J. Phys. Chem. 1985,89, 2630-2635
2630
and 2.21 s-l torr-'. The similarity of the slopes indicates similar energy-transfer characteristics for all three systems, even though the reactants differ considerably in size and polarity. Another interesting experimental result is the observed phase dependence of the kinetic parameters. Previous gas-phase conformational studies of cyclohexane3 revealed that kinetic parameters can be phase dependent in the absence of strong electrostatic interactions. Tetrahydropyran, a polar molecule resembling cyclohexane except the -CH2- group is replaced by an oxygen atom, also displays phase-dependent kinetics associated with the ring inversion process. Gas-phase unimolecular ring inversion rates of THP are ca. 4 times slower, and the activation parameters are slightly higher than those obtained in liquid-phase studies. AC* for conversion of the chair conformer to the twist-boat transition state is 10.5 (0.2) kcal/mol in the gas phase and 9.9 kcal/mol in CS2 solutions (see Table 111). The observed phase dependence is compatible with a negative activation volume, A P , for the process.26 The internal pressure of an ideal gas is 0 and that of (25) Slater, N. B. 'Theory of Unimolecular Reactions"; Cornell University Press: Ithaca, NY, 1959.
liquid CS2 at 298 K is 3714 atm.27 When these values and a AAG* of 600 cal/mol are used, a A P value of ca. -5 cm3/mol is obtained. This is, similar to previously observed values for cyclohexane and is also consistent with high-pressure liquid studies of cyclohexane which also determined a negative activation volume for the ring inversion process. THP and cyclohexane show strong similarities in their gas-phase ring inversion kinetics, demonstrating that this process is not significantly affected when the -CH2- group is replaced by an oxygen atom. Hence the lone pairs on the oxygen atom play the same role as the hydrogen atoms in the methylene groups in the stabilization of the chair conformation.
Acknowledgment. We are pleased to acknowledge support from the National Science Foundation (Grant CHE82- 10844 and CHE83-51698(PYI)) and the National Institutes of Health (Grant PHSGM 29985). Registry No. Tetrahydropyran, 142-68-7. (26) Ouellette, R. J.; Williams, S. H. J. Am. Chem. SOC.1971, 93, 466-47 1. (27) Dack, M. R. T. Chem. SOC.Rev. 1975, 4, 211-229.
Quenching of OH(A*z+) by Alkanes and Chlorofluorocarbons at Room Temperature Robert D. Stephens Environmental Science Department, General Motors Research Laboratories, Warren, Michigan 48090-9055 (Received: November 30, 1984)
This paper presents the rate constants for the electronic quenching of OH(A22+)by a series of alkanes. The quenching measurements were obtained from discharge flow-resonance fluorescence investigations of OH(X211) reaction kinetics and are the first values reported for these species. An empirical relationship was found to exist between the measured rate constants for quenching of OH(A22+) by alkanes and the number of primary, secondary, and tertiary hydrogens present in the alkane. The rate constants measured are closely approximated by the expression kq = (0.1lNp+ 0.41Ns + 2.05NT) X lo+' cm3molecule-l s-l, where Np, Ns, and NT.are respectively the number of primary, secondary, and tertiary hydrogens in the alkane. An examination of previously reported rate constants for quenching of OH(A*Z+) by halocarbons also reveals a relationship between the rate constantsand the number and type of substituents in the halocarbons. These rate constants can be approximated by the expression k(C,F,HbCI,) = (0.47~+ 0.92b + 2.86~)X 1O-Io cm3 molecule-l s-l. For alkanes, the quenching rate constants are correlated with C-H bond dissociation energies, whereas for halocarbons the correlation is with C-CI bond dissociation energies. These relationships suggest the possibility of reactive quenching of OH(A2Z+) by these species.
Introduction The importance of the hydroxyl radical in combustion and atmospheric chemical processes has made it the subject of study for many years. Extensive measurements of rate constants for both the O H ( X 2 n ) reaction kinetics'-' and, to a lesser degree, OH(A28+)quenching kinetics&10have been reported. Quenching (1) N. R. Greiner, J . Chem. Phys., 53, 1070 (1970). (2) R. P. Overend, G. Paraskevopoulos, and R. J. Cvetanovic, Can. J . Chem., 53, 3374 (1975). (3) R. D. Stephens, J. Phys. Chem., 88, 3308 (1984). (4) L. G. Anderson and R. D. Stephens, "XV Informal Conference on Photochemistry", Stanford, CA, 1982. See also General Motors Research
Publication GMR-4087. (5) F. P. Tully, A. R. Ravishankara, and K. Carr, Int. J. Chem. Kine?., 15, 1111 (1983). (6) M.-T. Leu and R. H. Smith, J. Phys. Chem., 85, 2570 (1981). (7) M. A. A. Clyne and P. M. Holt, J. Chem. Soc.,Faraday Trans. 2,75, 569 (1979). (8) K. Schofield, J. Phys. Chem. ReJ Data, 8, 763 (1979). (9) I. S. McDermid and J. B. Laudenslager, J . Chem. Phys., 76, 1824 (1982). (10) C. Y.Chan, R. J. OBrien, T. M. Hard, and T. B. Cook, J. Phys. Chem., 87, 4966 (1983).
0022-3654/85/2089-2630$01.50/0
of OH(A2Z+) is thought to occur via energy-transfer processes; however, chemical reactions have not been ruled out in most cases. Quenching processes are typically very efficient, occurring at rates near gas kinetic values. For this reason, quenching measurements have been useful in the quantification of O H concentration measurements made via laser-induced fluorescence.' 1 ~ 1 2 A wider range of quenching measurements could also be useful in providing an improved understanding of quenching processes. This paper presents measurements on the quenching of OH(A2Z') by a series of alkanes. No previous literature values are available for the quenching of OH(A2Z+) by these species. The quenching rate constants are highly correlated to the number of primary, secondary, and tertiary hydrogens in the alkanes. We also report for the first time a relationship between the quenching of OH(A22+) by halocarbons and the number and type of substituents in the halocarbons. These relationships may be due to the variation in the bond dissociation energies for C-H bonds in the alkanes and C-Cl bonds in the halocarbons. Such relationships (11) J. H. Bechtel and R. E. Teets, Appf. Opt., 18, 4138 (1979). (12) C. Morley, Symp. (In?.)Combust. [Proc.],18th, 23 (1981).
0 1985 American Chemical Society
Quenching of OH(A2Z+) by Alkanes and Halocarbons
The Journal of Physical Chemistry, Vol. 89, No. 12, 1985 2631
suggest the possibility of abstraction reaction mechanisms for quenching of OH(AZZ+).
Experimental Section Since a detailed description of the experimental apparatus has been provided previ~usly,~ a general review only will be provided here. The apparatus consists of a discharge-flow tube of 2.5 cm i.d. internally coated with halocarbon wax (Halocarbon Products Corp. Series 15-00). The O H radicals were generated by the reaction of atomic hydrogen (produced by microwave dissociation of molecular hydrogen) with an excess of nitrogen dioxide. This fast OH formation reaction occurred approximately 30-cm upstream of the first of eight injection ports equally spaced at 10-45 cm from a resonance fluorescence cell. Fluorescence was induced by a microwave resonance lamp containing approximately 2 torr (0.266 P a ) of a mixture of 1% water in helium. The steady-state OH fluorescence was monitored via a photomultiplier tube (EM1 9789QA) operated in the photon-counting mode. Fluorescence was collected from a region approximately 0.5 cm in diameter through a series of lenses and apertures and an interference filter with peak transmittance at 3088 A and full-width half-maximum transmittance at 3055 and 3172 A. The alkanes used in this study (99.5% purities except ethane at 99.0%) were analyzed by gas chromatography-mass spectrometry and were found to meet the manufacturers specifications for purity. The helium (99.999%), hydrogen (92 ppm) in helium, nitrogen dioxide (274 ppm) in helium, and alkanes were used without further purification. All gas flows were controlled and measured via calibrated electronic mass flow controllers. The flow rates, pressure, temperature, and photon count rates were collected and analyzed via computer. Statistical tests were applied to these measurements to verify that data collection occurred during stable experimental conditions. A series of eight measurements of each parameter was made a t each reaction time and the data were remeasured if the standard deviation in any measurement exceeded 3%. This system was used to measure the ground-state OH reaction kinetics at 1 torr (0.133 kPa) for five reactions of the type OH alkane HzO alkyl radical (1)
+
-
+
as discussed e l ~ e w h e r e .The ~ alkanes investigated were ethane, propane, n-butane, isobutane, and 2,2-dimethylpropane. It is from these ground-state O H reaction kinetics experiments that the values for the OH(AZZ+)quenching rate constants reported herein were derived. The ground-state O H reaction rates were obtained by measuring OH fluorescence intensity in the absence of the alkane reactant ( I t ) , then measuring the fluorescence intensity during the sequential addition of alkane through each of eight injection ports, and, finally, after completion of reactant addition, remeasuring the fluorescence intensity in the absence of reactant. This last measurement of fluorescence intensity verifies that experimental parameters (e.g., resonance lamp intensity and OH concentration) have remained stable between the beginning and end of each experiment. These measurements provide information on relative OH concentration (fluorescence intensity) vs. reaction time (determined by converting reaction distance to reaction time). Gas flow velocities used in all experiments were approximately 20 m S-'
.
The alkane gas flow rate was always less than 3% of total gas flow and did not, therefore, perturb the concentration of O H or other quenchers present in the experiments. All experiments were performed under pseudo-first-order conditions where [alkane] >> [OH] < 3 X 10" molecules cm-j. A plot of In (fluorescence intensity) vs. reaction time yields a slope which equals the observed pseudo-first-order decay rate, Kobsd,of O H , where Kobsd= k,[alkane] k,. The k , term represents the loss of OH at the flow tube wall. Figure 1 shows examples of data plotted in this fashion for several concentrations of dimethylpropane. An extrapolation of such data to zero reaction time provides a value for the OH fluorescence intensity (If)in the absence of removal of OH(X211) by reaction with alkane but in the presence of added
+
11
I
g9 -c
0
10
5
15
20
25
Reaction Time (msec) Figure 1. Plots of the natural log of OH fluorescence signal vs. reaction time at 296 K. The curves shown correspond to dimethylpropanecon1.30 X loL4,and 1.90 X lo1' molecules cm-3, centrations of 4.54 X respectively, from top to bottom. Statistical uncertaintiesshown are two standard deviations. Note that 1, (Y intercept) decreases with increasing reactant concentration. alkane, which also quenches fluorescence. These values of I f were used in conjunction with the measured values of I? to calculate the quenching rate constants via a Stern-Volmer analysis.
Results The processes involved in the resonance fluorescence detection of OH are OH(X211)
+ hv
+
OH(A2Z+) (excitation by resonance lamp) (2)
+ hv (fluorescence) OH(X211) + A (quenching)
OH(A22+) A OH(X211) OH(AZZ+)+ A
kp"
(3) (4)
The term, k,, is the rate constant for the radiative decay of OH(A%+) and kqAis the rate constant for OH(A2Z+)quenching by any species A. Although no evidence is available, it is possible that the OH(AZZ+)also chemically reacts with the quencher. For the alkanes this might occur analogously to reaction 1: OH(A2Z+)
+ alkane
kd4
HzO
+ alkyl radical
(5)
Under the experimental conditions employed, the primary species responsible for OH(AZZ+)quenching are helium, nitrogen dioxide, and alkane. In the absence of alkane, the OH(AzZ+) fluorescence lifetime (T?) is T?
= (k,
+ kqHC[He]+ kqN02[N02])-'
In the presence of alkane, the fluorescence lifetime (Tf) is Tf
= (k,
+ kqHC[He]+ kqNo2[NOz]+ kqA[alkane])-'
At any given hydroxyl concentration, the ratio of O H fluorescence intensity in the absence of alkane to the fluorescence intensity in the presence of alkane is I?/If =
T P / T ~= 1 + ~ ? k , ~ [ a l k a n e ]
Using the radiative lifetime of OH(A2Z+)and the rate constants for quenching of OH(AZZ+)by nitrogen dioxide and helium as summarized by Schofield,8 the OH(AzZ+) fluorescence lifetime in the absence of alkane is essentially equal to the radiative lifetime of OH(A2Z+),Le., approximately 0.76 X 10" s . ~Given the gas flow velocities employed in the experiments and the spatial resolution of fluorescence detection, the time resolution of fluorescence detection is long relative to processes 2-5, hence these processes are in steady state. By plotting the ratio, Ip/If,vs. concentration of alkane, a value for rpkqAis obtained as the slope. The values of kqAare, then,
2632 The Journal of Physical Chemistry, Vol. 89, No. 12, 1985
Stephens
TABLE I: Summary of the Reported Quenching Measurements Showing the Ratio I:/In Alkane Concentrations, the Values Obtained for $kqA, and the Ratio of the Quenching Rate Constants Obtained to the Hard-Sphere Collision Rate Constants" concn
dk,
IPlI'
kolkmii
concn
&,
PIIf
12.3 18.7 24.9 31.2 37.3 42.9 55.2 61.5 62.3 67.5 67.7 68.3 72.4 79.8 86.9 87.1 93.1
1.05 f 0.01 1.11 f 0.03 1.09 f 0.04 1.18 f 0.02 1.12 f 0.08 1.15 f 0.03 1.33 f 0.03 1.23 f 0.06 1.16 f 0.08 1.36 f 0.04 1.27 f 0.03 1.21 f 0.05 1.35 f 0.03 1.35 f 0.04 1.36 f 0.02 1.26 f 0.06 1.39 f 0.03
1.55 3.10 4.64 6.16 7.70 9.25 10.8 13.9 15.4
1.01 f 0.03 1.04 f 0.03 1.06 f 0.02 1.13 f 0.03 1.07 f 0.03 1.06 f 0.05 1.11 f 0.09 1.23 f 0.05 1.11 f 0.06
0.91 1.83 2.74 2.75 2.98 3.65 4.18 4.48 4.57 5.37 5.49 6.34 6.56 6.81 7.75 8.40 9.53 10.2
1.01 f 0.09 1.06 f 0.04 1.11 f 0.03 1.02 f 0.04 1.05 f 0.04 1.12 f 0.02 1.06 f 0.04 1.11 f 0.04 1.18 f 0.02 1.07 f 0.03 1.18 f 0.05 1.22 f 0.05 1.09 f 0.03 1.12 f 0.09 1.07 f 0.05 1.10 f 0.09 1.27 f 0.06 1.33 f 0.06
(4.15 f 0.70) X
1.6 f 0.3
Propane
(1.17 f 0.72) X
4.0 f 2.5
Butane
(1.77 f 1.40) X
k. lkcnii
Isobutane
Ethane
4.8 f 3.8
1.09 2.54 2.54 2.91 3.29 3.31 3.31 3.91 4.09 4.40 4.83 4.83 5.41 5.49 5.50 5.82 6.21 6.56 6.57 6.60 7.40 7.64 7.71 8.03 8.59 8.60 8.78 9.29 9.71 9.73 9.87 9.98 10.7
1.03 f 0.04 1.04 f 0.03 1.09 f 0.04 1.04 f 0.04 1.09 f 0.02 0.95 f 0.09 1.06 f 0.03 1.07 f 0.03 0.97 f 0.08 1.08 f 0.03 1.06 f 0.02 1.08 f 0.03 1.03 f 0.05 1.14 f 0.04 1.13 i 0.03 1.10 f 0.04 1.11 f 0.04 1.11 f 0.03 1.16 f 0.05 1.14 f 0.03 1.18 f 0.02 1.16 f 0.7 1.17 f 0.04 1.22 f 0.08 1.18 f 0.06 1.23 f 0.04 1.18 f 0.04 1.21 f 0.03 1.22 f 0.06 1.23 f 0.05 1.25 f 0.07 1.17 f 0.04 1.29 f 0.05
(2.35 f 0.42) X
7.4 f 1.3
2.57 4.29 4.54 6.04 7.21 7.72 9.02 10.6 12.5 13.0 16.0 16.0 19.0 22.0 24.9
2,2-Dimethylpropane 1.01 f 0.07 0.99 f 0.03 1.04 f 0.02 1.07 f 0.02 1.04 f 0.04 1.11 f 0.06 1.03 f 0.10 1.14 f 0.05 0.87 f 0.59 1.08 f 0.02 1.12 f 0.04 1.14 f 0.01 1.19 f 0.02 1.20 f 0.03 1.27 f 0.03 (1.05 f 0.20) X
2.9 f 0.6
"Uncertainties listed are two standard deviations. Concentration units listed are 10'j molecules cm-3 and units of $kqA are cm3 molecule-'.
derived relative to the radiative lifetime of OH(AZZ+). Figure 2 shows the data plotted in this fashion. The values obtained in this study for IfOlI,, alkane concentration, and resultant .r/)kqA's are tabulated in Table I. Also listed is the value of the ratio of kqAto the approximate hard-sphere collision rate constant. From Table I it can be seen that this ratio exceeds one for each alkane studied. The hard-sphere collision rate constants were calculated by using an OH(A2Z+)radius13of 1.01 X 10" cm and the radius of the alkanes as calculated from gas viscosity v a 1 ~ e s . l ~ Discussion Lin et al.15 developed a model which relates quenching cross sections to the attractive forces between an excited species (A*) and quencher (M). The intermolecular well depth between these (13) G. Herzberg, 'Molecular Spectra and Molecular Structure", 2nd ed, Van Nostrand-Rheinhold, New York, 1950. (14) W. Braker and A. L. Mossman, 'Matheson Gas Data Book", 6th ed, Matheson Co., 1980. (15) H.-M. Lin, M. Seaver, K. Y. Tang, A. E. W. Knight, and C. S. Parmenter, J . Chem. Phys., 70, 5442 (1979).
=(
species was approximated by the relationship In
uqM =
In C
~
~
.
~
.
resulting t ~ ~ )in~
+
where uqMis the quenching cross section and @ = ( t A * A * ) I i z / k p. Fairchild et a1.I6 applied this relationship to the measurements of OH(A2Z+)quenching cross sections available in the literature and reported a "reasonable correlation" between the cross sections obtained at room temperature and (tMM/k)l12. Although a correlation coefficient was not reported for the data employed in their analyses, we calculate a value for r of 0.81. With the inclusion of the data reported herein, the value of r is reduced to 0.77. The use of /3 and In C obtained from the analysis by Fairchild et al. underestimated most cross sections obtained at elevated temperatures, however. Experimental evidence on the fate of the A*M collision pair postulated for OH(AZZ+)15is nonexistent. The possibility that (16) P. W. Fairchild, G. P. Smith, and D. R. Crosley, J . Chem. Phys., 79, 1795 (1983).
/
~
The Journal of Physical Chemistry, Vol. 89, No. 12, 1985 2633
Quenching of OH(A2Z+) by Alkanes and Halocarbons
-
?
/
30r
! I
/
r
I 13r
I
I
25
50
1 1
75
-
1
100
125
Figure 3. A plot of quenching rate constants calculated from eq b (k;) vs. the measured quenching rate constants ( k J .
.-
09
I
14 131
1
T
I
0
I
1
5
!
10
I
I
I
I
I
15
20
25
-1 8
Concentration ( 1 0 1 3 molecules c m - 3 )
89
Figure 2. Stern-Volmer plots; (a) ethane, note the different concentration scale for this plot only; (b) propane; (c) n-butane; (d) isobutane, error bars have been omitted for clarity; (e) 2.2-dimethylpropane. The slopes ( T P I C ~are ~ ) calculated from the weighted least-squares fit of the data. Error bars in all plots are statistical uncertainties of f 2 a .
reactive pathways exist has not been ruled out. Indeed, Zipf" has evidence that OH(A2Z+) reacts with molecular nitrogen to form nitrous oxide. One possible mechanism for reactive quenching of OH(A2Z+)by alkanes is reaction 5. If this reaction occurs, the quenching rate constants might demonstrate a dependence upon the number and type of C-H bonds present in the alkane. Such a dependence has often been observed for abstraction of hydrogen atoms by other atomic and free radical s p e c i e ~ . ' ~ J ~ For example, the abstraction of hydrogen atoms from alkanes via OH(X211) has been demonstrated to follow a dependence upon the number and type of C-H bonds.'s4 Anderson and Stephens4 have found that their OH(X211) + alkane reaction rate constant measurements can be accurately estimated by the expression
k' = [1.30Np exp(-1001/7') + 1.21Ns exp(-354/T) + (3.31NT) exp(-281/T)] cm3 molecule-' s-' (a) where Np, Ns, and NT are the number of primary, secondary, and tertiary C-H bonds, respectively. For quenching of OH(AZZ+),a similar regression analysis can be performed. However, in these experiments the temperature dependence of quenching was not determined. Hence, the regression analysis yielded
k,' = (O.llNp + 0.41Ns
+ 2.05NT) X
cm3 molecule-'
s-l
(b) Although the number of quenching rate constants measured in this study is limited to five, the correlation coefficient ( r ) between the quenching rate constants estimated via eq b, k,', and the measured rate constants, k,*, exceeds 0.99. This empirical relationship then accurately estimates the OH(A2Z+)quenching rate constants reported herein and might be useful as a model for predicting rate constants for quenching of OH(A2Z+) by other alkanes. A plot of the relationship between predicted and measured rate constants is shown in Figure 3. (17) E.C. Zipf, private communication. (18) J. M. Tcdder, Q.Rev. Chem. Soc., 14, 336 (1960). (19) J. T. Herron and R. E. Huie, J . Phys. Chem., 73, 3327 (1969).
90
91
92 9 3 94 95 96 D (C-H) in Kcal.Mole-1
97 9 8
Figure 4. A plot of the natural log of the quenching rate constants per C-H bond type, as calculated via eq b, vs. the bond dissociation energies, D(C-H), for primary, secondary, and tertiary C-H bonds.
Equations a and b demonstrate that the rate constants for both quenching of OH(A22+) and reactions of OH(X211) share a similar dependence upon the number of primary, secondary, and tertiary hydrogens. Furthermore, the relative contribution of each type of hydrogen to the rate constants follows the same ordering tertiary > secondary > primary For reactions of OH(X211) with alkanes this relationship is attributable to the relative rate of abstraction of each type of hydrogen in the alkane. There is, therefore, a high correlation between rate constants for abstraction of each type of hydrogen and their respective bond dissociation energies.IFz2 An equally high correlation exists (r = 0.99) between these bond dissociation energies23and the contribution to the quenching rate constants made by each type of C-H bond as predicted via eq b. This relationship is shown as Figure 4. Although our measurements do not provide direct evidence, the dependence of the quenching rate constants upon the type and number of hydrogens present in an alkane is suggestive of reactive quenching of the type shown in reaction 5. These results do not conflict, however, with the explanation of the role of attractive forces in quenching processes. Indeed, a reasonable correlation exists ( r = 0.75)between In uqM,as measured for the alkanes, and ( e M M / k ) ' / ' . It would be interesting and possibly worthwhile to use a larger data base with which to examine the relationship between quenching rate constants and substituents in quenching molecules. Unfortunately, the only available measurements for alkane quenchers are those reported herein and that of methane at elevated temperatures.I6 However, a number of quenching rate constants are available for one and two carbon halogenated alkanes as measured by Cylne and Holt.' To examine the possible relationship between substituents in these chlorofluorocarbons and the rate constants for OH(A22+)quenching, a regression analysis was again performed. For the most part, the measured quenching R. Atkinson, Inr. J . Chem. Kinet., 12, 761 (1980). J. Heicklen, In?. J . Chem. Kiner., 13, 651 (1981). J. S. Gaffney and S. 2.Levine, h r . J . Chem. Kine?.,11, 1197 (1979). S. W. Benson, 'Thermochemical Kinetics: Methods for the Estimation of Thermochemical Data and Rate Parameters", Wiley, New York, 1968. (20) (21) (22) (23)
2634 The Journal of Physical Chemistry, Vol. 89, No. 12, 1985 TABLE II: Comparison of the Rate Constants (k,') Calculated via Eq c and the Values Measured (k,) by Clyne and HolP
k, 10.84 10.53 6.63 1.16 2.21 8.00 8.63 3.05 10.11 7.26 5.16 6.32 4.1 1 2.53 2.63 5.05 4.63 4.53
compd CCI4 CFC13 CF2C12 cF4 CHF, CHFZCI CHFC12 CH2F2 CHJCCII CH9CF2CI CH3CF2H CF9CHC12 CF3CH2C1 CF,CH, CF3CF2H CFjCFH, CFZHCFZH CF2HCFH2
k,'
k,'lk,
11.43 9.04 6.66 1.89 2.32 4.72 7.1 1 2.78 11.32 6.55 4.61 8.05 6.1 1 4.17 3.28 3.72 3.72 4.17
1.05 0.86 1.oo 1.63 1.06 0.59 0.82 0.91 1.12 0.90 0.89 1.27 1.49 1.65 1.25 0.74 0.80 0.92
-22
-In k,' (predl
20.639 20.673 21.134
20.589 20.824 21.130 21.574 20.724 21.375 20.600 21.300 21.130 20.775 20.998
20.638
D(C-CI) 70.3 74.0 82.6 86.2 71.0 83,O 70.0 83.6 82.8 76.2 78.0
ref 33 36 36 36 35 37 38 33 33 35 34
+
Table I1 compares the values calculated by the above expression to those experimentally measured. The correlation coefficient ( r ) between predicted and measured values is 0.89. The extremely good agreement between the rate constants approximated by eq c and the measured values strongly suggests that this relationship could be reliably employed to predict rate constants for OH(A2Z+)quenching by other halogenated alkanes. It is interesting to note that a similar empirical relationship has been previously reported for quenching of O(lD) by chlorof l u o r o c a r b o n ~ ,where ~ ~ quenching occurs to a large degree by chemical r e a ~ t i o n . ~ ~ - ~ l Although no direct evidence is available for reactive quenching of OH(A2Z+)by chlorofluorocarbons, reactions analogous to those of O(lD) are thermodynamically possible. For example, the reaction OH(A2ZC) + CF,C12 HOC1 C F 2 C l (6) is approximately 71 kcal mol-' exothermic.
-
0 75
0 80 D (C-CI)
85
90
Figure 5. A plot of the natural log of the rate constants for quenching of OH(A22+) by halocarbons vs. the C-CI bond dissociation energies. Solid data points are values of k,' as calculated via eq c and the open points represent the values of k, measured by Clyne and Holt.'
rate constants could be accurately represented by the following simple expression: k,'(C,F,HbCl,) = ( 0 . 4 7 ~ 0.926 2 . 8 6 ~ X) cm3 molecule-' s-' (c)
+
2
70
TABLE 111 Compilation of C-Cl Bond Dissociation Energies, in kcal mol-', along with the Respective Measured and Predicted Rate Constants for Quenching of OH(A22+)
-In k. (meas)
i
-
"Reference 7. Rate constant units are loTiocm3 molecule-I s-'.
compd CCI, CFCI, CF2C12 CF,CI CF$CI, CF,CF2CI CH3CCIj CH$l CH2C12 CHCl, CClF2CF2Cl
Stephens
+
(24) J. A. Davidson, H. I. Schiff, T. J. Brown, and C. J. Howard, J. Chem. Phys., 69,4277 (1978). (25) H. M. Gillespie and R. J. Donovan, Chem. Phys. Lett., 37, 468 (1976). (26) A. P. Force and J. R. Wiesenfeld, J. Phys. Chem., 85, 782 (1981). (27) I. S. Fletcher and D. Husain, J . Phys. Chem., 80, 1837 (1976). (28) M. C. Addison, R. J. Donovan, and J. Garraway, Discuss. Faraday Soc., 67,286 (1979). (29) J. A. Pitts, Jr., H. L. Sandoval, and R. Atkinson, Chem. Phys. Lett., 29,31 (1974). (30) R.Atkinson, G. M. Breuer, J. A. Pitts, Jr., and H. L.Sandoval, J . Geophys. Res., 81, 5765 (1976). (31) R. K. M. Jayanty, R. Simonaitis, and J. Heicklen, J . Photochem., 4, 203 (1975).
If reaction 6 is the predominant mechanism for OH(A2Z+) quenching by halocarbons, the rate constants would be expected to correlate with C-Cl bond dissociation energies. Indeed, a high correlation exists (r = 0.96) between the quenching rate constants and the available C-C1 bond dissociation energies (see Table 111). A plot of this relationship is shown in Figure 5. The presence of OH(A2Z+)in the atmosphere makes chlorine abstraction reactions analogous to ( 6 ) possible sinks for chlorofluorocarbons. These reactions can be shown to be unimportant, however. The potential impact of such reactions can be crudely estimated from the approximate OH(A2Z+)production rate and the known loss processes for this radical in the atmosphere. It is estimated that, at an altitude of 50 km, the resonance absorption of UV photons results in the production of OH(A2Z+)at a rate of 2 X lo4 molecules cm-3 s-I.l7 If one assumes steady-state conditions in the daytime stratosphere and assumes the major loss processes for OH(A2Z+) at 50 km to be radiative decay and quenching by nitrogen and oxygen, then the concentration of OH(A2Z+) can be determined by the following equation: [OH*],, = 2 x lo4 molecules cm-3 s - ' / ( k , + kqN2[N21+ kqo2[02l) If the rate constants for quenching and radiative decay as given by Schofield are used,8 this yields a steady-state OH(A2Z+) concentration of 9 X 10-3molecules cm-3 at 50 km. A lower limit on the chlorofluorocarbon lifetimes in the atmosphere due to reaction with OH(A22+) can now be estimated by using the measured quenching rate constants as upper limits for reaction rate constants. For CF2C12,this yields a minimum atmospheric lifetime of approximately 38 000 years due to reaction 6 , if it occurs. At an estimated lifetime of 120 years38 this represents only an additional 0.3%loss process for this chlorofluorocarbon. Analogous reactions for the remaining chlorofluorocarbons can also be shown to be unimportant as atmospheric loss processes for these molecules. Although the speculative reactions of OH(A2Z+) with halocarbons (and alkanes) are of no atmospheric importance, they represent potentially interesting reactions for which no information has been previously available in the literature. The measured rate constants for quenching of OH(A22+) by alkanes reported herein are the first available for these compounds. The values obtained are noteworthy due to their magnitude. Propane, n-butane, isobutane, and dimethylpropane have values larger than any quenching rate constants previously reported for OH(A2Z+) de(32) M. Welssman and s. w. Benson, J. Phys. Chem., 87, 243 (1983). (33) R. Foon and K. B. Tait, J. Chem. Soc., Faraday Trans. I , 68, 1121 (1972). (34) G. Baruch, L. A. Raibenbach. and A. Horowitz. fnt. J . Chern. Kinet.. 13, 473 (1981). (35) E. Illenberger, H.-U. Scheuemann, and H. Baumgartel, Chem. Phys., 37,21 (1979). (36) J. W. Coomber and E. Whittle. Trans. Faraday Soc... 63.. 2656 (1967 j. (37) J. A. Franklin and G. H. Huybrechts, f n t . J . Chem. Kinef., 1, 3 (1969). (38) "Causes and Effects of Changes in Stratospheric Ozone: Update 1983", National Academy Press, Washington, DC, 1984.
J. Phys. Chem. 1985,89, 2635-2638 spite the large values typical for quenching processes. Also noteworthy is the apparent dependence of quenching rate constants upon the type of substituents in the quenching molecules. The relationships described should be useful for predicting rate constants for quenching of OH(A2Z+) by alkanes and haloalkanes where measurements are not available.
Acknowledgment. The author thanks Dr. Jerry Rogers and Patricia Korsog for the time and assistance which they offered
2635
during many helpful discussions. Regisby NO. OH, 3352-57-6; CCId, 56-23-5; CFCls, 75-69-4; CF2CI2, 75-71-8; CF4, 75-73-0; CHF,, 75-46-7; CHF,Cl, 75-45-6; CHFC12, 7543-4; CHZF2,75-10-5; CH,CCl,, 71-55-6; CH3CF2C1,75-68-3; CH3CFzH, 75-37-6; CF3CHC12, 306-83-2; CF3CH2C1, 75-88-7; CF,CH,, 420-46-2; CF~CFZH, 354-33-6; CFPCFHZ,8 11-97-2; CFZHCFZH, 35935-3; CF2HCFH2, 430-66-0; isobutane, 75-28-5; 2,2-dimethylpropane, 463-82-1; ethane, 74-84-0; propane, 74-98-6; butane, 106-97-8.
Production of OH on Polycrystalline Nickel Studied by Thermal Desorption/Laser- I nduced Fluorescence J. T. Keiser,f M. A. Hoffbauer, and M. C. Lin* Chemistry Division, Naval Research Laboratory, Washington, D.C. 20375-50000 (Received: December 10, 1984)
The production of hydroxyl radical following either the reaction of H2/02mixtures or the dissociationof H 2 0on polycrystalline nickel has been studied in a flow system. The hydroxyr radicals were detected in the gas phase by laser-induced fluorescence (LIF) following desorption from the catalyst surface at temperatures above 850 K. The apparent activation energy for OH desorption from nickel varied from 26 to 40 kcal/mol depending upon the O/Hratio. The effect of the partial pressure of O2and H2 on the OH production rate was measured and a mechanism to explain these results is proposed. The reaction of CO H2over polycrystallinenickel was also studied. However, no OH radicals were detected desorbing from the surface at temperatures up to 1350 K.
+
Introduction There is considerable interest in the interactions of 0 2 , H2, and H 2 0 with nickel surfaces and particularly with regard to the presence of a Ni-OH species.’-” It is known that water will absorb on nickel a t room temperature only in the presence of oxygen.6*1° Most researchers have attributed this to the formation of some type of OH surface species. However, two oxygen 1s peaks are often observed in the XPS spectra when O2 and H 2 0 are coadsorbed on nickel, implying two types of oxygen are present. This has been explained by assuming either the presence of a “defect” nickel oxide2 or partial bonding of water to an absorbed ~ x y g e n .Recently, ~ Benndorf et al. have reported observing only a single oxygen 1s peak by XPS and suggested that earlier measurements of an oxygen doublet were caused by a mixture of 0, O H , and H 2 0 surface species.1° Several groups have detected Ni,(OH)+ and OH- fragments from 02/H2/Ni systems by SIMS.S*8*11 Benninghoven et al. have studied the reactions of O2and H2with polycrystalline nickel by S I M S and thermal desorption/mass spectrometry.” OH was detected by SIMS following the reaction of H2 with oxygen covered polycrystalline nickel. Upon heating, the desorption of water and a loss of the S I M S OH- signal was observed. Interestingly, the temperature at which this process occurred depended upon the degree of oxidation of the surface. This was presumed to be due to variations in the Ni-OH bond strength. The reactions ok CO and H2over nickel are very important since this reaction is of industrial importance in the production of Of the several mechanisms which have been proposed for the methape synthesis reaction, none involve an Ni-OH intermediate.I2 Low-temperature studies of coadsorbed CO and H2on nickel to date have revealed no OH bonds by EELS or HREELS.’7.18 However, at reaction temperatures both CO and H2 are known to be. dissociated on nickel surfaces and both 0 and H are present on the surface. It is therefore possible that the Ni-OH species is present in this system and may be an in-
’
Present address: Chemistry Department, University of Richmond, Richmond, VA 23173.
termediate in a competing reaction, i.e., the production of water.” In the present work we have utilized thermal desorption/laser-induced fluorescence (TD/LIF) to investigate the production of OH from polycrystalline nickel. Previously we have used this method to observe the desorption of O H from polycrystalline platinum and to study its production rate, apparent activation energy of desorption, and energy a c c ~ m m o d a t i o n . ~ ~This -~~ (1) FOPa recent review see: Norton, P. R. “The Chemical Physics of Solid Surfaces and Heterogeneous Catalysis”,Vol. 4, King, D. A., Woodruff, D. P., Eds.; Elsevier: Amsterdam, 1982. (2) Norton, P. R.; Tapping, R. L.; Goodale, J. W. Surf. Sci. 1977,65, 13. (3) Benndorf, C.; Nobl, C.; Thieme, F. Surf. Sci. 1982, 121, 245. (4) $endulic, K. D.; Winkler, A. Surf. Sci. 1978, 74, 318. (5) Hopten, H.; Brundle, C. R. J . Vac. Sci. Technol. 1979, 16, 548. (6) Benndorf, C.; Nobl, C.; Rusenberg, M.; Thieme, F.Surf. Sci. 1981, 111, 87. (7) Benndorf, C.; Nobl, C.; Rusenberg, M.; Thieme, F. Appl. Surf. Sci. 1982, 11/12, 803. (8) Bourgeois, S.;Perdereau, M. Surf. Sci. 1982, 117, 165. (9) Benndorf, C.; Nobl, C.; Thieme, F. Surf. Sci. 1982, 121, 249. (10) Benndorf, C.; Nobl, C.; Madey, T. E. Surf. Sci. 1984, 138, 292. (11) Benninghoven, A.; Beckmann, P.; Muller, K. H.;Schemmer, M. SurJ Sci. 1979, 89, 701. (12) Somorjai, G. A. Catal. Rev. Sei. Eng. 1981, 23, (1+2), 189. (13) Erley, W.; Ibach, H.; Lehwald, S.;Wagner, H. Surf. Sci. 1979,83, 585. (14) Rabo, J. A.; Risch, A. P.; Poutsma, M. L. J . Carol. 1978, 53, 295. (15) Underwood, R. P.; Bennett, C. 0. J . Curd. 1984, 86, 245. (16) Goodman, D. W. J. Vac. Sci. Technol. 1982, 20, 522. (17) Mitchell, G. E.; Gland, J. L. Surf. Sci. 1983, 131, 167. (18) White, J. M. J . Phys. Chem. 1983, 87, 915. (19) Kelley, R. D.; Goodman, D. W. Surf. Sci. 1982, 123, L743. (20) Klose, J.; Boerns, M. J . Cutul. 1984, 85, 105. (21) Labohm, F.; Gijzeman, 0. L.; Geus, J. W. Surf. Sci. 1983, 135,409. (22) Pannell, R. B.; Chung, K. S.;Bartholomew, C. H. J . Cutal. 1977,46, 340. (23) Tevault, D. E.; Talley, L. D.; Lin, M. C. J . Chem. Phys. 1980, 72, 3314. (24) Talley, L. D.; Sanders, W. A.; Bogan, D. J.; Lin, M. C. J . Chem. Phys. 1981, 75, 3107. (25) Talley, L. D.; Lin, M. C. AIP Conf. Proc. 1980, 61, 297. (26) Talley, L. D.; Lin, M. C. ‘Proceedings of the International Conference on Lasers 79”, Corcoran, V. J., Ed.; STS: Virginia, 1980; p 297.
This article not subject to US.Copyright. Published 1985 by the American Chemical Society
