J. Phys. Chem. 1900, 84, 1664-1674
1664
Kinetics Studies of the Reactions of CH,02 with NO, NO2, and CH302at 298 K S. P. S a n d e r * and R. T. Watson Molecular Physics and Chemistry Section, Jet Propulsion Laboratory, Pasadena, California 9 1 103 (Received October 5, 1979) Publicatlon costs assisted by the National Aeronautics and Space Administration
-
-
-
The flash photolysis/ultraviolet absorption technique was used to measure the rate constants for the reactions CH3O2 + NO CH30 + NOz (l),CH302+ NOz + M CH3OZNOz+ M (21, and CH302+ CH302 products (3) at 298 K over the pressure range 50-700 torr. Values for kl and k 3 were determined to be (7.1 f 1.4) X cm3molecule-' s-l and (3.6 f 0.7) x cm3molecule-' s-l, respectively,where k3 is defined by the relation -d[CH302]/dt = 2h3[CH30z]2;hz was found to vary strongly with pressure, indicating that the reaction occurs primarily by addition and is in the falloff region between second- and third-order kinetics. Experimentally determined parameters describing the shape of the falloff curve for kz were in reasonable agreement with those obtained by using theoretical methods developed by Troe and co-workers. An upper limit of 7 x cm3 molecule-' s-l was determined for the rate constant for the reaction CH302+ CO products.
-
Introduction Recent interest in the gas-phase reactions of the methylperoxy radical, CH3O2, has been stimulated by its role in both atmospheric and combustion-related processes. Kinetic mechanisms describing the chemistry of urban smog and the natural and polluted troposphere and stratosphere all include the reactions of CH302.' It is therefore important to elucidate as clearly as possible the reactions of this species with other atmospheric constituents. Much of the early work on CH3O2 kinetics was performed by using static photolysis with ratios of rate constants being measuredSz4The recent discovery of a strong UV absorption band of CH302between 210 and 280 nm and the detection of CH302 by mass spectroscopy7 have provided new techniques for the direct study of CH302 kinetics. Flash photolysis/ultraviolet absorption (FP/ UV),s-12 molecular modulation ultraviolet absorption spectroscopy (MMS),l3-I5and discharge flow/mass spectroscopy (DF/MS)'J6 have now been used to measure rate constants of CH3Oz reactions. These reactions include CH302 + NO CH,O + NO2 (1) CH302 + NOz + M CH3O2NOZ+ M (2)
+
CH302 + CH3O2
2CH30 + 0, CHzO + CH3OH
-+
(34 + 02 (3b)
+
CH3OOCH3 02 CH3O2 + HO2 CH300H + 02 CH302 SOz CH30 + SO3
(3c) (4) (5) Reaction 1 is an important step in the atmospheric photochemical cycle in which NO is converted to NO2, a process which ultimately results in the formation of ozone. This reaction has been shown to be almost as rapid as the analogous process involving HOZl7(unless otherwise indicated, rate constants cited in the text are from ref 18) HOZ + NO OH + NO2 (6) +
+
-
-+
k, = 8 X
lo-''
cm3 molecule-l s-'
and has proved difficult to study because of the rapidity of the reaction and the low detection sensitivity for CH302 by all techniques. Although several studies using FP/UV, MMS, and DF/MS have been published,10J6JGagreement has not been particularly good and, in one measurement, only the lower limit to the rate constant could be measured.'l 0022-3654/80/20841664$01 .OO/O
Reaction 2 may play a role as a sink reaction for NO, in urban smog and atmospheric methane oxidation in a manner analogous to pernitric acid, HOZNOz,and peroxyacetylnitrate (PAN), CH3C(0)02N02.Because these compounds have a large activation energy for thermal decomposition (20-26 kcal/mol) the position of the formation/destruction equilibrium is strongly temperature dependent, which may account for some of the unusual temperature effects observed in smog chambers.lg In the only reported direct study of this reaction, Cox and TyndalP5found that the rate constant was either independent of total pressure or depended only very weakly on total pressure between 50 torr of Ar/CH4 and 540 torr of N2, implying that the reaction is near its high-pressure limit at a relatively low pressure. Reaction 3 may play a role in the methane oxidation process occurring in the unpolluted troposphere where NO, levels are reasonably low. This reaction, which has three thermodynamically possible product channels, has been studied by FP/UV and MMS with good agreement being obtained for the overall rate c o n ~ t a n t . ~ ~Because ~ J - ' ~ it is a second-order process, however, the rate constant measurement relied upon the determination of the absolute CH302concentration. The rate constant is then derived from the product of two experimental observables, k 3 / c and c,where u is the CH302absorption cross section. At the moment, substantial disagreement exists between measurements of u by FP/UV and MMS. In addition, the branching ratios for channels 3a, 3b, and 3c are poorly defined. In this study, rate constants for reactions 1,2, and 3 were measured by using the FP/UV technique. Because of several improvements in the apparatus design, the disappearance of CH302 could be observed over a wide range of reactant concentrations and total pressure. It is shown that, contrary to previous measurements, reaction 2 is still in the falloff region at 700 torr and that the high-pressure limit is attained only at pressures exceeding several atmospheres. Experimental Section The apparatus has been described in detail previously.z0 The flash assembly consists of four concentric tubes comprising the reaction cell, photolyzing light filter, xenon flash lamp, and cooling/heating jackets. Two identical cells were constructed-one made of quartz and the other of Pyrex. Reagents are premixed and continuously flowed through the reaction cell (2.54 cm i.d., -95 cm long) with a resi0 1980 American Chemical Society
Kinetics Studies; of CH302
The Journal of Physical Chemistry, Vol. 84, No. 13, 1980
dence time of 15-30 51. This permits complete replacement of cell contents between flashes to minimize the photolysis and reaction of stable products from the previous flash and to ensure thait the starting reagents will not be depleted. Flow meters were individually calibrated for each gas by using a bubblle calibrating system (Hastings Model HBM1). The optical train consists of a 150-W xenon arc lamp, a White 8-pass mirror system (1 = 720-cm Pyrex cell and 769-crn quartz cell), and a McPherson Model 216.5 0.5-m monochromator (slit width = 300 pm, 0.6-nm resolution FWHM). Dichroic mirrors are used throughout to minimize reflection losses in the 220-270-nm spectral region. Proper calibration of both the flow system and the optical system was verified by comparing concentrations of NOz measured on the basis of partial flows and UV absorption. Agreement was typically better than 2%. Signals from the photomultiplier tube (EM1 9659 QA) are amplified and stored in a signal averager (TracorNorthern 570A) operating in the analogue mode. Multiple flashes (30-500) are employed to improve the signal-tonoise ratio (S/N). Further improvements in S/N are obtained by treating the data with a five-point smoothing algorithm.21Because of scattered light from the photolysis flash, the acquisition of data was delayed for -50 ps after the flash. Methylperoxy radicals were produced either by photolysis of chlorine in Clz-CH4-Oz mixtures in the quartz or Pyrex cell or by photolysis of azomethane, (CH3)zNz,with oxygen in the quartz cell. CH302 is produced by reactions 7-9 in the ClZ-CH4--O2system Cl2 + hv 2C1 (7) C1 -t CHI HC1+ CH3 (8) hB = 1.0 :K 10-l3 cm3 molecule-l s-l
-
CH3 + Oz + M CH3O2 + M cm3 molecule-' s-l h9 = 2.0 x
(9)
-+
2z
--
and by reactions 9 and 10 in the (CH3)zN2-02system (10) (CH3)zN2+ hu 2CH3 + O2 (9) CH3 Oz + M CH302 + M The ranges of reagent concentrations were the following 1.1-7.9; [CH,] X (in molecules ~ m - ~ [Cl,] ): X 2.0-15; [O,] X W7,1.1-4.8; [(CH&NZ] X 1.7-5.4; 4.8-23; [NO2]X lo-',, 4.0-42. [CH3O2I0(in [NO] X For molecules ~ m - ranged ~) from 3.4 X 1013to 2.0 X all experiments, CH302was formed on a time scale (20-70 ps, 95% formation) a t least ten times shorter than its loss (>190 ps, 95% removal). The photolyzing light filters on both quartz and Pyrex cells were filled with absorbing gases to isolate certain spectral regions. In the CH3O2 + NO study, it was desirable to inhibit the formation of O3 because of its overlapping absorption with CH3O2. This was accomplished by filling the filter cell with 760 torr of SOz, which prevented the photolysis of Oz. Since K 2 is several times smaller than hl at low pressures, a concerted attempt was made to minimize the photolysis of NOz in the CH302 NOz study. For this experiment, the filter cell was filled with a ClZ-Br2mixture which, at equilibrium, contained 60 torr of a BrCl and 220 torr of Clz. BrC1, which has an absorption band centered at 370 nm intermediate in intensity between Br2 and Clz, absorbs a portion of the photolysis light which would otherwise result in NO2 photolysis. Relatively low flash energies (e700 J/flash) were used in these experiments. Repeated flashing of static N02-He mixtures indicated that NOz photolysis was limited to 0.1% or less per flash.
+
+
1665
The CH302absorption was detected a t 245 nm for the CH3Oz + NOz and CH3O2 + CH3Oz reactions and at 270 nm for the CH30z+ NO and CH3O2 + CH3Oz reactions. The temperature of the reaction cell was maintained at 298 f 1 K by circulating methanol through the outer cell jacket from a constant-temperature circulator. Azomethane was synthesized by the method of Renaud and L e i t ~ h N02-02 .~~ mixtures were made by reacting small amounts of NO (Matheson C.P. Grade, 99.0% purity) with a large excess of Oz and allowing sufficient time for complete conversion. N2O4 corrections were negligible. Diluent gases had the following stated purities: He (Linde TJHP Grade, 99.999%), Nz (Linde UHP Grade, 99.999%), O2 (Linde UHP Grade, 99.99%), SF6(Matheson instrument purity, 99.99%). Chlorine (Matheson research purity, 99.96%) and methane (Matheson purity, 99.99%) were used without further purification.
Results CH302+ NO CH30 + NO2. The rate constant for the CH302+ NO reaction was measured over a wide range of reactant concentrations and total pressure. In early experiments where CH302was monitored at X = 245 nm, a substantial residual absorption of the analyzing light was observed after each flash, which was attributed to the formation of methyl nitrite by reaction 11. The identity CH30 + NO + M CH30N0 + M (11) --+
-
kll = I x 1O-l' cm3 molecule" s-l 24 of the product was verified by observing that the magnitude of the residual absorption could be predicted from a measurement of [CH3OZl0and the methyl nitrite absorption cross section.z5 Moreover, the shape of the absorption spectrum of the product closely matched the literature spectrum of CH30N0. Formation of CH30N0, which occurs on the same time scale as the disappearance of CH302,severely interferes with the detection of methylperoxy radicals. Correcting for this absorption is both difficult and inaccurate because the true value of Io is not well-known. For this reason, the analysis wavelength was shifted to 270 nm. Although this reduces the detection sensitivity for CH3Oz by a factor of 2, the ratio crCHaO2/ OCH,ONO is increased from 1.7 at 245 nm to 7.1 at 270 nm, significantly reducing the interference and simplifying the data analysis. Methylperoxy radicals were always formed on a time scale at least ten times faster than their removal. The ratios of [NO], to [CH3O2I0ranged from 7.1 to 23.4 with an average of 13.0, resulting in good pseudo-first-order conditions. Decays of CH3O2 were typically observed over a factor of 8-10 in concentration with a detection limit . corresponds to a around 4 X 10l2 molecules ~ m - ~This minimum detectabIe absorption of 0.5% which is attained after 50-500 flashes. A typical first-order decay plot is shown in Figure 1. The values of k; obtained in each kinetic run were corrected for two effects-absorption of the analytical beam by the product, CH30N0, and minor departures from true pseudo-first-order kinetics. A correction for the small residual absorption was evaluated from both computer simulations of the complete mechanism and an analytical expression derived from a simplified mechanism. The correction amounted to a 2.9% effect a t X = 270 nm. The first-order rate constant was also corrected for the minor consumption of NO by reaction with CH302(reaction 1)and CH30 (reaction 11). The correction varied with the ratio [NOlO/[CH3Oz],and averaged 4.7%. The correction for the bimolecular disproportionation of CH302
1666
Sander and Watson
The Journal of Physical Chemistry, Vol. 84, No. 13, 1980 C H 3 0 0 t NO
l!k
c
20
x
k
-
1
l
-
llimI
Q
,I 1 0
\ 200
3M)
5CQO
400
1
/
/.
I
1M)
PRODUCTS
298K, M He
2t -
+
fl
TOTAL PRESSURE
=
350 torr
T I M E (microseconds1
Figure 1. Pseudo-first-order decay of CH 00 radicals reacting with NO. [NO] = 1.07 X 10" molecules cm-?4 and [CH,OO], E 1.0 X molecuies ern-,.
was negligible (1.0 3.0 f 7.1 f
16 15 11 10 this work
0.2 1.4
kle = 7.5 x
k14
-+
k4 = 6.5
CH3 + O2 + M
k g = 2.6
CH3 + HCO
cm3 molecule-' s-'
= 6.8 X
+
(14) 29
CH3O2 + M
(9)
cm6 molecule-2s',M = N2
X
+
cm3 molecule-' s-'
HO2 + CH3O2
In the study of Plumb et al., a number of secondary reactions take place which were not mentioned by the authors. These reactions, along with possibly inadequate first-order conditions, make the results of their system somewhat difficult to interpret. In their study, CH3O2 is formed in the following sequence of reactions:
0 + CzH4
Molecules ~ m ' ~ .
+
technique
DF/MS MM/UV FP/UV FP/UV FP/UV
2.0 2.0
10-l' cm2.
-
CH3O2 in the region of the flow tube before NO is injected by the reactions HOz + HO2 H202 0 2 (18)
TABLE VI: Summary of Previous Work on the CH,O, + NO Reaction ref
10-17. [OJb 1.7 1.9 2.3 2.8 2.5 2.6 2.6 2.2 2.2 2.2 2.2 0.27 0.31 0.83 2.1 2.1 2.1 6.6 1.9 1.9 0.25 2.4
10-52k3/ U , cm 1013k,, 10-18. 10.13. molecm3 mole[COIb [CH3O,lOb cule-ls-' cu1e-I s-' 0 5.2 2.26 3.41 2.02 3.05 5.7 4.4 2.28 3.44 4.9 2.08 3.14 7.8 1.91 2.88 5.2 2.21 3.34 4.2 2.10 3.17 10 4.75 3.56 22 5.89 4.42 6.3 5.55 4.16 3.3 5.00 3.75 10 7.44 5.58 3.0 5.70 4.28 14 6.51 4.88 3.2 5.31 3.99 17 5.47 4.10 17 6.21 4.66 19 5.99 4.49 1.2 16 5.73 4.30 2.3 17 5.67 4.25 2.6 2.6 5.35 4.01 5.2 17 5.47 4.10
+
CH300H
+ 02
30
(4)
cm3 molecule-' s-'
X
In this region, CH302 will be removed by reactions 3 and 4 and CH3 will be removed by disproportionation CH3 + CH3
M +
C2HG
(19)
k19 = 5.2 X 10-l' cm3 molecule-' s-' Simulations of the reactions that occur in the flow tube, using conditions given in the paper, indicate that at the NO injection point, [CH302]= (1.1-1.6) X 10l2molecules cm-3 and [HO,] = (1.3-2.0) X 10l2molecules ~ m - After ~. the injection point, several reactions occur which remove NO: CH3Oz + NO CH30 + NO2 (1) HO2 + NO OH + NO2 (6)
- + +
The authors considered the reactions
0 + HCO
+
CO
+ OH
(15a)
+COz+H k15a
+ k15b = 2.1 x CH3 + HCO
(15b)
cm3 molecule-' s-' +
CH4 + CO
29
(16)
as possible loss processes far HCO. However, with [Ozl = 2.2 X 10l6molecules cm-3 and [O], N 4 X 10l2molecules the reaction HCO kl, = 5.1
X
+ 02 lo-''
+
HOz + CO
cm3 molecule-l s-'
(17) 29
will dominate the removal of HCO. The H 0 2 formed in reaction 17 will have time to react with itself and with
CH30 + NO
(MI
CH30N0
H 0 2 is also regenerated by CH30 + O2 H 0 2 HCHO cm3 molecule-' s-' k13 = 6.0 X
(11)
(13)
In the Plumb et al. study, [NO] ranged from 4.0 X 10l2to while the initial oxygen atom 31 X 10l2 molecules concentration was held at 4.0 X 10l2molecule cm-3.31 In ), the worst case ([NO], = 4.0 X 10l2molecules ~ m - ~computer simulations of the reactions occurring after the NO injection point, using initial CH302and H02 concentrations calculated above, indicate that almost 60% of the NO is removed by reactions 1, 6 and 13 at the l / e point for CH302. The absence of good first-order conditions would be expected to result in curvature in the plots of In [CH302]vs. time in addition to that already attributed by
Kinetics Studios of CH,O,
The Journal of Physical Chernktry, Vol. 84, No.
73, 7980 1671
k,(T) is the rate constant in the second-order regime, and the authors to a residual ion fragment at m / e = 47. F, is a parameter which characterizes the broadening of Computer simulations indicate that, under these conditions ) , rate constant ([NO] = (4-8) X lo1' molecules ~ m - ~the the falloff curve due to the energy dependence of the rate would be underestimated by -25% over the 30-ms reacconstant for the decomposition of the vibrationally excited tion period. The lack of curvature observed by Plumb et intermediate.3zi34 al. in either the CH3Oz decay plots or the plot of kl' vs. Equation I is derived from a semiempirical fit to the [NO] is therefore !surprising. Two possibilities can be RRKM formalism. The utility of fitting experimental data suggested: (1)overestimation of [O], and therefore the to such an expression is that the rate constant for any value number of CH3O2 and HOz radicals initially formed and of total pressure can be determined (for the purpose of (2) the presence of an additional unknown removal atmospheric modeling, for example) once the parameters mechanism for CH3O2 which is effective before or after the are determined. This is much easier than computing the NO injection point. rate constant from a full RRKM calculation or from tables Two flash photolysis studies of this reaction have been of Kassel integrals. reported.lOJx Because of their low detection sensitivity for The parameters Ito, k,, and F, were derived in this study CH302,Anastasi et, al.ll could report only a lower limit, by nonlinear least-squares curve fitting to eq I. For a given kl > 1.0 X cm3 molecule-l s-l, for the rate constant. diluent gas (He, Nz, or SF,) multiple sets of solutions can The FP/UV study of Adachi and Basco suffers from be obtained which give nearly the same fit to the data. complications due l;o secondary chemistry. Although the However, the constraint is posed that the values of k, and experimental technique and reaction mechanism are simF, (neglecting weak collision effects) must be the same for ilar to this work, there exists an important difference. In each diluent gas. The procedure adopted in this study was both systems, CH30 radicals from reaction 1 will quickly to map the entire domain of reasonable parameter values, react with NO to form CH30N0. In analyzing for CH3O2 holding k, and F, constant, and determine ko by least at 245 nm, the technique of Adachi and Basco is suscepsquares. The parameters finally selected were the ones tible to the interference caused by the formation of that resulted in overlapping regions of minimum xzfor all CH30N0 on the same time scale as the CH3Oz disapthree data sets. It should be emphasized that at least two, pearance. This has the effect of reducing the apparent and preferably three, different third-body data sets are CH302loss rate, which is consistent with their much lower required to assign unambiguously the three parameters. result for kl. However, since their Io value was derived This is particularly true for a reaction such as CH3Oz + from a blank cell, the CH30N0 formation should have NOz where the rate constant at the highest measured resulted in both nonlinear first-order decay plots and a pressure is well below the high-pressure limit, yet where large residual absorption. Neither of these effects is there is significant falloff at the lowest measured pressure. mentioned by the authors, and they cannot be explained. As indicated above, our study circumvents the problem by The parameters which gave the best fit to all three data analyzing for CH302at 270 nm where CH30N0 does not cm6 sets were the following: ko = (1.19 f 0.06) X absorb strongly. molecule-z s-l, M = He; (2.33 f 0.08) X cma molecule-2 C H 3 0 2 NOz + M. The single previous direct kinetic s-l, M = N,; (3.94 f 0.13) X cm6 molecule-z s-l, M = study of this reaction is by Cox and Tyndall15who report SF,; k, = (8.0 f 1.0) X cm3 molecule-l s-l; and F, = kz = (1.2 f 0.3) X cm3 molecule-ls-l at 50 torr of Ar 0.4 f 0.10. CHI and kz = (2.6 f 0.3) X cm3 molecule-l 5-l at A procedure for the estimation of F, from structural 540 torr of Nz.The small pressure dependence observed information about the adduct has been developed by Troe is used to argue that kz is near the second-order limit at and ~ o - w o r k e r s .F,~ ~is ~given ~ ~ by a product of strong50 torr total pressure. Although their measurement at 50 collision and weak-collision broadening factors, F,BC and torr of Ar/CH4 is in good agreement with this study at 50 FC? torr of N2, the marked pressure dependence observed in F ~ ~ t t r e - ~ r o n g l ~ ~ ~ ~ ~ *cause - ~ ~ ~ ~ r ~ n c l u s i o ~F, = F,BCFCWc of the relatively long time scales encountered in their molecular modulation experiment, several sources of sysThe overall expression for the rate constant, applicable tematic error may appear which account for their results. in all pressure regimes, is then given by These include depletion of NOz along the length of the cell and thermal decomposition of the adduct, CH3OZNO2, in the period required for cell transit. As indicated above, because the time scale of the flash photolysis experiment where FLHis the Lindemann-Hinshelwood factor is short and the cell contents are totally replaced between flashes, difficulties due to CH3OZNOzdecomposition are avoided. The experimental falloff curves obtained in this study show the typical behavior expected for the formation of Values of F,R"have been determined by Luther and a stabilized adduct from a vibrationally excited intermeT r ~ ine tabular ~ ~ form as a function of the parameters Sk diate. The vibrational quenching rate increases with the and Bk Skis given by size, complexity, and total pressure of the diluent gas. ~ ~shown , ~ ~ that Recent work by Troe and c o - w ~ r k e r shas SkN Seff+ 1 or Seff+ 2 the rate constant falloff curves of addition reactions can be describeld by a khree-parameter equation: where Seffis the effective number of transition state oscillators. Seffcan be estimated from the vibrational partition function of the adduct molecule from the relation
+
+
where ko(T)is the rate constant in the third-order regime,
1672
The Journal of Physical Chemistry, Vol. 84,
Sander and Watson
No. 13, 1980
TABLE VII: Estimated Vibrational Frequencies for CH,O,NO, frequency, cm-I t Y Pe modela 3006 2930 2000 1759 1467 1464 1301 1182 1050 1049 928 804 709 633 455 450 303 120
CH, degen stretch CH, sym stretch OON sym stretch NO, asym stretch CH, degen def CH, sym def NO, sym stretch CH, rock OON asym stretch CO stretch 00 stretch NO, scis NO, wag NO, stretch NO, rock OON bend OC bend OC torsion
CH,F CH,F N,O FONO, CH,F CH,F FONO, CH,F N, 0 CH,F FONO, FONO, FONO, FONO, FONO, N2O FONO, FONO,
a Frequencies taken from ref 37 with minor corrections where appropriate.
where S = number of internal modes including hindered rotations (3N - 6). The parameter Bk is given by
Bk
B'(Sk - 1) s-1
where
B' = Eo + a(Eo)E, kT and where Eo = the critical energy for the unimolecular decomposition of CH302N02,E , = the zero-point energy of vibrations, and a(Eo)= the Whitten-Rabinovitch factor.34 Higher-order approximations to Bk are derived by Luther and T r ~ e . ~ ~ Since a complete normal-coordinate analysis of CH3OzN02is lacking, vibrational frequencies were estimated by comparison with model compounds (FON02,CH3F,and N20). These frequencies are listed in Table VII. At 300 K, S k N 3.8, E , = 41.6 kcal mol-l, and a(Eo)N 1.0. Eo is assumed to be 20 kcal mol-l by analogy with H02N02 thermal decomp~sition.~'The resulting value of Bk is 14, and from the tables in ref 33, F,SC = 0.51. The weak-collision broadening factor, F,", can be approximated by a function of only one parameter, the is given by collision-efficiency factor /3,. Fcwc N
FcwcN p, 0.14
and kp = the low-pressure-limiting strong-collision rate constant; kOSCcan be estimated by methods derived by Troe.33i35Values of kOscfor several third-order reactions of atmospheric interest have recently been computed,36and the value for reaction 3 has been estimated to be 1.4 X cm6 molecule-2 s-l for M = Nz. From this, values of kOsC can be calculated for the other diluent gases by correcting the Lennard-Jones collision frequency. The resulting values of p, are 0.044 (M = He), 0.17 (M = N2),and 0.39 (M = SF,). Since FcwC is a relatively weak function of p,, weak-collision effects in F, are masked by experimental for the three diluent gases error. The average value of Fcwc is 0.77. The composite broadening factor, F,, is calculated to be 0.51 X 0.77 = 0.39, in good agreement with the value of 0.40 f 0.10 derived from the experimental data. At low temperatures (300-600 K), Skis small relative to S, but FF is relatively sensitive to sk. As a result, small errors in the estimation of Sk resulting, for example, from uncertainties in the estimation of the low-frequency vibrational modes have a large effect on F,. Another source of error in Sk is the uncertainty in the difference between Sk and Sefp The exact difference is33 E,, - Eo kT where E,, is the Arrhenius activation energy at the highpressure limit. Since E,, is often experimentally unobtainable, the energy difference must be obtained from a complete RRKM calculation or by estimation, as above. CH302 + C H 3 0 2 . Recently, a large number of studies on the CH3O2 + CH302 reaction have been publ i ~ h e d . ~ ~The ~ ~ lresults l - ~ ~ of these studies are listed in Table VIII. Although the room-temperature rate constants are in reasonable agreement, the agreement is misleading. The fundamental parameter measured, k3/ 0, cannot be directly compared because most of the studies were carried out at different wavelengths. Of the three measurements of the CH302absorption spectrum, two are in excellent agreement8s9whereas the third is scaled upward by about 50%.14 If the measured spectrum of Hochanadel et al. is used to derive values of k , from all the different to 4.1 X studies, the spread in k, ranges from 2.6 X cm3molecule-l s-l. The rate constant obtained in this study, (3.7 f 0.7) X cm3 molecule-l is in excellent agreement with the results of Kan et al.? Sanhueza et al.,12 and Hochanadel et ala8 The interpretation of the kinetic data from this and other studies of the CH302+ CH302reaction is complicated by the possibility of a reaction branch forming CH30, i.e. CH302+ CHBOZ ZCH30 + O2 (34
-
where
AH = +3.5 kcal mol-l In addition there are two exothermic reaction paths: TABLE VIII: Summary of Previous Work on the CH,O, 1 0 - 4 k , / ~ cm ,
ref 13 14 8 11 9 12
this work
A.
nm
220-290 231 235 238 265 253.7 245 270
+ CH,O,
Reaction 1013k3,cm3 m o l e c u l e - I ~ - ~ 1018u.a cmz molecule-1s-1 8.4 11 8.0 20 15 11 28
5.5 i 1.0 3.3 5.5, ref 1 4 2.0 2.5, ref 8
2.2-4.4 4.6 i: 1.2 3.8 i: 0.7 4.4 t 1.1 4.1 i 0.5 3.7 i: 0.3
1 0 ' * o , ~cmz
3.2 3.3 3.2 1.8 2.5 3.0 1.5
Reference number given if taken from the literature. a CH30, cross section used to derive h,. k , derived from measured k,/a and U . ~ sured by Hochanadel e t a1.8
1013k,,Ccm3 molecule-'^“ 2.7 3.8 2.6 3.7 3.7 3.2 4.3
c 0.7
0.7 0.6 t 0.4 i: 0.3 i: 0.2 c 0.5 i i
CH,O, cross section mea-
Kinetics Studies of CH,02
The Journal of Physical Chemistry, Vol. 84, No. 13, 1980
-
CH30a + CHS02
CH20 + CH30H + 0,
AH = -77.4 kcal mol-l CH302CH3 + 0
2
(3b)
t 3c)
AH = -42.8 kcal mol-l
--
Secondary removal of CH302by the processes CH:,02 + CH30 CH20 CH300H CH3O + 02 HO2 + CH2O CH302 it HO2 CH302H + 0 2
+
(12)
-
(13) (4) occurs subsequent l,o the formation of methoxy radicals. Several radical termination steps also occur which reduce the probability of secondary CH302removal. These include CH30 CH30 products (20) CH30 -+ HO2 CH30H + 02 (21) HO2 t HO2 H202 + 02 (18) Kan et al. have considered the effect of the above reactions on the overall CH302removal rate.g Their conclusion, based on the assumption that
-
+
-+
-k3a
k3a
+ k3b + k3c
= 0.33
was that the secondary consumption of CH302would result in a maximum error of 12% in the determination of k3. This is smaller than the error factor of 1.33 that would be expected if every CH30 radical were effective in removing a CH302 radical. The relationship between the true rate constant for CH302disproportionation, k3, and the observed rate constant, kQ, is strongly affected by the relative and absolute rates of the secondary reactions which consume CH3O2 CH302+ CH30 CH2O + CHBOOH (12) CH3O2 + HO2 CH302H + 02 (4)
--
the termination reactions for CH30 and H02,and the rate of conversion of CH30 to H02. The relative importance of these processes can be qualitatively determined by varying [02].In the limiting cases in which reactions 12 and 4 are both extremely rapid or extremely slow compared to competing processes, varying [O,] will have no effect on kQ, and k3//k3 will be (k3a + k 3 ) / k 3or 1.0, respectively. If reaction 12 is fast and reaction 4 is slow, increasing [a,]will reduce k3’. If reaction 4 is fast and reaction 12 is slow, increasing [O,] will increase kQ. Parked4 observed no change in kQ as [O,] was varied from 3 X 1015to 2.5 X loxgmolecules ~ m - In ~ ,this study, however, kQ decreased by about 15% as [O,] was varied from 2.7 X 10l6to 6.7 X 1017molecules ~ m - ~ . Addition of CO to the system resulted in a slight decrease in kQ. CO may have reacted directly with CH302 by the reaction CH302+ CO CH30 + C 0 2 (22) or with CH30 CH30 + CO CH3 + C02 (234 HCHO + HCO (23b) thereby competing with reaction 12. If the decrease in k i is attributablle solely to reaction 22, one can show that kz2 5 7 X 10-ls crn3molecule-’ s-l. Only one study of reaction 23 has been conducted which gave kz3/k11 = 5 X 10-14.2g If this measurement is correct, the presence of CO in the concentration range (1-5) X l0ls molecules ~ r n should -~ have resulted in almost complete removal of CH30 by this
--
1673
reaction. However, since neither the rate constant nor the branching ratio of reaction 23 is known with certainty, little can be said about the effect of adding CO except that it suggests that CH30 is present and that kSa is probably nonzero. The same conclusion can be reached concerning the effect of varying [O,]. To date, data on the temperature dependence of reaction 3 are limited to the observation by Parkes14and Anastasi et a1.l1 that kQ was independent of temperature between 288 and 298 K and 298 and 325 K, respectively. According to the most recent evaluation of kinetic and thermochemical data on CH302,18reaction 3a is endothermic by 3.5 kcal/mol. Even if reaction 3a had a 3.5 kcal/mol activation energy, it would be difficult to observe a temperature dependence for k, over the limited range they employed. Because of the large uncertainty on the CH3O2 enthalpy of formation, however, reaction 3a may be thermoneutral or even slightly exothermic. Acknowledgment. The authors are grateful to W. B. DeMore, G. W. Ray, J. C. Brock, and J. Margolis for helpful discussions and to S. L. Manatt and C. Kan (The Ohio State University) and M. Kalloo (Merck, Sharp and Dohme Ltd.) for assistance in the preparation of azomethane. The research described in this paper was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under NASA Contract NAS7-100.
Appendix Effect of Reaction Product Absorption on the Measured First-Order Rate Constant, k’. For a reaction such as CH302N02+ M CH302+ NOz + M in which both the minor reactant species (CH3O2) and a product (CH302N02)absorb a fraction of the analytical light, the effect of the product absorber on k’ must be considered. From Beer’s law It = Io exp(-l(a,x + ayy)) (A-1) where It = the transmitted light intensity, Io = the “true” Io, i.e., intensity of the light beam emerging from the evacuated cell, 1 = the path length, x = [CH302],y = [CH302N02],0; = the CH3O2 absorption cross section, and cy = the CHg02N02absorption cross section. For first-
order conditions x = xo exp(-k’t)
(A-2)
where
k’= k[NOz] and (-4-3)
y=xo-x
Substituting for x and y in eq A-1 gives It = IO exp(-lxo((a, - cy)expt-k’t)
+ cy))
(A-4)
Under experimental conditions, Io is usually taken to be the transmitted light intensity at the end of the reaction. However, with a stable product absorber, this will be somewhat less than the “true” Io,Le. I{ = lim It = Io exp(-lxoay) (A-5) t-The apparent concentration of the minor reactant, x ’, is then x’ = In (&‘/It)
Io exp[-lxoayl Io exp(-lxo((ax - ay) exp(-k’t) = Ixo(a, - uy) exp(-k’t)
= In
+ uz)) (A-6)
J. Phys. Chem. 1980, 84, 1674-1681
1674
A plot of In x’vs. t will have slope -h’, the true first-order rate constant. If x’is calculated by using lorather than I,,’, plots of In x’vs. t will be nonlinear, and the rate constant thus obtained will be in error. When the product absorber comes from the second of two consecutive first-order reactions, e.g. CH302+ NO CH30 + NOz
+ -
CH30 + NO M CH30N0 + M the rate constant and the simple relationship derived above no longer apply, and the correction to the first-order rate constant must be obtained by numerical techniques. Note Added in Proof. Another study of the CH3O2+ NOz + M reaction has recently been published (H. Adachi and N. Basco, Int. J. Chem. Kinet.,12, 1 (1980)). While the value for M = O2 at 108 torr total pressure ((1.57 f 0.30) X cm3 molecule-’ s-’) is in excellent agreement with the results of this study at 100 torr for M = Nz, no pressure dependence over the range 108-583 torr is observed, which is contrary to our measurements. References and Notes (1) (a) T. A. Hecht, J. H. Seinfeki, and M. C. Dodge, Environ. Sci. Techno/., 8, 327 (1974); (b) J. A. Logan, M. J. Prather, S. C. Wofsy, and M. 6. McEiroy, Phibs. Trans. R. Soc. London, Ser. A, 290, 187 (1978); (c) W. L. Chameides, Geophys. Res. Lett., 5, 17 (1978). (2) C.W. Spicer, A. Villa, H. A. Wiebe, and J. Heickien, J. Am. Chem. SOC., 95, 13 (1973). (3) R. Simonaitis and J. Heickien, J. Phys. Chem., 78, 2417 (1974). (4) C. T. Pate, 6. J. Finiayson, and J. N. Pitts, Jr., J. Am. Chem. Soc., 96, 6554 (1974). (5) R. A. Cox, R. G. Derwent, P. M. Holt, and J. A. Kerr, J. Chem. SOC., Faraday Trans. 1, 72, 2444 (1976). (6) R. Simonaitis and J. Heickien, J . Phys. Chem., 79, 298 (1975). (7) (a) R. Atkinson, B. J. Finiayson, and J. N. Pitts, Jr., J. Am. Chem. Soc.,96, 6554 (1974); (b) N. Wash& and K. D. byes, Int. J. Chem. Kinet., 8, 777 (1976). (8) C. J. Hochanadei, J. A. Ghormley, J. W. Boyie, and P. J. Ogren, J. Phys. Chem., 81, 3 (1977). (9) C. S. Kan, R. D. McQuigg, M. R. Whitbeck, and J. G. Caivert, Int. J. Chem. Kinet., 11, 921 (1979). (10) H. Adachi and N. Basco, Chem. Phys. Lett., 63, 490 (1979). (1 1) C. Anastasi, I. W. M. Smith, and D. A. Patkes, J. Chem. Sm., Faraday Trans. 1 , 74, 1693 (1978).
Kinetics Study of the CI(,P) 4- C1,O G. W. Ray, L. F. Keyser, and
I?.
-
(12) E. Sanhueza, R. Simonaitis, and J. Heickien, Int. J. Chem. Kinet., 11, 907 (1979). (13) D. A. Parkes, D. M. Paul, C. P. Quinn, and R. C. Robson, Chem. Phys. Lett., 23, 425 (1973). (14) D. A. Parkes, Int. J. Chem. Kinet., 9, 451 (1977). (15) R. A. Cox and G. S.Tyndali, Chem. Phys. Lett., 65, 357 (1979). (16) I. C. Plumb, K. R. Ryan, J. R. Steven, and M. F. R. Muicahy, Chem. Phys. Lett., 63, 255 (1979). (17) (a) C. J. Howard and K. M. Evenson, Geophys. Res. Lett., 4, 437 (1977); (b) C. J. Howard, W. M. 0. Symposium on the Geophysical Aspects and Consequences of Changes in the Composition of the Stratosphere, Toronto, June 1978. (18) CODATA Task Group on Chemical Kinetics, J. Phys. Chem. Ref. Data, in press. (19) W. P. L. Carter, A. M. Winer, K. R. Darnall, and J. N. Pitts, Jr., Environ. Sci. Techno/., 13, 1094 (1979). (20) R. T. Watson, S.P. Sander, and Y. L. Yung, J. Phys. Chem., 83, 2936 (1979). (21) A. Savitsky and M. J. E. Goiay, Anal. Chem., 36, 1627 (1964). (22) High pressure limit based on the work of Hochanadei et al. (ref 8). Allowance has been made for the fact that the reaction is in the transition region between second- and third-order kinetics between 50 and 700 torr. (23) R. Renaud and L. C. Leitch, Can. J. Chem., 32, 545 (1954). (24) High pressure limit based on the work of L. Batt and G. N. Rattray, to be published. The uncertainty in this rate constant may be as high as a factor of 3 but the exact value does not affect our results. See ref 22. (25) J. 0. Caivert and J. N. F‘itts, Jr., “Photochemistry”, Wiley, New York, 1966. (26) R. A. Graham, A. M. Winer, and J. N. Pitts, Jr., Geophys. Res. Lett., 5, 909 (1978). (27) R. A. Graham, A. M. Winer, and J. N. Pitts, Jr., J. Chem. Phys., 68, 4505 (1978). (28) R. Simonaitis and J. Heickien, Chem. Phys. Lett., 65, 362 (1979). (29) R. F. Hampson, Jr., and D. Garvin, Natl. Bur. Stand. U . S . Spec. Pub/., No. 513 (1978). (30) k , , is estimated from the data of Burrows et ai. (ref 18) and R. A. Cox and J. P. Burrows, J. Phys. Chem., 83, 2560 (1979). (31) I.Plumb, M. Mulcahy, personal communications. (32) J. Troe, J. Phys. Chem., 83, 114 (1979). (33) K.Luther and J. Troe, “Weak Coiilsion Effects in DissociationReactions at High Temperatures”, presented at the 17th International Symposium on Combustion, Leeds, Aug, 1978. (34) P. J. Robinsonand K. A. Holbrook, “Unimoiecular Reactions”, Wiley, London, 1972. (35) J. Troe, J. Chem. Phys., 66, 4758 (1977). (36) NASA Panel for Data Evaluation, “Chemical, Kinetic and Photochemicai Data for Use in Stratospheric Modeling, EvaluationNo. 2”. Jet Propulsion Laboratory Publication 79-27, Apr, 1979. (37) T. Shimanouchi, J. Phys. Chem. Ref. Data, 6, 993 (1977).
CI, 4- CIO Reaction at 298 K
T. Watson”
Molecular Physics and Chemistry Section, Jet Propulsion Laboratory, California Institute of Techno/ogy, Pasadena, California 9 1 103 (Received January 3, 1980) Publication costs assistedby the NationalAeronautics and Space Adrninistratlon
The low-pressure discharge flow technique has been used in conjunction with collision-free sampling mass spectrometry (DF/MS) and atomic resonance fluorescence (DF/RF) to study the kinetic behavior of atomic chlorine with chlorine monoxide (ClzO)at 298 K. In order to minimize complications caused by secondary kinetic processes, we used pseudo-first-order conditions in each study: excess atomic chlorine in the DF/MS experiments and excess ClzO in the DF/RF experiments. The reaction and the two values measured for the rate coefficient can be written as follows: C1(T) + ClzO Clz+ C10 (l),kl = (9.33 f 0.54) X lo-” cm3molecule-’ s-’ (DF/MS), kl = (10.3 f 0.8) X lo-’’ cm3 molecule-'^^^ (DF/RF). The reported value for k l is (9.8 f 0.8) x cm3molecule-’ s-’. This is the unweighted average of the two measurements. These results are compared with previous measurements for hl and used to reevaluate the results of some studies of the flash photolysis of OClO + ClzO mixtures.
-
Introduction In recent years there has been considerable interest in the kinetic behavior of the clO(2n) radical. Most of this interest has been stimulated by the need to understand the role of this radical in the atmospheric chlorine cycle 0022-3654/80/2084-1674$01 .OO/O
where it is currently thought to play a major role in photochemically controlling the stratospheric ozone layer.’ Laboratory studies of c10 kinetic behavior require methods of generating the radical in a rapid, clean manner. There are several possible reactions which can be used to 0 1980 American Chemical Society
