- Email: [email protected]
Rate constants for the reaction of CF3O radicals with hydrocarbons at 298 K
Volume 207, number 4,5,6
CHEMICAL PHYSICS LETTERS
28 May 1993
Rate constants for the reaction of CF30 radicals with hydrocarbons at 298 K Christina Kelly, Jack Treaty, Howard W. Sidebottom ’ Department of Chemistry,UniversityCollege Dublin, Dublin, Ireland
Ole John Nielsen ’ ChemicalReactivity Section, ES&T Department, Rise National Laboratory DK-4000 Roskilde, Denmark Received 18 February 1993
Rateconstant ratios of the reactions of CFsO radicals with a number of hydrocarbons have been determiued at 298 + 2 K and atmospheric pressure using a relative rate method. Using a previously determined value k(CF,O+ CsHs) = I .2 x lo-i2 cm3 molecule-’ s-i these rate constant mtios,provide estimates of tbe rate constants: k(CFsO+CH,)= (1.2kO.l) x 10-14, k(CF,O+ c-C,He,)=(3.6~0.2)x1013, k(CF,O+C,Hs)=(4.7+0.7)~10‘*, k(CF~O+(CHS),CH)=(7.2-f0.5)x10-I*, k(CF,O+ C~H4)=(3.0~0.1)~10-1’andk(CF,O+C~)=(3,6~0.1)~lO-L’cm3mol~ule~’~~‘.Theimportanceofthereactionsof CFsO radicals with hydrocarbons under atmospheric conditions is discussed.
1. Introduction
CF3O2 t NO+ CFS0 t NO;, ,
The catalytic destruction of ozone in the stratosphere by chlorine atoms is now well established [ 11. In order to reduce the concentration of chlorine-containing compounds diffusing into the stratosphere it has been proposed that fully halogenated alkanes should be substituted by hydrofluorocarbons (HFCs) and hydrocblorofluorocarbons (HCFCs). These compounds are, at least in part, oxidized in the troposphere via reaction with OH radicals, and hence their environmental impact depends on the oxidation products of the radical species produced following the reaction with OH radicals. The available evidence suggests that CFS radicals are produced as reaction intermediates in the atmospheric oxidation of compounds containing CF3groups [ 2,3] _In the atmosphere the CFJ radicals will react with oxygen and consecutively with NO [ 2,4] : CF,+Og SM+CF,O,+M
,
’ To whom correspondence should be addressed.
498
(1)
(2)
to give the corresponding alkoxy radical, CF,O. The fate of the CF90 radical under atmospheric conditions is unclear. Formation of CF20 by either unimolecular decay with loss of a fluorine atom, CF30-+CFz0+F
(A&E +24 kcalmol-‘),
(3)
or by reaction with 02,
(A&= t 13 kcalmol-‘)
,
(4)
is improbable, since both reactions are appreciably endothermic [ 5,6]. A number of experimental studies on the reactions of CF30 radicals have recently been reported. Li and Francisco [ 7 ] observed laser-induced fluorescence from the radical following infrared multiphoton dissociation of CF300CF3 and showed that the decay rate of CF30 increased after addition of NO to the system. Product analysis by infrared spectroscopy indicated that FNO was a product of the reaction between CF30 and NO. Chen et al. [ 81 investigated
0030-4018/93/S 06.00 (D 1993 Elsevier Science Publishers B.V. All rights reserved.
CHEMICAL PHYSICS LETTERS
Volume 207, number 4,5,6
the photolysis of mixtures containing CFSNO and NO in 0,/N, mixtures at about 1 atm total pressure. Analysis by long-path FTIR spectroscopy showed that CF20, PNO and NO2 were formed in stoichiometric amounts based on the loss of CFJNO. The authors rationalized their observations in terms of reaction ( 1) and (2 ) and the reactions: CF3NOthv(12>400nm)-CFS+N0,
(5)
CF,O+NO-CF,O+FNO.
(6)
Chen et al. [ 9 ] have also found evidence that the CF30 radical can abstract a hydrogen atom from alkanes CF,O+RH+CF,OH
t R.
(7)
Photolysis of CF,NO/NO mixtures in air in the presence of ethane or propane gave CF,OH and acetaldehyde or acetone, respectively. As previously reported [ 10] CF30H was found to be unstable and decomposes to give CF,O and HF. Recent work by Soathoff and Zellner [ 111, Sehested and Wallington [ 121 and by Bevilaqua et al. [ 131 have provided further evidence that H atom abstraction by CF,O radicals is a relative facile process. The purpose of the present work was to further investigate the kinetics and mechanism for the reactions of CFJO radicals with hydrocarbons. The results provide information on the importance of these reactions in the atmospheric chemistry of CFsO radicals.
2. Experimental Kinetic and product analysis studies were carried out at 298 f 2 K and atmospheric pressure in an approximately 50 I FEP teflon cylindrical reaction chamber. The chamber was surrounded by 20 germicidal lamps (Phillips TUV 15W) providing UVradiation with maximum intensity around 250 nm. Light intensity was varied by switching off various groups of lamps. A uniform reaction temperature was achieved by the use of electric fans positioned below the reaction chamber. All pressure measurements were made using MIC!SBaratron capacitance manometers. Measured amounts of the reactants were flushed from calibrated Pyrex bulbs into the teflon
28 May 1993
reaction chamber by a stream of zero-grade nitrogen (Air Products), which was then filled with either zero-grade nitrogen or ultra-pure air (Air Products). Reaction mixtures were allowed to mix for at least 30 min prior to photolysis. Quantitative analyses were carried out using gas chromatography (Gow Mac series 750 gas chromatograph equipped with a flame ionization detector) and PTIR spectroscopy (Mattson, Galaxy 5000, 0.5 cm-’ resolution). Gas tight syringes (Hamilton) or a Valco gas sampling valve were used to remove samples of the reaction mixtures from the teflon bag for chromatographic analysis. FTIR spectra were obtained using an evacuable 2 Lteflon-coated Wilks cell, containing a multipass White mirror arrangement (10 m path-length), mounted in the cavity of the spectrometer. The reaction mixtures were expanded into the cell through 3 mm inner diameter teflon tubing, after various periods of photolysis. CFsOOCFs ( > 99%, Fluorochem Ltd. ) was further purified by vacuum distillation. Hydrocarbon gases were research grade products from Matheson and were used as received. Benzene ( > 99%, Aldrich) was trap to trap distilled prior to use.
3. Results and discussion Photolysis of lo-60 ppm CF300CF3 ( 1 ppm =2.46X 1013molecule cmm3 at 298 K) in air or nitrogen at atmospheric pressure for 5 h resulted in about a 15% depletion of reactant and to the formation of stoichiometric amounts of CF,O. CF,O was stable in the dark for up to about 24 h and photolysis at 250 nm was negligible over the time period of the present experiments. In the presence of a hydrocarbon, the loss of CF300CF3 increased siguificantly, x30% in 1 h. For reactions carried out in air, the products detected by FTIR spectroscopy were CF,O and carbonyl compounds. In the initial stages of photolysis of CFSOOCF3 in the presence of isobutane in air, the yields of CF,O and CH3COCH3 corresponded to the loss of CFsOOCF,, ACF,O/ and ACH&OCHJ -ACF300CF3 = 2.12 0.3 - ACF,OOCF, = 2.0 f 0.2. At longer reaction times CH&OCH3 was lost by photolysis. Similar results were obtained from the photolysis of CFiOOCF3 with ethane and propane, where the carbonyl compounds 499
were acetaldehyde and acetone, respectively. The results are in agreement with Chen et al. [ 91, who also have provided evidence for the abstraction of hydrogen atoms from C2H6 and C3Hs by CFjO radicals. These results are consistent with the following reactions: CF300CF3+hv(l~250nm)-r2CF30,
63)
2CF~O-+CF300CF3,
(9)
CF,O+wall-+CF,O, CF,O+ (CH3)&H+CF30H+
(10)
(CHJ)&,
(11)
CF,OH-CF,O+HF,
(12)
(CH&C+02
(13)
tM+(CHs),C02tM,
2(CH,)sCOz-+2(CH9)3COt02,
(14)
(CH3)&O-+CH3COCH~ +CH3.
(15)
Combination of CF,O radicals in reaction (9) to reform the peroxide, appears to be the major reaction channel for these radicals after photolysis of CF300CF3 in air or nitrogen. The small amounts of CF20 detected are likely to arise from heterogeneous loss of CFsO at the chamber walls. Reaction of CF30 radicals with alkanes presumably involves a hydrogen abstraction process with the alkyl radical, (CH3) $ being oxidized in air to form acetone. In our system CFsOH was not observed directly, but decomposed to give C&O and HF. Kliiter and Seppelt [lo] have previously shown that the condensed phase CF90 decomposes rapidly to form CFzO and HF. However, in the gas phase at 12 Torr, the lifetime appeared to be somewhat longer with a half-life of at least 5 min. From their gas-phase experiments at 297 K Chen et al. [ 81 observed a lifetime for CFsOH of about 12 min. Decomposition of CF30H to give CFzO and HF is estimated to be approximately 5 kcal mol-’ exothermic [ 5,6]. In a recent theoretical treatment on the unimolecular decay of CF,OH, Francisco 1141 showed that the reaction had an appreciable barrier of around 50 kcal mol- I, suggesting that the loss of CFBOH observed by KlSter and Seppelt [ lo] and by Chen et al. [ 91 is probably surface catalyzed. In the present experiments the long photolysis times required to obtain measurable decays of CFJOOCF, precludes detection of CFBOH 500
28 May 1993
CHEMICAL PHYSICS LETTERS
Volume 207, number 4,5,6
with the result that CFzO is the observed product of the reaction of CFjO with alkanes. Relative rate constants for the reaction of a number of hydrocarbons with CF90 radicals were determined by comparison of the decay of the reactant hydrocarbon and that of a reference compound; reaction (7) and CF,O t reference+products .
(16)
Assuming that reaction with CFIO is the only significant loss process for both the reactant and the reference compound, it can be shown that In(@)=
2
In(s),
(17)
where the subscripts 0 and t indicate concentrations at the beginning of the experiment and at time t, respectively. Under the experimental conditions employed photolytic loss of the hydrocarbon was unimportant and the reaction mixtures were stable in the dark for at least 24 h. Photolysis of CF,OOCF,/ hydrocarbon mixtures were carried out in either air or nitrogen at 298+2 K and 1 atm total pressure; [CF,OOCF,],,= lo-60 ppm, lRH],= l-10 ppm. Concentration-time data for the hydrocarbon reactions investigated are plotted in the form of eq. ( 17 ) in figs. 1 and 2. The rate constant ratios k,/k,,, where CzH6was the reference compound, are given in table 1. These ratios were independent of reaction time, relative reactant concentration and light intensity in agreement with the proposed mechanism. At least five individual runs were carried out for each substrate and in order to test internal consistency of the rate constant ratios, each compound was run against another member of the series. In all cases the relative rate values were in excellent agreement with each other. Values of the rate constant ratios determined for each reactant pair were equal within experimental error for experiments run both in air and nitrogen. The eventual products arising from the alkyl radicals formed in these CFsO initiated reactions would be expected to vary depending on whether air or N2 was the bath gas. Thus complications arising from either product interference in the analytical procedure or enhanced removal of one of the reactant components due to product radical reactions are unlikely to be significant. It is possible that some loss
Volume207, number 4,5,6
CHEMICALPHYSICSLETTERS
28 May 1993
S.OE-2
7.0
0.0 4.OE-2
5.0 h ,' g 4.0
h +XOE-2
? ,o T 62 3.0 ==' I
5
K
z z Z.OE-2 s
2.0
1.0
0.0 0.00
0.05
0.25
0.30
Fig. 1. Concentration-time data for the reaction of CF,O radicals with various hydrocarbons: (0) CJH,; (0) i-C&,; (A)
0.00
0.05
0.10
0.15
0.20
0.25
Fig. 2. Concentration-time data for the reaction of CF,O radicals with (0) CH, and (0) c-C,H,.
Cd&; (0) Cd%.
of hydrocarbon in our relative rate studies may arise from reactions of F atoms, which may be formed in heterogeneous reactions at the chamber wall. However, the degree of selectivity shown is inconsistent with a F atom loss process for the hydrocarbon. Furthermore, the addition of up to 100 ppm of CH4 did not change the rate constant ratios obtained for a pair of hydrocarbons. Addition of CH4 to the system would be expected to scavenge F atoms, k(F+ CH4)=8~10-11 cm3molecule-‘s-r [16]. The rate constants for reaction of CF,O radicals with the hydrocarbons studied relative to the rate constant for reaction with ethane are given in table 1. Chen et al. [ 91 have previously reported the rate constant ratio k( CF,O+ CjHI) /k( CF,O t C2HQ)= 2.5 in reasonable agreement with our result. The rate constant ratios determined in this work for CF,O reactions are compared to the values for the corresponding OH radical reactions in table 1. The data show that CF,O radicals have about the same degree
of selectivity as the OH radical and thus may be expected to have the same level of reactivity. This rather surprising result is consistent with a recent laser flash photolysis - laser-induced fluorescence study by Soathoff and Zellner [ 111 on the reaction of CF30 radicals with C,H6, for which a value of the rate constant,k(CF,0tC2Hb)=(1.2+0.2)X10-’*cm3 molecule-’ s-t was obtained. Combining this rate constant with the rate constant ratios determined in this work provides estimates of the rate constants for the reaction of CF30 radicals with the series of hydrocarbons investigated, table 1. Bevilaqua et al. [ 13] employed a flow reactor coupled to a chemical ionization mass spectrometric detector to study the reaction of CF,O radicals with isobutane. The preliminary estimate of k(CF,O+fC,H,,) = (5 It 3) x 1O-'*cm3 molecule-’ s-’ is in excellent agreement with the rate constant determined from the present work. Rate constants per abstracted H atom from CH,, primary, secondary and tertiary sites in hydrocar501
Volume 207, number 4,5,6
CHEMICAL PHYSICS LETTERS
28 May 1993
Table 1 Rate constant data for the gas-phase reactions of CF,O radicals with hydrocarbons at 298 + 2 K and 1 atm total pressure Compound
k(CF,O+RH)/L(CF,O+CIHs)
a)
lO%(CF,O+RH)
0.010~0.001 0.30 + 0.02 1.0 3.9io.5 6.0f0.4 2551 30+1
CH4
c-cd6
WI GHs (CH&CH W4
(36
0.012+0.001 0.36 + 0.02 1.2 4.7kO.7 7.2f0.5 30+1 36kl
b,
k(OH+RH)/k(OH+CIH6)
‘)
0.031 0.26 1.0 4.3 8.1 32 4.6
‘) Errors quoted represent + 2a and represent precision only. b, Calculated using k(CF,O+C,HJ = (1.2* 0.2) x lo-” cm’ molecule-’ s-’ [ 111, errors in this reference rate constant are not included. ‘) Evaluated data taken from ref. [ 151.
bons can be derived from the rate data in table 1: p=&zl$
kW=1~&B-k2~6);
kWt=
z C2H6_ As expected for H atom abstrac(Cli3 ) 3CH tion reactions the rate constants for reaction of CF30 radicals with alkanes correlate well with the C-H bond strengths [ 171, see fig. 3. The high reactivity of both CzH4 and C6H6 towards reaction with CFJO
-20
-23
l&-25 Yb *c
-29
-32
-35
90
95
100
D
105
110
(kcal/mol)
Fig. 3. Rate constants per abstractable H of the specified type as a function of the C-H bond strength.
502
indicates that addition to the unsaturated system is the major reaction channel in these systems. It is interesting to compare the rate constants for reactions of CFaO radicals with alkanes to those for the corresponding CH,O reactions. At 298 K the rate constant for reaction of CF,O with ethane is over 5 orders of magnitude higher than that for CH,O, k(CH30tCZH,)=2.5x lo-” cm3 molecule-’ s-l [ 181. This result can be rationalized in terms of strengthening of the O-H bond following fluorine substitution of the methyl group in CH,OH, D(CF,O-H) ~109 kcal mol-’ [5] and D(CH,OH) = 104 kcal mol-’ [ 161. It is surprising however that the CF,O radical is slightly more reactive towards hydrocarbons than the OH radical despite the fact that the reactions are about 10 kcal mol- ’less exothermic. Presumably, the high level of reactivity of the CF30 radical is a consequence of both the strength of the O-H bond in CF,OH and the high electronegativity of the radical. As previously discussed, unimolecular decay with the loss of a F atom or reaction with O2 are unlikely to be important loss processes for CF30 in the atmosphere. Reaction of CFJO with oxides of nitrogen or hydrocarbons are possible sinks for CFJO in the troposphere. The kinetic data obtained in this work, and in that of Chen et al. [ 9 1, Soathoff and Zellner [ 111 and Bevilaqua et al. [ 13 1 indicate that reaction with hydrocarbons may be important. Bevilaqua et al. [ 131 have determined a value of the rate constant for the reaction of CF30 with NO, k(CFsO+NO) = (2+ 1)X lo-” cm3 molecule-’ s-l. Thus, aspreviously suggested by Chen et al. [9], at the hydrocarbon to NO mixing ratios typically found in the
Volume 207, number 4,5,6
CHEMICAL PHYSICS LETTERS
troposphere, the reaction with hydrocarbons will be a major sink for CFBO radicals giving CFSOH. Although CF30H is relatively unstable under laboratory conditions, the available data indicate that the decay may be mainly due to heterogeneous processes. It seems likely, therefore that CFsOH loss in the troposphere will largely take place at aerosol surfaces or by hydrolysis in cloud water.
Acknowledgement Thanks are due to Dr. C. Howard, Dr. T. Wallington, and Dr. R. Zellner for communicating their data prior to publication. The authors are grateful to AFEAS and the Commission of the European Communities for financial support.
[ 1] ES. Rowland, Ann. Rev. Phys. Chem. 42 ( 1991) 73 I. [2] Scientific Assessment of Stratospheric Ozone, World Meteorological Organization Global Ozone Research and Monitoring Project, Report No. 20, Vol. 2, Appendix: AFEAS Report (1989) ch. 6.
28May 1993
[ 31 Scientific Assessment of Stratospheric Ozone, World Meteorological Organization Global Ozone Research and Monitoring Project, Report No. 25 (1991) ch. 5. [ 41 A.M. Dognon, F. Caralp and R. Lesclauz, J. Chim. Phys. 82 (1988) 349. [ 5 ] L. Batt and R. Walsh, Intern. J. Chem. Kinetics 14 ( 1982) 933; 15 (1983) 605. [6] S.W. Benson, Thermochemical kinetics, 2nd Ed. (Wiley, NewYork, 1976). [ 71 Z. Li and J.S. Francisco, Chem. Phys. Letters 1a6 ( 1991) 336. [S] J. Chen, T. Zhu and H. Niki, J. Phys. Chem. 96 (1992) 6115. [9] J. Chen, T. Zhu, H. Niki and G.J. Mains, Geophys. Res. Letters 19 (1992) 2215. [IO] G. Kliiter and K. Seppelt, J. Am. Chem. Sot. 101 (1979) 347. [ I 1 ] H. Soathoff and R. Zcllner, Chem. Phys. Letters 206 ( 1993) 349. [ 121J. Sehested and T.J. Wallington, Environ. Sci. Technol. 27 (1993) 146. [ 131T.J. Bevilaqua, D.R. Hanson and C.J. Carlton, J. Phys. Chem., in press. [ 141J.S. Francisco, Chem. Phys. 19 ( 1991) 150. [IS] R. Atkinson, J. Phys. Chem. Ref. Data, Monograph 1 (1989). [ f6] W.B. DeMore, S.P. Sander, D.M. Golden, M.J. Molina, R.F. Hampson, M.J. Kurylo, C.J. Howard and A.R. Ravishankara, JPL Publication 961 (1990). [ 17] D.F. McMillen and D.M. Golden, Ann. Rev. Phys. Chem. 87 (1983) 1812. [ ! 8 ] W. Tang and R.F. Hampson, J. Phys. Chem. Ref. Data I 5 (1986) 1087.
so3
CHEMICAL PHYSICS LETTERS
28 May 1993
Rate constants for the reaction of CF30 radicals with hydrocarbons at 298 K Christina Kelly, Jack Treaty, Howard W. Sidebottom ’ Department of Chemistry,UniversityCollege Dublin, Dublin, Ireland
Ole John Nielsen ’ ChemicalReactivity Section, ES&T Department, Rise National Laboratory DK-4000 Roskilde, Denmark Received 18 February 1993
Rateconstant ratios of the reactions of CFsO radicals with a number of hydrocarbons have been determiued at 298 + 2 K and atmospheric pressure using a relative rate method. Using a previously determined value k(CF,O+ CsHs) = I .2 x lo-i2 cm3 molecule-’ s-i these rate constant mtios,provide estimates of tbe rate constants: k(CFsO+CH,)= (1.2kO.l) x 10-14, k(CF,O+ c-C,He,)=(3.6~0.2)x1013, k(CF,O+C,Hs)=(4.7+0.7)~10‘*, k(CF~O+(CHS),CH)=(7.2-f0.5)x10-I*, k(CF,O+ C~H4)=(3.0~0.1)~10-1’andk(CF,O+C~)=(3,6~0.1)~lO-L’cm3mol~ule~’~~‘.Theimportanceofthereactionsof CFsO radicals with hydrocarbons under atmospheric conditions is discussed.
1. Introduction
CF3O2 t NO+ CFS0 t NO;, ,
The catalytic destruction of ozone in the stratosphere by chlorine atoms is now well established [ 11. In order to reduce the concentration of chlorine-containing compounds diffusing into the stratosphere it has been proposed that fully halogenated alkanes should be substituted by hydrofluorocarbons (HFCs) and hydrocblorofluorocarbons (HCFCs). These compounds are, at least in part, oxidized in the troposphere via reaction with OH radicals, and hence their environmental impact depends on the oxidation products of the radical species produced following the reaction with OH radicals. The available evidence suggests that CFS radicals are produced as reaction intermediates in the atmospheric oxidation of compounds containing CF3groups [ 2,3] _In the atmosphere the CFJ radicals will react with oxygen and consecutively with NO [ 2,4] : CF,+Og SM+CF,O,+M
,
’ To whom correspondence should be addressed.
498
(1)
(2)
to give the corresponding alkoxy radical, CF,O. The fate of the CF90 radical under atmospheric conditions is unclear. Formation of CF20 by either unimolecular decay with loss of a fluorine atom, CF30-+CFz0+F
(A&E +24 kcalmol-‘),
(3)
or by reaction with 02,
(A&= t 13 kcalmol-‘)
,
(4)
is improbable, since both reactions are appreciably endothermic [ 5,6]. A number of experimental studies on the reactions of CF30 radicals have recently been reported. Li and Francisco [ 7 ] observed laser-induced fluorescence from the radical following infrared multiphoton dissociation of CF300CF3 and showed that the decay rate of CF30 increased after addition of NO to the system. Product analysis by infrared spectroscopy indicated that FNO was a product of the reaction between CF30 and NO. Chen et al. [ 81 investigated
0030-4018/93/S 06.00 (D 1993 Elsevier Science Publishers B.V. All rights reserved.
CHEMICAL PHYSICS LETTERS
Volume 207, number 4,5,6
the photolysis of mixtures containing CFSNO and NO in 0,/N, mixtures at about 1 atm total pressure. Analysis by long-path FTIR spectroscopy showed that CF20, PNO and NO2 were formed in stoichiometric amounts based on the loss of CFJNO. The authors rationalized their observations in terms of reaction ( 1) and (2 ) and the reactions: CF3NOthv(12>400nm)-CFS+N0,
(5)
CF,O+NO-CF,O+FNO.
(6)
Chen et al. [ 9 ] have also found evidence that the CF30 radical can abstract a hydrogen atom from alkanes CF,O+RH+CF,OH
t R.
(7)
Photolysis of CF,NO/NO mixtures in air in the presence of ethane or propane gave CF,OH and acetaldehyde or acetone, respectively. As previously reported [ 10] CF30H was found to be unstable and decomposes to give CF,O and HF. Recent work by Soathoff and Zellner [ 111, Sehested and Wallington [ 121 and by Bevilaqua et al. [ 131 have provided further evidence that H atom abstraction by CF,O radicals is a relative facile process. The purpose of the present work was to further investigate the kinetics and mechanism for the reactions of CFJO radicals with hydrocarbons. The results provide information on the importance of these reactions in the atmospheric chemistry of CFsO radicals.
2. Experimental Kinetic and product analysis studies were carried out at 298 f 2 K and atmospheric pressure in an approximately 50 I FEP teflon cylindrical reaction chamber. The chamber was surrounded by 20 germicidal lamps (Phillips TUV 15W) providing UVradiation with maximum intensity around 250 nm. Light intensity was varied by switching off various groups of lamps. A uniform reaction temperature was achieved by the use of electric fans positioned below the reaction chamber. All pressure measurements were made using MIC!SBaratron capacitance manometers. Measured amounts of the reactants were flushed from calibrated Pyrex bulbs into the teflon
28 May 1993
reaction chamber by a stream of zero-grade nitrogen (Air Products), which was then filled with either zero-grade nitrogen or ultra-pure air (Air Products). Reaction mixtures were allowed to mix for at least 30 min prior to photolysis. Quantitative analyses were carried out using gas chromatography (Gow Mac series 750 gas chromatograph equipped with a flame ionization detector) and PTIR spectroscopy (Mattson, Galaxy 5000, 0.5 cm-’ resolution). Gas tight syringes (Hamilton) or a Valco gas sampling valve were used to remove samples of the reaction mixtures from the teflon bag for chromatographic analysis. FTIR spectra were obtained using an evacuable 2 Lteflon-coated Wilks cell, containing a multipass White mirror arrangement (10 m path-length), mounted in the cavity of the spectrometer. The reaction mixtures were expanded into the cell through 3 mm inner diameter teflon tubing, after various periods of photolysis. CFsOOCFs ( > 99%, Fluorochem Ltd. ) was further purified by vacuum distillation. Hydrocarbon gases were research grade products from Matheson and were used as received. Benzene ( > 99%, Aldrich) was trap to trap distilled prior to use.
3. Results and discussion Photolysis of lo-60 ppm CF300CF3 ( 1 ppm =2.46X 1013molecule cmm3 at 298 K) in air or nitrogen at atmospheric pressure for 5 h resulted in about a 15% depletion of reactant and to the formation of stoichiometric amounts of CF,O. CF,O was stable in the dark for up to about 24 h and photolysis at 250 nm was negligible over the time period of the present experiments. In the presence of a hydrocarbon, the loss of CF300CF3 increased siguificantly, x30% in 1 h. For reactions carried out in air, the products detected by FTIR spectroscopy were CF,O and carbonyl compounds. In the initial stages of photolysis of CFSOOCF3 in the presence of isobutane in air, the yields of CF,O and CH3COCH3 corresponded to the loss of CFsOOCF,, ACF,O/ and ACH&OCHJ -ACF300CF3 = 2.12 0.3 - ACF,OOCF, = 2.0 f 0.2. At longer reaction times CH&OCH3 was lost by photolysis. Similar results were obtained from the photolysis of CFiOOCF3 with ethane and propane, where the carbonyl compounds 499
were acetaldehyde and acetone, respectively. The results are in agreement with Chen et al. [ 91, who also have provided evidence for the abstraction of hydrogen atoms from C2H6 and C3Hs by CFjO radicals. These results are consistent with the following reactions: CF300CF3+hv(l~250nm)-r2CF30,
63)
2CF~O-+CF300CF3,
(9)
CF,O+wall-+CF,O, CF,O+ (CH3)&H+CF30H+
(10)
(CHJ)&,
(11)
CF,OH-CF,O+HF,
(12)
(CH&C+02
(13)
tM+(CHs),C02tM,
2(CH,)sCOz-+2(CH9)3COt02,
(14)
(CH3)&O-+CH3COCH~ +CH3.
(15)
Combination of CF,O radicals in reaction (9) to reform the peroxide, appears to be the major reaction channel for these radicals after photolysis of CF300CF3 in air or nitrogen. The small amounts of CF20 detected are likely to arise from heterogeneous loss of CFsO at the chamber walls. Reaction of CF30 radicals with alkanes presumably involves a hydrogen abstraction process with the alkyl radical, (CH3) $ being oxidized in air to form acetone. In our system CFsOH was not observed directly, but decomposed to give C&O and HF. Kliiter and Seppelt [lo] have previously shown that the condensed phase CF90 decomposes rapidly to form CFzO and HF. However, in the gas phase at 12 Torr, the lifetime appeared to be somewhat longer with a half-life of at least 5 min. From their gas-phase experiments at 297 K Chen et al. [ 81 observed a lifetime for CFsOH of about 12 min. Decomposition of CF30H to give CFzO and HF is estimated to be approximately 5 kcal mol-’ exothermic [ 5,6]. In a recent theoretical treatment on the unimolecular decay of CF,OH, Francisco 1141 showed that the reaction had an appreciable barrier of around 50 kcal mol- I, suggesting that the loss of CFBOH observed by KlSter and Seppelt [ lo] and by Chen et al. [ 91 is probably surface catalyzed. In the present experiments the long photolysis times required to obtain measurable decays of CFJOOCF, precludes detection of CFBOH 500
28 May 1993
CHEMICAL PHYSICS LETTERS
Volume 207, number 4,5,6
with the result that CFzO is the observed product of the reaction of CFjO with alkanes. Relative rate constants for the reaction of a number of hydrocarbons with CF90 radicals were determined by comparison of the decay of the reactant hydrocarbon and that of a reference compound; reaction (7) and CF,O t reference+products .
(16)
Assuming that reaction with CFIO is the only significant loss process for both the reactant and the reference compound, it can be shown that In(@)=
2
In(s),
(17)
where the subscripts 0 and t indicate concentrations at the beginning of the experiment and at time t, respectively. Under the experimental conditions employed photolytic loss of the hydrocarbon was unimportant and the reaction mixtures were stable in the dark for at least 24 h. Photolysis of CF,OOCF,/ hydrocarbon mixtures were carried out in either air or nitrogen at 298+2 K and 1 atm total pressure; [CF,OOCF,],,= lo-60 ppm, lRH],= l-10 ppm. Concentration-time data for the hydrocarbon reactions investigated are plotted in the form of eq. ( 17 ) in figs. 1 and 2. The rate constant ratios k,/k,,, where CzH6was the reference compound, are given in table 1. These ratios were independent of reaction time, relative reactant concentration and light intensity in agreement with the proposed mechanism. At least five individual runs were carried out for each substrate and in order to test internal consistency of the rate constant ratios, each compound was run against another member of the series. In all cases the relative rate values were in excellent agreement with each other. Values of the rate constant ratios determined for each reactant pair were equal within experimental error for experiments run both in air and nitrogen. The eventual products arising from the alkyl radicals formed in these CFsO initiated reactions would be expected to vary depending on whether air or N2 was the bath gas. Thus complications arising from either product interference in the analytical procedure or enhanced removal of one of the reactant components due to product radical reactions are unlikely to be significant. It is possible that some loss
Volume207, number 4,5,6
CHEMICALPHYSICSLETTERS
28 May 1993
S.OE-2
7.0
0.0 4.OE-2
5.0 h ,' g 4.0
h +XOE-2
? ,o T 62 3.0 ==' I
5
K
z z Z.OE-2 s
2.0
1.0
0.0 0.00
0.05
0.25
0.30
Fig. 1. Concentration-time data for the reaction of CF,O radicals with various hydrocarbons: (0) CJH,; (0) i-C&,; (A)
0.00
0.05
0.10
0.15
0.20
0.25
Fig. 2. Concentration-time data for the reaction of CF,O radicals with (0) CH, and (0) c-C,H,.
Cd&; (0) Cd%.
of hydrocarbon in our relative rate studies may arise from reactions of F atoms, which may be formed in heterogeneous reactions at the chamber wall. However, the degree of selectivity shown is inconsistent with a F atom loss process for the hydrocarbon. Furthermore, the addition of up to 100 ppm of CH4 did not change the rate constant ratios obtained for a pair of hydrocarbons. Addition of CH4 to the system would be expected to scavenge F atoms, k(F+ CH4)=8~10-11 cm3molecule-‘s-r [16]. The rate constants for reaction of CF,O radicals with the hydrocarbons studied relative to the rate constant for reaction with ethane are given in table 1. Chen et al. [ 91 have previously reported the rate constant ratio k( CF,O+ CjHI) /k( CF,O t C2HQ)= 2.5 in reasonable agreement with our result. The rate constant ratios determined in this work for CF,O reactions are compared to the values for the corresponding OH radical reactions in table 1. The data show that CF,O radicals have about the same degree
of selectivity as the OH radical and thus may be expected to have the same level of reactivity. This rather surprising result is consistent with a recent laser flash photolysis - laser-induced fluorescence study by Soathoff and Zellner [ 111 on the reaction of CF30 radicals with C,H6, for which a value of the rate constant,k(CF,0tC2Hb)=(1.2+0.2)X10-’*cm3 molecule-’ s-t was obtained. Combining this rate constant with the rate constant ratios determined in this work provides estimates of the rate constants for the reaction of CF30 radicals with the series of hydrocarbons investigated, table 1. Bevilaqua et al. [ 13] employed a flow reactor coupled to a chemical ionization mass spectrometric detector to study the reaction of CF,O radicals with isobutane. The preliminary estimate of k(CF,O+fC,H,,) = (5 It 3) x 1O-'*cm3 molecule-’ s-’ is in excellent agreement with the rate constant determined from the present work. Rate constants per abstracted H atom from CH,, primary, secondary and tertiary sites in hydrocar501
Volume 207, number 4,5,6
CHEMICAL PHYSICS LETTERS
28 May 1993
Table 1 Rate constant data for the gas-phase reactions of CF,O radicals with hydrocarbons at 298 + 2 K and 1 atm total pressure Compound
k(CF,O+RH)/L(CF,O+CIHs)
a)
lO%(CF,O+RH)
0.010~0.001 0.30 + 0.02 1.0 3.9io.5 6.0f0.4 2551 30+1
CH4
c-cd6
WI GHs (CH&CH W4
(36
0.012+0.001 0.36 + 0.02 1.2 4.7kO.7 7.2f0.5 30+1 36kl
b,
k(OH+RH)/k(OH+CIH6)
‘)
0.031 0.26 1.0 4.3 8.1 32 4.6
‘) Errors quoted represent + 2a and represent precision only. b, Calculated using k(CF,O+C,HJ = (1.2* 0.2) x lo-” cm’ molecule-’ s-’ [ 111, errors in this reference rate constant are not included. ‘) Evaluated data taken from ref. [ 151.
bons can be derived from the rate data in table 1: p=&zl$
kW=1~&B-k2~6);
kWt=
z C2H6_ As expected for H atom abstrac(Cli3 ) 3CH tion reactions the rate constants for reaction of CF30 radicals with alkanes correlate well with the C-H bond strengths [ 171, see fig. 3. The high reactivity of both CzH4 and C6H6 towards reaction with CFJO
-20
-23
l&-25 Yb *c
-29
-32
-35
90
95
100
D
105
110
(kcal/mol)
Fig. 3. Rate constants per abstractable H of the specified type as a function of the C-H bond strength.
502
indicates that addition to the unsaturated system is the major reaction channel in these systems. It is interesting to compare the rate constants for reactions of CFaO radicals with alkanes to those for the corresponding CH,O reactions. At 298 K the rate constant for reaction of CF,O with ethane is over 5 orders of magnitude higher than that for CH,O, k(CH30tCZH,)=2.5x lo-” cm3 molecule-’ s-l [ 181. This result can be rationalized in terms of strengthening of the O-H bond following fluorine substitution of the methyl group in CH,OH, D(CF,O-H) ~109 kcal mol-’ [5] and D(CH,OH) = 104 kcal mol-’ [ 161. It is surprising however that the CF,O radical is slightly more reactive towards hydrocarbons than the OH radical despite the fact that the reactions are about 10 kcal mol- ’less exothermic. Presumably, the high level of reactivity of the CF30 radical is a consequence of both the strength of the O-H bond in CF,OH and the high electronegativity of the radical. As previously discussed, unimolecular decay with the loss of a F atom or reaction with O2 are unlikely to be important loss processes for CF30 in the atmosphere. Reaction of CFJO with oxides of nitrogen or hydrocarbons are possible sinks for CFJO in the troposphere. The kinetic data obtained in this work, and in that of Chen et al. [ 9 1, Soathoff and Zellner [ 111 and Bevilaqua et al. [ 13 1 indicate that reaction with hydrocarbons may be important. Bevilaqua et al. [ 131 have determined a value of the rate constant for the reaction of CF30 with NO, k(CFsO+NO) = (2+ 1)X lo-” cm3 molecule-’ s-l. Thus, aspreviously suggested by Chen et al. [9], at the hydrocarbon to NO mixing ratios typically found in the
Volume 207, number 4,5,6
CHEMICAL PHYSICS LETTERS
troposphere, the reaction with hydrocarbons will be a major sink for CFBO radicals giving CFSOH. Although CF30H is relatively unstable under laboratory conditions, the available data indicate that the decay may be mainly due to heterogeneous processes. It seems likely, therefore that CFsOH loss in the troposphere will largely take place at aerosol surfaces or by hydrolysis in cloud water.
Acknowledgement Thanks are due to Dr. C. Howard, Dr. T. Wallington, and Dr. R. Zellner for communicating their data prior to publication. The authors are grateful to AFEAS and the Commission of the European Communities for financial support.
[ 1] ES. Rowland, Ann. Rev. Phys. Chem. 42 ( 1991) 73 I. [2] Scientific Assessment of Stratospheric Ozone, World Meteorological Organization Global Ozone Research and Monitoring Project, Report No. 20, Vol. 2, Appendix: AFEAS Report (1989) ch. 6.
28May 1993
[ 31 Scientific Assessment of Stratospheric Ozone, World Meteorological Organization Global Ozone Research and Monitoring Project, Report No. 25 (1991) ch. 5. [ 41 A.M. Dognon, F. Caralp and R. Lesclauz, J. Chim. Phys. 82 (1988) 349. [ 5 ] L. Batt and R. Walsh, Intern. J. Chem. Kinetics 14 ( 1982) 933; 15 (1983) 605. [6] S.W. Benson, Thermochemical kinetics, 2nd Ed. (Wiley, NewYork, 1976). [ 71 Z. Li and J.S. Francisco, Chem. Phys. Letters 1a6 ( 1991) 336. [S] J. Chen, T. Zhu and H. Niki, J. Phys. Chem. 96 (1992) 6115. [9] J. Chen, T. Zhu, H. Niki and G.J. Mains, Geophys. Res. Letters 19 (1992) 2215. [IO] G. Kliiter and K. Seppelt, J. Am. Chem. Sot. 101 (1979) 347. [ I 1 ] H. Soathoff and R. Zcllner, Chem. Phys. Letters 206 ( 1993) 349. [ 121J. Sehested and T.J. Wallington, Environ. Sci. Technol. 27 (1993) 146. [ 131T.J. Bevilaqua, D.R. Hanson and C.J. Carlton, J. Phys. Chem., in press. [ 141J.S. Francisco, Chem. Phys. 19 ( 1991) 150. [IS] R. Atkinson, J. Phys. Chem. Ref. Data, Monograph 1 (1989). [ f6] W.B. DeMore, S.P. Sander, D.M. Golden, M.J. Molina, R.F. Hampson, M.J. Kurylo, C.J. Howard and A.R. Ravishankara, JPL Publication 961 (1990). [ 17] D.F. McMillen and D.M. Golden, Ann. Rev. Phys. Chem. 87 (1983) 1812. [ ! 8 ] W. Tang and R.F. Hampson, J. Phys. Chem. Ref. Data I 5 (1986) 1087.
so3