Document ykqaXRqvkMVQLRQQJVMQwj1g6

Final Report July 2003 AR226-1521 STUDY OF THE ATMOSPHERIC FATE OF FLUORINATED ALCOHOLS RECEIVED OPPT NCIC 03 SEP 30 PM 13:30 CNRS-LCSR Contract with RAND Corp. for the Telomer Research Program Study of the atmospheric fate of fluorinated alcohols FOREWORD This work has been carried out by Georges LE BRAS (principal investigator, CNRSLCSR), Valrie BOSSOUTROT (Post-doc, CNRS-LCSR) and Isabelle MAGNERON (Post-doc, CNRS-LCSR). The authors thank the following persons : Pr. H. SIDEBOTTOM and T. KELLY from University College Dublin, Dr. K. WIRTZ from CEAM (Valencia) for their contributions related to this research and J. M. LIBRE from TotalFinaElf (TFE) for his fruitful cooperation within the joint CNRS-TFE collaboration. - 1 - Study of the atmospheric fate of fluorinated alcohols SUMMARY CF3CH2CH2OH C6F13CH2CH2OH C8F17CH2CH2OH Trifluoropropanol (TFP) Tridecafluoro-octanol (TDO) 2-perfluoro-n-octyl ethanol (2POE) The central aim of the study was to characterize the atmospheric fate of C8F17CH2CH2OH (2POE). Preliminary tests made in the photoreactor of Orlans showed that this substance could not be properly investigated directly because if its too low vapor pressure. The alternative approach was to investigate shorter chain fluorinated alcohols as surrogates of 2POE. In this respect, the rate constant of CF3CH2CH2OH (TFP) and C6F13CH2CH2OH (TDO) with OH radical, the major atmospheric oxidant, have been measured and the OH-initiated oxidation mechanism of TFP in air has been determined at room temperature. The studies have been carried out using two photoreactors : a 150 L Teflon bag irradiated by lamps at CNRS-Orlans for the kinetics and mechanistic studies, and the 200 m3 European photoreactor, EUPHORE, irradiated by sunlight at Valencia (Spain) also for the mechanistic study. The rate constant measured using a relative method are k = (1,06 0,10) x 10-12 for OH + TFP reaction and k = (0,99 0,18) x 10-12 for the OH + TDO reaction at 298 K (units of cm3 molecule-1 s-1). The same values obtained indicate that the nature of the perfluorinated group has no influence on the reactivity of the alcohol. Therefore, the above value of the rate constant can be recommended for the reaction OH + 2POE, i.e. k = (1,0 0,2) x 10-12 cm3 molecule-1 s-1 at 298 K. For the OH-initiated oxidation of TFP, the major products observed are the fluoroaldehydes CF3CH2CHO and CF3CHO, in the presence and absence of NOX. PAN like compounds were also observed in the presence of NOX. Fluoroacids were not observed under both conditions. The yield of the primary aldehyde, CF3CH2CHO, was found close to 50 % within the uncertainty of measurements, both in the presence and absence of NOX. However a yield close to unity can not be excluded considering the different uncertainties, specially that on the rate constant of the OH + CF3CH2CHO reaction which has to be taken into account to determine the yield of CF3CH2CHO. The obtained results are consistent with a mechanism where the primary OH-reaction proceeds by H-atom abstraction from the CH2 group in -position of the -OH group. Then CF3CH2CHO and CF3CHO are successively formed, CF2O being the end product. PAN compounds can also be formed in the presence of NOX. This mechanism can be generalised for the larger fluoroalcohols, including TDO and 2POE. In this case the intermediate aldehydes are CnF2n+1CH2CHO and CnF2n+1CHO. Regarding the atmospheric implications of the obtained results, the atmospheric lifetime ( = 1/ k[OH]) of TFP, TDO and 2POE is estimated to be 11 days. Then, these fluoroalcohols could be transported over long distances before being oxidised. However, this lifetime is too short to make these compounds significant contributors to radiative forcing. Concerning the oxidation products, the fluoroaldehydes CnF2n+1CH2CHO will have a short lifetime (~ 4 days) resulting from their reaction with OH (photolysis would be less important as suggested by EUPHORE experiments). The fluoroaldehydes subsequently formed, CnF2n+1CHO, will react much - 2 - Study of the atmospheric fate of fluorinated alcohols more slowly with OH, the lifetime of the reference compound CF3CHO toward OH being ca 20 days (k OH + CF3CHO ~ 5 x 10-13 cm3 molecule-1 s-1 at 298 K, Scollard et al. 1993). This lifetime is comparable to that of the uptake of such soluble species by clouds (WMO, 1994). However, the importance of cloud processing of the fluoroaldehydes will depend on their loss rate by photodissociation which needs to be better defined. (The EUPHORE experiments have suggested a lifetime of the fluoroaldehydes higher than 6 days). The gas phase degradation of CnF2n+1CHO (OH reaction or photolysis) will ultimately produce CF2O which is relatively soluble and will be removed by uptake in cloud or rainwater. The importance of the alternative PAN like compounds will depend on their stability. However their contribution as NOX reservoir should be negligible regarding to PAN which is produced from hydrocarbons emitted in much larger amounts compared to fluoroalcohols. - 3 - Study of the atmospheric fate of fluorinated alcohols CONTENTS I - Bibliography ..................................................................... 5 I- 1/ State of knowledge ................................................................. 5 I- 2/ Context of the study ................................................................. 6 II - Experimental methods II- 1/ Experimental technique ............................................ 6 ....................................................... 6 II- 2/ Experimental conditions of the kinetic study ............................... 9 II- 3/ Experimental conditions of the mechanistic study .............................. 11 III - Kinetic and mechanistic results .................................... 13 III- 1/ Kinetic results ............................................................................ 13 III- 2/ Mechanistic results ................................................................... 16 IV - Atmospheric implications ............................................. 22 IV-1 / Summary of the results ....................................................... 22 IV-2 / Atmospheric impact ................................................................ 23 References ....................................................................... 25 - 4 - Study of the atmospheric fate of fluorinated alcohols I- Bibliography I- 1/ State of knowledge The existing studies on fluorinated alcohols include only short chain compounds (CF3OH, CF3CH2OH, CH2FCH2OH, CHF2CH2OH et CF3CH2CH2OH). The studies only consist in the determination of the reaction rate constants with OH radicals or chlorine atoms. No study has been published so far on the OH-initiated oxidation mechanism of this class of compounds. However mechanistic data as well as kinetic data have been recently obtained for CF3CH2OH by Kelly et al. [private communication] at University College of Dublin (UCD) (Professor Sidebottom). The results for CF3CH2CH2OH are presented below. CF3CH2OH Kinetic study The rate constant of the reaction CF3CH2OH + OH has been determined at room temperature by two groups : Wallington et al. [1988] : k OH = 9.55 x 10-14 cm3 molecule-1 s-1 Tokuhashi et al. [1999] : k OH = 1.07 x 10-13 cm3 molecule-1 s-1 The recent study performed at UCD provides at room temperature : k OH = 1.12 x 10-13 cm3 molecule-1 s-1 The average value k OH ~ 1 x 10-13 cm3 molecule-1 s-1 implies a mean atmospheric lifetime of CF3CH2OH towards OH of approximately three months. This value of k OH is significantly different from the estimated value obtained by the Structure-Activity-Relationship (SAR) of Atkinson et al. [1994] : k OH = 4 x 10-13 cm3 molecule-1 s-1. Mechanistic study The study of the OH + CF3CH2OH by Kelly et al. [private communication] was carried out at the EUPHORE facility in Valencia (Spain) in air in the absence of nitrogen oxides. The observed produtcs are CF3CHO, CF2O and CF3O3CF3. The suggested mechanism is based on the predominant initial attack of OH at the -CH2- group. Other compounds studied - 5 - Study of the atmospheric fate of fluorinated alcohols Recently Tokuhashi et al. [1999] have determined the rate constant for the reaction CF3CF2CH2OH + OH. The obtained result (kOH = 1.05 x 10-13 cm3 molecule-1 s-1 at room temperature) indicates that the addition of one CF2 group does not influence the -CH2- group reactivity. The length of the perfluorinated chain seems to have negligible influence on the reactivity of the -CH2-group bonded to the alcohol function. Besides, the Cl-initiated oxidation of CF3CH2CH2OH in air has been studied in Teflon bags at UCD (Kelly et al. [private communication]). The observed products are CF3CH2CHO, CF3CHO and CF2O. I- 2/ Context of the study The review of the literature shows that very few data exists for the gas phase oxidation of fluorinated alcohols, specially for the OH-initiated oxidation which is expected to be the predominant atmospheric loss process of these compounds. Moreover there is no data available for the CnH2n+1CH2CH2OH to which belongs C8F17CH2CH2OH (2POE) the central compound of the project. Considering that similar reactivity and mechanism are expected for this class of compounds, and that experimental investigation of 2POE might be difficult as a solid at room temperature, it seems appropriate to investigate also CF3CH2CH2OH (TFP) and possibly C6F13CH2CH2OH (TDO) as surrogates of 2POE. II- Experimental methods II- 1/ Experimental technique II- 1/ 1 Kinetic study The rate constants have been measured in a simulation chamber (built within the collaboration with TotalFinaElf, figure II-1). The chamber is a Teflon bag of 150 L irradiated by UV lamps. The OH radicals are produced from the photolysis of H2O2 at 254 nm. The species are sampled at different reaction times and analysed by Fourier Transformed InfraRed (FTIR) Spectroscopy and Gas phase Chromatography coupled with a Flame Ionisation Detection (GC/FID). The time profiles of reactants consumption and of products formation can be established in this way. The FTIR and GC-FID analysis performed at Orlans are automatic allowing a high number of samples (around 20 for an experiment of 4 hours). - 6 - Study of the atmospheric fate of fluorinated alcohols The IR analysis is made in a Nicolet White gas cell of 2 L and 25 cm length with an optical multipath of 10 m. The cell is equipped with valves to work either under flow or static conditions. We have worked under static conditions, with samplings of 2 L from the reaction mixture for each analysis at atmospheric pressure. The cell can be heated (from 20 to 185 C) for cleaning purpose when some surface adsorption problems are encountered. A volume of around 100 cm3 is sampled from the chamber and flushed into the overall chromatographic system. The flame ionisation detector (FID) allows detection of a large variety of organic compounds except short chain compounds such as formaldehyde (HCHO). The mass spectrometer coupled to GC allows to follow the evolution of the ionised fragments with time. This type of detection gives some qualitative information about the products formed. Concentration time profiles of reactants can be obtained by FTIR or GC analysis. The reaction rate constant measured from these profiles is obtained from the relative rate method using a reference compound. Liquid reactants are introduced into the bag through a flow of purified air by flushing a known amount of liquid placed in a bubbler. The photoreactor is then filled to its full capacity at atmospheric pressure with purified air. The alcohol, the reference compound and H2O2 are successively introduced into the reactor. Before starting the reaction by switching on the lamps, at least three analysis are performed to check the homogeneity and the stability of the gas mixture. Then the consumption of the reactants is measured. The OH radicals formed by photolysis react with both reactants : OH + Ref products OH + VOC products kref (known) kVOC (to be determined) Ref : reference compound for which reaction rate constant with OH is known, VOC : Volatil Organic Compound under study, kVOC : rate constant to be measured. By definition, and - d[Re f ] = kref[OH][Re f ] dt - d[VOC] = kVOC[OH][VOC] dt 1 d[Re f ] = 1 d[VOC] kref [Re f ] kVOC [VOC] Considering [VOC]0 and [Ref]0 as the initial concentrations, and [VOC]t and [Ref]t concentrations at reaction time t , and by integration, we obtain : 1 [Ref ]t d ln[Re f ] = 1 [ VOC ] t d ln[VOC] kref [Ref ]0 kVOC [VOC]0 - 7 - Study of the atmospheric fate of fluorinated alcohols 1 ln [Re f ]0 = 1 [VOC]0 ln k ref [Re f ]t kVOC [VOC]t ln [VOC]0 = kVOC ln [Re f ]0 [VOC]t kref [Re f ]t Chromatographic or spectrographic peak areas being proportional to compound concentrations, one can plot : A0 A0 ln VOC t = f ln Re f t AVOC ARef where AX0 et AXt are the peak areas of compound X at reaction time 0 and t. The slope of this fit corresponds to the ratio kVOC/kref, yielding kVOC. To check the reliability of the measured rate constants, many experiments have been performed varying the reactant and the OH concentrations through the number of lamps switched on. Moreover, for each VOC two reference compounds have been used. The rate constant recommended is the average of all the values obtained with the two reference compounds. The selected reference compounds have reaction rate constant with OH close to that expected for the fluorinated alcohols studied (TFP and TDO). II- 1/ 2 Mechanistic study The experiments have been performed using the simulation chamber at Orlans (figure II1) described above and the european photoreactor EUPHORE, installed at Valencia in Spain, (figure II-2). The EUPHORE facility consists of two independent hemispherical outdoor simulation chambers, made of FEP foil, with a volume of 200 m3 each (http://www.gva.es./ceam). The equipment is composed of many instruments (FTIR, GC-FID, HPLC, ...) for in-situ analysis or after sampling. The detection sensitivity is high enough to use VOC and NOX concentration ranges close to those present in the atmosphere. - 8 - Study of the atmospheric fate of fluorinated alcohols Figure II-1 : Teflon chamber and FTIR spectrometer. Figure II-2 : Photoreactor irradiated by sunlight (EUPHORE). The high volume of EUPHORE allows to make many samplings without any perturbation of the reaction mixture. The low area/volume ratio (1 m-1) limits the wall artifacts. However, some experimental parameters (temperature, sunlight) can not be controlled. In addition the experiment preparation (cleaning, getting optimized sunlight conditions) is time consuming. This limits the number of experiments for each compound (one per day). The complementarity between EUPHORE and the chamber of Orlans enables to compare the obtained results with the two setups under different conditions. Moreover, the number of experiments made at EUPHORE being limited and expensive, the preliminary experiments performed at Orlans enable to define the optimum conditions for EUPHORE experiments. The quantification of the products were obtained using the calibration performed in both experimental set-up. The calibrations were made either by introducing known concentrations of authentic samples (CF3CH2CH2OH and CF3CH2CHO) or by using our IR library. Two campaigns (two weeks) were performed at EUPHORE in june 2002 and december 2002. - 9 - Study of the atmospheric fate of fluorinated alcohols II- 2/ Experimental conditions of the kinetic study CF3CH2CH2OH (TFP) The rate constant for the reaction of OH with CF3CH2CH2OH have been measured using the relative rate method at 298 K. The two reference compounds were : n-hexane and n-butyl formate, which have reaction rate constants with OH close to that expected for CF3CH2CH2OH. The reactants were detected using FTIR and GC/FID analysis. The collected spectra using FTIR are obtained by the accumulation of 50 scans and with 1 cm-1 of resolution. The GC analysis is performed using a DB-1 column with He as a carrier gas. The temperature used are the following : 80C or 60 C for the column oven, 180C for the injector and 280 C for the detector. The retention time of TFP under these conditions is 2.6 and 3.5 minutes at 80 and 60 C respectively. The injected volume of H2O2 using a seringua through a septum is 50 L. The photolysis of TFP was negligible (lower than 1 % per hour) under the experimental conditions used (Table II-1). The TFP wall loss was also negligible (< 2 % per hour). Such observations were also made for the reference compounds used. Experience Source of OH Reference compound [TFP]0 (ppm) [Rf]0 (ppm) Experiment time 1 H2O2 2 H2O2 3 H2O2 n-hexane n-hexane n-hexane 48 47 3.5 hours 47.8 89 3 hours 67.5 66.7 3 hours 4 H2O2 n-butyl formate 63.9 77.1 3 hours 5 H2O2 n-butyl formate 56.6 55 4 hours Table II-1 : Experimental conditions for the kinetic study of the reaction CF3CH2CH2OH + OH performed at Orlans. C6F13CH2CH2OH (TDO) The rate constant has been determined using the relative rate method at 298 K at Orlans. The study has been performed using FTIR spectrometry and GC-FID under the same experimental conditions as for TFP study. - 10 - Study of the atmospheric fate of fluorinated alcohols C6F13CH2CH2OH having a very low vapor pressure (< 1 torr at 298 K), it could not be introduced directly into the Teflon bag in the gas phase. Then a known volume of liquid TDO was deposited on sintered glass flushed through purified air into the bag (around 20 L of liquid corresponds to 20 ppm of gaseous TDO in the bag). As for TFP, photolysis and wall loss of TDO were negligible (Figure II-3) under the experimental conditions used (Table II-2). 10000 8000 GC/FID area (arbitrary unit) 6000 4000 2000 0 0 20 40 60 80 100 Exposition time (minutes) Figure II-3 : Wall losses of C6F13CH2CH2OH. Experience OH source Reference compound [Ref]0 (ppm) Experiment time 1 H2O2 n-hexane 87.3 3 hours 2 H2O2 n-hexane 78.3 3.5 hours 3 H2O2 formate de n-butyle 72.7 3 hours 4 H2O2 formate de n-butyle 97 4 hours Table II-2 : Experimental conditions of the kinetic study of the reaction C6F13CH2CH2OH + OH performed at Orlans. II- 3/ Experimental conditions of the mechanistic study CF3CH2CH2OH (TFP) The OH-initiated oxidation mechanism of CF3CH2CH2OH has been studied in the absence of NOx (two experiments) and in the presence of NOx (one experiment) under conditions given in Table II-3. - 11 - Study of the atmospheric fate of fluorinated alcohols Experience OH source [TFP]0 (ppm) [NOX]0 (ppm) H2O2 injected volume Lamps at 254 nm Experiment time 1 H2O2 66.8 / 50 L 4 3.5 hours 2 H2O2 53.9 / 50 L 3 5 hours 3 H2O2 60 57 50 L 3 4.5 hours Table II-3 : Experimental conditions of the mechanistic study of the reaction CF3CH2CH2OH + OH performed at Orlans. C6F13CH2CH2OH (TDO) The OH-initiated oxidation mechanism of C6F13CH2CH2OH has been studied in the presence of NOx (two experiments). The homogeneity of TDO in the chamber was obtained after one hour. Many samples have confirmed that there was no wall loss of C6F13CH2CH2OH. OH radicals were produced from the photolysis of H2O2 using 3 lamps at 254 nm. The analyses were performed using FTIR spectrometry and GC/FID. The infrared spectrum of C6F13CH2CH2OH is shown on figure II-4. Absorbance 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0 3924 3683 3442 3201 2960 2719 2478 2237 1996 1754 1513 1272 1031 790 Wavenumber(cm-1) Figure II-4 : IR spectrum of C6F13CH2CH2OH obtained at Orlans. - 12 - Study of the atmospheric fate of fluorinated alcohols Some difficulties have been noticed during spectral analysis due to the strong absorption of C-F bonds at 1250 cm-1. For absorbance higher than unity, concentration is not directly proportionnal to the absorbance. Moreover, the saturation of the detector at 1250 cm-1 makes the subtraction of the infrared spectra of TDO difficult and not reliable. In contrast, the sensitivity of GC/FID for TDO is very low. The concentration of TDO (~20 ppm) introduced in the bag is the maximum compatible with linearity of absorbance with concentration. But for this concentration the FTIR sensitivity is insufficient to detect the products. II- 3/ 1 Study at EUPHORE CF3CH2CH2OH (TFP) Two experiments of the OH-initiated oxidation of trifluoropropanol have been performed in the EUPHORE photoreactor in june 2002 in collaboration with UCD. HONO, the OH radical precursor, was added during the overall experiment in order to increase the OH concentration to get sufficient TFP consumption. OH source [TFP]0 (ppb) Irradiation time 27 june 2002 HONO 939 6 hours 28 june 2002 HONO 1709 5 hours Table II-4 : Experimental conditions of the OH-initiatied oxidation study of TFP in the presence of NOX at EUPHORE. C6F13CH2CH2OH (TDO) et C8F17CH2CH2OH (2POE) During the campaign performed in decembre 2002, the experiments on the TDO and 2POE oxidation did not provide valuable information on the oxidation mechanism of these compounds. This was due to a too low consumption of the alcohols resulting from the low solar radiation available in this period of the year. Nevertheless, 2POE was successfully introduced into the chamber at a concentration around 120 ppb (by strong heating). III- Kinetic and mechanistic results III- 1/ Kinetic results Reaction OH + TFP - 13 - Study of the atmospheric fate of fluorinated alcohols The rate constant of the OH + TFP reaction has been measured with the two reference compounds : n-hexane and n-butyl formate. Figures III-1 and III-2 show examples of relative kinetic plots obtained. 0.25 kTFP/kRf = 0.200.01 0.2 ln ([TFP]0/[TFP]t) 0.15 0.1 0.05 0 0 0.4 0.8 1.2 ln ([n-hex]0/[n-hex]t) Figure III-1 : Reaction OH + CF3CH2CH2OH : Relative decay plot (reference compound : n-hexane). 0.25 kTFP/kRf = 0.300.01 0.2 ln ([TFP]0/[TFP]t) 0.15 0.1 0.05 0 0 0.2 0.4 0.6 0.8 ln ([FB]0/[FB]t) Figure III-2 : Reaction OH + CF3CH2CH2OH : Relative decay plot (reference compound : nbutyl formate). - 14 - Study of the atmospheric fate of fluorinated alcohols The rate constant data are summarised in Table III-1. Reference Compound n-hexane n-butyl formate kTFP / kRf 0.200.05 0.290.03 kRf (10-12 cm3 molecule-1 s-1) 5.450.16 a 3.540.52 b k TFP (10-12cm3 molecule-1 s-1) 1.080.14 1.030.06 Table III-1 : Reaction rate constants for OH + TFP obtained using a relative rate method. a Donahue et al. 1998, b Le Calv et al. 1997. The mean value obtained for the rate constant of the OH + TFP reaction at 298 K is : kTFP = (1.06 0.10) x 10-12 cm3 molecule-1 s-1 Reaction OH + TDO The rate constant has been determined at (298 3) K using hexane and n-butyl formate as reference compounds. Figure III-3 presents two examples of C6F13CH2CH2OH decays versus 0.25 n-butyl formate n-hexane 0.2 ln([C6F13CH2CH2OH]0/[C6F13CH2CH2OH]t) 0.15 0.1 0.05 0 0 0.4 0.8 1.2 ln([reference]0/[reference]t) reference compounds decays. The obtained data are reported in Table III-2. Figure III-3 : Reaction OH + CF3CH2CH2OH : Relative decay plots (reference compounds : n-hexane and n-butyl formate). - 15 - Study of the atmospheric fate of fluorinated alcohols Reference Compound n-hexane n-butyl formate Analytical technique FTIR GC-FID FTIR GC-FID FTIR GC-FID FTIR GC-FID kTDO / kRef 0.15 0.01 0.26 0.02 0.17 0.02 0.21 0.02 0.22 0.01 0.29 0.02 0.27 0.03 0.24 0.05 k TDO (10-12cm3 molecule-1 s-1) 0.820.08 1.42 0.15 0.93 0.14 1.14 0.14 0.78 0.15 1.03 0.22 0.96 0.25 0.85 0.30 Table III-2 : Rate constant data for the reaction OH + TDO using the relative rate method at Orlans. The given error is calculated according to : k TDO = k TDO x [((kTDO/kRef) / kTDO/kRef) + (kRef / kRef)] where (kTDO/kRef) is equal to twice standard deviation for the slope (k TDO / kRef) and kRef, is the error on the rate constant of the OH reaction with the reference compound. From these results, the average rate constant for the reaction OH + C6F13CH2CH2OH is : k = (0.99 0.18) x 10-12 cm3 molecule-1 s-1 at 298 K No kinetic data for the reaction OH + C6F13CH2CH2OH has been published so far. However, we can notice that the rate constant for this reaction ((0.99 0.20) x 10-12 cm3 molecule-1 s-1) is very close to that for the reaction of OH with CF3CH2CH2OH ((1.06 0.10) x 10-12 cm3 molecule-1 s-1, this work). Therefore the nature of the perfluorinated group has no influence on the reactivity of the compound. Similarly, Tokuhaschi et al. [1999] have shown that CF3CH2OH (k = 1.0x10-13 cm3 molecule-1 s-1) and CF3CF2CH2OH (k = 1.02x10-13 cm3 molecule1 s-1) have the same reactivity towards OH. Besides, the much higher reactivity of CF3CH2CH2OH and C6F13CH2CH2OH compared to CF3CH2OH and CF3CF2CH2OH confirms that fluorinated groups strongly deactivate the -CH2- group in -position. This indicates that for CF3CH2CH2OH and C6F13CH2CH2OH the OH attack occurs predominantly on the -CH2- group in -position to the -OH group of the alcohol. These results indicate that the reactivity of C8F17CH2CH2OH with OH will be similar to that of CF3CH2CH2OH and C6F13CH2CH2OH. Then we recommended for the rate constant of OH + C8F17CH2CH2OH at 298 K: k (1.0 0.2) x10-12 cm3 molecule-1 s-1. - 16 - Study of the atmospheric fate of fluorinated alcohols III- 2/ Mechanistic results III- 2/ 1 OH-initiated oxidation mechanism of CF3CH2CH2OH (study at LCSR) TFP oxidation without NOX Trifluoropropanol and its oxidation products have been analysed using FTIR. Spectra collected at the beginning and at the end of the reaction are presented on figure III-4. The spectrum obtained at t=0 shows TFP, which has been quantified from its absorption bands between 1238 and 1275 cm-1. The major oxidation products observed are trifluoropropanaldehyde (CF3CH2CHO) and trifluroroacetaldehyde (CF3CHO). t = 0 TFP Absorbance t = 200 min ar 29 11:56:26 2002 CF3CH2CHO TFP 3000 2500 2000 1500 Wavenumber (cm-1) 1000 Figure III-4 : OH-initiated oxidation of TFP without NOX : FTIR spectra before reaction (t = 0) and at the end of the reaction (t = 200 min). Temporal profiles are presented on figure III-5. The calculation of the yield of the primary product CF3CH2CHO, [CF3CH2CHO]formed / [TFP]consumed, needs to correct the CF3CH2CHO concentration for the reaction of CF3CH2CHO with OH which is significant within the time constant of the experiment, as shown by the shape of the concentration-time profiles of CF3CH2CHO. However, the rate constant of the OH + CF3CH2CHO reaction is not well known, but it can be postulated that this latter is definitely lower than that of the OH + CH3CHO (k = 1.5 x 10-11 cm3 molecule-1 s-1) due to the inhibiting effect of the -CF3 group in CF3CH2CHO. Recently preliminary data obtained by Kelly et al. [private communication] indicate a value of k ~ 3 x 10-12 cm3 molecule-1 s-1 for the OH + CF3CH2CHO reaction. The corrected yield of - 17 - IR area (TFP, CF3CHO, CF2O) (arbitrary unit) d[CF3CH2CHO] (ppm) Study of the atmospheric fate of fluorinated alcohols CF3CH2CHO using k OH + CF3CH2CHO = (3.0 0.6) x 10-12 cm3 molecule-1 s-1 is found to be (42 13)% (Figure III-6). Considering the given error on the yield and the additionnal error on the calibration of CF3CH2CHO and the present uncertainty on k OH + CF3CH2CHO, a yield close to unity can not be excluded. The CF3CH2CHO calibration was done at EUPHORE by introducing a known volume of the compound in the reactor (3 points were obtained). 6 TFP CF3CH2CHO CF3CHO CF2O 4 0.3 10 CF3CH2CHO yield = (36 4) % 8 0.2 6 IR area CF3CH2CHO (arbitrary unit) 4 2 0.1 2 0 0 0 40 80 120 160 200 Time (minutes) Figure III-5 : OH-initiated oxidation of CF3CH2CH2OH in the absence of NOX : Time-profiles of the main products. 0 0 10 20 30 -d[TFP] (ppm) Figure III-6 : Concentration of CF3CH2CHO formed (corrected for the OH+CF3CH2CHO reaction) versus concentration of TFP consumed. The formation of the OH-initiated oxidation products of TFP in the absence of NOX can be explained by the mechanism given in Figure III-7. CF3CH2CHO is formed in the Habstraction pathway from the -CH2- group in -position to the alcohol fonction. The corresponding oxidation produces first CF3CH2CHO and the HO2 radical as co-product. The reaction of the primary aldehyde with OH gives the secondary one, CF3CHO from H-abstraction from the -CH2- group in a chain involving peroxy (RO2) and oxy (RO) radicals. CF3CHO reacts with OH in a chain producing the end product CF2O which has been observed. Concerning the trioxyde (CF3O3CF3) also observed, it is expected to be produced following the mechanism of Figure III-7. However, in the natural atmosphere (low level of NOX), the CF3O2 radical (precursor of CF3O3CF3 in the laboratory) will react with HO2 or CH3O2, the main peroxyl radicals present to produce the peroxides such as CF3OOH or CF3OOCH3. These two will go to CF3O from photolysis or reaction with OH and then to CF2O. - 18 - Study of the atmospheric fate of fluorinated alcohols CF3 CH2 CH2 OH + OH CF3 CH2 CHOH + H2O O2 CF3 CH2 CH OH OO + O CF3 CH2 C H OH + O CF3 CH2 C HO2 H2O O2 CF3 CH2 O C OO CF3CH2COO2 + O CF3 CH2 C O2 O CF3 CH2 + CO2 O2 CF3 CH2 OO CF3CH2O2 CF3 + O2 CF3O2 CO CF3O2 CF3 CH2 + O2 O O2 O CF3 C + H OH HO2 O CF3 C + H2O CF3O + O2 parois CF3O2 CF2O CF3O3CF3 Figure III-7 : OH-initiated oxidation scheme of CF3CH2CH2OH in the absence of NOx. The products in bold have been observed. TFP oxidation in the presence of NOX - 17 - Study of the atmospheric fate of fluorinated alcohols The study in the presence of NOX have shown that the main products formed from the oxidation of CF3CH2CH2OH are trifluoropropanaldehyde (CF3CH2CHO) with a yield of (57 5) % and perfluorinated PAN like compounds (CF3CH2C(O)O2NO2, and CF3C(O)O2NO2). In the spectra, in the presence of NOX, after subtraction of TFP, CF3CH2CHO, other infrared structures are still present. The residual peaks are at 1840, 1750, 1300, 1051 and 791 cm1, and are similar to those of PAN (CH3C(O)OONO2) (see figure III-8). Moreover, refering to the study of Chen et al. [1992] and [1993] about reactions of CF3O2 and CF3O with NO2 and CF3O with NO, the bands attributed to CF3O2NO2 are at 1764, 1296 et 787 cm-1 and those of CF3ONO2 mainly at 1744 and 791 cm-1. The secondary aldehyde (CF3CHO) detected in this work in the absence of NOX, could not be observed in the presence of NOX. Its infrared absorption bands could be ultimately covered by those of the PAN like compounds. Futhermore, CF2O and CF3O3CF3 are not observed indicating that the reactions CF3CH2O2 + NO et CF3CH2O + NO2 are favoured compared to the mutual recombinaison which occurs in the absence of NOX. The proposed mechanism is shown in figure III-9. We also observe the formation of HNO3 which very likely comes from the reactions NO + HO2 OH + NO2 and OH + NO2 HNO3. HO2 radicals being the co-products of the reaction producing CF3CH2CHO. e d c HNO3 b a NO2 CF3CH2CHO HNO3 TFP 1900 1600 1300 1000 700 Wavenumbers (cm-1) Figure III-8 : FTIR spectra collected during the OH-initiated oxidation of TFP in the presence of NOX. Spectrum a) at t = 0, b) at t = 208 mn from which TFP have been subtracted, c) subtraction of HNO3 from previous spectrum, d) after subtraction of CF3CH2CHO, e) spectrum of PAN. - 18 - Study of the atmospheric fate of fluorinated alcohols CF3 CH2 CH2 OH + OH CF3 CH2 CH OH + H2O O2 OO CF3 CH2 CH OH + H CF3 CH2 C O HO2 H2O OH + CF3 CH2 C O NO2 O2 O CF3 CH2 C NO OO NO2 CF3 CH2 O C NO2 O CF3 CH2 C OO NO2 O CF3 CH2 C O CF3CH2 + CO2 O2 CF3 CH2 OO NO CF3 CH2 O + H CF3 C HO2 OH O CF3 C O O2 CO + CF3 O2 CF3 OO O2 NO RH CF3 O NO O CF3 C NO OO NO2 O CF3 C O CF3 + CO2 CF3OH CF2O + FNO O CF3 C OO NO2 Figure III-9 : OH-initiated oxidation scheme of CF3CH2CH2OH in the presence of NOx. The products in bold have been observed. III- 2/ 1 OH-initiated oxidation mechanism of C6F13CH2CH2OH (study at LCSR) Two experiments have been conducted in the laboratory to study the OH-initiated oxidation of C6F13CH2CH2OH at room temperature, atmospheric pressure and in the presence of nitrogen oxides. - 19 - Study of the atmospheric fate of fluorinated alcohols The response by FTIR of C6F13CH2CH2OH was very good compare to its response by GC. However, the FTIR signal at the 1250 cm-1 band is not specific of C6F13CH2CH2OH since its fluorinated oxidation products have very similar bands. Then, the subtraction of these bands led to large errors in the quantification of the signal due to C6F13CH2CH2OH, specially because the initial amount (20L ~ 20 ppm) of C6F13CH2CH2OH introduced was limited to maintain linearity of the concentration-absorbance plots. On the other hand, this concentration of 20 ppm is not high enough to accurately monitor the formation of the products. III- 2/ 2 OH-initiated oxidation mechanism of CF3CH2CH2OH (study at EUPHORE) The experiments performed in june 2002 confirm the observations made at Orlans, i.e. the formation of CF3CH2CHO and CF3CHO as the major products. CF2O was not observed because the OH source used, the photolysis of HONO, produces NO. In the presence of NOX, CF2O is not formed as we observed at Orlans in agreement with the mechanism of figure III-9. Calibration of CF3CH2CHO has also been performed at EUPHORE allowing to determine the yield of CF3CH2CHO by correcting the CF3CH2CHO concentration-time profiles for the OH reaction as described above. The obtained yield of CF3CH2CHO, (47 7) %, is similar to that obtained at Orlans. However, as discussed previously, the rate constant used to correct CF3CH2CHO concentrations is not well known and a yield close to unity can not be excluded. Temporal profiles of trifluoropropanol, CF3CH2CHO and of CF3CHO IR peak areas are presented on figure III-10. Concentration-time profiles of NO, NO2 and ozone are shown on figure III-11. The plot of the yield CF3CH2CHO is given figure III-12. 2000 1600 1200 800 2 400 Trifluoropropanol Trifluoropropanaldehyde Trifluoroacetaldehyde 1.6 300 1.2 200 0.8 NO NO2 Ozone 400 0.4 100 0 0 50 100 150 200 Temps d'irradiation (min) 0 250 Figure III-10 : OH-initiated oxidation of CF3CH2CH2OH : - 20 - 0 0 100Irradiation tim2e0(0min) 300 Temporal profiles of TFP and products. Figure III-11 : Concentrations (ppb) Aire CF3CHO Concentrations (ppb) Study of the atmospheric fate of fluorinated alcohols OH-initiated oxidation of CF3CH2CH2OH : Temporal profiles of NO, NO2 and ozone. 200 CF3CH2CHO yield = (48 2)% 160 d[CF3CH2CHO] (ppb) 120 80 40 0 0 100 200 300 400 -d[TFP] (ppb) Figure III-12 : Concentration of CF3CH2CHO formed (corrected for the OH + CF3CH2CHO reaction) versus concentration of TFP consumed in the OH-initiated oxidation of TFP (EUPHORE). During this campaign, some photolysis experiments of the two aldehydes CF3CH2CHO and CF3CHO, produced from the reaction OH + CF3CH2CH2OH, have been performed in order to determine their photodissociation rates under atmospheric conditions. The total loss rates for the aldehydes have been found of the order of the dilution rate in the photoreactor measured from the first order loss of SF6 added in the reaction mixture. Therefore, only an upper limit of the photolysis rate of the aldehydes can be given : J < 2 x 10-6 s-1. This corresponds to an atmospheric lifetime for the two aldehydes towards photolysis ( = 1/J) higher than six days. This J value is much lower than that calculated for CF3CHO (J = 0.7 x 10 -4 s-1) from the known value of the absorption cross sections , and assuming a dissociation quantum yield of unity (Rattigan et al. 1998). Our J value suggests a much lower quantum yield. It has to be mentionned that CCl3CHO would behave differently since a J value of ~ 4 x 10-5 s-1 has been estimated from the known values (which are close to those of CF3CHO in the actinic spectrum region) and assuming a photodissociation quantum yield of unity from the measured yield of unity for Cl atoms at = 308 nm (Talukdar et al. 2001). Clearly, further study is required to better define the photodissociation rate of CF3CHO. In this respect, further studies under simulated irradiation conditions at EUPHORE are suggested where the monitoring of CF2O produced in the photodissociation of CF3CHO or CnF2n+1CHO will yield a more accurate value of J compared to that obtained in the present study from the decay of the aldehydes. The results obtained at EUPHORE are consistent with the mechanism of Figure III-9 discussed above. The ozone formation is due to the NO to NO2 conversion in different steps of - 21 - Study of the atmospheric fate of fluorinated alcohols the mechanim and further photolysis of NO2 to give O atoms which recombine with O2 to yield ozone. III- 2/ 2 OH-initiated oxidation mechanism of C8F17CH2CH2OH (study at EUPHORE) During the campaign performed in decembre 2002, the experiments on the 2POE oxidation did not provide valuable information on the oxidation mechanism. This was due to a too low consumption of the alcohol resulting from the low solar radiation intenisty available in this period of the year. Nevertheless, 2POE was successfully introduced into the chamber at a concentration around 120 ppb (by strong heating of the solid sample). In summary, the mechanistic studies performed at Orlans and at EUPHORE show that the main OH-initiated oxidation products of TFP are CF3CH2CHO and CF3CHO in the absence and presence of NOX, whereas PAN like compounds (CF3CH2C(O)OONO2, CF3C(O)OONO2) could also be significant in the presence of NOX. IV- Atmospheric implications IV- 1/ Summary of results The main results obtained in this study at ambiant temperature are summarised in table IV-1. - 22 - Study of the atmospheric fate of fluorinated alcohols Nature of the study Kinetics reaction TFP + OH Kinetics reaction TDO + OH Results kTFP = (1.06 0.10) x 10-12 cm3 molecule-1 s-1 kTDO = (0.99 0.18) x 10-12 cm3 molecule-1 s-1 Observations Two reference compounds Two reference compounds Mechanism TFP + OH in air Products : CF3CH2CHO = (42 13) %* CF3CHO Products : CF3CH2CHO = (51 9) %* without NOX (LCSR) with NOX (LCSR and EUPHORE) PAN like Table IV-1 : Kinetic and mechanistic results. (* see text) The results obtained allows to propose the following simplified mechanism for the atmospheric OH-initiated oxidation of fluoroalcohols CnF2n+1CH2CH2OH, including in particular TFP, TDO and 2POE. CnF2n+1CH2CH2OH + OH CnF2n+1CH2CHO CnF2n+1CH2C(O)OONO2 CnF2n+1CHO CnF2n+1C(O)OONO2 n CF2O IV- 2/ Atmospheric impact The rate constants measured in this study for the reaction OH + CF3CH2CH2OH (TFP) and C6F13CH2CH2OH (TDO) allow to estimate the atmospheric lifetime for these compounds towards OH. Considering a mean tropospheric concentration for OH radicals, [OH] = 1 x 106 radicals cm-3 (average for 24 hours), we obtain : - 23 - Study of the atmospheric fate of fluorinated alcohols TFP = 1 / kTFP[OH] 11 days TDO = 1 / kTDO[OH] 11 days Since the rate constant for the reaction OH + C8F17CH2CH2OH (2POE) is expected to be the same as for the reactions of OH with TFP and TDO, as previoulsy discussed, the lifetime of C8F17CH2CH2OH towards OH will be the same as those of TFP and TDO : 2POE = 1 / k2POE[OH] 11 days The reaction with OH being very likely the major loss process, this lifetime indicates that these fluorinated alcohols, including 2POE, could be transported over long distances before being oxidised. They may have a regional impact. However this lifetimes is too short to make these compounds significant contributors to radiative forcing. Concerning the oxidation products, the fluoroaldehydes CnF2n+1CH2CHO will have a short lifetime (~ 4 days) resulting from their reaction with OH (photolysis would be less important as suggested by EUPHORE experiments). The fluoroaldehydes subsequently formed, CnF2n+1CHO, will react much more slowly with OH, the lifetime of the reference compound CF3CHO toward OH being ca 20 days (k OH + CF3CHO ~ 5 x 10-13 cm3 molecule-1 s-1 at 298 K, Scollard et al. 1993). This lifetime is comparable to that of the uptake of such soluble species by clouds (WMO, 1994). However, the importance of cloud processing of the fluoroaldehydes will depend on their loss rate by photodissociation which needs to be better defined. (The EUPHORE experiments have suggested a lifetime of the fluoroaldehydes higher than 6 days). The gas phase degradation of CnF2n+1CHO (OH reaction or photolysis) will ultimately produce CF2O which is relatively soluble and will be removed by uptake in cloud or rainwater. The importance of the alternative PAN like compounds will depend on their stability. However their contribution as NOX reservoir should be negligible regarding to PAN which is produced from hydrocarbons emitted in much larger amounts compared to fluoroalcohols. - 24 - Study of the atmospheric fate of fluorinated alcohols References Wallington T. J., Dagaut P., Kurylo M. J. Correlation between gas-phase and solution-phase reactivities of hydroxyl radicals toward saturated organic compounds. J. Phys. Chem. 92, 5024, 1988. Tokuhashi K., Nagai H., Takahashi A., Kaise M., Kondo S., Sekiya A., Takahashi M., Gotoh Y., Suga A. Measurement of the OH reaction rate constants for CF3CH2OH, CF3CF2CH2OH, and CF3CH(OH)CF3. J. Phys. Chem. A: 103, 2664 (1999). Kelly T., Sidebottom H. Private communication (2002). Donahue N. M., Anderson J.G., Demerjian K.L. New rate constants for ten OH alkane reactions from 300 to 400 K: an assessment of accuracy. J. Phys. Chem. A: 102, 3121-3126 (1998). Le Calv S., Le Bras G., Mellouki A. Temperature dependence for the rate coefficients of the reactions of OH radical with a series of formates. J. Phys. Chem. A: 101, 5489-5493 (1997). Chen J., Young V., Niki H. FTIR spectroscopic study of the reaction of CF3O with NO: evidence for CF3O + NO and CF2O + FNO. J. Phys. Chem. 96, 6115 (1992). Chen J., Young V., Niki H. Long path fourier transform infrared spectroscopic study of the reactions of CF3OO and CF3O radicals with NO2. J. Phys. Chem. 97, 11696 (1993). Rattigan O. V., Wild O., Cox R. A. UV absorption cross-sections and atmospheric photolysis lifetimes of hlogenated aldehydes. J. Photochem. and Photobiol. 112, 1 (1198). Talukdar R. K., Mellouki A., Burkholder J. .B, Gilles M. K., Le Bras G., Ravishankara A. R. Quantification of the tropospheric removal of chloral (CCl3CHO): rate coefficient for the reaction with OH, UV absorption cross sections, and quantum yields. J. Phys. Chem., 105, 5188 (2001). Scollard D. J., Treacy J. J., Sidebottom H. W, Balestra-Garcia C., Laverdet G., Le Bras G., Mac Leod H., Tton S. - 25 - Study of the atmospheric fate of fluorinated alcohols Rate constants for the reactions of hydroxyl radicals and chlorine atoms with halogenated aldehydes. J. Phys. Chem., 97, 4683 (1993). Scientific Assessment of Ozone Depletion : 1994 Report n 37, World Meteorological Organization. - 26 -