligand in [V2O5(18OCH3)]− (m/z 215) revealed unexpected 16O/18O exchange prior to decomposition for eqs 3b and 4 (Figure S2 (Supporting Information) and Table 1). Although is was not possible to obtain quantitative ratios for these exchange processes due to subsequent ion−molecule reactions (cf. Scheme 2), the combined losses of H2C16O (eq 5a) and H2C18O (eq 5b) and C16O18O (eq 6a) and C16O2 (eq 6b) following H2 elimination clearly highlight that 16O/18O exchange between the methoxo ligand and the vanadium oxide occurs rapidly prior to subsequent fragmentation of the ligand. Thus, any mechanisms for the fragmentation of the G



Article

Theoretical calculations predicted that the two unpaired electrons of triplet [V2O4(OH)(OCH2)]− (7h) were located on the ketyl ligand (with some interaction with a vanadium d orbital) and on the vanadium center (Figure 7c). This is consistent with [V2O4(OH)(OCH2)]− (7h) being described as containing a radical ketyl ligand and a single V(IV) center (as well as a second V(V) center). In the triplet product [V2O4(OH)]− (m/z 183) (13a), the two unpaired electron were located on a V(III) center neighboring a VV center with a hydroxo ligand (Figure 7c). However, an analysis of the singlet transition state TS 7a−7i also highlighted in Figure 7, shows that the reaction requires more energy to occur. Hence, transition state TS 7a−7i is 0.43 eV higher in energy than its corresponding transition state on the triplet surface (TS 7a− 7h). Therefore, the reaction at the singlet surface is not expected to occur experimentally. Possible Isomers Relevant to H2 Elimination and 16O/18O Exchange. Various possible isomers and spin states of [V2, O6, C, H3]− (m/z 213) considered in Figure 3 might be involved in the H2 elimination (eq 3a) and 16O/18O exchange. Importantly, the relative energy of all of these isomers was significantly below that of the transition state for HAT in methoxo activation, suggesting that they might be accessible under the CID conditions. Mechanism of 16O/18O Exchange in the Formaldehyde Loss Channel (Equations 5a and 5b). The isomer [V2O 3(OH)(O2CH2)] (7d), containing two tetrahedral V(IV) centers and a bidentate methanediolate ligand, was related to [V2O4(OH)(OCH2)]− (7h) by attack of the ketyl radical at a neighboring oxo ligand. It is worth noting that related bidentate methanediolate ligands bridging two vanadium centers have been proposed as intermediates associated with the formation of byproducts in the oxidation of methanol over vanadium oxide catalysts.32 The equivalent nature of both oxygen groups in this ligand suggested that it was an obvious lead to explain the observed 16O/18O exchange (Figure S2, Supporting Information). The overall proposed mechanism is summarized in Scheme 3. For the experimentally observed 16O/18O exchange to be possible via this mechanism, all of the isomers and activation barriers would need to be lower than the competing formaldehyde dissociation pathway. Indeed, this was predicted to be the case, suggesting that this pathway provides a reasonable explanation for the observed 16O/18O exchange (Figure 8). Hence, from the intermediate structure 7h, a transition state TS 7h−7d was found whereby the CH2 of the ketyl is oriented toward an oxo ligand of the other vanadium center with an energy ∼0.24 eV lower than that for structure 13a. Intermediate 7d can rearrange via TS 7d−7j with an energy ∼0.20 eV lower than that of 13a to generate an isomer of 7 (i.e., 7j) where the oxygen of the ketyl was “swapped”. The new [V2O4(OH)(OCH2)]− (7j) can rearrange further to generate yet another isomer of [V2O4(OH)(OCH2)]− (7b). The energies of all of these processes fall below that of 13a.

Figure 7. (a) B3LYP/SDD6-311++G(d,p) calculated potential energy diagram for the loss of formaldehyde from [V2O5(OCH3)]− (7a) via HAT to an oxo ligand of the same vanadium center. Energies for reactants, transition states, and products on both the singlet (blue) and triplet (red) diagrams are given. (b) Structures, spin values, and energies relative to singlet [V2O5(OCH3)]− (7a) (eV) are given for transition states and products. Imaginary frequencies (cm−1) for the transition states are listed. (c) Singly occupied molecular orbitals of (A) [V2O4(OH)(OCH2)]− and (B) [V2O5(OCH3)]−.

species.31 The ketyl ligand could then dissociate as formaldehyde, to yield the product anion [V2O4(OH)]− (13a), which was also predicted to favor a triplet state. The energy of this formaldehyde dissociation pathway set an expected maximum energy level for other competing pathways that were also observed experimentally. Scheme 3. Mechanism for 16O/18O Exchange

H



Article

the competing formaldehyde loss pathway, depicted by the dotted line in Figure 9, consistent with it also being observed experimentally along with H2CO loss (Figure 9).

Figure 9. (a) B3LYP/SDD6-311++G(d,p) calculated potential energy diagram for the loss of H2 from the binuclear anion [V2O5(OCH3)]− via the intermediate [V2O3(μ2-O2CH2)]−. The energy level of the competing formaldehyde dissociation pathway is shown in blue. (b) Structures, spin values, and energies relative to singlet [V2O5(OCH3)]− (7a) (eV) are given for the transition states, intermediate, and product.

Figure 8. (a) B3LYP/SDD6-311++G(d,p) calculated potential energy diagram for the mechanism for 16O/18O exchange and carbon mobility via intermediate [V2O4(OH)(OCH3)]−. The energy level of the competing formaldehyde dissociation pathway is shown in blue. (b) Structures, spin values, and energies relative to singlet [V2O5(OCH3)]− (7a) (eV) are given for transition states and intermediates. Imaginary frequencies (cm−1) for the transition states are listed.

Mechanism of the CO2 Loss (Equation 4). Experimental results indicated that the product [V2O4(O2CH)] (m/z 211) (12a) underwent further loss of CO233 to yield the product [V2O4(H)]− (m/z 167) (14). The mechanism of this reaction was predicted to begin with the bidentate formate ligand cleaving at one V−O bond and rotating 12f to allow hydride transfer to the vanadium center (TS 12f−14). The loosely bound CO2 would then dissociate to yield the product anion triplet [V2O4(H)]− (m/z 167) (14a) (Figure 10). As the gasphase elimination of H2 was an irreversible reaction, the further reactivity to dissociate CO2 was not in competition with the much lower energy formaldehyde dissociation pathway. However, the endothermicity of the reaction is quite significant in comparison to the stable [V2O4(O2CH)] (12a) or HAT activation of the methoxo ligand (TS 7a−7h). The experimental observation of CO2 loss might be rationalized by the very similar masses of [V2O5(OCH3)]− (m/z 213) (7a) and [V2O4(O2CH)]− (m/z 211) (12a), with the latter ion potentially undergoing additional activation to provide the extra energy required for decarboxylation to yield [V2O4(H)]− (m/z 167) (14a). Discussion of the Key Intermediate. [V2O3(OH)(O2CH2)] (7d) was proposed as a key intermediate in the decomposition of [V 2 O 5 (OCH 3 )] − (m/z 213) (7a). It contained a

Mechanism of the H2 Elimination Reaction (Equation 3a). The observed H2 elimination reaction suggested the formation of an isomer with a hydride ligand and a hydroxo ligand, which would combine to eliminate H2. The hydride-containing isomer [V2O3(H)(OH)(O2CH)] (7f) described above (Figure 3) was an obvious candidate. In addition, this isomer also contained a formate ligand, equivalent to that predicted for the product of H2 elimination [V2O4(O2CH)] (12a). Methanediolate ligands [H2CO2]2− and the formate ligand [HCO2]− are related by formal removal of hydride from the former to give the later (i.e., [H2CO2]2− → [HCO2]− + H−). This suggested that the hydride-containing isomer could be formed from [V2O3(OH)(O2CH2)] (7d) by hydride transfer from the methanediolate ligand to a vanadium center, producing the required formate and hydride ligands via a transition state TS 7d−7f requiring 0.59 eV to generate 7f. Elimination of H2 was predicted to occur from reaction of the hydride and hydroxo ligands of [V2O3(H)(OH)(O2CH)] (TS 7f−12a), yielding the formate-containing product anion [V2O4(O2CH)] (m/z 211) (12a). Each of the steps in this process was predicted to occur at lower energies than those of I



Article

cycle shown in Scheme 4, involving three steps, (i) reactions with methanol, (ii) CID to liberate oxidized products and Scheme 4. Gas-Phase Catalytic Cycle for the Oxidation of Methanol by [V2O5(OH)]−

reduce the vanadium oxide anion, and (iii) reoxidation of the reduced vanadium cluster. [V2O4(O2CH)]− (m/z 211) can generate [V2O4(H)]− (m/z 167) upon CID, which in turn can generate [V2O4(OH)]− (m/z 183) upon ion−molecule reactions with water, confirming that these intermediate species can also complete the catalytic cycle shown in Scheme 4.

■

Figure 10. (a) B3LYP/SDD6-311++G(d,p) calculated potential energy diagram for the loss of CO2 from [V2O4(μ2-O2CH)]−. The energy level of the formaldehyde dissociation pathway is shown in blue. (b) Structures, spin values, and energies relative to singlet [V2O5(OCH3)]− (7a) (eV) are given for the transition state, intermediate, and product.

CONCLUSIONS By studying the gas-phase reactions of the anions [VO3]−, [VO2(η2-O2)]−, [VO2(OH)2]−, and [V2O5(OH)]−, we have been able to examine the role of oxo, peroxo, and hydroxide groups and the effect of an adjacent vanadium center on the oxidation of methanol. Only [VO3]−, [VO2(η2-O2)]−, and [V2O5(OH)]− directly react with methanol to give the methoxides, which decompose in different ways. [VO2(OH)(OCH3)]− decomposes via water loss to give [VO2(η2OCH2)]− (step b of cycle 2 in Scheme 1). The methoxide ligand of [VO(O2)(OH)(OCH3)]− fragments via formaldehyde loss (step e of cycle 2 in Scheme 1). In contrast, the methoxide ligand of the binuclear [V2O5(OCH3)]− fragments via three competing pathways, including loss of H2 (eq 3a), formaldehyde (eq 3b), and H2 + CO2 (eqs 3a and 4). In both the mono- and binuclear systems, the subsequent reactions of the products of methoxide ligand decomposition can regenerate the starting catalysts, thereby producing the catalytic cycles shown in Schemes 1 and 4. These cycles highlight the incredible sensitivity of methanol reactivity toward the local chemical structure of the vanadium center(s). The cycle involving mononuclear sites is “cleaner”, producing formaldehyde via the following oxidation: 2 CH3OH + O2 → 2 H2CO + 2 H2O. In contrast, binuclear sites oxidize methanol via two different product channels. Thus, formaldehyde is produced via CH3OH + O → H2CO + H2O, while hydrogen is produced via CH3OH + O → 2H2 + CO2. This suggests that more than one type of catalytic cycle might occur on the surface of vanadium catalysts. The experimentally observed differences in reactivity of [VO2(OH)2]− (1) and [V2O5(OH)]− (2) toward methanol highlight the importance of bridging oxo ligands for reactivity, consistent with condensed-phase studies.29 This study supports

methanediolate ligand bridging the two vanadium centers. This provided an explanation for the 16O/18O exchange observed in the CH318OH experiments; [V2O3(OH)(16OCH218O)]− (7d) could cleave at either of the equivalent V−O bonds, resulting in loss of either CH216O or CH218O. Furthermore, the methanediolate ligand of [V2O3(OH)(O2CH2)]− (7d) also facilitated dehydrogenation. In this case, hydride transfer from the methanediolate ligand to vanadium generated [V2O3(H)(OH)(O2CH)]− (7f) containing a bidentate formate ligand. This could then undergo loss of H2 to give the product [V2O4(O2CH)]− (12a). This dehydrogenation provides a plausible mechanism that might be considered for some of the complex dehydrogenation reactions observed for organic substrates on vanadium centers. Catalytic Cycle for the Oxidation of Methanol Catalyzed by [V2O5(OH)]−. As mentioned earlier, ion trap mass spectrometers have been used previously to uncover catalytic cycles, including the oxidation of methanol.16 In this work, the reduced vanadium oxide anions, [V2O4(OH)]− and [V2O4(H)]−, formed upon loss of formaldehyde and the combined losses of H2 and CO2, could be reoxidized to their active forms by introducing oxidants into the ion trap. Indeed, we have found that [V2O4(OH)]− (m/z 183) (13a) can be successfully oxidized by the oxidants O2, CO2, N2O, and MeNO2 (data not shown) to regenerate [V2O5(OH)]− (m/z 199) (2). These reoxidations of the reduced [V2O4(OH)]− suggest the possibility of generating the gas-phase catalytic J



■

the relevance of the mononuclear mechanisms proposed for methoxo activation and formaldehyde dissociation in the presence of neighboring vanadium oxide centers. Calculations predicted that the methoxo ligand of [V2O5(OCH3)]− (7a) was activated by HAT to a terminal oxo ligand, yielding the anion [V2O4(OH)(OCH2)]− (17h). The ketyl ligand of [V2O4(OH)(OCH2)]− (17h) could then dissociate as formaldehyde, in a reaction analogous to that proposed for the mononuclear systems.31 A key finding was a novel 16O/18O exchange process between the methoxo ligand and vanadium oxide prior to loss of formaldehyde, with implications for the interpretation of isotope labeling studies on vanadium oxide surfaces. This scrambling process was not observed for the mononuclear system,16i,j DFT calculations suggested that the key intermediate of this process was [V2O3(OH)(O2CH2)]− (7d), containing a bidentate methanediolate ligand with two equivalent V−O bonds. The proposed intermediates and mobility of the surface “CH2” group around the cluster may be relevant to the formation of side products observed in industrial processes, that is, dimethyl ether, methyl formate, and dimethoxymethane.32 Interestingly, the same intermediate [V2O3(OH)(O2CH2)]− (7d) is responsible for H2 loss from the methoxide ligand. The importance of this intermediate suggests that it should be considered when interpreting condensed-phase processes particularly where dehydrogenation of alkoxide intermediates is relevant (e.g., the oxidation of butane to maleic anhydride). Finally, given that a wide range of polynuclear vanadium oxide cluster anions can be formed via ESI/MS, it is will be interesting to examine their reactivity toward methanol. Preliminary studies into the reactivity of the larger cluster anions [V3O8]− and [V4O10(OH)]− toward methanol suggests differences to those of [VO2(OH)2]− (1) and [V2O5(OH)]− (2) described above. Thus, [V3O8]− (m/z 281) did not eliminate water after methanol addition and instead formed [V3O7(OH)(OCH3)]−. In contrast, [V4O10(OH)]− (m/z 381) reacted in an equivalent manner to the binuclear anion to form [V4O10(OCH3)]− via loss of water. Decomposition of the product anion, [V4O10(OCH3)]− under CID, suggested that it was a more selective catalyst than [V2O5(OH)]−, with only a single pathway, involving formaldehyde dissociation, being experimentally observed. Importantly, H2 loss was not observed, in contrast to [V2O5(OH)]− (2). Accordingly, the reaction of [V4O10(OH)]− with CH318OH (and the potential 16 O/18O exchange) could be used to probe the similarities of decomposition between the binuclear and tetranuclear anions. Such studies are underway.

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (G.N.K.); rohair@unimelb. edu.au (R.A.J.O.). Fax: +613 9347-5180. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank the Australian Research Council for financial support via Grants to T.W. (#DP0772053), G.N.K. (#DP1096134), and R.A.J.O. (#DP0558430 and the ARC CoE Program). B.L.H., T.W., and G.N.K. thank the Victorian Institute for Chemical Sciences High Performance Computing Facility for allocation of computer time. Dedicated to our friend and colleague Peter B. Armentrout on the occasion of his 60th birthday and in recognition of his outstanding work in metal-mediated chemistry, which has been an ongoing source of inspiration.

■

REFERENCES

(1) (a) Centi, G.; Cavani, F.; Trifirò, F. Selective Oxidation by Heterogeneous Catalysis; Kluwer Academic: New York, 2001; pp 1−24; (b) Thomas, J. M.; Thomas, W. J. Principles and Practice of Heterogeneous Catalysis; Wiley-VCH: Weinheim, Germany, 1996; pp 1−64. (2) (a) Centi, G.; Perathoner, S. Int. J. Mol. Sci. 2001, 2, 183−196. (b) Grasselli, R. K. Top. Catal. 2002, 21, 79−88. (3) Bond, G. C.; Tahir, S. F. Appl. Catal. 1991, 71, 1−31. (4) (a) Hess, C. Highly dispersed vanadium oxide catalysts. In Nanostructured Catalysts: Selective Oxidation; Hess, C., Schlögl, R., Eds.; RSC Publishing: Cambridge, U.K., 2011; pp 299−325. (b) Appl. Catal., A 1997, 157, 1−425 (Special issue devoted to vanadia catalysts for selective oxidation of hydrocarbons and their derivatives). (5) Weckhuysen, B. M. Catal. Today 2003, 78, 25−46. (6) Dyrek, K.; Che, M. Chem. Rev. 1997, 97, 305−331. (7) Grasselli, R. K. Top. Catal. 2002, 21, 79−88. (8) Vedrine, J. C. Top. Catal. 2002, 21, 97−106. (9) Ozkan, U. S.; Watson, R. B. Catal. Today 2005, 100, 101−114. (10) Bañares, M. A. Top. Catal. 2009, 52, 1301−1302. (11) Zemski, K. A.; Justes, D. R.; Castleman, A. W., Jr. J. Phys. Chem. B 2002, 106, 6136−6148. (12) (a) Böhme, D. K.; Schwarz, H. Angew. Chem., Int. Ed. 2005, 44, 2336−2354. (b) O’Hair, R. A. J.; Khairallah, G. N. J. Cluster Sci. 2004, 3, 331−363. (13) Zhao, Y.-X.; Wu, X.-N.; Ma, J.-B.; He, S.-G.; Ding, X.-L. Phys. Chem. Chem. Phys. 2011, 13, 1925−1938. (14) A recent review has highlighted studies on the oxidation of methane; see: (a) Schwarz, H. Angew. Chem., Int. Ed. 2011, 50, 10096−10115. For selected studies on the reactions of metal oxide ions with alkanes, see: (b) Zemski, K. A.; Justes, D. R.; Castleman, A. W., Jr. J. Phys. Chem. A 2001, 105, 10237−10245. (c) Engeser, M.; Schlangen, M.; Schröder, D.; Schwarz, H.; Yumura, T.; Yoshizawa, K. Organometallics 2003, 22, 3933−3943. (d) Feyel, S.; Schröder, D.; Schwarz, H. J. Phys. Chem. A 2006, 110, 2647−2654. (e) Feyel, S.; Döbler, J.; Schröder, D.; Sauer, J.; Schwarz, H. Angew. Chem., Int. Ed. 2006, 45, 4681−4685. (f) Feyel, S.; Schröder, D.; Rozanska, X.; Sauer, J.; Schwarz, H. Angew. Chem., Int. Ed. 2006, 45, 4677−4681. (g) Cassady, C. J.; McElvany, S. W. Organometallics 1992, 11, 2367−2377. (h) Kretzschmar, I.; Fiedler, A.; Harvey, J. N.; Schröder, D.; Schwarz, H. J. Phys. Chem. A 1997, 101, 6252−6264. (i) Dinca, A.; Davis, T. P.; Fisher, K. J.; Smith, D. R.; Willett, G. D. Int. J. Mass Spectrom. 1999, 182/183, 73−84. (15) For selected studies on the reactions of metal oxide ions with alkenes, see: (a) Bell, R. C.; Zemski, K. A.; Kerns, K. P.; Deng, H. T.; Castleman, A. W., Jr. J. Phys. Chem. A 1998, 102, 1733−1742. (b) Bell, R. C.; Castleman, A. W., Jr. J. Phys. Chem. A 2002, 106, 9893−9899. (c) Justes, D. R.; Mitric, R.; Moore, N. A.; Bonacic-Koutecky, V.;

ASSOCIATED CONTENT

S Supporting Information *

Full citation details for ref 24. Table S1: rate constant measurements; Figures S1−S9: mass spectra showing ion− molecule reactions with isotopically labeled methanol; Figure S10: singlet versus triplet transition states for formaldehyde loss. Cartesian coordinates, energies, and vibrational frequencies for reactants, intermediates, products, and transition states. This material is available free of charge via the Internet at http://pubs.acs.org. K



Article

Castleman, A. W., Jr. J. Am. Chem. Soc. 2003, 125, 6289−6299. (d) Moore, N. A.; Mitric, R.; Justes, D. R.; Bonacic-Koutecky, V.; Castleman, A. W., Jr. J. Phys. Chem. B 2006, 110, 3015−3022. (16) For selected studies on the reactions of metal oxide ions with alcohols, see: (a) Justes, D. R.; Moore, N. A.; Castleman, A. W., Jr. J. Phys. Chem. B 2004, 108, 3855−3862. (b) Cao, Y.; Zhao, X.; Xin, B.; Xiong, S.; Tang, Z. J. Mol. Struct.: THEOCHEM 2004, 683, 141−146. (c) Engeser, M.; Schröder, D.; Schwarz, H. Chem. Eur. J. 2005, 11, 5975−5987. (d) Fialko, E. F.; Kikhtenko, A. V.; Goncharov, V. B.; Zamaraev, K. I. J. Phys. Chem. B 1997, 101, 5772−5773. (e) Fialko, E. F.; Kikhtenko, A. V.; Goncharov, V. B. Organometallics 1998, 17, 25− 31. (f) Oliveira, M. C.; Marcalo, J.; Vieira, M. C.; Ferreira, M. A. A. Int. J. Mass Spectrom. 1999, 187, 825−835. For two catalytic cycles involving the oxidation of methanol, see: (g) Waters, T.; O’Hair, R. A. J.; Wedd, A. G. Chem. Commun. 2000, 225−226. (h) Waters, T.; O’Hair, R. A. J.; Wedd, A. G. J. Am. Chem. Soc. 2003, 125, 3384−3396. (i) Waters, T.; Khairallah, G. N.; Wimala, S. A. S. Y.; Ang, Y. C.; O’Hair, R. A. J.; Wedd, A. G. Chem. Commun. 2006, 4503−4505. (j) Waters, T.; Wedd, A. G.; O’Hair, R. A. J. Chem. Eur. J. 2007, 13, 8818−8829. (17) For decomposition of metal complexes with “bare” methoxo ligands, see: (a) Schlangen, M.; Schwarz, H. Chem. Commun. 2010, 46, 1878−1880. (b) Schlangen, M.; Schwarz, H. ChemCatChem 2010, 2, 799−802. (18) For a review, see: Waters, T.; O’Hair, R. A. J. Metal Ion Complexes: Formation and Reactivity. In The Encyclopedia of Mass Spectrometry: Fundamentals of and Applications to Organic (and Organometallic) Compounds; Gross, M. L., Caprioli, R., Nibbering, N. M. M., Eds.; Elsevier: Amsterdam, The Netherlands, 2005; Vol. 4, Chapter 6, pp 604−612. (19) (a) Burcham, L. J.; Briand, L. E.; Wachs, I. E. Langmuir 2001, 17, 6175−6184. (b) Badlani, M.; Wachs, I. E. Catal. Lett. 2001, 75, 137−149. (20) (a) McCarron, E. M., III; Harlow, R. L.; Li, Z. G.; Suto, C.; Yuen, Y. J. Solid State Chem. 1998, 136, 247−252. (b) Ardissone, D. E.; Valente, N. G.; Cadús, L. E.; Arrúa, L. A. Ind. Eng. Chem. Res. 2000, 39, 2902−2909. (c) Faliks, A.; Yetter, R. A.; Floudas, C. A.; Bernasek, S. L.; Fransson, M.; Rabitz, H. J. Phys. Chem. A 2001, 105, 2099−2105. (d) Burcham, L. J.; Briand, L. E.; Wachs, I. E. Langmuir 2001, 17, 6164−6174. (e) Farneth, W. E.; Ohuchi, F.; Staley, R. H.; Chowdhry, U.; Sleight, A. W. J. Phys. Chem. 1985, 89, 2493−2497. (f) Farneth, W. E.; Staley, R. H.; Sleight, A. J. Am. Chem. Soc. 1986, 108, 2327−2332. (21) Deo, G.; Wachs, I. E. J. Catal. 1994, 146, 323−334. (22) (a) Reid, G. E.; O’Hair, R. A. J.; Styles, M. L.; McFadyen, W. D.; Simpson, R. J. Rapid Commun. Mass Spectrom. 1998, 12, 1701−1708. (b) O’Hair, R. A. J. Chem. Commun. 2006, 1469−1481. (c) Gronert, S. J. Am. Soc. Mass Spectrom 1998, 9, 845−848. (23) (a) Su, T.; Bowers, M. T. Int. J. Mass Spectrom. Ion Phys. 1973, 12, 347−356. (b) Lim, K. F. Quantum Chem. Program Exch. 1994, 14, 3−13. (24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; et al., Gaussian 03, revision B.04l Gaussian, Inc.L Wallingford, CT, 2004. (25) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (b) Lee, C.; Yang, W.; G. Parr, R. Phys. Rev. B 1988, 37, 785−789. (26) (a) Dolg, M.; Wedig, U.; Stoll, H.; Preuss, H. J. Chem. Phys. 1987, 86, 866−872. (b) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213−222. (c) Clark, T.; Chandrasekhar, J.; Schleyer, P. v. R. J. Comput. Chem. 1983, 4, 294−301. (d) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650−654. (e) Gill, P. M. W.; Johnson, B. G.; Pople, J. A.; Frisch, M. J. Chem. Phys. Lett. 1992, 197, 499−505. (27) (a) Hack, M. D.; Maclagan, R. G. A. R.; Scuseria, G. E.; Gordon, M. S. J. Chem. Phys. 1996, 104, 6628−6630. (b) Vyboishchikov, S. F.; Sauer, J. J. Phys. Chem. A 2000, 104, 10913−10922. (c) Asmis, K. R.; Santambrogio, G.; Bruemmer, M.; Sauer, J. Angew. Chem., Int. Ed. 2005, 44, 3122−3125.

(28) (a) Zhang, X.; Schwarz, H. Chem. Eur. J. 2010, 16, 1163−1167. (b) Vyboishchikov, S. F.; Sauer, J. J. Phys. Chem. A. 2001, 105, 8588− 8598. (c) Asmis, K. R.; Sauer, J. Mass Spectrom. Rev. 2007, 26, 542− 562. (29) (a) Weckhuysen, B. M.; Keller, D. E. Catal. Today 2003, 78, 25−46. (b) Tian, H.; Ross, E. I.; Wachs, I. E. J. Phys. Chem. B 2006, 110, 9593−9600. (c) Limberg, C. Eur. J. Inorg. Chem. 2007, 3303− 3314. (d) Hoppe, E.; Limberg, C.; Ziemer, B. Inorg. Chem. 2006, 45, 8308−8317. (30) Dobler, J.; Pritzsche, M.; Sauer, J. J. Am. Chem. Soc. 2005, 127, 10861−10868. (31) (a) Schroeder, D.; Schwarz, H.; Shaik, S. Struct. Bonding (Berlin) 2000, 97, 91−123. (b) Schwarz, H. Int. J. Mass Spectrom. 2004, 237, 75−105. (32) (a) Forzatti, P.; Tronconi, E.; Elmi, A. S.; Busca, G. Appl. Catal. 1997, 157, 387−408. (b) Tatibouet, J. M.; Lauren-Pernot, H. J. Mol. Catal. A: Chem. 2001, 171, 205−216. (33) For examples of gas-phase decarboxylation of metal carboxylate ions, see: (a) Fiedler, A.; Schröder, D.; Zummack, W.; Schwarz, H. Inorg. Chim. Acta 1997, 259, 227−235. (b) Bachrach, S. M.; Hare, M.; Kass, S. R. J. Am. Chem. Soc. 1998, 120, 12646−12649. (c) O’Hair, R. A. J. Chem. Commun. 2002, 20−21. (d) O’Hair, R. A. J.; Vrkic, A. K.; James, P. F. J. Am. Chem. Soc. 2004, 126, 12173−12183. (e) Khairallah, G. N.; O’Hair, R. A. J. Int. J. Mass Spectrom. 2006, 254, 145−151. (f) Rijs, N. J.; Khairallah, G. N.; Waters, T.; O’Hair, R. A. J. J. Am. Chem. Soc. 2008, 130, 1069−1079. (g) Rijs, N. J.; O’Hair, R. A. J. Organometallics 2009, 29, 2684−2692. (h) Khairallah, G. N.; Thum, C.; O’Hair, R. A. J. Organometallics 2009, 28, 5002−5011. (i) Rijs, N. J.; Sanvido, G. B.; Khairallah, G. N.; O’Hair, R. A. J. Dalton Trans. 2010, 39, 8655−8662. (j) Meyer, M. M.; Chan, B.; Radom, L.; Kass, S. R. Angew. Chem., Int. Ed. 2010, 49, 5161−5164. (k) Attygale, A. B.; Chan, C.-C.; Axe, F. U.; Bolgar, M. J. Mass Spectrom. 2010, 45, 72−81. (l) Chan, C.-C.; Axe, F. U.; Bolgar, M.; Attygale, A. B. J. Mass Spectrom. 2010, 45, 1130−1138. (m) Butschke, B.; Schwarz, H. Organometallics 2010, 29, 6002−6011. (n) Woolley, M.; Khairallah, G. N.; Donnely, P. S.; O’Hair, R. A. J. Rapid Commun. Mass Spectrom. 2011, 25, 2083−2088. (o) Leeming, M. G.; Khairallah, G. N.; da Silva, G.; O’Hair, R. A. J. Organometallics 2011, 30, 4297−4307. (p) Brunet, C.; Antoine, R.; Broyer, M.; Dugourd, P.; Kulesza, A.; Petersen, J.; Röhr, M. I. S.; Mitrić, R.; Bonačić-Koutecký, V.; O’Hair, R. A. J. J. Phys. Chem. A 2011, 115, 9120−9127. (q) Rijs, N. J.; Yoshikai, N.; Nakamura, E.; O’Hair, R. A. J. J. Am. Chem. Soc. 2012, 134, 2569− 2580. (r) Rijs, N. J.; O’Hair, R. A. J. Dalton Trans. 2012, 41, 3395− 3406. (s) Sraj, L. O.; Khairallah, G. N.; da Silva, G.; O’Hair, R. A. J. Organometallics 2012, 31, 1801−1807. (t) Röhr, M. I. S.; Petersen, J.; Brunet, C.; Antoine, R.; Broyer, M.; Dugourd, P.; Bonačić-Koutecký, V.; O’Hair, R. A. J.; Mitrić, R. J. Phys. Chem. Lett. 2012, 3, 1197−1201.

L


Gas-Phase Reactions of [VO2(OH)2]â and [V2O5 ... - ACS Publications

Recommend Documents

Gas-Phase Reactions of [VO2(OH)2]â and [V2O5 ... - ACS Publications