J. Phys. Chem. 1983, 87, 4467-4470
for permitting the use of their photon counting system. The cost of the present investigation was partly defrayed by the Japanese Ministry of Education, Science, and Culture through a Grant-in-Aid (56540261) to H.M. and a Grant-in-Aid for Special Project Research on Photo-
Rate Constant of the OH -tH,O,
-
HO,
4467
biology to N.M. and H.M. Registry No. 9-Ethylphenanthrene, 3674-75-7; 1,3-di(9phenanthryl)propane,82901-39-1;poly[2-(9-phenanthryl)ethyl vinyl ether] (homopolymer), 86900-66-5; poIy(9-vinyIphenanthrene) (homopolymer), 86885-30-5.
+ H20 Reaction
John J. Lamb,+ Luisa 1. Moiina,* Craig A. Smith, and Mario J. Molina"' Department of ChemlStW, Universtty of California, Imine, California 92717 (Received: October 25, 1982)
-
Absolute rate constants for the reaction OH + H202 H 0 2 + H 2 0 have been measured by using the flash photolysis resonance fluorescence technique over the temperature range 241-413 K and at 1 atm of helium. H202concentrations were measured by ultraviolet and infrared absorption spectrophotometry. The rate constant at 294 K was found to be kl = (1.8 i 0.3) X cm3molecule-l s-l, in agreement with recent literature values. The results yield a curved Arrhenius plot, with a slight increase in the rate constant at the lower temperatures.
Introduction The gas-phase reaction of OH with H202is an important component of the atmospheric HO, cycle: OH
+ H202
+
HO2 + H2O
(1)
H202is formed in the atmosphere mainly by the recombination reaction of H 0 2 radicals (reaction 2) and its HO2 + HOP
H202 + 0
2
(2)
principal loss processes are reaction with OH radicals (reaction 1)and photodissociation (reaction 3). In adH202 + hv
+
OH
+ OH
(3)
dition, rainout and washout processes represent significant sinks for H202in the troposphere. Several previous studies of reaction 1 have been reported: early measurements of the rate constant using steady-state photolysis methods,1*2flash ph0tolysis,4~and discharge flow techniques5 gave results in the (0.62-1.2) X cm3 molecule-l s-l range at 298 K. More recently, measurements by two different methods-discharge floW69' and flash photolysis resonance fluorescenceel0-yielded rate constants that are about a factor of 2 larger at 298 K and which showed linear Arrhenius plots, with the rate constant decreasing by -10% at 250 K. In this work we essentially confirm these new rate constant values using the flash photolysis resonance fluorescence technique at 1atm of He; however, we observe a curved Arrhenius plot, with the rate constant increasing below room temperature and becoming 10% larger at 250 K.
Experimental Section Instrumentation and Chemicals. An all-Pyrex jacketed reaction cell with an internal volume of -150 cm3was used in the experiments. The cell was maintained at constant temperature by flowing a thermostated liquid through the outer jacket surrounding the cell. The temperature was monitored a t five different positions on the cell with copper-constantan thermocouples. 'Current address: Hydrocarbon Research Institute, University of Southern California, Los Angeles, CA 90089-1661. Current address: Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109.
*
0022-365418312087-4467$01.50/0
OH radicals were produced by flash photolysis of H202 or HNOBusing a nitrogen flash lamp equipped with a Suprasil quartz lens and operated around 15 kV. An adjustable iris controlled the amount of photolysis light reaching the cell without affecting the beam diameter. An OH microwave-dischargelamp excited resonance fluorescence in the 0-0 band of the A2Z+-X211system (A 308 nm). This fluorescence was monitored perpendicular to both the flash lamp and the resonance lamp radiation flux by an EMI-9782QA photomultiplier fitted with an interference filter (290-320 nm, 20% T at.307 nm) and a Corning 7-54 filter. The resonance lamp and the photomultiplier were also fitted with quartz lenses which mildly focused the radiation. The signals, in the form of photon counts, were processed by a signal averager-computer system (Inotech Ultima 11,Data General-Nova 3) using a nonlinear leastsquares routine to analyze the decays; data in the first few channels ( t < 500 ps) were not considered in the analysis. For each decay rate sufficient flashes (usually 400) were averaged to obtain a well-defined temporal profile of OH concentration over at least three l / e times. The residuals and the autocorrelation coefficients were plotted routinely; in no case was there statistical evidence for departures from the single exponential decay expected for pseudofirst-order conditions. In order to avoid accumulation of products and to minimize the decomposition of H202,the reactant mixture was flowed through the reaction cell with a linear velocity of about 8 cm s-l. The flash lamp repetition rate was about 1 Hz, so that the cell was refilled with a fresh reactant mixture for every flash. H202was introduced into the system by passing He gas through a glass bubbler containing several milliliters of
-
(1) R. A. Gorse and D. H. Volman, J.Photochem., 1, 1 (1972). (2)J. F.Meagher and J. Heicklen, J. Photochem., 3, 455 (1979). (3)N. R. Greiner, J. Phys. Chem., 72, 406 (1968). (4)G.Harris and J. N. Pitta, Jr., J. Chem. Phys., 70, 2581 (1979). (5)W. Hack, K. Hoyermann, and H. Gg. Wagner, Int. J. Chem. Kinet. Symp., 1, 329 (1975). (6) U. C. Sridharan, B. Reimann, and F. Kaufman, J. Chem. Phys., 73,1286 (1980). (7)L. F. Keyser, J.Phys. Chem., 84,1659 (1980). (8)P.H. Wine, D. H. Semmes, and A. R. Ravishankara, J. Phys. Chem., in press. (9)M. J. Kurylo, J. L. Murphy, G. S. Haller, and K. D. Cornett, Int. J. Chem. Kinet., 14, 1149 (1982). (lo)W. J. Marinelli and H. S. Johnston, J. Chem. Phys., 77, 1225 (1982).
0 1983 American Chemical Society
4488
Lamb et al.
The Journal of Physical Chemistty, Vol. 87, No. 22, 1983
concentrated H202solution at 0 "C. For some runs HN03 was added in small amounts (5-10 mtorr) in order to provide an alternate more efficient photolytic OH source. A separate bubbler was assembled containing several milliliters of aqueous HN03 at 0 OC. Helium gas was slowly bubbled through the liquid and the mixture was added to the H202/He flow downstream from the UV spectrophotometer. All connections and tubing after the bubblers were either glass or Teflon. Helium gas (Matheson Gas Products) was passed through a molecular sieve trap kept at liquid-nitrogen temperature. Hydrogen peroxide (go%, FMC Corp.) was used without further purification. Nitric acid (70%, Mallinckrodt, Analytical Reagent grade) and C2H6 (Matheson Gas Products, CP grade) were also used without further purification. The flow rates were measured by using Matheson mass flowmeters. Pressures were monitored with an MKS capacitance manometer (0-100-torr range) and with a Bourdon-type Wallace-Tiernan gauge (0-800-torr range). H 2 0 2Concentration Determination. The H202concentrations were measured by using a Perkin-Elmer 552 W-vis spectrophotometer equipped with a Wcm cell with folded optics, to give an optical path length of 100 cm. Due to the low partial pressures of H202used (0.8-13 mtorr), this UV measurement was made before the final dilution with He. The absorbance of the room-temperature gas mixture was monitored in the 205-220-nm range, and the concentrations were calculated by using the absorption cross section data of Molina and Molina.'l For most of the runs, the H202concentrations were also measured after the reaction cell by using a Nicolet-7199 Fourier-transform infrared spectrometer (FTIR) equipped with a standard Nicolet data-handling system and a liquid-nitrogen-cooled HgCdTe detector. A 55 cm long, Teflon-coated absorption cell with a -6-cm2 cross section and maintained at room temperature was interfaced to the FTIR. The cell was fitted with internal mirrors and KC1 windows, with an optical path of -20 m. The HN03 concentrations were also monitored with the FTIR cell. The IR band strengths for H202and HN03were calibrated as described earlier.ll The H202concentrations calculated from UV and IR measurements agreed to within a few percent, except at the highest temperature (413 K), where the discrepancy indicated about 15% decomposition in the resonance fluorescence cell. Sensitivity and Control Experiments. The OH detection sensitivity of the resonance fluorescence system was calibrated by photolyzing HN03/He mixtures a t 222 nm with a KrCl excimer laser whose power was measured with a thermopile: a t 1atm of He, this sensitivity was -lo-' counts cm3molecule-l s-l. When this figure was used, the initial OH concentration for the flash photolysis runs was estimated to be at most loll molecules ~ m - ~ . Some control experiments were carried out to determine the OH + C2H6 rate constant in our apparatus, as a function of temperature, using HN03 as an OH photolytic source, and using a C2H6/He mixture analyzed by gas chromatography. The observed rate constant decreased by a factor of 2 between 298 and 250 K, and the results agreed to within 10% at these two temperatures with the values recommended by the NASA Panel for Data Evaluation. l2 (11)L. T.Molina and M. J. Molina, J. Photochem., 15, 97 (1981). (12)W. B.DeMore, D. M. Golden, R. F. Hampaon, C. J. Howard, M. J. Kurylo, M. J. Molina, A. R. Ravishankara, and R. T. Watson, JPL Publ. NO.82-57 (1982).
t
f
3
>-
1021
E
z t
t IO0
0
1
2
3
4
5
6
7
8
T I M E (ms) Figure 1. Typical OH temporal profile following photolysis of an H,O2/HN0,/H20/He mixhrre. T = 241 K P = 760 torr; [H202]= 5.3 mtorr; [HNO,] = 4.4 mtorr.
700 -
600
-
-
T 294K
-
500 -
h
Ivl
400-
I
Y
300 -
i
1 [HzOz]
(
molecule ~ m - ~ )
Fl~ure2. Plot of k' vs. [bo2]at room temperature and atmospheric pressure (first 294 K entry In Table I). Line Is least-squares fit of experimental points.
Results and Discussion Chemical Kinetics. The OH H202 reaction was studied over the temperature range 241-413 K and at a He pressure of 760 torr. A total of 25 bimolecular rate constants were measured. The results are shown in Table I. All experiments were carried out under pseudo-firstorder conditions:
+
In ([OH],/[OH],) = (kl[H202]+ kd)t
k't
where k d is the first-order rate constant for loss of OH in the absence of H202and k' is the pseudo-first-order rate constant. kl was obtained from the slope of plots of k' vs. [H202]. A typical OH decay is shown in Figure 1, and Figure 2 shows a typical plot of k' vs. [H202]. The individual rate constants were determined from a linear
-
Rate Constant of the OH + H202
H02
+ H20 Reaction
The Journal of Physical Chemism, Vol. 87, No. 22, 1983 4489
TABLE I: Rate Constant Data for the Reaction of OH with H,O,
HP, press.
HNO, no. press., of mtorr runs 0 13 5 19
temp, K
range,
241 241
1.3-6.8 2.4-6.0
251 251 251
0.7-5.6 0.9-9.7 0.9-7.9
0 0 0
12 10
294 294 294 294 294 294 294 294 294
1.1-11.3 1.8-7.1 1.9-6.3 1.2-12.0 5.0-8.7 1.7-9.9 1.3-11.0 1.3-10.6 2.4-8.2
0 0 5 0 0 0
15 22 12 11 4
333 333 333
2.4-11.8 1.1-9.4 0.8-8.3
0 0
mtorr
T ( K)
10'%,
av:
15 av:
0
0 0
7 6 6 4 av:
0
21 12 9 av:
365 365
2.9-9.2 2.4-9.8
5 5
17 12
413 413
2.8-6.8 2.9-9.2
5 10
9 9
?
lo, cm3
molecule-I s-'
av: av:
2.10 ? 0.13 2.04 ? 0.07 2.07 * 0.10 1.95f 0.06 1.90 * 0.05 1.90 ? 0.05 1.92? 0.05 1.88 ? 0.11 1.86 t 0.10 1.81 * 0.06 1.85 ? 0.04 1.71 ? 0.03 1.81 ? 0.04 1.66 ? 0.05 1.81 ? 0.05 1.77 ? 0.05 1.80 * 0.06 1.66 * 0.08 1.76 ? 0.08 1.75 f 0.04 1.72 ?r 0.07 1.76 f 0.13 1.79 ?r 0.15 1.78? 0.14 2.04 ? 0.30 1.64 ? 0.12 1.84t 0.21
least-squares analysis of these latter plots. The value of the intercept was typically 20-80 s-', in very good agreement with the value for kd obtained in the absence of HzOz (with H20as the photolytic source of OH). The uncertainties listed in Table I represent one standard deviation, that is, about &3% at 294 K. We estimate that systematic errors raise the uncertainty to f15-20% yielding k1(294 K) = (1.8 f 0.3) X 10-l2cm3 molecule-' s-' For runs containing HN03, contributions to the observed k' from the reaction OH + HN03 were subtracted out by using a rate constant obtained from the Arrhenius equation
k = (9.4 X 1015)exp(778/T) cm3molecule-' s-l, taken from the recommendation of the NASA Panel for Data Evaluation.lZ This correction amounted in all cases to less than 10% of the observed pseudo-fmborder rate constant. The addition of HN03 provided about a threefold increase in the signal-to-noise ratio for an equivalent number of flashes. Secondary Reactions. Given the relatively small radical concentration ([OH], < lo1' cmS) complications from secondary reactions should be minimal. As a check, the initial [OH] was varied by more than a factor of 2 by changing the intensity of the photolysis light; no significant change was observed in k'. Also, as may be seen in Table I, the addition of HN03 yielded essentially the same results. Furthermore, no departures from an exponential decay in OH were evident in any of the experiments. The fast reaction between H atoms and HOz radicals is potentially a source of error since OH radicals can be regenerated by this route, thus decreasing the OH decay rate H + HO2 OH + OH (4) -+
Some H atoms might have been produced by the flash photolysis of water vapor present in our sample in spite of the 165-nm Suprasil cutoff, but the expected amounts were too small to significantly perturb the rate constant measurement: by comparison with signals generated with
IO e
-
lyl
-
l
I
I
s --
H2O o Sridhoron e t o l (1980) A K e y s e r (19801 A Wine e t o l (1982) o Kurylo elol. (19821
'a - 6¶
u
-
$ 4 -
E " -
0 0
*e
x-
1
-
O H +H202-H02+
-
-
Morinelli ond Johnston This w o r k
I 3
2
IOOO/T
,
4
(K-')
Figure 4. Arrhenius plot of k values from the most recent studies.
HzO alone, we estimate [H],/[OH], < 0.03 for typical runs without HN03 added, and [H],/[OH], < 0.01 in the presence of HNOP If the H atoms produced in the photolysis were to be converted initially to HOz (perhaps by reaction with Oz present as an impurity) the reaction between OH and H02 would then be a more important source of error, with an effect opposite to that of reaction 4; i.e., the OH decay rate would increase: OH
+ HOz
-
H2O
+0 2
(5)
However, as with reaction 4 a simple calculation shows that at the radical levels present in our experiment the effect of reaction 5 should be minimal, i.e., at most 2% error in the rate constant measurements. Detailed computer simulations of the kinetics incorporating these as well as other potentially interferring secondary reactions confirmed the simple calculations; the results of these simulations essentially duplicate those presented by Kurylo et al.,9and hence they will not be repeated here. Temperature Dependence. The temperature dependence of the rate constant for the reaction under study is shown in Figure 3, and also in Figure 4 along with the results of the four groups which have studied this reaction recently. A linear least-squares fit to our results yields the expression k = (1.47 f 0.20) X 10-l2 exp[(7O f 4l)/T'J cm3 molecule-' s-' (6) where the uncertainty represents one standard deviation. However, a clear departure from simple Arrhenius behavior
Lamb et al.
The Journal of Physlcal Chemistry, Vol. 87, No. 22, 1983
4470
TABLE 11: Comparison of Recent Rate Measurements for the OH + H,O,Reaction 10'2h,,a cm3 molecule" s-l temp, K 24 1 251 294 333 365 413
b
C
d
f
e
1.50 i 1.54 i 1.69 i 1.85 f 1.89 i 1.99 ?.
0.30 1.49 i 0.48 1.26 i 0 . 2 9 1.49 i 0.31 0.30 1.52 * 0.39 1.31 i 0.28 1.53 i 0.31 0.26 1.64 f 0.32 1.53 i 0.28 1.68 i 0.30 1.81 i 0.24 0.32 1.72 ?. 0.36 1.69 t 0.34 1.79 i 0.34 0.36 1.78 i 0.45 1.81 i 0.40 1.87 i 0.38 0.42 1.85 i 0.52 1.97 i 0.47 1.97 i 0.42 * Computed from the Arrhenius expression and error bars given by each group. The error bars for T # estimated as explained by the NASA Panel for Data Evaluation (ref 12). Reference 6. Reference 7. e Reference 9. - f Reference 10. T = 298 K. g This work.
is apparent in the figure, and the results are better fitted by the expression k = [(7.0 f 2.0) X 10-20]T2.5 exp[(838 f 8 6 ) / T ] cm3 molecule-' s-l (7) As expected, the fit is insensitive to the exact value of the power n to which T i s raised (the minimum standard deviation was actually obtained for n = 2.8). Equation 6 reproduces the experimental rate constant values to within 7%, and eq 7 to within 3%. Curvature in the Arrhenius plot can be rationalized by assuming the existence of two different reaction mechanisms: a direct H-atom abstraction predominating at the higher temperatures, and an addition mechanism, Le., a "complex-mode" bimolecular reaction (see, for example DeMore et al.),12predominating at the lower temperatures. Recently Leu and Smith13found similar, even more pronounced curvature in the Arrhenius plot for the reaction OH + H2S, which they also interpreted in terms of the dual-mechanism postulate. Thus, it appears that at low tempertures the negative activation energies are quite prevalent for reactions of the OH radical with polar molecules which may form hydrogen bonds, e.g., HN03, HN04, H g , and H202.12The H2O2molecule is nonplanar, so that a plausible structure for the intermediate adduct is a hydrogen-bonded five-membered ring:
?\
0
I I
8
2.04 i 1.97 i 1.79 i 1.75 i 1.77 i 1.85 i
0.51 0.46 0.30 0.35 0.41 0.49
294 K were Reference 8.
Marinelli and Johnstonlosuggested instead that the OH
+ H202reaction might be an elementary bimolecular re-
action with a transition state consisting of a zig-zag sixmembered chain with a linear 0--H- -0segment. Perhaps one way to distinguish between these alternate reaction mechansims is to measure the rate constant as a function of total pressure: complex-mode reactions proceeding through a relatively long-lived intermediate are expected to have rate coefficients which increase slightly with increasing pres~ure.'~Such experiments are in progress in our laboratory. Comparison with Other Measurements. Inspection of Table I1 and Figure 4 shows that our measured rate constants for the OH H202reaction at 273 K and above are in very good agreement with those reported most recentlfl0 but are larger by about 25% at 250 K. The discrepancy appears to lie just within the combined error limits. Sridharan et al. suggested a possible leveling off at the low-temperature end of their Arrhenius plot, a trend which appears also in the data of Wine et al., and, if one ignores the lowest temperature run, it is borne out in the data of Keyser as well. In summary, it is now well established that earlier measurements underestimated the rate constant for the OH HzO reaction by about a factor of 2 for reasons that have been discussed elsewhere at l e ~ ~ g t The h . ~results ~ ~ of the most recent flash photolysis and discharge flow studies are in very good agreement with each other, except possibly at 250 K and below, where our numbers show an increase in the rate constant with decreasing temperatures.
+
+
\ \ \
\ \
V
(13) M.-T. Leu and R. H. Smith, J. Phys. Chem., 86,73 (1982).
Acknowledgment. This research was supported by the Department of Transportation, High Altitude Pollution Program, under Contract DTFA 01-80-C-10084. Registry No. Hydroxyl, 3352-57-6; hydrogen peroxide, 7722-84-1. (14) J. J. Margitan and R. T. Wataon, J . Phys. Chem., 86,3819(1982).
