J. Phys. Chem. 1980, 84, 819-821
The nature of the discrepancy is such that ascribing it to an error in the ratio of aO3a t the two wavelengths would not be unreasonable if the spectrum of ozone were not so well established. Since this explanation is not reasonable, we can only attribute the discrepancy t o some unknown source of error. One further aspect of the results which warrants discussion is ths variation of [HO,], and [O(3P)]owith [O,], shown in Table 11. The increase in [O(3P)]owith increasing [02] is expected, but we are unable to present any quantitative explanation as to why [HO,], should increase with increasing [OZ]. There does not seem to be any appreciable “sink” for H atoms other than reaction with Oz under the conditions of‘these experiments. We can only suggest the possibility that a sequence of ion-molecule reactions results in a greater efficiency of utilization of energy for H (or HOJ formation as [O,] is increased.
Acknowledgment. We thank Dr. C. Jonah for the least-squares fitting program.
819
References and Notes ( I ) Work performed under the auspices of the Office of Basic Energy Sciences of thle U.S. Department of Energy.
(2) Department of Chemlstry, Malcolm X College, Chicago, IL 60612. (3) E. J. Hamilton, Jr. and R. R. Lii, Int. J. Chem. Kinet., 9, 875 (1977). (4) R. R. LII, R. A. Gorse, Jr., M. C. Sauer, Jr., and S. Gordon, J . Phys. Chem., 83, 1803 (1979). (5) R. R. Lii, R. A. Gorse, Jr., M. C. Sauer, Jr., and S. Gordon, J . Phys. Chem., preceding paper In this issue. (6)R. R. Lii, R. A. Gorse, Jr., M. C. Sauer, Jr., and S. Gordon, J. Phys. Chem., following paper in this issue. (7)J. P. Burrows, G.W. Harris, and 8. A. Thrush, Nature (London),267,
233 (1977). (8) S.Gordon, W. Mubc, arid P. Nangia, J. phys. Chem., 75, 2087 11971). (9) R. F. Hampsori, Jr. and D. Garvin, Eds., Nat. Bur. Stand. (U.S.), Spec. Pub/., No. 513 (1978). (IO) T. T. Paukert and H. S. Johnston, J. Chem. Phys., 56, 2824 (1972). (1 1) Based on the value at 230 nm from ref 10 and relative values from C. J. Hochanaclel, J. A. Ghormley, and P. J. Ogren, J. Chem. Phys., 56, 4426 (1972). (12) M. Griggs, J. IChem. Phys., 49,857 (1968). (13)J. G. Calvert and J. N. Pitts, Jr., “Photochemistry”, Wiley, New York, 1967,p 201. (14) A. R. Ravishankara, P. H. Wine, and A. 0.Langford, J. Chem. Phys., 70, 984 (1979).
Rate Constant for the Reaction of QH with H02’ Ruey-Rong Lii,’ Robert A. Gorse, Jr,,3 Myran C. Sauer, Jr.,” and !Sheffield Gordon Chemistry Division, Argonne National Laboratory, Argonne, Illinois 60439 (Received October 4, 1979) Publication costs assisted by Argonne National Laboratory
-
Spectrophotometric observation of both HOz and OH in pulse-irradiated Ar-Hz0-02 systems leads to a value cm3 molecule-I s-l at 308 K for the reaction OH + HOz HzO + 0 2 . of hl = (0.99 f 0.12) X
Introduction There is particular interest in the reaction between OH and HOz (eq 1)from the standpoint of models of reactions OH + HOz HzO + 02 (1) occurring in the stratosphere. Recent investigations4p5of this reaction under the low-pressure conditions typical of discharge flow techniques indicate a value of kl N 3 X cm3 molecule-l s-l, but at higher pressure and in the cm3 presence of water vapor, a value of kl in excess of molecule-l s-l has been determined,6 and the suggestion was put forward that the difference could be related to a pressure effect or water vapor effect (complexing of the HO,). Earlier work on the determination of kl, which is cited in ref 4 -6, shows a similarly wide range of values for
-
TABLE I: Kinetic Modela reaction no.
Experimental Section The general characteristics of the apparatus and techniques used in this investigation have been described in detail e a r l i e ~ r .Improvements ~~~ in the method have been Oxygen (99.99%) detailed in a more recent publi~ation.~ and argon (99.999%) from Matheson were used without purification. The water was triply distilled and thoroughly degassed. The decays of H 0 2and OH were monitored by kinetic spectrophotometry at 230 nm (spectral resolution 1.7 nm) and 308.7 nm (spectral resolution 0.08 nm), respectively. The light source for 230 nm was a 1-kW Hg-Xe high-pressure lamp. The lamp used a t 308.7 nm was a 450-W xenon lamp, intensified by applying an approxi-
OH t HO,
H,O + 0, 0 + H,O
1 2
OH + OH
3
OH
4
0t 0 : , 3 0 ,
0.00156
110
5 6 8 9 10 11 12
Ht O:,~HO, HO, + HO, 0,t H,O, OH + H,O, HO, t H,O OH + 10, HO, + 0, O t H O , + O H + 0, H + H10, H, + 0, H + H10, -+ OH + OH H + HID, 0 + H,O
0.095 0.45d 0.088 0.0089 7 1.34 1.93 1.64
110 8 10 11 12 10 110 10
13 14
H t O.KsH,O O+OH-+H+O,
1.00 4.0
10
7
kl.
reaction
10” x rate constant, cm3 molecule” s’-’ ref
-+
-+
+ OHM’H,O,
-+ -+
-+
-+
+
b
9.9 0.22
10
0.83
c
10
For 3.9 X 10’’ molecules of Ar ~ m -3.9 ~ ,x 10’’ molecules of H,O cm-3,and about 1.6 X 10’’molecules of 0, ~ r n at - ~308 K. Determined by the fitting procedure. See text. Thlisvalue takes into account the known effect of water vapor on reaction 6.8 a
mately square wave voltage pulse lasting for up to 2 ms. All experiments were carried out at 35 f 0.5 “C. Oxygen gas (2,4,6, or 8 torr) was introduced into the cell, followed by 12 torr of water vapor. Ar was then added to bring the total pressure up to 1200 torr (3.9 x iO19 molecules ~ m - ~ ) .
0022-3654/80/2084-0819$01.00/00 1980 American Chemical Society
820
The Journal of Physical Chemistry, Vol. 84, No. 8, 1980
I 0.02
1-f
"
Lii et al.
"
determined on the basis of the observed absorbance a t 308.7 nm by using the concentration of OH as determined from V,. Other information used in the fitting procedures is the following: (1) 0 ~ 3 = 2 ~4.49 ~ X cm2 molecule-l, base e;14 (2) OH = 2.21 X cm2 molecule-', base e;15(3) the optical path length, I , is 40 cm, and the absorbance at 308.7 nm ( A = log [IO/It,],where Io is the incident light and Itris the transmitted light) due to OH is related to the concentration of OH by A = (V3X COHX 1)o.6. The latter relationship has been shown to hold under the conditions of these experiments, and the reason that the normal Beer-Lambert law does not apply has been discussed.16 In the fitting procedure, all of the kinetic data shown in Figures 1and 2 were fit by using the same value of the concentration parameter, V,. This was found to be an acceptable procedure despite the following. As the O2 pressure is increased from 6 X 10I6to 2.5 X 1017molecules cm-3 in a system containing 3.9 X lOI9 molecules of Ar ~ , fraction of the and 3.9 X 1017molecules of H 2 0 ~ m -the energy deposited by the electron beam which eventually leads to ionization and excitation of O2would be expected to increase, and conversely, the fraction leading to ionization and excitation of HzO would be expected to decrease. The lack of dependence of V, on [O,] indicates that energy which is transferred to 0, from excited and ionized argon species must subsequently be fairly efficiently transferred to water, causing dissociation of the water. This can be rationalized as follows. On the basis of information on ion-molecule reactions occurring in this system,17 if 02+ is formed by charge transfer from ionized argon species, it will react very efficiently via Oz+ H,O -k Ar O2+-HzO Ar (tlj2 1 ns). The O2+.HzOspecies might, upon neutralization, result in dissociation of the HzO into H and OH. Furthermore, known reactions of the O2+.H2Ospecies with HzO result in the dissociation of water.17 However, O,+.(H,O), and other ionic-complex molecules are involved in the steps leading to dissociation of HzO, and uncertainty arises because of the fact that the products of neutralization of such complexes, which is in competition with the reactions leading to dissociation of HzO, are not known. However, the aforementioned situation with respect to V2 indicates that dissociation of HzO is a frequent result of complex formation involving water and 02+. A t oxygen pressures greater than about 10 torr, the absorbance at 230 nm reaches an appreciable plateau value at about 200 ws, indicating that an appreciable amount of ozone is being formed via reaction 4 from oxygen atoms originating from 0,.Therefore, a small contribution due to this process is built into the kinetic model by assuming that the concentration of oxygen atoms from 0, is given by [O] = V,{0.2[02]/(0.2[02]+ [H20])1. In other words, on a per molecule basis, dissociation of O2 into two oxygen atoms is assumed to be only one-tenth as efficient as dissociation of H 2 0 into H and OH. The oxygen atoms produced from 0, in this way have only a small effect on the calculated curves in Figures 1and 2; the resulting O3 formed serves to raise the "plateau" 230-nm absorbance slightly, in agreement with the observed experimental trend. The values of the three parameters corresponding to the "best-fit'' calculated curves shown in Figures 1 and 2 are v1= hl = (0.99 f 0.12) x 10-10 cm3 molecule-l S-I, VZ = (1.11 0.16) X 1015molecules ~ m - and ~ , V3 = (2.2 f 0.2) ~ X cm2 molecule-l, where the error limits are two standard deviations. The standard deviations are determined by the fitting program. Their relatively small values
02230
O
0
L
'
'
'
100 TIME ( p s )
200
Flgure 1. Absorbance (log [ I o / I 1) at 230 nm (mainly HO ) for samples at 308 K containing 3.9 X ' 0 1 molecules of Ar cm- B, 3.9 X lof7 molecules of H20 ~ m - and ~ , the following O2concentrations: 0.8 X lo1' molecules cm-3 ((0) experimental, calculated); 1.3 X IOi7 molecules cm3 ((0) experimental, (--) calculated); 1.9X 10'~ molecules ~ m ((U) - ~experimental, (- -) calculated); and 2.6 X loi7 molecules ~ m ((A) - ~experimental, (- -) calculated). (-a)
-
-+
0
100
200
TIME
300
(PSI
Figure 2. Absorbance (lo [ I O / I v ] )at 308.7 nm (OH) for samples at 308 K containing 3.9 X 10' molecules of Ar cm3, 3.9 X loi7molecules of H20 ~ m - and ~ , the following O2 concentrations: 0.65 X loi7 molecules cm3 ((0) experimental, calculated); 1.3 X loi7molecules ~ m ((0) - ~experimental, (-) calculated); 1.9 X loi7 molecules cm-3
B
(..a)
((0)experimental, (---) calculated); and 2.6 X ((A)experimental, (--) calculated).
loi7molecules
Results and Discussion The decay of absorbance in samples at 35 "C containing 3.9 X 1019molecules of Ar ~ m -3.9 ~ , X 1017molecules of H20 cmW3, with various concentrations of O,, was measured both a t 230 nm (HO,) and 308.7 nm (OH). Typical HOz data are shown in Figure 1 and OH data in Figure 2. These data were analyzed by an iterative, nonlinear least-squares fitting program which fit the data to the kinetic model given in Table I. The differential equations corresponding to this model were integrated numerically. The variable parameters (V,) were V , (+), V,, the initial concentration of OH radicals produced by the pulse (and an equal concentration of H atoms), and V3,the effective optical absorption cross section for OH. The data were, therefore, treated in terms of a threeparameter fit; the concentration parameter is essentially determined by the observed absorbance at 230 nm and the known13 absorption cross section of H 0 2 ( ~ ~ = ~2.172 X 2 cm2 molecule-', base e ) . The parameter V3 is an effective absorption cross section for OH at 308.7 nm under the experimental conditions employed and is essentially
~
+
+
-
+
J. Phys. Chem. 1980, 84, 821-826
indicate that the three parameters are not strongly correlated. The quality of the fit does not change significantly unless one (or more) of the three parameters is fixed at a value more than two standard deviations removed from the best-fit value. The value of V,, the effective optical absorption cross section of OH, can be shown to be consistent with earlier measurements in this laboratory on the OH radical in pulse-irradiated water vapor.16 Despite the complexity of the kinetic model, we believe that the value of kl obtained is reasonably accurate and supports the "high" values of kl previously mentioned. Most of the reactions in Table I are of relatively minor importance, and errors in their rate constants do not markedly affect kl. The most significant rate constant, other than h,, in terms of its effect on the fit is h3. A value of 0.5 X lo-"' for h3would be expected on the basis of the value k3 with N2 as t2 third body18 if we use the reportedTg relative third-body efficiencies for N2,Ar, B20,and 02. However, WI: find that a value of h3 N 0.8 X cm3 molecule-l results in a better fit. Making k3 even larger causes the H02decay to be too slow over the first 20 /LS. Hence, under the conditions of these experiments a h3 value of (0.8 f 0.2) X 1O-I' cm3 molecule-l s-l is indicated. The effect of'the value used for h3 on the "best-fit" value of kl was not large, however. The change in kl over the was less than 5%. range h3 = 0.5 X 10-11-1.1 X Because of the known effect of [H20] on the value of k6,6~8~20 it is reasonable to ask whether the presence of 3.9 X loT7molecules of H 2 0 has an effect on the value obtained for hl. Therefore, the H02decay was examined ~ , X loT7molecules of a t 3.9 X loT9molecules of Ar ~ m - 2.6 O2~ m - and ~ , concentrations of H20from 6 X 1016-5 X lo1' molecules CI+. No evidence of a significant change in the H02 decay rate was observed, indicating that reaction 1
821
is apparently not appreciably affected by the complexing of H02by H20. The fact that hl is about 20 times greater than k6 is the reason that the known variation of h, with [H20]does not appreciably influence the observed H02 decay in the latter series of experiments.
Achnowledgment. We thank Dr. C. Jonah for the least-squares fitting program.
References and Notes (1) Work performed under the auspices of the Office of Basic Energy Sciences of the U S . Department of Energy. (2) Department of Chemistry, Malcolm X College, Chicago, IL 60612. (3) Scientific Research Laboratory, Ford Motor Company, Dearborn, MI 48121. (4) J. S.Chang and F. Kaufman, J . Phys. Chem., 82, 1683 (1979). (5) W. Hack, A. W. Preuss, and H. Gg.Wagner, Ber. Bunsenges. Phys. Chem., 82, 1'167 (1978). (6) W. 8. DeMore, J. Phys. Chem., 83, 1113 (1979). (7) S. Gordon, W. LAulac, and P. Nangia, J. phys. Chem., 75, 2087 (1971). ( 8 ) E. J. Hamilton, Jr., and R. R. Lii, Int. J . Chem. Kinet., 9, 875 (1977). (9) R. R. Lii, R. A. Gorse, Jr., M. C. Sauer, Jr., and S.Gordon, J . Phys. Chem., companion paper in this issue. 10) R. F. Hampson, Jr., and D. Garvin, Eds., Nat. Bur. Stand. (U.S.), Spec. Publ., No. 513 (1978). 11) A. R. Ravishankara, P. H. Wine, and A. 0. Langford, J . Chem. Phys., 70, 984 (1979). 12) R. R. Lii, M. C. Sauer, Jr., and S.Gordon, J. Phys. Chem., preceding paper in this issue. 13) T. T. Paukert and H. S.Johnston, J. Chem. Phys., 56, 2824 (1972). 14) M. Griggs, J . Chem. Phys., 49, 857 (1968). 15) J. G. Caivert and J. N. Pitts, Jr., "Photochemistry", Wiley, New York 1966. (16) S.Gordon andl W. A. Mulac, Int. J . Chem. Kinet., Symp., No. 1, 289-299 (1975). (17) P. Kebarle in "Ion-Molecule Reactions", J. Franklin Ed., Plenum Press, New York, 1972, Chapter 7, pp 328-332. (18) D. W. Trainor and C. W. von Rosenberg, Jr., J . Chem. Phys., 61, 1010 (1974). (19) G. Black and G. Porter, Proc. R . SOC.London, Ser. A , 266, 185 (1962). (20) R. A. Cox and J. P. Burrows, J. Phys. Chem., 83, 2560 (1979).
Production of OH from Photolysis of HOC1 at 307-309 nm Mario J. Molina,*st Takashi Ishiwata, and Luisa 1. Molina Department of Chemistry, University of California, Irvine, California 927 17 (Received September 7, 1979) Publication costs assisted by the National Aeronautics and Space Administration
Laser-induced fluorescence has been used to measure the amount, of OH produced in the laser photolysis of HOCl at -310 nm. The HOCl was generated by the C120 + H20 and by the HC1+ ClzO reactions. The laser technique samples OH on a microsecond time scale so that secondary radical reactions cannot contribute significantly to the observed signals. The fluorescence signals were calibrated with 03-H20 mixtures, the reactions O3 f hu O(lD) + O2and O('D) + H20 20H yielding known amounts of OH. The value for the absorption cross section of HOCl at 310 nm has been determined to be -6 X cm2,assuming K,, = 0.082 for the ClzO + H20 ~ ' 2HOC1 t system and assuming unit quantum yield for the production of OH. The present investigation further supports earlier experimental work concluding that HOCl will photodissociate rapidly and will not be an important holding tank for chlorine in the stratosphere. These results are in disagreement with theoretical calculations of the 1JV absorption spectrum of HOCl in the 300-350-nm wavelength region.
-
-
Introduction The ultraviolet spectrum of HOCl has been the subject of several recent investigations because of the potential role that this species may play in stratospheric chemistry. The UV spectrum is needed to estimate its rate of phoDreyfus Teacher-Scholar. 0022-3654/80/2084-0821$01.00/0
todecomposition in the stratosphere where HOCl is formed by the rapid reaction between HOz and C10 radicals.' The theoretical calculations of Jaffe and LanghoffZ and some experiments done by Timmons' and by Hisatsune3indicate negligible HOCl absorption at wavelengths longer than 300 nm; in contrast,, our earlier work4 as well as the experimental results of Jaffe and DeMore5 and of Knauth et a1.6 0 1980 American
Chemical Society
