Synchrotron Photoionization Measurements of OH-Initiated

May 25, 2012 - Department of Chemistry, University of San Francisco, San Francisco, ...... by Sandia Corporation, a Lockheed Martin Company, for the N...
1 downloads 0 Views 2MB Size
Article pubs.acs.org/JPCA

Synchrotron Photoionization Measurements of OH-Initiated Cyclohexene Oxidation: Ring-Preserving Products in OH + Cyclohexene and Hydroxycyclohexyl + O2 Reactions Amelia W. Ray,†,‡ Craig A. Taatjes,§ Oliver Welz,§ David L. Osborn,§ and Giovanni Meloni*,† †

Department of Chemistry, University of San Francisco, San Francisco, California 94117, United States Combustion Research Facility, Mailstop 9055, Sandia National Laboratories, Livermore, California 94551, United States

§

S Supporting Information *

ABSTRACT: Earlier synchrotron photoionization mass spectrometry experiments suggested a prominent ring-opening channel in the OH-initiated oxidation of cyclohexene, based on comparison of product photoionization spectra with calculated spectra of possible isomers. The present work reexamines the OH + cyclohexene reaction, measuring the isomeric products of OH-initiated oxidation of partially and fully deuterated cyclohexene. In particular, the directly measured photoionization spectrum of 2-cyclohexen-1ol differs substantially from the previously calculated Franck−Condon envelope, and the product spectrum can be fit with no contribution from ring-opening. Measurements of H2O2 photolysis in the presence of C6D10 establish that the addition−elimination product incorporates the hydrogen atom from the hydroxyl radical reactant and loses a hydrogen (a D atom in this case) from the ring. Investigation of OH + cyclohexene-4,4,5,5-d4 confirms this result and allows mass discrimination of different abstraction pathways. Products of 2-hydroxycyclohexyl-d10 reaction with O2 are observed upon adding a large excess of O2 to the OH + C6D10 system.



INTRODUCTION As increasing use is made of nontraditional petroleum sources, such as oil-sands derived fuels, there is a greater need to fully understand the mechanisms of oxidation of naphthenic compounds (cycloalkanes) that are found in higher concentration in these fuels.1 Differences in fuel composition are dependent on the original source, such as terrestrial plants (coal) compared to marine or fresh water algae (oil shale). Mono- and multicyclic alkanes make up a substantial fraction of pyrolysis liquid from prehistoric-algae derived fuel stocks such as oil-sands (asphalt ridge tar sand) and Colorado oil shale.1 As one of the smallest and simplest of the naphthenic compounds, the oxidation of cyclohexane has been well studied.2−8 One important aspect in the oxidation of cyclohexane is the formation of benzene via successive dehydrogenation through cyclohexene and cyclohexadiene intermediates. The benzene and 1,3-cyclohexadiene products have been observed in several cyclohexane oxidation studies using various techniques such as a closed reactor,3 a rapid compression machine,4 a jet-stirred reactor,5 and laminar premixed flames.7 This process is distinct from the formation of aromatic rings via more familiar routes involving small radical species such as C3H3 and C4H5.9 The formation of aromatics is the first step in the formation of soot, and cyclic alkenes such as cyclopentene have previously been shown to have a high tendency to soot.10,11 As such, there has been recent interest in the oxidation of benzene12,13 as well as © 2012 American Chemical Society

the OH-initiated oxidation of the cycloalkane reaction intermediates cyclopentene and cyclohexene.14 The hydroxyl radical plays an important role in oxidation of alkenes in both combustion15,16 and in the atmosphere.17 The major reaction channel consists of OH addition to the double bond forming a hydroxyalkyl intermediate.18 The minor hydrogen abstraction channel becomes more prominent with increasing temperature.18 Previous work on the OH-initiated oxidation of small, cyclic alkenes14 showed that cyclic alkenes behave like their acyclic counterparts, with the major reaction channel arising from OH addition to the double bond.17 Dissociation of the intermediate adduct, or reaction of the stabilized hydroxyalkyl radical with O2, can result in ketone/ aldehyde formation. Recent studies of flames using tunable vacuum ultraviolet photoionization mass spectrometry19 indicate the formation of the corresponding enols from the OH + alkene reactions. The presence of both enol and ketone products was suggested in recent studies of the OH-initiated oxidation of cyclopentene and cyclohexene.14 In the OH + cyclohexene reaction, a ring-opening pathway resulting in the formation of straight-chain aldehydes was indicated in addition to formation of the cyclic unsaturated alcohols. The corresponding ring-opening product was not observed for the Received: March 7, 2012 Revised: May 19, 2012 Published: May 25, 2012 6720

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730

The Journal of Physical Chemistry A

Article

Neutral species from the reacting mixture effuse through a pinhole and are skimmed into the differentially pumped ionization region of the mass spectrometer. The tunable vacuum ultraviolet synchrotron radiation of the ALS is used to ionize the sampled species, and the ions are detected by means of an orthogonal-acceleration time-of-flight mass spectrometer with a pulse rate of 50 kHz. Scanning the photon energy of the synchrotron during data acquisitions yields a three-dimensional block data as a function of kinetic time, mass-to-charge ratio, and photoionization energy. After subtraction of the prephotolysis background and normalization for the photon flux of the ionization beam, time-resolved mass spectra are obtained by integrating over a range of photon energies, and photoionization spectra by integration over a range of kinetic time. An example of such a photoionization spectrum is shown in Figure 1, which depicts the spectra of the m/z = 102 product

cyclopentene system. The assignment of a ring-opening pathway in the cyclohexene reaction was made because the experimental photoionization product spectrum was not consistent with contributions from cyclic species alone, but could be successfully reproduced by additionally including the ring-opening product hex-5-enal. The analysis was based on simulated spectra of the cyclic unsaturated alcohol products 1cyclohexen-1-ol and 2-cyclohexen-1-ol using Franck−Condon factor calculations. The current work uses a direct experimental calibration spectrum of the 2-cyclohexen-1-ol isomer; using this experimental spectrum for analysis, there is no indication of ring-opening in the OH + cyclohexene reaction. This work also uses isotopically labeled cyclohexenes to more fully characterize the OH-initiated oxidation of cyclohexene, including the reaction of the 2-hydroxycyclohexyl radical with O2.



EXPERIMENTAL SECTION Experiments are carried out at the Chemical Dynamics Beamline of the Advanced Light Source (ALS) at Lawrence Berkeley National Laboratory using a side-sampled slow-flow reactor coupled to a multiplexed photoionization mass spectrometer, described in greater detail elsewhere.20,21 Hydrogen peroxide (H2O2) produced by thermal decomposition of urea-hydrogen peroxide adduct (97%) is used as the hydroxyl radical photolytic precursor. The H2O2 source also produces some O2, possibly from surface-mediated decomposition of H2O2, so the minimum O2 concentration in these experiments is approximately 1012 cm−3. Reactions are carried out at 298 K and 4 Torr. The cyclohexene reactants consist of perdeuterated cyclohexene (C6D10; 98 atom % D) and partially deuterated cyclohexene-4,4,5,5-d4 (98 atom % D; see Scheme 1). Scheme 1

Figure 1. Left panel: Photoionization spectra for the m/z = 102 product, formed via the OH addition/hydrogen loss pathway in the C6H6D4 + OH reaction (open circles) and its isomeric composition. The individual isomer photoionization spectra are shown for reference, with the 2-cyclohexen-1-ol offset to depict its contribution to the fit. The spectrum of the disfavored m/z = 101 product, from loss of a deuterium atom after OH addition, is also displayed. Right panel: Photoionization spectrum of the analogous (m/z = 107) product formed via the OH reaction with C6D10, which has a similar isomeric composition to the m/z = 102 product in the left panel. In both cases, the best fit contained a small, statistically insignificant component of cyclohexanone, the spectrum of which is shown for reference. The magnitude of the continued rise in the experimental spectrum toward higher energy cannot be reproduced by cyclohexanone.

Cyclohexene and H2O2 are diluted in helium and flowed through a quartz slow-flow reactor, where the mixture is photolyzed by a pulsed 248 nm (KrF) excimer laser. The cyclohexene concentration is ∼3 × 1013 cm−3. The measured depletion of cyclohexene for these conditions is ∼6%. The maximum possible concentration of H2O2, if it were present at its saturation vapor pressure22,23 in the flow out of the ureahydrogen peroxide bubbler, would be ∼5 × 1015 cm−3. The H2O2 concentration is not directly measured, but is estimated to be between 1 × 1014 cm−3 and 5 × 1014 cm−3 based on the measured cyclohexene depletion, the rate coefficients for OH reactions with cyclohexene24 and with H2O2,25 and the measured ∼1.5% depletion of H2O2. The majority of the OH produced in the photolysis (∼3 × 1012 cm−3) reacts with cyclohexene to initiate the reaction, but some side reaction with the H2O2 precursor also occurs to form HO2 radicals. Some single-photon-energy experiments were carried out at lower (6 × 1012 cm−3) and higher (1.3 × 1014 cm−3) cyclohexene concentrations.

(corresponding to OH addition to the cyclohexene double bond followed by the loss of an H atom) initiated by the reaction of OH with cyclohexene-4,4,5,5-d4. The spectra can be compared to measured or calculated spectra of the various isomers to derive isomeric branching fractions, as described previously.26,27 The photoionization spectrum of commercially obtained 2-cyclohexen-1-ol was measured relative to that of propene (see Supporting Information), following a similar procedure to that described elsewhere.28,29 Figure S1 shows that the measured photoionization cross section spectrum of 2cyclohexen-1-ol exhibits a much broader envelope than the Franck−Condon simulations employed in the previous study.14



COMPUTATIONS Electronic structure calculations for both the neutral and cation of cyclohexanol, 2-cyclohexen-1-one, and 2,3-epoxycyclohexan6721

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730

The Journal of Physical Chemistry A

Article

Scheme 2

Table 1. Major Secondary Product Isomers Considered in This Worka

B3LYP/6-311G(d,p) optimized bond distances are in Å. CBS-QB3 adiabatic ionization energies (AIE) are in eV. The enthalpy of formation ΔfH°0 and the heat content function H°298 − H°0 are in kJ mol−1. Calculations for 1-cyclohexen-1ol, 2-cyclohexen-1-ol, and cyclohexanone are given in earlier work.14 a



RESULTS AND DISCUSSION The initial reaction of OH with cyclohexene proceeds either by addition of OH to the double bond, forming a 2hydroxycyclohexyl radical or by abstraction of an H/D atom to form a cyclohexenyl radical. Under the conditions of the present experiments, these initial radical products can react to form closed-shell products. The time profiles of the OH addition adduct, the 2-hydroxycyclohexyl-d10 radical (m/z = 109), and the m/z = 107 product in the OH + C6D10 reaction, are shown in Figure 2. Formation of m/z = 107 corresponds to the loss of a deuterium atom from 2-hydroxycyclohexyl-d10. The 2-hydroxycyclohexyl radical is relatively rapidly removed under these conditions, and the m/z = 107 product is formed rapidly. The OH radicals also react with the H2O2 precursor, forming HO2 radicals; both HO2 and OH can react with the radical products of OH + cyclohexene. Furthermore, reactions with the unavoidable residual O2 from the H2O2 source can produce some of the observed products. Varying the initial concentrations of reactants can help disentangle these pathways, as shown in Figure 3. A summary of the experimental findings is given in Table 2. The contributions of reactions of 2-hydroxycyclohexyl radicals with OH or HO2 to the observed product signals can be qualitatively isolated by increasing the concentration of the cyclohexene reactant. At higher cyclohexene concentrations the removal of OH becomes more and more dominated by the

1-ol were carried out using the composite method CBSQB3,30,31 which provides reliable energetics.31 Calculations of stationary points (wells and transition states; note that we did not attempt to locate a van der Waals complex in the addition reaction) on the C6H11O surface were performed at the G3MP2B3 level.32 The Gaussian09 suite of programs33 was employed for quantum chemistry calculations. The predicted onset of ionization is determined by the adiabatic ionization energy (AIE), which is defined as the electronic transition from the vibronic ground state of the neutral to the vibronic ground state of the cation. Spectral simulations of the Franck−Condon (FC) overlap of these states are performed using the PESCAL program.34 Selected bond lengths and the energetics of the species of interest (Scheme 2) are presented in Table 1, and results for several other associated species are given in the Supporting Information. The thermochemistry of these species is calculated relative to atomization enthalpies, as described previously.14 The reliability of these calculations is discussed elsewhere14 and they are found to be in good agreement with literature determinations. Based on the experimental errors and computational accuracy of 4−5 kJ mol−1 given by Montgomery et al.,31 the total uncertainty for the calculated thermodynamic quantities is estimated to be 10 kJ mol−1. 6722

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730

The Journal of Physical Chemistry A

Article

Table 2. Products Identified in This Worka

Figure 2. Time traces of reaction products, measured at 10.1 eV photoionization, initiated by the reaction of OH with cyclohexene-d10, with an initial C6D10 concentration of 3 × 1013 cm−3.

reaction with cyclohexene and the reactions of OH with H2O2 (producing HO2) or with 2-hydroxycyclohexyl are suppressed. For larger concentrations of cyclohexene the decay of 2hydroxycyclohexyl and the rise of products are slower, as shown in Figure 4, suggesting that a component of the 2hydroxycyclohexyl removal at lower cyclohexene concentrations may be due to reaction with HO2 or OH. Scaling the product-formation and reactant-depletion signals by the measured photoionization cross sections shows that, for experiments with a cyclohexene-d10 concentration of 3 × 1013

a

Uncertainty in product ratios is estimated as ±30%.

cm−3, the production of 2-cyclohexen-1-ol represents approximately 5−10% of the depletion of cyclohexene, and the sum of

Figure 3. Dependence of varying reaction pathways on changes in O2 or cyclohexene concentration. Examples are shown for the oxidation initiated by OH reaction with C6D10, primary products of which include 2-hydroxycyclohexyl-d10 (m/z = 109) and cyclohexen-3-yl-d9 radical (m/z = 90). Subsequent unimolecular, bimolecular, or heterogeneous reaction of these primary products will result in the other product species observed. If reaction with O2 is responsible, the product will increase with increasing O2 concentration. If reaction with OH or HO2 is responsible, the product will be decreased by increasing O2 concentration (because the reaction with O2 will compete with the reaction with OH or HO2) and will decrease relative to the primary products as cyclohexene concentration is increased (because OH and HO2 concentrations are reduced, both because OH preferably reacts with cyclohexene with increasing cyclohexene concentration). Products that are diminished by increasing O2 concentration, but unchanged relative to primary products upon increasing cyclohexene concentration, must arise from other reactions (including possible unimolecular dissociation) of the primary products. 6723

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730

The Journal of Physical Chemistry A

Article

Figure 4. Measured time traces of m/z = 107 products initiated by the reaction of OH with cyclohexene-d10, as a function of initial C6D10 concentration. The traces are taken with 10.1 eV photoionization and are normalized by the total cyclohexene depletion. Higher cyclohexene concentration increases the yield of the m/z = 107 product relative to the cyclohexene consumed.

Figure 5. Comparison of relative product signals initiated by OH + cyclohexene-d10, taken at 10.1 eV photon energy and integrated over 0−60 ms after photolysis, as a function of cyclohexene-d 10 concentration. The signals are scaled to coincide at the peak of the m/z = 107 signal. Higher cyclohexene concentrations reduce the contribution of 2-hydroxycyclohexyl reactions with OH and HO2. Increasing cyclohexene concentration reduces the contributions of m/z = 104 and 108 relative to m/z = 107, whereas the relative contributions of the other masses shown remain approximately constant.

2-cyclohexen-1-ol and 1-cyclohexen-1-ol is estimated to be ∼10−20% of cyclohexene depletion. The sum of all observed products at low cyclohexene concentration is well below that required for molar balance; the “missing” cyclohexene may form low-volatility products that are not detected in the mass spectrum. Other experiments in our laboratory that attempted to measure room-temperature oxidation of slightly larger unsaturated ring hydrocarbons, for example, limonene, using similar techniques to the present work, failed because the reactions formed nonvolatile products. Indeed, oxidation of terpenes is a source of aerosols in the troposphere.35 However, the overall yield of observed products, relative to the cyclohexene depletion, increases substantially with increasing cyclohexene concentration, as shown in Figure 4 for the m/ z = 107 product, indicating that whatever chemistry leads to the undetected, possibly nonvolatile, products may be related to secondary chemistry involving OH or HO2. Figure 5 shows the relative integrated product signals at m/z = 104−110 from the photolysis of H2O2 in the presence of varying concentrations of cyclohexene-d10, scaled to the m/z = 107 peak (D atom loss from the OH-addition adduct). The relative yield of C6D10O at m/z = 108 is suppressed by higher cyclohexene-d 10 concentrations, suggesting that this mass is also formed as a product of reactions of 2-hydroxycyclohexyl with OH or HO2. The exact nature of the reaction that forms the observed final products from the initially formed 2-hydroxycyclohexyl is not completely clear. Figures 6 and 7 show calculated pathways for unimolecular reaction of 2-hydroxycyclohexyl radicals with the lowest-lying barriers to bimolecular product channels. More pathways are included in Figures S2 and S3 in the Supporting Information. All unimolecular decomposition pathways are calculated to traverse transition states that lie higher in energy than the OH + cyclohexene asymptote, but given the uncertainty in the G3MP2B3 calculations, some of the channels might be accessible. Nevertheless, the rise of m/z = 107 products (∼600 s−1 for [C6D10] = 1.3 × 1014 cm−3) is much faster than would be expected for thermal decomposition of 2hydroxycyclohexyl radicals. In any case, the removal of 2hydroxycyclohexyl and formation of final products is not via

Figure 6. Stationary point energies at 0 K for C−H scission pathways of the 2-hydroxycyclohexyl radical, calculated at the G3MP2B3 level. The isomerization to the cyclohexoxy radical is also shown. More pathways including other hydrogen-shift reactions from 2-hydroxycyclohexyl and formation of energetically disfavored bimolecular products are included in Figure S2 in the Supporting Information. A full description of the decomposition pathways of the cyclohexoxy radical is given by Hoyermann et al.37 A potential van der Waals complex in the OH + C6H10 entrance channel has not been characterized here.

reaction with residual oxygen − when oxygen is added ([O2] = 1016 cm−3), the 2-hydroxycyclohexyl becomes nearly undetectable, and the reaction products change substantially, as shown in Figure 8. The change in the observed product spectrum tends to confirm that the products arise from reactions of the initial radical products of OH + cyclohexene, but also implies that the C6H10O products, which are diminished by addition of O2, arise from a reaction other than that with molecular oxygen. 6724

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730

The Journal of Physical Chemistry A

Article

subsequent loss of a second D atom, and at m/z = 107, corresponding to addition of OH and subsequent loss of D. The contribution of the m/z = 108 product, which would arise from the addition of OH and subsequent loss of the hydroxyl H, is much smaller than the m/z = 107 product signal, showing that the dominant loss of hydrogen atom from the 2hydroxycyclohexyl radical is from the ring. Similarly, as seen in Figure 1, the m/z = 101 product initiated by OH + cyclohexene-4,4,5,5-d4 is much smaller than the m/z = 102 product, implying that the atom lost from the ring is from a carbon either in or adjacent to the cyclohexene double bond. The isomeric nature of the observed products can be determined from the photoionization spectra at each mass. Products Initiated by OH Addition to Cyclohexene. The earlier study of OH-initiated cyclohexene oxidation measured products with the sum formula C6H10O, which were assigned to cyclic unsaturated alcohols (1-cyclohexen-1-ol and 2cyclohexen-1-ol) and (mistakenly, as it turns out) the acyclic aldehyde hex-5-enal.14 A satisfactory fit to the C6H6D4O (m/z = 102) product spectrum initiated from the reaction of OH with cyclohexene-4,4,5,5-d4, shown in Figure 1, requires only cyclic products. The contribution of the 1-cyclohexen-1-ol is clear because it has the lowest ionization energy. The remainder of the signal can be explained by a combination of 2cyclohexen-1-ol and cyclohexanone. An unconstrained fit of the spectrum to the cyclic isomers returns a small contribution of cyclohexanone, with uncertainty estimates on the same order as or larger than the coefficient. A fit including only 2-cyclohexen1-ol and 1-cyclohexen-1-ol has an only slightly larger chisquared value. We conclude that the dominant C6H10O isomers are 1-cyclohexen-1-ol and 2-cyclohexen-1-ol. It can be seen from Figure 6 that these products correspond to the two overall lowest-energy pathways for C−H bond scission of the 2hydroxycyclohexyl radical, although both pathways are predicted to lie higher in energy than the OH + cyclohexene reactants. Formation of cyclohexanone + H is more exothermic, but must traverse a higher barrier for isomerization of 2hydroxycyclohexyl to the cyclohexoxy radical. Of the products initiated by OH addition to cyclohexene-d10, C6D9HO products (involving the loss of deuterium from the ring) would appear at m/z = 107, while C6D10O (resulting from the loss of hydrogen from the hydroxyl group) appears at m/z = 108. In the experiments with 3 × 1013 cm−3 of cyclohexened10, the remaining signal at m/z = 108 after subtraction of the 6.4% 13C isotopic fraction of C6D9HO is only 10% of the m/z = 107 signal, indicating a strong preference for loss from the ring. The photoionization spectrum for the C6D10O product at m/z = 108 (corrected for the 13C isotopic fraction of C6D9HO) is consistent with either cyclohexanone or 2-cyclohexen-1-ol (Figure 9), with an onset above 9 eV and fairly steep rise. Cyclohexanone formation is possible if the hydroxyl H-atom is transferred to the ring to form a cyclohexoxy radical; studies by Hoyermann and co-workers have shown that cyclohexanone + H is the dominant bimolecular product channel of cyclohexoxy decomposition.37 As seen in Figure 6, the transition state for the 2-hydroxycyclohexyl ↔ cyclohexoxy radical isomerization reaction is calculated to lie well above the reactant energy, and in this case it would still be expected that a D atom is lost for OH + C6D10 at the site of the initial OH attack. The apparent formation of cyclohexanone-d10 in this reaction must occur by a more complex mechanism or by secondary reactions. Moreover, as discussed above, the response of the m/z = 108 signal to changes in cyclohexene

Figure 7. Stationary point energies at 0 K for C−C scission pathways of the 2-hydroxycyclohexyl radical, calculated at the G3MP2B3 level. Only the energetically lowest-lying H-loss channels of the C−C scission products are shown. More pathways are included in Figure S3 in the Supporting Information. A potential van der Waals complex in the OH + C6H10 entrance channel has not been characterized here.

Figure 8. Integrated product mass spectra (integrated over 8.5−10.2 eV and 0−60 ms following photolysis) for the reaction of OH with cyclohexene-d10, taken with (blue dashed line) and without (red solid line) added O2. The spectra are normalized by the cyclohexene depletion.

Identification of Products. The formation of C6H10O products in the OH-initiated oxidation of cyclohexene are connected with OH addition to the double bond, followed by loss of an H atom either from the ring or from the hydroxyl group. Measuring the oxidation of perdeuterated cyclohexene (C6D10) initiated by OH permits the products of these two hydrogen-loss pathways to be distinguished by mass, and measurements with partially deuterated cyclohexene were further employed to distinguish products from different abstraction channels. The mass spectrum of the products of the OH + C6D10 reaction, integrated over kinetic time and photoionization energies from 8.2 to 10.2 eV, is shown in Figure 8. Additional measurements at higher photon energies (10.7 and 11.0 eV) were carried out to search for possible acetaldehyde (ionization energy (IE) = 10.23 eV), ethene (IE = 10.51 eV), or formaldehyde (IE = 10.88 eV) products.36 A relatively small amount of partially and completely deuterated formaldehyde was observed, but negligible acetaldehyde or ethene formation was seen. The major products are at m/z = 88, corresponding to abstraction of D from cyclohexene-d10 and 6725

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730

The Journal of Physical Chemistry A

Article

Figure 10. Experimental photoionization spectra for m/z = 82 and 83 products from the C6H6D4 + OH reaction (the m/z = 84 spectrum has a similar shape and magnitude to m/z = 83), compared with the photoionization spectra for 1,4-cyclohexadiene and 1,3-cyclohexadiene.38 Smoothed experimental spectra (small solid symbols) are also given for comparison.

Figure 9. Literature photoionization spectrum for cyclohexanone14 (solid line) superimposed on the experimental curve for C6D10O product (m/z = 108) from C6D10 + OH (open circles), that is, the product resulting from hydrogen loss from the OH group. The 2cyclohexen-1-ol spectrum (dashed line) is a similar fit to the experimental spectrum.

corresponding to 1,4-cyclohexadiene-d2, and m/z = 83/84, which are −d3 and −d4 isotopologs of 1,3-cyclohexadiene.38 The observed ionization energies (∼8.2 eV for 1,3-cyclohexadiene and 8.8 eV for 1,4-cyclohexadiene) are in good agreement with literature values.36 Based on the empirical method of Bobeldijk et al.,39 the cross sections for these two isomers should be nearly identical. A comparison of signal amplitudes indicates that the contribution from 1,3-cyclohexadiene is more than 10 times that of 1,4-cyclohexadiene, that is, >90% of cyclohexadiene consists of the 1,3-cyclohexadiene isomer. If the formation of cyclohexadiene is initiated by abstraction of an H atom from cyclohexene-4,4,5,5-d4 to form the allylic cyclohexen-3-yl radical, then subsequent loss by dissociation or bimolecular reaction of either an H or a D atom can form the 1,3-cyclohexadiene isomer. The signals for m/z = 83 and 84 are approximately equal, suggesting little preference for H or D removal from the isotopically labeled cyclohexen-3-yl radical. Formation of the 1,4-cyclohexadiene isomer is presumably initiated by the abstraction of a D atom from cyclohexene4,4,5,5-d4, which is three times less favorable (see Table 2) because of the likely kinetic isotope effect and because the intermediate radical, the nonresonance-stabilized cyclohexen-4yl radical, is less stable. Moreover, the small contribution of 1,4cyclohexadiene (m/z = 82 is approximately 1/7 of m/z = 83) indicates a substantial preference for H atom loss from the cyclohexen-4-yl-4,5,5-d3 radical (Scheme 3). As in the case of the C6D10O product (m/z = 108) described above, the decrease in the relative contribution of the cyclohexadiene product (m/z = 88) with increasing cyclohexene concentration, shown in Figure 11, suggests that the loss of the second H (or D) atom to form cyclohexadiene occurs by reaction with OH or HO2. In these experiments with C6D10, the signal-to-noise does not permit the contribution of 1,4-cyclohexadiene to be distinguished from the more abundant 1,3-cyclohexadiene product at the same mass. Secondary and Side Chemistry. Other minor products from OH + cyclohexene-4,4,5,5-d4 are observed at m/z = 104 and m/z = 100. The m/z = 104 product, one mass unit higher than the OH-addition adduct (m/z = 103), corresponds to the

concentration (Figure 5) is different than the response of m/z = 107. The decreasing relative contribution of m/z = 108 with increasing cyclohexene concentration suggests that it is a secondary product arising from reactions of the addition product with OH or HO2. The C−C bond scission channels following OH addition are largely unobserved, contrary to the conclusions of the earlier work.14 Although several transition states for C−C scission from 2-hydroxycyclohexyl are calculated to lie relatively low in energy (Figure 7 and Figure S3 in the Supporting Information), the ring-opened radical products have no low-energy dissociation pathways available and they are likely to simply reform the 2-hydroxycyclohexyl radical. Products Initiated by Abstraction. Products initiated by the H-abstraction channels, primarily the cyclohexadienes, were also observed. Reaction with cyclohexene-4,4,5,5-d4 allows the products from abstraction at different sites to be distinguished, as depicted in Scheme 3. Figure 10 shows the experimental product photoionization spectra for both m/z = 82, Scheme 3

6726

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730

The Journal of Physical Chemistry A

Article

product mass spectra with and without added O2 is shown in Figure 8, normalized to the loss of cyclohexene. As discussed above, the yields of the C6D10O and C6D9HO products from OH addition to cyclohexene-d10, followed by H or D atom loss, are decreased by the addition of O2. A higher-mass product (m/z = 123) is observed when O2 is added, with an increasing yield with increased O2 concentration. Its photoionization spectrum is shown in Figure 12. The shape of the photo-

Figure 11. Comparison of relative product signals initiated by OH + cyclohexene-d10, taken at 10.1 eV photon energy and integrated over 0−60 ms following photolysis, as a function of cyclohexene-d10 concentration. The signals are scaled to coincide at the peak of the m/z = 107 signal. Higher cyclohexene concentrations reduce the formation of cyclohexadienes (m/z = 88) and suppress the removal of the cyclohexenyl radical isomers at m/z = 90, resulting in a larger m/z = 90 signal.

isotopic product that appears at m/z = 110 in the reaction of OH with cyclohexene-d10. The photoionization spectrum is in good agreement with a calculated spectrum of cyclohexanol (Figure S4 in the Supporting Information), although the observed threshold at 9.45 eV is a poor match to the ionization energies previously reported in literature (9.75−10 eV).36 The product at m/z = 100 in the OH + cyclohexene-4,4,5,5-d4 system corresponds to the same species as the m/z = 104 product observed in the C6D10 oxidation. The onset of ionization of this product is in reasonable agreement with the AIE of 2-cyclohexenone (AIE 9.26 ± 0.05 eV),40 a possible product from O2 or HO2 reaction with the resonance-stabilized cyclohexen-3-yl radical produced in the abstraction channel of the OH + cyclohexene reaction. The contribution of the cyclohexanol product at m/z = 110 from OH + cyclohexene-d10 initiation is decreased by addition of O2, as shown in Figure 8, suggesting that it arises from a bimolecular reaction of the 2-hydroxycyclohexyl radical. However, the cyclohexanol contribution is not diminished by increasing cyclohexene-d10 concentration (Figure 5), implying that reactions with OH and HO2 are not the dominant source of cyclohexanol. The m/z = 104 product (2-cyclohexenone), on the other hand, is increased by addition of O2 and diminished by increasing cyclohexene-d10 concentration. The 1,3-cyclohexadiene-d8 product at m/z = 88 is decreased by addition of O2 as well as by increased cyclohexene concentration. The cyclohexadiene is presumably a product of reactions of the allylic cyclohexen-3-yl radical that is produced by abstraction of a D atom from the cyclohexene reactant. Reaction of cyclohexen-3-yl with O2 or with HO2 to produce 2-cyclohexenone would be consistent with the observed dependence of the m/z = 88 and 104 signals on cyclohexene and oxygen concentrations. Reaction of the 2-Hydroxycyclohexyl Radical with O2. To discriminate addition−elimination products from products of secondary reaction with O2, the reaction of OH with cyclohexene-d10 was carried out in the presence of 1016 cm−3 of added oxygen. A comparison of the observed integrated

Figure 12. Calculated photoionization spectrum for 2,3-epoxycyclohexan-1-ol (solid line) and the integrated photoelectron spectrum41 for 2-hydroxycyclohexanone (dashed line) superimposed on the experimental photoionization spectrum for m/z = 123 (open circles) from the C6D10 + OH system. The m/z = 124 signal (corrected for 13 C contributions from m/z = 123) is shown for comparison.

ionization spectrum and the onset near 9.5 eV are in good agreement with a Franck−Condon simulation based on CBSQB3 calculations for 2,3-epoxycyclohexan-1-ol, which is one possible product from the reaction of O2 with the stabilized 2hydroxycyclohexyl radical (Scheme 4). Figure 12 also shows the integral of the reported photoelectron spectrum41 of 2-hydroxycyclohexanone, another conceivable product from the reaction of 2-hydroxycyclohexyl with O2 (Scheme 4). The photoionization spectrum of the m/z = 123 product from OH-initiated C6D10 oxidation agrees only slightly more closely with 2,3-epoxycyclohexan-1-ol than with 2-hydroxycyclohexanone. However, the four-membered ring transition state for 2-hydroxycyclohexanone formation is expected to be disfavored relative to the five-membered ring that leads to 2,3-epoxycyclohexan-1-ol. Nevertheless, it is certainly possible that both species contribute to the m/z = 123 product. The rise of the m/z = 123 signal is rapid, ruling out self-reaction of stabilized 2-hydroxycyclohexylperoxy radicals as the source of that product. Other potential epoxycyclohexanol products have substantially different AIE values and photoionization spectra. The CBS-QB3 AIE values are calculated as 8.32 eV for 1,2epoxycyclohexan-1-ol, 8.80 eV for 2,4-epoxycyclohexan-1-ol, 8.97 eV for 2,5-epoxycyclohexan-1-ol, and 7.85 eV for 2,6epoxycyclohexan-1-ol. The cation of 2,6-epoxycyclohexan-1-ol is unstable with respect to ring-opening via C−C bond fission breaking the C1−C2 bond. 1,2-Epoxycyclohexan-1-ol is a three-membered-ring lactol (i.e., contains an oxirane ring with an OH group on the carbon atom in α-position to the ethereal O atom). Its cation is calculated to be unstable with respect to ring-opening as well, in agreement with calculations on other 6727

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730

The Journal of Physical Chemistry A

Article

Scheme 4

three-membered-ring lactols.29 As a consequence, the potential 1,2-epoxycyclohexan-1-ol (and 2,6-epoxycyclohexan-1-ol) products may have poor photoionization cross sections near threshold or exhibit substantial dissociative ionization, and the lack of signal at the parent mass may not necessarily imply that they are absent. Another possible product of the reaction of 2-hydroxycyclohexyl with O2 is hexanedial (calculated AIE 9.72 eV), formed via internal abstraction of the hydroxyl H atom in the 2hydroxycyclohexylperoxy adduct (the “Waddington mechanism”).42,43 This product would appear at m/z = 124 in the C6D10 oxidation system. Although there is a small m/z = 124 signal, it is less than 10% of the m/z = 123 signal after correcting for 13C contributions. Based on this observation and assuming that hexanedial photoionization results in a significant ion signal at its parent mass, it appears that the hexanedial formation is substantially disfavored relative to the epoxidation channel at 298 K. Considering the previous erroneous interpretation of the photoionization spectrum of C6H10O products,14 one would clearly urge caution for assignments of overlapping continua based solely on Franck−Condon calculations. However, in this case, the relative lack of Waddington product is inferred from the isotope ratio in the products, and does not rely on the assignment of photoionization spectra. No authentic photoionization spectrum of hexanedial is available; it is conceivable that the hexanedial undergoes dissociative ionization, reducing the amount of signal observed at the parent mass. The major dissociative ionization channels for an aldehyde44 include α-C−H scission, which would form an ion at m/z = 122 in the present case (by D atom loss); α-C−C scission, which would form an ion at m/z = 30 in the present case (DCO+); β-cleavage, which would form an ion at m/z = 78 (C4D7O+) or m/z = 46 (C2D3O+); and the transfer of a deuterium from the 4-carbon to the carbonyl oxygen (“McLafferty rearrangement”),45,46 which would produce a

perdeuterated ethenol cation at m/z = 48 (Scheme 5). At most, very small amounts of these cation masses are observed. Loss of Scheme 5

D2O after ionization of hexanedial-d8 would leave a cation at m/z = 104; however, the energetic onset of the m/z = 104 signal is almost 0.5 eV lower than the calculated adiabatic ionization energy of hexanedial, and no feature is observed at higher energy. Nevertheless, there might be other dissociative ionization channels available for hexanedial, and we are therefore cautious in ruling out formation of hexanedial via the Waddington mechanism. For example, loss of ethene47 would place the dissociative ionization product ion at the same nominal mass as the cyclohexene reactant. Finally, a product at m/z = 76 is also increased by addition of O2. The exact mass corresponds to C4D6O (see Figure S5 in the Supporting Information), and its photoionization spectrum displays an onset near 9.7 eV. This may be a butenal isomer or may arise from dissociative ionization of m/z = 123 or 124 products; note that the onset is close to the calculated AIE of hexanedial of 9.72 eV. This product would appear at m/z = 70 for the perprotio system. Earlier experiments14 in which the mass resolution was far poorer and the O2 background was higher noted more apparent dissociative ionization signal from hex-5-enal product, assigned at m/z = 69, than the directly measured dissociative ionization fraction from a calibration 6728

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730

The Journal of Physical Chemistry A

Article

Present Address

measurement of hex-5-enal, suggesting that some other product could make m/z = 69 in that system. Reanalysis of the original mass spectra (both the reactive data and the hex-5-enal calibration) shows that this peak in the perprotio system is in fact at m/z = 70, and it appears plausible that at least some of the excess signal in the earlier OH + cyclohexene measurements may have been affected by m/z = 70 product from reaction of the 2-hydroxycylohexyl radical with residual O2. It is conceivable that the prominent m/z = 76 (m/z = 70 in the reaction of C6H10) arises from dissociative ionization of hexanedial. It, however, seems unlikely that the charge would reside on the C4D6O fragment in Scheme 5 because the ionization energy of 3-butenal (calculated to be 9.68 eV at CBS-QB3 level, in agreement with 9.65 eV estimated from isodesmic reactions48) is substantially higher than that of ethenol (anti-CH2CHOH AIE = 9.17 eV; syn-CH2CHOH AIE = 9.3 eV).49,50 The Waddington mechanism is prominent in the reactions of other β-hydroxyalkyl radicals with O2,29,51 so its apparently minor role in 2-hydroxycyclohexyl-d10 + O2 would be somewhat surprising, particularly as the kinetic isotope effect should disfavor the D-atom transfer that leads to the epoxide product.



Department of Chemistry and Biochemistry, University of CaliforniaSan Diego, 9500 Gilman Drive, MC 0332, La Jolla, CA 92093−0332. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS G.M. acknowledges the University of San Francisco Faculty Development Fund for financial support and the Advanced Light Source (ALS) division at the Lawrence Berkeley National Laboratory for beamtime allocation. The ALS is supported by the Director, Office of Science, Office of Basic Energy Sciences of the U.S. Department of Energy under Contract No. DEAC02-05CH11231. This work is also supported by the Division of Chemical Sciences, Geosciences, and Biosciences, the Office of Basic Energy Sciences, and the U.S. Department of Energy. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the National Nuclear Security Administration under Contract No. DEAC04-94-AL85000.





CONCLUSIONS The products from the OH-initiated oxidation of cyclohexene have been characterized, including analysis of product formation mechanisms using isotopically labeled precursors. The data are consistent with both addition−elimination and Habstraction mechanisms for the OH + cyclohexene reaction. Direct measurement of the photoionization spectrum of 2cyclohexen-1-ol shows that, in contrast to earlier interpretations, the C6H10O products from OH + cyclohexene all retain the C6 ring. The present results show a particular need for caution when interpreting overlapping contributions from multiple isomers using only calculated Franck−Condon factors. Although calculated ionization energies are reasonably reliable, and computed Franck−Condon factors often give correct shapes for photoionization spectra, these assume pure continuum ionization to one cationic state, and direct measurements are inherently more accurate. Both 1,3- and 1,4-cyclohexadiene products were observed as a result of hydrogen abstraction to form cyclohexenyl radical. The 1,3cyclohexadiene isomer was found to be dominant, more than 90% of the total cyclohexadiene. Reaction of the 2hydroxycyclohexyl radical with O2 produces 2-hydroxycyclohexanone and 2,3-epoxycyclohexan-1-ol (the latter identified by its calculated adiabatic ionization energy and Franck−Condon simulation of the photoionization spectra) and possibly other epoxidized hydroxycyclohexanes. The product isotope ratios in the OH-initiated oxidation of cyclohexene-d10 suggested only a small role for the hexanedial “Waddington” product. However, this channel might be masked if dissociative ionization substantially reduces the ion signal at the parent mass.



(1) Khan, M. R.; Seshadri, K. S.; Kowalski, T. E. Energy Fuels 1989, 3, 412−420. (2) Knepp, A. M.; Meloni, G.; Jusinski, L. E.; Taatjes, C. A.; Cavallotti, C.; Klippenstein, S. J. Phys. Chem. Chem. Phys. 2007, 9, 4315−4331. (3) Gulati, S. K.; Walker, R. W. J. Chem. Soc., Faraday Trans. 2 1989, 85, 1799−1812. (4) Lemaire, O.; Ribaucour, M.; Carlier, M.; Minetti, R. Combust. Flame 2001, 127, 1971−1980. (5) Voisin, D.; Marchal, A.; Reuillon, M.; Boettner, J. C.; Cathonnet, M. Combust. Sci. Technol. 1998, 138, 137−158. (6) Bennett, P. J.; Gregory, D.; Jackson, R. A. Combust. Sci. Technol. 1996, 115, 83−103. (7) Law, M. E.; Westmoreland, P. R.; Cool, T. A.; Wang, J.; Hansen, N.; Taatjes, C. A.; Kasper, T. Proc. Combust. Inst. 2007, 31, 565−573. (8) Silke, E. J.; Pitz, W. J.; Westbrook, C. K.; Ribaucour, M. J. Phys. Chem. A 2007, 111, 3761−3775. (9) Miller, J. A.; Pilling, M. J.; Troe, J. Proc. Combust. Inst. 2005, 30, 43−88. (10) Gomez, A.; Sidebotham, G.; Glassman, I. Combust. Flame 1984, 58, 45−57. (11) McEnally, C. S.; Pfefferle, L. D. Combust. Sci. Technol. 1998, 131, 323−344. (12) Ghigo, G.; Tonachini, G. J. Am. Chem. Soc. 1998, 120, 6753− 6757. (13) Taatjes, C. A.; Osborn, D. L.; Selby, T. M.; Meloni, G.; Trevitt, A. J.; Epifanovsky, E.; Krylov, A. I.; Sirjean, B.; Dames, E.; Wang, H. J. Phys. Chem. A 2010, 114, 3355−3370. (14) Meloni, G.; Selby, T. M.; Osborn, D. L.; Taatjes, C. A. J. Phys. Chem. A 2008, 112, 13444−13451. (15) Taatjes, C. A.; Hansen, N.; McIlroy, A.; Miller, J. A.; Senosiain, J. P.; Klippenstein, S. J.; Qi, F.; Sheng, L. S.; Zhang, Y. W.; Cool, T. A.; et al. Science 2005, 308, 1887−1889. (16) Taatjes, C. A.; Hansen, N.; Miller, J. A.; Cool, T. A.; Wang, J.; Westmoreland, P. R.; Law, M. E.; Kasper, T.; Kohse-Höinghaus, K. J. Phys. Chem. A 2006, 110, 3254−3260. (17) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F.; Kerr, J. A.; Rossi, M. J.; Troe, J. J. Phys. Chem. Ref. Data 1997, 26, 521−1011. (18) Hippler, H.; Viskolcz, B. Phys. Chem. Chem. Phys. 2000, 2, 3591−3596. (19) Cool, T. A.; Nakajima, K.; Mostefaoui, T. A.; Qi, F.; McIlroy, A.; Westmoreland, P. R.; Law, M. E.; Poisson, L.; Peterka, D. S.; Ahmed, M. J. Chem. Phys. 2003, 119, 8356−8365.

ASSOCIATED CONTENT

S Supporting Information *

Additional analytical data and figures. This material is available free of charge via the Internet at http://pubs.acs.org.



REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. 6729

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730

The Journal of Physical Chemistry A

Article

(20) Meloni, G.; Zou, P.; Klippenstein, S. J.; Ahmed, M.; Leone, S. R.; Taatjes, C. A.; Osborn, D. L. J. Am. Chem. Soc. 2006, 128, 13559− 13567. (21) Osborn, D. L.; Zou, P.; Johnsen, H.; Hayden, C. C.; Taatjes, C. A.; Knyazev, V. D.; North, S. W.; Peterka, D. S.; Ahmed, M.; Leone, S. R. Rev. Sci. Instrum. 2008, 79, 104103 1−10. (22) Egerton, A. C.; Emte, W.; Minkoff, G. J. Discuss. Faraday Soc. 1951, 10, 278−282. (23) Maass, O.; Hiebert, P. G. J. Am . Chem. Soc. 1924, 46, 2693− 2700. (24) Atkinson, R. Atmos. Chem. Phys. 2003, 3, 2233−2307. (25) Sander, S. P.; Friedl, R. R.; Ravishankara, A. R.; Golden, D. M.; Kolb, C. E.; Kurylo, M. J.; Huie, R. E.; Orkin, V. L.; Molina, L. T.; Moortgat, G. K.; Finlayson-Pitts, B. J. Chemical Kinetics and Photochemical Data for Use in Stratospheric Modeling: Evaluation Number 16; The Jet Propulsion Laboratory: Pasadena, CA., 2010. (26) Goulay, F.; Osborn, D. L.; Taatjes, C. A.; Zou, P.; Meloni, G.; Leone, S. R. Phys. Chem. Chem. Phys. 2007, 9, 4291−4300. (27) Goulay, F.; Trevitt, A. J.; Meloni, G.; Selby, T. M.; Osborn, D. L.; Taatjes, C. A.; Vereecken, L.; Leone, S. R. J. Am. Chem. Soc. 2009, 131, 993−1005. (28) Cool, T. A.; Wang, J.; Nakajima, K.; Taatjes, C. A.; McIlroy, A. Int. J. Mass. Spectrom. 2005, 247, 18−27. (29) Welz, O.; Zádor, J.; Savee, J. D.; Ng, M. Y.; Meloni, G.; Fernandes, R. X.; Sheps, L.; Simmons, B. A.; Lee, T. S.; Osborn, D. L.; Taatjes, C. A. Phys. Chem. Chem. Phys. 2012, 14, 3112−3127. (30) Montgomery, J. A.; Frisch, M. J.; Ochterski, J. W.; Petersson, G. A. J. Chem. Phys. 1999, 110, 2822−2827. (31) Montgomery, J. A.; Frisch, M. J.; Ochterski, J. W.; Petersson, G. A. J. Chem. Phys. 2000, 112, 6532−6542. (32) Baboul, A. G.; Curtiss, L. A.; Redfern, P. C.; Raghavachari, K. J. Chem. Phys. 1999, 110, 7650−7657. (33) 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, Rev. A.02, Gaussian, Inc.: Wallingford, CT, 2009. (34) Ervin, K. M. PESCAL, Fortran program. 2009. (35) Hoffmann, T.; Odum, J. R.; Bowman, F.; Collins, D.; Klockow, D.; Flagan, R. C.; Seinfeld, J. H. J. Atmos. Chem. 1997, 26, 189−222. (36) NIST Chemistry WebBook, NIST Standard Reference Database Number 69; Linstrom, P. J., Mallard, W. G., Eds.; National Institute of Standards and Technology: Gaithersburg, MD, 2003. (37) Hoyermann, K.; Maarfeld, S.; Nacke, F.; Nothdurft, J.; Olzmann, M.; Wehmeyer, J.; Welz, O.; Zeuch, T. Phys. Chem. Chem. Phys. 2010, 12, 8953−8967. (38) Wang, J.; Yang, B.; Cool, T. A.; Hansen, N.; Kasper, T. Int. J. Mass Spectrom. 2008, 269, 210−220. (39) Bobeldijk, M.; van der Zande, W. J.; Kistemaker, P. G. Chem. Phys. 1994, 179, 125−130. (40) Holmes, J. L.; Terlouw, J. K.; Vijfhuizen, P. C.; A’Campo, C. Org. Mass. Spectrom. 1979, 14, 204−212. (41) Brown, R. S. Can. J. Chem. 1976, 54, 3203−3205. (42) Ray, D. J. M.; Diaz, R. R.; Waddington, D. J. Proc. Combust. Inst. 1973, 14, 259−266. (43) Ray, D. J. M.; Waddington, D. J. Combust. Flame 1973, 20, 327− 334. (44) Morgan, R. P.; Derrick, P. J.; Loudon, A. G. J. Chem. Soc., Perkin Trans. 2 1980, 306−312. (45) Kingston, D. G. I.; Bursey, J. T.; Bursey, M. M. Chem. Rev. 1974, 74, 215−242. (46) Gilpin, J. A.; McLafferty, F. W. Anal. Chem. 1957, 29, 990−994. (47) Liedtke, R. J.; Djerassi, C. J. Am . Chem. Soc. 1969, 91, 6814− 6821. (48) Takhistov, V. V.; Ponomarev, D. A. Org. Mass. Spectrom. 1994, 29, 395−412. (49) Ruscic, B.; Berkowitz, J. J. Chem. Phys. 1994, 101, 10936−10946. (50) Matti, G. Y.; Osman, O. I.; Upham, J. E.; Suffolk, R. J.; Kroto, H. W. J. Electron Spectrosc. Relat. Phenom. 1989, 49, 195−201.

(51) Zádor, J.; Fernandes, R. X.; Georgievskii, Y.; Meloni, G.; Taatjes, C. A.; Miller, J. A. Proc. Combust. Inst. 2009, 32, 271−277.

6730

dx.doi.org/10.1021/jp3022437 | J. Phys. Chem. A 2012, 116, 6720−6730