155
J . Phys. Chem. 1994, 98, 155-169
Kinetics and Mechanism of the C10 + CIO Reaction: Pressure and Temperature Dependences of the Bimolecular and Termolecular Channels and Thermal Decomposition of Chlorine Peroxide Scott L. Nickolaisen,t Randall R. Friedl, and Stanley P. Sander. Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91 I09 Received: July 6, 1993; In Final Form: October 18, 1993'
+
The kinetics and mechanism of the C10 C10 reaction and the thermal decomposition of ClOOCl were studied using the flash photolysis/long path ultraviolet absorption technique. Pressure and temperature dependences were determined for the rate coefficients for the bimolecular and termolecular reaction channels and for the thermal decomposition of ClOOCl. In order to determine channel-specific rate coefficients and to minimize complications associated with secondary chemistry, the reaction was studied over wide ranges of initial reactant stoichiometry and temperature. The rate coefficient for the termolecular association channel in the lowpressure limit, C10 C10 (+M) ClOOCl (+M) ( l ) , with N2 as a third body was measured over the temperaturerange 195-390Kandresultedinkl,~,(T)=(1.22*0.15) X 1&33exp{(833f 3 4 ) / T J ~ m ~ m o l e c u l e - ~ s-1 ( f 2 a error bounds). The 300 K rate coefficient for reaction 1 was measured for a number of bath gases. The results are k l , (X10-32 ~ cm6 molecule-2 s-l) = 0.99 f 0.05, 1.24 f 0.09,1.71 f 0.06, 2.00 f 0.27, 2.60 f 0.17, 3.15 f 0.14, and 6.'/ f 3.6 for He, 0 2 , Ar, N2, CF4, SF6, and C4, respectively. The effective collision efficiency for M = C12 is very large and is likely due to a chaperone mechanism. Below 250 K, the reaction was in the falloff regime between second- and third-order kinetics. From the falloff data, the rate constant in the high-pressure limit, k,3w, was estimated to be (6 f 2) X cm3 molecule-' s-'. The Arrhenius expressions for the three bimolecular channels, C10 C10 Cl2 + 0 2 (2), ClOO C1(3), and OClO C1 (4), over the temperature range 260-390 K are kz(7') = (1.01 f 0.12) X exp(-(1590 f lOO)/TJ cm3 molecule-' s-1, k3(7') = (2.98 f 0.68) X 10-11 exp{-(2450 f 330)/Tj cm3 molecule-' s-l, and k4(T) = (3.50 f 0.31) X 10-13 exp{-(1370 f 150)/Tj cm3 molecule-' s-l. These expressions lead to a value of (1.64 f 0.35) X lO-14cm3 molecule-1 s-1 for the overall bimolecular rate constant (k2 k3 k4) at 298 K. The rate coefficient expression for ClOOCl thermal decomposition was determined to be k-l( 7') = (9.81 f 1.32) X exp{-(7980 f 320)/Tj cm3 molecule-' s-1 over the range 260-310 K. From a Third Law analysis using equilibrium constants derived from measured values of kland k-1, the enthalpy of formation (AHOr(298)) of ClOOCl was determined to be 30.5 f 0.7 kcal mol-'. The equilibrium constant expression from this analysis is K,(T) = (1.24 f 0.18) X 10-27 exp((8820 f 440)/Tj cm3 molecule-'. From the observed activation energy for reaction 4 and the literature activation energy for reaction -4, the OClO enthalpy of formation was calculated to be 22.6 f 0.3 kcal mol-'.
-
+
-
+
+
+
+ +
Introduction
c10 + c10
In the past few years, there has been considerable interest in reactions that proceed over complex potential energy surfaces. Because long-livedcollision complexes may form in these reactions, the experimental rate coefficients may show significant pressure dependences, negative temperaturedependences(which vary with pressure), and product distributions which depend on the pressure and temperature. The self-reactionsof halogen monoxide radicals are interesting examples of reactions which proceed via collision complexes. Of the 10 reactions of the type
XO + XO XO + YO
-
-
products
t NASA/NRC
+
+
+
M
ClOOCl
ClOO
(3)
c1+OClO
(4)
is perhaps the most complex because of the multiplicity of channels leading to stable products. The association (termolecular) channels form ClOOCl and ClOClO via surfaceswith essentially no potential barrier. The corresponding bimolecular product channels proceed through loose transition states to give CI ClOO and C1+ OClO across surfaceswith a significant potential barrier. Finally, the molecular elimination channel produces Clz('2,3II) 0 2 through a tight transition state originating in the ClOOCl intermediate. Because of the complexity of this system, the experimental description of the pressure and temperature dependences of the product branching ratios is far from complete. The C10 C10 reaction also plays an important role in atmosphericchemistry. C1 atoms, which are produced from the photochemical degradation of chlorofluorocarbons, react with 0,to produce C10 radicals from the reaction
+
+
products
where X,Y = F, C1, Br, I, there are experimental data on five (FO FO, C10 CIO, C10 + BrO, BrO BrO, and IO IO). Of these, the C10 C10 reaction,
+
-c1+ -
+
Resident Research Associate.
* To whom all correspondence should be addressed.
Abstract published in Aduance ACS Absrracrs, December 1, 1993.
0022-365419412098-0155$04.50/0
+
c1+0,- c10 + 0 2 Under normal stratospheric conditions, the self-reaction of c10 @ 1994 American Chemical Society
156 The Journal of Physical Chemistry, Vol. 98, No. 1 , 1994
+ 0 2 ~hanne1.I~There are no time-resolved studies of the branching ratios in M = N2 with which to compare these results. Equilibrium constants for the reaction
radicals cannot compete with reactions such as
0 + c10 and
+
+
-
Cl + 0,
ClOOCl
+ - c1+ c1+ - + M
ClOOCl
hv
ClOO
M
ClOO 2
0,
2(C1+ 0, C10 net:
+
+
2 c 1 0 ClOOCl have been measured by Basco and Hunt6 and Cox and Derwentls at room temperature and by Cox and Hayman20 over the temperature range 233-303 K. These results are in reasonable agreement and lead to a calculated 0-0 bond dissociation energy of 17.3 kcal mol-' assuming ClOOCl to be the predominant form of the C10 dimer. In this paper we present the results of an extensive study of the C10 + C10 reaction with the goal of characterizing the rates and mechanism of both the bimolecular and termolecularchannels over a wide range of temperature and pressure using several different collision partners. Evidenceis presented for the existence of a long-lived ClO-Cl2 complex which greatly enhances the apparent rate of the recombination channel. Results are also presented on the rate of thermal decomposition of ClOOCl. Experimental Section
0,)
203-30,
Since the C10 + C10 reaction is partially rate determining in this cycle (the photolysis of ClOOCl also plays a role), it is important to determine the temperature and pressure dependences of the rate coefficients under atmospheric conditions. During the day, photolysis of ClOOCl is much faster than unimolecular decomposition. Under nighttime conditions, the removal of ClOOCl is most likely dominated by thermal decomposition, and the rate ofthis processmustbeknown tocalculate the partitioningbetween chlorine-containing species. Because thermally stabilized ClOOCl has at least three unimolecular decomposition channels (C10 C10, C1+ C100, C12 02), and these channels have different atmospheric consequences under nighttime conditions, it is also important to establish the thermal decompositionbranchingratios under atmospheric conditions. The C10 C10 reaction has been studied extensively by direct and indirect methods for several decades. Early work on the photolysis of Cl2-03 mixturesl.2 suggested the importance of the C10 C10 reaction in propagating the chain destruction of 03. The first direct studies of reaction 1 did not come about until the development of the flash photolysis technique and the subsequent identification of the ultraviolet absorption spectrum of ClO.394 Since these early studies, rate coefficients for reaction 1 have been measured using several kinetic techniques including flash photolysis-ultraviolet absorption,5-8 discharge flow-ultraviolet absorption,"' discharge flow-mass spectrometry,'z and molecular modulation-ultraviolet absorption.l3-1* Rate coefficients for the termolecular component have been obtained by Basco and Hunt,6 Hayman et al.,17 Sander et aL,7 and Trolier et a1.8 The latter two studies were carried out in the temperature range relevant to the polar stratosphere and are in reasonably good agreement, although Trolier et al. observed nonzero intercepts in the rate coefficient falloff curves which were not observed by Sander et al. By comparison, the bimolecular channels are poorly understood. Recent results for the overall bimolecular rate coefficient at 298 K (k2 + k3 + k4) have ranged over a factor of about 4, and there is only one study of the temperaturedependence of this parameter. There are no temperature dependence studies of the branching ratios for the bimolecular channels. Steady-state photolysis studies of the chlorine-photosensitizeddecomposition of ozone have consistently given values of about 6 for the quantum yield for ozonedestructioninN2 bath gas at 298 K, implying a branching ratio of about 2 for the ClOO + C1 channel relative to the C12
+
M
+
BrO C10 Br ClOO as an important rate-limiting step in the catalytic destruction of 03. In the highly perturbed conditions present in the polar stratosphere, however, where catalytic processes on the surfaces of polar stratosphericcloud particles can result in the repartitioning of inorganic chlorine into predominantly active forms, the following cycle may play an important role in 0 3 destruction:
c10 + c10
Nickolaisen et al.
+
The experimentsdescribed in this paper were carried out using the flash photolysis/ultraviolet absorption technique. The apparatus has been described in detail previously.21The flash lamp/ reactor is a unit consisting of four concentric Pyrex tubes, approximately 1 m long, comprisingthe reaction cell (1-in. i.d,), photolyzing light filter, xenon flash lamp, and cooling/heating jacket. The cell was operated in the continuouslyflowing mode, with all reagent and carrier gas flows being measured with calibrated mass flow meters. The analytical light source was a 150-W xenon arc lamp which was collimated and coupled into the reaction cell through eight-pass White-type optics with external mirrors. The optical path length was 720 cm. The exit beam was transferred via a switching mirror to a 0.32-m-focal length spectrograph (1 50-rm slit width, 0.18-nm resolution) equipped with a 1024-channel optical multichannel analyzer (OMA) for the detection of C10 and OClO or to a 0.5-m monochromator (150-Fm slit width, 0.13-nm resolution) and photomultiplier for the detection of C10 and C120. The photomultiplier output was amplified, low pass filtered, and digitized with a signal averager interfaced to a microcomputer. In experiments which used the photomultiplier, C10 radicals were monitored by their absorption at the peak of the 12-0 f l = 3/2 subband (A X)at 275.5 nm. In the OMA experiments, five vibrational bands (10-0,11-0,12-0,13-0, and 14-0) were monitored over the wavelength range 270-280 nm. OClO was monitored using the OMA over the wavelength range 340400 nm which encompassed seven vibrational bands [a(6), a(7), a(8), a(g),a(lO),a(l l),anda(l2)]. TheOMAwasnormallyoperated in the time-resolved modes2' In this mode, up to 500 sweeps could be recorded per flash with each sweep consisting of a 1024channel spectrum with a minimum 16-ms averaging time. In a few experiments, the OMA time resolution was increased by operating the OMA with a reduced number of channels (200 channels, 4-ms resolution), which also reduced the spectral coverage. OClO and C10 spectra from the OMA were processed by first coadding and converting to absorbance the spectra from 5 to 20 successive flashes. Spectra were convolved with a sevenpoint triangular kernel to match the exit slit function of the spectrograph. Because of the formation and removal of species such as Cl2Oand ClOOC1, which have slowly varying continuum absorptions which overlapthe species of interest, it was necessary to eliminatetime-varying base line offsets. This was accomplished by subtracting from the smoothed spectrum a third-order polynomial obtained from a curve fit to the spectrum. Absolute C10 and OClO concentrations were obtained by fitting each observed spectrum to calibrated literature spectra using the
-
Kinetics of the C10
+ C10 Reaction
The Journal of Physical Chemistry, Vol. 98, No. 1, 1994 157
conjugate gradient method. For these experiments, the OMA was preferred over the PMT because of several important advantages. Because both C10 and OClO have absorption spectra with distinguishing vibrational band features, retrieval of these features provided a means to discriminate against time-varying base line shifts from continuum absorbers. This was particularly important for the high-temperatureexperimentswith excess C120 because of the effect of chain C120 removal on theC10 absorption base line. The second benefit stems from the OMA multiplex advantage which arises from the use of 1024 separate detectors in the focal plane. This results in a signal-to-noise ratio improvement of more than a factor of lOover the photomultiplier. The detection limit for both C10 and OClO using the OMA was about 2 X 1O1Omolecules cm-3. Experimental signalswere convertedto absorbanceunits using Beer's law,
the case of OC10, the expression was ~ w l o ( T=' ) ~wIo(296)(1.733- 0.002487')
where uo~lo(296)= 1.275 X l e 1 ' cm2 molecule-I is the cross section for the a( 11) line. This expression was derived by fitting the a(l1) cross sections reported in ref 26 versus temperature to a straight line. The mechanism employed in the study of the C10 + C10 reaction used the photolysis of ClrCl20 mixtures at wavelengths longer than 300 nm (Pyrex cutoff). In this system, C10 radicals were formed by the reactions C1,
+ hu
c1+C1,O
--*
2C1
-c10 + c1,
(5)
and to a minor extent from the photolysis of Cl20,
A = Nul = -ln(Z/Zo) where N is the species concentration, u is the absorption cross section, and 1 is the path length. IO was determined from the preflash signals for PMT and OMA data recorded at higher temperatures ( T 2 250 K). The signal as t was used as IO for low-temperature (T < 250 K) PMT data. For PMT data collected at 275.5 nm, four species (ClO, Cl20, ClOOCl, C12) have significant absorptions with the following absorption cross sections at 298 K: UCIO = 8.4 X 10-18cm2,21 UCI,O = 1.24 X cm2,22ucl-l= 2.45 X 10-18 cm2,23and gcll = 2.19 X 10-20 cm2.24 The measured absorption cross section of C10 shows a temperature dependence;21however, thevariation of u~10overthe temperature range of the excess chlorine atom experiments (260 K IT I 400 K) is less than 7% so no correction was made for this variation in the fitting routine. At 195 K, the lowest temperature of the excess Cl20 PMT data, the C10 cross section is extrapolated to be 10.6 X 10-18 cm2. OMA reference spectra were obtainedby madding 1000spectra collected under conditions which maximized either the C10 or OClO signals. For C10, 100 spectra at a time resolution of 4 ms were collected immediately following the flash when the C10 concentration was greatest. This cycle was repeated 10 times to achieve a total of 1000 averaged spectra. The absolute C10 concentration of the resulting spectrum was calculated by comparingthe absorbance differencebetween successivemaxima and minima of the absorptionspectrum to the relative cross section differences of the same maxima and minima determined previously.21 By using relative cross sections between maxima and minima of the absorption spectrum instead of absolute cross sections, possible errors caused by background absorption from Cl20, C12, or ClOOCl were minimized. For OC10,lOOO spectra were collected at a time resolution of 20 ms beginning 5 s following the flash. This allowed OClO formation to reach a maximum. This reference spectrum was converted to absolute concentration in the same manner as described for C10 by comparison of the relative absorbances to the OClO cross sections measured by Wahner et al.25 Because of the improved signal to noise of the OMA data which allowed the rate coefficients to be determined more precisely, the temperature dependences of the C10 and OClO absorption cross sections were used when calculating the absolute concentration of the reference spectrum. For C10, the expression used tocalculate the cross section at a given temperature was
-
uclo(T) = uc10(298){1.01 1 - 104.9/7'+ 3 0 3 3 0 / p ] where aclo(298) = 8.4 X 10-18cm2molecule-1 is the cross section at 275.5 nm. This expression was derived by fitting the C10 cross sections reported by Sander and Fried121 versus 1 / T to a quadratic equation. The use of a quadratic equation dependent on inverse temperature was somewhat arbitrary and was chosen because it provided the best visual interpolation of the data. In
C1,O
+ hv
+
C1+ C10
As discussed in the Results section, conditions for reaction 5 were employed such that either C1 or Cl20 was present in excess. In either case, the production of C10 typically took place on a time scale 3-4 orders of magnitude faster than its removal. The exception to this condition occurred when 0 2 was used as the buffer gas. In this case, C10 was also formed by the reactions
c1+0,
M
c1+ClOO
ClOO
-
2c10
which occurredon a longer time scale. This formationmechanism will be discussed below. C120was prepared by the method of cad^.^^ Residualchlorine from the synthesiswas removed by distillation at -1 12 OC. C120 was introduced into the flash photolysis cell by flowing helium at 5 psia through a bubbler containing the pure liquid at -78 OC. Ultrahigh-purity chlorine, helium, argon, nitrogen, oxygen, CF,, and SFs were used as received.
Results The primary objectives of this work were to measure the rate coefficients for all of the known bimolecular and ttrmolecular channels of the C10 + C10 reaction and to measure rate coefficients for the unimolecular decomposition of ClOOCl to 2C10 over as wide a range of pressureand temperature as possible. The major difficulty in determining channel-specific rate parameters for the C10 self-reaction arises from the rapid regeneration of C10 from C1 and ClOO radicals produced in reactions 3 and 4. In addition, at temperatures above about 250 K, the thermal decomposition of ClOOCl back to C10 plays a major role in determiningthe C10 time dependence. These regeneration mechanisms complicate the kinetic analysis of both the primary C10 decay and the formation of OClO and have been a major source of uncertainty in previous studies. Mechanism. The complete set of reactions considered in the kinetic analysis is given in Table 1. For completeness, the mechanism explicitly considersall the known reactions that involve C10 formation and removal, reactions of C100, OC10, and ClOOCl, and Cl termination. However, not all these reactions are important under all conditions. The stoichiometric and temperature regimes in which individual reactions play an important role are indicated in Table 1. Most of the rate coefficientswere taken from the NASA kinetics data evaluation26 including the expressions for k-1 and k-7 which were derived by combining the tabulated temperature-dependent expressions for the reverse reactions and equilibrium constants. Flow through the cell was also included in the mechanism and was modeled by
158 The Journal of Physical Chemistry, Vol. 98, No. 1. 1994
Nickolaisen et al.
TABLE 1: Summarv of Reactions Used in Fittine of Data to the Reaction Mechanism ~~
~
~~
experimentalconditions
+ - + - ++ + + - c10 +
excess ClzO (195-250 K)
excess C1 (250-400 K)
excess ClzO (250-400 K)
8
8
8
ki' k-1 = 6.0 X 1V (T/300)-3.9exp(-8450/'I)* kz' k3' kL3 2.3 X 1 6 " k't3 1.2 X 161' b
8
8
8
0 0 0 0 0
ha
0
8 8 8 8 8 8 8 8
8 8 0 0 0 8
0
reaction
rate coefficient
c1+ClZO c10 + c12 CIO + c10 2 ClooCl termolecular channel: C l o o C l E c10 c10 bimolecular channels: c10 + CIO 4 CI2 02 c10 + c10 C l o o + CI ClOO + c1- Clz + 02 initiation:
c10 c10
CIO c10 OClO c1 OClO CI CIO ClOOCl CI ClOO + c12
+
secondary chemistry:
+
c1+0 2 E ClOO C l o o E CI + 02 CI + CIO E Cl20
+
CI CI E Cl2 CIO, CIOOC1, etc.
mass transport:
-
leave cell
ks
9.8 X 1 6 1 1 b
*
k 4 = 3.4 X 1 6 I l exp(l60/q* kg = 1.0 x 1 6 "
0 0
k7 = 2.7 X 10-33(T/300}-'.5b
0
8
0
k-7 = 4.7 X
0
8
0
ka=6X1632d
0
8
0
k9 = 7.5 X exp(906/7 Tflow = 0.01-0.1
0
8
0 8
T/300)-1,5exp(-2500/ T j b
0
0
0 Determined in fitting procedure. Estimated by combining ko from ref 7 with the equilibrium constant from ref 26. e From ref 36. Estimated using Troe's method (see text). 8 indicates reaction used in fitting routine; 0 indicates reaction not used in fitting routine.
TABLE 2
Summary of Conditions for Each Set of Experiments and Absorption Cross Sections' [ClZOl (1.0-1.5) x 1015 (6-x81013 ) (1.0-1.5) X loi5
excess C120 (195-250 K) excess C1(250-400 K) excess C120 (250-400 K)
[Cll (2.0-5.0) x 1013 (2-7) x 1014 (4.0-5.0) x 1013
[MI (1-30) X 10'' (0.5-6) X 1018 (0.5-22) X 10''
AbsorDtion Cross Section at 275.5 nm 1298 K) UClr = 2.19x m i 0 UCIO(T) ~ ~ ~ 0 ( 2 9 8 ) ( 1 . 0 1104.9/T+ 130330/P}
8.4 X 16'' uc1p = 1.24 X 1 6 " U C I ~=~2.45 ~ I X 16'* UCIO
a
[Chi (2.0-5.0) X 10l5 (5-18) X 10l6 (1.0-1.2) x 1016
Concentrations in cm3 molecule-l and cross sections in cm2 molecule-1.
a single-exponential function with a time constant, qOw. The value of q, was in the range 0.1-0.01 s-I. The effect of flow on the observed signals was minimal as the residence time of reactants and products in the cell was adjusted to be 5-20 times longer than the time intervalover which useful data were acquired. For the experimentswhich employed excess C1, it was necessary to estimate the time dependence of the chlorine atom concentration. C1 loss can occur by the direct recombination reaction, M
c1+c1- c1,
(9)
and also by ClO catalysis,
M
c1+c10 c1,o c1+C1,O
-c1, + c10
(8) (5)
M
net:
C1+ C1- C1,
For the latter mechanism, reaction 8 is the rate-limiting step. Since the Cl C10 recombination reaction has not been studied experimentally, an estimate of the low-pressure limiting rate constant was made using the method developed by Troe.28 A value of 6 X lo-', cm6 molecule-2 s-1 was estimated at 298 K. In order to minimize the complications arising from secondary chemistry, the reaction was studied in three different regimes of initial reactant stoichiometry (C120:Cl) and temperature. These regimes were (1) excess Cl20, T I 250 K,(2) excess C1,250 K IT I400 K,and (3) excess C120,250 K I T I 400 K. The ranges of reactant concentrationsfor each of these kinetic regimes are summarized in Table 2. The experimental conditions as well
+
as the mechanisms used in the analysis of data from the three regimes will be described in detail below. For some of the reaction conditionsemployed, the rateequation for the C10 time dependence did not have an analytical solution. In this case, rate coefficients were determined by obtaining the best fits between experimental absorbance data and numerical solutions of the reaction mechanism in which the rate coefficients were varied. Two separate computer codes were used to obtain the fits. One employed the conjugate gradient method and was used for very rapid evaluation of mechanisms of limited size. The second code was the Harwell program FACSIMILE29 which allowed symbolic input of the reaction mechanism at the expense of increased execution time. Excess azo, Low Temperature ( T I 250 K). At low temperature ( T I 250 K),the dominant channel for C10 reaction is ClOOCl formation. Under these conditions, the rate of the addition channel is at least a factor of 25 times faster than the sum of the bimolecular reaction rates. The rate of C10 dimer decomposition is also very slow; the l / e lifetime of the dimer is greater than 20 s at T = 250 K. By neglecting the contributions from bimolecular reactions and dimer decomposition and assuming that reaction 5 is fast compared to the rate of dimer formation, the mechanism reduces to a simple second-order loss process for C10, i.e., d[ClO]/dt = -2k,[M][ClO]' l/[ClO] = 2kl[M]t
+ l/[CIO],
(10)
(11)
The results of the experimentscamed out under these conditions were reported in a previous publication7and will not be repeated here. Due to space limitations in the earlier paper, it was not possible to give the measured values of kl in tabular form. Values
Kinetics of the C10
+ CIO Reaction
The Journal of Physical Chemistry, Vol. 98, No. 1, 1994 159
TABLE 3 Summary of Experimental Rate Coefficients, 4, as a Function of Nt Concentration under Conditions of Excess (320 and T < 250 K and Results of Fits to Troe’s Falloff Expression’ 10-’8[N2] temp (molecules 1013kl (cm3 molecule-1 s-1) (K) cm-3) 195 0.81 1.oo 0.94 1.25 1.22 1.67 1.70 2.50 1.75 2.62 1.88 3.00 4.23 2.51 2.94 5 .oo 6.94 3.72 4.80 10.0 5.59 11.3 7.77 18.2 0.67 1 .oo 208 0.96 1.62 1.58 2.65 2.28 4.27 2.98 6.98 4.72 11.4 8.22 18.4 11.2 30.0 0.48 220 1 .oo 0.75 1.61 1.29 2.65 1.36 3.00 1.90 4.28 2.72 6.91 11.4 3.87 5.85 18.4 9.64 30.0 0.43 233 1 .oo 0.68 1.62 1.01 2.65 1.58 4.28 2.45 6.98 3.49 11.4 5.05 18.4 7.96 30.1 0.3 1 247 1 .oo 0.53 1.62 0.80 2.65 1.04 3 -00 1.48 4.28 2.13 6.98 3.15 11.5 5.00 18.4 6.73 30.0 a The
Second,C1rapidly convertsall of the ClOOCl produced in reaction 1 to C12 and ClOO (which rapidly decomposes to C1+ 02).The net reaction is the C1-catalyzed recombination of CIO to C12,
-c1+ + M
iO32kOb (cm6 molecule-2 s-1) 8.8
10I2k-c (cm3 molecule-1 8 ) 3.8
c10 + c10 ClOOCl c1+ClOOCl c1, + ClOO
(1)
(6)
M
0,
ClOO net:
7
6.6
C10
C10
C1,
(-7)
+ 0,
Third, the OClO produced in reaction 4 rapidly regenerates C10 by reaction with excess CI (reaction -4). Reaction 4 therefore does not contribute to the loss of C10. With the rapid removal of C120, ClOOCl, and OClO by C1, there is no secondary regeneration of C10 on the time scale of the C10 C10 reaction. Under these conditions, the loss of C10 in the excess C1 system is given by
+
+ k, + k,)[C10I2
d[ClO]/dt = -2(k,[M]
6
5.3
(12) where C10 formation (reaction 5 ) is assumed to be instantaneous relative to C10 loss and transport out of the cell has been neglected. [ClO] should follow a second-order rate law with the effective rate coefficient being given by
,k 4.6
5
3.7
7
= k,[M]
+ k, + k,
Plots of koh vs [MI should therefore be linear with slope kl and y intercept k2 k3. When 02 was used as the bath gas, the time dependence of the C10 formation displayed a slower secondary increase in addition to the rapid rise due to reaction 5 . This secondary formation of CIO was caused by the association reaction of chlorine atoms with oxygen to form the chlorine peroxy radical followed by reaction with atomic chlorine to form C10, Le.,
+
c1+0, ClOO
M
ClOO
+ c1-
2c10
ko was determined by fitting the data at each temperature of the falloff equation, Fitting ko( T ) to the expression ko( T ) = koMO( T/300)4 yielded kow = (1.8 0.5) X 10-32 cm6 molecule-2 s-I and n = 3.6 1.O using Fc = 0.6. k, was determined by fitting the data at each temperature to the falloff equation. Fitting k,(T) to the expression k,(T) = k300 (T/300)* yielded k,m = (6.0 & 2.0) X 10-l2 om3 molecule-I s-I and m = 0.0 & 1.0 using FE = 0.6.
*
of kl at each pressure and temperature are therefore given in Table 3 as well as the parameters derived from fitting these data to the falloff expression given by Troe.28 Excess Atomic Chlorine, High Temperature (250 K ITI 400 K). A set of experiments was conducted under conditions such that photolysis of CI2 produced excess C1 over C120. To a first approximation, C1 remained in large excess over the time scale of the C10 + CIO reaction. The presence of excess C1 affects the reaction mechanism in three ways: first, CI rapidly and stoichiometrically converts all of the ClzO to C10. There can therefore be no secondary regeneration of C10 by reaction 5 .
(-3”)
The secondary C10 formation process could be readily modeled by including these reactions in the mechanism used to fit the data. Because the formation of C10 from reaction -3” was much faster than its removal by the C10 C10 reactions, there was no impact on the uncertainty of the derived rate constants. It was found that C10 temporal decay profiles could not be fit precisely by a second-order rate law. This was due to a partial breakdown of the excess chlorine atom assumption at long reaction times due to direct C1 recombination (reaction 9) and C10catalyzed recombination (reactions 8 and 5 ) . Under these conditions, additional terms must be included in the C10 rate equation accounting for regeneration from dimer decomposition (reaction -1) and the reaction of C1 with OClO (reaction -4). Mechanism simulationsindicate that thecontribution from dimer decomposition dominates especially at high temperature and pressure. (ClOOCl decomposition increases by over 4 orders of magnitude between T = 260 K and T = 400 K,26 while the rate of dimer formation decreases by a factor of only 3.) Gas flow out of the reaction cell also made a small contribution to the C10 loss rate. Because of these additional complications, the C10 rate equation does not have an analytical solution. Rate coefficients were determined using FACSIMILE to fit single-wavelength (275.5 nm) absorbancedata from thePMT to the timadependent absorbance change calculated by the model using the reaction
+
falloff equation is given by28
(7)
160 The Journal of Physical Chemistry, Vol. 98, No. 1 , 1994
mechanism in Table 1 and absorption cross sections in Table 2. Thecalculated absorbancechange at this wavelength is primarily due to C10 but includes minor (
