Document LYJ44qMEp25MaLrgdQ5XnVdQ

Environ. Sd. Technol. 2002, 36, 2395-2402 Atmospheric Chemistry of HFE-7500 [n-C3F7CF(0C2H5)CF(CF3)2]: Reaction with OH Radicals and CI Atoms and Atmospheric Fate of n-C3F7CF(OCHNCF(CF3)2 and n-C3F7CF(OCH2CH20.)CF(CF3)2 Radicals M. GOTO, Y. INOUE, AND M. KAWASAKI Department of Molecular Engineering, Kyoto University, Kyoto 606-8501, Japan A. G. GUSCHIN, L. T. MOLINA, AND M. J . MOLINA Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 T. J . WALLINGTON* AND M. D. HURLEY Ford Research Laboratory,SRL-3083, Ford Motor Company, P.O. Box 2053, Dearborn, Michigan 48121-2053 Relative rate techniques were used to measure k(OH + HFE-7500) = (2.6 0.6) x 1O-14, k(CI + HFE-7500) = (2.3 0.7) x 10-12, k[CI + n-C3F7CF(OC(0)H)CF(CF3)2] = (9.7 1.4) x 10-1s, and k[CI + n-C3F7CF(OC(0)CH3)CF(CF3)2] < 6 x 1O-17 c m3 molecule-1 s-1 at 295 K [HFE-7500 = n-C3F7CF(0C2H5)CF(CF3)2]. From the value of k(OH + HFE-7500) an estimate of 2.2 years for the atmospheric lifetime of HFE7500 is obtained.Two competing loss mechanisms for n-C3F7CF(OCHOCH3)CF(CF3)2 radicals were identified in 700 Torr of N2/02 diluent at 295 K; reaction with 02 and decomposition via C --C bond scission with kojkdecomp = 0.013 0.006 Torr-1. The CI atom initiated oxidation of HFE7500 in N2/02 diluent gives n-C3F7CF(OC(0)CH3)CF(CF3)2 as the major product and n-C3F7CF(OC(0)H)CF(CF3)2 as a minor product. The atmospheric oxidation of HFE-7500 gives n-C3F7CF(OC(0)CH3)CF(CF3)2 and n-C3F7CF(OC(0)H)CF(CF3)2 as oxidation products. The results are discussed with respect to the atmospheric chemistry and environmental impact of HFE-7500. 1. Introduction Recognition of the adverse environmental impact of chlorofluorocarbon (CFC) release into the atmosphere (1, 2) has led to an international effort to replace CFCs with environmentally acceptable alternatives. Hydrofluoroethers (H FEs) are a class of fluid compounds that have been developed to replaceCFCsin applicationssuch asthecleaningofelectronic equipment, heat transfer agents, and carrier fluids for lubricant deposition (3). HFE-7000(n-C3F7OCH3), HFE-7100 [supplied commerciallyasa mixtureof 35%n-C4F9OCH3 and *Correspondin phone: fax: M; e-mail: @ford.corn. 10.1021/es0113798 CCC: $22.00 Published on Web 04/25/2002 2002 American Chemical Society 65% (CF3)2CFCF2OCH3], and HFE-7200 [supplied commercially as a mixture of 35% n-C4F9OC2H 5 and 65% (CF3)2CFCF2OC2H5]haveattracted commercial interestand wehave reported assessments of their atmospheric chemistry (4-6). HFE-7500 [n-C3F7CF(OC2H5)CF(CF3)2, 2-trifluoromethyl-3ethoxydodecafluorohexane] is a new member of this class of compounds. H FE-7500 is a volatile liquid (bp 130 IC) with a vapor pressure of 16 Torr at 25 IC (7) and will probably be released into the atmosphere. Priorto its large-scale industrial use, an assessment of the atmospheric chemistry, and hence environmental impact, of HFE-7500 is needed. We report the results of the first study of the atmospheric chemistry of H FE-7500. CH3 CH2 CF3 CF2 HFE-7500 CF3 F cF3 The atmospheric oxidation of HFE-7500 will be initiated by reaction with OH radicals. The alkyl radicals produced in reaction 1 will add O2 rapidly to give peroxy radicals: n-C3F7CF(OCH 2CH3)CF(CF3)2 + OH n-C3F7CF(OCH.CH 3)CF(CF3)2 + H2O (la) n-C3F7CF(OCH 2CH3)CF(CF3)2 + OH n-C3F7CF(OCH2CH2.)CF(CF3)2 + H2O (lb) n-C3F7CF(OCH.CH3)CF(CF3)2 + O2 + M n-C3F7CF(OCHOOCH3)CF(CF3)2 + M (2a) n-C3F7CF(OCH 2CH2.)CF(CF3)2 + O2 + M n-C3F7CF(OCH 2CH 2OO.)CF(CF3)2 + M (2b) The peroxy radicals derived from HFE-7500 will react with NO, NO2, HO2, and other peroxy radicals in the atmosphere (8, 9): n-C3F7CF(OC2H4OO.)CF(CF3)2 +NO n-C3F7CF(OC2H4O.)CF(CF3)2 + NO2 (3a) n-C3F7CF(OC2H4OO.)CF(CF3)2 +NO + M n-C3F7CF(OC2H4ONO2)CF(CF3)2 + M (3b) n-C3F7CF(OC2H4OO.)CF(CF3)2 + NO2 M n-C3F7CF(OC2H4O2NO2)CF(CF3)2 + M (4) C3F7CF(OC2H4OO.)CF(CF3)2 + HO2 -- products (5) C3F7CF(OC2H4OO.)CF(CF3)2 + R'O2 products (6) Experiments have been performed in our laboratories to elucidate the atmospheric chemistryof HFE-7500. A relative rate method was used at MIT to measure the kinetics of the reaction of OH radicals with H FE-7500 and henceto provide an assessment of its atmospheric lifetime. Relative rate methods were used at Ford to measure the kinetics of the reaction of CI atoms with HFE-7500 and its oxidation products. Fourier transform infrared (FTIR) spectroscopy coupled to a smog chamber at Ford was employed to VOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2395 investigate the atmospheric fate of the alkoxy radicals C3F7CF(OCHOCH3)CF(CF3)2 and C3F7CF(OCH2CH2O)CF(CF3)2. The results are reported herein and discussed with respect to the environmental impact of HFE-7500. 2. Experimental Section The two experimental systems used are described in detail elsewhere(10, 11). All samplesof HFE-7500used in thiswork were supplied by the 3M Company at a purity >99% and were used without further purification. The uncertainties reported in this paper are two standard deviations unless otherwise stated. Standard error propagation methods were used to combine uncertainties where appropriate. 2.1. FTIR-Photolysis System at MIT. The rate constant for theOH + HFE-7500reaction wasobtained bymonitoring thedisappearancerateof HFE-7500relativeto CH4and CH3Cl in the presence of OH radicals at 295 K. The decay of the sample was measured by infrared spectroscopy (10). The concentration of HFE-7500 was monitored by spectral subtraction. OH radicals were generated by photolysis of ozone at 254 nm in the presence of water vapor: O3 + h f O(1D) + O2 (7) O(1D) + H2O f 2OH (8) The long-path absorption cell, made of Pyrex glass, had a volumeof 7.6L and abaselength of 60cm,which wasadjusted to give a total of 24 passes and an optical path length of 14.4 m. The concentrations of the reactants and products were monitored with an FTIR spectrometer (Nicolet 20SX). The mercury photolysis lamp (Ace Hanovia 450-Watt mediumpressuremercurylamp) wasenveloped in aVycor tube,which transmits254-nm radiation but absorbsthe185-nm Hgline, and was placed inside the absorption cell. No decay of the HFE-7500 sample was observed upon irradiation in the chamber in the absence of ozone. The organic reactants were mixed with helium in a 3-L glass reservoir to yield mole fractions of 1%. Ozone was prepared by first trapping the effluent from an ozonizer in cold silica gel, and then desorbing the sample into a 12-L glass reservoir and subsequently mixing it with helium. The experiments were performed at room temperature in 200 Torr of helium as a buffer gas in the presence of 3-5 Torr of ozone and 2-3 Torr of water vapor. 2.2. FTIR-SmogChamber System at Ford Motor Company. Experiments were performed in a 140-L Pyrex reactor interfaced to a Mattson Sirus 100 FTIR spectrometer (11). The reactor was surrounded by 22 fluorescent blacklamps (GE F15T8-BL), which wereused to photochemicallyinitiate theexperiments. Theoxidation of HFE-7500wasinitiated by reaction with Cl atoms, which were generated by the photolysis of molecular chlorine in air diluent at 700 Torr total pressure at 295 ( 2 K. Cl2 + h f 2Cl (9) Cl + n-C3F7CF(OC2H5)CF(CF3)2 f n-C3F7CF(OC2H4)CF(CF3)2 + HCl (10) n-C3F7CF(OC2H4)CF(CF3)2 + O2 + M f n-C3F7CF(OC2H4O2)CF(CF3)2 + M (2) The loss of HFE-7500 and the formation of products were monitored byFTIR spectroscopywith an infrared path length of 27.4 m and a resolution of 0.25 cm-1. Infrared spectra were derived from 32 coadded interferograms. Two sets of experiments were performed. First, relative ratetechniqueswereused to determinetherateconstant for 2396 9 ENVIRONM ENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002 FIGURE 1. Decay of HFE-7500 versus CH4 (2) and CH3Cl (b) in the presence of OH radicals in 200 Torr of helium at 295 K. the reaction of Cl atoms with HFE-7500 and n-C3F7CF(OC(O)H)CF(CF3)2. Second, the products of the atmospheric oxidation of HFE-7500wereinvestigated byirradiatingHFE7500/ Cl2/ O2/ N2 mixtures. Initial concentrations of the gas mixtures for the relative rate experiments of HFE-7500 were 1.7 mTorr of HFE-7500, 7-30 mTorr of the reference compounds [C2H5Cl or CH3OC(O)H], and 100 mTorr of Cl2 in 700 Torr of air diluent. In the study of the oxidation of HFE-7500, reaction mixtures consisted of 1.7mTorr of HFE-7500, 9-100mTorr of Cl2 and 10-550Torr of O2at atotal pressureof 700Torr in N2diluent. All experiments were performed at 295 K. 3. Results and Discussion 3.1. Relative Rate Study of the Reaction of OH with HFE7500. The kinetics of reaction 1 were measured relative to reactions 11 and 12 by use of the experimental system at M I T. OH + HFE-7500 f products (1) OH + CH4 f products (11) OH + CH3Cl f products (12) Loss of HFE-7500versus the reference compounds is shown in Figure 1. The error bars on the data points in Figure 1 represent theuncertaintyassociated with theanalysis. Linear least-squares analysis of the data in Figure 1 gives k1/ k11 ) 2.96 ( 0.34 and k1/ k12 ) 0.73 ( 0.05. Using k11 ) (6.3 ( 0.6) 10-15 and k12 ) (3.6 ( 0.7) 10-14 cm3 molecule-1 s-1 (12), we derive k1 ) (1.86 ( 0.29) 10-14 and (2.63 ( 0.56) 10-14 cm3 molecule-1 s-1. Within the experimental uncertainties, the values of k1 derived from the two different reference compounds are indistinguishable. Close examination of Figure 1 reveals that there is more data scatter evident in the experiments obtained with CH4 as reference than in those experiments with CH3Cl. The reason for the greater scatter in the CH4 data lies in the fact that CH4 is significantly less reactive than HFE-7500, while CH3Cl hasareactivitycomparableto thatof HFE-7500.Hence, as seen from Figure 1, the fractional loss of CH4 spans the range 8-21% while that of CH3Cl spans the range 23-57%. It isdifficult to measurean 8%lossof CH4with great precision and this probably goes some way to explain why the first data point in the CH4 series lies above the regression line, which has been forced to fit through the origin. Given the scatter in the CH4 data discussed above, we choose to cite a final value for k1 that is based upon the data obtained with CH3Cl as reference. Hence, k1 ) (2.6 ( 0.6) 10-14 cm3 molecule-1 s-1; the quoted uncertainty includes statistical uncertaintyin k1/ k12and possiblesystematicerrorsassociated with the uncertainty in k12, and hence reflects the accuracy of the measurement of k1. Thisresult can becompared to k(OH + C2H5OC2H5) ) 1.3 10-11 (13), k(OH + n-C4F9OC2H5) ) (6.4 ( 0.7) 10-14 (5), and k(OH + (CF3)2CFCF2OC2H5) ) (7.7 ( 0.8) 10-14 cm3 molecule-1 s-1 (5). The 295-fold reduction in the reactivity of the OC2H5 group in HFE-7500compared to that in diethyl ether presumablyreflectstheincreasein C-H bond strength associated with the presence of multiple fluorinated substituents (14). As might be expected for such structurally similar molecules,thereactivitiesof OH radicalstoward HFE7500, n-C4F9OC2H5, and (CF3)2CFCF2OC2H5 are broadly comparable. The somewhat lower reactivity of HFE-7500 compared to n-C4F9OC2H5 and (CF3)2CFCF2OC2H5 probably reflects a somewhat greater C-H bond strength in HFE7500. The value of k1 can be used to provide an estimate of the atmospheric lifetimeof HFE-7500. Assumingan atmospheric lifetimefor CH3CCl3with respect to reaction with OH radicals of 5.7 years (15) and a rate constant for the CH3CCl3 + OH reaction of 1.0 10-14 cm3 molecule-1 s-1 (12) leads to an estimate of the atmospheric lifetime of HFE-7500 of (1.0 10-14/ 2.6 10-14) 5.7) 2.2years. Theoptimal temperature for such a scaling analysis is 277 K (16) (rather than 295 K used here) but we do not have any data for k1 at 277 K. By analogy to other HFEs (12), the temperature dependence of reaction 1 is expected to be very similar to that for reaction of OH radicals with CH3CCl3. Hence, the use of 295 K rather than 277 K is not expected to have any material impact on the estimated atmospheric lifetime. 3.2. RelativeRateStudyof theReaction of Cl with HFE7500. Prior to investigating the atmospheric fate of n-C3F7CF(OC2H4O)CF(CF3)2 radicals, relativerateexperiments were performed with the FTIR system at Ford Motor Company to investigate the kinetics of reaction 10. The techniques used are described in detail elsewhere (17). Cl atoms were generated by photolysis of molecular chlorine: Cl2 + h f 2Cl (13) Cl + n-C3F7CF(OC2H5)CF(CF3)2 f n-C3F7CF(OC2H4)CF(CF3)2 + HCl (10) The kinetics of reaction 10 were measured relative to the reactionsof Cl atomswith ethyl chlorideand methyl formate (reactions 14 and 15): Cl + C2H5Cl f products (14) Cl + CH3OC(O)H f products (15) Reaction mixtures consisted of 1.7-1.9 mTorr of HFE-7500, 100-110mTorr of Cl2, and 29mTorr of C2H5Cl or 7-8mTorr of CH3OC(O)H in 700Torr of air diluent. Theobserved losses of HFE-7500versusthoseof referencecompoundsareshown in Figure 2A. Linear least-squares analysis gives k10/ k14 ) 0.32( 0.04and k10/ k15 ) 1.44( 0.20. Usingk14 ) 8.04 10-12 (18) and k15 ) 1.4 10-12 (19) gives k10 ) (2.6 ( 0.3) 10-12 and k10 ) (2.0( 0.3) 10-12 cm3 molecule-1 s-1, respectively. FIGURE 2. (A) Decay of HFE-7500 versus C2H5Cl (b) and CH3OC(O)H (2) in the presence of Cl atoms in 700 Torr of N2 at 295 K. (B) Decay of n-C3F7CF(OCH2CH2O2)CF(CF3)2 versus CHD3 (O) and CD4 (0) in the presence of Cl atoms in 700 Torr of N2 at 295 K. Weestimatethat potential systematic errors associated with uncertainties in the reference rate constants could add an additional 10% uncertainty range for k10. Propagating this additional uncertainty gives k10 ) (2.6 ( 0.4) 10-12 and k10 ) (2.0 ( 0.4) 10-12 cm3 molecule-1 s-1. We choose to cite a final value of k10 that is an average of those determined with the two different reference compounds, together with error limits that encompass the extremes of the individual determinations. Hence, k10 ) (2.3 ( 0.7) 10-12 cm3 molecule-1 s-1. This result can be compared to k(Cl + C2H5OC2H5) ) (3.6 ( 0.3) 10-10 (20) and k(Cl + n-C4F9OC2H5) ) (2.7 ( 0.6) 10-12(5).Aswith OH radicals,thereactivityof Cl atomstoward the -OC2H5 group in HFE-7500 is substantially lower than that in diethyl ether. The magnitude of the decrease in reactivity for the Cl atom reactions is somewhat less than that for OH radicals (a factor of 78 compared to a factor of 295), reflecting the fact that Cl atoms are, in general, less discriminating than OH radicals in their reactions with organic compounds. The less discriminating nature of Cl atoms is also evidenced by the fact that, in contrast to the situation for OH radicals discussed in the previous section, there is no discernible difference between the reactivity of Cl atoms toward n-C4F9OC2H5 and HFE-7500. The average concentration of Cl atoms in the atmosphere is much lower [by a factor of approximately 1000 (21)] than that of OH VOL. 36, NO. 11, 2002 / ENVIRONM ENTAL SCIENCE & TECHNOLOGY 9 2397 radicals and reaction with Cl atoms will not be a substantial atmospheric loss of HFE-7500. Finally, we can use the rate constants for reactions of Cl atoms and OH radicals with HFE-7500 to test the empirical relationship log [k(OH + HFE)] ) (0.74 ( 0.07) log [k(Cl + HFE)] - (4.4 ( 0.9) proposed recently to estimate k(OH + HFE) from k(Cl + HFE) or vica versa (22). Substituting a value of k10 ) 2.3 10-12 into the above expression leads to a prediction of k1 ) 9.7 10-14 cm3 molecule-1 s-1. This prediction is a factor of 3.7 greater than the measured value of k1 ) (2.6( 0.6) 10-14 cm3 molecule-1 s-1, which provides somemeasureof theuncertainties inherent in theempirical relationship. 3.3. Relative Rate Studies of the Reactions of Cl with n-C3F7CF(OC(O)H)CF(CF3)2 and n-C3F7CF(OC(O)CH3)CF(CF3)2. As discussed in section 3.4, n-C3F7CF(OC(O)H)CF(CF3)2 and n-C3F7CF(OC(O)CH3)CF(CF3)2 were readily generated in theFord chamber byUV irradiation of gasmixtures of 2mTorr of HFE-7500, 100-200mTorr of Cl2, and 6-7Torr of O2 in 700 Torr of N2 diluent. After all (>97%) of the HFE7500 was consumed, a reference compound was added and the UV irradiation was resumed: n-C3F7CF(OC(O)H)CF(CF3)2 + Cl f n-C3F7CF(OC(O))CF(CF3)2 + HCl (16) n-C3F7CF(OC(O)CH3)CF(CF3)2 + Cl f n-C3F7CF(OC(O)CH2)CF(CF3)2 + HCl (17) CHD3 + Cl f products (18) CD4 + Cl f products (19) The rate constants k16 and k17 were derived by observing therelativelossratesof n-C3F7CF(OC(O)H)CF(CF3)2, n-C3F7CF(OC(O)CH3)CF(CF3)2, and the reference compounds. Figure 2b shows loss of n-C3F7CF(OC(O)H)CF(CF3)2 versus lossesof referencecompounds. Linear least-squaresanalysis of thedata in Figure2b gives k16/ k18 ) 0.42( 0.04and k16/ k19 ) 1.58 ( 0.12. Using k18 ) 2.32 10-14 (17) and k19 ) 6.1 10-15 cm3 molecule-1 s-1 (17), we derive k16 ) (9.7 ( 0.9) 10-15 and (9.6( 0.5) 10-15 cm3 molecule-1 s-1. Weestimate that potential systematic errorsassociated with uncertainties in thereferencerateconstants contributean additional 10% uncertainty range for k15. Propagating this additional uncertaintygivesk16 ) (9.7( 1.4) 10-15 and (9.6( 1.2) 10-15 cm3 molecule-1 s-1. We choose to cite a final value for k16 thatistheaverageof thosedetermined with thethreedifferent reference compounds, together with error limits that encompass the extremes of the individual determinations. Hence, k16 ) (9.7 ( 1.4) 10-15 cm3 molecule-1 s-1, and the quoted uncertainty reflects the accuracy of the measurements. Therateconstant k16 is 237times lower than k10. This is consistent with the behavior of other formates, which are typically significantly less reactive toward Cl atoms than the ethers from which they are derived [e.g., k(Cl + CF3OCH3)/ k(Cl + CF3OC(O)H) ) 143 ( 27 (23)]. During the present series of experiments no discernible loss (<2%) of n-C3F7CF(OC(O)CH3)CF(CF3)2 was observed while consumptions of CHD3 and CD4 were 5-90%. From this observation we are able to derive an upper limit of k17 < 6 10-17 cm3 molecule-1 s-1. 3.4. Study of theMechanism of theAtmospheric OxidationofHFE-7500.Theatmospheric degradation mechanism of HFE-7500 was studied by UV irradiation of HFE-7500/ Cl2/ O2/ N2 mixtures in the FTIR-smog chamber system at Ford. Experiments were performed at a constant total pressure of 700 Torr with the O2 partial pressure varied over therange10-700Torr. Therelativeintensitiesof theinfrared 2398 9 ENVIRONM ENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002 FIGURE 3. Infrared spectra of HFE-7500 (A), X ) n-C3F7CF(OC(O)CH3)CF(CF3)2 (B), and Y ) n-C3F7CF(OC(O)H)CF(CF3)2 (C). FIGURE 4. Observed yields of X ) n-C3F7CF(OC(O)CH3)CF(CF3)2 (b) and Y ) n-C3F7CF(OC(O)H)CF(CF3)2 (2) versus the O2 partial pressure follow ing the UV irradiation of HFE-7500/Cl2/N2/O2 mixtures at 700 Torr total pressure and 295 K. The curves are fits to the data; see text for details. features of the products of the Cl atom initiated oxidation of HFE-7500 varied with the O2 partial pressure. A careful comparison of the IR features formed in low and high [O2] experiments revealed that two distinct products, or sets of products, were formed in the chamber. We will label these two products X and Y. Figure 3 shows IR spectra of HFE7500, X, and Y. IR features attributable to X were observed at984,1133,1139,1167,1257,1304,and 1841cm-1.IRfeatures attributable to Y were observed at 980, 1160, 1257, 1305, and 1812 cm-1. The increases in X and Y scaled linearly with the loss of HFE-7500 over the range of HFE-7500 consumptions of 10-90%, suggesting the absence of significant secondary loss processes for X and Y. As shown in Figure 4, as the O2 partial pressurewasincreased, theyield of Y decreased while that of X increased. For O2 partial pressures greater than 400 Torr there was little change in the yields of X and Y. The reaction of Cl atoms with HFE-7500 in the presence of O2 givesriseto two different peroxyradicals, which in turn will undergo self- and cross-reaction to give the corresponding alkoxy radicals: n-C3F7CF(OCHO2CH3)CF(CF3)2 + RO2 f n-C3F7CF(OCHOCH3)CF(CF3)2 + RO + O2 (20a) n-C3F7CF(OCH2CH2O2)CF(CF3)2 + RO2 f n-C3F7CF(OCH2CH2O)CF(CF3)2 + RO + O2 (20b) There are several possible fates of these alkoxy radicals. For n-C3F7CF(OCHOCH3)CF(CF3)2 radicals, thepossibilitiesare n-C3F7CF(OCHOCH3)CF(CF3)2 + M f n-C3F7CF(OC(O)H)CF(CF3)2 + CH3 + M (21) n-C3F7CF(OCHOCH3)CF(CF3)2 + M f n-C3F7CF(OC(O)CH3)CF(CF3)2 + H + M (22) n-C3F7CF(OCHOCH3)CF(CF3)2 + O2 f n-C3F7CF(OC(O)CH3)CF(CF3)2 + HO2 (23) while for n-C3F7CF(OCH2CH2O)CF(CF3)2 radicals the possibilities are n-C3F7CF(OCH2CH2O)CF(CF3)2 + M f n-C3F7CF(OCH2)CF(CF3)2 + HCHO + M (24) n-C3F7CF(OCH2CH2O)CF(CF3)2 + M f n-C3F7CF(OCH2C(O)H)CF(CF3)2 + H + M (25) n-C3F7CF(OCH2CH2O)CF(CF3)2 + O2 f n-C3F7CF(OCH2C(O)H)CF(CF3)2 + HO2 (26) Theincreasein theyield of Xand decreasein Y with increasing [O2] shows that reactions 23and/ or 26become important at high [O2]. The unknown X is either n-C3F7CF(OC(O)CH3)CF(CF3)2, n-C3F7CF(OCH2C(O)H)CF(CF3)2, or both. Thereare four aspects of the IR spectrum which suggest that X is the acetate n-C3F7CF(OC(O)CH3)CF(CF3)2. First, the observed carbonyl stretching frequency of 1841 cm-1 is typical for a fluorinated ester (4-6). Second, there is no aldehydic C-H stretchingfeaturein the2720-2820cm-1region. Third, there isno significant -CH2- scissorsfeatureat approximately1450 cm-1. Finally, there is no -C(O)H deformation band at approximately 1400 cm-1. The kinetic behavior of X is also consistent with its assignment as n-C3F7CF(OC(O)CH3)CF(CF3)2rather than n-C3F7CF(OCH2C(O)H)CF(CF3)2.Aldehydes react rapidly with Cl atoms with rate constants of typically 10-11-10-10 cm3 molecule-1 s-1 (12). In contrast, esters are relatively unreactive toward Cl atoms; for example, k(Cl + CH3C(O)OCH3) ) 2.2 10-12 cm3 molecule-1 s-1 with more than 95% of the reaction occurring at the -OCH3 site (24). It was determined in the present work (see section 3.3) that Cl atoms are relatively unreactive toward X (rate constant < 6 10-17 cm3 molecule-1 s-1); thus the kinetic behavior of X is consistent with its identification as n-C3F7CF(OC(O)CH3)CF(CF3)2. In thefollowingwewill proceed on theassumption that X is the acetate n-C3F7CF(OC(O)CH3)CF(CF3)2. As noted aboveand illustrated in Figure4, increasein the O2 partial pressure led to an increase in the yield of product X and a corresponding decrease in the yield of product Y. The simplest explanation of this observation is that product Y is the formate n-C3F7CF(OC(O)H)CF(CF3)2 formed via reaction 21, which is in competition with the O2-dependent reaction 23. The observed dependence of the yields of the formate and acetate on [O2] then reflects a competition FIGURE 5. Schematic depiction of the reactions discussed in section 3.4 leading to formation of X and Y. Reaction numbers are given in parentheses. between reactions21and 23. For claritythereactionsleading to X and Y are shown schematically in Figure 5. As shown in Figure 4, while the yield of n-C3F7CF(OC(O)H)CF(CF3)2 decreases as the O2 partial pressure is increased from 3 to 400 Torr, further increase in [O2] has little or no effect and the n-C3F7CF(OC(O)H)CF(CF3)2 yield remains at approximately 20%. This behavior suggests that there is a source of n-C3F7CF(OC(O)H)CF(CF3)2 which is independent of [O2]. A likely source for this n-C3F7CF(OC(O)H)CF(CF3)2 is the n-C3F7CF(OCH2CH2O)CF(CF3)2 alkoxy radical. Therearethreepossiblefatesof then-C3F7CF(OCH2CH2O)CF(CF3)2 radical: reactions 24-26. Decomposition via reaction 24 gives the carbon-centered radical n-C3F7CF(OCH2)CF(CF3)2. By analogy to the behavior of analogous radicals[e.g., CF3OCH2 (23), C2F5OCH2 (25), C3F7OCH2 (6), and (CF3)2CFCF2OCH2 (4)] the C3F7CF(OCH2)CF(CF3)2 radical isexpected to react to givetheformate, C3F7CF(OC(O)H)CF(CF3)2. Reactions 25 and 26 produce C3F7CF(OCH2C(O)H)CF(CF3)2, which containsan aldehydic functional group and, as discussed above, is expected to be at least 1 order of magnitude more reactive than HFE-7500 toward Cl atoms and will beconverted into n-C3F7CF(OCH2)CF(CF3)2radicals and henceinto n-C3F7CF(OC(O)H)CF(CF3)2viareactions2730. Therewereno IR product features that could beascribed to thealdehydeC3F7CF(OCH2C(O)H)CF(CF3)2 in thesystem, but thisdoesnot precludeitsformation and subsequent loss via reaction 27: n-C3F7CF(OCH2C(O)H)CF(CF3)2 + Cl f n-C3F7CF(OCH2C(O))CF(CF3)2 + HCl (27) n-C3F7CF(OCH2C(O))CF(CF3)2 + O2 + M f n-C3F7CF(OCH2C(O)OO)CF(CF3)2 + M (28) n-C3F7CF(OCH2C(O)OO)CF(CF3)2 + RO2 f n-C3F7CF(OCH2C(O)O)CF(CF3)2 + RO + O2 (29) n-C3F7CF(OCH2C(O)O)CF(CF3)2 + M f n-C3F7CF(OCH2)CF(CF3)2 + CO2 + M (30) The decrease in the yield of n-C3F7CF(OC(O)H)CF(CF3)2 and increaseof n-C3F7CF(OC(O)CH3)CF(CF3)2with increasing [O2] shown in Figure 4 reflects a competition between reactions 21 and 23 for the available n-C3F7CF(OCHOCH3)CF(CF3)2 radicals. We can place the yield of n-C3F7CF(OC(O)CH3)CF(CF3)2 on an absolute basis by equating the increase in its yield to the decrease in the yield of n-C3F7CF(OC(O)H)CF(CF3)2and byassumingthat thesetwo species account for all of the observed loss of HFE-7500. It was on thisbasisthat then-C3F7CF(OC(O)CH3)CF(CF3)2 and n-C3F7- VOL. 36, NO. 11, 2002 / ENVIRONM ENTAL SCIENCE & TECHNOLOGY 9 2399 CF(OC(O)H)CF(CF3)2 yields in Figure 4 were calibrated. The dependence of the product yields in Figure 4 on [O2] can be used to extract a value for the rate constant ratio k23/ k21. Assuming that the fate of n-C3F7CF(OCHOCH3)CF(CF3)2 radicals is either unimolecular decomposition via reaction 21 or bimolecular reaction with O2 (reaction 23) and that all n-C3F7CF(OCH2CH2O)CF(CF3)2 radicals are converted into n-C3F7CF(OC(O)H)CF(CF3)2, then theyield of n-C3F7CF(OC(O)H)CF(CF3)2 can be related to the molar yields of the two alkoxy radicals and the rate constant ratio k23/ k21: Y[n-C3F7CF(OC(O)H)CF(CF3)2] ) Y[n-C3F7CF(OCHOCH3)CF(CF3)2] (k23/ k21)1[O2] + 1 + Y[n-C3F7CF(OCH2CH2O)CF(CF3)2] (I) while the yield of n-C3F7CF(OC(O)CH3)CF(CF3) is given by Y[n-C3F7CF(OC(O)CH3)CF(CF3)] ) Y[n-C F CF(OCHOCH )CF(CF ) ] (k23/ k21)[O2] + 37 3 3 2 (k23/ k21)[O2] + 1 C (II) The term C in expression II was added to account for the nonzero intercept in the n-C3F7CF(OC(O)CH3)CF(CF3) yield plot and is discussed later. The curves in Figure 4 are fits of the above expressions to the experimental data. A leastsquaresfit to then-C3F7CF(OC(O)H)CF(CF3)2 datagivesY[nC3F7CF(OCHOCH3)CF(CF3)2] ) 0.91 ( 0.05, Y[n-C3F7CF(OCH2CH2O)CF(CF3)2] ) 0.04 ( 0.03, and k23/ k21 ) 0.011 ( 0.002 Torr-1, while a fit to the n-C3F7CF(OC(O)CH3)CF(CF3) data gives Y[n-C3F7CF(OCHOCH3)CF(CF3)2] ) 0.88 ( 0.07, k23/ k21 ) 0.015 ( 0.004 Torr-1, and C ) 0.09 ( 0.06. It is gratifying to note that least-squares fits of expressions I and II to thevariation of theyieldsof n-C3F7CF(OC(O)H)CF(CF3)2 and n-C3F7CF(OC(O)CH3)CF(CF3) with [O2] give indistinguishable values of k23/ k21. We choose to quote a final value of k23/ k21 that is an average of the two determinations, with error limits that encompass the extremes of the individual determinations; k23/ k21 ) 0.013 ( 0.006 Torr-1. In 1 atm of air at 295 K, the O2 partial pressure is 160 Torr and 68% of the n-C3F7CF(OCHOCH3)CF(CF3)2 radicals will react with O2 while the remainder will decompose via reaction 21. In theatmospheremost of theoxidation of HFE-7500will occur at elevated altitudes and hence temperatures and total pressures below that employed here. Lower temperature favors reaction 23, while lower total pressure (and hence O2 partial pressure) favors reaction 21. Reaction 21 is a unimolecular decomposition reaction and will haveasubstantial activation barrier. As with other similar competitions [e.g., in the case of the CF3CFHO radical (8)], it is expected that the temperature effect will greatly dominate the pressure effectand thattheatmosphericfateof n-C3F7CF(OCHOCH3)CF(CF3)2 radicals will be dominated by reaction with O2. At this point we need to return to the inclusion of factor C in expression II. It is apparent from Figure 4 that n-C3F7CF(OC(O)CH3)CF(CF3) hasanonzero y-axisintercept. There are several possible causes of this intercept: (i) reaction 21 is important, (ii) there is a contribution due to a molecular channel of theperoxyradical self- and cross-reactionswhich gives the ester, or (iii) the ester is formed in the n-C3F7CF(OCHO2CH3)CF(CF3)2 + HO2 reaction. We are not able to distinguish between these possibilities at the present time. It should be noted that the important conclusion from this study, namely, that the atmospheric fate of n-C3F7CF(OCHOCH3)CF(CF3)2radicalsisdominated byreaction with O2, is insensitive to the cause of the intercept. To quantify the effect of factor C on the value of k23/ k21 obtained in the 2400 9 ENVIRONM ENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002 FIGURE 6. Secondary alkoxy radicals derived from HFE-7500 (I), (CF3)2CFCF2OC2H5 (main component of HFE-7200) (II), bis(2,2,2trifluoroethyl) ether (III), ethyl tert-butyl ether (IV), and diethyl ether (V). analysis, a modified version of expression II (without C) was fitted to the data in Figure 4. Omission of C leads to a value k23/ k21 ) 0.021 ( 0.007 Torr-1, which is not substantially different from the value of k23/ k21 ) 0.015 ( 0.004 Torr-1 obtained from the complete expression II. Finally, we note that factor C doesnot need to beincorporated into expression I as the factor Y[n-C3F7CF(OCH2CH2O)CF(CF3)2] serves to account for the fact that there appears to be a non-O2dependent source of the formate Y. In section 3.2 it was concluded that the atmospheric fate of HFE-7500 is reaction with OH radicals and that reaction with Cl atomsisof littleatmospheric significance. Thisraises the question, "Of what help is a study of the products of Cl atom-initiated oxidation of HFE-7500 to understanding the products of the OH radical-initiated oxidation?" To answer thisquestion weneed to consider two piecesof information. First, the molar yield of n-C3F7CF(OCHOCH3)CF(CF3)2 radicals derived from the data in Figure 4 is approximately 0.90, indicating that the majority of the reaction of Cl atoms with HFE-7500 occurs at the -CH2- site. Second, as evident from sections 3.1 and 3.2, OH radicals are less reactive and more discriminating than Cl atoms. It is likely that >90% of thereaction of OH radicalswith HFE-7500occursat the-CH2site and that Cl atoms provide a good surrogate for OH radicalsin understandingthemechanism of theatmospheric oxidation of HFE-7500. 4. Implications for Atmospheric Chemistry 4.1. Atmospheric Fate of Alkoxy Radicals of the General Formula ROCH(O)CH3. As discussed in section 3.4, two competing loss processes were identified for n-C3F7CF(OCHOCH3)CF(CF3)2 radicals: reaction with O2 and decomposition (k23/ k21 ) kO2/ kdiss ) 0.013 ( 0.006 Torr-1). We can compare this behavior with the literature data for that of structurally similar secondary alkoxy radicals formed duringtheoxidation of HFE-7200(II), CF3CH2OCH2CF3 (III), ethyl tert-butyl ether (IV), and diethyl ether (V): see Figure 6. In 1atm of air at 295K, reaction with O2and decomposition arecompetinglossprocessesfor alkoxyradicalsderived from HFE-7200 (II) with kO2/ kdiss ) 0.026 ( 0.010 Torr-1 (5). In contrast, decomposition via C-C bond scission is the sole fate of the alkoxy radicals (III) (26), (IV) (27), and (V) (27). Given the obvious structural similarity of species I and II evident in Figure 6, it is not too surprising that they have indistinguishablevaluesof kO2/ kdiss.Comparingthechemistry of species I and II with those of IV and V, it is evident that fluorination of the R moiety in ROCH(O)CH3 radicals leads to an increasein therelativeimportanceof thereaction with O2. It is of interest to contrast theatmospheric fateof ROCH(O)CH3 radicalsderived from ethersto theatmospheric fate of RCH(O)CH3 radicals derived from alkanes. At room temperature, RCH(O)CH3 radicals react with O2 with a rate constant of kO2 ) 7.5 10-15 cm3 molecule-1 s-1 and undergo decomposition into RCHO + CH3 with a rate constant of kdiss ) 28 s-1 (28, 29). In 1 atm of air, kO2[O2]/ kdiss ) 1400. In contrastto thefateof ROCH(O)CH3radicals,theatmospheric fate of RCH(O)CH3 radicals is completely dominated by reaction with O2. It can be concluded that the ether functionality in ROCH(O)CH3 radicals either decreases the rate of O2 reaction or increases the rate of the C-C bond scission, or both. Introduction of the ether functionality will changethebond strengthsin theradical. Thedecomposition channel will bemoresensitivethan theO2reaction to changes in chemical bond strength. The simplest explanation for the differentbehavior of ROCH(O)CH3and RCH(O)CH3radicals is that the presence of the ether functionality decreases the C-CH3bond strength,therebyfacilitatingthedecomposition channel. High level ab initio calculations are needed to provide further insight into the fundamental cause of the different behavior of the two types of alkoxy radical. Such calculations are beyond the scope of the present work. 4.2.AtmosphericChemistryandEnvironmental Impact of HFE-7500. We present herein a large body of kinetic and mechanistic data pertaining to the atmospheric chemistry of HFE-7500. The atmospheric lifetime of HFE-7500 is determined byitsreaction with OH radicalsand isestimated to be2.2years. Theatmospheric oxidation of HFE-7500gives a fluorinated acetate, n-C3F7CF(OC(O)CH3)CF(CF3)2, and a fluorinated formate, n-C3F7CF(OC(O)H)CF(CF3)2. The atmospheric fate of these two esters will probably be hydrolysis, which will produce n-C3F7CF(OH)CF(CF3)2 and CH3C(O)OH, and n-C3F7CF(OH)CF(CF3)2 and HC(O)OH. Acetic acid and formic acid are ubiquitous naturally occurring compounds in the atmosphere and the additional burden associated with the oxidation of HFE-7500 is of no consequence. In contrast, n-C3F7CF(OH)CF(CF3)2isnot anaturally occurring compound. By analogy with the well-known behavior of CF3OH, we expect that the fluorinated alcohol n-C3F7CF(OH)CF(CF3)2 will undergo heterogeneous decomposition to give HF + n-C3F7C(O)CF(CF3)2. The perfluoroketonen-C3F7C(O)CF(CF3)2will not react with OH radicals, NO3 radicals, or ozone. While there are no available data concerning the rate of photolysis or hydrolysis of n-C3F7C(O)CF(CF3)2, such data exist for hexafluoroacetone. Hexafluoroacetone (CF3COCF3) is believed to be removed from the atmosphere via both photolysis and hydrolysis on a time scale of 5-10 days (30). It is likely that a similar fate awaits n-C3F7C(O)CF(CF3)2; further work is needed to clarify the atmospheric fate of this compound. With regard to theenvironmental impact of HFE-7500we can make the following three statements. First, HFE-7500 does not contain any chlorine and will not contribute to stratospheric ozone depletion via the well-established chlorinebased chemistry. Aswith all hydrofluorocarbons(HFCs) and hydrofluoroethers(HFEs) theozonedepletion potential of HFE-7500iszero.Second,theatmosphericlifetimeof HFE7500isapproximately2.2years. Usingthemethod of Pinnock et al. (31) and the IR spectra of HFE-7500 shown in Figure 3A and CFC-11 reported elsewhere (6), we calculate instantaneous forcings for HFE-7500 and CFC-11 of 0.37 W/ m2 and 0.26 W/ m2, respectively. Values of the GWP (global warming potential) for HFE-7500 (relative to CFC-11) can then be estimated (32): GWPHFE-7500 ) IFHFE-7500 IFCFC-11 H FE-7500M CFC-11 CFC-11M H FE-7500 1 - exp(-t/ HFE-7500) 1 - exp(-t/ CFC-11) where IFHFE-7500, IFCFC-11, MHFE-7500, MCFC-11, HFE-7500, and CFC-11 are the instantaneous forcings, molecular weights, and atmospheric lifetimes of the two species and t is the time horizon over which the forcing is integrated. Using HFE-7500 ) 2.2 years and CFC-11 ) 50 years (31), we estimate that theGWP of HFE-7500is 0.06for a20-year horizon and 0.02 for a 100-year time horizon. Because of its short atmospheric lifetime, HFE-7500 has a small GWP. Emission of HFE-7500 into the atmosphere will not contribute significantly to radiative forcing of global climate change. Third, the atmospheric oxidation of HFE-7500 will produce a fluorinated acetate, n-C3F7CF(OC(O)CH3)CF(CF3)2, and a fluorinated formate, n-C3F7CF(OC(O)H)CF(CF3)2, neither of which is expected to be persistent or pose any significant environmental hazard. Acknowledgments Wethank John Owens(3M SpecialtyMaterials) for supplying thesamplesof HFE-7500used in thiswork, Roc Carter (Ford) for help in interpretation of IRspectra,3M SpecialtyMaterials for funding the work performed at MIT, and the Japanese Governmentfor aNEDO grantand aMonbusho sciencegrant for the priority field "Radical Reactions" which made this collaborative research project possible. Literature Cited (1) Molina, M. J.; Rowland, F. S. Nature 1974, 249, 810. (2) Farman, J. D.; Gardiner, B. G.; Shanklin, J. D. Nature1985, 315, 207. (3) J. Owens, 3M SpecialtyMaterials, privatecommunication, 1999. (4) Wallington, T. J.; Schneider, W. F.; Sehested, J.; Bilde, M.; Platz, J.; Nielsen, O. J.; Christensen, L. K.; Molina, M. J.; Molina, L. T.; Wooldridge, P. W. J. Phys. Chem. A. 1997, 101, 8264. (5) Christensen L. K.; Sehested, J.; Nielsen, O. J.; Bilde, M.; Wallington, T. J.; Guschin, A.; Molina, L. T.; Molina, M. J. J. Phys. Chem. A 1998, 102, 4839. (6) Ninomiya, Y.; Kawasaki, M.; Guschin, A.; Molina, L. T.; Molina, M.; Wallington, T. J. Environ. Sci. Technol. 2000, 34, 2973. (7) 3M Technical Information Sheet 2001. (8) Wallington, T. J.; Schneider, W. F.; Worsnop, D. R.; Nielsen, O. J.;Sehested,J.;Debruyn,W.J.;Shorter,J.A. Environ.Sci.Technol. 1994, 28, 320. (9) Tyndall, G. S.; Cox, R. A.; Grainer, C.; Lesclaux, R.; Moortgat, G. K.; Pilling, M. J.; Ravishankara, A. R.; Wallington, T. J. J. Geophys. Res. 2001, 106, 12157. (10) Tang,Y.AtmosphericFateof VariousFluorocarbons.MSThesis, Massachusetts Institute of Technology, Cambridge, MA, 1993. (11) Wallington, T. J.; Japar, S. M. J. Atmos. Chem. 1989, 9, 399. (12) DeMore W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.; Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina, M. J. JPL Publication No. 94-26; NASA Jet Propulsion Lab: Pasadena, CA, 1997. (13) Atkinson, R. J. Phys. Chem. Ref. Data. Monogr. 2, 1994. (14) Stein, S. E.;Rukkers, J. M.;Brown, R. L. NIST Standard Reference Database 25: Structures and Properties; NIST: Gaithersburg, MD, 1991. (15) Prinn, R. G.; Weiss, R. F.; Miller, B. R.; Huang, J.; Alyea, F. N.; Cunnold, D. M.; Fraser, P. J.; Hartley, D. E.; Simmonds, P. G. Science 1995, 269, 187. (16) Spivakovsky, C. M.; Logan, J. A.; Montzka, S. A.; Balkanski, Y. L.; Foreman-Fowler, M.; Jones, D. B. A.; Horowitz, L. W.; Fusco, A. C.; Brenninkmeijer, C. A. M.; Prather, M. J.; Wofsy, S. C.; McElroy, M. B. J. Geophys. Res. 2000, 105, 8931. (17) Wallington, T. J.;Hurley, M. D. Chem. Phys. Lett. 1992, 189, 437. (18) Wine, P. H.; Semmes, D. H. J. Phys. Chem. 1983, 87, 3572. (19) Wallington, T. J.; Hurley, M. D.; Ball, J. C.; Jenkin, M. E. Chem. Phys. Lett. 1993, 211, 41. (20) Wallington, T. J.; Skewes, L. M.; Siegl, W. O.; Wu, C.-H.; Japar, S. M. Int. J. Chem. Kinet. 1988, 20, 867. (21) Wingenter, O. W.; Blake, D. R.; Blake, N. J.; Sive, B. C.; Rowland, F. S.; Atlas, E.; Flocke, F. J. Geophys. Res. 1999, 104, 21819. (22) Goto,M.;Kawasaki,M.;Wallington, T.J.;Hurley, M.D.;Sharratt, A. P. Int. J. Chem. Kinet. 2002, 34, 139. VOL. 36, NO. 11, 2002 / ENVIRONM ENTAL SCIENCE & TECHNOLOGY 9 2401 (23) Christensen, L. K.; Wallington, T. J.; Guschin, A.; Hurley, M. D. J. Phys. Chem. A 1999, 103, 4202. (24) Christensen, L. K.; Ball, J. C.; Wallington, T. J. J. Phys. Chem. A. 2000, 104, 345. (25) Nohara, K.; Toma, M.; Kutsuna, S.; Takeuchi, K.; Ibusuki, T. Environ. Sci. Technol. 2001, 35, 114. (26) Wallington, T. J.; Guschin, A.; Stein, T. N. N.; Platz, J.; Sehested, J.; Christensen, L. K.; Nielsen, O. J. J. Phys. Chem. A 1998, 102, 1152. (27) Japar, S. M.; Wallington, T. J.; Richert, J. F. O.; Ball, J. C. Int. J. Chem. Kinet. 1990, 22, 1257. (28) Atkinson, R. Atmos. Environ. 1990, 24A, 1. (29) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F., Jr.; Kerr, J. A.; Rossi, M. J.; Troe, J. J. Phys. Chem. Ref. Data 1999, 28, 191. (30) Mgelberg, T. E.;Platz, J.;Nielsen, O. J.;Sehested, J.;Wallington, T. J. J. Phys. Chem. 1995, 99, 5373. (31) Pinnock, S.;Hurley, M. D.;Shine, K. P.;Wallington, T. J.;Smyth, T. J. J. Geophys. Res. 1995, 100, 23227. (32) Houghton, J. T.; Meira Filho, L. G.; Bruce, J.; Lee, H.; Callander, B. A.; Haites, E.; Harris, N.; Maskell, K. In ClimateChange1994: Radiative Forcing of Climate Change and an evaluation of the IPCC IS92 Emission Scenarios; Houghton, J. T., Ed.; Cambridge University Press: Cambridge, U.K., 1995; p 222 (Report of Working Groups I and III of the Intergovernmental Panel on climate change). Received for review October 23, 2001. Revised manuscript received March 6, 2002. Accepted March 12, 2002. ES0113798 2402 9 ENVIRONM ENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002