Interactive comment on “ Tropospheric photooxidation of CF 3 CH 2 CHO and CF 3 ( CH 2 ) 2 CHO initiated by Cl atoms and OH radicals ”

Abstract. The absolute rate coefficients for the tropospheric reactions of chlorine (Cl) atoms and hydroxyl (OH) radicals with CF3CH2CHO and CF3(CH2)2CHO were measured as a function of temperature (263–371 K) and pressure (50–215 Torr of He) by pulsed UV laser photolysis techniques. Vacuum UV resonance fluorescence was employed to detect and monitor the time evolution of Cl atoms. Laser induced fluorescence was used in this work for the detection of OH radicals as a function of reaction time. No pressure dependence of the bimolecular rate coefficients, kCl and kOH, was found at all temperatures. At room temperature kCl and kOH were (in 10−11 cm3 molecule−1 s−1): kCl(CF3CH2CHO) = (1.55±0.53); kCl(CF3(CH2)2CHO) = (3.39±1.38); kCl(CF3CH2CHO) = (0.259±0.050); kCl(CF3(CH2)2CHO) = (1.28±0.24). A slightly positive temperature dependence of kCl was observed for CF3CH2CHO and CF3(CH2)2CHO, and kOH(CF3CH2CHO). In contrast, kOH(CF3(CH2)2CHO) did not exhibit a temperature dependence over the range investigated. Arrhenius expressions for these reactions were: kCl(CF3CH2CHO) = (4.4±1.0)×10−11 exp{−(316±68)/T} cm3 molecule−1 s−1 kCl(CF3(CH2)2CHO) = (2.9±0.7)×10−10 exp{−(625±80)/T} cm3 molecule−1 s−1 kOH(CF3CH2CHO) = (7.8±2.2)×10−12 exp{−(314±90)/T} cm3 molecule−1 s−1 The atmospheric impact of the homogeneous removal by OH radicals and Cl atoms of these fluorinated aldehydes is discussed in terms of the global atmospheric lifetimes, taking into account different degradation pathways. The calculated lifetimes show that atmospheric oxidation of CF3(CH2)x CHO are globally dominated by OH radicals, however reactions initiated by Cl atoms can act as a source of free radicals at dawn in the troposphere.


Introduction
Initially, among the candidates for replacing CFCs were the hydrofluorocarbons (HFCs) and hydrochlorofluorocarbons (HCFCs) (Atkinson et al., 1989).Although HFCs have a zero Ozone Depletion Potential (ODP), HCFCs do affect stratospheric ozone because of the presence of chlorine atoms in these species.Moreover, both of them are in general very strong greenhouse gases with high Global Warming Potentials (GWP), commonly associated with long tropospheric lifetimes and strong absorption in the IR region.Recently, hydrofluoroethers (HFEs) and partially fluorinated alcohols, like the family of CF 3 (CH 2 ) x CH 2 OH, have been suggested as alternatives for HFCs and HCFCs.
CF 3 CH 2 OH and CF 3 CH 2 CH 2 OH have negligible GWPs (Sellevåg et al., 2004a, Waterland et al., 2005) and a zero ODP, since the −CF 3 group does not cause significant ozone depletion (Ravishankara et al., 1994).On the other hand, atmospheric photooxidation of these fluorinated alcohols, initiated by hydroxyl (OH) radicals and chlorine (Cl) atoms, is known to yield fluorinated aldehydes as major products (Kelly et al., 2005;Hurley et al., 2004 and2005;Sellevåg et al., 2004a).Therefore, a complete understanding of the atmospheric fate of CF 3 (CH 2 ) x CHO would be needed in order to better evaluate the environmental implications of these hydrofluoroalcohols as CFCs substitutes.
As far as we know, only kinetic studies on the reactions of CF 3 CHO and CF 3 CH 2 CHO with the main oxidizing agents in the atmosphere, OH radicals and Cl atoms, have been reported almost exclusively by relative rate methods at room temperature (Atkinson et al., 2008;Kelly et al., 2005;Hurley et al., 2005;Sellevåg et al., 2004b).Neither OH and Cl kinetic data on other trifluoroaldehydes nor kinetic data on the reactions with NO 3 and O 3 were found in the literature.So, the aim of this work is to determine the absolute rate coefficients for the OH-and Cl-reaction with 3,3,3trifluoropropanal (CF 3 CH 2 CHO) and 4,4,4-trifluorobutanal (CF 3 (CH 2 ) 2 CHO) as a function of temperature (T = 263-371 K): Pulsed Laser Photolysis (PLP) techniques were employed in this work to generate the transient species, Cl atoms and OH radicals, which were detected and monitored by Resonance Fluorescence (RF) and Laser Induced Fluorescence (LIF), respectively.An expression for the temperature dependence of k Cl and k OH is recommended in this paper, which allows the derivation of the rate coefficients, and the corresponding lifetime due to the homogeneous removal, for these reactions at tropospheric temperatures.The possible atmospheric impact of these secondary pollutants will be discussed in terms of their gas-phase reactivity and other degradation routes.

Experimental and procedure
Kinetic measurements on the set of reactions 1 and 2 have been performed by UV laser pulsed photolysis techniques described earlier (Albaladejo et al., 2002 and2003;Jiménez et al., 2005 and2007).A brief description of the PLP/RF experimental set-up is given in Sect.2.1.The PLP/ LIF experimental system, described briefly in Sect.2.2, was modified in order to measure in situ the fluoroaldehyde concentration.A detailed description of this latter set-up is presented in Sect.2.3.Analysis procedure of the kinetic data is described in Sect.2.4.

Cl kinetics: PLP/RF system
A Pyrex jacketed reaction cell with an internal volume of about 200 cm 3 was employed in the kinetic measurements of reactions 1 as a function of temperature.Cl atoms ((2.6-4.8)×10 11radical cm −3 ) were produced by the photolysis of Cl 2 ((2.1-3.7)×10 14molecule cm −3 ) at 308 nm using a pulsed XeCl excimer laser (Lambda Physik LPX 105i, at 10 Hz).Laser fluence ranged from 6.1 to 9.5 mJ pulse −1 cm −2 .The concentration of Cl atoms was monitored as a function of time by resonance fluorescence at λ∼135 nm.Ground state Cl( 3 P 3/2 ) was excited by the 135 nm radiation from a microwave-driven lamp, situated perpendicular to the photolysis laser beam, through which He containing a low concentration of Cl 2 (0.2-0.3%) was flowed.The resonance fluorescence (RF) radiation was recorded with a solar blind photomultiplier, PMT (Hamamatsu R6835) positioned perpendicular to both the resonance lamp and photolysis laser beam, after passing through two CaF 2 lenses.The regions between the two lenses and between the last lens and the photomultiplier were pumped in order to remove ambient O 2 , which absorbs the RF at 135 nm.
Signals were obtained using photon counting electronics and multichannel scaling data storage.The output signal from the PMT was first inputted into a fast preamplifier and then collected with a multichannel scaler (EG&G Instruments T914P) and stored in a coupled microcomputer for subsequent analysis.Data acquisition by the multichannel scaler was triggered between 0.1 ms and 10 ms (ca.10% of the total temporal scale) prior to the photolysis laser in order to obtain the background signal which was subtracted from the post photolysis signal to obtain the temporal profile of the Cl atoms.The Cl temporal profiles following 5000 to 30 000 laser pulses were coadded to improve the signal-to-noise ratio.
All experiments were carried out in a large excess of buffer gas (He).Diluted fluoroaldehydes and Cl 2 /He mixtures were slowly flowed from 10-L bulbs into the reaction cell by means of electronic mass flow controllers.Concentrations of the different gases were measured from their mass flow rates and the total pressure.The total flow rate (180-470 sccm) was fast enough to prevent the accumulation of photolysis or reaction products, but slow enough for considering the mixture static during the measurements.

OH kinetics: PLP/ LIF system
A schematic diagram of the experimental set-up modified for this work is shown in Fig. 1.The Pyrex reaction cell was identical to that used in the Cl kinetics.Total flow rate ranged from 211 to 542 sccm.OH(X 2 ) radicals were generated by the pulsed photolysis of H 2 O 2 (g) at 248 nm by using a KrF excimer laser which operates at a repetition rate of 10 Hz.Gas-phase H 2 O 2 was introduced in the reaction cell by bubbling the carrier gas (He) through a pre-concentrated liquid sample of hydrogen peroxide (Jiménez et al., 2005).Upper limits for H 2 O 2 concentrations in the reaction cell, were estimated as described in our previous paper (Jiménez et al., 2007), and were in the range of (0.7-4.8)×10 14 molecule cm −3 .Therefore, OH concentration can be estimated from the upper limit of the H 2 O 2 concentration, taking into account the absorption cross section of H 2 O 2 at 248 nm, the quantum yield for OH  production (Sander et al., 2006), and the photolysis laser fluence ((2.0-8.4)mJ pulse −1 cm −2 ).OH radical concentration was around 3.0×10 11 molecule cm −3 in all experiments.At different reaction times, OH radicals were excited at 282 nm (OH(A 2 + , v = 1)←OH(X 2 , v = 0)) by using the frequency-doubled output of a dye laser pumped by a Nd-YAG laser.LIF signal from OH(A 2 + ) was detected at λ≥282 nm by a photomultiplier tube placed orthogonal to both light sources.LIF emissions were selected by a bandpass filter (90% transmittance at 350 nm and 150 nm of FWHM) and transferred to a personal computer for analysis.

Optical measurement of [CF
An accurate measurement of CF 3 (CH 2 ) x CHO concentrations is essential to report a rate coefficient within reasonable quoted uncertainties.As in our previous works (Jiménez et al., 2005 and2007;Lanza et al., 2008), firstly, the absorption cross sections at 185 nm (σ λ=185nm ) were measured in order to use them in the determination of reactant concentrations during the kinetic experiments.These experiments were conducted at pressures inside the absorption cell between 0.2 and 8.4 Torr of pure fluorinated aldehydes.σ λ=185nm were obtained from the slope of the plots of absorbance (in base e) versus pressure and using a large number of experiments (30 runs).σ λ=185nm were found to be (4.82±0.20)×10−20 cm 2 molecule −1 and (9.42±0.19)×10−20 cm 2 molecule −1 for CF 3 CH 2 CHO and CF 3 (CH 2 ) 2 CHO, respectively.The stated uncertainty was ±2σ (only statistical error).Some examples of the Beer-Lambert plots at 185 nm are given in Fig. 1S of the supplementary information http://www.atmos-chem-phys.net/10/1911/2010/acp-10-1911-2010-supplement.pdf for both aldehydes.The low absorption cross section (several orders of magnitude lower than non-fluorinated aldehydes) did not al- low the measurement of aldehyde concentrations during the kinetic experiments, since the decrease in the intensity of the incident radiation was not measurable (absorbances much lower than 0.1%).
For that reason, concentrations of CF 3 (CH 2 ) x CHO were measured in situ during the kinetic experiments by Fourier transform infrared (FTIR) spectroscopy, mainly at the exit of the reaction cell, by using the experimental set-up presented in Fig. 1.In some experiments, CF 3 (CH 2 ) x CHO concentration was measured before entering the reaction cell and no difference outside the experimental error was observed.These concentration measurements were performed in a borosilicate glass multipass cell of variable pathlength from 1 to 8 m (Specac, Cyclone C5) sealed with ZnSe windows (transmittance range: 450-17 000 cm −1 ).The FTIR cell is provided with protected gold mirrors and vacuum/ gas inlet and outlet taps.The path length was set by means of a low power diode laser (0.8 mW at 635 nm) (Specac, P/N 24500 Benchmark™ Series Laser Alignment Accessory).All experiments were performed with a fixed pathlength of 8 m.A FTIR spectrometer (Bruker, TENSOR 27), with a cube corner interferometer which eliminates the alignment errors, was used to record the FTIR spectra of both aldehydes.A N 2 -liquid cooled mercury cadmium telluride (MCT) detector was used during the quantification process.The FTIR spectra of diluted fluorinated aldehydes during the calibration were recorded at a total pressure between 1.3 and 96 Torr in He.The dilution factors in the storage bulbs, f , defined as An example of the FTIR spectrum range of diluted CF 3 CH 2 CHO and CF 3 (CH 2 ) 2 CHO recorded in the 500-4000 cm −1 region is shown in Fig. 2S of the supplementary information http://www.atmos-chem-phys.net/10/1911/2010/acp-10-1911-2010-supplement.pdf.The four absorption bands used in the quantification of the fluorinated aldehydes during the kinetic experiment are listed in  integrated IR absorbance for a given band between ν1 and ν2 , A int , is expressed according to the Beer-Lambert law by: where A(ν) is the absorbance at a wavenumber ν, σ (ν) is the absorption cross section at that wavenumber, and is the optical pathlength.The integrated absorption cross section, S int , can be defined as follows: Thus, the expression of the Beer-Lambert law is simplified to: S int was then obtained from the linear least-squares analysis of the plots of Reation (R5).An example of such plots is presented in Fig. 2. For each band and fluorinated aldehyde, S int are listed in Table 1.During the kinetic experiment an average of the aldehyde concentration extracted from these bands was taken into account in order to derive the OH rate coefficients.The aldehyde concentration measured by FTIR was usually lower than that measured from the flow rate (up to 20% difference), probably due to some heterogeneous process in the storage bulb and the tubing.

Kinetic data analysis
Reactions (R1) and (R2) were studied under pseudo-first order conditions with fluoroaldehydes ([CF 3 (CH 2 ) x CHO] = (0.1-8.8)×10 14 molecule cm −3 ) and precursor concentrations in large excess with respect to the initial concentration of the oxidants.In both systems, the temporal profiles of the photochemically generated oxidant can be described by the following kinetic scheme: k oxid is the bimolecular rate coefficients for the reaction of Cl and OH with CF 3 (CH 2 ) x CHO, i.e., k Cl and k OH , respectively.In the absence of fluoroaldehyde, only Reactions (R7) and (R8) are taking place, where k prec is the bimolecular rate coefficient for that reaction and k diff is the first-order diffusion rate coefficient of the oxidant out of the detection zone.The reaction of Cl atoms with its photochemical precursor, Cl 2 , is negligible (Hutton and Wright, 1965), but the reaction of OH radicals with H 2 O 2 is a significant contribution to the decay of OH concentration (Sander et al., 2006, Jiménez et al., 2004).So, under these conditions, the oxidant concentration profile should follow a simple exponential rate law: where k is the pseudo-first order rate coefficient obtained from the analysis of Reaction k 0 ranged from 91 to 866 s −1 in the case of the reactions of OH radicals and k range was 234-2696 s −1 for CF 3 CH 2 CHO and 627-7606 s −1 for CF 3 CH 2 CH 2 CHO.For chlorine atoms, k 0 = k diff (Cl) which ranged from 20 to 46 s −1 at p T = 50-200 Torr of He.The bimolecular rate coefficients k oxid were obtained from the slopes of the plots of k against the fluoroaldehyde concentration (Eq.10) or the plots of Eq. ( 13): Examples of k -k diff and k -k 0 versus [CF 3 (CH 2 ) x CHO] plots are presented in Fig. 3a and b, respectively, at 298 K and total pressures between 50 and 215 Torr.At other temperatures (T = 263-371 K), k was measured for the titled reactions at a total pressure of 100 Torr for Cl-reactions and at total pressures between 50 and 215 Torr for OH-reactions.Several examples of the plots of Eq. ( 13) at three different temperatures are given in Fig. 4 for the Cl and OH reactions, respectively.
Temporal profiles of the OH LIF signal were always monoexponential.In contrast, preliminary kinetic experiments showed that Cl decays in the presence of the fluoroaldehydes were biexponential.This behaviour was attributed to the regeneration of Cl via the secondary reaction: where CF 3 (CH 2 ) x CO is the likely radical formed in the initial Cl-reaction of CF 3 (CH 2 ) x CHO.To make this secondary reaction negligible, oxygen ((1.5-2.2)×10 16molecule cm −3 ) was added to the reaction mixture through a mass flow controller.The Cl atom decay was then monoexponential, showing that the radical CF 3 (CH 2 ) x CO reacts predominantly with O 2 under these experimental conditions.
This observation is consistent with a rate ratio Maricq et al., 1995;Wallington et al., 1994).On the other hand, the addition of O 2 decreased the detection sensitivity of Cl atoms, as O 2 is an efficient quencher and absorber of the fluorescence of Cl atoms.Despite the reduction of the signal-to-noise ratio, [O 2 ] 0 was set between 1.0 and 2.5×10 16 molecule cm −3 over the T-range studied in order to ensure the minimization of Cl atom regeneration.The lowest concentration of O 2 was considered to be suitable to prevent the CF 3 (CH 2 ) x CO + Cl 2 reaction without decreasing the detection sensitivity of Cl atoms too much.Likewise, no difference on k Cl outside the error limits was observed at 298 K and other temperatures.Chemicals.Helium (UHP >99.999% from Praxair and Air Liquid) was employed as supplied.Cl 2 (99.9%,Praxair), CF 3 CH 2 CHO (>97%, Apollo-Scientific), and CF 3 CH 2 CH 2 CHO (95%, Apollo-Scientific) were degassed several times at 77 K before use by repeated freeze, pump, and thaw cycles.Purity of CF 3 (CH 2 ) x CHO samples was checked by GC-MS.

Absolute rate coefficients at 298 K and comparison with previous kinetic studies
As shown in Fig. 3, no pressure dependence of k Cl and k OH was observed between 50 and 215 Torr of helium.The same trend was observed at temperatures below and above room temperature.So, weighted averages of k Cl and k OH are reported in Tables 2 and 3, respectively.The weighting factor was taken as the inverse of the variance of each indi-vidual measurement.Uncertainties stated in Tables 2 and 3 are only statistical errors (±2σ ).Systematic errors account mainly for the uncertainties in the measurement of the aldehyde concentration, since errors in the measurement of temperature and pressure are much lower than 1%.Since in the Cl-kinetics aldehyde concentrations were not optically measured, an additional 20% was added to the statistical uncertainty.Thus, at room temperature the averaged rate coefficients obtained were (±2σ ):  As pointed out in the introduction section, only a few measurements of k Cl and k OH for CF 3 CH 2 CHO are found in the literature at room temperature (Hurley et al., 2005;Kelly et al., 2005;Sellevåg et al., 2004b).These previous studies were performed almost exclusively by relative methods.Therefore, k Cl measured in this work represents the first absolute value of that rate coefficient, as far as we know.Furthermore, our kinetic study on the Cl and OH reactions with CF 3 (CH 2 ) 2 CHO constitutes the first rate coefficients reported for this partially fluorinated aldehyde.Hurley et al. (2005).Some discrepancy observed in k Cl (CF 3 CH 2 CHO) in previous relative studies can be related with the use of different reference compounds, as discussed by Hurley et al. (2005).Regarding the OH-rate coefficient for CF 3 CH 2 CHO, Kelly et al. (2005) reported the only absolute k OH data employing the PLP/LIF technique, which is in agreement with the value reported in this paper and the relative one from Hurley et al. (2005), who used ethane and ethene as reference compounds.In this case, the discrepancy between the relative studies of Sellevåg et al. (2004b) and Hurley et al. (2005) remains unclear.
Room temperature k Cl and k OH for fluorinated aldehydes and the corresponding non-fluorinated aldehydes are compared in Tables 5 and 6, respectively.As can be seen, the fluorine substitution in aldehydes clearly reduces the Cl and OH reactivity.The rate constant ratio k(CF 3 (CH 2 ) x CHO)/k(CH 3 (CH 2 ) x CHO) ranges from 0.02 for the reaction of CF 3 CHO with Cl atoms to 0.51 for the OH-reaction with CF 3 (CH 2 ) 2 CHO.The observed reduction in reactivity (somewhat higher for the Cl reactions than for the OH reactions) must be associated with the inductive electron withdrawing of the -CF 3 group.This slight difference may be associated with the difference in the mechanisms for Cl and OH reactions.Mechanistic studies on the Cl-reaction with CF 3 CH 2 CHO confirmed that the main reaction pathway is H-atom abstraction from aldehydic group (Hurley et al., 2005).No product studies on the Cl and OH-reactions with CF 3 (CH 2 ) 2 CHO have been reported, as far as we know.Also as Tables 5 and 6 show, k Cl and k OH significantly increase with the length of the hydrocarbon chain, compared to the slight increase observed for the corresponding reactions with non-fluorinated aldehydes (Atkinson et al., 2006).Rate constant ratios k(CF 3 (CH 2 ) x CHO)/k(CF 3 (CH 2 ) x−1 CHO) range from ca. 2 to 9, while k(CH 3 (CH 2 ) x CHO) is only 20-50% larger that k(CH 3 (CH 2 ) x−1 CHO).

Absolute rate coefficients k Cl and k OH as a function of temperature
A summary of the absolute second-order rate coefficients for the reactions of Cl and OH with the studied aldehydes at all temperatures is given in Tables 2 and 3, respectively.The influence of temperature over the 263-371 K range on k Cl and k OH can be seen in the pseudo-first order plots given in Fig. 4 for each reaction and the corresponding Arrhenius plots presented in Fig. 5.The temperature dependence observed for the Cl-and OH-reactions with CF 3 CH 2 CHO was slightly positive, i.e. the rate coefficients slightly increase with temperature: This behaviour was also observed for the reaction of Cl atoms with CF 3 (CH 2 ) 2 CHO: In contrast, the corresponding reaction with OH radicals seems to proceed with no energy barrier, within the experimental errors: In this case, we recommend the weighted average of k OH between 263 and 358 K: A summary of the Arrhenius parameters, E a /R and A, and the room temperature rate coefficients for the reactions of Atmos.Chem.Phys., 10, 1911Phys., 10, -1922Phys., 10, , 2010 www.atmos-chem-phys.net/10/1911/2010/CF 3 (CH 2 ) x CHO and CH 3 (CH 2 ) x CHO (x = 0, 1, and 2) with Cl and OH is presented in Tables 5 and 6, respectively.As Table 6 shows, a positive E a /R was also reported for the OHreaction with CF 3 CHO (Laverdet et al., 1993).In the case of the reaction with Cl atoms no kinetic studies were found as a function of temperature.Positive activation energies are usually indicative that the major reaction pathway proceeds via a H-atom abstraction mechanism (Atkinson and Arey, 2003).In fact, product studies on Cl-and OH-reactions with CF 3 CHO (Sellevåg et al., 2004b;Sulbaek Andersen et al., 2004) and CF 3 CH 2 CHO (Hurley et al., 2005) confirm that the abstraction of the hydrogen atom from the -CHO group is the major reaction pathway, as mentioned above.In contrast, while k Cl (CH 3 CHO) did not show a temperature dependence (Atkinson et al., 2006), most of the reactions with non-fluorinated aldehydes exhibit a negative temperature dependence and similar E a / R factors.A negative temperature dependence is frequently associated with the formation of a long-lived addition complex before the hydrogen abstraction by OH radicals and subsequent decomposition of the com-plex, as suggested by Smith and Ravishankara (2002) for oxygenated organic compounds.

Tropospheric lifetimes of CF
Global tropospheric lifetimes of CF 3 (CH 2 ) x CHO due to homogeneous reactions with OH radicals and Cl atoms, τ homog , were calculated by using the rate coefficient k OH and k Cl and a 24-hour average oxidant concentration in the atmosphere of 1×10 6 radical cm −3 for OH (Krol et al., 1998) and 10 3 atom cm −3 for Cl (Singh et al., 1996).Under these conditions, τ OH for CF 3 CH 2 CHO was estimated to be 4 days at the surface and 6 days in the upper troposphere.In contrast, a constant τ OH for CF 3 (CH 2 ) 2 CHO of one day along the troposphere is reported here, while τ Cl for both CF 3 (CH 2 ) x CHO were estimated to be more than one year.Globally, the tropospheric removal of both fluorinated aldehydes is dominated by the reaction with OH radicals, i.e., τ homog is almost equal to τ OH .
In coastal areas and some industrialized zones, the Cl concentration can reach a peak value of 1.3×10 5 atom cm −3  ( Spicer et al., 1998) at dawn, where [OH] and solar actinic flux are very low.Under these circumstances, Cl atoms can contribute to the removal of CF 3 (CH 2 ) x CHO.At night the atmospheric degradation of fluorinated aldehydes is expected to be of minor importance, since lifetimes due to the reaction with NO 3 radicals are likely to be much longer than several days.
Other degradation routes to take into account in order to evaluate the atmospheric lifetimes of these CF 3 (CH 2 ) x CHO can be UV photolysis in the actinic region (λ>290 nm) and aqueous-phase atmospheric degradation.
The atmospheric lifetime of CF 3 CH 2 CHO with respect to UV photolysis has been estimated to be longer than 15 days (Sellevåg et al., 2004b;Chiappero et al., 2006).If the photolysis quantum yield ( λ ) of the fluorinated aldehyde depends on total pressure, the photolysis of CF 3 CH 2 CHO could be more important than reported at higher altitudes in the troposphere.As far as we know, no photodegradation study on CF 3 (CH 2 ) 2 CHO has been reported.Thus, a similar photolysis rate to CF 3 CH 2 CHO was assumed in this work for 4,4,4trifluorobutanal.Further studies on the pressure dependence of λ of both fluoroaldehydes are needed in order to better evaluate this removal process throughout the troposphere.
Little information is available for the aqueous-phase removal of fluorinated aldehydes.The atmospheric lifetime due to the water phase removal depends on the Henry's law coefficients k H,cp (=C/p, in M atm −1 ) and the hydrolysis rate coefficients, k hyd .Henry's law coefficients for some fluorinated oxygenated compounds have been compiled by Sander (1999).k H,cp ranges from 3 M atm −1 for FCHO to 160 M atm −1 for CHF 2 CF 2 CH 2 OH.No measurements of k H,cp have been found in the literature for other fluorinated aldehydes rather than FCHO.Based upon the air-water partition coefficients given above and assuming an annual rainfall rate of 580 mm yr −1 , as in previous works (Jiménez et al., 2009), the lifetime of CF 3 (CH 2 ) x CHO with respect to wet deposition can be assessed to be up to 50 years in the boundary layer (ca. 2 km) and 300 years in the free troposphere (ca. 10 km).Heterogeneous kinetics on these fluorinated aldehydes have not been studied, but it is expected that the uptake of CF 3 (CH 2 ) x CHO on or into the aqueous component of cloud/fog droplets or aqueous aerosol particles is not likely to be an important atmospheric sink for these compounds.However, further studies are needed in order to better evaluate this degradation route, since it can be a source of fluorinated acids, CF 3 (CH 2 ) x C(O)OH, in the atmosphere.
In conclusion, the removal of CF 3 CH 2 CHO by UV photolysis in troposphere might not be negligible compared to that by OH radicals.However, the removal of CF 3 (CH 2 ) 2 CHO by OH radicals might be its major atmospheric sink.Therefore, the products of their reactions with OH radicals and UV photolysis may significantly influence air quality.Similarly to Cl atoms (Hurley et al., 2005), major products of the OH-reactions with CF 3 CH 2 CHO and CF 3 (CH 2 ) 2 CHO are likely to be the corresponding fluorinated acids, CF 3 (CH 2 ) x C(O)OH, and CF 3 CHO and CF 3 CH 2 CHO, respectively, in clean atmospheres.Like CF 3 CHO, UV photolysis products of CF 3 (CH 2 ) x CHO at 308 nm (near the absorption maximum) are likely to be those from the radical channel, yielding formyl radical (HCO) and a fluoroalkyl radical, CF 3 (CH 2 ) x (Chiappero et al., 2006).Further reactions of all the species formed in such degradation processes may have a slight contribution to smog formation.However, additional photochemical studies would be needed to determine the products formed after the absorption of actinic radiation and to better evaluate the impact of these fluorinated aldehydes.

Conclusions
In this paper, we provide for the first time a detailed study on the temperature dependence of the absolute rate coefficients k Cl and k OH for reactions with CF 3 CH 2 CHO and CF 3 (CH 2 ) 2 CHO in the gas phase between 263 and 371 K. Activation energies (E a ) and pre-exponential factors (A) of reactions 1 and 2 have been reported here for the first time.Similar slightly positive activation energies of 2.6 kJ/ mol were obtained for the reactions of CF 3 CH 2 CHO with Cl and OH.The observed kinetic behaviour of CF 3 (CH 2 ) 2 CHO with respect to temperature seems to be different for both radicals.E a (Cl) is 5.2 kJ/ mol, while the OH-reaction has no energy barrier, within the error limits.We can observe an enhancement in the reactivity of higher fluorinated aldehydes by comparison with similar aldehydes possibly by a lower influence of the deactivation effect of CF 3 group, which seems to favour H-atom abstraction from the aldehydic group.The evaluation of the atmospheric lifetimes of the secondary pollutant CF 3 (CH 2 ) x CHO under global and specific conditions indicates that they are rapidly degraded in the troposphere.This implies that other pollutants, such as smaller fluorinated aldehydes, and free radicals are generated, which can contribute to air pollution.Also, the atmospheric degradation of these partially fluorinated aldehydes can be a source of CF 3 (CH 2 ) x C(O)OH in the troposphere.Lastly an important degradation process to be taken into account is the UV photolysis of CF 3 (CH 2 ) x CHO which seems to be an important removal route for these species.Future studies on the photolysis of CF 3 (CH 2 ) x CHO in the actinic region are needed to better quantify this removal process.

Figure 1 .
Figures and Captions

Fig. 1 .
Fig. 1.Experimental set-up for the on-line measurement of the concentration of CF 3 (CH 2 ) x CHO by FTIR spectroscopy during the OH-kinetic experiments.MFC: mass flow controller.

Fig. 2 .
Fig. 2. Plots of the Beer-Lambert law to obtain the integrated absorption cross sections, S int , for CF 3 CH 2 CHO and CF 3 CH 2 CH 2 CHO.

Figure 3 .Fig. 3 .
Figure 3. Examples of the pseudo-first order plots at room temperature for a) Cl reactions and b) OH reactions with CF 3 CH 2 CHO (white symbols) and CF 3 (CH 2 ) 2 CHO (black symbols).

Figure 4 .
Figure 4. Examples of the plots of k'-k diff (for Cl reactions) and k'−k 0 (for OH reactions) versus the concentration of aldehydes studied in this work at several temperatures.Solid lines are obtained from the least-squares analysis of the data.

Fig. 4 .
Fig. 4. Examples of the plots of k − k diff (for Cl reactions) and k − k 0 (for OH reactions) versus the concentration of aldehydes studied in this work at several temperatures.Solid lines are obtained from the least-squares analysis of the data.

Figure 5 .
Figure 5. Arrhenius plot of the experimental results obtained.The error bars of the individual points are ±2σ.

Fig. 5 .
Fig. 5. Arrhenius plot of the experimental results obtained.The error bars of the individual points are ±2σ .

Table 1 .
Integrated absorption cross sections, S int , for the four bands employed in the determination of fluorinated aldehyde concentrations.

Table 1
Plots of the Beer-Lambert law to obtain the integrated absorption cross sections, S int , for CF 3 CH 2 CHO and CF 3 CH 2 CH 2 CHO.

Table 2 .
Summary of experimental conditions and the obtained bimolecular rate coefficients for the reaction of Cl in the studied temperature range.

Table 3 .
Summary of experimental conditions and the obtained rate coefficients for the reaction of OH with the studied compounds in the range of temperature studied.

Table 4 .
Summary of the absolute second-order rate coefficients obtained in this work at 298 K together with those previously reported in the literature for the reaction of Cl atoms and OH radicals.

Table 4
summarizes our results at 298 K together with those previously reported for the reaction of Cl and OH with CF 3 CH 2 CHO.As can be seen in the table, the absolute rate coefficient k Cl (CF 3 CH 2 CHO) obtained in this work agrees with that of

Table 5 .
Cuevas et al. (2006)6)us parameters for the reaction of Cl atoms with the studied aldehydes obtained in this work along with those found in the literature for other fluorinated aldehydes and non-fluorinated aldehydes.This work, uncertainties stated in the text; cAtkinson et al. (2006), k Cl independent of temperature; dCuevas et al. (2006).

Table 6 .
Summary of the Arrhenius parameters for the reactions of OH radicals with the studied aldehydes obtained in this work along with those found in the literature for other fluorinated and non-fluorinated aldehydes.