The Overall Reaction Process of Ozone with Methacrolein and

Article pubs.acs.org/JPCA

The Overall Reaction Process of Ozone with Methacrolein and Isoprene in the Condensed Phase Jian-guo Deng, Jian-hua Chen,* Chun-mei Geng, Hong-jie Liu, Wei Wang, Zhi-peng Bai, and Yi-Sheng Xu State Key Laboratory of Environmental Criteria and Risk Assessment, Chinese Research Academy of Environmental Science, Beijing 100012, China S Supporting Information *

ABSTRACT: The reaction of isoprene and methacrolein with ozone was investigated at different stages in the condensed phase at temperatures from 15 to 265 K by IR spectroscopy. The results revealed the following overall reaction process: the generation of primary ozonide (POZ), then its decomposition, and finally conversion into secondary ozonide (SOZ), which supported the Criegee mechanism. In the POZ and SOZ of isoprene, ozone cyclo-added preferentially to the double-bond that is not substituted by the methyl group. For methacrolein, the mainly detected SOZ is claimed to be MACSII formed by recombination of the intermediate CH2OO radical with aldehyde carbonyl of methylglyoxal in stead of the ketone carbonyl group. Theoretical calculations were performed at the B3LYP//MP2/6-311++G (2d, 2p) level to analyze the resulting spectrum. The good agreement between the calculated infrared spectra of POZ and SOZ and the experimental spectra supports the abovedescribed findings.

■

INTRODUCTION Isoprene, emitted by vegetation, is estimated at 600 Tg yr−1 and accounts for 50% of global nonmethane hydrocarbon (NMCH).1 NMCH can result in ozone formation in rural and urban areas as their peroxide radicals convert NO into NO2. Isoprene rapidly reacts with ozone, which eventually influences the atmospheric chemical processes in two ways. First, ozonolysis can produce an OH radical, which is the most important oxidant and determines the lifetime of other substances in the atmosphere. A number of studies have reported an OH radical yield of about 0.25. Quantum mechanical calculations have demonstrated that OH radical was produced via a vinyl hydroperoxide channel.2−5 Second, ozonolysis can serve as a significant source of secondary organic aerosol (SOA), which causes deterioration of visibility and air quality.6 On the basis of current knowledge, SOA formation involves homogeneous and heterogeneous processes, including the homogeneous condensation of low-volatility products, such as multifunctional acids or aldehyde (e.g., glyoxal, methylglyoxal) and acid-catalyzed heterogeneous reactions in aerosol phase.7,8 However, this mechanism is still far from being well-understood. For example, many of the chemical compositions of SOA from isoprene ozonolysis have not yet been recognized, and some important components such as organic acids, oligomer, and humic-like substances can not be explained by established reaction mechanisms.9 Methacrolein (MAC), a first-generation carbonyl product of isoprene ozonolysis, still has a double bond reaction with ozone.10 In particular, it is suggested that MAC is a key intermediate in SOA formation from isoprene.6 On the basis of product studies and theoretical investigation, the gas phase reaction of ozone with isoprene and MAC also © 2012 American Chemical Society

followed the Criegee mechanism, concluded from alkenesozone reaction in condensed phase.4 The Criegee mechanism can be described simply in three steps as Scheme 1: first, Scheme 1

1,3-dipolar cycloaddition of ozone to the alkene double-bond forms 1,2,3-trioxolane, referred to as primary ozonide (POZ); second, the POZ splits into a carbonyl compound and a reactive carbonyl-oxide (i.e., Criegee intermediate (CI)); third, CI and the coproduct recombine to create secondary ozonide (SOZ). The difference in gas phase is the fate of CI; much of it undergoes unimolecular dissociation to yield an OH radical or reacts with other substances, such as water.11 Received: October 20, 2011 Revised: December 26, 2011 Published: January 13, 2012 1710

dx.doi.org/10.1021/jp2100864 | J. Phys. Chem. A 2012, 116, 1710−1716

The Journal of Physical Chemistry A

■

Matrix isolation supplies a very low temperature (T < 20 K) and an inert atmosphere, and is a powerful method for reactive intermediate studies. Employing this technique, some studies successfully detected the POZ and SOZ of several olefins such as ethene, propene, and butane. However, CI had never been detected until Hoops et al., who reported that they observed the CI of cyclopentene and cyclopentadiene.12 Generally, when the matrix sample (consisting of ozone, alkene, and Ar) is annealed to 35 K (80 K for Xe), both POZ and SOZ are observed.13,14 Using a different apparatus (liquid nitrogen cooled), Hull and Epstein found that POZ decomposed at higher temperature.15,16 Inspired by Hull and Epstein’s works, we studied the ozonolysis of isoprene and MAC in condensed phase by slowly heating the matrix sample from 15 to 265 K. To our knowledge, not many experiments have been done on isoprene and MAC, except Feltham et al., who have found that SOZ was formed as a stable compound in isoprene−ozone gas phase reaction.17 In this study, ozonolysis processes at different stages were observed, and IR spectra of their POZ and SOZ were obtained. The IR spectrum of SOZ is useful for recognizing its existence in SOA. We consider SOZ as one component of SOA since the surface of particles may stabilize SOZ.18,19

■

Article

CALCULATION DETAILS

Theoretical calculations of the likely ozonides were carried out using Gaussian 03 and 03W suite programs.20 Initially, the geometry of ethene POZ and SOZ were optimized using HF, MP2, and DFT (B3LYP) methods at deferent basis sets from small to large and compared with the experimental data. It showed that the results from MP2 and B3LYP converged well and were highly close to the experimental data (Tables S1−S3, S5−S7). The calculated vibrational frequencies of ethene POZ and SOZ using MP2 and B3LYP/6-311++G (2d, 2p) levels were in good agreement with vibrational frequencies obtained in our laboratory. The accuracies of B3LYP and MP2 were 0.98 and 0.97, respectively (Tables S4 and S8). Therefore both MP2 and B3LYP/6-311++G (2d, 2p) levels were used to locate the energy minima and to compute vibrational spectra of the likely ozonides of isoprene and MAC.

■

RESULTS The infrared spectra of the resulting products were recorded as the temperature rose. The main differences took place in the region of 600−1800 cm−1 where the bands due to POZ and SOZ appear. Isoprene + Ozone. Compared to the isoprene and ozone blank spectra, no new bands were observed after codeposition of isoprene and ozone samples (Figure 1a). Annealing to 35 K resulted in a very weak new band at 678 cm−1, indicating that a reaction occurred. As the temperature continually rose, a series of new bands (at 678, 746, 797, 817, 848, 962, 1078, 1137, 1168, 1205, 1332, 1654 cm−1) appeared and grew quickly at a ratio of ca. 150%, and finally stabilized. Meanwhile, the reactant bands (1033 cm−1 for ozone, 902, 1600, and 1636 cm−1 for isoprene) reduced substantially in intensity (Figure 1a). Additionally, the bands appearing at 1724 and 1499 cm−1 indicated that formaldehyde was produced; likewise, the bands appearing at 1692, 1311, 1024, and 817 cm−1 indicated that MAC was formed, which could be verified by its blank spectrum (Figure 2a). Upon annealing to 155 K, the new bands underwent two distinct variations: the bands labeled P decreased greatly; at the same time, the bands labeled S increased sharply (Figure 1b. It was seen that almost all of the bands throughout the spectrum were S-type bands up to 185 K. With the temperature continually rising, the S-type bands declined gradually. However, at 215 K, additional bands at 1742, 1086, 1057, 1174, 1120, and 970 cm−1 appeared and increased. Finally, all of the bands disappeared when the temperature reached 265 K (Figure 1c). MAC + Ozone. Compared to the MAC and ozone blank spectra, no changes were noted after the samples of MAC and ozone were deposited onto the 15 K cold window and annealed to 65 K. Warming the matrix up to 70 K brought about a perceptible absorption band at 1742 cm−1 (Figure 2a). With temperature continually increasing, the parent bands reduced rapidly, and a large number of new bands sprung up (the bands at 674, 717, 795, 846, 900, 923, 1087, 1117, 1137, and 1742 cm−1), which gradually grew to 125 K. In particular, the “carbonyl region” showed marked changes: the 1742 cm−1 band grew substantially, while the absorption bands of CO in MAC rapidly reduced (Figure 2b). From 155 K, two types of results arrived in these new bands: the bands labeled P decreased substantially; at the same time, the bands labeled S increased dramatically. When MAC was consumed by reaction and evaporation, some of its bands (e.g., 962, 1378, and 1453 cm−1) remained constant, but suddenly dropped at 155 K. Therefore, these

EXPERIMENTAL DETAILS

Isoprene (99%, Tokyo chemical industry company) was used directly as received. MAC (95%, Aldrich) was degassed by repeated trap-to-trap distillation at 77 K. Ozone was produced by a Tesla coil discharge of O2 (99.995%, Zhengyuan chemical industry company, Beijing) and trapped by liquid nitrogen to remove the residual O2 and other trace gases. Argon (99.9996% Zhengyuan chemical industry company, Beijing) was used as the matrix gas without further purification. Each gas was mixed separately with Ar to the desired ratio (M/Ar = 1:100) in a 5 L Pyrex bulb using standard manometric technique. The bulb was connected to the chamber of cryostat via a 1/4 in. Teflon tube and a 1/16 in. peek tube. The matrix isolation equipment used for this experiment was a conventional matrix isolation apparatus based on a closed-cycle helium refrigerator (APD Cryogenics, model DE-204NE). The high vacuum (10−7mbar) was maintained by a turbo-molecular pump backed by an oil diffusion pump. The temperature of the CsI window was controlled by a Lake Shore Cryogenics temperature controller (model 331). The ozone/Ar and alkene/Ar were codeposited (at a ratio of 2 mmol/h) onto a 15 K cold window from two separate lines. After 4 h of deposition, the matrix was annealed to 35 K (the temperature of Ar diffusion) and held at that temperature for 1 h. Then it was further warmed up to 40−45 K gradually over 3 h; at this moment, the Ar began to evaporate rapidly, and the Ar vapor produced was pumped away quickly. Finally, the temperature was maintained at 47−50k for a period until there was almost no Ar on the cold window; it could be detected from the vacuum gauge. Although some of the reactants were taken away by Ar gas, most of them remained on the window to form a neat film. After that, the compressor was turned off and allowed the Cryostat to warm up slowly (about 1 K/min). The infrared spectrum of the resulting neat film was recorded every increase of 10 K by a vacuum Fourier transform spectrometer (Bruker, vertex 80v) at the desired temperature. Data collection was performed in the range of 600−4000 cm−1 with 0.5 cm−1 resolution, with 32 scans. All of the resulting spectra were corrected at their baseline by OPUS (Bruker). 1711



Article

Gauss-view and the output by OPUS after normalization, and scaled by a factor (0.98 for B3LYP, 0.97 for MP2).

■

DISCUSSION Isoprene + Ozone. The bands labeled P are assigned to the POZ of isoprene, and the bands labeled S are assigned to SOZ. This means that during the warming up of the isoprene and ozone matrix, the POZ was formed first, then decomposed at 155 K, and finally converted into the SOZ, which accounts for the observed phenomenon that the decrease of P-type bands is followed by the increase of S-type bands. There are three types of evidence to support the above hypothesis. First, the characteristic peaks for POZ are in the ranges of 600−800 cm−1 and 900− 1100 cm−1, corresponding to O−O−O antisymmetric stretching vibration and C−O symmetric stretching vibration, respectively. The SOZ has an important characteristic peak in the range of 1000−1200 cm−1, corresponding to the C−O−C antisymmetric stretching vibration.13−15,24−27 Therefore, the strong 678 and 962 cm−1 bands observed in this study imply the formation of POZ. Likewise, the strongest 1078 cm−1 band implies the formation of SOZ. Second, using the same method, we also studied the reactions of ethene and propene with ozone. The same experimental phenomenon was observed that the decrease of P-type bands was followed by the increase of S-type bands. The frequencies of P- and S-type bands were compared with the reported values of their POZ and SOZ, respecticvely.13−15 It is found that P-type bands are indeed the IR absorption bands of POZ, and S-type bands are indeed the IR absorption bands of SOZ (Figures S1 and S2 and Tables S9−S12), which supports the assertion that the assignments for isoprene are valid. Although Feltham et al. have reported the absorption bands of isoprene SOZ, these bands are very limited (i.e., only at 1234(w), 1223(w), 1124(s), 998(w), and 930(m) cm−1). The relatively strong 1124, 998, and 930 cm−1 bands correspond to 1135, 991, and 928 cm−1 bands observed in this experiment, but the strongest 1078 cm−1 band has not been clearly observed in their experiment, which may be caused by the differences between gas phase reaction and condensed phase reaction or by the complexity of the spectra of gas phase reaction results. Finally, the infrared spectra of isoprene POZ and SOZ were predicted by high-level theoretical calculations at the B3LYP//MP2/6-311++G (2d, 2p). Isoprene is a conjugated diene and will yield two POZs (referred to as ISPPI, ISPPII) and two normal SOZs (referred to as ISPSI, ISPSII) when reacting with ozone according to the Criegee mechanism (as illustrated in Figure 3). Therefore the calculated spectra of the four ozonides were compared with the experimental spectra. It was found that the bands of ISPPI were consistent with the P-type bands observed, and the bands of ISPSI were consistent with the S-type bands (Figure 4a,b), which confirmed that isoprene POZ was formed first, and then decomposed and converted into SOZ in the condensed phase. As shown in Figure 4a, it appears that the spectrum of ISPPII matches with the experimental spectra; in particular, the strongest 945 cm−1 band (Table S13) is in the vicinity of 962 cm−1. However, the calculations show that the 945 cm−1 band is due to the wagging of =CH2, not the C−O stretching vibration of ISPPII (Table 1). In ISPPI, the wagging of the =CH2 is also a strong band and calculated at 931 cm−1, but the corresponding band is not observed near 931 cm−1 in the experimental spectrum. The hypothesis is that this band is probably overlapped with the broad parent band at 905 cm−1, which is due to the wagging of the =CH2 of isoprene.

Figure 1. (a) Infrared spectra of (1) separately deposited isoprene/Ar, (2) the matrix of codeposition isoprene−ozone−Ar (1:1:200) at 15 K, and (3) the reaction results during warming up. From 55 K, a neat film of isoprene and ozone was formed by evaporating Ar from the matrix. (b) Infrared spectra of the reaction products of isoprene with ozone from 115 to 175 K. (c) Infrared spectra of the reaction products of isoprene with ozone from 185 to 265 K.

bands also belong to the P-type band. A new band appeared at 1727 cm−1, which was combined with the surge in the 1360 cm−1 band and indicated the production of methylglyoxal on the grounds that methylglyoxal was isolated in the Ar matrix21 (Figure 2c). With the temperature continuously rising, The S-type bands gradually dominated the whole spectrum at 185 K and finally disappeared at 265 K (Figure 2d). Results of Calculation. Seven ozonides were computed, and the final structures are displayed in Figure 3. All of the POZs possessed the O-envelope conformation in which the unique oxygen was bent out of the plane defined by the four remaining atoms. All of the SOZs possessed the O−O half hair conformation in which the two adjacent oxygen atoms were twisted out of the C−O−C plane, indicating that the molecular structures located were lowest in energy.22,23 With these results, we can refer to their calculated frequencies to determine which one is formed. Their calculated spectra were simulated by 1712



Article

Figure 2. (a) Infrared spectra of (1) separately deposited MAC/Ar, (2) the matrix of codeposition MAC−ozone−Ar (1:1:200) at 15 K, and (3) the reaction results during warming up. From 60 K, a neat film of MAC and ozone was formed by evaporating Ar from the matrix. (b) Infrared spectra of the reaction products of MAC with ozone from 75 to 125 K. (c) Infrared spectra of the reaction products of MAC with ozone from 135 to 185 K. (d) Infrared spectra of the reaction products of MAC with ozone from 205 to 265 K.

1623, 1400, 1364, 1251, and 1183 cm−1), but the characteristic bands of MAC as described above. Second, except for the normal SOZ, no cross SOZ was detected, such as ethene SOZ. In the study of propene reaction with ozone, we also have not detected either ethene or 2-butene SOZ (Figure S2). Third, all attempts to identify absorption bands of CI and its possible products failed. The C−O and O−O stretching vibrational absorption of CI should be in the 900−1200 cm−1 range.28 However, some bands which were observed in this range in the resulting spectrum still exist at higher temperature (i.e., 205K) and are not related to CI. Thus the possibility of CI presence was excluded. Because the computed spectrum of the diperoxide was far removed from the experimental spectrum (Figure S3), significant formation of the diperoxide from two CI (CH2OO and CH2C(CH3)CHOO) was ruled out. Also, any other products derived from CI were not detected, such as hydrogen peroxide including double bond, ester, or dioxiranes. Finally, the same conclusion was obtained for isoprene ozonolysis in the gas-phase on the basis of analysis of stable compounds,29 except in the work of Zhang et al., who assume that the initial two O3 addition pathways are nearly equally accessible via quantum chemical study.4 It is considered that the steric hindrance of the methyl group leads to the ozone preferential cycloaddition to the double-bond without substitution by the methyl group. The bands (at 1742, 1086, 1057, 1174, 1120, and 970 cm−1) appearing at 215 K are tentatively assigned to MACSOZ. The assignment is based on the discussion below. The reason of the formation of MACSOZ requires more studies. MAC + Ozone. Similarly, the bands labeled P are ascribed to the POZ of MAC, and the bands labeled S are ascribed to the SOZ of MAC. From this we can see that POZ is formed

Figure 3. The sequences of reaction outlined by the optimized structure of ozonides, ISPP, and ISPS stand for POZ and SOZ of isoprene, and MACP and MACS stand for methacrolein.

The ISPPI and ISPSI are mainly detected ozonides of isoprene, that is, ozone preferentially cyclo-adds to the double-bond without substitution by methyl group. In addition to the agreement between experimental and calculated vibrational frequencies (Figure 4a,b; Table 2), there are four facts that indirectly support this conclusion. First, the detected stable carbonyl compound was MAC, rather than methyl vinyl ketone (MVK). We have not observed the characteristic bands of MVK (generally at 1700, 1713



Article

Table 1. Observed Absorptions Due to the POZ of Isoprene and Calculated Vibrational Frequencies of Its Two Likely POZs at B3LYP/6-311++G (2d, 2p) calculationb

experiment POZ cm−1

IR int.a (a.u.)

635

15

678

79

706

18.3

746

16

744

16.1

797 848

34 26

781

10.8

931

41.5

962

1332

Figure 4. (a) Comparison between the isoprene POZ spectra predicted by Gaussian 03 at the B3LYP//MP2/6-311G++ (2d, 2p) level and the spectrum measured at 135 K. (b) Comparisons between the isoprene SOZ spectra predicted by Gaussian 03 at the B3LYP// MP2/6-311G++ (2d, 2p) level and the spectrum measured at 195 K. At this moment almost all of the bands are S-type bands.

1654

first, then decomposes, and finally converts into SOZ. This hypothesis better explains the result observed. The reaction initiating at approximately 70 K indicates that MAC reaction with ozone requires higher activation energy. Afterward, the 1742 cm−1 band grew substantially, while the parent band of 1692 cm−1 rapidly decreased (Figure 2b). It can be explained by the formation of POZ. MAC is propene substituted by an aldehyde group, and the carbon−oxygen double bond conjugates with the carbon−carbon double bond. The conjugation and the possible aggregation lead to a downward shift of CO to 1692 cm−1. When the reaction of MAC with ozone produces POZ (referred to as MACP shown in Figure 3), the conjugated system is broken, and the CO group is released and results in the absorption at 1742 cm−1. Thus, this band and its correlative P-type bands are associated with MACP. The spectrum of MACP predicted by two methods succeeded in modeling the P-type bands (Figure 5a). Analogous to the POZ of propene (Table S12), MACP is expected to have two characteristic absorptions, corresponding to two C−O stretching vibrations. These two absorptions bands are predicted by theoretical calculation (B3LYP) at 963 (end C−O) and 1133 cm−1 (middle C−O), with the intensity of 21 km/mol and 25 km/mol, respectively. They match reasonably well with the 962 and 1134 cm−1 bands in the experimental spectrum, except for a slightly different intensity distribution (Table S13). In the low frequency region, the absorption band of O−O−O of MACP is calculated at 717 cm−1, which matches well with the strong P-type band at 717 cm−1.

100

14

28

ISPPI cm−1

961

calc. int. (km/mol)

ISPPII cm−1

calc. int. (km/mol)

700

15.7

725

36.9

778

10.9

873

21.0

945 951

36.9 9.2

987 1016 1114 1164 1271 1327

15.1 12.0 22.0 9.8 7.2 3.9

1386 1428 1466

9.6 11.6 6.7

1666 2991

6.9 15.2

3056 3083 3163

10.5 6.8 8.0

52.1

1334 1394

3.6 7.4

1453 1478

6.7 15.2

1671 2977 2988 3002 3025 3079 3157

13.0 14.8 13.6 40.0 10.9 5.8 10.4

assignment ? δ O−O−O τ CH2 υ O−O−O as υ O−O−O as τ CH2 υC−C−O s ring def. υ O−C mid ω CH2 ω CH2 ρ CH2 υ C−O υ O−C end ω −CH υ O−C mid. υ C−C−C(s) υ C−C−C(as) ω >CH2 δ C−H δ CH3(s) δ CH2 δ CH3(as) δ CH3(as) + δ CH2 υ CC υ CH3(s) υ C−H (as) υ C−H (s) υ CH3(as) υ CH2(s) υ CH2(as)

a Normalized by the most intense absorption at 1115 cm−1. bThe vibrational frequencies were scaled by 0.98.

Upon the dropping of P-type bands, the substantial growth of 1086 cm−1 and its correlative S-type bands demonstrate the rearrangement of MACP into SOZ. According to the Criegee mechanism, MACP has two ways of splitting: one is CH3C(CHO)OO and formaldehyde; the other is CH2OO and methylglyoxal. If the first way works, the MACSI in Scheme II will be formed (Figure 3). If in the second way, CH2OO reacts with the concomitant ketone carbonyl of methylglyoxal, MACSI will also be formed. However, the calculated spectrum of MACSI is very different from the experimental spectrum (Figure 5b). Hence, the possibility of a significant formation of MACSI was ruled out. The apparent production of methylglyoxal suggests that MACP mainly decomposes in the second way. As we know, the methylglyoxal has an alde-carbonyl group that can also react with CH2OO. If it happens, the MACSII will be formed (Figure 3). As shown in Figure 5b (and Table S14), the calculated spectrum of MACSII predicted the experimental spectrum at 185 K, where almost all of the bands throughout the spectrum are attributed to the SOZ. Therefore, MACSII was the mainly detected SOZ. Because of the inertness of ketone carbonyl, CH2OO recombines with aldehyde carbonyl rather than the ketone carbonyl of methylglyoxal into MACSII. 1714



Article

Table 2. Observed Absorptions Due to the SOZ of Isoprene and Calculated Vibrational Frequencies of Its Two Likely SOZs at B3LYP/6-311++G (2d, 2p) calculationb

experiment

calculationb

experiment

a

SOZ IR int. ISPSI calc. int. ISPSII calc. int. cm−1 (a.u.) cm−1 (km/mol) cm−1 (km/mol) 700

6.9

714

4.0 776

817 837

7.6 2.5

823 854

7.6 6.1 845 875 909.4

928 951

24.6 19.8

940 941

15.5 47.0

986

1078

18.1 20.9 28.7

1024 1035 1055

100.0 1068

1122

7.3

1135

9.6

1125

14.2 7.6

27.5 41.9 35.0 1046

142.8

1098

49.1

167.2 6.1 1127

a

36.8 51.9

53.0 1003 1010

1024 1039 1056

5.6 19.3 18.5

22.1 41.8 948.6 957

971 991

8.4

10.2

assignmentc

SOZ IR int.a ISPSI cm−1 (a.u.) cm−1

τ CH2 υ O−C(end)−O(s) υ C−C−C(s) υ O−O υ O−O υ O−C(mid)−O(as) υ Oe−C−C(as) ω CH2 υ O−C−O(as) υ O−C(end)−O(as) ω =CH2 ? υ C−O−C(s) ρ C−H ρ CH2 υ C−C−C(as) υ O−C−O(s) ρ CH3 υ C−O−C(as) υ C−O−C(as) υ O−C(mid)−O(as) ρ >CH2 ρ >CH2 ?

1171

calc. int. ISPSII (km/mol) cm−1

calc. int. (km/ mol)

1161

95.6

1205 1253

13.0

1363 1383

5.5 22.9

1426

19.2

1463 1469 1671

3.4 4.4 0.7

2957

78.5

2999 3046 3077 3088 3168

6.8 14.2 8.2 5.8 6.6

7.5

1312

7.6

1301 1348

14.2 22.1

1378 1415

39.3 9.6

1383

5.7

1435

8.5

1463 1659 2896 2924

14.2 2.3 12.8 6.1

1432 1450 1475 1677 2961

7.9 8.6 22.1 7.4 70.5

2961 2977

5.5 5.7

2981 2985

16.4 38.3

3090

1.8

3042 3056 3080 3159

18.6 15.2 5.9 8.8

assignmentc δ OCOC (as) ? τ >CH2 υ C−C−C(as) δ C−H δ CH2+ δ C−H ω >CH2 δ CH3(s) ? δ CH2 δ CH2+ δ C−H δ CH3(as) δ CH3(as) + δ CH2 υ CC υ >CH2(s) ? υ >CH2(s) υ CH3(s) υ C−H υ CH3(s) υ >CH2(as) υ CH3(as) υ CH2(s) υ CH2(as)

Normalized by the most intense absorption at 1078 cm−1. bThe vibrational frequencies were scaled by 0.98. cOp is a peroxide oxygen; Oe is a ether oxygen.

Actually, Criegee discovered that CI does not react with ketone, even a concomitant in general, except for ketone in a high concentration or special electronic structure.30 The production of MACSII indicates that the reaction of MAC with ozone follows this rule. Additionally, we failed to find any track for the presence of CI or its derivative products (Figure S4). Finally, it is worth noting that the POZ and SOZ are yielded at the same time (the P- and S-type bands appeared simultaneously), in agreement with Samuni’s observation.14 One possible explanation is the formation of numerous vacancies during the preparation of the matrix sample. If the vacancy is too large to dissipate excess energy of POZ in time, it will lead to the decomposition of POZ into SOZ. Also, the SOZ formed in the matrix was mostly normal-SOZ, as opposed to the cross-SOZ often found in the liquid phase. This phenomenon may be caused by the weak movability of CI in the matrix.

■

CONCLUSIONS We have studied ozone reaction with isoprene and MAC in the condensed phase by slowly heating the matrix sample. The different stages of ozonolysis were observed: the formation of POZ, then splitting, and finally converting into SOZ, which supported the Criegee mechanism. However, the CI from two reactions were not detected. In the reaction of ozone with isoprene, ozone preferentially cyclo-adds to the double-bond without substitution by methyl group and forms POZ (ISPPI) and SOZ (ISPSI). For MAC, the POZ (MACP) mainly splits into CH2OO radical and methylglyoxal. Then the CH2OO radical recombines with the aldehyde carbonyl of methylglyoxal rather than the ketone carbonyl into MACSII, which suggests

Figure 5. Comparisons between the MAC POZ spectra predicted by Gaussian 03 at the B3LYP//MP2/6-311G++ (2d, 2p) level, and the spectrum measured at 135 K. (b) Comparison between the MAC SOZ spectra predicted by Gaussian 03 at the B3LYP//MP2/6-311G++(2d,2p) level and the spectrum measured at 185 K. At this moment almost all of the bands are S-type bands. 1715



Article

K. N.; Burant, J. C.; et al. Gaussian 03, revision A.1, Gaussian, Inc.: Pittsburgh, PA, 2003. (21) Mucha, M.; Mielke, Z. Chem. Phys. 2009, 361, 27−34. (22) Lattimer, R. P.; Kuczkowski, R. L.; Gillies, C. W. J. Am. Chem. Soc. 1974, 96, 348−358. (23) Gillies, J. Z.; Gillies, C. W.; Suenram, R. D.; Lovas, F. J. J. Am. Chem. Soc. 1988, 110, 7991−7999. (24) Sarnuni, U.; Haas, H. Spectrochim. Acta, Part A 1996, 52, 1479− 1492. (25) Garvin, D.; Schubert, C. J. Am. Chem. Soc. 1956, 60, 807−808. (26) Sablinskas, V.; Hegelund, F.; Ceponkus, J.; Bariseviciute, R.; Aleksa, V.; Nelander, B. J. Phys. Chem. A 2005, 109, 8719−8723. (27) Fajgar, R.; Vítek, J.; Haas, Y.; Pola, J. J. Chem. Soc., Perkin Trans. 1999, 2, 239−248. (28) Sander, W. Angew. Chem., Int. Ed. Engl. 1990, 29, 344−354. (29) Paulson, S. E.; Flagan, R. C.; Seinfeld, J. H. Int. J. Chem. Kinet. 1992, 24, 103−125. (30) Bailey, P. S. Ozonation in Organic Chemistry; Academic Press: New York, 1978; Vol. 1.

the inertness of ketone carbonyl. The calculated frequencies of POZ and SOZ both agreed well with the experimental frequencies, which further confirm the above conclusions. It is hoped that the infrared features of isoprene and MAC SOZ are helpful for recognizing its existence in the SOA in the future.

■

ASSOCIATED CONTENT

S Supporting Information *

The spectra of ethene, propene reaction with ozone, and the calculated spectrum of the possible products from alternative intermolecular reactions between the intermediates are available. Tables of the geometry parameters and vibrational frequencies of ethene, propene, isoprene, and MAC are available. This material is available free of charge via the Internet at http:// pubs.acs.org

■

AUTHOR INFORMATION

Corresponding Author

*Tel: 0086-10-84934330; Fax: 0086-10-84934330; E-mail address: [email protected].

■

ACKNOWLEDGMENTS This work was supported by National Natural Science Foundation of China (Grant 40775075) and Ministry of Science and Technology of the People’s Republic of China (200809052).

■

REFERENCES

(1) Guenther, A.; Karl, T.; Harley, P.; Wiedinmyer, C.; Palmer, P. I.; Geron, C. Atmos. Chem. Phys. 2006, 6, 3181−3210. (2) Neeb, P.; Moortgat, G. K. J. Phys. Chem. A 1999, 103, 9003− 9012. (3) Kuwata, K. T.; Valin, L. C.; Converse, A. D. J. Phys. Chem. A 2005, 109, 10710−10725. (4) Zhang, D.; Zhang, R. J. Am. Chem. Soc. 2002, 124, 2692−2703. (5) Malkin, T. L.; Goddard, A.; Heard, D. E.; Seakins, P. W. Atmos. Chem. Phys. 2010, 10, 1441−1459. (6) Carlton1, A. G.; Wiedinmyer, C.; Kroll, J .H. Atmos. Chem. Phys. Discuss. 2009, 9, 8261−8305. (7) Claeys, M.; Wang, W.; Ion, A. C.; Kourtchev, I.; Gelencsér, A.; Maenhaut, W. Atmos. Environ. 2004b, 38, 4093−4098. (8) Jang, M.; Czoschke, N. M.; Lee, S.; Kamens, R. M. Science 2002, 298, 814−817. (9) Nguyen, T. B.; Bateman, A. P.; Bones, D. L.; Nizkorodov, S. A.; Laskin, J. Atmos. Environ. 2010, 44, 1032−1042. (10) Aschmann, S. M.; Arey, J.; Atkinson, R. Atmos. Environ. 1996, 30, 2939−2943. (11) Johnson, D.; Marston, G. Chem. Soc. Rev. 2008, 37, 699−716. (12) Hoops, M. D.; Ault, B. S. J. Am. Chem. Soc. 2009, 131, 2853− 2863. (13) Andrews, L.; Kohlmiller, C. K. J. Phys. Chem. 1982, 86, 4548− 4557. (14) Samuni, U.; Fraenkel, R.; Haas, Y.; Fajgar, R.; Pola, J. J. Am. Chem. Soc. 1996, 118, 3687−3693. (15) Hull, L. A.; Hisatsune, I. C.; Heicklen, J. J. Am. Chem. Soc. 1972, 94, 4856−4864. (16) Epstein, S. A.; Donahue, N. M. J. Phys. Chem. A 2008, 112, 13535−13541. (17) Feltham, E. J.; Almond, M. J.; Marston, G.; Ly, V. P.; Wiltshire, K. S. Spectrochim. Acta, Part A 2000, 56, 2605−2616. (18) McIntire, T. M.; Ryder, O.; Finlayson-Pitts, B. J. J. Phys. Chem. C 2009, 113, 11060−11065. (19) Griesbaum, K.; Volpp, W.; Greinert, R.; Greunig, H. J.; Schmid, J.; Henke, H. J. Org. Chem. 1989, 54, 383−389. (20) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A.; Vreven, T. ; Kudin, 1716


The Overall Reaction Process of Ozone with Methacrolein and

Recommend Documents