Atmospheric Degradation Initiated by OH Radicals of the Potential

Article pubs.acs.org/est

Atmospheric Degradation Initiated by OH Radicals of the Potential Foam Expansion Agent, CF3(CF2)2CHCH2 (HFC-1447fz): Kinetics and Formation of Gaseous Products and Secondary Organic Aerosols Elena Jiménez,*,†,‡ Sergio González,† Mathieu Cazaunau,§ Hui Chen,§ Bernabé Ballesteros,†,‡ Véronique Dael̈ e,§ José Albaladejo,†,‡ and Abdelwahid Mellouki*,§ †

Departamento de Química Física, Facultad de Ciencias y Tecnologías Químicas, Universidad de Castilla-La Mancha (UCLM), Avenue de Camilo José Cela, s/n, 13071 Ciudad Real, Spain ‡ Research Institute on Combustion and Atmospheric Pollution (UCLM), Camino de Moledores 13071 Ciudad Real, Spain § Centre National de la Recherche Scientifique, Institut de Combustion Aérothermique Réactivité et Environnement ICARE/OSUC, CNRS 1C, Avenue de la Recherche Scientifique, 45071 Orléans cedex 02, France S Supporting Information *

ABSTRACT: The assessment of the atmospheric impact of the potential foam expansion agent, CF3(CF2)2CHCH2 (HFC-1447fz), requires the knowledge of its degradation routes, oxidation products, and radiative properties. In this paper, the gas-phase reactivity of HFC-1447fz with OH radicals is presented as a function of temperature, obtaining kOH (T = 263−358 K) = (7.4 ± 0.4) × 10−13exp{(161 ± 16)/ T} (cm3·molecule−1·s−1) (uncertainties: ±2σ). The formation of gaseous oxidation products and secondary organic aerosols (SOAs) from the OH + HFC-1447fz reaction was investigated in the presence of NOx at 298 K. CF3(CF2)2CHO was observed at low- and high-NOx conditions. Evidence of SOA formation (ultrafine particles in the range 10−100 nm) is reported with yields ranging from 0.12 to 1.79%. In addition, the absolute UV (190−368 nm) and IR (500−4000 cm−1) absorption crosssections of HFC-1447fz were determined at room temperature. No appreciable absorption in the solar actinic region (λ > 290 nm) was observed, leaving the removal by OH radicals as the main atmospheric loss process for HFC-1447fz. The major contribution of the atmospheric loss of HFC-1447fz is due to OH reaction (84%), followed by ozone (10%) and chlorine atoms (6%). Correction of the instantaneous radiative efficiency (0.36 W m−2·ppbv−1) with the relatively short lifetime of HFC-1447fz (ca. 8 days) implies that its global warming potential at a time horizon of 100 year is negligible (0.19) compared to that of HCFC-141b (782) and to that of modern foam-expansion blowing agents (148, 882, and 804 for HFC-152a, HFC-245fa and HFC-365mfc, respectively).

■

INTRODUCTION Since the Kyoto Protocol identified hydrofluorocarbons as industrial gases requiring emissions control because of their high global warming potential (GWP), several producers have started the development of hydrofluoroolefins (HFOs). For instance, 3,3,4,4,5,5,5-heptafluoro-1-pentene (CF3(CF2)2CH CH2, HFC-1447fz) is being considered as a potential CFC replacement of diclorofluoroethane (CH3CFCl2, HCFC-141b) as an expansion agent in polymeric foams including polyurethane. HCFC-141b is a potent greenhouse gas with a GWP at a horizon time of 100 years (GWP100‑yr relative to CO2) of 7821 and an ozone-depletion potential (ODP relative to CFC-11, CFCl3) of 0.11.2 For that reason, in the European Union or the United States, the use of HCFC-141b in foam applications has been banned, and its consumption will be gradually reduced from 2015 to 2030, when it will be completely phased out. An appropriate alternative to HCFC-141b, apart from the suitable physical properties, must present zero ODP and low GWP. © 2015 American Chemical Society

HFC-1447fz has a zero ODP; however, it may contribute to GWP due to the presence of C−F bonds that may absorb infrared (IR) radiation in the atmospheric window (800−1400 cm−1). For the evaluation of the potential contribution of HFC1447fz to global radiative forcing, its radiative properties need to be known together with its atmospheric lifetime (τ). Atmospheric lifetimes can be estimated from individual lifetimes due to all degradation routes, i.e., reaction with the main diurnal atmospheric oxidants (OH, Cl, O3, and NO3), UV photolysis and wet−dry deposition). No information regarding the radiative properties is currently available in the literature. Therefore, in this work, we report the first determination of the IR absorption cross-sections (σν̃) of HFC-1447fz in the midReceived: Revised: Accepted: Published: 1234

September 9, 2015 December 16, 2015 December 24, 2015 December 24, 2015 DOI: 10.1021/acs.est.5b04379 Environ. Sci. Technol. 2016, 50, 1234−1242

Article

Environmental Science & Technology infrared region (ν̃ = 500−4000 cm−1). In addition, the absolute rate coefficients of the reaction of HFC-1447fz with OH radicals (R1) are also reported for the first time as a function of total pressure (50−650 Torr of He) and temperature (T = 263−358 K). OH + CF3(CF2)2 CH=CH 2 → products

were filled with pure HFC-1447fz (pHFC‑1447fz = 4.2−9.2 Torr) or diluted mixtures of the HFC in He prepared and stored in 10 L glass bulbs (pT,bulb = 4.9−435 Torr; mixing ratios, f (= pHFC‑1447fz/pT,bulb), of 0.2−2.6%), respectively. In the IR absorption cell, total pressure (pT) ranged from 15 to 91 Torr and f for HFC-1447fz was 1.8%, i.e., concentrations of HFC-1447fz, [HFC-1447fz], ranged from 1.3 × 1015 to 7.8 × 1016 molecules cm−3 at room temperature. The UV detection system is formed by a 0.5 m focal-length spectrograph that has a 300 grooves/mm grating and a coupled-charged device cooled at 253 K by a Peltier system. The FTIR spectrometer has a liquid nitrogen cooled mercury cadmium telluride (MCT) detector. IR spectra were recorded at an instrumental resolution of 0.5 and 1 cm−1. At ICARE-CNRS, HFC-1447fz was introduced in a Teflon atmospheric simulation chamber (7300 L) and diluted by flowing purified dry air (0.05−0.085) Torr over atmospheric pressure.14 Diluted mixtures of the HFC (491−1122 ppbv) were prepared in situ by adding the liquid sample of HFC1447fz via an impinger and driven into the chamber at room temperature. The IR spectrum of HFC-1447fz was recorded at an instrumental resolution of 1 cm−1 using a FTIR spectrometer (Nicolet 5700 Magna) with a White-type mirror system providing a total path length of 146 m. Additionally, the recorded IR spectrum was used as a reference spectrum for quantifying HFC-1447fz during the product studies and for determining the wall-loss rate of the HFC in the chamber (see below). PLP-LIF System for Kinetic Studies. Hydroxyl radicals were produced in situ at the center of a Pyrex jacketed cell (200 cm3) by photodissociation at 248 nm of gaseous H2O2 or HNO3. Ranges of the upper limit of [H2O2] and [HNO3] are given in the Supporting Information. The photolysis energy, which varied from 2.9 to 9 mJ pulse−1 (10 Hz) from a pulsed KrF excimer laser was measured with a calorimetric disk at the exit of the reaction cell. Orthogonally to the photolysis beam, OH radicals were excited at 282 nm by using the doubled output of a Nd:YAG-pumped dye-laser. The laser-induced fluorescence (LIF) from the X2Π(ν″ = 0) ← A2Σ+(ν′ = 0) transition at ca. 309 nm was used to monitor the ground-state OH radicals. The OH LIF emission was detected by an optically isolated (band-pass filter centered at 309 nm with a fwhm of 10 nm) photomultiplier tube oriented perpendicular to the both photolysis and excitation beams. The time-resolved fluorescence was then used to monitor the evolution of OH concentration as a function of the reaction time, given by the delay time between the photolysis and the probe laser pulses. The reactor was heated or cooled by circulating a fluid (water and ethanol, respectively) from a thermostated bath through its external jacket. Mixing ratios of HFC-1447fz in the storage bulb ranged from 0.4 to 2.6%. Apart from flow measurements, the HFC concentration was also measured by FTIR spectroscopy at room temperature before and after passing the gas through the reactor, as described in the Supporting Information. Corrections to account the difference in total pressure and temperature between the IR cell and the reaction cell were made to derive [HFC-1447fz] used in the kinetic analysis for obtaining kOH(T). It was verified that [HFC-1447fz] from flow measurements differs in less than 14% and, typically around 7%, from those from the IR measurements ((0.14−1.6) × 1015 molecules cm−3).

k OH(T ) (R1)

The IR absorption cross-sections obtained allow the calculation of the radiative efficiency, while kOH at 298 K allows the estimation of τ for this route and the photochemical ozone creation potential (εPOCP) due to an increment in the emission of HFC-1447fz by the method developed by Derwent et al.3 and Jenkin.4 Regarding the UV photolysis of HFC1447fz, no measurement of the UV absorption cross-sections, σλ, have been performed up to date to our knowledge. For C2− C3 HFOs, Orkin et al.5 reported σλ in the vacuum UV between 160 and 220 nm, giving an upper limit of 10−22 cm2 molecule−1 in the solar actinic region (λ > 290 nm). To evaluate whether photolysis is important for HFC-1447fz, we determined σλ to be between 190 and 368 nm for the first time here. Moreover, to quantify the indirect environmental impact of the atmospheric degradation of HFC-1447fz, the oxidation reaction products need to be known as well. In this work, the identification of secondary products formed after the initiation reaction R1 has been performed in an atmospheric simulation chamber at room temperature and atmospheric pressure. The gaseous reaction products were identified by Fourier transform infrared (FTIR) spectroscopy, while the number-size distribution of the secondary organic aerosols (SOAs) formed from reaction R1 has been carried out by using a scanning mobility particle sizer (SMPS). The obtained kinetic, spectroscopic and product information allow us to assess the potential atmospheric impact of emissions of HFC-1447fz.

■

EXPERIMENTAL SECTION A total of four different, but complementary, experimental setups were employed in this work. The first two experimental systems, based on UV and mid-IR spectroscopy, were used in the determination of the absorption cross-sections in those spectral regions.6−10 The second setup is based on the pulsed laser photolysis/laser-induced fluorescence (PLP-LIF) technique previously used in kinetic studies on OH radicals.6−13 Finally, the atmospheric simulation chamber used at ICARECNRS (acronym of Institut de Combustion Aérothermique Réactivité et Environnement Centre National de la Recherche Scientifique) for conducting the product formation studies, both gaseous and SOAs, is briefly described because it was previously described in detail by Bernard et al.14 More details of the experimental procedures and methodology is provided in the Supporting Information. Ultraviolet and Fourier Transform Infrared Spectroscopy. Basically, these setups consist of a continuous radiation source, a gas cell with a fixed optical path length, and a specific detection system for monitoring the transmitted radiation. At UCLM (acronym of University of Castilla−La Mancha), the irradiation source is a D2 lamp for UV measurements and a Globar lamp for IR measurements using a FTIR spectrometer (Bruker, model Tensor 27).6−10 The UV and IR gas cells are made of Pyrex and stainless steel, respectively, and are sealed with quartz and ZnSe windows, respectively. The optical path lengths were S = 107 cm (UV) and 10 cm (IR). These cells 1235

DOI: 10.1021/acs.est.5b04379 Environ. Sci. Technol. 2016, 50, 1234−1242

Article

Environmental Science & Technology

Figure 1. UV absorption cross-sections of HFC-1447fz. The spectral solar actinic flux, calculated at ground level for a zenith angle of 16° (summer midday at Ciudad Real), is also depicted as a dashed line.

ICARE Smog Chamber for Product Studies. Through the chamber a controlled air flow is used to introduce the gasphase reactants, through a calibrated cylinder (0.9 L), and to maintain a stable overpressure. Homogenization of the gas mixture inside the chamber was achieved by the use of two fans. A set of 14 and 28 UV−vis lamps are coupled in the inner wall of the chamber close to the Teflon bag. The emission of these lamps are centered at 254 and 365 nm, respectively. The temperature and relative humidity (RH = 9−15%) were monitored by a combined sensor. Concentrations of gaseous compounds (reactants and products) were recorded every 4−5 min using the same FTIR spectrometer used in the determination of the IR absorption cross-sections. Levels of O3 and NOx (NO + NO2) were measured before starting the irradiation using ozone (Horiba, APOA 360) and NOx (Horiba, APNA 360) monitors, respectively. Ozone and NOx concentrations ranges were 2.1−121 ppb and 13.7−1376.3 ppb, respectively. In situ FTIR spectroscopy was used to determine the fraction of HFC-1447fz reacted (Δ[HFC-1447fz] = [HFC1447fz]0−[HFC-1447fz]t]) and the formation of gaseous reaction products. Experiments to detect the SOA formed in reaction R1 were performed under low- and high-NOx conditions. The time evolution of SOA formed was monitored by a SMPS (TSI, model 3080), consisting of a differential mobility analyzer, DMA (TSI, model 3081), and a condensation particle counter, CPC (TSI, model 3022A). SMPS determined the particle size distribution in the size range between 10 and 500 nm. In the presence of added NOx, the chemical system CH3ONO−NO−air was used as the photo-

chemical OH precursor. Initial concentrations of HFC-1447fz, CH3ONO, and NO were 539−593 ppb, 11−13 and 1.30 ppm, respectively. Under low NOx concentrations, UV photolysis of H2O2 was used as OH precursor. Initial H2O2 and HFC-1447fz concentrations ranged from 18 to 33 ppm and from 491 ppb to 1.12 ppm, respectively.

■

RESULTS Absolute UV and IR Absorption Cross-Sections of HFC-1447fz. Following the methodology described in the Supporting Information, a series of independent experiments was carried out at several [HFC-1447fz] values to derive the absolute UV and IR absorption cross-sections from the plots of equation (EII) of the Supporting Information. Figure 1 shows the UV σλ (depicted in log scale) of the investigated HFC as a function of wavelength (data above 225 nm are below 10−23 cm2molecule−1 and are omitted). Uncertainties are not included in the figure for clarity either. The inset figure shows the UV absorption spectrum between 190 and 220 nm corresponding to the tail of the π → π* transition of the double bond. σλ of HFC-1447fz strongly decreases from (3.52 ± 0.19) × 10−20 cm2 molecule−1 at 191.3 nm to ca. 10−23 cm2 molecule−1 in the UV actinic region, i.e. HFC-1447fz exhibits a very low absorption at the photolysis and probe wavelengths (σ248 nm ∼ 5 × 10−23 cm2 molecule−1 and σ282 nm < 10−23 cm2 molecule−1). Therefore, no loss of HFC-1447fz by UV photolysis is expected during the kinetic experiments at the laser fluences employed in this work. Absorption cross-sections 1236


Article


kOH(T) were fitted to the Arrhenius equation (kOH(T) = A exp{−Ea/RT}) yielding the following expression (±2σ):

between 191.3 and 225.3 nm are listed in Table S1 at 1 nm intervals. IR spectra are necessary to be recorded for determining the HFC-1447fz concentration in the kinetic and product studies. Figure S1 shows the mid-IR absorption spectrum of HFC1447fz obtained independently at UCLM and ICARE-CNRS using setups with different optical path lengths. As can be seen in Figure S1a, no significant IR absorption is observed between 4000 and 1500 cm−1. The absorption bands corresponding to C−F vibrational modes lie in the 1500−700 cm−1 spectral region. For that reason, in Figure S1b, the IR absorption crosssections are shown in a reduced wavenumber range. The IR absorption spectra obtained in both laboratories are in excellent agreement (when σν̃ from ICARE-CNRS is plotted against σν̃ from UCLM between 1300 and 700 cm−1 the slope of the linear regression is 1.01 with r2 = 0.998), which leads confidence to the IR absorption cross-sections reported. In addition, the instrumental resolution is not influencing the obtained values of the IR absolute absorption cross-sections (Figure S1c). Therefore, in Table S2, the averaged absolute absorption cross-sections obtained between 500 and 1500 cm−1 are listed. HFC-1447fz strongly absorbs in the IR atmospheric window (720−1250 cm−1); therefore, this gas could contribute to warming of Earth. Rate Coefficients of the OH + HFC-1447fz Reaction as a Function of Temperature. The analysis of the OH LIF temporal profiles, described in the Supporting Information, yields the pseudo-first-order rate coefficients in the absence (k0) and presence of varying HFC concentrations (k′). Table S3 lists the ranges of the obtained k0 and k′ as a function of temperature and total pressure in the reaction cell (pT,cell). Individual rate coefficients kOH(T), listed in Table S3, were obtained from the plots of eq IV of the Supporting Information. Because k0 can slightly change from one experiment to another at a single temperature, to compare different kinetic experiments, we transformed eq IV of the Supporting Information into the following equation, where k′ is corrected with k0: k′ − k 0 = k OH(T ) × [HFC − 1447fz]

k OH(T ) = (7.4 ± 0.4) × 10−13 exp(161 ± 16)/T ) (cm 3· molecule−1· s−1)

The obtained rate coefficient for the OH+HFC-1447fz reaction exhibits a slightly negative temperature dependence, which is in agreement with the observations for other HFOs. For comparison purposes, in Figure S3, the reported temperature dependencies for CF3CHCH2,15 CF3CF2CHCH2,16 CF3(CF2)3CHCH2,17 and CF3(CF2)5CHCH2.17 Arrhenius parameters A and Ea/R values are also listed in Table S4 for the OH reactions with CF3CHCH2,5,15 CF3CF2CH CH2,11,16 CF3(CF2)3CHCH2,17 and CF3(CF2)5CHCH2.17 For reaction R1, the Arrhenius parameters obtained in this work are in good agreement with those reported in the bibliography for CF3(CF2)x=0−3CHCH2, where Ea/R values are slightly negative or even zero, within the uncertainties. Gas-Phase Products. Under high-[NOx] conditions, i. e., in the HFC-1447fz−CH3ONO−NO−air system, an example of initial and final FTIR spectra is depicted in Figure 2 (panels A and B). CF3(CF2)2CHO was observed, which is in excellent agreement with previous studies with other HFOs,17−19 together with formaldehyde (either from the decomposition of α-hydroxy alkoxy radical or from the CH3O+O2 reaction in the OH formation scheme from CH3ONO) and HNO3 (presumably formed through the reaction between OH and NO2 and through heterogeneous processes involving the chamber wall). For the identification of CF3(CF2)2CHO, an IR spectrum from Chiappero et al. was used as a reference.20 After the spectral subtraction of all these identified products, some bands are still present in the residual spectrum (Figure 2, panel C). IR features of expected CF3(CF2)2C(O)O2NO2 were not identified.21 In addition, Figure S4 shows an example of the initial, final, and residual spectra recorded for a typical experiment performed using H2O2 as OH precursor (low [NOx]). Evidence of Particle Formation in the OH + HFC1447fz reaction: Number and Size Distributions. For each NOx level, three different experiments were performed: (A) dark experiments to evaluate the wall losses (kwall,HFC) of HFC1447fz; (B) HFC-1447fz + UV−vis radiation to evaluate the photolysis process of the HFC; and (C) HFC-1447fz + OH precursor + UV−vis radiation to characterize the formation of SOA from the OH-reaction. In A experiments, kwall,HFC (±2σ) for HFC-1447fz due to wall losses, dilution effects, etc., was measured to be (1.33 ± 0.14) × 10−5 s−1, while in B experiments, the averaged loss rate coefficient was (1.63 ± 0.17) × 10−5 s−1. Under the irradiation conditions employed in this work, the loss rate of HFC-1447fz by UV photolysis is similar to wall losses. In C experiments, particle formation was observed in the photolysis as it is illustrated in Figure 3a,b. This figure shows the general behavior of aerosol number concentrations over the course of the reaction. Figure 3a shows an example of the log-normal distributions of the number density of newly formed particles, nn(log Dp) = dN/d(log Dp), in the H2O2(+hν)/HFC-1447fz system. Under our experimental conditions, Dp for fine particles varied between 10 nm and 0.1 μm (Aitken mode) depending on the time after photolysis commenced. Typically, the SOA formation was observed 5−10 min after the UV photolysis commenced. As an example, Figure 3a shows the observed shift

(E1)

The plots of k′ − k0 versus [HFC-1447fz] were linear over the entire concentration range and no significant change in kOH(T) with total pressure (50−650 Torr) was observed, out of the experimental uncertainty limits (see Figure S2). Therefore, an average rate coefficient of individual kOH(T) in the measured pressure range is reported in Table 1 and depicted in Figure S3. Table 1. Averaged Absolute Rate Coefficientsa kOH(T) as a Function of Temperature between 50 and 650 Torr of He

a

T (K)

[HFC-1447fz] (1014molecules cm−3)

range of k′ (s−1)

263 270 278 287 298 308 323 338 358

1.5−11.6 1.4−10.8 2.8−9.9 1.5−11.4 2.2−15.7 2.1−12.0 2.1−10.2 2.2−10.9 2.1−10.4

196−1510 199−1366 263−1408 180−1382 270−1384 247−1407 241−1264 227−1302 218−1217

kOH(T) × 1012 (cm3 molecule−1 s−1) 1.36 1.34 1.32 1.29 1.24 1.24 1.21 1.19 1.16

± ± ± ± ± ± ± ± ±

(E2)

0.12 0.14 0.14 0.12 0.14 0.12 0.12 0.10 0.10

Uncertainties are ±2σ of the precision of the fit. 1237


Article


Figure 2. FTIR spectra recorded in the HFC-1447fz (1126 ppb)−CH3ONO (593 ppb)−/NO (1000 mL) system. (A) Initial spectrum before irradiation, (B) final IR spectrum after irradiation, and (C) after spectral subtraction of unreacted HFC-1447fz, CF3(CF2)2CHO, HC(O)OH, and HNO3.

1447fz]) both expressed in μg/m3. M0 was corrected by taking into account the wall loss process of particles measured with the lights off and/or OH precursors completely consumed, kwall,particle= (5 ± 1) × 10−5 s−1. The yield (Y) of SOA formation in the investigated reaction was obtained for each experiment using eq E3:

in the distribution toward larger diameters at longer times, suggesting a nucleation process. In this experiment, the maximum particle number concentration was observed at 13.5 min after the photolysis. In Figure 3b, the time evolution of N and [HFC-1447fz] is presented. Colored circles correspond to the distribution curves from four reaction times between 9 and 51.7 min. After 5−10 min, nucleation stops, and the number of particles decreases as their size increases because of the combined effects of condensation, coagulation, and wall loss. Time Evolution of SOA Mass Concentration, M0, and Determination of SOA Yields. Aerosol mass concentration, M0, was obtained by the SPMS system from the mobility diameters determined by this instrument, assuming a spherical shape and an aerosol density of 1.2 g cm−3. Figure S5 illustrates the temporal profiles of particle mass formation from the OHreaction with the investigated HFC using H2O2 and CH3ONO as sources of OH radicals. A short induction time was observed in less than 20 min that can be interpreted as the initiation stage of the homogeneous nucleation process. As can be seen in Figure S5, the particle mass obtained under low NO x conditions is slightly higher than at high NOx. The formation of SOAs was quantified by measuring the ratio of the mass concentration (M0) and the reacted fluorolefin (Δ[HFC-

Y=

M0 Δ[HFC‐1447fz]

(E3)

The SOA yield can then be determined from the slopes of M0 versus Δ[HFC-1447fz] shown as solid lines in Figure 3c. In Table 2, a summary of the results obtained is presented together with the reagent concentrations used. As can be observed in Table 2, in the experiments in which H2O2 is used as a source of OH, ignoring the differences in NOx and O3 levels, a clear decrease of the SOA yield can be observed as the concentration of HFC-1447fz increases. This may indicate that at the levels of NOx and O3 of these experiments, the variation of HFC concentration is causing the major effect on the SOA yield. For a similar initial concentration of HFC-1447fz (ca. 500−600 ppb), the SOA formation yields obtained from lowNOx experiments have been found to be slightly higher (around 2%) than in the experiments conducted in the presence of high NOx concentrations (0.12−0.94%). Under 1238


Article


Table 2. Initial Conditions and SOA Yields (Y) Obtained in the Gas-Phase Reaction of OH Radicals with CF3(CF2)2CHCH2 at Room Temperature and Atmospheric Pressure of Air source of OH H2O2(hν = 254 nm) H2O2(hν = 254 nm) H2O2(hν = 254 nm) CH3ONO (hν = 365 nm)/NO CH3ONO (hν = 365 nm)/NO

[HFC1447fz]0/ppb

[NOx]0/ ppb

[O3]0/ ppb

491 514 1122 539

16.4 268.1 56.1 1348.0

87.6 114.4 120.9 2.1

593

1375.7

16.7

Y% 1.79 1.61 0.37 0.12

± ± ± ±

0.04 0.05 0.01 0.01

0.94 ± 0.08

with O2, preferably react with RO2 or HO2, forming low- and semivolatile species and contributing to SOA growth. At high NOx concentrations, there is a competitive chemistry of RO2 radicals between NO and HO2 radicals with the RO2 + NO reaction producing products of higher volatility than the RO2+ HO2 reaction and consequently leading to a reduction in the SOA yield. The same trend on SOA yields was observed in the ozonolysis of α-pinene as a function of NOx.22,23 However, the effect of NOx on the SOA yield from the reaction of O3 with large sesquiterpenes is just the opposite, as discussed by Ng et al.23 In our work, the effect of ozone on the SOA yield can be evaluated by comparing the results obtained for different [NOx]0/[O3]0 ratios at a similar HFC initial concentration. As can be derived from Table 2, the SOA yield exponentially decreases when increasing [NOx]0/[O3]0. Therefore, here we report that the SOA yield from reaction R1 between 16 ppb and 1.37 ppm of NOx is low, ranging from 0.12 to 1.79%. Note that wall losses of semivolatile products were not considered and could lead to an underestimation of SOA yield.

■

DISCUSSION The present study constitutes the first report of the gas-phase kinetics of the OH + HFC-1447fz reaction, the absorption cross-sections in the IR and UV regions, and end-product analysis and SOA formation from the oxidation reaction R1. Lifetime Estimations and Calculation of GWP for HFC1447fz. In the estimation of the tropospheric lifetime of HFC1447fz (τ) all possible removal processes must be considered, e.g., homogeneous reactions with atmospheric oxidants (O3, OH, NO3, Cl, etc.) and photolysis in the UV actinic region (λ > 290 nm). The lifetime due to reaction with OH radicals, τOH (=1/kOH[OH]), is 9.3 days, considering the room temperature kOH and [OH]24‑h = 1 × 106 cm−3.24 In this estimation, it is assumed that the spatial distribution of the emitted HFC1447fz is homogeneous and it is well-mixed in the atmosphere, which is not really true for short-lived species. No kinetic information is available for the Cl- and NO3-reactions with HFC-1447fz. However, Sulbaek Andersen et al.16 reported that the rate coefficients for the reactions of O3 and Cl with CxF2x+1CHCH2 (x = 1, 2, 4, 6, and 8) at room temperature were kO3= 2 × 10−19 (cm3·molecule−1·s−1) and kCl= 9.07 × 10−11 (cm3·molecule−1·s−1), respectively. Assuming the upper limit reported by Singh et al.25 for the global annually averaged Cl atom concentration is 103 cm−3 in the troposphere, the lifetime due to the Cl reaction (τCl = 1/kCl[Cl]) is around 4 months. Assuming a 24 h average ozone concentration of 7 × 1011 cm−3,26 the tropospheric lifetime due to the O3 reaction (τO3 = 1/kO3[O3]) is around 3 months. Because photolysis of

Figure 3. (a) SPMS particle-size distributions in the chamber as a function of time after inception of photolysis. Experimental conditions: [CF3(CF2)2CHCH2]0 = 514 ppb, [H2O2]0 = 292 ppb, and RH = 13.9%. (b) Evolution of the particle concentration at different reaction times. The colored circles correspond to the examples of the size distribution illustrated in (a). (c) SOA yield plot when H2O2 is used as an OH precursor.

low-NOx conditions, organic peroxy radicals (RO2), formed in the reaction of the addition complex (produced in reaction R1) 1239


Article


radicals although much slower than HFC-1447fz does (∼6 × 10−13 (cm3·molecule−1·s−1) at room temperature) as reported by Sulbaek Andersen et al.21 and Solignac et al.28 The main conclusion drawn by these authors is that the major atmospheric removal pathway for CF3(CF2)2CHO will be UV photolysis, which, under low-NOx conditions, may be a source of fluorinated carboxylic acids in the troposphere, while in the presence of NOx, C(O)F2 is the main degradation product and small quantities (molar yields

Atmospheric Degradation Initiated by OH Radicals of the Potential

Recommend Documents