Gas-Phase Fragmentation of Aluminum Oxide Nitrate Anions Driven

Feb 26, 2016 - Gas-phase metal nitrate anions are known to yield a variety of interesting metal oxides upon fragmentation. The aluminum nitrate anion ...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/JPCA

Gas-Phase Fragmentation of Aluminum Oxide Nitrate Anions Driven by Reactive Oxygen Radical Ligands Johnny Lightcap, Thomas H. Hester, Kurt Kamena, Rachael M. Albury, Carrie Jo M. Pruitt, and Daniel J. Goebbert* Department of Chemistry, The University of Alabama, Tuscaloosa, Alabama 35487, United States ABSTRACT: Gas-phase metal nitrate anions are known to yield a variety of interesting metal oxides upon fragmentation. The aluminum nitrate anion complexes, Al(NO3)4− and AlO(NO3)3− were generated by electrospray ionization and studied with collision-induced dissociation and energy-resolved mass spectrometry. Four different decomposition processes were observed, the loss of NO3−, NO3•, NO2•, and O2. The oxygen radical ligand in AlO(NO3)3− is highly reactive and drives the formation of AlO(NO3)2− upon loss of NO3•, AlO2(NO3)2− upon NO2• loss, or Al(NO2)(NO3)2− upon abstraction of an oxygen atom from a neighboring nitrate ligand followed by loss of O2. The AlO2(NO3)2− fragment also undergoes elimination of O2. The mechanism for O2 elimination requires oxygen atom abstraction from a nitrate ligand in both AlO(NO3)3− and AlO2(NO3)2−, revealing the hidden complexity in the fragmentation of these clusters. anion, and that copper was not oxidized upon O•− abstraction due to relative energy differences of the valence orbitals in copper and oxygen. The fragmentation of Cr(NO3)4− yielded a series of chromium oxide fragments, CrOn(NO3)4−n− (n = 1− 4), formed by consecutive loss of NO2•. Theoretical studies confirmed that chromium was oxidized by O•− because the metal valence orbitals are higher in energy than those on oxygen.15 Thus, oxygen ligands in CuO(NO3)− are reactive radical anions, whereas in CrOn(NO3)4−n− they are unreactive ligands. The goal of this work was to study the decomposition of Al(NO3)4− and investigate the properties of the aluminum oxide nitrate fragments, AlOn(NO3)4−m−. The core metal oxides, AlOn, decorated by nitrate ligands, act as a model for aluminum oxide systems with known composition and charge state. The electronic structure of aluminum oxide nitrate anions is simpler than most transition metals, and unlike many transition metals, Al 3+ cannot be oxidized upon O •− abstraction. These properties greatly simplify computational demands and interpretation of theoretical results relative to analogous transition metal oxides. Two previous fragmentation studies have been carried out on Al(NO3)4−.12,13 The original study reported three fragments, NO3−, AlO(NO3)3−, and AlO2(NO3)2−,12 whereas a more recent study detected the same major products along with two less intense products assigned as Al(NO2)(NO3)2− and Al(NO3)2−.13 The less intense products correspond to O2 elimination from AlO(NO3 )3 − and AlO2(NO3) 2−. The

1. INTRODUCTION γ-Alumina, γ-Al2O3, is an important catalyst and support due to its high surface area, thermal stability, and low cost. The two most common catalytic processes utilizing γ-Al2O3 include alcohol dehydrogenation1−4 and the Claus process for desulfurization of petroleum.5 Alcohol dehydration has received considerable interest for the production of organic starting materials such as ethylene, from renewable biomass. The elementary reaction mechanisms and active sites of γ-Al2O3 are poorly understood at the molecular level because the bulk structure has not been well characterized. Fundamental studies on small model aluminum oxide complexes have the potential to reveal integral structural motifs regulating the reactivity of the bulk material. Several theoretical studies have recently been carried out on model γ-Al2O3 clusters,1,4,6,7 along with a number of benchmark experimental studies on model systems.8−11 These studies suggest aluminum oxide reactivity involves cooperative effects of Lewis acid/base sites with possible contributions from surface Brønsted base sites. Despite the previous research, few experiments have studied Al−O bonding and reactivity in small isolated AlOn complexes. Mass spectrometry is one of the few experimental methods to allow direct investigation of isolated materials with known composition under controlled conditions. Generation of gasphase metal oxides is challenging, but previous mass spectrometric studies have demonstrated that transition metal nitrate complexes yield metal oxide fragments upon collisioninduced dissociation.12−15 Our group has investigated copper nitrate and chromium nitrate ions by tandem mass spectrometry.14,15 Fragmentation of Cu(NO3)2− complexes yielded CuO(NO3)− and CuO2−. Theoretical studies predicted the oxygen atom in the CuO(NO3)− fragment remained a radical © XXXX American Chemical Society

Received: December 18, 2015 Revised: February 17, 2016

A

DOI: 10.1021/acs.jpca.5b12417 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

2.3. Theoretical Calculations. All calculations were carried out using Gaussian 09.23 Structures were optimized at the B3LYP/6-311+G(d) level of theory. Open-shell species were calculated using the unrestricted, (U)B3LYP, method. Various conformations and/or electronic spin states of the complexes were explored, and the lowest energy structures are presented in this study. Vibrational frequencies were calculated to ensure the optimized structures converged to a minimum, with no imaginary frequencies, or a transition state, with a single imaginary frequency corresponding to motion along the reaction coordinate. Theoretical charges were obtained from natural population analysis, NPA, calculations.24−26

elimination of O2 is generally not observed in the dissociation of metal oxide nitrates12−15 because metal−oxide bond energies are typically large. The elimination of O2 from aluminum oxide is unexpected because this process may involve unfavorable metal reduction. We were interested in repeating these measurements to confirm the previously reported O 2 elimination reaction, and to investigate the mechanism of this unique process. These experiments provide fundamental information related to aluminum−oxygen bonding and reactivity.

2. EXPERIMENTAL METHODS 2.1. Sample Preparation. Aluminum nitrate hydrate was purchased from VWR and used without further purification. Solutions were prepared with a concentration of 1.00 mM in a mixture of HPLC grade methanol (50% vol) with HPLC grade acetonitrile (50% vol). 2.2. Mass Spectrometry. All experiments were carried out using a tandem quadrupole-octopole-quadrupole mass spectrometer TSQ-7000 (Finnigan MAT, San Jose, CA). Ions were generated by electrospray ionization (ESI). The ESI voltage was maintained at −4 kV with a current of 1 μA. The electrospray was directed toward a heated capillary inlet (200 °C). The pressure of the first differentially pumped region was about 800 mTorr. Ions were focused by a tube lens through a skimmer and sent into a second differentially pumped region with a pressure of about 1 mTorr, where they were collimated by an octopole ion guide. The voltage applied to the tube lens was minimized to limit internal excitation from collisions with background gas in the high pressure region. However, the tube lens voltage could also be increased to generate fragments for investigation. The collimated ion beam was focused by a pair of electrostatic lenses into the high vacuum chamber containing the tandem quadrupole assembly. Ions of interest were mass selected by the first quadrupole. Collision-induced dissociation, CID, was carried out in an octopole collision cell with 0.50 mTorr Ar target gas. Major fragment ions were studied by energy-resolved mass spectrometry, ERMS, where each product was monitored as a function of the collision energy offset voltage applied to the collision cell. The collision cell pressure was reduced to about 0.15 mTorr in ERMS experiments to minimize the probability of multiple collisions (pressure-dependent studies could not be carried out due to drift in the pressure gauge). The lab-frame collision energy was calibrated by scanning the collision cell offset voltage to locate the absolute zero ion kinetic energy of the precursor ion. The center-of-mass collision energy, Ecm, was calculated as Ecm = Elab[M/(m + M)], where Elab is the zero kinetic energy corrected lab-frame collision energy, M is the mass of the argon target gas, and m is the ion mass. Appearance energies were determined by extrapolation of the linearly increasing portion of the curves to the energy-axis intercept. Absolute dissociation energies were not modeled from the ERMS16−20 measurements because the internal energy of the precursor ion is unknown, the ion kinetic energy distribution is relatively broad, and a small fraction of ions undergo multiple collisions or collisions outside the collision cell resulting in broad low intensity signal in the threshold region.14,21 Relative appearance energies remain useful for analyzing and interpreting dissociation dynamics and for comparison with theoretical calculations. We have successfully used this approach to study the fragmentation of several ions.14,15,21,22

3. RESULTS AND DISCUSSION 3.1. Collision-Induced Dissociation. A representative collision-induced dissociation mass spectrum for Al(NO3)4− is shown in Figure 1a. This spectrum was recorded under multiple

Figure 1. Representative CID mass spectra of (a) Al(NO3)4−, (b) AlO(NO3)3−, and (c) AlO2(NO3)2− recorded at 40 eVlab collision energy using argon collision gas.

collision conditions (PAr = 0.50 mTorr) at a collision energy of 40 eVlab. The most intense fragment, m/z 229, corresponds to AlO(NO3)3−. This fragment is formed upon O•− abstraction from one of the nitrate ligands and elimination of NO2•. The intense fragment at m/z 62 is NO3−, yielding the neutral aluminum nitrate, Al(NO3)3. The peak at m/z 183 corresponds to the loss of NO2• from m/z 229, and the peak at m/z 151 corresponds to elimination of O2 from m/z 183. Two additional low intensity fragments are located at m/z 197 and 167. The spectrum in Figure 1 is in good agreement with the recently published spectrum by Frański and co-workers using a quadrupole time-of-flight mass spectrometer with a T-wave collision cell.13 Slight differences in relative product intensities are attributed to instrumental variations and collision conditions. The spectrum in Figure 1 confirms the formation B

DOI: 10.1021/acs.jpca.5b12417 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

shapes for m/z 151 and 183 are characteristic of sequential dissociation products.15,22 The m/z 62 fragment has an onset energy similar to that of m/z 151, but with lower relative intensity. The appearance energy for the m/z 121 fragment is around 3 eV, and it has the lowest relative intensity of the fragment ions studied by ERMS. 3.3. Theoretical Results. The proposed sequential decomposition of Al(NO3)4− is shown in Figure 3. The formation of the low intensity fragments such as m/z 75 and 105 is not considered because they can form by several different pathways and these species were not studied by ERMS. Theoretical studies of O2 elimination are examined separately; the corresponding products are not indicated in Figure 3. The optimized geometry for Al(NO3)4− in Figure 3 shows the nitrate ligands bind to aluminum via a single oxygen atom in a nearly tetrahedral arrangement. The theoretical metal charge (+1.971) is less than the formal oxidation state of +3; the difference is due to electron donation from the ligands to the metal.14,15 The lowest energy structure for Al(NO3)4− in our study is slightly different from a previously published structure where the ligands were indicated as having a less symmetric arrangement.27 The Al(NO3)4− precursor undergoes loss of NO3− or NO2•. No other direct products were detected. The calculated energy difference for the loss of NO3− is 1.99 eV whereas the energy difference for the loss of NO2• is 2.23 eV. Calculated energy differences are greater than the experimental appearance energies, about 1.3 eV for both ions, Figure 2a, because the internal energy of Al(NO3)4− from electrospray ionization reduces the energy required for fragmentation.14,15,21,28 The elimination of NO3− yields an undetectable neutral aluminum complex, this this reaction is ignored in the following discussion. The O2N−O− bond energy of the isolated nitrate ion is 4.69 eV.14 Thus, the energy required to induce N−O bond dissociation of a nitrate ligand is significantly reduced in the presence of the Lewis acid, Al3+. The lower energy is the result of simultaneous formation of the ionic Al−O bond upon elongation of the O2N−O− bond. The AlO(NO3)3− product, m/z 229, is the precursor for all other aluminum-containing fragments, Figure 1b. Theory shows the optimized structure for AlO(NO3)3− is similar to Al(NO3)4−, Figure 3. The unique Al−O bond is slightly shorter than the metal bonds to nitrate ligand oxygen atoms. A qualitative molecular orbital diagram in the zero-order approximation for the [AlO]2+ core is shown in Figure 4. The oxygen is a radical anion with an unpaired electron in a π type orbital, and the metal oxide core is best described as an Al3+[O•−] ion-contact pair. The electronic structure of the Al3+[O•−] core is useful for understanding the fragmentation reactions of AlO(NO3)3−. The AlO(NO3)3− ion yields three primary aluminum-containing fragments corresponding to the loss of NO3•, NO2•, and O2, and the reactive oxygen radical in AlO(NO3)3− plays a central role in all three processes. The loss of NO3• from AlO(NO3)3− yields m/z 167, AlO(NO3)2−. This is a reduction reaction involving transfer of an electron from one of the nitrate ligands. The electron could transfer to Al3+ or O•− in the [AlO]2+ core. Theory predicts a large negative charge, −1.399, on oxygen in AlO(NO3)2−. The oxygen ligand is reduced from O•− to O2− upon the loss of NO3• because the lowest energy unfilled orbital in AlO(NO3)3− is the π (nonbonding oxygen 2p) orbital, Figure 4. The increased charge on oxygen results in a shorter, 1.64 Å, and stronger ionic bond in the AlO(NO3)2− product with a

of the two O2 elimination products, m/z 197 and 151 from Al(NO3)4−. In-source fragmentation of Al(NO3)4− was used to generate m/z 229, AlO(NO3)3−, and m/z 183, AlO2(NO3)2−. These ions were mass-selected and subjected to CID. Representative fragmentation spectra are shown in Figure 1b,c. More fragments were detected in the CID spectra of AlO(NO3)3− and AlO2(NO3)2−, compared to the spectra of Al(NO3)4−, but the major fragments are common to all spectra. Fragmentation of AlO(NO3)3− yields intense products at m/z 62, 151, and 183, along with a series of less intense peaks located at m/z 75, 105, 121, 137, 167, and 197. This spectrum shows all aluminum-containing fragments originate from AlO(NO3)3−. The unique oxygen atom ligand is clearly responsible for the formation of the large number of fragment ions because Al(NO3)4− only yields two direct fragments whereas AlO(NO3)3− can yield four direct products. Fragmentation of AlO2(NO3)2−, Figure 1c, shows many of the same fragments as AlO(NO 3 ) 3 − . The most intense primary fragment of AlO2(NO3)2− is m/z 151 corresponding to elimination of O2. 3.2. Energy-Dependent Fragmentation. The major fragments from Al(NO3)4− were studied by ERMS, Figure 2a.

Figure 2. Energy-resolved mass spectra for dissociation of (a) Al(NO3)4− and (b) AlO(NO3)3−.

The appearance energy for both AlO(NO3)3− m/z 229 and NO3−, m/z 62 is about 1.3 eV. The appearance energy of m/z 183 is around 3 eV, whereas the appearance energy of m/z 151 is approximately 3.5 eV. The relative intensity of m/z 151 is initially lower than m/z 183, but the two fragments have similar intensities around 6 eV. This behavior is typical for sequential dissociation products. Energy-dependent fragmentation was also carried out for AlO(NO3)3−, Figure 2b. The presence of an intense AlOH(NO3)3− peak at m/z 230 required the use of higher resolution conditions to avoid interference from the higher mass ion, and this limited precursor ion transmission. The low absolute intensity of the precursor resulted in a noisier spectrum. The relative intensities of the four major fragments, m/z 183, 151, 121, and 62 were monitored as a function of collision energy. The m/z 183 fragment has the lowest appearance energy, about 0.6 eV and reaches a maximum intensity around 4 eV. The appearance energy of m/z 151 is about 1.8 eV, and m/z 151 becomes the most intense product above 5 eV. The curve C

DOI: 10.1021/acs.jpca.5b12417 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 3. Proposed decomposition of Al(NO3)4−. Optimized geometries were obtained at the (U)B3LYP/6-311+G(d) level of theory. Energy differences include zero-point corrections. Theoretical charges from NPA calculations are listed, along with select bond lengths or average metal− oxygen nitrate ligand bonds. Pink atoms are aluminum, red are oxygen, and blue are nitrogen.

anions form a bond upon the second NO2• elimination. The formation of the O−O bond explains the lower relative energy of the AlO2(NO3)2− product from AlO(NO3)3−. Elimination of O2 from AlO(NO3)3− yields the m/z 197 fragment, but AlO(NO3)3− cannot lose 3O2 directly. Figure 5 shows a slightly higher energy AlO2(NO2)(NO3)2− isomer that could lose 3O2 to yield Al(NO2)(NO3)2−. The rearrangement of AlO(NO3)3− to AlO2(NO2)(NO3)2− requires oxygen atom transfer from a nitrate ligand to the oxygen radical anion yielding a superoxide, O2•−, ligand. The oxygen−oxygen bond distance is 1.36 Å in the optimized structure for AlO2(NO2)(NO3)2−, to the bond length of 1.35 Å in isolated O2•−.31 The barrier for oxygen atom transfer is calculated to be about 2.46 eV. The AlO2(NO2)(NO3)2− isomer can eliminate 3 O2 by electron transfer and reduction of the metal to yield Al(NO2)(NO3)2−. The net energy required to lose 3O2 from AlO(NO3)3− is 2.92 eV, consistent with the low relative intensity of m/z 197 in Figure 1. Theoretical charges in Figure 5 show the aluminum atom is reduced in the Al(NO2)(NO3)2− product, and the local geometry of the aluminum atom in Al(NO2)(NO3)2− is trigonal pyramidal. The Al(NO2)(NO3)2− product is not the most stable isomer of m/z 197 because of the unusual oxidation state of the metal. Theory shows the lowest energy m/z 197 isomer is AlO(NO2)2(NO3)−, formed upon subsequent oxygen atom transfer from another nitrate ligand. The lower energy of the AlO(NO2)2(NO3)− isomer is attributed to the strong ionic bond formed upon electron transfer from the aluminum to oxygen as indicated by the theoretical charges in Figure 5. The AlO2(NO3)2− fragment, m/z 183, has the lowest appearance energy and greatest relative intensity of all three direct products from AlO(NO3)3−. This ion is the main precursor for the lower mass aluminum-containing fragments as shown in Figure 1c. Elimination of NO3• from AlO2(NO3)2− yields the m/z 121 fragment, loss of NO2• yields m/z 137, and

Figure 4. Orbital interaction diagram for the combination of Al3+ with O•− in the zero-order approximation. The core metal oxide is best described as Al3+[O•−]. The unfilled π orbital is essentially a 2p oxygen atom orbital.

Al3+[O2−] core. The formation of a strong ionic Al3+[O2−] bond yields low energy products (as opposed to the alterative Al2+[O•−] core), and this explains why reduction occurs at the oxygen rather than the metal. The calculated energy difference of 2.57 eV for the loss of NO3• is consistent with the low relative intensity of the m/z 167 product in Figure 1b. However, single determinant methods do not properly describe the electronic structure of NO3•,4,29 resulting in a larger possible error for the energy reported in Figure 3 for the loss of NO3•. Higher order methods were not applied because this reaction is not of primary interest in this study, and experimental energies for the formation of m/z 167 were not recorded due to its low intensity. The loss of NO2• from AlO(NO3)3− yields m/z 183, AlO2(NO3)2−. This is one of the most intense products in the CID spectra in Figure 1. Theory predicts the net energy difference for loss of NO2• is 1.79 eV, consistent with the low relative appearance energy for the m/z 183 product in Figure 2. The two oxygen atoms in AlO2(NO3)2− have a relative distance of 1.60 Å, characteristic of the peroxide, O22−, ligand.30 This indicates the unpaired electrons on the two oxygen radical D

DOI: 10.1021/acs.jpca.5b12417 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 5. Potential energy diagram for elimination of 3O2 from AlO(NO3)3− at the (U)B3LYP/6-311+G(d) level of theory. Select bond distances and NPA charges are listed for various species. Several specific charges are listed in red font. Relative energies include zero-point corrections.

Figure 6. Potential energy diagram for elimination of 1O2 from AlO2(NO3)2− at the (U)B3LYP/6-311+G(d) level of theory. Select bond distances and NPA charges are listed for various species. Several specific charges are listed in red font. Relative energies include zero-point corrections. The energies of the final products are listed as a range where the lower energy value is from theoretical calculations and the higher energy value, indicated with an *, is a hybrid theoretical and experimental estimate (see text for explanation).

loss of 1O2 yields m/z 151. The lowest energy structures for the m/z 121 and 137 fragments are shown in Figure 3. Theory predicts the loss of NO2• to yield m/z 137, AlO3(NO3)−, is lower in energy than loss of NO3• to yield m/z 121, AlO2(NO3)−. The CID spectrum in Figure 1c shows greater relative intensity for the theoretically higher energy AlO2(NO3)− product than it does for AlO3(NO3)−. There are several possible explanations for this. First, as previously mentioned, the energy calculated for the loss of NO3• is a crude estimate. Second, dissociation barriers were not located and the

relative energies of the transition states could account for the modulation in product intensity of AlO 3 (NO 3 ) − and AlO2(NO3)−. Finally, secondary dissociation reactions must be considered for CID spectra recorded under multiple collision conditions at high energy. For example, the intensity of the AlO3(NO3)− ion can be diminished by the loss of O2 to yield m/z 105 or NO3• to yield m/z 75. The most interesting AlO2(NO3)2− fragment corresponds to loss of O2, and experiment, Figure 2, shows this is a low energy process. E

DOI: 10.1021/acs.jpca.5b12417 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

driven by the oxygen radical anion ligand. Elimination of 3O2 from AlO(NO3)3− first requires oxygen atom transfer from a nitrate ligand to the oxygen radical anion yielding a superoxide ligand. The loss of NO3• from AlO(NO3)3− results in electron transfer to the oxygen radical. The loss of NO2• from AlO(NO3)3− yields the AlO2(NO3)2− product, where the oxygen radical anions combine to form a peroxide ligand. The theoretical study of 1O2 elimination from AlO2(NO3)2− revealed the reaction does not proceed via direct loss of the peroxide ligand, as might be expected, because double reduction of the aluminum atom yields a high energy Al(NO3)2− product. A lower energy mechanism for 1O2 elimination involves oxygen atom transfer from a nitrate ligand to the metal, coupled with electron transfer from the peroxide ligand to the nascent oxygen atom. This rearrangement yields a reactive oxygen radical anion ligand that can accept a second electron upon elimination of 1O2. Thus, a strong ionic bond is formed in the lower energy AlO(NO2)(NO3)− product. Finally, this work demonstrates that Al3+ is resistant to reduction in aluminum oxide nitrate complexes. For example, in the electron transfer reaction AlO(NO3)3− → AlO(NO3)2− + NO3•, the oxygen atom gains the electron, and in the reaction AlO2(NO3)2− → AlO(NO2)(NO3)− + 1O2, a complex rearrangement involving a nitrate ligand circumvents reduction of the metal. Aluminum is reduced in the reaction AlO(NO3)3− → Al(NO2)(NO3)2− + 3O2, but the Al(NO2)(NO3)2− product is less stable than the AlO(NO2)2(NO3)− isomer, where the metal returns to the favored +3 oxidation state upon oxygen atom transfer from a nitrate ligand. Alternative reactions that either increase metal oxide bond strength or form new metal oxide bonds appear to be preferred over metal reduction.

Theoretical energies should reflect the low energy for the formation of m/z 151 relative to m/z 121 and 137. The structure of AlO2(NO3)2− suggests the m/z 151 fragment could form by direct loss of 1O2 via transfer of two electrons and reduction of Al to the +1 oxidation state in the corresponding Al(NO3)2− product, Figure 6. This is the formula of the m/z 151 product assigned in the earlier fragmentation study.13 Care must be taken in the analysis of this process because the AlO2(NO3)2− precursor is a singlet, and conservation of spin requires elimination of 1O2.32 Theoretical calculations predict the direct elimination of 1O2 from AlO2(NO3)2− requires about 3.82 eV, Figure 6. However, there is significant spin contamination in the 1O2 wave function at this level of theory (S2 = 1.004), and the calculated singlet− triplet splitting for O2 is 0.45 eV compared to the experimental energy of 0.98 eV.31 Theoretical calculations on the triplet ground state of oxygen, 3O2, are more reliable. A hybrid experimental/theoretical estimate for 1O2 elimination was calculated by combining the experimental term energy with the theoretical energy for the (spin-forbidden) loss of 3O2 from AlO2(NO3)2−. The hybrid calculation suggests elimination of 1 O2 requires about 4.35 eV, as indicated in Figure 6. The energy calculated for direct elimination of 1O2 from AlO2(NO3)2− is not consistent with experimental results. These results suggest the m/z 151 fragment does not correspond to Al(NO3)2−. It is important to note, theory predicts the energy difference for elimination of the spin-forbidden product, 3O2, is 3.37 eV. The calculated energy is well above the relative energies for the formation of AlO 3 (NO 3 ) − and AlO 2 (NO 3 ) − from AlO2(NO3)2−. An alternative process for 1O2 elimination from AlO2(NO3)2− must be considered to explain the experimental energy trends for the m/z 151 product. Figure 6 shows the AlO(NO2)(NO3)− isomer of m/z 151 is much lower in energy than the Al(NO3)2− isomer. Formation of this product requires rearrangement of AlO2(NO3)2−. A p l a u s i b l e in t e r m e d i a t e f o r t h i s r e a c t i o n is t h e AlO(O2)(NO2)(NO3)− isomer shown in Figure 6. This isomer is formed upon oxygen atom transfer from a nitrate ligand to aluminum. A transition state for the rearrangement was found to be 2.18 eV higher in energy than AlO2(NO3)2−. Theoretical charges indicate an electron is transferred from the peroxide ligand in AlO 2 (NO 3 ) 2 − to the lone oxygen atom in AlO(O2)(NO2)(NO3)−. The new oxygen radical anion ligand formed by this rearrangement can accept the remaining electron from the superoxide ligand in AlO(O2)(NO2)(NO3)− resulting in loss of 1O2. The AlO(NO2)(NO3)− product is favored by the formation of a strong ionic bond, Al3+[O2−], in the metal oxide. The energy required for elimination of 1O2 is lower than the barrier for rearrangement. Therefore, oxygen elimination is spontaneous after isomerization. The calculated barrier for the isomerization of AlO2(NO3)2− is lower than theoretical energies for the loss of NO2• and NO3• reported in Figure 3, consistent with experimental fragment intensities and energetic trends in Figures 1 and 2. Thus, elimination of 1O2 from AlO2(NO3)2− proceeds via a complex rearrangement involving oxygen atom transfer from a nitrate ligand to the metal.

■

AUTHOR INFORMATION

Corresponding Author

*D. J. Goebbert. E-mail: [email protected]. Phone: (205) 348-2206. Fax: (205) 348-9104. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This project was supported by a grant from the University of Alabama (College Academy for Research and Creative Activity). We thank Mr. Clair Bragg (GenTech Scientific) for graciously providing us with electronics needed to make repairs to the instrument used in this work.

■

REFERENCES

(1) Christiansen, M. A.; Mpourmpakis, G.; Vlachos, D. G. DFTdriven multi-site microkinetic modeling of ethanol conversion to ethylene and diethyl ether on γ-Al2O3(1 1 1). J. Catal. 2015, 323, 121−131. (2) Kang, M.; DeWilde, J. F.; Bhan, A. Kinetics and mechanism of alcohol dehydration on γ-Al2O3: effects of carbon chain length and substitution. ACS Catal. 2015, 5, 602−612. (3) DeWilde, J. F.; Bhan, A. Kinetics and site requirements of ether disproportionation on γ-Al2O3. Appl. Catal., A 2015, 502, 361−369. (4) Fang, Z.; Wang, Y.; Dixon, D. A. Computational study of ethanol conversion on Al8O12 as a model for γ-Al2O3. J. Phys. Chem. C 2015, 119, 23413−23421. (5) Khanmamedov, T. K.; Weiland, R. H. How sulphur really forms on the catalyst surface. Sulphur 2013, 345, 62−68. (6) Comas-Vives, A.; Valla, M.; Copéret, C.; Sautet, P. Cooperativity between Al sites promotes hydrogen transfer and carbon-carbon bond

4. CONCLUSIONS This study revealed several unexpectedly complex processes in the decomposition of Al(NO3)4− and AlO(NO3)3−. The AlO(NO3)3− intermediate is the precursor to all other aluminum-containing fragments, and sequential dissociation is F

DOI: 10.1021/acs.jpca.5b12417 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A formation upon dimethyl ether activation on alumina. ACS Cent. Sci. 2015, 1, 313−319. (7) Kostestkyy, P.; Yu, J.; Gorte, R. J.; Mpourmpakis, G. Structureactivity relationships on metal-oxides: alcohol dehydration. Catal. Sci. Technol. 2014, 4, 3861−3869. (8) Johnson, G. E.; Tyo, E. C.; Castleman, A. W., Jr. Oxidation of CO by aluminum oxide cluster ions in the gas phase. J. Phys. Chem. A 2008, 112, 4732−4735. (9) Wang, Z.-C.; Weiske, T.; Kretschmer, R.; Schlangen, M.; Kaupp, M.; Schwarz, H. Structure of the oxygen-rich cluster cation Al2O7+ and its reactivity toward methane and water. J. Am. Chem. Soc. 2011, 133, 16930−16937. (10) Arakawa, M.; Kohara, K.; Terasaki, A. Reaction of aluninum cluster cations with a mixture of O2 and H2O gases: Formation of hydrated-alumina clusters. J. Phys. Chem. C 2015, 119, 10981−10986. (11) Kretschmer, R.; Wang, Z.-C.; Schlangen, M.; Schwarz, H. Single and double N-H bond activation of ammonia by [Al2O3]•+: Room temperature formation of the aminyl radical and nitrene. Angew. Chem., Int. Ed. 2013, 52, 9513−9517. (12) Li, F.; Byers, M. A.; Houk, R. S. Tandem mass spectrometry of metal nitrate negative ions produced by electrospray ionization. J. Am. Soc. Mass Spectrom. 2003, 14, 671−679. (13) Frański, R.; Sobieszczuk, K.; Gierczyk, B. Mass spectrometric decomposition of [Mn+(NO3−)n+1]− anions originating from metal nitrates M(NO3)n. Int. J. Mass Spectrom. 2014, 369, 98−104. (14) Pruitt, C. J. M.; Goebbert, D. J. Experimental and theoretical study of the decomposition of copper nitrate cluster anions. J. Phys. Chem. A 2015, 119, 4755−4762. (15) Albury, R. M.; Pruitt, C. J. M.; Hester, T. H.; Goebbert, D. J. Fragmentation of Cr(NO3)4−: Metal oxidation upon O•− abstraction. J. Phys. Chem. A 2015, 119, 11471−11478. (16) Armentrout, P. B.; Ervin, K. M.; Rodgers, M. T. Statistical rate theory and kinetic energy-resolved ion chemistry: Theory and applications. J. Phys. Chem. A 2008, 112, 10071−10085. (17) Armentrout, P. B. Statistical modeling of sequential collisioninduced dissociaiton thresholds. J. Chem. Phys. 2007, 126, 234302. (18) Armentrout, P. B. Mass spectrometry - not just a structural tool: The use of guided ion beam tandem mass spectrometry to determine thermochemistry. J. Am. Soc. Mass Spectrom. 2002, 13, 419−434. (19) Rodgers, M. T.; Armentrout, P. B. Noncovalent metal-ligand bond energies as studied by threshold collision-induced dissociation. Mass Spectrom. Rev. 2000, 19, 215−247. (20) Rodgers, M. T.; Ervin, K. M.; Armentrout, P. B. Statistical modeling of collision-induced dissociation thresholds. J. Chem. Phys. 1997, 106, 4499−4508. (21) Pruitt, C. J. M.; Benham, K.; Bandyopadhyay, B. G.; D, J. Fragmentation of deprotonated nitrophenol anions. Chem. Phys. Lett. 2014, 614, 192−198. (22) Hester, T. H.; McCraney, K. K.; Castillo, D. E.; Goebbert, D. J. The sequential dissociation of protonated polyethyelene glycols. J. Mass Spectrom. 2013, 48, 459−464. (23) 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. Gaussian 09; Gaussian Inc.: Wallingford, CT, 2009. (24) Maseras, F.; Morokuma, K. Application of the natural population analysis to transition-metal complexes. Should the empty p orbitals be included in the valence space? Chem. Phys. Lett. 1992, 195, 500−504. (25) Reed, A. E.; Weinhold, F. Natural localized molecular orbitals. J. Chem. Phys. 1985, 83, 1736−1740. (26) Reed, A. E.; Weinhold, F. Natural bond order analysis of nearHartree-Fock water dimer. J. Chem. Phys. 1983, 78, 4066−4073. (27) Lucena, A. F.; Lourenço, C.; Michelini, M. C.; Rutkowski, P. X.; Carretas, J. M.; Zorz, N.; Berthon, L.; Dias, A.; Oliviera, M. C.; Gibson, J. K.; et al. Synthesis and hydrolysis of gas-phase lanthanide and actinide oxide nitrate complexes: a correspondence to trivalent metal ion redox potentials and ionization energies. Phys. Chem. Chem. Phys. 2015, 17, 9942−9950.

(28) Pruitt, C. J. M.; Goebbert, D. J. The C-N dissociation energies of nitrobenzene and nitrotoluene radical anions and neutrals. Chem. Phys. Lett. 2013, 580, 21−27. (29) Kaufman, S. H.; Weber, J. M. Photodissociation Spectroscopy of the Anionic Copper Nitrate Association Complex Cu(NO3)3−. J. Phys. Chem. A 2014, 118, 9687−9691. (30) Miessler, G. L.; Tarr, D. A. Inorganic Chemistry, 2nd ed.; Prentice-Hall, Inc.: Hoboken, NJ, 1998. (31) Ervin, K. M.; Anusiewicz, I.; Skurski, P.; Simons, J.; Lineberger, W. C. The only stable state of O2− is the X2Πg ground state and it (still!) has an adiabatic electron detachment energy of 0.45 eV. J. Phys. Chem. A 2003, 107, 8521−8529. (32) Burgert, R.; Schnöckel, H.; Grubisic, A.; Li, X.; Stokes, S. T.; Bowen, K. H.; Ganteför, G. F.; Kiran, B.; Jena, P. Spin conversion accounts for aluminum cluster anion reactivity patterns with O2. Science 2008, 319, 438−442.

G

DOI: 10.1021/acs.jpca.5b12417 J. Phys. Chem. A XXXX, XXX, XXX−XXX