Rate Acceleration of the Heterogeneous Reaction of Ozone with a

Feb 21, 2013 - Rate Acceleration of the Heterogeneous Reaction of Ozone with a Model Alkene at the Air–Ice Interface at Low Temperatures. Debajyoti ...
1 downloads 6 Views 620KB Size
Article pubs.acs.org/est

Rate Acceleration of the Heterogeneous Reaction of Ozone with a Model Alkene at the Air−Ice Interface at Low Temperatures Debajyoti Ray,† Joseph K’Ekuboni Malongwe,† and Petr Klán*,†,‡ †

Research Centre for Toxic Compounds in the Environment, Faculty of Science, Masaryk University, Kamenice 5, 62500 Brno, Czech Republic ‡ Department of Chemistry, Faculty of Science, Masaryk University, Kamenice 5, 62500 Brno, Czech Republic S Supporting Information *

ABSTRACT: The kinetics of the ozonation reaction of 1,1diphenylethylene (DPE) on the surface of ice grains (also called “artificial snow”), produced by shock-freezing of DPE aqueous solutions or DPE vapor-deposition on pure ice grains, was studied in the temperature range of 268 to 188 K. A remarkable and unexpected increase in the apparent ozonation rates with decreasing temperature was evaluated using the Langmuir− Hinshelwood and Eley−Rideal kinetic models, and by estimating the apparent specific surface area of the ice grains. We suggest that an increase of the number of surface reactive sites, and possibly higher ozone uptake coefficients are responsible for the apparent rate acceleration of DPE ozonation at the air−ice interface at lower temperatures. The increasing number of reactive sites is probably related to the fact that organic molecules are displaced more to the top of a disordered interface (or quasi-liquid) layer on the ice surface, which makes them more accessible to the gas-phase reactants. The effect of NaCl as a cocontaminant on ozonation rates was also investigated. The environmental implications of this phenomenon for natural ice/snow are discussed. DPE was selected as an example of environmentally relevant species which can react with ozone. For typical atmospheric ozone concentrations in polar areas (20 ppbv), we estimated that its half-life on the ice surface would decrease from ∼5 days at 258 K to ∼13 h at 188 K at submonolayer DPE loadings.



INTRODUCTION Various (semi)volatile natural and anthropogenic organic compounds, including unsaturated hydrocarbons, have been found in natural snow.1,2 In general, organic molecules cannot be incorporated within the ice lattice when their solutions are frozen. Instead they are trapped within highly concentrated layers on the surface and may also form micropockets or microveins in which the contaminants are less available to reactive gases.3,4 Experimental, as well as computational studies, describe a nanoscale region of ice surface disorder, usually referred to as a quasi-liquid layer (QLL) and called here a disordered interface5 (DI), the thickness of which decreases with decreasing temperature but may increase with the presence of impurities.5−7 The ice surface nature allows for adsorption of (semi)volatile organic compounds.8,9 Polar adsorbates can be buried into the DI layer,10 whereas nonpolar aromatic species can acquire a flat orientation at the air−ice interface because of strong electrostatic interactions between the delocalized π-system and surface water hydrogens.11−15 This interaction is much stronger than that of intramolecular (mutual) π−π stacking.13 Many factors, such as light absorption properties, phase properties of the host matrix, temperature, and the presence of other (reactive) species in/on ice or in the surrounding atmosphere, influence the chemical fate of the snow contaminants.1,2,5,16 © XXXX American Chemical Society

Surface ozone plays an important role in the chemistry occurring at the air−ice interface.17−19 Alkenes, relatively abundant trace contaminants in the environment,1 readily undergo oxidative cleavage by ozone via a well-established mechanism. The 1,3-dipolar cycloaddition of ozone on a CC bond yields primary ozonides, which subsequently fragment to the Criegee intermediate and the corresponding aldehyde or ketone.20 In our previous work, we studied the kinetics of the ozonation reaction of 1,1-diphenylethylene (DPE) on ice surfaces at 258 K.19 The kinetic data were used to evaluate the specific surface area of ice grains, produced by a shock-freezing technique, as a measure of the availability of molecules on the surface for chemical reaction with gaseous species. The experimental results were consistent with the Langmuir− Hinshelwood (LH) type reaction mechanism,21 analogous to those obtained for the reactions of, for example, PAHs with O3 at the ice-air18,19 or some other22,23 interfaces. Shiraiwa and coSpecial Issue: Rene Schwarzenbach Tribute Received: November 25, 2012 Revised: February 4, 2013 Accepted: February 6, 2013

A

dx.doi.org/10.1021/es304812t | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Preparation of Ice Samples. Ice (artificial snow) samples were produced4,19 by (i) spraying pressurized aqueous solutions of DPE (c0DPE = 3 × 10−7 − 3 × 10−5 mol L−1), in some cases containing NaCl (cNaCl = 35 or 175 g L−1), into liquid nitrogen, or (ii) adsorbing DPE vapor at the given temperature (268 to 188 K; this temperature was always the same as that of subsequent ozonation) on ice grains prepared by spraying pure water or NaCl solutions (cNaCl = 35 g L−1) into liquid nitrogen. The size of the individual snow grains formed using this technique (50−300 μm) was determined in our previous work.4 Each sample was homogenized and divided into four parts; one part was analyzed directly; the remaining three parts were exposed to ozone and analyzed afterward. Ozonation. The experiments were performed according to the procedures described earlier.19 A snow sample (50 g) was placed on a porous porcelain plate, which was located in a glass desiccator kept inside a stainless-steel freezer at a given temperature (268−188 K) for 30 min. It was then exposed to ozone for a given time. Ozone was produced by UV irradiation of oxygen in a homemade generator. The flow rate of the incoming oxygen was adjusted by a flow meter. The resulting flow of an O3/O2 mixture was led through the desiccator bottom.19 The unreacted ozone was directed into a KI solution via a glass tube outlet located at the top of the desiccator. The flow-tube ozone concentration (∼1.6 × 1016 molecules cm−3; flow 0.25 L min−1) was determined iodometrically using UV− vis spectrometry. In the control experiments, selected samples were exposed to a stream of pure nitrogen or oxygen in order to correct the extent of ozonolysis for losses caused by evaporation, which was negligible at lower temperatures, or by oxidation by O2. Sample Analysis. The ice samples (50 g) were melted at ∼23 °C and extracted with dichloromethane (3 × 15 mL). The extracts were concentrated under a stream of nitrogen to 10 mL, dried with anhydrous sodium sulfate (∼5 g), filtered through glass wool, concentrated again to 1 mL under gentle stream of nitrogen gas, and analyzed using GC-MS. All data are based on at least 3 independent experiments; the relative standard deviation of the mean is given.

workers have demonstrated that the LH kinetics can also be explained by a multistep mechanism which involves the formation of long-lived reactive oxygen intermediates.24 The LH approach21 involves (i) equilibrium air−ice surface partitioning of ozone and (ii) a heterogeneous reaction between an organic molecule and adsorbed ozone; the I pseudofirst-order rate constant of the reaction (kobs ) is expressed according to eq 125−27 I kobs = k II[SS]

K O3[O3]g 1 + K O3[O3]g

(1)

where kII is the second-order rate constant, [SS] is the number of active surface sites, KO3 is the adsorption equilibrium constant of ozone, and [O3]g is the gas-phase ozone concentration. At saturated surface concentration of ozone I k II[SS] = k max

(2)

where kImax is the maximum pseudo-first-order rate constant. In contrast, when ozone reacts directly from the gas phase with a molecule adsorbed on the surface, the reaction kinetics is usually referred to as an Eley−Rideal (ER) mechanism.21 The rate is expressed by a standard second-order kinetic equation d[S]ad = −k O3[O3]g [S]ad dt

(3)

where [S]ad is the substrate (DPE) surface load and kO3 is the second-order rate constant of the reaction. Fitting the experimental pseudo-first-order rate constants kobs(= kO3[O3]g)as a function of [O3]g should provide a linear plot with a slope of kO3. This work studies the ozonation kinetics of DPE on the surface of ice grains (or artificial snow), produced by shockfreezing of DPE solutions and DPE vapor-deposition on pure ice.4,19 DPE has been selected as an environmentally relevant snow contaminant which undergoes an ozonation reaction. The apparent rate constants were measured as a function of temperature in the range of 268 to 188 K. The influence of NaCl as a cocontaminant on the ozonation kinetics was also investigated. Mechanistic considerations using both LH and ER kinetic models, evaluation of the apparent specific surface area, and environmental implications are provided.



RESULTS AND DISCUSSION Ice (artificial snow) samples containing DPE were prepared by the two methods described previously:4,19 shock freezing of DPE aqueous solutions sprayed into liquid nitrogen, or DPE vapor deposition on pure ice. We have already demonstrated that shock-freezing of the DPE aqueous solutions produced ice grains in which over 80% of DPE molecules are in reactive contact with ozone at submonolayer coverages (∼2 × 10−5 mol kg−1) at 258 K.19 Alternatively, a vapor deposition technique in principle localized DPE molecules on the ice surface (DPE concentrations were adjusted by the exposition time). The ice samples were subsequently placed into a reaction vessel (desiccator) and exposed to an O3/O2 mixture or pure nitrogen (the gas flow was always the same). Thermal desorption of DPE was found negligible during the period of the experiment.19 The amounts of DPE were determined by GC-MS after the compound was extracted from the melted samples by dichloromethane. Accessibility of Ozone to DPE. The availability of DPE for the ozonation reaction to form benzophenone (Figure 1, inset) at the air−ice interface was examined by prolonged exposure of the samples to ozone. Figure 1 shows the plot of DPE disappearance in ice samples produced by both (solution



EXPERIMENTAL SECTION Materials. Reagents and solvents (DPE, sodium chloride, potassium iodide, potassium iodate, anhydrous sodium sulfate, dichloromethane) of the highest purity available were used as purchased. Water was purified on an Osmonics 2 and a Millipore Simplicity 185. Instrumentation. A GC-MS system consisting of a HP 6890 gas chromatograph, equipped with a J&W Scientific DB5MS fused silica column (60 m × 0.25 mm × 0.25 μm stationary phase film; 5% phenyl/95% methylpolysiloxane) and a HP 5973 mass selective detector (Agilent), was used; the carrier gas was helium (flow rate of 1.5 mL min−1). Injector and transfer line temperature was 280 °C. Samples (1 μL) were injected in a split mode (1:20) for a full-scan analysis or a splitless mode (SIM). Initial temperature was held at 80 °C for 1 min, subsequently raised to 180 °C by 10 min−1, and then to 300 °C by 15 min−1 (held for 5 min). B

dx.doi.org/10.1021/es304812t | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 1. Decrease in DPE concentrations during ozonation ([O3]g = 1016 molecules cm−3) in ice samples prepared from shock-frozen DPE solutions (sol; ozonation temperature, red full (268 K) or blue empty (188 K) squares; c0DPE ≈ 8.7 × 10−6 mol kg−1) or by DPE vapor deposition (ad, ozonation temperature, red full (268 K) and blue empty (188 K) circles; c0DPE ≈ 1 × 10−5 mol kg−1) on pure ice grains (sample layer thickness = 1.5 cm). Inset: DPE produces benzophenone (BP) on ice in the presence of ozone.19 The error bars represent the standard deviations.

Figure 2. The effect of temperature on the observed initial pseudofirstorder rate constants (kobs). The samples were prepared from shockfrozen DPE solutions (ksol obs, red full squares) or DPE solutions (1) , c1NaCl = 35 g L−1, red empty squares; containing NaCl (kad,NaCl obs ad,NaCl (2) 2 −1 , cNaCl = 175 g L , green empty squares), or by DPE vapor kobs deposition on pure ice (kad obs, full blue circles) or ice made from shock, cNaCl= 35 g L−1, empty blue circles). frozen NaCl solutions (kad,NaCl obs The concentration of ozone ([O3]g) was 1016 molecules cm−3. c0DPE corresponded to submonolayer coverages (2−0.07 × 10−5 mol kg−1; see text) in all cases. The error bars represent the standard deviations.

shock-freezing and vapor-deposition) techniques at 268 and 188 K. Both the initial DPE concentrations, c0DPE = 8.7 × 10−6 mol kg−1 (sol; shock freezing) and ∼1 × 10−5 mol kg−1 (ad; vapor deposition), corresponded to submonolayer loadings.19 A constant ozone concentration during the experiments ([O3]g = 1016 molecules cm−3) showed that the reaction can be described by a pseudofirst-order kinetics. 19 The DPE concentration leveled off to approximately 80 and 90% at 268 K in 2 h for both shock-freezing- and vapor-depositionproduced samples, respectively (red symbols). This is analogous to our previous results from experiments carried out at 258 K; we suggested that 10−20% of DPE molecules are overlaid by another molecular layer of DPE or buried in the ice matrix because the kinetic analysis identified a slower kinetic component at high reaction conversions.19 In contrast, faster and almost complete DPE disappearance was observed at 188 K (Figure 1; blue symbols), which indicates that essentially all alkene molecules are in reactive contact with O3 regardless of the sample preparation technique. Temperature Dependence of the Rate Constants. Figure 2 shows dependence of the observed rate constants ad (ksol obs: a solution shock-freezing method; kobs: a vapor-deposition (adsorption) method) of DPE ozonation on temperature and ice sample properties. The constants were determined by plotting ln(cDPE/c0DPE) vs reaction time using the least-squares method (Supporting Information (SI) Figure S2). The plots were linear at DPE conversions lower than ∼70%; therefore, the initial reaction followed pseudofirst-order kinetics. Nevertheless, the reaction conversions were kept low (below 40%) to keep relatively constant DPE surface loads during the experiment and avoid interference from the products. The sample layer thickness in the desiccator was kept as low as possible (1.5 cm) because the ozonation efficiency of DPE on the subjacent grains can be restricted by ozone diffusion through the top ice sample layers as determined previously.19 In general, the values of the observed rate constants obtained in the range of 268 to 188 K increased at lower temperatures (Figure 2, SI Table S1). The sample preparation method and

the presence of NaCl as a cocontaminant (2 different NaCl concentrations: c1NaCl = 35 g L−1, c2NaCl = 175 g L−1) had also a noteworthy effect on the rate constant magnitudes. At 268 K, −4 −1 ad,NaCl kad s ) was higher than those of ksol , and obs (2.28 × 10 obs, kobs sol,NaCl kobs by several factors; this difference was observed for the whole temperature range and was even more pronounced between 258 and 218 K. An increase by an order of magnitude ad in the kobs value was found between 268 and 238 K; temperature had no effect on kad obs below 238 K. The rate ad,NaCl constant ksol , and ksol,NaCl values increased at temperobs, kobs obs atures below ∼228 K. The NaCl concentration had a significant influence on kobs for the samples produced by a vaporad,NaCl deposition method (kad ). The rate constants were obs vs kobs essentially same in the case of the samples prepared by shockad,NaCl freezing method (ksol ). The NaCl loading had a obs, kobs (2) relatively small effect: the kad,NaCl values were noticeably obs ad,NaCl (1) lower than those of kobs at the lowest temperatures. The rate constants were essentially not influenced by initial DPE concentrations (c0DPE) in the range of 2 × 10−5 to 7 × 10−7 mol kg−1, i.e., when the surface coverage was kept below one monolayer. Kinetic Models. Figure 3 shows dependences of kobs on [O3]gfor three representative c0DPE submonolayer19 concentrations and two different temperatures (258 and 188 K). The kobs values leveled off at high ozone concentrations, which is consistent with the Langmuir−Hinshelwood type reaction mechanism21 describing the bimolecular reaction of two species adsorbed on a surface. While the LH fit to the data measured at 258 K19 provided the coefficient of determination R2 = 0.99, the regression obtained for the measurements at 188 K was less stringent, especially at higher [O3]g (R2 = 0.92). A linear fit corresponding to the Eley−Rideal mechanism was far from perfect (R2 = 0.85). Nevertheless, we calculated KO3 and kImax values based on the assumption that the data obtained at 188 K can be explained by the Langmuir−Hinshelwood model: for C

dx.doi.org/10.1021/es304812t | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 3. Pseudo-first-order rate constants (kobs) as a function of gasphase ozone concentration for the reaction of DPE and ozone on ice at 258 and 188 K. Fitting of both the Langmuir−Hinshelwood (solid lines) and Eley−Rideal (dashed line) equations to the data gave the mean values of KO3 and kmax; the coefficients of determination are given. The data obtained at 258 K are taken from our previous work.19 The error bars represent the standard deviations.

Figure 4. Apparent specific surface area (ASSA) of ice grain samples prepared by DPE vapor deposition on pure ice (ad, full blue circles) or ice made from shock-frozen NaCl solutions (ad, c1NaCl = 35 g L−1, empty blue circles), or from shock-frozen DPE solutions in the absence (sol, full red squares) and the presence of NaCl as a cocontaminant (sol, c1NaCl = 35 g L−1, empty red squares). The dashed line is shown to guide the eye. The concentration of ozone ([O3]g) was 1016 molecules cm−3.

the submonolayer concentration c0DPE = 5.2 × 10−7 mol L−1, KO3 = 4.1 × 10−8 cm−3, and kImax = 0.29 s−1. These are much higher values when compared to those obtained previously at 258 K (c0DPE = 3.6 × 10−7 mol L−1, KO3 = 1.5 × 10−15 cm−3 and kImax = 1.1 × 10−3 s−1).19 Temperature Dependence of the Apparent Specific Surface Area. Surface coverage, that is, the fraction of the adsorption sites occupied by molecules of ice or snow can be determined from the Langmuir adsorption isotherm11 or calculated from the solid surface area and the area occupied by an adsorbed molecule,4,19 which subsequently serves for estimating the area accessible to gases per unit mass, that is, the specific surface area (SSA). We recently evaluated the SSA of ice grains (artificial snow) using the DPE ozonation reaction (∼70 cm2 g−1 at 258 K) as a measure of the availability of the surface molecules for a given chemical reaction with gaseous species.19 DPE molecules on snow grains reacted with O3 at approximately the same rate constant on the surface covered by a monolayer (at the Langmuir concentration, cLDPE ≈ 2 × 10−5 mol kg−1) or less. This monolayer surface concentration was identified by a sudden decrease of kobs with increasing c0DPE19,22 (the subjacent layers of DPE are much less exposed to ozone; for example, SI Figures S3 and S4). Because the SSA is dependent on experimental conditions, we use the term apparent SSA (ASSA). Analogous to our previous study,19 the ASSA values in this work were determined for ice samples produced by both vapordeposition and solution shock-freezing techniques in the temperature range of 268 to 198 K (Figure 4). Here it should be emphasized that the ice grain samples were always prepared by spraying pressurized water or aqueous solutions into liquid nitrogen, i.e., under the same conditions (temperature). A higher temperature was subsequently set for DPE vapor deposition and ozonation. The ASSA values were found higher at lower temperatures, following a similar tendency as that observed with the rate constants (Figure 2). At 268 K, ASSA (ad = vapor-deposition) was 320 cm2 g−1, whereas the remaining ASSA values (the samples containing NaCl and, for sol = solution shock-freezing, without or with NaCl) were

lower by a factor of 3 (∼100 cm2 g−1). An increase in ASSA values at lower temperatures was observed in all cases (by a factor 3 when compared to the values obtained at 198 K). The solution shock-freezing samples and those containing NaCl had nearly identical ASSA values over the whole temperature range. The experimental results described above provide evidence that the apparent rates of DPE ozonation increase with decreasing temperature. This counterintuitive finding, which cannot be attributed to a simple concentration effect triggered by freezing28 because the ozonation experiments were carried out at a constant temperature and the samples were prepared under the same experimental conditions, are discussed in the following paragraphs. Accessibility of Gas-Phase Ozone to DPE. O 3 accessibility to DPE molecules on the ice surface was evaluated by prolonged exposure of contaminated ice grains to ozone at different temperatures and sample preparation techniques (Figure 1). We have already demonstrated that only a small portion of DPE molecules (∼10−20%) on the ice surface were restricted to ozonation reaction at 268 K because they are hidden in the DI layer or deeper in the ice grains.19 In contrast, Figure 1 shows that complete reaction conversion was accomplished at 188 K, thus the molecules had to be displaced fully to the air−ice interface as the DI layer becomes negligible at this temperature,5 and their reaction with ozone had no significant restrictions. On the other hand, the ice production technique had only a minor effect on DPE reaction conversions; thus the adsorbed DPE molecules from the gas phase and ozonized at 268 K had sufficient time to become partially immersed into the DI layer (thus be less reactive). The Kinetic Model. Using molecular dynamics simulations, Hung and co-workers found that both naphthalene and phenanthrene tend to be displaced to the air−ice interface in supercooled aqueous solutions during freezing at 260 and 270 K, whereas small benzene molecules are partially trapped inside the ice lattice.29 Although DPE is a relatively water-soluble (∼3 × 10−5 mol L−1) and flexible molecule, we have already D

dx.doi.org/10.1021/es304812t | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

decreasing temperature which may also be related to a change of the kinetic model (LH → ER). Diffusion on the Ice Surface. Pöschl and co-workers proposed a scenario for the ozonation reaction on ice, which assumed that the reaction is slowed down at lower temperatures due to the decrease of the diffusion coefficient of the reaction partners.30 Indeed, the DI viscosity was found to be considerably larger than that of bulk water, though affected only slightly by temperature.38 The ice surface diffusion of organic compounds has scarcely been studied to date, for example, hydrophobic butane migrates rapidly on the ice surface with a diffusion coefficient of D ≈ 10−6 cm2 s at 131 K.39 Diffusion of small organic hydrophobic molecules, such as benzyl radicals, restricted by the surrounding “cage” of water molecules of ice surface, was found considerably limited below ∼250 K.4 Although a precise reaction mechanism of DPE ozonation on the ice surface and the rate determining steps need to be resolved, it is obvious that restricted diffusion at low temperatures had either no or very little effect on the course of ozonation in a way that would overrule the observed rate acceleration at lower temperatures in our experiments. NaCl as a Cocontaminant. At high polar contaminant concentrations, a new phase called brine is formed,40,41 which subsequently influences the physical (such as the partitioning of gas-phase species) and chemical processes of these species.7 For example, Le Calve and co-workers have reported that the adsorption of organic compounds, such as acetaldehyde, is driven by their brine solubility.42 Indeed, the effect of the brine layer affected the ozonation ad,NaCl , and ksol,NaCl were of the kinetics. The rate constants ksol obs, kobs obs same magnitude at all experimental temperatures (Figure 2). In contrast to kad obs, which exhibited an increase already at an environmentally relevant temperature of ∼270 K, an increase in these rate constants occurred at ∼220 K. The presence of NaCl appears to cause DPE to dissolve more or perhaps deeper into the surface layer compared to that of pure ice. The vapor deposition technique initially located DPE on the ice surface, but the molecules were subsequently partially buried into the brine layer after equilibration occurred. Analogous to samples prepared by the shock-freezing method, the DPE molecules equilibrated in the brine layer after they were segregated from the ice phase during freezing. The differences in all rate constants below 200 K (Figure 2) became small because both the pure ice surface and the brine layer solidified. Of course, the presence of brine or possibly NaCl crystals on ice should also be an important fact when considering the thermodynamic partition constant and adsorption isotherm of ozone. Abbatt and co-workers found that no reaction was observed between ozone and an aromatic compound adsorbed to solid NaCl particles, and they concluded that the partitioning of ozone onto NaCl particle substrates is very low.43 Therefore, the presence of NaCl crystals on the ice surface below the eutectic temperature should not increase the ozone uptake coefficients. Ozonation Mechanism. McNeill and co-workers have demonstrated that the substrate structure can influence the ozonation reaction of surface-bound aromatic hydrocarbons, which form nonplanar intermediates that partially desorb in the course of the reaction.44 This energy penalty may not be significant for a flexible DPE molecule, the reactive CC double bond of which can freely rotate out of the phenyl ring plane; however, restrictions to the conformational flexibility of the short alkyl chain of valerophenone in frozen solutions have

provided evidence that it is buried into the DI layer only to a small extent at 268 K.19 Now we can simplify a model proposed by Pöschl and co-workers for reactions on atmospheric particles26,30,31 when considering that ozonation reaction occurs primarily on the surface of a thin DI layer, that is, in a sorption layer. Calculations performed by Hung and collaborators revealed that naphthalene and ozone are adsorbed at the air−ice interface, rather than being dissolved into the DI layer or incorporated to the crystal lattice.14 Assuming that the reaction rate is proportional to the number of naphthalene−O3 contacts, these authors found that when the naphthalene concentrations increase, the DI layer becomes thinner and surface adsorption of ozone (and thus the reaction rate) is linearly increased. A nonlinear Langmuir−Hinshelwood-type dependence was found as the number of O3 molecules increased at a constant number of naphthalene molecules. Interestingly, temperature had no significant effect on the reaction rate in the range of 250 and 270 K in these calculations. Similarly, Kahan and Donaldson have studied phenanthrene ozonation kinetics on ice, which showed no temperature dependence between 263 and 243 K.18 When we extended the temperature range in our experiments, we found that the apparent rates are significantly enhanced at lower temperatures (Figure 2). If the reaction is described by the LH model, the equilibrium air-ice surface partitioning of ozone, its concentration, the second-order rate constant, and the number of reaction sites on the surface (eqs 1 and 2) are important parameters. Our data obtained at 258 K can indeed be correlated with this mechanism (Figure 3). However, the kobsvalues obtained at 188 K correlated partially with both this (LH) as well as the ER model, according to which O3 reacts directly from the gas phase (via chemisorption24). In other words, the collision ER mechanism becomes important for the description of the reaction kinetics at low temperatures. Ozone Adsorption. Adsorption of gaseous molecules is usually efficient due to the highly dynamic nature of the ice surface32 but is not, in general, the rate determining step of a heterogeneous chemical reaction; thus the adsorption equilibrium is typically described by an air−ice partitioning coefficient.5 The uptake coefficient γ then describes the probability that a gaseous molecule undergoing a kinetic collision with a surface will be taken up under the given conditions.33 It has been shown that the interaction of ozone with pure ice surfaces is weak (γ in the order of 10−8 to 10−10 at 223−258 K);34 however, O3 loss from the ice surface in the range of 195 to 262 K is insignificant.35 Ozone and water may form a complex due to a Coulomb interaction between the positively charged hydrogen atom and two negatively charged oxygen atoms of O3 or between the oxygen atom of the water molecule and the central atom of ozone.36,37 The binding (adsorption) energy was estimated to be ∼4 kcal mol−1; the O3 desorption from the ice surface was shown to have a kinetic first-order dependence on temperature.37 For this work, the γ coefficients for the reaction at 258 K were calculated using equation S1 and are shown in Figure S1 (SI). Using kImax and KO3 at high ozone concentrations, γ was 1.5 × 10−9, which is in good agreement with the values mentioned above. However, because the kinetic model could not be assigned with confidence to data obtained at 188 K (Figure 3), an estimation of the uptake coefficient shown in SI Figure S1 (γ ≈ 6.6 × 10−7) derived from the LH model must be taken with caution. Interestingly, its value increased with E

dx.doi.org/10.1021/es304812t | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

been reported.45 The temperature dependence of the ozonation rate constants of various alkenes in the gas phase has been established and the activation barriers are 1−5 kcal mol−1;46 chemisorption of O3 by PAHs has been estimated by DFT calculations to be an order of magnitude higher.47 According to the Arrhenius law, the rate of the DPE reaction should decrease by an order of magnitude in the given temperature range provided that ozone addition is the rate determining step. Therefore, this step is either not rate determining or is overruled by a different, slower process. It is also possible that, if reactive oxygen intermediates 24 are involved in the transformation, a higher probability of their formation at lower temperatures (by prolonging their lifetimes) affects the observed ozonation rates. The ASSA as a Measure of the Reaction Probability. Various experimental methods have been used to evaluate the specific surface area of ice.4,5,11,19,48 For example, Abbatt and collagues have found the saturation surface coverage by benzene to increase with increasing temperature.49 The experiments did not provide any evidence that this dependence is related to temperature-dependent molecular mobility in the DI, but the authors argued that a more rigid surface can accommodate fewer adsorbed molecules in saturation. In this work, we calculated the ASSA (e.g., SI Figures S3 and S4) as the amount of DPE located on an ice surface not as a function of its partial pressure or concentration, but as a function of its specific ability to react with a gas-phase reagent− ozone (Figure 4). The ASSA was found to increase with decreasing temperature by a factor of 3 within the studied temperature range, thus this dependence followed the very similar tendency of the rate constants shown in Figure 2. Here we must emphasize that kobs and the ASSA values were obtained using different experimental strategies. While the former variable is an initial rate constant, the latter is related to the DPE surface concentration at which the initial rate constants values start to change their magnitude. We relate this value to the number of gas-accessible (reactive) surface sites ([SS], eqs 1 and 2): A higher ASSA clearly indicates a higher specific surface coverage for the given DPE amounts under corresponding experimental conditions. As a result, the corrected kobs,corr values at 198 K for the same [SS] number as that at 268 K (i.e., the corresponding kobs from Figure 2 was divided by a factor calculated by a fraction of ASSA (198 K)/ ASSA (268 K) obtained from the data in Figure 4) are kad obs,corr ≈ −4 −1 ad,NaCl s , kobs,corr ≈ 7 × 10−4 s−1, and 8 × 10−4 s−1, ksol obs,corr ≈ 5 × 10 ≈ 9 × 10−4 s−1 (see also SI Table S1). These ksol,NaCl(1) obs,corr corrected rate constants are higher than those measured at 268 K, always by approximately the same factor (3−5). Nevertheless, the calculation reflects similar trends shown in Figures 2 and 4 and brings evidence that the gas-accessible surface site number plays an essential role in the observed rate acceleration at lower temperatures. Our previous work suggested that the orientation of organic molecules on the ice surface changes with temperature.13 It is possible that while the DI layer becomes thinner at lower temperatures, DPE surface concentration increases due to a more efficient separation of the contaminants and, at the same time, the mutual π−π stacking interactions4 cause the molecules to occupy a smaller effective surface without losing their reactivity toward ozone. In other words, the entropy factor clearly opposes the enthalpy change. In conclusion, we suggest that an increase of the number of surface reactive sites and possibly higher ozone uptake coefficients are responsible for the observed rate acceleration

of DPE ozonation at the air−ice interface at lower temperatures. The increased number of reactive sites is probably related to the fact that organic molecules are displaced more to the layer surface which makes them more accessible to the gasphase reactants (ozone). The higher ozone uptake coefficient estimated for lower temperatures can in part be attributed to the occurrence of a different kinetic scenario, the Eley−Rideal model in which the ozone adsorption is chemisorption. The (semi)liquid properties of the ice surface also play an important role: a brine layer formed by the NaCl addition allows for a more efficient diffusion of organic contaminants, and may also change their orientation and location at the air/ice interface. Environmental Implications. Simple alkenes, such as ethylene, propene, or 1-butene were found to be produced in the amounts of units of pptv above the snowpack at Summit, Greenland, possibly by photofragmentation of the corresponding carbonyl compounds.50 These species can readily react with atmospheric ozone. DPE was selected as an example of environmentally relevant alkenes; it is sufficiently large and lipophilic to be well separated from the ice phase. The ozonation reaction rate constants in water are influenced by the chemical structure of the reacting alkenes.51 While the electron withdrawing groups, such as Cl or carboxyl, attached to the CC bond considerably decrease the rate constant, the reactivities of aliphatic and aromatic alkenes toward ozone are comparable and high (in the order of 105 M−1 s−1). However, our results suggest that it is not the rate constant of the ozonation reaction, but the temperature-dependent number of surface reactive sites that is responsible for the observed rate acceleration. The physical properties of alkenes, such as their solubility in the DI layer or volatility, would certainly influence the reaction. Further studies are now being conducted in our laboratory to determine the ozonation reaction kinetics of other environmentally relevant substrates. The ozone concentrations used in this work (0.1−500 ppmv) are substantially higher than those of tropospheric ozone in polar areas (∼20−50 ppbv detected at some Northern Hemisphere stations52). Previously we calculated that the halflife of DPE on the surface of artificial snow grains would be 5.1 and 2.2 days at cO3 = 20 and 50 ppbv, respectively, at submonolayer DPE loadings and 258 K.19 Using the new results shown in Figure 3, and assuming that the LH model can be applied at such low ozone concentrations, the DPE half-life is estimated to be ∼13 h at 188 K; that is, four times shorter. It has been shown that photochemical depletion is the key removal mechanism of ozone in the snowpack interstitial air in polar areas;53 local surface ozone concentration probably cannot be influenced by the reactions with various substrates when their concentrations are orders of magnitude lower. The sodium and chloride ions present in the polar snowpack54 are predominantly derived from ocean spray. Our results show that the presence of NaCl may affect the rate constants of chemical processes occurring on the ice surface, when it creates a brine layer or forms crystals. The unexpected increase in the reactivity of alkenes on ice surfaces with gaseous ozone with decreasing temperature is a remarkable phenomenon that deserves further investigations.



ASSOCIATED CONTENT

* Supporting Information S

Uptake coefficient calculation, determination of the observed rate constants, determination of the ASSA, and the observed F

dx.doi.org/10.1021/es304812t | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

(11) Domine, F.; Cincinelli, A.; Bonnaud, E.; Martellini, T.; Picaud, S. Adsorption of phenanthrene on natural snow. Environ. Sci. Technol. 2007, 41, 6033−6038. (12) Ardura, D.; Kahan, T. F.; Donaldson, D. J. Self-association of naphthalene at the air−ice interface. J. Phys. Chem. A 2009, 113, 7353−7359. (13) Heger, D.; Nachtigallova, D.; Surman, F.; Krausko, J.; Magyarova, B.; Brumovsky, M.; Rubes, M.; Gladich, I.; Klan, P. Selforganization of 1-methylnaphthalene on the surface of artificial snow grains: A combined experimental−computational approach. J. Phys. Chem. A 2011, 115, 11412−11422. (14) Liyana-Arachchi, T. P.; Valsaraj, K. T.; Hung, F. R. Molecular simulation study of the adsorption of naphthalene and ozone on atmospheric air/ice interfaces. J. Phys. Chem. A 2011, 115, 9226−9236. (15) Chen, J.; Ehrenhauser, F.; Liyana-Arachchi, T. P.; Hung, F. R.; Wornat, M. J.; Valsaraj, K. T. Adsorption of gas-phase phenanthrene on atmospheric water and ice films. Polycycl. Aromat. Com. 2011, 31, 201−226. (16) Klan, P.; Holoubek, I. Ice (photo)chemistry. Ice as a medium for long-term (photo)chemical transformations - environmental implications. Chemosphere 2002, 46, 1201−1210. (17) Donaldson, D. J.; Valsaraj, K. T. Adsorption and reaction of trace gas-phase organic compounds on atmospheric water film surfaces: A critical review. Environ. Sci. Technol. 2010, 44, 865−873. (18) Kahan, T. F.; Donaldson, D. J. Heterogeneous ozonation kinetics of phenanthrene at the air-ice interface. Environ. Res. Lett. 2008, 3, 045006. (19) Ray, D.; Kurkova, R.; Hovorkova, I.; Klan, P. Determination of the specific surface area of snow using ozonation of 1,1-diphenylethylene. Environ. Sci. Technol. 2011, 45, 10061−10067. (20) Bunnelle, W. H. Preparation, properties, and reactions of carbonyl oxides. Chem. Rev. 1991, 91, 335−362. (21) Masel, R. I. Principles of Adsorption and Reaction on Solid Surfaces; John Wiley & Sons: New York, 1996. (22) Pöschl, U.; Letzel, T.; Schauer, C.; Niessner, R. Interaction of ozone and water vapor with spark discharge soot aerosol particles coated with benzo[a]pyrene: O3 and H2O adsorption, benzo[a]pyrene degradation, and atmospheric implications. J. Phys. Chem. A 2001, 105, 4029−4041. (23) Kwamena, N. O. A.; Staikova, M. G.; Donaldson, D. J.; George, I. J.; Abbatt, J. P. D. Role of the aerosol substrate in the heterogeneous ozonation reactions of surface-bound PAHs. J. Phys. Chem. A 2007, 111, 11050−11058. (24) Shiraiwa, M.; Sosedova, Y.; Rouviere, A.; Yang, H.; Zhang, Y. Y.; Abbatt, J. P. D.; Ammann, M.; Pöschl, U. The role of long-lived reactive oxygen intermediates in the reaction of ozone with aerosol particles. Nat. Chem. 2011, 3, 291−295. (25) Ammann, M.; Pöschl, U.; Rudich, Y. Effects of reversible adsorption and Langmuir−Hinshelwood surface reactions on gas uptake by atmospheric particles. Phys. Chem. Chem. Phys. 2003, 5, 351−356. (26) Pöschl, U.; Rudich, Y.; Ammann, M. Kinetic model framework for aerosol and cloud surface chemistry and gas-particle interactions Part 1: General equations, parameters, and terminology. Atmos. Chem. Phys. 2007, 7, 5989−6023. (27) Ammann, M.; Pöschl, U. Kinetic model framework for aerosol and cloud surface chemistry and gas-particle interactions - Part 2: Exemplary practical applications and numerical simulations. Atmos. Chem. Phys. 2007, 7, 6025−6045. (28) Takenaka, N.; Ueda, A.; Maeda, Y. Acceleration of the rate of nitrite oxidation by freezing in aqueous solution. Nature 1992, 358, 736−738. (29) Liyana-Arachchi, T. P.; Valsaraj, K. T.; Hung, F. R. Ice growth from supercooled aqueous solutions of benzene, naphthalene, and phenanthrene. J. Phys. Chem. A 2012, 116, 8539−8546. (30) Shiraiwa, M.; Ammann, M.; Koop, T.; Pöschl, U. Gas uptake and chemical aging of semisolid organic aerosol particles. Proc. Nat. Acad. Sci. U.S.A. 2011, 108, 11003−11008.

rate constants and the ASSA values. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; phone: +420-549494856; fax: +420-549492443. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The project was supported by the Grant Agency of the Czech Republic (P503/10/0947), 7FP EU ArcRisk (226534), and the project CETOCOEN (CZ.1.05/2.1.00/01.0001) granted by the European Regional Development Fund. The authors express their thanks to Markus Ammann, Dominik Heger, and Jakob Wirz for fruitful discussions, and Hana Lišková for helping with the GC-MS analyses.



REFERENCES

(1) Grannas, A. M.; Jones, A. E.; Dibb, J.; Ammann, M.; Anastasio, C.; Beine, H. J.; Bergin, M.; Bottenheim, J.; Boxe, C. S.; Carver, G.; Chen, G.; Crawford, J. H.; Domine, F.; Frey, M. M.; Guzman, M. I.; Heard, D. E.; Helmig, D.; Hoffmann, M. R.; Honrath, R. E.; Huey, L. G.; Hutterli, M.; Jacobi, H. W.; Klan, P.; Lefer, B.; McConnell, J.; Plane, J.; Sander, R.; Savarino, J.; Shepson, P. B.; Simpson, W. R.; Sodeau, J. R.; von Glasow, R.; Weller, R.; Wolff, E. W.; Zhu, T. An overview of snow photochemistry: Evidence, mechanisms and impacts. Atmos. Chem. Phys. 2007, 7, 4329−4373. (2) Grannas, A. M.; Bogdal, C.; Hageman, K. J.; Halsall, C.; Harner, T.; Hung, H.; Kallenborn, R.; Klán, P.; Klánová, J.; Macdonald, R. W.; Meyer, T.; Wania, F. The role of the global cryosphere in the fate of organic contaminants. Atmos. Chem. Phys. Discuss. 2012, 12, 16923− 17000. (3) Betterton, E. A.; Anderson, D. J. Autoxidation of N(III), S(IV), and other species in frozen solution - A possible pathway for enhanced chemical transformation in freezing systems. J. Atmos. Chem. 2001, 40, 171−189. (4) Kurkova, R.; Ray, D.; Nachtigallova, D.; Klan, P. Chemistry of small organic molecules on snow grains: The applicability of artificial snow for environmental studies. Environ. Sci. Technol. 2011, 45, 3430− 3436. (5) McNeill, V. F.; Grannas, A. M.; Abbatt, J. P. D.; Ammann, M.; Ariya, P.; Bartels-Rausch, T.; Domine, F.; Donaldson, D. J.; Guzman, M. I.; Heger, D.; Kahan, T. F.; Klan, P.; Masclin, S.; Toubin, C.; Voisin, D. Organics in environmental ices: sources, chemistry, and impacts. Atmos. Chem. Phys. 2012, 12, 1−26. (6) Wettlaufer, J. S. Impurity effects in the premelting of ice. Phys. Rev. Lett. 1999, 82, 2516−2519. (7) Bartels-Rausch, T.; Jacobi, H.-W.; Kahan, T.; Thomas, J.; Thomson, E. S.; Abbatt, J.; Ammann, M.; Blackford, J. R.; Bluhm, H.; Boxe, C.; Domine, F.; Frey, M. M.; Gladich, I.; Guzman, M. I.; Heger, D.; Huthwelker, T.; Klan, P.; Kuhs, W. F.; Kuo, M.; Maus, S.; Moussa, S.; McNeill, V. F.; Newberg, J. T.; Pettersson, J. B. C.; Roeselova, M.; Sodeau, J. Relationship between the microstructure of snow and its physical processes and chemical reactions. Atmos. Chem. Phys. Discuss. 2012, 12, 30409−30541. (8) Roth, C. M.; Goss, K. U.; Schwarzenbach, R. P. Sorption of diverse organic vapors to snow. Environ. Sci. Technol. 2004, 38, 4078− 4084. (9) Wania, F.; Hoff, J. T.; Jia, C. Q.; Mackay, D. The effects of snow and ice on the environmental behaviour of hydrophobic organic chemicals. Environ. Pollut. 1998, 102, 25−41. (10) Ullerstam, M.; Abbatt, J. P. D. Burial of gas-phase HNO3 by growing ice surfaces under tropospheric conditions. Phys. Chem. Chem. Phys. 2005, 7, 3596−3600. G

dx.doi.org/10.1021/es304812t | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

(31) Shiraiwa, M.; Garland, R. M.; Poschl, U. Kinetic double-layer model of aerosol surface chemistry and gas-particle interactions (K2SURF): Degradation of polycyclic aromatic hydrocarbons exposed to O3, NO2, H2O, OH, and NO3. Atmos. Chem. Phys. 2009, 9, 9571− 9586. (32) Abbatt, J. P. D. Interactions of atmospheric trace gases with ice surfaces: Adsorption and reaction. Chem. Rev. 2003, 103, 4783−4800. (33) Crowley, J. N.; Ammann, M.; Cox, R. A.; Hynes, R. G.; Jenkin, M. E.; Mellouki, A.; Rossi, M. J.; Troe, J.; Wallington, T. J. Evaluated kinetic and photochemical data for atmospheric chemistry: Volume VHeterogeneous reactions on solid substrates. Atmos. Chem. Phys. 2010, 10, 9059−9223. (34) Langenberg, S.; Schurath, U. Ozone destruction on ice. Geophys. Res. Lett. 1999, 26, 1695−1698. (35) Dlugokencky, E. J.; Ravishankara, A. R. Laboratory measurements of direct ozone loss on ice and doped-ice surfaces. Geophys. Res. Lett. 1992, 19, 41−44. (36) Sennikov, P. G.; Ignatov, S. K.; Schrems, O. Complexes and clusters of water relevant to atmospheric chemistry: H2O complexes with oxidants. ChemPhysChem 2005, 6, 392−412. (37) Borget, F.; Chiavassa, T.; Allouche, A.; Aycard, J. P. Experimental and quantum study of adsorption of ozone (O3) on amorphous water ice film. J. Chem. Phys. B 2001, 105, 449−454. (38) Goertz, M. P.; Zhu, X. Y.; Houston, J. E. Exploring the liquidlike layer on the ice surface. Langmuir 2009, 25, 6905−6908. (39) Livingston, F. E.; Smith, J. A.; George, S. M. General trends for bulk diffusion in ice and surface diffusion on ice. J. Phys. Chem. A 2002, 106, 6309−6318. (40) Kuo, M. H.; Moussa, S. G.; McNeill, V. F. Modeling interfacial liquid layers on environmental ices. Atmos. Chem. Phys. 2011, 11, 9971−9982. (41) Cho, H.; Shepson, P. B.; Barrie, L. A.; Cowin, J. P.; Zaveri, R. NMR investigation of the quasi-brine layer in ice/brine mixtures. J. Chem. Phys. B 2002, 106, 11226−11232. (42) Petitjean, M.; Mirabel, P.; Le Calve, S. Uptake measurements of acetaldehyde on solid ice surfaces and on solid/liquid supercooled mixtures doped with HNO3 in the temperature range 203−253 K. J. Phys. Chem. A 2009, 113, 5091−5098. (43) Kwamena, N. O. A.; Thornton, J. A.; Abbatt, J. P. D. Kinetics of surface-bound benzo a pyrene and ozone on solid organic and salt aerosols. J. Phys. Chem. A 2004, 108, 11626−11634. (44) Chu, S. N.; Sands, S.; Tomasik, M. R.; Lee, P. S.; McNeill, V. F. Ozone oxidation of surface-adsorbed polycyclic aromatic hydrocarbons: Role of PAH surface interaction. J. Am. Chem. Soc. 2010, 132, 15968−15975. (45) Klan, P.; Janosek, J.; Kriz, Z. Photochemistry of valerophenone in solid solutions. J. Photochem. Photobiol. A 2000, 134, 37−44. (46) Leather, K. E.; McGillen, M. R.; Percival, C. J. Temperaturedependent ozonolysis kinetics of selected alkenes in the gas phase: An experimental and structure-activity relationship (SAR) study. Phys. Chem. Chem. Phys. 2010, 12, 2935−2943. (47) Maranzana, A.; Serra, G.; Giordana, A.; Tonachini, G.; Barco, G.; Causa, M. Ozone interaction with polycyclic aromatic hydrocarbons and soot in atmospheric processes: Theoretical density functional study by molecular and periodic methodologies. J. Phys. Chem. A 2005, 109, 10929−10939. (48) Domine, F.; Taillandier, A. S.; Simpson, W. R. A parameterization of the specific surface area of seasonal snow for field use and for models of snowpack evolution. J. Geophys. Res., Earth Surf. 2007, 112, F02031. (49) Abbatt, J. P. D.; Bartels-Rausch, T.; Ullerstam, M.; Ye, T. J. Uptake of acetone, ethanol and benzene to snow and ice: effects of surface area and temperature. Environ. Res. Lett. 2008, 3, 045008. (50) Swanson, A. L.; Blake, N. J.; Dibb, J. E.; Albert, M. R.; Blake, D. R.; Rowland, F. S. Photochemically induced production of CH3Br, CH3I, C2H5I, ethene, and propene within surface snow at Summit, Greenland. Atmos. Environ. 2002, 36, 2671−2682.

(51) Hoigne, J.; Bader, H. Rate constants of reactions of ozone with organic and inorganic compounds in water. 1. Non-dissociating organic compounds. Water Res. 1983, 17, 173−183. (52) Helmig, D.; Oltmans, S. J.; Carlson, D.; Lamarque, J. F.; Jones, A.; Labuschagne, C.; Anlauf, K.; Hayden, K. A review of surface ozone in the polar regions. Atmos. Environ. 2007, 41, 5138−5161. (53) Helmig, D.; Boulter, J.; David, D.; Birks, J. W.; Cullen, N. J.; Steffen, K.; Johnson, B. J.; Oltmans, S. J. Ozone and meteorological Summit, Greenland, boundary-layer conditions at during 3-21 June 2000. Atmos. Environ. 2002, 36, 2595−2608. (54) de Caritat, P.; Hall, G.; Gislason, S.; Belsey, W.; Braun, M.; Goloubeva, N. I.; Olsen, H. K.; Scheie, J. O.; Vaive, J. E. Chemical composition of arctic snow: Concentration levels and regional distribution of major elements. Sci. Total Environ. 2005, 336, 183−199.

H

dx.doi.org/10.1021/es304812t | Environ. Sci. Technol. XXXX, XXX, XXX−XXX