Fragmentation of [Ni(NO3)3 ... - ACS Publications

Jun 22, 2016 - atoms with the goal of better understanding metal−oxygen bonding .... with no imaginary frequencies, whereas transition-state structu...
1 downloads 0 Views 1MB Size
Article pubs.acs.org/IC

Fragmentation of [Ni(NO3)3]−: A Study of Nickel−Oxygen Bonding and Oxidation States in Nickel Oxide Fragments Thomas H. Hester, Rachael M. Albury, Carrie Jo M. Pruitt, and Daniel J. Goebbert* Department of Chemistry, The University of Alabama, Tuscaloosa, Alabama 35487, United States S Supporting Information *

ABSTRACT: Gas-phase nickel nitrate anions are known to produce nickel oxide nitrate anions, [NiOx(NO3)y]− upon fragmentation. The goal of this study was to investigate the properties of nickel oxide nitrate complexes generated by electrospray ionization using a tandem quadrupole mass spectrometer and theoretical calculations. The [Ni(NO3)3]− ion undergoes sequential NO2• elimination to yield [NiO(NO3)2]− and [NiO2(NO3)]−, followed by elimination of O2. The electronic structure of the nickel oxide core influences decomposition. Calculations indicate electron density from oxygen is delocalized onto the metal, yielding a partially oxidized oxygen in [NiO(NO3)2]−. Theoretical studies suggest the mechanism for O2 elimination from [NiO2(NO3)]− involves oxygen atom transfer from a nitrate ligand to yield an intermediate, [NiO(O2)(NO2)]−, containing an oxygen radical anion ligand, O•−, a superoxide ligand, O2•−, and a nitrite ligand bound to Ni2+. Electron transfer from superoxide partially reduces both the metal and oxygen and yields the energetically favored [NiO(NO2)]− + O2 products.

1. INTRODUCTION Modern industrial chemistry relies on heterogeneous catalysts composed of metals or metal oxides to increase yields and reduce waste and energy costs. Surprisingly, the most economical approach to identify catalysts for industrial application involves combinatorial or trial-and-error searches. These methods can be used to select viable catalysts, but they offer little insight as to why a particular catalyst works at the molecular level. The development of new catalysts tailored for applications with increased activity, specificity, turnover, minimal waste generation, and low cost requires a fundamental understanding of metal and metal oxide bonding and properties. Experiments that provide fundamental atomic-level information on catalyst properties promise to usher in a new era of advanced high-performance materials. A variety of experimental methods are needed to realize these goals. Gas-phase ion chemistry allows direct investigation of model size-selected metal complexes under carefully controlled conditions.1−5 These studies make it possible to observe changes that depend on stoichiometry, oxidation state, spin state, and size. For example, research on nanoclusters has repeatedly demonstrated the addition or removal of a few atoms can drastically alter reactivity and reaction rates.1,3,5−10 We are interested in studying the properties of metal oxides composed of a few atoms with the goal of better understanding metal−oxygen bonding and the role of electronic structure in reactivity. Experimental and theoretical studies on small model systems provide a conceptual foundation that improves understanding © XXXX American Chemical Society

of bonding, electronic structure, and reactivity in larger complex clusters.11−21 Previous work by our group on small metal oxide complexes has emphasized the link between electronic structure and fragmentation. For example, dissociation of [Cr(NO3)4]− results in metal oxidation upon sequential O•− abstraction from the nitrate ligands to yield a series of [CrOn(NO3)4−n]− (n = 1−3) fragments.22 Oxidation of the metal yields relatively inert oxo- [O2−] ligands bound to the metal. The conceptual molecular orbital diagram in Figure 1 shows metal oxidation is driven by the high-energy 3d electrons of Cr3+ relative to the partially filled 2p orbital on O•−. In contrast, the fragmentation of [Cu(NO3)3]− involves metal reduction to yield [Cu(NO3)2]− + NO3• by electron transfer from a nitrate ligand to copper.23 The reduction reaction is facilitated by the lowenergy unfilled 3d orbital in Cu2+. Dissociation of [Cu(NO3)2]− yields [CuO(NO3)]− by O•− abstraction from a nitrate ligand. Theory suggested copper is not oxidized, and the oxygen ligand is a reactive radical anion. The conceptual molecular orbital diagram in Figure 1 indicates Cu+ is not oxidized by O•− because of the lower relative 3d orbital energies typical of later transition metals. The metal oxide core is best described as a Cu+[O•−] ion contact pair. Secondary dissociation of [CuO(NO3)]− to yield [CuO2]− results in partial oxidation of the metal. These studies have demonstrated the importance of electronic structure in metal−oxygen Received: April 4, 2016

A

DOI: 10.1021/acs.inorgchem.6b00812 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

ties of [NiO(NO3)2]− depend on the electronic structure of the [NiO]+ core. Theoretical calculations were compared with experimental fragmentation and energy measurements to better understand nickel−oxygen bonding within nickel oxide fragments.

2. EXPERIMENTAL METHODS 2.1. Sample Preparation. Nickel nitrate hexahydrate (Alfa Aesar, 98%) was purchased from VWR and used without further purification. Solutions were prepared with a concentration of 1.00 mM in a mixture of high-performance liquid chromatography (HPLC) grade methanol (50% vol) with HPLC grade acetonitrile (50% vol). 2.2. Mass Spectrometry. All experiments were performed using a tandem quadrupole mass spectrometer TSQ-7000, (Finnigan MAT, San Jose, CA) with an electrospray ionization, ESI, source. Solutions were infused with a syringe pump (5 μL/min). Nitrogen was used as a nebulizing gas, 30−60 psi. The ESI voltage was maintained at −4 kV with a current of ∼1 μA. The electrospray was directed toward a heated capillary held at 200 °C. Ions entered a differentially pumped region, where they were focused by a tube lens through a skimmer into a second differentially pumped region. Ions were collimated by an ion guide and transferred into the high-vacuum chamber. The tandem mass spectrometer used in this experiment has a quadrupole−octopole−quadrupole configuration used for collisioninduced dissociation (CID) and energy-resolved mass spectrometry (ERMS). Fragment ions from CID were mass analyzed by scanning the third quadrupole and detected by a channeltron electron multiplier. Fragment ions were further studied by ERMS, where the intensity of each product was monitored as a function of the collision energy offset voltage applied to the collision cell. The collision cell pressure was ≤0.15 mTorr to minimize the probability of multiple collisions during these measurements. The lab-frame collision energy was calibrated by a retarding potential analysis to locate the voltage offset equal to the zero ion kinetic energy. The center-of-mass collision energy was calculated as Ecm = Elab[M/(m + M)], where Ecm is the center of mass energy, Elab is the lab-frame collision offset voltage corrected for zero kinetic energy, M is the mass of the argon target gas, and m is the precursor ion mass. Approximate appearance energies were determined by extrapolation of the linearly increasing portion of the curves to the x-axis intercept. Threshold modeling27−30 was not performed because the precursor ion internal energy from the ESI source is not well-characterized, the ion kinetic energy distribution is relatively broad, and a small fraction of ions undergo multiple collisions and/or collisions outside the collision cell resulting in broad low intensity background signal in the threshold region. Pressuredependent studies and extrapolation to zero pressure would reduce several sources of error, but attempts to include this correction have been hampered by pressure gauge drift. Interpretation relies on comparisons between relative appearance energies and trends with theoretical results; we have successfully used this approach to study the fragmentation of several ions.23,31,32 2.3. Theoretical Calculations. All calculations were performed using Gaussian 09.33 Structures were optimized at the unrestricted, (U)B3LYP/6-311+G(3df) level of theory. Various structures with different electronic spin states were optimized for all complexes to determine the lowest-energy species. Vibrational frequencies were calculated to ensure the optimized structures converged to a minimum with no imaginary frequencies, whereas transition-state structures were characterized by a single imaginary frequency corresponding to motion along the reaction coordinate. Reported energies include zero-point correction. A natural population analysis (NPA)34−36 was performed at the (U)B3LYP/6-311+G(3df) level of theory to estimate atomic charges and electron configurations. The density functional theory calculations were compared with experiment to provide a general understanding of metal−oxygen bonding. Several differences between experiment and theory arise, because accurate electronic structure calculations require inclusion of relativistic effects and multireference methods;37 however, such approaches are outside the scope of the current research. The main conclusions from this work are in good

Figure 1. Hypothetical orbital interaction diagrams illustrating (top) oxidation of high-energy 3d electrons on, an earlier transition metal, Cr3+ by O•−. In contrast, the combination of a late transition metal, Cu+ with O•− (bottom), does not result in metal oxidation due to the low-energy 3d orbitals.

bonding and fragmentation.22,23 The conceptual molecular orbital diagrams in Figure 1 show the location of unpaired electrons, on the metal or oxygen atoms, influences properties and reactivity. In this paper we report our investigation of [Ni(NO3)3]− dissociation. Previous collision-induced dissociation studies on [Ni(NO3)3]− have reported preferential elimination of NO2•.24,25 The Ni2+ cation at the core of this complex is interesting because it has triplet ground state with a 3d8 configuration, and the relative 4s−3d splitting and energies should be similar to that of copper, Figure 1. The two partially filled low-energy 3d metal orbitals could accommodate an electron upon reduction and elimination of NO3•, similar to the reactions observed for the [Cu(NO3)3]− ion.23 The metal could also abstract O•− from a nitrate ligand, a common reaction for most metal nitrate anions.22−25 Two different nickel−oxygen bonding schemes are possible in the resulting [NiO(NO3)2]− complex. The first involves pairing of electrons on the metal and oxygen radical anion, similar to the bonding in the [CrO(NO3)3]− complex.22 The [NiO(NO3)2]− complex would have a doublet ground state with a single unpaired electron remaining on the metal. The second bonding scheme involves donation of a lone pair of electrons from the oxygen atom into the empty 4s orbital of the metal, similar to the bonding in the [CuO(NO3)]− ion.23 In this scenario, the unpaired electron on the oxygen radical anion ligand and two unpaired electrons from the metal would occupy nearly degenerate orbitals, yielding a quartet ground state in the [NiO(NO 3 ) 2 ] − complex.26 The sequential dissociation, reactivity, and properB

DOI: 10.1021/acs.inorgchem.6b00812 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

form through several pathways, including loss of NO3• from [NiO2(NO3)]−, loss of NO2• from [NiO(NO3)]−, or possibly loss of O from [NiO3]−. 3.2. Energy-Resolved Fragmentation. The most intense fragments from [Ni(NO3)3]− were studied using ERMS. A representative spectrum is shown in Figure 3a. The lowest-

agreement with recent high-level calculations on [NiO]+.26 Survey calculations for select complexes using other levels of theory do not change the important conclusions from this study. Geometries, energies, and transition-state frequencies for the nickel complexes are provided in the Supporting Information.

3. RESULTS 3.1. Collision-Induced Dissociation. The fragmentation spectrum of [Ni(NO3)3]− is shown in Figure 2a. The spectrum

Figure 3. Energy-dependent fragmentation spectra of (a) [Ni(NO3)3]− and (b) [NiO(NO3)2]−. Figure 2. CID mass spectra of (a) [Ni(NO3)3]− and (b) [NiO(NO3)2]− recorded at 40 eVlab with 0.50 mTorr pressure of Ar collision gas.

energy products correspond to [NiO(NO3)2]− and NO3−. The [NiO(NO3)2]− fragment is more intense than NO3− under the lower collision gas pressure (