Ozonolysis of Cyclic Alkenes as Surrogates for Biogenic Terpenes

Jun 25, 2010 - Ozonolysis of Cyclic Alkenes as Surrogates for Biogenic Terpenes: .... by nuances in the molecular structure of volatile organic compou...
0 downloads 0 Views 2MB Size
Download PDF
J. Phys. Chem. A 2010, 114, 7509–7515

7509

Ozonolysis of Cyclic Alkenes as Surrogates for Biogenic Terpenes: Primary Ozonide Formation and Decomposition Scott A. Epstein and Neil M. Donahue* Center for Atmospheric Particle Studies, Carnegie Mellon UniVersity, Pittsburgh, PennsylVania 15213 ReceiVed: March 10, 2010; ReVised Manuscript ReceiVed: May 11, 2010

Alkene ozonolysis reactions proceed through an unstable intermediate, the primary ozonide (POZ). POZ decomposition controls the complex mechanism. We probe the kinetics of primary ozonide decomposition using temperature programmed reaction spectroscopy (TPRS), revealing primary ozonide decomposition barrier heights of 9.1 ( 0.4, 9.4 ( 0.4, and 11.9 ( 1.2 kcal mol-1 for cyclohexene, 1-methyl-cyclohexene, and methylene-cyclohexane, respectively. We compare experimental decomposition spectra with spectral predictions using density functional theory (DFT) to reveal decomposition products resembling vinyl-hydroperoxides and dioxiranes. We do not find evidence of secondary ozonides. Additional computations with DFT, scaled with the TPRS barrier height, yield barrier heights ranging from 9.4 to 12.1 kcal mol-1 for the four competing decomposition pathways of the 1-methyl-cyclohexene POZ. Entropic differences were minimal, indicating that POZ decomposition branching is controlled purely by enthalpic variations. These kinetic computations were used to calculate a hydroxyl radical yield for 1-methyl-cyclohexene ozonolysis of 0.85 at 298 K. Introduction Alkenes comprise approximately 50% of total volatile organic carbon (VOC) emissions to the atmosphere.1 Reaction with ozone is the most important sink for many alkenes and thus determines their eventual fate in the atmosphere. Gas-phase ozonolysis reactions can be a significant source of low-volatility products, leading to secondary organic aerosol (SOA) formation,2,3 and they also produce significant quantities of hydroxyl radical4-8sthe most important oxidant in the troposphereseven at night. The ozonolysis mechanism is complex, involving several branch points and in some cases yielding large suites of productsssome of which cannot be identified.9,10 Furthermore, ozonolysis reactions are highly exothermic, with several very weakly bound and reactive intermediates among the initial reaction products. In combination, these facts mean that the gasphase and condensed-phase mechanisms can differ, as collisional energy transfer in the gas phase can be slower than the swift unimolecular reactions.11 One objective of atmospheric chemists is to understand how bulk observables such as the SOA mass yield vary with the presence or type of an OH radical scavenger,12 NOx levels,13,14 or temperature.15,16 Another is to understand how OH yields vary for different alkenes. These objectives intersect with a desire to understand the fundamental reaction dynamics, because those dynamics control the product distribution.17 The fundamental understanding of ozonolysis began with the Criegee mechanism.18 The first step in every ozonolysis mechanism consists of ozone attacking the alkene double bond via 1,3-dipolar cycloaddition to form an unstable 1,2,3-trioxolane called the primary ozonide (POZ). The POZ then undergoes an exothermic decomposition to form a highly reactive intermediate with both zwitterionic and biradical properties (the Criegee intermediate, or carbonyl-O-oxide, referred to here as the COO) and either an aldehyde or ketone (the carbonyl coproduct, CCP), depending on the nature of the alkene. The COO and CCP can * To whom correspondence should be addressed. E-mail: [email protected].

then react with each other in a second 1,3-dipolar cycloaddition, forming a 1,2,4-trioxalane called the secondary ozonide (SOZ). However, the COO can also proceed down several unimolecular and bimolecular pathways leading to dioxiranes, vinyl-hydroperoxides, and hydroxyl radicals.19 This work focuses on the first branch point in this complex sequence: POZ decomposition. Decomposition of the POZ is generally thought to be a concerted cycloreversion, with simultaneous O-O and C-C bond scission,19 although in some cases a stepwise sequence starting only with O-O scission has been proposed.20 Either O-O bond in the POZ can break, leading to different CCP/COO pairs in the concerted cycloreversion for asymmetric systems. This can have dramatic consequences for both SOA formation21 and OH formation.22 In the most extreme case, a POZ formed at a terminal double bond can decompose into either formaldehyde and a high carbon number COO, or formaldehyde-O-oxide and a high carbon number aldehyde. The first case might lead to high OH radical and SOA yields, while the second might lead to very low yields of both. The general wisdom is that the COO favors the more substituted product,19 but the preference is often not strong. We strive to constrain the ozonolysis mechanism by understanding the branching between these different CCP/COO pairs and whether the selectivity of POZ cycloreversion is controlled more by different barrier heights among possible pathways or by different pre-exponential (steric) factors. In other words, we hope to discover whether energy or entropy governs the reaction mechanism. We are especially interested in cyclic alkenes because the cycloreversion products in these systems are tethered, leading to higher carbon number and thus lower vapor pressure reaction products. Many monoterpenes, such as R-pinene, are endocyclic, and they are known to be efficient SOA sources.2,3 Terpenes are highly asymmetrical and thus form a rich array of products.23 Here we shall begin by investigating the ozonolysis of cyclohexene, which serves as a symmetrical model system for cyclic alkene ozonolysis. Many research groups have examined the cyclohexene mechanism and product distribution,11,24-26 but our

10.1021/jp102177v  2010 American Chemical Society Published on Web 06/25/2010

7510

J. Phys. Chem. A, Vol. 114, No. 28, 2010

Epstein and Donahue 1,2,3-trioxilane (the primary ozonide, MECYHEpoz). The MECYHEpoz has two potential cycloreversion channels: cleavage of the O-O bond distal from the methyl subsistent or cleavage of the O-O bond proximal to the methyl subsistent. The first channel leads to two possible tethered CCP/COO species6 that can go on to produce similar aldehydic vinylhydroperoxides or cyclize to form a secondary ozonide (MECYHEsoz). The second channel produces both a syn and anti tethered CCP/COO species. The syn species can undergo a hydrogen shift to form a ketonic vinyl-hydroperoxide (MECYHEvhpk), and the anti species can isomerize to form a dioxirane (MECYHEdiok). Both the anti and syn COO species can cyclize to form the secondary ozonide (MECYHEsoz), although the relative speed of SOZ formation may be very sensitive to the COO conformation.11 A principal objective of this study is to constrain the relative barrier heights of the two POZ cycloreversion channels and, ultimately, to understand the physics controlling this branching. Methylene-cyclohexane involves the oxidation of an exocyclic double bond leading to a significantly different mechanism than the previous two species. Ozonolysis of MECYHA produces a primary ozonide (MECYHApoz), which ultimately decomposes in two different ways to produce two distinct CCP/COO pairs. Each CCP/COO pair can undergo recombination to yield the secondary ozonide (path 3, MECYHAsoz), or the COO can react to form a vinyl-hydroperoxide (path1, MECYHAcyhp) or a dioxirane (path 2, MECYHAcdio and path 4, dio). Since decomposition of the POZ determines the subsequent product distribution, we strive to capture the kinetics of this step with temperature programmed reaction spectroscopy (TPRS) and computational chemistry predictions using DFT.

Figure 1. (a) Cyclohexene, (b) 1-methyl-cyclohexene (MECYHE), and (c) methylene-cyclohexane (MECYHA) ozonolysis mechanism.

focus is on the POZ. We combine experiments studying the decomposition of cyclohexene POZ with Density Functional theory (DFT) calculations to elucidate the behavior of this model system. Subsequent oxidation experiments and DFT simulations with 1-methyl-cyclohexene serve as a surrogate for β-pinene ozonolysis; additional experiments with methylene-cyclohexane serve as a surrogate for R-pinene ozonolysis. Experiments with all three species enable the comparison of POZ decomposition barrier heights and product distributions. Mechanism Ozonolysis mechanisms for cyclohexene (CYHE), 1-methylcyclohexene (MECYHE), and methylene-cyclohexane (MECYHA) are shown in Figures 1a, 1b, and 1c, respectively. Ozonolysis of CYHE produces a 1,2,3-trioxolane (CYHEpoz), which undergoes unimolecular decomposition forming a tethered CCP and COO in which the O-oxide is either syn or anti relative to the carbon adjacent to the carbonyl-O-oxide.11 The anti configuration can oligomerize, decompose, or isomerize to form a dioxirane (CYHEdio);19,22,27,28 several previous studies indicate that dioxirane formation is the preferred pathway for anti carbonyl-O-oxides,11,29,30 although this is still debated. The syn configuration allows for a hydrogen shift leading to a vinylhydroperoxide (CYHEvhp).12 In the gas phase, the vinylhydroperoxide is thought to be the major immediate precursor for OH formation,20,22,29 and so varying OH yields are generally attributed to different syn-anti branching ratios in different systems.31 Both the syn and anti configuration can form a 1,2,4trioxilane (the secondary ozonide, CYHEsoz).11 1-Methyl-cyclohexene ozonolysis is more complicated due to its asymmetric nature. Like all alkenes, MECYHE forms a

Experimental Methods We perform and observe ozonolysis chemistry on a transparent coldfinger. Details of the experimental methods are presented elsewhere,32 but here we provide a brief summary. A cryogenic zinc-selenide surface, cooled by thermal contact with liquid nitrogen, is used to isolate the POZ after the condensed precursor alkene is reacted with ozone gas directed at the cold window. Significant alkene coatings on the order of 10 000 molecular layers thick are used to minimize surface effects. We warm the infrared-transparent cryogenic surface and passively observe products with Fourier transform infrared (FTIR) spectroscopy. We track selected peak areas as a function of temperature to generate a temperature programmed reaction spectroscopy (TPRS) signal. We select peaks by comparing observed spectra to those predicted using Gaussian 0333 at the DFT B3LYP 6-311G++(2d,2p) level of theory. Our previous publication32 details the POZ peak selection process. In addition, agreement with the previously observed thermal behavior of the cyclohexene POZ in Hull et al.34 helps verify the presence of the POZ. DFT B3LYP 6-311G++(2d,2p) accurately predicts the tetramethylethene POZ spectrum.32 The raw spectra evolve considerably during an experiment, presumably due to condensation of water vapor or other impurities on the ZnSe windows. Consequently, it is not possible to use a background subtraction to generate absorbance spectra; instead, we draw a linear baseline under each absorption peak in the raw spectrum and track the integrated peak areas. The peak area at each temperature is proportional to the concentration of that species on the surface at low total absorbance. Peak area curves exhibit sigmoidal increases if the species that they represent is being formed; sigmoidal decreases indicate that the species of interest is either reacting away or desorbing off of the window.

Ozonolysis of Cyclic Alkenes

J. Phys. Chem. A, Vol. 114, No. 28, 2010 7511

Figure 2. Adsorption peak area as a function of temperature for cyclohexene, MECYHE, and MECYHA POZ decomposition. Dashed green lines indicate the peak reaction temperature for these experiments. The peaks that correspond to each of these curves are presented in the Supporting Information. In Figure 2b, the dashed cyan line indicates the theoretical peak reaction temperature (Tp ) 153 K) of the second most energetically favorable MECYHE POZ decomposition transition state.

We use a modified form of the Redhead equation35 to transform the experimental data into useful kinetic parameters:

( ) ( )

ln

Ea β AR ) ln 2 Ea RTP TP

(1)

where TP, the peak reaction temperature, is defined as the inflection point of the sigmoid that tracks decomposition of the POZ. β is the heating rate of the experiment corresponding to each TP value, Ea is the activation barrier, A is the pre-exponential factor in the Arrhenius form of the rate constant, and R is the gas constant. The precision of the experimental data does not enable an accurate determination of the preexponential factor; consequently, we calculate the A factor using computational transition-state properties and transitionstate theory. Several experiments are fit using a linear regression to determine the activation energy for each POZ decomposition reaction. Materials. Cyclohexene and MECYHE were obtained from Sigma-Aldrich at 99% and 97% purity, respectively. Methylenecyclohexane was obtained from Acros Organics at 98% purity. All reagents were used without further purification. Ozone was produced with a (Pacific Ozone Technologies) corona discharge ozone generator in an ozone/oxygen mixture of an unknown composition. Experimental Results Several experiments using different heating rates were conducted for each of the three precursor alkenes. The integrated peak area as a function of temperature for a representative cyclohexene experiment is presented in Figure 2a. Peaks corresponding to the POZ are shown in red, and peaks corresponding to POZ decomposition products are in blue. The peak ranges used to calculate each of these curves are presented in the Supporting Information. Comparable traces for the MECYHE POZ and the MECYHA POZ are presented in Figures 2b and 2c, respectively. All three POZ species exhibit sharp sigmoidal decreases when the POZ decomposes. Con-

Figure 3. Determination of barrier heights for cyclohexene, MECYHE, and MECYHA POZ decomposition. The y-intercepts “ln(AR/Ea)”, presented in the corresponding colors above the plot, were determined iteratively with calculated pre-exponential factors. Even a large error in A-factor produces a small change in the activation energy. The slope of each of these curves is -Ea/R. Each shape indicates a specific peak that was integrated to track POZ decomposition. No trend is shown with specific peaks indicating that the scatter is due to experimental noise.

versely, the POZ decomposition products exhibit sharp sigmoidal increases at roughly the same temperature. The peak reaction temperatures of these reactions are indicated with the vertical dashed green lines in Figure 2. The low-frequency drift in some of the TPRS curves is a consequence of the strong variation with temperature of the overall IR intensity. However, the sigmoidal changes are welldefined and reproducible. Several well-defined transitions occur after POZ decomposition; however, due to the increasing complexity as the mechanism branches, we are unable to identify the resulting product distribution. There are two obvious features in Figure 2. The peak desorption temperatures for the first two POZ species formed from endocyclic cycloalkenes (Figure 2a and 2b) are similar and much lower than the peak desorption temperature for the POZ formed from the exocyclic alkene. There clearly are substituent effects to POZ decomposition, and we can resolve them. Despite this, all systems show only a single strong transition even though multiple decomposition channels are possible. Either the competing channel is negligible or the two cycloreversion barriers are similar (leading to the same peak reaction temperature). We shall turn to DFT calculations below to explore this question. A Redhead plot of ln(β/TP2) vs the inverse peak desorption temperature for each precursor reveals the decomposition barrier heights. Figure 3 shows these curves for all three POZ decomposition data sets. The slope (-Ea/R) was determined iteratively because the activation energy is present in the y-intercept term (ln(AR/Ea)). The pre-exponential factor for cyclohexene POZ decomposition was obtained from Fenske et al. 20005 with DFT at the B3LYP/ 6-31G*(d,p) level of theory. The pre-exponential factor for MECYHA POZ decomposition was not available in the literature; therefore, a range of A-factors were used from 1012 to 1014 s-1. In all cases, the large A-factor range did not lead to a significant perturbation in the activation energy, but the larger uncertainty in the MECYHA barrier height reflects this A-factor estimation. The A-factor of MECYHE was determined compu-

7512

J. Phys. Chem. A, Vol. 114, No. 28, 2010

Epstein and Donahue

TABLE 1: Experimental Barrier Heights of POZ Decomposition with Corresponding A-Factor precursor alkene

Ea (experiment) [kcal/mol]

A-factor [1/s] @ 298 K

A factor notes

cyclohexene MECYHE MECYHA

9.1 ( 0.4 9.4 ( 0.4 11.9 ( 1.2

7.90 × 1012 4.4 × 1013 1.00 × 1013

theoretical from Fenske6 theoretical from this work varied from 1012 to 1014 s-1

tationally using DFT B3LYP 6-31G(d). Details of this calculation are presented in the computational methods section. We can also use the computational results to predict the location of higher-energy transition-state decompositions in the TPRS spectra. For example, in an experiment with a heating rate of 0.12 K/s, decomposition along the second lowest energy MECYHE decomposition pathway would occur at a peak reaction temperature of 153 K (indicated with the cyan dashed vertical line in Figure 2b). According to Figure 2b, all of the POZ has decomposed at a lower temperature. Therefore, we would not expect to see any competing decomposition pathways in the MECYHE TPRS spectra. The measured barrier heights and corresponding preexponential factors are presented in Table 1. This confirms what was evident from the raw data: the endocyclic alkenes yield POZ species with significantly lower cycloreversion barriers than the exocyclic alkene. Computational Methods The TPRS data are complemented with DFT calculations, which we used for two main tasks: predicting spectra to enable product identification, and elucidation of the complete MECYHE POZ decomposition pathway. We used Gaussian 0333 at the DFT B3LYP/6-311G++(2d,2p) level of theory to calculate IR spectra for each of the primary ozonides and their corresponding decomposition products. Frequencies were corrected by a factor of 0.9613 as suggested by Wong.36 We calculated difference spectra for POZ decomposition product identification in order to minimize the effects of computational errors. We explored the complete MECYHE POZ system with DFT B3LYP/6-31G(d), including stable configurations and transition states. Slight structural differences in the MECYHE POZ produced large energy differences; therefore, the lowest-energy POZ structure was identified among several stereoisomers. With knowledge of the most probable POZ structure, the four POZ decomposition transition-state structures could be optimized. The pre-exponential factors and barrier heights were calculated from these structures. Computational barrier heights are seldom accurate, and the physics driving inaccuracies in the barrier heights also extends to loose transition-state frequencies associated with bending deformations correlated with separate motion of the two reaction fragments (i.e., translation and relative rotation of the COO and the CPP in this case).37 We thus corrected the computational results for all four pathways following the procedure described in Sage and Donahue 2005.38 We adjusted both the computational barrier height and selected frequencies by a common factor, s. This is an iterative process because we used computational A factors in the Redhead equation, and so the observed activation energy is weakly sensitive to the correction factor s. We used the same correction factor to scale the corresponding frequencies and the barrier heights in all of the decomposition pathways. There was some ambiguity in identifying the (five) frequencies that are related to the decomposition reaction coordinate, but an identification error does not have a significant effect on the revised activation energy; eliminating the frequency

Figure 4. Experimental difference spectra taken before and after cyclohexene POZ decomposition (124-149 K) are shown in black. Computationally predicted difference spectra for the three possible decomposition pathways corresponding to Figure 1 are shown in blue, green, and red. Positive peaks indicate formation, whereas negative peaks indicate consumption.

correction altogether yields an activation energy estimation that is well within the uncertainty of the experimental results. We are ultimately interested in the branching ratios for POZ decomposition, which we express as a decomposition selectivity. This is controlled by the difference of Gibbs free energy between each transition state. We define the differential selectivity of path “n” as the rate of reaction through path n divided by the total rate of consumption of the reactant.39 For a unimolecular decomposition, the differential selectivity (referred to as selectivity in this manuscript) is simply a function of the rate constants:

Sn )

kn #paths

∑

(2) kp

p)1

Defining the rate constant in terms of the transition state Gibbs free energy, GTS, yields an expression for the selectivity of path n:

1 - Sn ) Sn

(∑

#paths p)1

(

e

TS GTS p -Gn

RT

)) - 1

(3)

where p represents the path. GpTSis the Gibbs free energy of the transition state that is encountered in decomposition path p. In the case of MECYHE POZ decomposition, p ranges from 1 to 4. Equation 3 is derived in the Supporting Information. Computational Results One of the major objectives of this work is to identify the POZ decomposition products for the three alkenes of interest. Product spectra were too complicated to compare directly to predicted spectra. Therefore, experimental difference spectra were compared to computational difference spectra. The experimental difference spectra were obtained from adsorption spectra before and after POZ decomposition. The computational difference spectra were calculated for each of the possible decomposition pathways. Positive peaks in the difference spectra indicate the formation of products after decomposition while negative peaks indicate the disappearance of a particular product.

Ozonolysis of Cyclic Alkenes

Figure 5. Experimental difference spectra taken before and after MECYHE POZ decomposition (130-160 K) is shown in black. Computationally predicted difference spectra for the four possible decomposition pathways corresponding to Figure 1 are shown in blue, green, cyan, and red.

Figure 4 shows the experimental and computational difference spectra comparison for cyclohexene. A strong carbonyl peak is present in the experimental spectrum around 1700 cm-1. Both the vinyl-hydroperoxide (VHP) and the dioxirane (DiO) pathway share this strong feature. However, neither VHP nor DiO possess any additional strong features separating one from the other. We do not see strong evidence of the SOZ: the strength of the carbonyl peak suggests that it is unlikely that the SOZ comprises a significant fraction of the products. It is difficult to compare the difference spectra at wavenumbers less than 1500 cm-1 due to the frequent occurrence of peaks in that region and the error inherent in computational spectra prediction at the level of theory employed. The difference spectra comparison for MECYHE POZ decomposition (Figure 5) is similar to cyclohexene decomposition. Experimental spectra were taken at 130 and 160 K: before and after POZ decomposition. As in the case of cyclohexene, a strong carbonyl peak is present around 1700 cm-1. All products except the SOZ share this peak. The accuracy of the computational spectrum is not sufficient to discriminate between an aldehyde or ketone. A weak positive peak is present at 2750 cm-1, which is most likely due to the aldehydic C-H bond from the aldehydic epoxide (EpOA) or the aldehydic vinyl hydroperoxide (VHPK); however, this does not allow us to rule out significant formation of the ketonic vinyl hydroperoxide (VHPK). As in the case of cyclohexene, it is unlikely that the SOZ is a significant product of the MECHYE POZ decomposition. Methylene-cyclohexane POZ decomposition can occur through four possible pathways. A comparison between computational and experimental difference spectra taken at 160 and 190 K for these four pathways is presented in Figure 6. A key characteristic of the experimental spectrum is a relatively strong peak at 1650 cm-1. This peak, likely due to a C-C double bond, is only present in the cyclo-vinyl hydroperoxide computational spectrum. Therefore, it is likely that the cyclo-vinyl hydroperoxide is a significant product of MECYHA POZ decomposition. The rest of the spectra are too congested to identify any additional products. We performed a complete computational study to examine MECYHE POZ decomposition more thoroughly. The four optimized POZ decomposition transition-state structures predicted with DFT B3LPY 6-31G(d) are presented in Figure 7. The transition-state structures are also labeled accordingly in Figure 1. TS1, TS2, and TS3 produce a syn-CCP/COO intermediate, leading to the formation of vinyl-hydroperoxides, which ultimately produce hydroxyl radicals in the gas phase. TS4 produces an anti-CCP/COO intermediate leading to more stable dioxiranes.

J. Phys. Chem. A, Vol. 114, No. 28, 2010 7513

Figure 6. Experimental difference spectra taken before and after MECYHA POZ decomposition (160-190 K) is shown in black. Computationally predicted difference spectra for the four possible decomposition pathways are shown in blue, green, cyan, and red.

Figure 7. Optimized transition state structures for MECYHE POZ decomposition. Dashed blue lines indicate bonds that break during decomposition. Red spheres indicate oxygen atoms, yellow spheres indicate carbon, and cyan spheres indicate hydrogen. Bond lengths are in angstroms.

The purely computational and experimentally corrected thermodynamic parameters at 298 K for MECYHE POZ decomposition are presented in Table 2. Correction of the computational parameters changes the enthalpy of reaction (∆H), the entropy of reaction (∆S), the energy of reaction (∆E0), and the electronic energy of reaction (∆Eelec) significantly. These thermodynamic parameters were also predicted as a function of temperature ranging from 100 to 325 K. Figure 8 indicates that at the temperature of decomposition, the competing decomposition pathway (TS4) is not a significant POZ sink. Therefore, the most energetically favorable decomposition pathway (TS2) should dominate the experimental spectra. Discussion Although limited to one decomposition pathway, experimental determination of POZ decomposition rate constants using TPRS is a novel and powerful tool to elucidate previously unknown kinetic parameters. Computational studies complement these experiments but some uncertainty remains in the determination of the decomposition products. The procedures outlined in this manuscript were successful in determining both the decomposition rate constant and the sole decomposition product for tetramethylethene (TME) decomposition in Epstein and Donahue 2008.32 However, the infrared (IR) signature of TME is sparsely populated with peaks below 1000 cm-1. This enables a more

7514

J. Phys. Chem. A, Vol. 114, No. 28, 2010

Epstein and Donahue

TABLE 2: Thermodynamic Parameter Predictions for 1-Methyl-cyclohexene POZ Decomposition Using DFT B3LYP 6-31G(d) and QST240 at 298 K for Transition State Calculations ∆H (experiment) [kcal/mol] TS1 TS2 TS3 TS4

9.4 ( 0.4

∆H (theoretical) [kcal/mol]

∆S (theoretical) [cal/mol/K]

selectivity (theoretical)

∆H (exp. corr.) [kcal/mol]

∆S (exp. corr.) [cal/mol/K]

selectivity (exp. corr.)

∆E0 (exp. corr.) [kcal/mol]

∆Eelec (exp. corr.) [kcal/mol]

18.2 14.2 18.8 15.9

2.3 0.48 0.40 0.44

0.00 0.90 0.00 0.10

11.8 9.4 12.1 10.2

5.8 3.9 3.8 3.9

0.03 0.78 0.01 0.18

11.8 9.4 12.1 10.2

13.1 10.7 13.3 11.5

robust comparison between computational and experimental spectra because TME POZ decomposition products strongly absorb IR at frequencies between 600 and 1000 cm-1. The more complex nature of the precursor alkenes examined here leads to richer IR spectra. These spectra are difficult to resolve when comparing them to computational spectra that are inherently imprecise. Nevertheless, the experimental decomposition rate constant measurements are robust and reproducible. Our measurements indicate that the additional methyl group in MECYHE does not have a significant effect on the POZ decomposition barrier height. On the other hand, the barrier height of MECYHA POZ decomposition is larger than its endocyclic counterparts but smaller than TME POZ decomposition32 (13.8 ( 1 kcal mol-1). We failed to see convincing evidence of secondary ozonides from any of the cyclic precursors investigated in this study. This is in contrast to our studies of tetramethylethene where the SOZ was the dominant decomposition product. This suggests that it is not energetically favorable for cyclic alkene precursors to cyclize into a five-member ring after POZ decomposition despite tethering of the COO and CCP. Therefore, in the gas phase, a stabilized COO may react with other speciesssuch as watersbefore it isomerizes. To our knowledge, no other experimental studies of POZ decomposition rate constants exist. However, a few researchers have done similar computational studies, notably Fenske et al.6 for the MECYHE system; their thermodynamic parameter estimations with DFT B3LYP 6-31G(d,p) differ from ours in magnitude, but still predict that a syn-CCP/COO intermediate will dominate over the anti-CCP/COO. Our results can be extrapolated to yield information on hydroxyl radical yields following gas-phase ozonolysis. These OH radicals are a direct product of the dissociation of vinylhydroperoxide. The syn-CCP/COO intermediate produces OH

radicals with near unit efficiency.28 Approximately 15% of the anti-CCP/COO intermediate forms OH radicals.31 For MECYHE ozonolysis, we predict an OH yield of 0.92 for the purely computational case and 0.85 for the experimentally corrected calculations. This compares well to published experimental OH radical yields of 0.91 ( 0.20 from Fenske et al.6 and 0.9 (+0.45/ -0.3) from Atkinson et al. 1995.41 Using the computational predictions from Fenske et al.6 in eq 3 to calculate the selectivity leads to an OH yield of 0.99, which is also within the uncertainty of the experimental measurements. Figure 8 shows that the lowest-energy pathway for MECYHE POZ cycloreversion remains dominant at room temperature, although the selectivity is progressively diminished. The selectivity is controlled by energy differences among the various transition states rather than entropy (A-factor) differences. The low-temperature TPRS data thus reveal the lowest-energy pathway exclusively. Conclusions The barrier height for decomposition of primary ozonides formed from endocyclic alkenes (cyclohexene and MECYHE) is lower than the barrier height for decomposition of the primary ozonide formed from MECYHA, an exocyclic alkene. In all three cases, we do not see significant evidence of SOZ formation. The preferential selectivity of the lowest energy decomposition pathways only revealed a single decomposition for each of the primary ozonides. Selectivity calculations, using DFT, of the four possible pathways for MECYHE POZ decomposition reveal that the selectivity of the dominant pathway decreases with increasing temperature. These calculations can also be used to predict an OH radical yield that agrees with previous experiments. Differences in selectivity for each of the four decomposition pathways are due solely to enthalpic variations. Acknowledgment. This work was supported by Grant CHE0717532 from the National Science Foundation. Supporting Information Available: A supplement to Figure 2 identifying all of the peaks used in the figure, the experimental and computational spectra of all three primary ozonides, a graphical representation of eq 3, the kinetic parameters for MECYHE POZ decomposition as a function of temperature, a derivation of eq 3, and raw DFT computational results for the MECYHE system. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes

Figure 8. Selectivity predictions for the four pathways of MECYHE decomposition as a function of temperature. Selectivity is predicted computationally in panel (a) and corrected with the experimental data in panel (b).

(1) Koppmann, R. Volatile Organic Compounds in the Atmosphere; Blackwell Publishing: Oxford, U.K., 2007. (2) Hallquist, M.; Wenger, J. C.; Baltensperger, U.; Rudich, Y.; Simpson, D.; Claeys, M.; Dommen, J.; Donahue, N. M.; George, C.; Goldstein, A. H.; Hamilton, J. F.; Herrmann, H.; Hoffmann, T.; Iinuma, Y.; Jang, M.; Jenkin, M. E.; Jimenez, J. L.; Kiendler-Scharr, A.; Maenhaut, W.; McFiggans, G.; Mentel, T. F.; Monod, A.; Prevot, A. S. H.; Seinfeld, J. H.; Surratt, J. D.; Szmigielski, R.; Wildt, J. Atmos. Chem. Phys. 2009, 9, 5155.

Ozonolysis of Cyclic Alkenes (3) Kroll, J. H.; Seinfeld, J. H. Atmos. EnViron. 2008, 42, 3593. (4) Donahue, N. M.; Kroll, J. H.; Anderson, J. G.; Demerjian, K. L. Geophys. Res. Lett. 1998, 25, 59. (5) Aschmann, S. M.; Arey, J.; Atkinson, R. Atmos. EnViron. 2002, 36, 4347. (6) Fenske, J. D.; Kuwata, K. T.; Houk, K. N.; Paulson, S. E. J. Phys. Chem. A 2000, 104, 7246. (7) Gutbrod, R.; Kraka, E.; Schindler, R.; Cremer, D. J. Am. Chem. Soc. 1997, 119. (8) Rathman, W. C. D.; Claxton, R. A.; Rickard, A. R.; Marston, G. Phys. Chem. Chem. Phys. 1999, 1, 3981. (9) Winterhalter, R.; Neer, P.; Grossmann, D.; Kolloff, A.; Horie, O.; Moortgat, G. J. Atmos. Chem. 2000, 35, 165. (10) Yu, J. Z.; Flagan, R. C.; Seinfeld, J. H. EnViron. Sci. Technol. 1998, 32, 2357. (11) Chuong, B.; Zhang, J.; Donahue, N. M. J. Am. Chem. Soc. 2004, 126, 12363. (12) Keywood, M. D.; Varutbangkul, V.; Bahreini, R.; Flagan, R. C.; Seinfeld, J. H. EnViron. Sci. Technol. 2004, 38, 4157. (13) Zhang, J.; Huff Hartz, K.; Pandis, S.; Donahue, N. M. J. Phys. Chem. A 2006, 110, 11053. (14) Presto, A. A.; Huff Hartz, K. E.; Donahue, N. M. EnViron. Sci. Technol. 2005, 39, 7036. (15) Pathak, R. K.; Stanier, C. O.; Donahue, N. M.; Pandis, S. N. J. Geophys. Res., Atmos. 2007, 112, D03201. (16) Saathoff, H.; Naumann, K. H.; Mohler, O.; Jonsson, A. M.; Hallquist, M.; Kiendler-Scharr, A.; Mentel, T. F.; Tillmann, R.; Schurath, U. Atmos. Chem. Phys. 2009, 9, 1551. (17) Donahue, N. M.; Hartz, K. E. H.; Chuong, B.; Presto, A. A.; Stanier, C. O.; Rosenhørn, T.; Robinson, A. L.; Pandis, S. N. Faraday Discuss. 2005, 130, 295. (18) Criegee, R. Lebigs. Ann. Chem. 1953, 583, 1. (19) Bailey, P. S. Ozonation in Organic Chemistry; Academic Press: New York, 1978; Vol. I. (20) Fenske, J. D.; Hasson, A. S.; Paulson, S. E.; Kuwata, K. T.; Ho, A.; Houk, K. N. J. Phys. Chem. A. 2000, 104, 7821. (21) Donahue, N. M.; Tischuk, J. E.; Marquis, B. J.; Huff Hartz, K. E. Phys. Chem. Chem. Phys. 2007, 9, 2991. (22) Kroll, J. H.; Sahay, S. R.; Anderson, J. G.; Demerjian, K. L.; Donahue, N. M. J. Phys. Chem. A 2001, 105, 4446. (23) Jang, M.; Kamens, R. M. Atmos. EnViron. 1999, 33, 459. (24) Kalberer, M.; Yu, J.; Cocker, D. R.; Flagen, R. C.; Seinfeld, J. H. EnViron. Sci. Technol. 2000, 34, 4894.

J. Phys. Chem. A, Vol. 114, No. 28, 2010 7515 (25) Aschmann, S.; Tuazon, E.; Arey, J.; Atkinson, R. J. Phys. Chem. A 2003, 107, 2247. (26) Ziemann, P. J. J. Phys. Chem. A 2002, 106, 4390. (27) Ziemann, P. J. J. Phys. Chem. A 2003, 107, 2048. (28) Paulson, S. E.; Chung, M. Y.; Hasson, A. S. J. Phys. Chem. A 1999, 103, 8125. (29) Kuwata, K. T.; Templeton, K. L.; Hasson, A. S. J. Phys. Chem. A 2003, 107, 11525. (30) Gutbrod, R.; Schindler, R. N.; Kraka, E.; Cremer, D. Chem. Phys. Lett. 1996, 252, 221. (31) Kroll, J. H.; Donahue, N. M.; Cee, V. J.; Demerjian, K. L.; Anderson, J. G. J. Am. Chem. Soc. 2002, 124, 8518. (32) Epstein, S. A.; Donahue, N. M. J. Phys. Chem. A 2008, 112, 13535. (33) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, ReVision D.01; Gaussian, Inc.: Wallingford CT, 2004. (34) Hull, L. A.; Hisatsune, I. C.; Heicklen, J. J. Am. Chem. Soc. 1972, 94, 4856. (35) Redhead, P. A. Vacuum 1962, 203. (36) Wong, M. W. Chem. Phys. Lett. 1996, 256, 391. (37) Donahue, N. M. J. Phys. Chem. A 2000, 105, 1489. (38) Sage, A. M.; Donahue, N. M. J. Photoc. Photobio. A 2005, 176, 238. (39) Butt, J. B. Reaction Kinetics and Reactor Design, 2nd ed.; Marcel Dekker, Inc: New York, NY, 2000. (40) Peng, C.; Schlegel, H. B. Isr. J. Chem. 1993, 33, 449. (41) Atkinson, R.; Tuazon, E. C.; Aschmann, S. M. EnViron. Sci. Technol. 1995, 29, 1860.

JP102177V