Reaction Mechanism of Atomic Oxygen with Hydrogen Sulfide at High

Aug 1, 1994 - of 0 atoms, that is, kl,/kl = 0.2 f 0.1 at 1520-1820 K. The rate constant kl .... 98, No. 34, 1994. Tsuchiya et al. stainless steel. Thi...
2 downloads 0 Views 560KB Size
J. Phys. Chem. 1994,98, 8419-8423

8419

Reaction Mechanism of Atomic Oxygen with Hydrogen Sulfide at High Temperature Kentaro Tsuchiya,'*t Keiichi Yokoyama,* Hiroyuki Matsui,r Masaaki Oya,? and Gabrielle Duprel Department of Reaction Chemistry, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113, Japan; National Institute for Resources and Environment, 16-3 Onogawa, Tsukuba, Zbaraki 305, Japan; Japan Atomic Energy Research Institute, Tokai-mura, Naka-gun, Zbaraki 319-1 1 , Japan; and Centre National de la Recherche Scientifque. 45071 Orleans, Cedex 2, France Received: April 14, 1994; Zn Final Form: June 23, 1994"

The reaction of atomic oxygen (3P)with hydrogen sulfide was investigated by the shock tube/laser photolysis method a t high temperatures (1 100-2000 K), where the time dependence of 0 and H atoms was monitored with atomic resonance absorption spectrometry (ARAS). 0 atoms were produced by the photolysis of SO2 by a K r F excimer laser behind reflected shock waves. The overall rate constant for the reaction 0 H2S products (1) was determined from the decay rate of the absorption of 0 atoms as kl = (2.5 f 0.6) X 10-10 exp[-(32 f 4) kJ mol-'/RT] cm3 molecule-' s-l. The branching fraction for the substitution channel 0 + H2S H HSO (IC) was determined by measuring the concentration of H atoms relative to the initial production of 0 atoms, that is, kl,/kl = 0.2 f 0.1 a t 1520-1820 K. The rate constant kl and the branching fraction kl,/kl were also evaluated based on the conventional transition state theory with the Wigner's tunneling correction for the potential energies. The branching fraction determined in this study was found to be consistent with TST calculation.

-

+

-

+

Introduction The reaction mechanism of O(3P)atoms with hydrogen sulfide is a key process of the initiation stageof oxidation and combustion of hydrogen ~ulfide.l-~This reaction has been studied e ~ t e n s i v e l y , ~but - ~ only ~ at temperatures below 500 K. Neither the rate constants at temperatures important in combustion systems nor the detailed reaction mechanism had been determined. HSOH was identified by IR absorption in a low-temperature (8 K) matrix reaction between H2S and 0 atom generated by the photolysis of 0 3 . 1 2 It has been often speculated from this evidence and by analogy with the reactions of oxygen atoms with organic sulfides that the reaction proceeds via the competition between direct abstraction ( l a ) and an addition path ( l b ) followed by decomposition of the excited adduct:9-*

o(3~) + H,S

-

OH

+ HS,

-

(H2SO)*

(la)

H + HSO

(1b)

The reactions of O(3P) atoms with thiols and sulfides are suggested to proceed by the electrophilic addition of the oxygen atoms to the sulfur. Gutman and co-workers examined the reactions of 0 atoms with a series of divalent sulfides (CH3SH, C ~ H S S Hand , CH3SCH3). They first analyzed the reaction products a t room temperature13 and then measured the rate constants for these reactions over the temperature range 250-500 K.9 The observed activation energies for these reactions are found to have a good linear relationship with the ionization potential, I,. Cvetanovic,l4 Huie and Herron,I5 and others have found that such a linear relationship also holds for the reactions of 0 atoms with a series of substituted olefins; this correlation has been interpreted as due to charge donation from the double bond to the 0 atom. Consequently, an activated complex of 0 atom and the olefin is formed as a reaction intermediate. It was suggested by Slagle et al. that 0 + organic sulfides reactions showed a good correlation with the 0 olefins.9 They attributed these observations as

+

t National Institute for Resources and Environment.

t Japan Atomic Energy Research Institute.

I The University of Tokyo. 1

Centre National de la Recherche Scientifique. published in Aduance ACS Abstracts, August 1, 1994.

e Abstract

indirect evidence of an electrophilic addition mechanism analogous to that for 0 +olefin reactions. However, the observed activation energy for the reaction 0 + H2S products was not consistent with the linear relationship between the organic sulfides and Z,. Thus, a clear indication of the reaction mechanism for (1) has not been given yet. Crossed molecular beam studies on the threshold collision energy have suggested that HSO is a direct products of the title reaction.I"l8 Takane and FuenoI9 have performed an ab initio quantum mechanical calculation of the potential energy surfaces for this reaction. The calculation indicates that instead of the additionchannel (lb), the reactionproceedsvia substitution ( IC),

-

O(3P) + H2S

-

HSO + H

(IC)

Le., no stable reaction intermediate H2SO is found along the reaction coordinate. However, the indicated energies of the transitionstates, 77 kJ mol-' for ( l a ) and 127 kJ mol-' for (IC), are much higher than experimental activation energies. As there has been no direct experimental information on the branching fractions for ( l ) , it may be meaningful to identify the reaction products and determine the branching fraction for (IC). In the present study, the overall rate constant kl was measured directly by monitoring the concentration of O(3P)atoms with the ARAS technique behind reflected shock waves over the temperatures 1100-2000 K, where 0 atoms were produced by laser photolysis (KrFexcimer laser; 248 nm) of S02. The concentration of H atoms is monitored by ARAS in order to determine the branching fraction kl,/kl. The accuracy for determination of the branching fraction was improved by measuring relative ARAS intensities of H and 0 atoms in the reference reaction of 0 2H2 H 2 0 2H, where 0 atoms were supplied by the ArF photolysis of SO2. By comparing the rate constant for the consumption of 0 atoms via ArF laser photolysis (193 nm) with that in KrF laser photolysis (248 nm), the side reaction of 0 atom with HS radical was also examined.

-

+

+

Experimental Section The details of the shock tube laser flash photolysis system were described elsewhere.20 The shock tube used in this study is a piston-actuated type. It was 5 cm id., 4 m long and made of

0022-3654/94/2098-8419$04.50/0 0 1994 American Chemical Society

8420

Tsuchiya et al.

The Journal of Physical Chemistry, Vol. 98, No. 34, 1994

stainless steel. This shock tube is advantageous for reproducing shock heated conditions, Le., high temperatures above 800 K can be reproduced within h 2 0 K accuracy behind the reflected shock wave region. The S / N ratio of ARAS signal is easily improved by signal averaging. O(3P) atoms were prepared behind reflected shock waves by the photolysis of SO2 using mainly a KrF (248 nm) excimer laser. The sample gas mixtures were irradiated by the laser beam through a rectangular quartz window located at the end plate of the shock tube. The laser was introduced with a 100 ps delay after the reflected shock wave passed through the observation section located at distance of 3 cm from the end plate. ArF excimer laser photolysis (193 nm) was also tried. In this case, due to the more efficient photodecomposition of H2S, higher initial concentrations of H S were produced. From the difference of the decay rates of 0 atom between ArF and KrF laser photolysis, the effect of the sidereaction, 0 + HS, was examined. In the ArFlaser photolysis, the successive production of H atoms was obscured because the amount of H atoms formed by photolysis was too large; thus, the ArF laser was not used to determine the branching fraction of the reaction channel for (IC). O(3P) and H(2S) atoms weredetected by using atomic resonant absorption spectrometry (ARAS). Resonant radiation at 130.5 nm for O(3P) atoms and 121.6 nm for H(2S) atoms was observed from a microwave discharge lamp (1% 0 2 in He or 1% H2 in H e are flowed, respectively). The atomic lines were isolated by using a 20 cm vacuum ultraviolet monochromater and then sensed by a solar-blind photomultiplier. Experiments were carried out on H2S/S02 mixtures highly diluted in Ar (H2S/S02 = 23-58/106-358 ppm). The initial concentration of 0 atoms was always kept low enough so that pseudo-first-order kinetic conditions for the decay of 0 atoms, i.e., (0)0/(H2S) 1800 K). These calibration curves were based on the absorption intensities for H and 0 atoms, whose concentrations were evaluated from kinetic simulations. It is desirable to examine the validity of the evaluated concentrations based on these calibration curves extrapolated to low temperature ranges. In particular, for determination of the branching fractions for (IC) from the measurement of the concentrations of 0 and H atoms the reliability of the calibration curves is essential. Thus, these calibration curves at low-temperature ranges were directly examined in this study by conducting ArF laser photolysis/ARAS measurements in S02/H2 mixtures. In such mixtures, an oxygen atom produced in the photolysis was successively converted into two H atoms through the reactions,

+ H2OH + H, 0

+

+H H 2 0+ H

OH

(i) (ii)

If [H2] >> [initially produced 0 atom], the H atom signal increased according to the reactions (i) and (ii) but soon arrived at a steady level. Side reactions other than (i) and (ii) were not sensitive to the consumption of 0 or production of H atoms; thus, ARAS intensities could bedirectly converted to the relative concentrations for these atoms at arbitrary temperatures. H2S (Takachiho, 99.5%) and SO2 (Takachiho, 99.0%) were used without further purification. Ar (Nihon Sanso, 99.9999%) was passed through a cold trap kept at -120 OC before use. The temperature range covered was 1100-2000 K. As the reaction channel of producing H atoms was found to be minor, ARAS signals for H Lyman-a (1 21.6 nm) had to be averaged over 4-6 times in order to improve the signal to noise ratio at the same shock wave condition.

-0.41

-0.8 -

'M 1 -200

o0

200

, 400

600 6

Time I p

Figure 1. Example of the time dependence of the 130.5 nm absorption for the KrF laser photolysis in a H2S/S02 mixture behind reflected shock wave [ 5 8 ppm H2S + 341 ppm SO2 diluted in Ar, T = 1597 K,p = 9.62

x 10'8 molecules cm-31.

Results Overall Rate Constant for the HzS + O(3P) Reaction. Absorption profiles of 0 atoms were measured over the temperature range 1100-2000 K at total densities of (0.5-1.6) X 1019molecules/cm3. In the KrF photolysis the yield of O(3P) from SO2 were estimated to be d o = 0.077 X 10-18 cmz at 1100 K and 0.53 X cm2 at 2000 K; in the ArF photolysis 3.4 X 10-18cm2 at 1100 K and 2.8 X 10-18cm2 at 2000 K. An example of the time dependence of absorption intensity at 130.5 nm is shown in Figure 1. The sudden increase of the absorption signal at the arrival of the incident and reflected shock waves was mainly due to SO2 compressed by shock waves. After 100 ps from the arrival of the reflected shock wave a t the observation port, the KrF excimer laser was fired. The absorption intensity increases abruptly due to the formation of 0 atoms by the laser photolysis. The background signal for the same sample gas under the same shock waveconditions was recorded separately, where theexcimer laser was not fired. The time profiles for the absorption of 0 atoms were obtained by simply subtracting the signal without laser photolysis from that with photolysis. The first-order rate constant for the O(3P) atom decay was determined by a leastsquares fit of the decay profile of 0 atoms. The Arrhenius plot of the decay rate for O(3P) atom parametric in the initial concentration of HIS is shown in Figure 2. To examine the effects of the consecutive reactions, the decay rate for O(3P) was measured in the initial concentration ratio of [H~S]O/[O]O = 8-16. The effect of the initial concentration of H2S is not clear in the KrF laser photolysis. However, it was found that the rate constants are 30-50% smaller than those of ArF laser photolysis, as shown in the same figure. Initial concentrations of O(3P) were (2-5) X 10'3 molecules/cm3 both in the KrF and ArF laser photolysis experiments. H(*S) production in 193 nm photolysis of H2S was found to be independent of temperature and was 8.5% yield with a fluence of 30 mJ/cm2; in contrast, it was found that the yield in 248 nm photolysis of H2S increased with elevating temperature but was limited to about 0.9% a t 1800 K with a fluence of 57 mJ/cm2. The initial amounts of H and HS produced in the 193 nm photolysis were (1-2) X lOI3 molecules/cm3, which is about 10 times larger than those produced in the 248 nm photolysis. It was also observed that the decay rate of 0 atoms increased in proportion to the laser fluence especially in ArF laser photolysis. It was also verified by using NO as an alternative precursor in the ArF photolysis that SO2 did not affect the decay rate of 0 atoms. Thus, the effect of the initial concentration of HS on the disappearance rate constants of 0 atoms was assigned as due to the side reaction of 0 + HS. Le.

~

The Journal of Physical Chemistry, Vol. 98, No. 34, 1994 8421

Reaction Mechanism of Atomic 0 with H2S

0

k, = 25x1 O"'exp(-32kJ/RT)

c~n~molecule~~s~~ \

-

Figure 2. Summaryof the experimental result on the overall rate constant for the reaction 0 + H2S products (1) measured via KrF [(0)43.5 22.9ppm ppmH2S/200ppmS02, ( 0 ) : 22.8 ppmH2S/106ppmS02, (0) H2S/358 ppm SO2, (A)58 ppm H2S/341 ppm SO21 and ArF [(X) 32.3 ppm H2S/99.2 ppm s 0 2 / 3 0 mJ/cm2 of the fluence, (+) 48.5 ppm H2S/ 193 ppm S02/ 18 mJ/cm2 of the fluence] laser photolysis behind reflected shockwaves.Thestraight lineis the kl valuederived from theexperimental data using the mechanism of Table 1.

0 + HS

-

products

(2)

The rateconstants for (1) and (2) weredetermined by adjusting kl and k2 in the reaction mechanism in Table 1 until theobserved relation between the 0 atom decay rate and the initial concentration of HS is satisfied. The experimental results were expressed by the Arrhenius form as

k, = (2.5 f 0.6)

X lo-''

-1 00

0

100

200

Figure 3. Example of the time dependence of H atom concentration: (A) 58 ppm H2S without addition of SOz; (B) with addition of 341 ppm SO2 to 58 ppm HzS; (C) net ARAS signal of H atom where the contribution of the initial production of H atoms is eliminated, i.e., [B -A]. The signals are averaged over for five shots at the same shock wave condition (58 ppm H2S + 341 ppm SO2 in Ar, T = 1610 k 20 K, p = 9.75 X 10l8 molecules ~ m - (~0 ), o = 4.6 X loL3molecules cm3). Solid lines in B and C: kl,/kl = 0.2 in the reaction mechanism of Table 1. Dotted lines: kl,/kl = 0.3. Dashed4otted lines: kl,/kl = 0.1.

TABLE 1: Elementary Reactions and the Arrhenius Rate Parameters in the 0 H2S Study System [k = A P X exp(-E/RT) Units in molecule, cm, and s]

+

X

exp[-(32 f 4) kJ mol-'/RT] cm3 molecule-' s-l

over temperatures of 1100-2000 IC, where the error limits are estimates of the total experimental uncertainties. As far as we know, this is the first measurement of the rate constants for (1) at temperatures above 1000 K. Also, themagnitudeofthe reaction rate for (2) obtained in this study is consistent with the value of 2.3 X 1O-Io cm3 molecule-l s-l at room temperature assigned by Singleton et ale2' Measurementof the Branching Fraction of the HydrogenAtom Channel. The absorption profiles of H and 0 atoms were repeatedlymonitoredat 1530,1610,1630,and 1820K. Examples of the time profile of H atoms with and without coexistence of 0 atoms are shown in Figure 3. Although the absorption coefficient of H2S at 248 nm was very small, H2S was still photodecomposed so that H atoms were produced as is shown in Figure 3A, where only H2S was photolyzed (no SO2). Without oxygen atoms, the concentration of H atoms produced in the photolysis of HIS decays exponentially due to the rapid consumption by the reaction

-.

H,

+ HS

- ++ + + + + + + + - + + + + + - + + elementary reaction

k, = (1.3 f 0.5) X lo-'' cm3 molecule-' s-l

H + H,S

300

Time Ips

(3)

In contrast, a gradual increase of the concentration of H atoms (following the initial production by the photolysis) was observed when SO2 was added to the above mixture; in this case, H and 0 atoms were simultaneously produced in the photolysis of HIS and SO2 by the KrF laser. The time dependence of H atoms is shown in Figure 3B. The contribution of the initial production by the photolysis and the consecutive decay for hydrogen atom shown in this figure was canceled simply by subtraction of the ARAS signals for H atom in the H2S/Ar mixtures from those of H2S/S02/Ar mixtures after averaging over for several shots at fixed shock wave conditions. An example of such data

la. 0 + HzS IC. 0 + H2S

HS OH HSO H 2 . 0 HS-SO H 3. H2S H - H 2 HS 4. H2S OH - H S H20 5. HSO M H SO M 6.SO O H - S 0 2 + H 7. H S + H S d H 2 S S 8. H S H H2 S 9. HS + S - H S2 10.0 OH-02 + H 11.OH+Hz+H2O+H 12.0 H 2 4 O H H

+ +

+

E A

E 0 0 0 0 0

(kJ/mol)

32.0 2.00 X 32.0 0.50 X 0.0 1.30 X 20.7 3.20 X lo-" 5.25 X 0.0 1.40 X l e 8 0 245.0 8.60 X lo-" 0 0.0 1.30 X lo-." 0 0.0 0.0 3.30 X 0 0.0 4.50 X 10-" 0 0.25 7.49 X -0.5 3.6OX1&I1 0 21.5 3.00 X 1.0 37.2

ref this study this study thisstudy 22 23

estimate 24 25 26 27 28 29 29

processing is shown in Figure 3C. As shown in this figure, an increase of H atoms a t the initial stage was clearly demonstrated indicating that H atoms were formed directly by the reaction of 0 atom with H2S (1). By taking account of the ab initio calculation,~g(IC)was assumed to be the initial source for emitting H atoms in this study. This process is followed by the successive reactions listed in Table 1. Some of the successive reactions were found to have sensitivities to the initial rise of H atoms, but these contributions are not significant. The observed decay part was well explained by the reaction of H + HIS (3) whose rate constant has been recently revised.22 The branching fraction for (IC)was determined by fitting the calculated profile with observed one, where the calibration measurement of H and 0 atoms in S02/H2/Ar mixtures supplies direct information regarding the relative sensitivities for the initial amount of 0 atoms prepared in the photolysis and that for H atoms produced in the successive reactions. An example of the kinetic simulation compared with the observed profile of H atom is shown by the solid curve in the same figure. The branching fraction for (IC) was determined to be kl,/kl = 0.2 f 0.1 over thetemperatures T = 1520,1610,1630,and 1820K. Noobvious

8422

The Journal of Physical Chemistry, Vol. 98, No. 34, 1994

Tsuchiya et al.

+

TABLE 2 Vibrational Analysis of the Transition Structure for O(3P) HS Reaction System [TSa, Transition Structure for (la); TSs, Transition Structure for (IC), UHF/6-31GS* at the UHF/6-31G** Geometries] vibrational frequency/cm-' rotational constant/GHz ZPE/U mol-1 TSa TSs HzS HSO SH OH

391 29 1 1341 838 2879 4054

473 569 2896 1178

482 672

2909 2816

1096 1216

2887 2901

3150i 1092i

291.81 287.65 327.62 306.75

6.09 14.63 264.50 19.37 29 1.99 584.57

0 0 -lOF\\;Y'

temperature dependence for the branching fraction was indicated. As has been suggested in the previous studies at room temperature? the fraction for producing HSO + H in (1) was confirmed to be minor.

I

'

'

I

"

5.96 13.92 146.35 18.22 291.99 584.57

"

I

'

I

'

'

"

32

34 43 29 17 24 I

"

"

I

,j

Discussion Ab initio calculations for the reaction of O(3P) + HIS were performed using GAUSSIAN 9230 to evaluate the magnitude of kl and kl,/kl. All structures of the reactant (HzS), the products (HS, HSO) and the transition states were determined by the analytical gradient method at the HF/6-31G** level. The transition states located were verified by the vibrational analysis at the same level, and the results are listed in Table 2. The transition statestructures were thesameas thosefound by Takane and Fueno.19 The vibrational frequencies and the rotational constants of the transition state of channel ( l a ) were suggested to be generally smaller than those of ( IC);however, the imaginary vibrational frequency for ( l a ) was found to be much larger than that for (IC) [3147 and 1091 cm-I, respectively]. At the MP2/ 6-311G (3df, 3pd) level by using the HF/6-31G** structures, the barrier heights for ( l a ) and (IC)were calculated to be 35 and 39 kJ/mol, respectively, including zero-point energy (ZPE) correction. By using this information, the rate constants for ( l a ) and (IC)werecalculated with conventional transition state theory with the Wigner's tunneling correction. It is found that the calculation with the energies for the transition states indicated above gives a branching fraction kl,/ kl in agreement with the observed one, but the magnitude of the overall rate constant kl is much smaller than that measured in the previous experiments at low temperatures; also, the temperature dependence of the overall rate constant evaluated by the TST calculation is found to be much larger than the observed ones. Thus, calculated barrier heights seem to be overestimated by several kJ/mol. The rate constants were evaluated also by lowering the energies of these transition states but keeping the calculated structures unchanged. Here, Wigner's tunneling correction is also included in the present calculation. The calculated rate constant kl is compared with the experimental data in Figure 4. Also, the calculated branching fraction kl,/kl is compared with the present experimental result in Figure 5 . In these figures, the energies of the transition states for the channels ( l a ) and (IC) are assumed as 24 and 25 kJ/mol, respectively, including ZPE correction. Three groups have measured the temperature dependence of the rate constant for (1) up to 500 K.*-"J Although a slight difference of the magnitudes of the rate constants has been reported among these measurements, all these studies give almost the same activation energies of about 17 kJ/ mol. In contrast, the present study gives an activation energy of 32 kJ/mol for the temperature range 1100-2000 K, which is about twice as large as that a t low-temperature ranges. Combination of all these experimental data indicates the nonlinear Arrhenius temperature dependence for (1) as can be seen in Figure 4. The present TST calculation is found to agree with the experimental results on kl for the whole temperature range. Such a skewed Arrhenius plot is probably due to the tunneling effect at low temperatures, as well as the contribution of the vibrational partition functions of the transition state in the high-temperature range. The agreement of the branching fraction determined in

.

-l&

'

'

1

"

2

'.'

. '.I

3

'

I

.

'

'

I

4

,'

T" I 103K" Figure 4. Comparison of the TST calculation with experimentalresults on the overall rate constant kl. [ ( O )Singletonet al.,1° ( 0 ) Slagle et a1.: (0)Whytok et a1.: (A) Hollinden et a1.,6 (+) Liuti et al.,s (X) Cuppit et al.'] Curves are the result of the TST calculation. (A) &(la, IC) = 24, 25 kJ/mol; (B) &(la, IC) = 24, 25 kJ/mol without the Wigner's tunneling correction. (C) &(la, IC) = 35, 39 U/mol. (D) Eo(la, IC) = 35, 39 kJ/mol without the Wigner's tunneling correction.

t Y'

0.6

B l . ~ Z ............................... .

1

2

T.' I

3

I

4

1 0 3 ~ "

Figure 5. Comparison of the TST calculation with experimentalresults on the branching fraction kl,/kl. Curves are the result of the TST calculation. Legends for curves are given in Figure 4.

this study with the TST calculation is excellent regardless of such adjustment. The result of the branching fraction for (IC) is also consistent with the conclusion given by Singleton,et a1.,4 who suggested that kl,/kl < 0.2 at room temperature. Even though the branching fraction for producing H atoms is concluded to be minor, the present study indicates a substantial contribution of this process to (1) for the first time. The concentration of OH was not monitored in this study; however, it may be reasonable toconclude that the direct abstraction channel ( l a ) is the major process for (1).

Acknowledgment. Wewish to thank Professor T. Fueno (Osaka University) for many helpful discussions. References and Notes (1) Bernez-Cambot, J.; Vovelle, C.; Delbourgo, R. 18rh Symp. Combusr. 1981, 777.

(hi.)

Reaction Mechanism of Atomic 0 with H2S (21 Bradlev. J. N.: Dobson. D. C. J. Chem. Phvs. 1976. 46. 2865. (3j Frenklkh,M.;Lee, J. H:; White, J. N.;Gardker, W.C., J;. Combust. Flame 1981, 41, 1. (4) Muller, C. H., 111; Schofield, K.; Steinberg, M.; Broida, H. P. 17th Symp. (Inr) Comb. 1978, 867. (5) Liuti, G.; Dondes, S.; Hartek, P. J. Am. Chem. SOC.1966,88,3212. (6) Hollinden, G. A.; Kurylo, M. J.; Timmons, R. B. J . Phys. Chem. 1970, 74, 988. (7) Cupitt, L. T.; Glass, G. P. Trans. Faraday SOC.1970,66, 3007. (8) Whytock, D. A.; Timmons, R. B.; Lee, J. H.; Michael, J. V.; Payne, W. A.; Stief, L. J. J . Chem. Phys. 1976, 65, 2052. (9) Slagle, I. R.; Balocchi, F.; Gutman, D. J . Phys. Chem. 1978, 82, 1333. (10) Singleton, D. L.; Irwin, R. S.; Nip, W. S.; Cvetanovic, R. J. J. Phys. Chem. 1979. 83. 2195. (11) Singleton, D. L.; Paraskevopoulos, G.; Irwin, R. S. J . Phys. Chem. 1982, 86, 2605. (12) Smardzewski, R. R.; Lin, M. C. J. Chem. Phys. 1977,66, 3197. (13) Slagle, I. R.; Graham, R. E.; Gutman, D. Int. J. Chem. Kinet. 1976, 8, 451. (14) Cvetanovic, R.J. J . Chem. Phys. 1959,30,19, Can. J. Chem. 1960, 38, 1678. (15) Huie, R. E.; Herron, J. T. Progress in Reaction Kinetics; Jennings, K. R., Kundal, R. B., Eds.; Pergammon Press: Oxford, 1975; Vol. 8, part 1. (16) Clemo, A. R.;Davidson, F. E.; Duncan, G. L.; Grice, R. Chem. Phys. Lett. 1981, 84, 509. (17) Davidson, F. E.;Clemo, A. R.;Duncan,G. L.; Browett,R. J.; Hobson, J. H.; Grice, R. Mol. Phys. 1982, 46, 33.

The Journal of Physical Chemistry, Vol. 98, No. 34, 1994 8423 (18) Balucani, N.; Beneventi, L.; Casavecchia, P.; Stranges, D.; Volpi, G. G. J. Chem. Phys. 1991, 94, 8611. (19) Takane, S.; Fueno, T. Bull. Chem. SOC.Jpn. 1993.66, 3633. (20) Koshi, M.; Yoshimura, M.; Fukuda, K.; Matsui, H.; Saito, K.; Watanabe, M.;Imamura, A.; Chen, C. J. Chem. Phys. 1990, 93, 8703. (21) Singleton, D. L.; Cvetanovic, R. J. J. Phys. Chem. Ref. Data 1988, 17, 1377. (22) Yoshimura, M.; Koshi, M.; Matsui, H. Chem. Phys. Letr. 1992,89, 199. (23) Perry, R. A.; Atokinson, R.; Pitts, J. N., Jr. J. Chem. Phys. 1976, 64, 3237. (24) Jourdain, J. L.; Le Bras, G.; Combourieu, J. Int. J. Chem. Kinet. 1979, 11, 569. (25) Bradley, J. N.; Trueman, S. P.; Whytock, D. A,; Zaleski, T. A. J . Chem. SOC.,Faraday Trans. 1 1973, 69,416. (26) Cupitt, L. T., Glass, G. P. Int. J. Chem. Kinet. Symp. 1976, I , 39. (27) Schofield, K. J. Chem. Phys. Ref. Data 1973, 2, 25. (28) Cohen, N.; Westberg, K.R. J. Phys. Chem. ReJ Data 1983,12,531. (29) Baulch, D. L.; Drysdale, D. D.; Horne,D. G.;Lloyd, A. C. Eualuated Kinetic Data for High Temperature Reactions; Butterworth: London, 1976 Vol. 1. (30) Frisch, M. J.; Trucks, G. W.; Head-Gordon, M.; Gill, P. W. L.; Wong, M. W.; Foresman, J. B.; Johnson, B.G.;Schlegel, H. B.; Robb, M.A.; Replogle, E. S.; Gomperts, R.; Andres, J. L.; Raghavachari, K.; Binkley, J. S.;Gonzales,

C.; Martin, R. L.; Fox, D. J.; Defrees, D. J.; Baker, J.; Stewart, J. J. P.; Pople, J. A. Gaussian, Inc., Pittsburg, PA, 1992.