Criegee intermediate CH2OO
Wei-Lun Ting, Chun-Hung Chang, Yu-Fang Lee, Hiroyuki Matsui, Yuan-Pern Lee, and Jim Jr-Min Lin
Citation: The Journal of Chemical Physics 141, 104308 (2014); doi: 10.1063/1.4894405
View online: http://dx.doi.org/10.1063/1.4894405
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/141/10?ver=pdfcov Published by the AIP Publishing
Articles you may be interested in
Ab initio chemical kinetics for the ClOO + NO reaction: Effects of temperature and pressure on product branching formation
J. Chem. Phys. 137, 014315 (2012); 10.1063/1.4731883
Reaction of C 2 ( a Π 3 u ) with methanol: Temperature dependence and deuterium isotope effect J. Chem. Phys. 133, 114306 (2010); 10.1063/1.3480395
Efficient dehalogenation of polyhalomethanes and production of strong acids in aqueous environments: Water-catalyzed O–H-insertion and HI-elimination reactions of isodiiodomethane ( CH 2 I–I ) with water
J. Chem. Phys. 120, 9017 (2004); 10.1063/1.1701699
The reaction of C 2 H with H 2 : Absolute rate coefficient measurements and ab initio study J. Chem. Phys. 116, 3700 (2002); 10.1063/1.1436481
Kinetics over a wide range of temperature (13–744 K): Rate constants for the reactions of CH(ν=0) with H 2 and D 2 and for the removal of CH(ν=1) by H 2 and D 2
Detailed mechanism of the CH
2I
+ O
2reaction: Yield and self-reaction
of the simplest Criegee intermediate CH
2OO
Wei-Lun Ting,1Chun-Hung Chang,1Yu-Fang Lee,2Hiroyuki Matsui,2Yuan-Pern Lee,1,2,a)
and Jim Jr-Min Lin1,2,3,a)
1Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 10617, Taiwan
2Department of Applied Chemistry and Institute of Molecular Science, National Chiao Tung University, 1001,
Ta-Hsueh Road, Hsinchu 30010, Taiwan
3Department of Chemistry, National Taiwan University, Taipei, Taiwan
(Received 29 May 2014; accepted 20 August 2014; published online 9 September 2014)
The application of a new reaction scheme using CH2I + O2 to generate the simplest Criegee in-termediate, CH2OO, has stimulated lively research; the Criegee intermediates are extremely impor-tant in atmospheric chemistry. The detailed mechanism of CH2I+ O2 is hence important in under-standing kinetics involving CH2OO. We employed ultraviolet absorption to probe simultaneously CH2I2, CH2OO, CH2I, and IO in the reaction system of CH2I+ O2 upon photolysis at 248 nm of a flowing mixture of CH2I2, O2, and N2 (or SF6) in the pressure range 7.6–779 Torr to inves-tigate the reaction kinetics. With a detailed mechanism to model the observed temporal profiles of CH2I, CH2OO, and IO, we found that various channels of the reaction CH2I + O2 and CH2OO + I play important roles; an additional decomposition channel of CH2I+ O2to form products other than CH2OO or ICH2OO becomes important at pressure less than 60 Torr. The pressure depen-dence of the derived rate coefficients of various channels of reactions of CH2I+ O2 and CH2OO + I has been determined. We derived a rate coefficient also for the self-reaction of CH2OO as k = (8 ± 4) × 10−11 cm3molecule−1s−1 at 295 K. The yield of CH
2OO from CH2I + O2 was found to have a pressure dependence on N2 and O2 smaller than in previous reports; for air un-der 1 atm, the yield of∼30% is about twice of previous estimates. © 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4894405]
I. INTRODUCTION
The reactions of O3 with alkenes are extremely impor-tant because they are responsible for the removal of both O3 and unsaturated hydrocarbons, and for the production of OH and organic aerosols in the troposphere.1–3The current model
indicates that cycloaddition of O3 to the C=C double bond of unsaturated hydrocarbons forms a primary cyclic ozonide, which rapidly cleaves its C–C and O–O bonds to form a carbonyl molecule and a carbonyl oxide that is commonly referred to as the Criegee intermediate.4–9 Previous under-standing of the mechanism of these reactions was based on in-direct laboratory observations of stable end products because the Criegee intermediates have eluded direct detection until recently.10,11
The recent applications of a new reaction scheme us-ing CH2I+ O2 to generate the simplest Criegee intermedi-ate, CH2OO, have stimulated lively research. Using this re-action scheme, CH2OO has been detected with photoioniza-tion mass spectrometry,12 ultraviolet (UV) depletion,13 UV
absorption,14,15 infrared (IR) absorption,16 and microwave
spectroscopy.17,18 With the use of some of these detection
methods, the kinetics of reactions of CH2OO with vari-ous compounds have been investigated experimentally.12,19–22
a)Authors to whom correspondence should be addressed. Electronic addresses: yplee@mail.nctu.edu.tw and jimlin@gate.sinica.edu.tw
Even though the reaction of CH2I + O2 has been investi-gated extensively,23–31 a detailed mechanism of this reaction
still needs to be established. The total rate coefficient for the reaction of CH2I + O2 was reported to be (1.28–1.82) × 10−12cm3molecule−1s−1,14,24,26,29,31but the branching of
various formation channels was not clearly characterized. The proposed mechanisms in the earlier studies were incomplete because only product channels for the formation of ICH2OO and H2CO+ IO, not CH2OO, were considered. The mecha-nisms employed in more recent reports include the formation of CH2OO but not the rapid self-reaction of CH2OO and var-ious channels of the reaction of CH2OO+ I; these reactions become important in some laboratory experiments involving large concentrations of CH2OO and I. A more detailed un-derstanding of the CH2I+ O2system covering experimental conditions over a wide range is thus desirable.
Using transient IR absorption to probe directly the decay of CH2OO, we found that the self-reaction CH2OO+ CH2OO was extremely rapid,21 and estimated the rate coefficient of
this self-reaction to be (4± 2) × 10−10 cm3molecule−1s−1 at 343 K. According to quantum-chemical calculations, this reaction is rapid because a cyclic dimeric intermediate is formed with large exothermicity (∼375 kJ mol−1) before fur-ther decomposition to 2 H2CO+ O2(1
g). The formation of this dimer, with the terminal O atom of one CH2OO bound to the C atom of the other CH2OO, reflects a unique prop-erty of the zwitterionic character of CH2OO in which the
0021-9606/2014/141(10)/104308/11/$30.00 141, 104308-1 © 2014 AIP Publishing LLC
terminal oxygen atom is partially negatively charged whereas the other oxygen atom and the C atom are partially posi-tively charged. While we were preparing this manuscript, a recent report by Buras et al. indicated that the rate coeffi-cient of this self-reaction of CH2OO, (6.0 ± 2.1) × 10−11 cm3molecule−1s−1 at 297 K, is much smaller than our pre-vious estimate at 343 K;32the kinetics were investigated with
probes of CH2OO at 375 nm and I atom at 1315 nm.
We have performed new experiments to probe simulta-neously the UV absorption of CH2I2, CH2OO, IO, and CH2I upon photolysis at 248 nm of a flowing mixture of CH2I2, O2, and N2 in the pressure range 7.6–779 Torr. The tempo-ral profiles of CH2OO and IO were analyzed simultaneously with a detailed reaction mechanism and the pressure depen-dence of the yield of CH2OO and rate coefficients of various channels of CH2I+ O2and CH2OO+ I has been character-ized. The rate coefficient of CH2OO+ CH2OO at 295 K was determined to be (8± 4) × 10−11cm3molecule−1s−1. II. EXPERIMENTS
The photolysis cell with transient UV absorption detec-tion has been described previously;33,34only relevant details are given here. The Criegee intermediate CH2OO was pro-duced from the reaction of CH2I with O2; CH2I was prepared by photodissociation of CH2I2at 248 nm.
The photolysis laser beam and the probe light (both di-ameter 1.8 cm) overlapped collinearly in the reaction cell (length 75 cm, inner diameter 2.0 cm), as shown in Fig.1. The photolysis laser beam at 248 nm from an excimer laser (Co-herent, CompExPro 205F, KrF, ∼50 mJ, 1 Hz) was intro-duced into the cell by reflection from an ultra-steep long-pass edge filter (Semrock LP02-257RU-25), which also limited the probe wavelength to be greater than 260 nm. The output of a high-brightness broadband light source (Energetiq, EQ-99) was collimated with a parabolic mirror (f= 50.8 mm) before entering the reaction cell. Another parabolic mirror (f= 101.6 mm) served to focus the light onto the slit of a spec-trometer (Andor SR303i) equipped with an intensified charge-coupled detector (iCCD, Andor iStar, DH320T-18F-E3). The resolution of the spectrometer was∼1.5 nm. The wavelength was calibrated with the emission spectrum of a low-pressure mercury lamp, with typical errors smaller than 0.5 nm.
The transient absorption spectra were recorded with the array detector at varied delays after photolysis. The delay (du-ration of reaction) was defined as the interval from the
pho-FIG. 1. Schematic experimental setup (not to scale). PM: parabolic mir-ror; HR248: highly reflective mirror at 248 nm; LP257: long-pass filter at 257 nm; PD: photodiode; and iCCD: image-intensified CCD camera.
tolysis laser pulse to the center of the detector gate (width = 1 μs). The reference spectrum was taken at 1 or 5 μs before the photolysis laser pulse. The integration interval at each delay was 1 μs; the delay was scanned automatically af-ter each photolysis pulse with a compuaf-ter program written in Andor Basic. The spectrum at a specific delay was typically averaged 60–200 times to achieve an adequate ratio of signal to noise.
A small fraction of CH2 might be produced through the photolysis of CH2I. We have tested the transient ab-sorption spectra with laser pulse energies varied from 30 to 100 mJ pulse−1. A small decrease (2%–4%) in the yield of CH2I was observed at laser energy greater than 100 mJ pulse−1, but an insignificant difference in the deduced spec-trum of CH2OO was found.
Liquid CH2I2was slightly heated to 305 K to ensure sat-uration of its vapor pressure and the CH2I2vapor was carried with a flow of N2 or O2. The mixing ratios of the reagent gases (CH2I2, N2, O2, SF6) were controlled with mass flow controllers (Brooks Instruments, 5850E). These gases were mixed in a Teflon tube before entering the reaction cell. A small stream (1%–2% of the total mass flow) of the N2/O2 mixture gas (with the same ratio as the reagent gas) was intro-duced to purge both windows of the cell to prevent undesired photochemistry at the window surfaces. This purging also de-creased the effective length of the sample from 75 to 72 cm, which was calibrated with the absorbance of CH2I2/N2gas of a known mixing ratio.
The linear flow velocity in the reaction cell was adjusted to be greater than 0.8 m s−1to allow refreshment of the sample gas between the photolysis laser pulses at a repetition rate of 1 Hz. The temperature of the reaction cell was 295± 1.5 K. The number density of CH2I2 was determined from its ab-sorption spectrum. The number densities of O2and N2 were deduced with the ideal gas law from the measured cell pres-sure and their mass flow rates.
III. RESULTS
A. Determination of[CH2I2], [CH2I], [CH2OO], and [IO]
A representative transient difference UV absorption spec-trum recorded 9 μs after photolysis of a flowing mixture of CH2I2/O2/N2 (0.044/10.4/90.7) at 101.1 Torr is shown in Fig.2. The spectrum was deconvoluted to spectra of CH2I2 (negative due to depletion), IO, CH2OO, and CH2I, as is discernible from the small residual after subtracting corre-sponding spectra of these four species. The absorption spectra of CH2OO and IO have characteristic progressions, whereas those of CH2I2and CH2I are broad but different in shape. On minimizing the residual between the simulated and experi-mental spectra in the region 265–480 nm with a least-squares fitting, the number densities of CH2OO, IO, and CH2I, and the decrease of that of CH2I2 after photolysis, were deduced according to the cross sections of these species (CH2I2 and IO,35 CH
2I,23 and CH2OO).15 The uncertainties in concen-tration measurements of CH2OO, IO, CH2I, and CH2I2 from deconvolution were estimated to be∼5%, 5%, 10%, and 5%,
FIG. 2. Comparison of experimental and simulated transient absorption spectra. The total simulation consists of absorbance of CH2I2 (depletion), IO, CH2OO, and CH2I. In this example (experiment no. 50),−[CH2I2] = 1.82 × 1014, [IO]= 6.56 × 1012, [CH
2OO]= 9.01 × 1013, [CH2I]= 4.98
× 1012molecule cm−3. Other experimental conditions are P= 101 Torr, P O2
= 10.4 Torr, PN2= 90.7 Torr, PCH2I2= 43.7 mTorr, and delay time = 9 μs.
respectively; errors throughout this report are 1σ in fitting un-less stated otherwise. In this case the uncertainties in cross sections are not included in the error estimates.
At higher pressure and at a later period of reaction, the product ICH2OO becomes important. The reported UV cross section of ICH2OO, with a maximum value ≥ 1.7 × 10−18 cm2molecule−1 (Ref. 27) or 2.5 × 10−18 cm2molecule−1 near 330 nm,23is much smaller than that of CH
2OO that has a maximal value 1.2× 10−17cm2molecule−1at 340 nm.15At
760 Torr, the concentration of ICH2OO might become twice that of CH2OO; consequently, using the reported cross sec-tion of ICH2OO, ICH2OO might contribute up to∼30% of observed UV absorption near 340 nm. However, according to theoretical calculations, ICH2OO is not the main carrier of the previously observed spectrum in the region 275–450 nm;36 absorption of CH2OO might have some contribution. To clar-ify this issue, we utilized the characteristic oscillatory pattern in the region 352–404 nm, likely due to a vibronic progres-sion of CH2OO, to show that the presence of ICH2OO at high pressure affects little our determination of CH2OO concen-trations from observed UV spectra, as discussed in Sec.Iand demonstrated in Fig. S1 of the supplementary material.37
Because the rate coefficient of the self-reaction of CH2OO requires accurate measurements of [CH2OO], the ac-curacy of the UV cross section of CH2OO is important. As compared in our previous paper,15 for CH
2OO near 295 K, the UV absorption spectrum reported by Sheps14agrees with our spectrum showing similar vibronic structures on the long-wavelength side, but decreases in absorbance much more rapidly than ours on the structureless short-wavelength side. More importantly, the cross sections reported by Sheps14
were approximately 3.3 times ours. We think that our mea-surements are more reliable for the following reasons: (1) Our technique has mass selectivity and employs a depletion method to compare with a known reference molecule. This method was proved reliable in the measurements of cross
sec-tions of ClOOCl.38 (2) We utilized both the SO2 scaveng-ing reaction and the self-reaction of CH2OO to extract the CH2OO spectrum and obtained consistent results with signal-to-noise ratios superior to other reports. (3) If we used the cross section of Sheps, not only the transient absorption spec-tra could not be deconvoluted satisfactorily, but also the yield of CH2OO from CH2I+ O2 at low pressure would become ∼0.25, much smaller than the values 0.87–1.0 reported by Huang et al.30 and 0.67–1.0 reported by Stone et al.31 With
the cross section determined by us, the yield 0.72–0.78 agrees with those in previous reports. (4) Buras et al. recently re-ported cross sections of CH2OO at 375 nm to be (7±7
3.5) × 10−18cm2molecule−1(Ref. 39) and (6.2± 2.2) × 10−18 cm2molecule−1,32 in satisfactory agreement with our value
of (7.6± 1.1) × 10−18cm2molecule−1with the uncertainty limits overlapping with each other.
B. Temporal profiles of CH2OO and IO in experiments with N2and O2
A summary of experimental conditions and fitted rate co-efficients of some representative experiments is presented in TableI; a complete list of a total of 64 experiments is available in Table SI of the supplementary material.37
Representative temporal profiles of CH2I, CH2OO, and IO recorded upon photolysis of a flowing mixture of CH2I2 (24.3 or 24.5 mTorr), O2 (both 10.5 Torr), and N2 (92.7 or 90.4 Torr) in two experiments (nos. 60 and 63 in TableI) near 100 Torr and 295 K with small [CH2OO]0are shown in Fig.3. The decrease in concentration of CH2I2upon laser irradiation, hence [CH2I]0, was−[CH2I2]= [CH2I]0∼= 1.17 mTorr (3.8 × 1013 molecule cm−3). The rapid increase of CH
2OO and decay of CH2I were due to the reaction of CH2I+ O2. After approximately 8 μs, CH2OO began to decay and the concen-tration of IO gradually increased. In this figure (and all oth-ers showing experimental temporal profiles) we also show the simulated temporal profiles of CH2I, CH2OO, IO, ICH2OO, H2CO, and I according to the proposed mechanism, as dis-cussed in Sec.IV A.
For comparison, Fig.4shows results of two experiments (nos. 49 and 52 in Table I) under similar conditions except for increased [CH2OO]0. The decrease in concentration of CH2I2upon laser irradiation, hence [CH2I]0, was−[CH2I2] = [CH2I]0 ∼= 3.54 mTorr (1.16 × 1014 molecule cm−3). In this example, [CH2OO] decreased more rapidly than in Fig. 3, indicating the second-order nature of the decay of CH2OO. Similarly, [IO] increased more rapidly than in Fig.3, indicating that some IO was produced from subsequent reactions of CH2OO.
The temporal profile of CH2I, CH2OO, and IO recorded upon photolysis of a flowing mixture of CH2I2(42.9 mTorr), O2 (51.1 Torr), and N2 (354.4 Torr) in an experiment (no. 2 in TableI) at high pressure (405.5 Torr) and 295 K is shown in Fig. S2 of the supplementary material.37 In this example, CH2OO increased rapidly because of larger [O2] employed and the yield of IO (and ICH2OO by simulation) relative to CH2OO was enhanced, indicating that some IO was produced from secondary reactions of ICH2OO of which the concentra-tion was expected to be enhanced at high pressure.
TABLE I. Representative experimental conditions, fitted rate coefficients, yield y and fraction of survival β of CH2OO in the CH2I+ O2system at 295 K.
Expt. PTotal PO
2 PN2 PCH2I2 −[CH2I2] k1a
k
1b k1c k3
no. (Torr) (Torr) (Torr) (mTorr) (1013)a (10−12)b (10−12)b (10−12)b (10−12)b yc βc
3 779.2 5.3 773.9 42.3 8.3 0.44 1.06 0.00 104 0.30 1.00 4 777.0 53.1 723.9 42.1 8.5 0.42 1.08 0.00 93 0.28 1.00 36 758.0 163.0 595.0 41.6 9.8 0.56 0.94 0.00 65 0.37 1.00 18 606.6 10.3 596.3 41.3 8.8 0.49 1.01 0.00 109 0.33 1.00 14 602.1 10.4 591.7 26.5 5.8 0.49 1.01 0.00 67 0.33 1.00 13 515.4 10.5 504.9 26.3 6.0 0.51 0.99 0.00 72 0.34 1.00 17 497.8 10.0 487.8 40.5 8.8 0.55 0.95 0.00 97 0.37 1.00 2 405.5 51.1 354.4 42.9 9.1 0.60 0.90 0.00 65 0.40 1.00 35 399.3 85.4 313.9 41.0 10.2 0.69 0.81 0.00 71 0.46 1.00 16 305.3 10.3 295.0 41.2 9.0 0.68 0.82 0.00 91 0.46 1.00 12 304.0 10.3 293.7 26.1 6.0 0.69 0.81 0.00 60 0.46 1.00 15 205.5 10.3 195.2 40.8 9.3 0.79 0.71 0.00 83 0.53 1.00 11 200.1 10.2 189.9 26.9 6.4 0.78 0.72 0.00 74 0.52 1.00 49 104.3 10.8 93.5 43.8 11.4 0.96 0.54 0.00 85 0.64 1.00 60 103.2 10.5 92.7 24.3 3.8 0.94 0.56 0.00 72 0.63 1.00 52 101.0 10.4 90.6 44.5 11.8 0.98 0.52 0.00 84 0.65 1.00 63 100.9 10.5 90.4 24.5 3.8 0.96 0.54 0.00 74 0.64 1.00 53 100.9 10.4 90.5 44.6 18.4 0.96 0.54 0.00 85 0.64 1.00 42 100.5 10.7 89.8 14.3 2.1 1.00 0.50 0.00 82 0.67 1.00 66 62.5 10.3 52.2 49.8 7.6 1.07 0.43 0.00 80 0.71 1.00 65 61.6 10.2 51.4 49.3 7.6 1.08 0.42 0.00 85 0.72 1.00 9 41.0 2.3 38.7 32.2 9.8 1.12 0.18 0.21 104 0.74 0.85 6 40.7 2.3 38.4 37.1 9.2 1.13 0.19 0.18 101 0.75 0.86 25 31.2 5.0 26.2 27.6 7.5 1.12 0.15 0.23 107 0.75 0.83 21 30.9 5.4 25.5 26.4 7.1 1.11 0.14 0.25 97 0.74 0.81 64 21.9 11.0 10.9 40.2 6.7 1.17 0.14 0.19 87 0.78 0.86 8 20.3 2.2 18.1 34.7 9.6 1.15 0.08 0.27 91 0.77 0.81 23 11.1 5.8 5.3 26.1 7.8 1.09 0.07 0.33 79 0.73 0.77 19 10.7 5.5 5.2 27.1 7.3 1.16 0.09 0.24 81 0.78 0.83 33 7.9 7.9 0.0 42.1 15.2 1.08 0.03 0.39 59 0.72 0.73 29 7.6 7.6 0.0 41.0 14.4 1.13 0.04 0.33 62 0.76 0.78 aIn unit of molecule cm−3. bIn unit of cm3molecule−1s−1. cy= k 1a/(k1a+ k1b+ k1c); β= k1a/(k1a+ k1c).
FIG. 3. Temporal profiles of concentrations of IO, CH2OO, and CH2I recorded upon photolysis of a flowing mixture of CH2I2(24.3 or 24.5 mTorr), O2 (both 10.5 Torr), and N2 (92.7 or 90.4 Torr) at 295 K in two experi-ments (nos. 60 and 63, symbols and∇). The decrease of concentration of CH2I2 was 1.17 mTorr (3.8× 1013molecule cm−3). Concentrations of IO
(red), CH2OO (black), CH2I (blue), ICH2OO (gray), I (magenta, scaled by 0.5), and H2CO (green, scaled by 0.5) simulated with a model described in Sec.IV Aare shown as lines. The inset shows a more detailed view in the 0–40 μs region.
FIG. 4. Temporal profiles of concentrations of IO, CH2OO, and CH2I recorded upon photolysis of a flowing mixture of CH2I2(43.8 or 44.5 mTorr), O2(10.8 or 10.4 Torr), and N2(93.5 or 90.6 Torr) at 295 K in two experi-ments (nos. 49 and 52, symbols and∇). The decrease of concentration of CH2I2was 3.53 mTorr (∼1.16 × 1014molecule cm−3). Concentrations of IO
(red), CH2OO (black), CH2I (blue), ICH2OO (gray), I (magenta, scaled by 0.5), and H2CO (green, scaled by 0.5) simulated with a model described in Sec.IV Aare shown as lines. The inset shows a more detailed view in the 0–40 μs region.
FIG. 5. Temporal profiles of concentrations of IO, CH2OO, and CH2I recorded upon photolysis of a flowing mixture of CH2I2(41.0 or 42.1 mTorr) and O2(7.6 or 7.9 Torr) at 295 K in two experiments (nos. 29 and 33, symbols
and∇). The decrease of concentration of CH2I2was 4.51 mTorr (∼1.48 × 1014molecule cm−3). The concentration of IO simulated with the model
described in Sec.IV Ais shown as dashed-dotted red line. Concentrations of IO (red), CH2OO (black), CH2I (blue), ICH2OO (gray), I (magenta, scaled by 0.5), and H2CO (green, scaled by 0.5) simulated with a model described in Sec.IV Bare shown as lines. The inset shows a more detailed view in the 0–40 μs region.
For experiments at low pressure, representative temporal profiles of CH2I, CH2OO, and IO recorded upon photolysis of a flowing mixture of CH2I2(41.0 or 42.1 mTorr) and O2(7.6 and 7.9 Torr) in two experiments (nos. 29 and 33 in TableI) at 295 K are shown in Fig.5. The decrease in concentration of CH2I2upon laser irradiation, hence [CH2I]0, was−[CH2I2] = [CH2I]0 ∼= 4.52 mTorr (1.48 × 1014molecule cm−3). The yield of IO (and ICH2OO by simulation) relative to CH2OO was diminished as compared with experiments at high pres-sure. Similar to all experiments with P≤ 20 Torr, the rise of [CH2OO] was slightly slower than simulated, to be discussed in Sec. IV B. Furthermore, the simulated profile of IO us-ing the same model (Sec.IV A) as that for high pressures (P ≥ 100 Torr) is shown in a dashed-dotted line in Fig.5; it is significantly greater than experimental observation, to be dis-cussed in Sec.IV B.
C. Experiments using SF6as a quencher
Five experiments using SF6as an efficient quencher in the pressure range 20.0–62.7 Torr at 295 K were performed; the experimental conditions and fitted rate coefficients are pre-sented in Table SII of the supplementary material.37A
repre-sentative temporal profile of CH2I, CH2OO, and IO recorded upon photolysis of a flowing mixture of CH2I2(38.3 mTorr), O2(5.2 Torr), and SF6(15.4 Torr) in an experiment (no. 75 in Table SII) at 20.6 Torr and 295 K is shown in Fig. S3 of the supplementary material.37 The decrease in concentration of
CH2I2upon laser irradiation, hence [CH2I]0, was−[CH2I2] = [CH2I]0∼= 3.02 mTorr (9.9 × 1013molecule cm−3). Unlike in experiments with N2 and O2 below 20 Torr (Fig.5), the rise of CH2OO has no delay and can be simulated satisfac-torily. Compared with experiments with O2 and N2, smaller
yields of CH2OO and IO due to efficient quenching of inter-nally excited ICH2OO were observed.
IV. DISCUSSION
A. Reaction mechanism and simulation of temporal profiles at P> 60 Torr
To describe the CH2I+ O2system in detail, we employed the following scheme:
in which ICH2OO∗ represents an energized adduct ICH2OO initially formed upon reaction of CH2I with O2. ICH2OO∗ might decompose to form the original reactants, proceed to form CH2OO + I (reaction (1a)), or become stabilized to ICH2OO (reaction(1b)), with branching ratios α1, α2, and (1 − α1 − α2), respectively. From the steady-state approxima-tion of [ICH2OO∗], described in detail in the supplementary material,37 α
1 and α2 were derived as functions of pressure and detailed rate coefficients indicated in grey in the scheme. The grey part in this scheme was not used explicitly in the kinetic fitting; only the effective reactions (solid parts) were used. Because of the large concentration of I atoms and the great reactivity of CH2OO, once produced, CH2OO reacts readily either with I atom (reaction (2)) or with itself (reac-tion (3)). According to theoretical calculations,21 the major
reactions of CH2OO with I atom proceed via three channels: attack of the terminal O atom of CH2OO by the I atom to form H2CO+ IO (reaction(2c)) and attack of the C atom by the I atom to form ICH2OO∗, followed by formation of CH2I + O2 (reaction (2a)) or stabilization to ICH2OO (reaction (2b)). ICH2OO further reacts with itself (reaction (4)) or I atoms (reaction(5)) to form ICH2O which subsequently de-composes to H2CO+ I (reaction(6)).
The mechanism is summarized below:
CH2I+ O2→ CH2OO+ I, k1a= α2k1, (1a) CH2I+ O2→ ICHM 2OO, k1b = (1 − α1− α2)k1, (1b) CH2OO+ I → CH2I+ O2, k2a = α1k2, (2a) CH2OO+ I→ ICHM 2OO, k2b = (1 − α1− α2)k2, (2b) CH2OO+ I → H2CO+ IO, k2c, (2c) CH2OO+ CH2OO→ 2H2CO+ O2(1g), k3, (3) This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
ICH2OO+ ICH2OO→ 2ICH2O+ O2 k4, (4) ICH2OO+ I → ICH2O+ IO, k5, (5)
ICH2O→ H2CO+ I, k6, (6) IO+ IO → products k7 (7) in which k4 = 9.0 × 10−11 cm3molecule−1s−1,27 k 5 = 3.5 × 10−11cm3molecule−1s−1, k 6= 1.0 × 105s−1,27and k7= 9.9× 10−11cm3molecule−1s−1were reported.40
Even though five rate coefficients are listed for reactions (1) and(2), we had to determine only k1a, k1b, and k2c be-cause K−1 = k2a/k1a = α1k2/α2k1 = α1k2b/α2k1b in which
K is the equilibrium constant of reaction (1a). According to calculations with the CCSD(T)//B3LYP/aug-cc-pVTZ-pp method, G = 10.4 kJ mol−1 for reaction(1a); hence K−1 = 71. Once k1a and k1b were determined, we calculated k2a and k2b from the relation k2a/k1a = k2b/k1b = K−1 on as-suming α1 = α2. As the total rate coefficient for reaction (1), k1a + k1b = (1 − α1) k1, was reported to be (1.28– 1.82) × 10−12 cm3molecule−1s−1 for pressure up to 250 Torr of Ar,14,24,26,29,31 we used the average value of k
1a +
k1b = 1.5 × 10−12 cm3molecule−1s−1 in the fitting; this value fits satisfactorily with the experimental decay of CH2I, when available, and the initial rise of CH2OO in all exper-iments in SF6 and those in N2 and O2 with P ≥ 60 Torr. The yield of CH2OO, y= α2/(1− α1)= [CH2OO]0/[CH2I]0, implies that k1a = y (k1a + k1b) and k1b = (1− y) (k1a +
k1b); k1a and k1b could be determined from y and k1a+ k1b. In the fit, the yield of CH2OO was initially estimated with
y = [CH2OO]0/[CH2I]0 in which [CH2OO]0 was estimated with a short extrapolation of the decay of [CH2OO] to t= 0 and [CH2I]0 = −[CH2I2] when we assumed that all pho-tolyzed CH2I2 produced CH2I + I. Subsequent fine adjust-ments of y were performed on fitting the observed [CH2OO] profile; k1aand k1bwere thereby derived.
The production of IO resulted mainly from two chan-nels: reaction (2c) from CH2OO+ I and reaction (5) from ICH2OO+ I. At low pressure, [CH2OO] is much greater than [ICH2OO], so that reaction (2c)is more important than re-action(5); in contrast, at high pressure, [ICH2OO] is much greater than [CH2OO], so that reaction(5)is more important than reaction (2c). Although the rate coefficient of k2c was predicted to be 5.5× 10−14cm3molecule−1s−1,21 we found
that we needed to set k2c = 9.0 × 10−12cm3molecule−1s−1 to fit the rise of IO properly, especially at low pressures. This deviation of k2cfrom a previous theoretical prediction is rea-sonable because this rate coefficient is sensitive to the barrier height and the calculated value of 7.9 kJ mol−1 might have been overestimated. From the sensitivity analysis shown in the supplementary material,37this rate coefficient has a
negli-gible effect on the rate coefficient k3in fitting the profiles of CH2OO.
It should be noted that the reaction
CH2OO+ CH2I→ C2H4I+ O2(3g−) (8) with k8 predicted to be 6.3 × 10−11 cm3molecule−1s−1 (Ref.21) is not included in the model because in our
exper-iments O2was in excess so that most CH2I was readily con-verted to CH2OO or ICH2OO. In experiments with significant [CH2I], this reaction has to be included in the model.
The fitting procedures thus became systematic on per-forming the following steps using either the Chemkin II program41 or a fitting program written with MATLAB; the
latter was more efficient and convenient to use; its validity was verified with the former. After the initial fitting of y with
k1a+ k1b fixed to 1.5× 10−12cm3molecule−1s−1 to derive
k1aand k1b, and subsequently k2aand k2busing an initial guess of K−1 = 70, the rate coefficient of self-reaction of CH2OO,
k3, was fitted with least squares to minimize the deviations of the simulated and observed temporal profiles of [CH2OO] and [IO]; k2c and k4–k7 were fixed in the fitting. Values of
k4–k6were taken from the literature, as listed previously, but
k7 = 1.5 × 10−10cm3molecule−1s−1 had to be used to ac-count for the decay of IO. The fit yielded k3 = (6.5 ± 2.1) × 10−11cm3molecule−1s−1, but k
3decreased with pressure for data in the pressure range 60–779 Torr. When K−1 was decreased to 40, we obtained k3 = (10.9 ± 2.2) × 10−11 cm3molecule−1s−1, with k
3increasing with pressure for data in the pressure range 100–779 Torr.
We then varied the value of K systematically and re-peated the fit, all data under varied experimental conditions were fitted satisfactorily to yield k3 independent of pressure only when K−1 was in the range 50–60; the best fits were with K−1= 55, which yields k3= (8.2 ± 1.4) × 10−11cm3 molecule−1s−1 for data in the pressure range 60–779 Torr. The value K−1 = 55 corresponds to G = 9.8 kJ mol−1 for reaction (1a), within expected uncertainties of the value 10.4 kJ mol−1 predicted with the CCSD(T)//B3LYP/aug-cc-pVTZ-pp method. Using this model, the simulated tempo-ral profiles of CH2I, CH2OO, IO, ICH2OO, I, and H2CO for the experiment are shown with thick lines in Figs.3and 4; these simulated profiles of CH2I, CH2OO, and IO agree satisfactorily with experiments. In contrast, the profile sim-ulated for IO in the experiment, shown as a dashed-dotted line (noted as “k1c= 0”) in Fig.5 using this model, is sig-nificantly greater than the experimental data, to be discussed in Sec.IV B.
B. Decomposition of ICH2OO∗or CH2OO∗ at P< 60 Torr
As indicated in Fig.5, after fitting the profile of CH2OO, the experimental temporal profile of IO disagrees with simu-lations using the mechanism discussed in Sec.IV A. The ob-served concentration of IO was significantly smaller than that simulated with this model and that of CH2OO showed a rise less rapid than simulation. These deviations were observed for all experiments performed below 60 Torr.
As discussed in Sec.IV A, IO could be produced from two channels: CH2OO+ I (reaction (2c)) and ICH2OO+ I (reaction(5)). At low pressure, as [ICH2OO] is much smaller than [CH2OO], most IO was produced from CH2OO + I. To reconcile the smaller [IO], we propose that some inter-nally excited ICH2OO∗or CH2OO∗ might have decomposed to form products other than CH2OO or ICH2OO at low pres-sure because of less efficient quenching. A similar mechanism
was also proposed by Eskola et al. for ICH2OO∗.29Hence, for experiments below 60 Torr, we modified the original reaction (1a)to include two channels,
CH2I+ O2→ CH2OO+ I, k1a= (1.5 × 10−12− k1b) β, (1a) CH2I+ O2→ products other than CH2OO or ICH2OO,
k1c= (1.5 × 10−12− k1b)(1− β) (1c) in which k1c = 0 and the fraction of survival of CH2OO
β = k1a/(k1a + k1c) = 1 for P > 60 Torr and β de-creases with pressure for P < 60 Torr. The decomposition of some ICH2OO∗or CH2OO∗at low pressure consequently ac-counted for the smaller concentrations of CH2OO and IO. The yield of CH2OO is consequently revised to be y= β k1a/(k1a + k1b) to include the data at low pressure; hence k1a= (k1a + k1b) y, k1b= (k1a+ k1b) (1− y/β), and k1c= (k1a+ k1b) y (1 − β)/β. This decomposition channel at low pressure is further supported by our observation of infrared absorption bands of CO and CO2 in the photolytic reaction of CH2I2 + O2 at 248 nm for pressures below 40 Torr.
With this revised model, we adjusted β and y to fit the observed temporal profiles of IO and CH2OO, with all other parameters determined the same way as in the case at high pressure. Subsequently, rate coefficient k3was fitted with least squares. After considering this decomposition channel, the temporal profiles of CH2OO and IO were simulated satisfac-torily, as shown in thick lines in Fig.5with β= 0.78 and 0.73 for experiment nos. 29 and 33, respectively. Twenty three ex-periments performed with total pressure below 60 Torr were fitted satisfactorily using this revised model with β = 0.73– 0.92 or k1cup to 3.9× 10−13cm3molecule−1s−1, 26% of the total rate coefficient of reaction (1).
The less rapid rise of CH2OO might be explained also by the decomposition mechanism. As at low pressure the quenching of ICH2OO∗ and CH2OO∗ is less facile, the pro-portion of the internally excited ICH2OO∗ or CH2OO∗ that decomposes increased, more at the earlier period of reaction. The proportion of the “loss” in CH2OO was hence largest at the beginning and decreased at the later stage of reaction. When an efficient quencher SF6 was employed, even at low pressure this delayed rise of CH2OO disappeared, as indicated in Fig. S3 of the supplementary material.37
C. Fitted rate coefficients and their dependence on pressure
This fitting procedure worked well for 69 sets of data in total in the pressure range 7.6–779 Torr in which the pres-sure of O2 was varied from 2.0 to 163.0 Torr, N2 from 0 to 773.9 Torr, SF6from 0 to 62.7 Torr, and [CH2I]0in the range (2.1–18.4) × 1013 molecule cm−3. The experimental condi-tions, the fitted rate coefficients, and the values of y and β thus derived for some representative experiments are summa-rized in TableI; a complete list is available in Tables SI (N2 and O2) and SII (SF6) of the supplementary material.37
FIG. 6. Dependence on pressure of the rate coefficient for the formation of CH2OO, k1a. (a) k1aas a function of total pressure P; the solid line is fit-ted according to Eq.(9); (b) (k1a+ k1b)/k1a= y−1as a function of P; k1a = k1a+ k1c.
1. Pressure dependence of k1aand k1b
The pressure dependence of k1ais shown in Fig.6(a); k1a represents the rate coefficient for the formation of CH2OO from CH2I+ O2. As expected, the rate coefficient decreases with pressure as the formation of ICH2OO becomes more im-portant. A plot of (k1a+ k1b)/k1aas a function of P is shown in Fig.6(b); in this case k1atakes into account the proportion that decomposes at low pressure. According to the rate expression shown in the supplementary material,37
k1a+ k1b k1a = 1 +
kq[M]
k−2 . (9)
The linear relation between (k1a + k1b)/ k1a and [M] (=P) and an intercept ∼1 are satisfactorily demonstrated in Fig. 6(b), supporting the validity of our model. Using this equation, we were able to derive kq/k−2= (1.1 ± 0.1) × 10−19 cm3molecule−1.
The dependence of k1b on pressure is shown in Fig.7(a); k1b represents the rate coefficient for the formation of ICH2OO from CH2I+ O2. A linear rise with pressure is observed for k1b at low pressure; k1b levels off at high pres-sure, characteristic of quenching stabilization of ICH2OO. A plot of (k1a + k1b)/k1b as a function of P−1 is shown Fig. 7(b); because this channel is important only at high pressure, we plot data only with P ≥ 100 Torr. According to the rate This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
FIG. 7. Dependence on pressure of the rate coefficient for the formation of ICH2OO, k1b. (a) k1bas a function of total pressure P; the solid line is fit-ted according to k1b= 1.5 × 10−12cm3molecule−1s−1 − k
1a; (b) (k1a+
k1b)/k1bas a function of P−1, P > 100 Torr.
expression shown in the supplementary material,37
k1a+ k1b k1b = 1 +
k−2
kq[M]. (10)
The linear relation between (k1a + k1b)/k1b and P−1 and an intercept∼1 are satisfactorily demonstrated in Fig.7(b). The satisfactory dependence of both k1aand k1bover a broad pres-sure range indicates that our model is adequate to describe the variation of concentration of CH2OO and IO. However, because we assumed that k1a + k1b = 1.5 × 10−12 cm3 molecule−1s−1, the equation k1b = 1.5 × 10−12 cm3 molecule−1s−1 − k1ainstead of Eq.(10)should be used for derivation of k1b.
Because of the equilibrium relationship, K−1 = k2a/k1a = α1k2b/α2k1b= 55, the pressure dependence of k2a and k2b follows that of k1aand k1b, respectively.
2. Pressure dependence of k1c
The dependence of k1con pressure is shown in Fig.8(a) and the dependence of β, the fraction of survived CH2OO, on pressure is shown in Fig.8(b); because of the smaller values, the errors in these measurements are greater than others. An initial increase in β with pressure before leveling off, similar to that of k1b, was observed; this is characteristic of quenching stabilization. A decrease with pressure is observed for k1c; the value becomes negligible near 60 Torr. The observed pressure dependence is consistent with a mechanism of stabilization of ICH2OO∗ and CH2OO∗ by collisional quenching. The β
FIG. 8. Dependence on total pressure of k1c(a) and the fraction of survival of CH2OO, β, in the pressure range 7–100 Torr (b); the solid line is fitted according to Eq.(11).
values are fitted to an equation
β = 1 − (0.47 ± 0.11)/[1 + (3.2 ± 1.2) × 10−18[M]]. (11) Because of the large uncertainties in β, the fitting reproduce
β values to only within 0.08.
3. Determination of k3
The fitted rate coefficient for self-reaction of CH2OO,
k3, ranged from 5.6 to 12.0 × 10−11 cm3molecule−1s−1, with an average of all data k3 = (8.2 ± 1.4) × 10−11 cm3 molecule−1s−1for experiments in N2and O2; the error limits represent one standard deviation in averaging. Some repre-sentative sensitivity analyses are shown in the supplementary material.37 At high pressure, the analysis clearly shows that
reactions (1a),(1b)and(2b)are more sensitive to the vari-ation of [CH2OO] than reaction(3), but the factors of reac-tions (1a) and(1b)have similar values with opposite signs because these reactions are competing with CH2I for the for-mation of either CH2OO or ICH2OO. The sensitivity of re-action (3), ∂x/∂(ln k3), in which x is the mass fraction of CH2OO, indicates that k3 has greater errors at high pres-sure than at low prespres-sure, mainly because [CH2OO] is small relative to [ICH2OO]. At lower pressure [CH2OO] is much greater than [ICH2OO] and the sensitivity of reaction(3) is greater than of reactions(2a)and(2b), but still smaller than of reactions (1a)and(1b). However, the effect of reactions (1a)and(1b)on k3cancels each other. Even under such con-ditions and in the critical range 20–150 μs, ∂(ln k3)/∂(ln ki)
∼
= 0.4–1.0 for i = 2a and 2b, indicating that k3is sensitive to values of k2aand k2b. Considering the large variations in k2a and k2bunder our experimental conditions, the consistency of values of k3throughout the pressure range supports the valid-ity of the mechanism including decomposition reaction(1c).
Two assumptions were made in these fits: k1a + k1b = 1.5 × 10−12cm3molecule−1s−1and α
1/α2= 1. The former was actually tested with our experimental data, even at high pressure, and we fixed this value simply to minimize the vari-ables in our fitting. The value of α1/α2might deviate slightly from 1; a deviation of 30% implies a deviation of k2bby 30%, which translates to∼20% in k3at low pressure. Considering the estimated relative error of 20% in concentration measure-ments of CH2OO, which translates to∼60% error in k3at high pressure and∼30% at low pressure based on sensitivity anal-ysis, 24% error induced by the uncertainty of K−1(=55 ± 15), 17% error in the least-square fit of each individual temporal profile, we report k3= (8 ± 4) × 10−11cm3molecule−1s−1 with the error representing the 95% confidence level.
Previously, using IR absorption to monitor CH2OO, we could only roughly estimate the rate coefficient of k3because of large uncertainties.21 Because the IR probe beam did not
follow the UV photolysis beam, the average concentration in the photolysis volume hence differed from that of the IR-probed volume. The conversion between concentrations of CH2OO in the UV-photolyzed volume and the IR-probed vol-ume had large errors. The initial concentration of CH2I in the photolyzed volume was estimated with the UV absorp-tion cross secabsorp-tion of CH2I2and the laser fluence, which was estimated from the energy and the size of the laser beam. The concentration of CH2OO measured from IR absorption also had large errors because the IR absorption cross section predicted with quantum-chemical computations might have large errors which affect the bimolecular rate coefficient by approximately the same factor. Furthermore, the mechanism employed previously ignored reactions(1b)and(1c), and also employed theoretically predicted rate coefficients for reac-tions(2a)–(2c). All these factors might result to an underesti-mated error bar of our previous estimate. The dependence of
k3 on temperature is expected to be small, so the previously reported value of k3 = (4 ± 2) × 10−10cm3molecule−1s−1 at 343 K might have been overestimated. Our determination of k3= (8 ± 4) × 10−11cm3molecule−1s−1 at 295 K in this work agrees with the recently reported value k3= (6.0 ± 2.1) × 10−11cm3molecule−1s−1at 297 K by Buras et al.,32even
though a simplified mechanism was employed in that work. D. Yield of CH2OO and its dependence on pressure
In Fig. 9 we plot the reciprocal yield of CH2OO,
y−1= (k1a+ k1b+ k1c)/k1a= (k1a+ k1b)/k1a, as a function of total density [M] for the pressure range 7.6–779.2 Torr. These data were fitted with the equations
y−1 = (1.24 ± 0.03) + (9.13 ± 0.33) × 10−20[M], (12) M= O2or N2
in which [M] is the density in molecule cm−3; the errors rep-resent one standard deviation in fitting. Our data are compared
FIG. 9. Reciprocal initial yield y of CH2OO from CH2I+ O2as a function of total density [M]. y−1= (k1a+ k1b)/k1a= (k1a+ k1b)/βk1a. The data were fitted to a line with intercept 1.24 and slope 9.1× 10−20cm3molecule−1 for M= O2or N2. The results of Huang et al.30for M= He, O
2, and N2,
with slopes 0.95, 1.14, and 2.41× 10−19cm3molecule−1, respectively, and
the result of Stone et al.31for M= O
2or N2, with a slope 1.90× 10−19
cm3molecule−1, are also shown for comparison.
with previous reports in Fig.9. The smaller yield of CH2OO (greater y−1) at low pressure is due to the decomposition of CH2OO∗/ICH2OO∗; this reduction of yield would not appear in the experiments with I-atom detection. It is unclear why Huang et al. observed significantly different pressure depen-dence on N2and O2;30 our results for M= O
2and N2agree satisfactorily with their report for M = He [slope = (0.93 ± 0.13) × 10−19] and O
2 [slope= (1.1 ± 0.1) × 10−19], but are much smaller than for M= N2 [slope= (2.4 ± 0.4) × 10−19].30Our results also have a pressure dependence much
less than that reported by Stone et al.31 with M= N2and O2 [slope= (1.90 ± 0.11) × 10−19for all data].
Our assumption of −[CH2I2] = [CH2I]0 might over-estimate [CH2I]0 by as much as 15%, but not significantly enough to explain the discrepancies. When we tested the power dependence of this discrepancy, we found that this dis-crepancy was not due to secondary photolysis of CH2I; it is likely due to the error of the reported cross section of CH2I. Some uncertainties in these previous reports might have re-sulted from the indirect methods employed by observation of I atoms and H2CO rather than CH2OO, and the uncertainties in analysis of observed temporal profiles with their models. It should be noted that we performed experiments at 248 nm whereas Huang et al. employed light at 355 nm.30Weather the
difference in internal energy of CH2I affects the stabilization of ICH2OO, even at high pressures, requires further investiga-tion. However, from the results of Stone et al. in which laser light at both 248 nm and 355 nm was used, the effect of pho-tolysis wavelength on yield of CH2OO is insignificant.31
The estimated yields of CH2OO from CH2I + O2 at 298 K and 760 Torr in air,∼15% reported by Huang et al.30
and 18% by Stone et al.,31 were proposed to have
signif-icant implications for the oxidation chemistry of halogen-containing organic compounds and for the atmospheric chem-istry in marine regions with large concentrations of CH2I2. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
TABLE II. Summary of the rate coefficients and yield of CH2OO derived from the CH2I+ O2system at 295 K.
Reaction/description Expressiona Conditions Assumptionsa
k1a CH2I+ O2→ CH2OO+ I 1.5× 10−12/{1+ (1.1 ± 0.1) × 10−19[M]} P≥ 60 Torr k1a+ k1b= 1.5 × 10−12 k1a CH2I+ O2→ CH2OO+ I β× 1.5 × 10−12/{1+ (1.1 ± 0.1) × 10−19[M]} P < 60 Torr k1a+ k1b+ k1c= 1.5 × 10−12 β Fraction of survival of CH2OO at low P,
β= k1a/k1a
1− (0.47 ± 0.11)/{1 + (3.2 ± 1.2) × 10−18[M]}b P < 60 Torr
k1b CH2I+ O2→ ICHM 2OO 1.5× 10−12− k1a k1a+ k1b= 1.5 × 10−12
k1c CH2I+ O2→ products other than CH2OO or ICH2OO k1a(1−β) P < 60 Torr k2a CH2OO+ I → CH2I+ O2 55 k1a K−1= 55 k2b CH2OO+ I→ ICHM 2OO 55 k1b K−1= 55, α1= α2 k2c CH2OO+ I → H2CO+ IO 9.0× 10−12 P-independent k3 CH2OO+ CH2OO→ 2 H2CO+ O2(1 g) (8± 4) × 10−11 P-independent y Yield of CH2OO, y−1= (k1a+ k1b + k1c)/k1a= (k1a+ k1b)/k1a y−1= (1.24 ± 0.03) + (9.13 ± 0.33) × 10−20[M], M= O2or N2 −[CH2I2]= [CH2I]0 aRate coefficient in cm3molecule−1s−1, [M] in molecule cm−3, and K, β, and y are dimensionless. K is the equilibrium constant of the reaction CH
2I+ O2= CH2OO+ I; K−1= 55
provides the best fitting.
bApproximate fitting of the scattered data below 60 Torr.
Our estimate of a yield∼30% at 295 K and 760 Torr would enhance these effects significantly.
Following the same method21to simulate the experimen-tal conditions in the laboratory investigations of the ozonol-ysis of C2H4 using the revised k3 value and a model re-ported previously,9 we found that, when the self-reaction of
CH2OO is included in the model, the simulated yield of hy-droperoxymethyl formate [HPMF, CH2(OOH)–O–CHO] and formic acid anhydride [FA, (HCO)2O] decreased by ∼7% and that of HCOOH increased by∼3%. The H2CO produced due to the self-reaction of CH2OO accounts for an additional yield ∼0.03. When we increased the rate coefficient for O3 + alkene from 1 × 10−18cm3molecule−1s−1 for C
2H4 to 1 × 10−16cm3molecule−1s−1 for larger alkenes,30 the
simu-lated yield of compounds due to the reaction of Criegee in-termediate+ HCOOH, corresponding to HPMF + FA in O3 + C2H4, decreased by 25%–30%, whereas that of HCOOH increased by 10%–12%. Furthermore, the additional carbonyl compounds produced from the self-reaction of the Criegee in-termediates account for a yield 0.11–0.14, explaining the ob-served stoichiometry ratio larger than unity. Even though the significantly reduced value of the rate coefficient k3decreased its effect on the laboratory ozonolysis experiments, it might play an important role when the concentration of CH2OO is large.
The rate coefficients k1a, k1a, k1b, k1c, k2a, k2b, k2c, k3, and values of β (fraction of survival of CH2OO) and y (yield of CH2OO) derived in this work are summarized in TableII.
V. CONCLUSION
To investigate the detailed kinetics of the CH2I+ O2 re-action, we monitored the UV absorption of CH2I2, CH2I, IO, and CH2OO simultaneously in the reaction system of CH2I+ O2 at 295 K upon photolysis of a flowing mixture of CH2I2, O2, and N2 (or SF6) at 248 nm. Using a detailed mechanism for the reaction, we simulated the temporal profiles of CH2OO and IO that agreed satisfactorily with experimental data over a wide range of experimental conditions with P = 7.6–779
Torr. We found that, at pressure below 60 Torr, some inter-nally excited ICH2OO∗or CH2OO∗decomposed; the fraction of survival β= k1a/(k1a+ k1c) was determined to be as small as∼0.75 near 7.8 Torr.
The feature of our mechanism is that we clearly spec-ified three channels for the reaction CH2I + O2 and three channels for CH2OO+ I, and include the self-reaction of CH2OO that becomes important when the concentration of CH2OO is large. The dependence of derived rate coefficients on pressure for the formation of CH2OO+ I (k1a), ICH2OO (k1b), and other products (k1c) from CH2I+ O2 and the for-mation of CH2I + O2 (k2a), ICH2OO (k2b), and H2CO + IO (k2c) from CH2OO+ I was determined; they conform to the expected behavior for enhanced stabilization of ICH2OO at higher pressure. We also determined a rate coefficient k3 = (8 ± 4) × 10−11 cm3molecule−1s−1 for the self-reaction of CH2OO, significantly smaller than our previous estimate at 343 K using IR absorption.
The dependence on pressure of the yield y of CH2OO from CH2I + O2 conforms to the equation y−1 = (1.24 ± 0.03) + (9.13 ± 0.33) × 10−20[M] in which [M]= O
2or N2is the total density in molecule cm−3. This dependence on pressure is smaller than in previous reports; the∼30% yield of CH2OO at 760 Torr much greater than values 15%–18% in previous reports might have a significant impact on the atmo-spheric chemistry of marine regions.
ACKNOWLEDGMENTS
Ministry of Science and Technology, Taiwan (Grant Nos. NSC102-2745-M009-001-ASP and NSC100-2113-M-001-008-MY3) and the Ministry of Education, Taiwan (“ATU Plan” of National Chiao Tung University) supported this work. The National Center for High-Performance Computing provided computer time.
1D. Johnson and G. Marston,Chem. Soc. Rev.37, 699 (2008).
2J. G. Calvert, R. Atkinson, J. A. Kerr, S. Madronich, G. K. Moortgat, T. J.
Wallington, and G. Yarwood, The Mechanisms of Atmospheric Oxidation
3O. Horie and G. K. Moortgat,Acc. Chem. Res.31, 387 (1998). 4R. Criegee and G. Wenner,Justus Liebigs Ann. Chem.564, 9 (1949). 5W. Sander,Angew. Chem. Int. Ed. Engl.29, 344 (1990).
6W. H. Bunnelle,Chem. Rev.91, 335 (1991).
7S. Hatakeyama and H. Akimoto,Res. Chem. Intermed.20, 503 (1994). 8G. Marston,Science335, 178 (2012).
9P. Neeb, O. Horie, and G. K. Moortgat,J. Phys. Chem. A102, 6778 (1998). 10C. A. Taatjes, D. E. Shallcross, and C. J. Percival,Phys. Chem. Chem. Phys.
16, 1704 (2014).
11C. A. Taatjes, G. Meloni, T. M. Selby, A. J. Trevitt, D. L. Osborn,
C. J. Percival, and D. E. Shallcross, J. Am. Chem. Soc. 130, 11883 (2008).
12O. Welz, J. D. Savee, D. L. Osborn, S. S. Vasu, C. J. Percival, D. E.
Shall-cross, and C. A. Taatjes,Science335, 204 (2012).
13J. M. Beames, F. Liu, L. Lu, and M. I. Lester,J. Am. Chem. Soc.134,
20045 (2012).
14L. Sheps,J. Phys. Chem. Lett.4, 4201 (2013).
15W.-L. Ting, Y.-H. Chen, W. Chao, M. C. Smith, and J. J. Lin,Phys. Chem.
Chem. Phys.16, 10438 (2014).
16Y.-T. Su, Y.-H. Huang, H. A. Witek, and Y.-P. Lee,Science340, 174 (2013). 17M. Nakajima and Y. Endo,J. Chem. Phys.139, 101103 (2013).
18M. C. McCarthy, L. Cheng, K. N. Crabtree, O. Martinez, T. L. Nguyen,
C. C. Womack, and J. F. Stanton,J. Phys. Chem. Lett.4, 4133 (2013).
19C. A. Taatjes, O. Welz, A. J. Eskola, J. D. Savee, D. L. Osborn, E. P. F.
Lee, J. M. Dyke, D. W. K. Mok, D. E. Shallcross, and C. J. Percival,Phys. Chem. Chem. Phys.14, 10391 (2012).
20D. Stone, M. Blitz, L. Daubney, N. U. M. Howes, and P. Seakins,Phys.
Chem. Chem. Phys.16, 1139 (2014).
21Y.-T. Su, H.-Y. Lin, R. Putikam, H. Matsui, M. C. Lin, and Y.-P. Lee,Nat.
Chem.6, 477 (2014).
22Y. Liu, K. D. Bayes, and S. P. Sander,J. Phys. Chem. A118, 741 (2014). 23J. Sehested, T. Ellermann, and O. J. Nielsen,Int. J. Chem. Kinet.26, 259
(1994).
24A. Masaki, S. Tsunashima, and N. Washida,J. Phys. Chem.99, 13126
(1995).
25S. Enami, J. Ueda, M. Goto, Y. Nakano, S. Aloisio, S. Hashimoto, and M.
Kawasaki,J. Phys. Chem. A108, 6347 (2004).
26S. Enami, Y. Sakamoto, T. Yamanaka, S. Hashimoto, M. Kawasaki, K.
Tonokura, and H. Tachikawa,Bull. Chem. Soc. Jpn.81, 1250 (2008).
27T. Gravestock, M. Blitz, W. Bloss, and D. E. Heard,ChemPhysChem11,
3928 (2010).
28V. G. Stefanopoulos, V. C. Papadimitriou, Y. G. Lazarou, and P.
Papagian-nakopoulos,J. Phys. Chem. A112, 1526 (2008).
29A. J. Eskola, D. Wojcik-Pastuszka, E. Ratajczak, and R. S. Timonen,Phys.
Chem. Chem. Phys.8, 1416 (2006).
30H. Huang, A. Eskola, and C. A. Taatjes, J. Phys. Chem. Lett.3, 3399
(2012); 4,3824(2013).
31D. Stone, M. Blitz, L. Daubney, T. Ingham, and P. Seakins,Phys. Chem.
Chem. Phys.15, 19119 (2013).
32Z. J. Buras, R. M. I. Elasmra, and W. H. Green,J. Phys. Chem. Lett.5,
2224 (2014).
33M.-N. Su and J. J. Lin,Rev. Sci. Instrum.84, 086106 (2013). 34M.-N. Su and J. J. Lin,GSTF J. Chem. Sci. (JChem)1, 52 (2013). 35S. P. Sander, J. Abbatt, J. R. Barker, J. B. Burkholder, R. R. Friedl,
D. M. Golden, R. E. Huie, C. E. Kolb, M. J. Kurylo, G. K. Moortgat, V. L. Orkin, and P. H. Wine, “Chemical kinetics and photochemical data for use in atmospheric studies,” Evaluation Number 17, JPL Publication 10-6, Jet Propulsion Laboratory, Pasadena, 2011, seehttp://jpldataeval.jpl.nasa.gov.
36E. P. F. Lee, D. K. W. Kok, D. E. Shallcross, C. J. Percival, D. L. Osborn,
C. A. Taayjes, and J. M. Dyke,Chem. Eur. J.18, 12411 (2012).
37See supplementary material at http://dx.doi.org/10.1063/1.4894405 for
analysis of the transient absorption spectra, kinetic analysis, additional tem-poral profiles, a complete list of experimental conditions and fitted results, and sensitivity analysis.
38H. Y. Chen, C.-Y. Lien, W.-Y. Lin, Y. T. Lee, and J. J. Lin,Science324,
781 (2009).
39Z. J. Buras, R. M. I. Elasamra, A. Jalan, J. E. Middaugh, and W. H. Green,
J. Phys. Chem. A118, 1997 (2014).
40R. Atkinson, D. L. Baulch, R. A. Cox, J. N. Crowley, R. F. Hampson, R. G.
Hynes, M. E. Jenkin, M. J. Rossi, and J. Troe,Atmos. Chem. Phys.7, 981 (2007).
41R. J. Kee, F. M. Rupley, and J. A. Miller, Chemkin-II, “A Fortran chemical
kinetics package for the analysis of gas-phase chemical kinetics,” Sandia Report, SAND89-8009B, 1995.
![FIG. 9. Reciprocal initial yield y of CH 2 OO from CH 2 I + O 2 as a function of total density [M]](https://thumb-ap.123doks.com/thumbv2/9libinfo/7575204.125765/10.918.476.835.75.342/fig-reciprocal-initial-yield-ch-function-total-density.webp)
