Simple Relationship between Oxidation State and Electron Affinity in

Article pubs.acs.org/JPCA

Simple Relationship between Oxidation State and Electron Affinity in Gas-Phase Metal−Oxo Complexes Sarah E. Waller,† Manisha Ray,† Bruce L. Yoder,‡ and Caroline Chick Jarrold*,† †

Department of Chemistry, Indiana University, 800 East Kirkwood Avenue, Bloomington, Indiana 47405, United States Laboratory for Physical Chemistry, ETH Zürich, Wolfgang-Pauli-Strasse 10, CH-8093 Zürich, Switzerland

‡

S Supporting Information *

ABSTRACT: The photoelectron spectra of WO3H− and WO2F− are presented and analyzed in the context of a series of previous similar measurements on MOy− (M = Mo, W; y = 0−3), MO4H− and AlMOy− (y ≤ 4) complexes. The electronic structures of the WO3H and WO2F anion and neutral complexes were investigated using the B3LYP hybrid density functional method. The spectra of WO3H−, WO2F−, and previously measured AlWO3− photoelectron spectra show that the corresponding neutrals, in which the transition metal centers are all in a +5 oxidation state, have comparable electron affinities. In addition, the electron affinities fit the general trend of monotonically increasing electron affinity with oxidation state, in spite of the WO3H−, WO2F−, and AlWO3− having closed shell ground states, suggesting that the oxidation state of the metal atom has more influence than shell closing on the electron affinity of these transition metal−oxo complexes. Results of DFT calculations suggest that the neutrals are pyramidal and the anions are planar. However, the barriers for inversion on the neutral surface are low, and attempts to generate simple Franck−Condon simulations based on simple normal coordinate displacement, ignoring the effects of inversion, are inadequate.

■

INTRODUCTION The electron affinity (EA) of a neutral is a direct measure of the relative energy between the neutral and its associated anion. Electronic shell closings or full orbital occupancy is associated with enhanced stability. A closed shell neutral will necessarily have an open shell anion, which typically results in small or negative EA’s. For example, closed-shell simple organic molecules generally have negative EA’s.1 On the other hand, organic radicals can have fairly high EA’s because the addition of an electron results in a relatively stable closed shell anion.2 Metal−oxo clusters tend to have positive EA’s, whether they are open or closed shell. Given the ionic character of metal oxide bonds, it follows that metal oxide species in stoichiometric oxidation states should have higher EA values than those in lower oxidation states, though it is not an absolute rule.3 How the variation in EA with oxidation state might be counteracted by changes in the orbital occupancy of anions relative to neutrals is difficult to measure in a perfectly systematic way; addition or removal of an O-atom from an MOy or MOy− (M = transition metal) species changes the number of electrons on the metal center by ±2, so an open shell metal-oxo anion will be open shell regardless of the value of y. A different type of ligand (e.g., OH, F, CN) is required to change the oxidation state of the metal center by ±1 increments, in which case the chemical environment of the metal center may not be directly comparable to a pure metal−oxo cluster. Research in our group has focused on the electronic structure of transition metal−oxo cluster anions and their associated © 2013 American Chemical Society

neutrals accessed via anion photodetachment. For clusters with more than one metal center and varying oxygen atoms, we have determined how cluster structure affects EA based on differences in stabilizing effects of structural motifs in anions versus neutrals.4−6 For species with a single transition metal atom, the manner in which the metal center can bond with an O-atom is limited, so structural effects no longer come into play. For a group 6 transition metal center such as W, the anions, W−, WO−, WO2−, and WO3− have W oxidation states of −1, +1, +3, and +5, respectively. All anions in this series have an odd number of electrons and are necessarily open shell. Therefore, though measuring the photodetachment of these spectra will show how EA varies with oxidation state, it fails to measure the impact of the even/odd electron count. The results of more recent studies on AlMoOy− and AlWOy− suggested that the electronic structures of the anions could be described as Al+[MOy]−2 for M = Mo7 and W.8 These complexes therefore have the same MO bonding environments as pure MOy−/MOy molecules, but the anions have an extra electron, which results in an even electron count, and their associated neutrals are necessarily open-shell. On the other hand, perturbation from the nearby Al+ center may Special Issue: Terry A. Miller Festschrift Received: October 1, 2013 Revised: October 29, 2013 Published: October 29, 2013 13919

dx.doi.org/10.1021/jp4097666 | J. Phys. Chem. A 2013, 117, 13919−13925

The Journal of Physical Chemistry A

Article

σ ⎡ ⎛3 1 ⎞⎤ ∂σ = total ⎢1 + β(E)⎜ cos2 θ − ⎟⎥ ⎝2 4π ⎣ 2 ⎠⎦ ∂Ω

influence the EA because of enhanced Coulomb stabilization in the anion relative to the neutral. In the present study, we substitute both a hydroxyl and a fluoride ligand for an oxo group on WO3− to form WO3H− and WO2F−, both of which are closed shell anions with the central metal center nominally in a +4 oxidation state. Calculations predict that both anions have singlet ground states, with triplet states lying significantly higher in energy. However, as will be described below, the stability of the closed-shell anion state does not appear to result in deviation from a trend of increasing EA with W oxidation state from +4 to +5 to +6. It therefore appears that the EA is simply related to the oxidation state of the metal center. We further examine the effects of neutrals with low inversion barriers on the appearance of the photoelectron spectra.

where β(E) is the asymmetry parameter, determined experimentally for various transitions from photoelectron intensity measurements using the equation β (E ) =

I0 − I90 1 I + I90 2 0

Whereas the correlation between β(E) and polyatomic molecular orbitals is complex, the asymmetry parameter is sometimes a useful tool in the analysis of complex PE spectra. Computational Methods. Calculations on WO3H and WO2F anions and neutrals were done using the Gaussian 09 program suite for electronic structure calculations.14 The B3LYP hybrid density functional method15 was used because of previous success in treating MoO3−/MoO3.16 The 60 core electrons of W are replaced with Stuttgart−Dresden relativistic pseudopotentials. The 14 valence and outer core electrons were described by the associated double-ζ basis set.17 These calculations were used to screen the viability of structures, after which triple-ζ quality results were obtained by using the aug-cc-pVTZ basis set for oxygen, fluorine, and hydrogen atoms and by adding two f-type functions (ζf = 0.256, 0.825), and one g-type function (ζg = 0.627) to the tungsten basis set.18 Because the anion wave function is expected to be more spatially extended than the neutral, diffuse s-, p-, and d-functions were added to the tungsten basis set using an exponent ratio of 0.3, as done in previous computational studies on transition metal suboxide systems.19 Frequency calculations were performed to confirm that minima were found. Singlet, triplet, quintet, and septet spin states were considered for the anions, and doublet, quartet, sextet, and octet spin states were considered for the neutrals. Because the experimental spectroscopic method involves anion to neutral transition energies, adiabatic detachment energies (ADE) were determined from the energy difference between the zero-point corrected energy of a particular anion and a one-electron accessible neutral, whereas the vertical detachment energies (VDE) were determined from the difference between the total energy of an anion and a single point energy calculated for a one-electron accessible neutral confined to the optimized structure of the anion. These values can be compared directly to experimental spectra: the ADE corresponds to the origin of the anion to neutral transition (and also the adiabatic electron affinity, or EA, of the neutral). The VDE corresponds to the energy at which the anion to neutral electronic transition has maximum intensity; so, the difference between the ADE and VDE values gives a measure of how extended an electronic transition is predicted to be. Results of the ADE/VDE calculations (below) suggested that the neutral may not be appropriately described by a single-well potential; so single point calculations were carried out for a range of structures on the neutral surface. Attempts to simulate spectra assuming single well potentials for the anion and neutral were made using home written codes (one code to properly align the anion and neutral Cartesian coordinates from the Gaussian output files and another to calculate mode displacements using the structures and vibrational frequencies obtained from the Gaussian output files) in conjunction with PESCAL,20 which calculates the Franck−Condon factors for individual

■

METHODS Experimental Methods. The experimental apparatus and procedures used in these studies have been described in detail previously.9 Briefly, isotopically pure 186W powder (Oak Ridge National Laboratory, Isotope Business Office) is compressed into a disk. To facilitate target disk formation, small amounts (1−2 mol %) of 98Mo (Trace Sciences International) are added to the 186W powder used to make the disk. WxOy− (x = 1−4; y ≤ 3x) clusters are generated in a pulsed laser ablation/pulsed molecular beam source.10 Exposing the clusters to a methanol/ He mixture [30 psi (gauge) backing pressure] in a room temperature high-pressure fast flow reactor resulted in the production of WO3H−, among other reaction products. Clusters exposed to a trace amount of ArF laser premix (in Ne buffer gas, Spectra Gases, Inc.) in the He carrier gas formed WO2F−, among other products. The mass distributions of the cluster anion precursors and subsequently generated reaction products were measured in a 1.2 m time-of-flight mass spectrometer. Prior to colliding with the dual microchannel plate ion detector situated at the end of the ion drift tube, the anions of interest were selectively photodetached using the third harmonic output of a Nd:YAG laser (355 nm, 3.49 eV). The drift times of a small fraction of the resulting photoelectrons that traveled the length of a 1 m field-free tube situated perpendicular to the laser and ion beam axes and collided with a 40 mm dual microchannel plate detector were recorded using a digitizing oscilloscope. Drift times were converted to electron kinetic energy (e−KE), calibrated against the well-known photoelectron (PE) spectra of O−, OH−, and WO2−.11 The spectra included here show electron counts plotted as a function of electron binding energy (e−BE): e−BE = hν − e−KE

which is independent of photon energy, and directly reflects the relative energies of the anion and neutral states. The spectra were collected over 2 240 000 laser shots for WO3H− and 240 000 laser shots for WO2F−. The resolution of the apparatus is 7 meV for an electron with 0.45 eV e−KE and deteriorates with (e−KE)3/2. Spectra were measured with the laser polarization both parallel to (θ = 0°) and perpendicular to (θ = 90°) the direction of the electron drift path. In atomic12 and simple molecular systems,13 the angular distribution of photoelectrons can be easily related to the symmetry of the orbital associated with electron detachment and follows the expression 13920

dx.doi.org/10.1021/jp4097666 | J. Phys. Chem. A 2013, 117, 13919−13925

The Journal of Physical Chemistry A

Article

used to collect the spectrum;21 thus only WO3H− is detached. Both spectra feature one distinct electronic transition. The electronic transition in the WO3H− PE spectrum is significantly broader than the transition observed in the WO2F− spectrum, and the origin of the former difficult to identify. We therefore place very large uncertainty on the 2.20(15) eV value. The band reaches maximum intensity at 2.54(3) eV. The shoulders on the high e−BE or low e−BE side of the band do not form an evenly spaced progression. The electron signal obtained with laser polarization perpendicular to the detected electron trajectory (shown in the Supporting Information) was significantly lower than that obtained with parallel polarization. The asymmetry parameter was determined to be 1.1(1), which is consistent with p-wave photoelectron detachment.13 The PE spectrum of WO2F− features one nearly vertical electronic transition. Because there are relatively few detachment channels compared to WO3H−, the WO2F− spectrum was acquired with significantly fewer laser shots, as noted in the Experimental Methods. The most intense peak by far is at 2.658(5) eV, and it is 0.024 eV full-width at half-maximum. Low-intensity partially resolved peaks are found to lower binding energy with a 167(10) cm−1 spacing, and several peaks are found at 680(50) and 1024(30) cm−1 to higher binding energy. As with WO3H−, the spectrum obtained with laser polarization perpendicular to the direction of electron detection (shown in the Supporting Information, SI 3) was much lower than with parallel polarization, and the asymmetry parameter was determined to be 1.1(1) for this detachment process, as well.13 Results of calculations on possible structures of both anions and neutrals suggest that the lowest energy structures for both WO3H− and WO2F− are planar, with WO3H− in a Cs 1A′ state, and WO2F− in an analogous C2v 1A1 electronic state. Figure 2a,b shows the lowest energy structures and their relative energies determined computationally for the two most favored spin states found for the anions and neutrals of WO3H and WO2F, respectively. All other states and structures are predicted to be at least 3 eV higher in energy relative to the

vibronic transitions and convolutes them with Gaussian functions.

Figure 1. PE spectra of (a) WO3H− and (b) WO2F− obtained using 3.49 eV photon energy with laser polarization parallel to the direction of electron detection. The red and blue lines indicate the calculated ADE and VDE values (listed in Table 1).

■

RESULTS AND ANALYSIS Parts a and b of Figure 1 show the PE spectra of WO3H− and WO2F−, respectively, collected with 3.49 eV photon energy. WO3H− and WO3− ions overlap slightly in the ion beam (mass spectra are included in the Supporting Information, SI 2). However, the EA of WO3 is greater than the photon energy

Figure 2. Calculated structures and relative energies of the most stable (a) WO3H and (b) WO2F anion and neutral structures. Red balls represent O, blue represents W, pale aquamarine represents F, and white represents H. 13921

dx.doi.org/10.1021/jp4097666 | J. Phys. Chem. A 2013, 117, 13919−13925

The Journal of Physical Chemistry A

Article

Table 1. Electronic Term Energies (Zero-Point Corrected) and Transition Energies of Computational Results on WO2F/ WO2F− and WO3H/WO3H− Including Electronic States and Relative Energiesa electronic state WO2F

relative energy T0 (eV)

relevant one-e− transition

ADE (eV)

VDE (eV)

expt ADE/VDE (eV)

A′ + e− ← 1A1 A′ + e− ← 3B1 4 B2 + e− ← 3B1

2.59 1.74 4.81

2.65 2.18 5.16

2.65(4)/2.69(1)

2

A + e− ← 1A′ A + e− ← 3A′ 4 A″ + e− ← 3A′

2.21 1.33 4.48

2.30 2.04 4.85

2.02(15)/2.54(3)

− 1

A1 B1

3

0 0.85

2

WO2F 2

A′ B2

4

WO3H−

1

A′ A′

3

0 3.07 0 0.88

2 2

WO3H 2

A′ A″

4

0 3.15

a ADEs and VDEs of relevant one electron transitions are also given along with the experimentally determined ADE and VDE of the metal oxo− monohydroxides.

Table 1. However, the calculated VDE value for the WO3H + e− ← WO3H−, particularly relative to the ADE value, is not consistent with the observed spectrum and with the calculated structure change. That is, the energy calculated for the neutral confined to the planar anion structure was found to be very close to the optimized neutral energy, which indicates a nearly vertical transition. Likewise, a nearly vertical transition is predicted for the WO2F− PE spectrum, which does indeed exhibit a nearly vertical transition, though the planar versus nonplanar differences in anion and neutral structures might be expected to generate a vibrationally broadened spectrum. The calculated ADE and VDE values are shown in red and blue, respectively, on the spectra shown in Figure 1a,b. Calculations on the WO3H and WO2F neutral doublet species confined to planar structures, but otherwise allowed to optimize, give energies that are very close to the fully optimized energy of the pyramidal structures with one imaginary frequency in the umbrella mode. For both the planar Cs WO3H and C2v WO2F neutrals, the electronic term energy, Te, is 0.03 eV (210 cm−1 for WO3H; 220 cm−1 for WO2F) relative to the total electronic energy of the optimized pyramidal structures. Figure 3 shows a schematic of the anion and neutral potential energy curves along the umbrella mode for WO2F based on force constants. The normal coordinate value of ±2.03 Å·amu1/2 corresponds to a 146° dihedral angle. An expanded view with the anion and neutral potentials shown superimposed is included in the Supporting Information (SI 6), along with depictions of the singly occupied neutral HOMO in the pyramidal and planar structures of both WO3H and WO2F. The umbrella mode of the pyramidal structure is calculated to be 114 cm−1 for WO3H and 117 cm−1 for WO2F, meaning that as few as three levels (e.g., the symmetric and antisymmetric combinations of υ′ = 0, and the symmetric combination of υ′ = 1) could be below the inversion barrier. The displacement along the umbrella coordinate in the anion to neutral transition is therefore zero, and only symmetric umbrella mode levels have nonzero overlap. Furthermore, overlap drops precipitously at higher υ′ levels due to dropping wave function amplitude at the center of the potential with increasing υ′. We approximated this effect by (1) calculating the Franck−Condon factors between the υ″ = 0 of the umbrella mode of the planar anion with two totally symmetric levels

respective anion and neutral isomers shown here. Table 1 summarizes the calculated transition energies between the various anion and neutral states. The lowest energy molecular and electronic structures of WO3H− and WO2F− are predicted to be very similar. The HOMO of the anions can be described as a W-local sdz2 hybrid, in which the z-axis runs perpendicular to the molecular plane through the metal center. Detachment from this sdz2 hybrid orbital is consistent with p-wave photoelectrons because it is localized on the W atom, and photodetachment of an electron from either an atomic s or an atomic d orbital yields p- (or p−f mixed) wave electrons. The HOMO is the only occupied orbital that is predominantly localized on the metal; the HOMO−1 and lower lying valence orbitals are more O- or OH- (or F-) localized, and significantly lower in energy. Depictions of WO3H and WO2F anion and neutral molecular orbitals and their relative energies can be found in the Supporting Information (SI 4−5). Both the ground and first excited triplet state of WO2F− have C2v symmetry. The ground and first excited triplet state of WO3H− both have Cs symmetry, but the 3A′ state is nonplanar, and the reflection plane runs along the W−OH bond, which is perpendicular to the plane defined by the WO2 portion of the molecule. The lowest energy molecular structures found for WO3H and WO2F are pyramidal structures with 146° dihedral angles formed between the WO2 plane and the W−OH or W−F bond, in 2A and 2A′ ground electronic states, respectively. Calculations on WO2F give harmonic frequencies that are in very good agreement with the observed vibrational progressions: The W−F stretch and WO2 symmetric stretch frequencies are calculated to be 675 and 1036 cm−1 (unscaled), in very good agreement with the observed 680 and 1024 cm−1 vibrational spacings. A compilation of the structural parameters, vibrational frequencies, and normal coordinates and a description of the normal modes are given for all the anion and neutral species included in this study is included in the Supporting Information (SI 11−18). The 2A + e− ← 1A transitions are one-electron allowed, and the calculated ADE values for both WO3H + e− ← WO3H− and WO2F + e− ← WO2F− transitions are in excellent agreement (within 0.1 eV) with the experimental values, as summarized in 13922

dx.doi.org/10.1021/jp4097666 | J. Phys. Chem. A 2013, 117, 13919−13925

The Journal of Physical Chemistry A

Article

Figure 3. Schematic of the potential energy curve along the lowfrequency umbrella coordinate for WO2F and WO2F−. The vibrational potentials for WO3H and WO3H− are nearly identical.

Figure 4. Calculation-based simulations of the (a) WO3H− and (b) WO2F− PE spectra, generated from calculated spectroscopic parameters for planar neutral structures (solid black traces) and fully optimized pyramidal structures, assuming a single-well potential along the umbrella mode (ca. 2 Å·amu1/2; blue traces). A complete list of the calculation-based simulation parameters is included in the Supporting Information.

below the barrier, assuming that the wave functions were symmetric superpositions of the υ′ = 0 and υ′ = 1 harmonic oscillator wave functions of the pyramidal structures, and (2) calculating the Franck−Condon factors between the anion for υ′ ≥ 4 levels assuming the neutral potential was delocalized over the breadth of the double-well potential (overlap with odd υ′ levels were, of course, identically zero). This simple model gave overlaps of 0.00015, 0.00114, 0.005, 0.0012, 0.0003, ... in order of increasing energy for υ′ = 0even, 1even, 4, 6, 8, .... respectively, giving one dominant transition at υ′ = 4, with much lower-intensity satellite transitions to both sides. A simulation of the umbrella mode progression based on these overlaps is shown in the Supporting Information (SI 6). Figure 4 shows simulations including all vibrational modes for the (a) WO3H− and (b) WO2F− PE spectra assuming a 300 K vibrational temperature, generated in two ways. The solid black traces were generated from calculated spectroscopic parameters generated from the calculated planar structures (which have one negative frequency associated with the umbrella mode), and the solid blue traces were generated from the calculated spectroscopic parameters associated with the optimized pyramidal structures, assuming a single-well potential. A complete list of the calculation-based simulation parameters is included in the Supporting Information (SI 7−8). Note that the experimental resolution is insufficient to resolve any splitting in the ca. 115 cm−1 vibrational progression in the umbrella mode, which, on the basis of the assessment above, would be dominated by a single peak. The simulation of the WO2F− spectrum based on the planar neutral is clearly in better agreement with the experimental spectrum. This is not the case for WO3H−, though for this molecule, the W−O−H torsion adds another inversion dimension. Indeed, given the similarities between the WO3H− and WO2F− electronic and molecular structures, the striking difference between the spectra must be associated with the properties of the hydroxyl group. The barrier for torsion around

the W−OH bond is low, 0.12 eV for WO3H−. A plot of energy as a function of torsion angle are included in the Supporting Information (SI 9), along with several levels shown at the harmonic energies (dashed lines). This additional nonharmonic vibrational coordinate must contribute significantly to the vibrational congestion of the spectrum.

■

DISCUSSION Oxidation State versus Shell Closing. The WO3H and WO2F neutrals contain a metal atom that is in a +5 oxidation state, accessed via detachment of an electron from the W center in a +4 oxidation state. Although the anion ground state is closed shell, the detachment energy for this closed-shell to open-shell transition appears to be in line with a simple plot of EA versus oxidation state, shown in Figure 5. Though the ADE of WO2F− is higher than the ADE of WO3H−, both values are between the ADE values of WO2 and WO3.11,21 Most importantly, neither EA is notably higher than the trend line, which is what would be expected if particular stability were furnished by the closed-shell anions. Also included on the plot is the EA of AlWO3, another complex with the W center in a +5 oxidation state and a closed shell electronic state,8 along with the EA of W 22 and several analogous molybdenum species,7,16,23−25 indicated with open symbols. On the basis of this comparison, it appears that the electronegativity of the ligand can have an effect on the neutral EA; the EA of WO2F is several tenths of an electronvolt higher than the EA’s of AlWO3 and WO3H. However, the dominant factor appears to be the oxidation state of the metal center. The closed shell electronic structure of WO3H− and WO2F− (and AlWO3−)8 certainly does not result in a significantly higher corresponding neutral EA resulting from particular electronic stability of the anion. 13923

dx.doi.org/10.1021/jp4097666 | J. Phys. Chem. A 2013, 117, 13919−13925

The Journal of Physical Chemistry A

Article

the low-frequency umbrella mode in the Cs pyramidal neutral features a very low-barrier inversion potential, and given the double-well potential, the normal coordinate displacement along the umbrella coordinate in the (planar) anion to neutral transition is zero, and only even levels are accessed. Franck− Condon factors calculated assuming a single-well pyramidal structure are inadequate for this type of system.

■

CONCLUSIONS The EA’s of open-shell WO3H and WO2F were determined from the PE spectra of WO3H− and WO2F−, in which transitions from closed-shell anions to the respective neutrals are observed. The W centers in the neutral species are nominally in a +5 oxidation state. Comparing the EA’s of these species with other tungsten−oxo complexes in which the neutral is closed shell and the anion is open shell revealed that the neutral complex EA’s follow a monotonic trend of increasing electron affinity with oxidation state, in spite of the WO3H−, WO2F−, and AlWO3− having closed shell ground states. This suggests that electronic stability in the closed-shell anion has a much less significant impact than oxidation state of the metal atom on the neutral electron affinity. In the W +4 to +6 oxidation range, there is also a correlation between oxidation state and the structure of the complex. In the stoichiometric WO3 cluster in which W is in a +6 oxidation state, the LUMO correlates to the singly occupied HOMO of both WO3H and WO2F (as well as WO3−), and doubly occupied HOMO of both WO3H− and WO2F−. This orbital, which can be described as a sdz2 hybrid orbital where z is essentially the pyramid axis, is the lowest energy W-local orbital in these complexes. The anions in which the W-local sdz2 orbital is doubly occupied are planar, whereas distortion to pyramidal structures increases as the occupancy decreases to one and zero. Computational investigations of the gas-phase metal-oxo complexes are imperative for the interpretation of the measured PE spectra. A comparison of experimental and calculated ADE/ VDE values is the first step in determining which of the predicted low-energy isomers actually gives rise to a particular PE spectrum. Calculation of normal coordinate displacements in the anion to neutral transition based on computationally determined spectroscopic parameters have been successfully used to simulate spectra for a more quantitative comparison of computational and experimental results. Though this has generally been a successful approach, there was very poor agreement between WO3H− experimental and theoretical VDE values, and the large normal coordinate displacement along the low-frequency umbrella mode calculated for the WO2F− spectrum were not consistent with the observed, nearly vertical transition. However, examination of the harmonic potentials compared to single-point calculations along the low-frequency umbrella mode pointed to the inadequacy of simulating the spectra based on a single-well potential. This result underscores the importance of careful scrutiny of “black box” computational results when used to interpret experimental results.

Figure 5. Plot of electron affinity as a function of Group 6 metal center oxidation state (n). Solid symbols represent tungsten complexes and open symbols represent analogous molybdenum complexes. Note that the W-center in the anionic complexes is in an n − 1 oxidation state. Data are drawn from refs 7 (AlMoOy), 8 (AlWOy), 11 (WO2), 16 (MoO3), 21 (WO3), 23 (W), 24 (Mo, MoO), and 25 (MoO3H).

Planar versus Pyramidal Structures. Comparing the orbital occupancies and the molecular structures of WO3H− and WO2F− and their neutrals, as well as previous results of WO3− photodetachment and computational studies,21 shows that the extent of distortion from planar to pyramidal can be related very simply to the orbital occupancy of the W-local sdz2 hybrid orbital. When the orbital is doubly occupied, a symmetric planar structure is formed. WO3H, WO2F, and WO3− all have a singly occupied analogous sdz2 hybrid orbital and have similar dihedral angles of approximately 145°. WO3, with no electrons occupying this lowest energy W-local orbital, was found to be significantly more distorted from a planar structure, with a dihedral angle approximately 120°.21 WO4H−, another complex with the W center in a +6 oxidation state (detachment of this anion involves a ligand local electron), was found to have a distorted tetrahedral structure, with the W− OH bond being significantly longer than the WO bonds.26 It is interesting to note at this point that in many W- (and Mo-) based enzymes, W ions are coordinated with both oxo and hydroxyl groups, among other ligands.27 In this study we have shown how the oxidation state of the W-center determines the electron affinity of the center, and also how occupancy affects the ligand structural arrangement in very simple ways. The ease with which a metal center can change between different oxidation states is an important feature in its catalytic activity.28 Simulating Low-Frequency Modes. In previous combined experimental and computational studies on the molecular structures of metal oxide clusters, we have found that using the harmonic oscillator and parallel mode approximations when spectra are simulated strictly on the basis of calculated normal coordinate displacements is problematic for the lowest frequency modes that generally correlated with the structural change between the anion and neutral. Because of anharmonicity, the energy level spacing between adjacent levels can converge quickly, so nontrivial Franck−Condon overlap occurs between the anion ground vibrational level and neutral levels spanning a much narrower energy range than predicted by harmonic models. In this current study, it was determined that

■

ASSOCIATED CONTENT

S Supporting Information *

Portions of the mass spectra showing the WO3−/WO3H− and WO3−/WOF− separation, laser polarization studies of WO3H− and WO2F−, depictions of the calculated valence orbitals of important species, a graph of the neutral and anion umbrella mode potentials with unperturbed vibrational levels, calculated spectroscopic parameters used in simulations of the WO3H− 13924

dx.doi.org/10.1021/jp4097666 | J. Phys. Chem. A 2013, 117, 13919−13925

The Journal of Physical Chemistry A

Article

and WO2F− PE spectra, O2−W−O3−H− torsion angle energy analysis, the PE spectrum of MoO3H− and MoO3−and calculated structural and vibrational details for the WO3H and WO2F anions and neutrals. This information is available free of charge via the Internet at http://pubs.acs.org.

■

(16) Yoder, B. L.; Maze, J. T.; Raghavachari, K.; Jarrold, C. C. Structures of Mo2Oy− and Mo2Oy (y = 2, 3, and 4) Studied by Anion Photoelectron Spectroscopy and Density Functional Theory Calculations. J. Chem. Phys. 2005, 122, 094313. (17) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Energy-Adjusted ab Initio Pseudopotentials for the 2nd and 3rd Row Transition Elements. Theor. Chim. Acta. 1990, 77, 123−141. (18) Martin, J. M. L.; Sundermann, A. Correlation Consistent Valence Basis Sets for use with the Stuttgart-Dresden-Bonn Relativistic Effective Core Potentials: The Atoms Ga-Kr and In-Xe. J. Chem. Phys. 2001, 114, 3408−3420. (19) Jensen, F. Introduction to Computational Chemistry, 2nd ed.; Wiley: Chichester, U.K., 2007. (20) Ervin, K.M. PESCAL, Fortran program (2010). (21) Zhai, H. J.; Kiran, B.; Cui, L.-F.; Li, X.; Dixon, D. A.; Wang, L.-S. Electronic Structure and Chemical Bonding in MOn− and MOn clusters (M = Mo, W; n = 3−5): A Photoelectron and ab Initio Study. J. Am. Chem. Soc. 2004, 126, 16134−16141. (22) Electron Affinities. In CRC Handbook of Chemistry and Physics, 93rd ed. (Internet Version); Haynes, W. M., Ed.; CRC Press/Taylor and Francis: Boca Raton, FL, 2013. (23) Wyrwas, R. B.; Yoder, B. L.; Maze, J. T.; Jarrold, C. C. Reactivity of Small MoxOy Clusters Toward Methane and Ethane. J. Phys. Chem. A 2006, 110, 2157−2164. (24) Gunion, R. F.; DixonWarren, S. J.; Lineberger, W. C.; Morse, M. D. Ultraviolet Photoelectron Spectroscopy of Molybdenum and Molybdenum Monoxide Anions. J. Chem. Phys. 1996, 104, 1765− 1773. (25) The unpublished PE spectrum of MoO3H− is included in the Supporting Information (SI 10). (26) Waters, T.; Wang, X. B.; Li, S. G.; Kiran, B.; Dixon, D. A.; Wang, L. S. Electronic Structure of the Hydroxo and Methoxo Oxometalate Anions MO3(OH)− and MO3(OCH3)− (M = Cr, Mo, and W). J. Phys. Chem. A 2005, 109, 11771−11780. (27) Gonzalez, P. J.; Rivas, M. G.; Mota, C. S.; Brondino, C. D.; Moura, I.; Moura, J. J. G. Periplasmic Nitrate Reductases and Formate Dehydrogenases: Biological Control of the Chemical Properties of Mo and W for Fine Tuning of Reactivity, Substrate Specificity and Metabolic Role. Coord. Chem. Rev. 2013, 257, 315−331. (28) Labinger, J. A. Alkane Functionalization via Electrophilic Activation. In Alkane C-H Activation by Single-Site Metal Catalysis; Pérez, P. J., Ed.; Springer: Dordrecht, Netherlands, 2012; pp 46−47.

AUTHOR INFORMATION

Corresponding Author

*C. C. Jarrold: e-mail, [email protected]; fax, 812-855-8300. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported in its entirety by Department of Energy Grant No. DE-FG02-07ER15889. REFERENCES

(1) Jordan, K. D.; Burrow, P. D. Temporary Anion States of Polyatomic Hydrocarbons. Chem. Rev. 1987, 87, 557−588. (2) Oakes, J. M.; Ellison, G. B. Photoelectron Spectroscopy of Radical Anions. Tetrahedron 1986, 42, 6263−6267. (3) Klingeler, R.; Lüttgens, G.; Pontius, N.; Rochow, R.; Bechthold, P. S.; Neeb, M.; Eberhardt, W. Photolectron spectra of small LaOn− clusters: decreasing electron affinity upon increasing the number of oxygen atoms. Euro. Phys. J. D 1999, 9, 263−267. (4) Mann, J. E.; Mayhall, N. J.; Jarrold, C. C. Properties of Metal Oxide Clusters in Non-Traditional Oxidation States. Chem. Phys. Lett. 2012, 525−26, 1−12. (5) Mann, J. E.; Waller, S. E.; Rothgeb, D. W.; Jarrold, C. C. Study of Nb2Oy (y = 2 - 5) Anion and Neutral Clusters Using Anion Photoelectron Spectroscopy and Density Functional Theory Calculations. J. Chem. Phys. 2011, 135, 104317. (6) Waller, S. E.; Mann, J. E.; Rothgeb, D. W.; Jarrold, C. C. Study of MoNbOy (y = 2 - 5) Anion and Neutral Clusters Using Photoelectron Spectroscopy and Density Functional Theory Calculations: Impact of Spin Contamination on Single Point Calculations. J. Phys. Chem. A 2012, 116, 9639−9652. (7) Waller, S. E.; Mann, J. E.; Hossain, E.; Troyer, M.; Jarrold, C. C. Electronic Structures of AlMoOy− (y = 1 − 4) Determined by Photoelectron Spectroscopy and DFT calculations. J. Chem. Phys. 2012, 137, 024302. (8) Mann, J. E.; Waller, S. E.; Jarrold, C. C. Electronic Structures of WAlOy and WAlOy− (y = 2 − 4) Determined by Anion Photoelectron Spectroscopy and Density Functional Theory Calculations. J. Chem. Phys. 2012, 137, 044301. (9) Moravec, V. D.; Jarrold, C. C. Study of NiO− and NiO Using Negative Ion Photoelectron Spectroscopy. J. Chem. Phys. 1998, 108, 1804−1810. (10) Waller, S. E.; Mann, J. E.; Jarrold, C. C. Asymmetric Partitioning of Metals Among Cluster Anions and Cations Generated via Laser Ablation of Mixed Aluminum/Group 6 Transition Metal Targets. J. Phys. Chem. A 2013, 117, 1765−1772. (11) Davico, G. E.; Schwartz, R. L.; Ramond, T. M.; Lineberger, W. C. An Experimental Study of the Low-lying Electronic States of WO2. J. Phys. Chem. A 1999, 103, 6167−6172. (12) Cooper, J.; Zare, R. N. Angular Distribution of Photoelectrons. J. Chem. Phys. 1968, 48, 942−943. (13) Reed, K. J.; Zimmerman, A. H.; Andersen, H. C.; Brauman, J. I. Cross Sections for Photodetachment of Electrons from Negative Ions Near Threshold. J. Chem. Phys. 1976, 64, 1368−1375. (14) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. G09, revision h08; Gaussian, Inc.; Wallingford, CT, 2009. (15) Lee, C. T.; Yang, W. T.; Parr, R. G. Development of the ColleSalvetti Correlation Energy Formula into a Functional of the Electron Density. Phys. Rev. B 1988, 37, 785−789. 13925

dx.doi.org/10.1021/jp4097666 | J. Phys. Chem. A 2013, 117, 13919−13925