J. Phys. Chem. 1994,98, 8780-8784
8780
Study of the Uptake of
N205
by Water and NaCl Solutions
Ch. George,+J. L. Ponche, and Ph. Mirabel' Facult.? de Chimie et Centre de Gkochimie de la Surface 28, rue Goethe, F-67083 Strasbourg Cedex, France
W. Behnke, V. Scheer, and C. Zetzsch' Fraunhofer-Institut fur Toxikologie und Aerosolforschung, Nikolai-Fuchs-Str. 1 , 0-30625 Hannover, Germany Received: March 30, 1994; In Final Form: June 24, 1994' The uptake of N2O5 by pure water and NaCl solution was studied as a function of temperature in the range from 262 to 278 K with the droplet train technique, where a highly controlled beam of droplets was exposed to N2O5 in a low-pressure flow tube reactor, and the formation of nitrate in the liquid phase was determined by ion chromatography. The uptake coefficients, y, for N2O5 on pure water are observed to decrease from 0.03 to 0.0 13 with increasing temperature. This behavior corresponds to the expected negative temperature dependence of mass accommodation leading to an enthalpy m o b , = (-9.6 f 1.6) kcal mol-' and to an entropy m o b s = (-43 f 6) cal mol-' K-l for the phase transfer, corresponding to a continuous nucleation process with a critical cluster size N* of about 2.4. A significantly lower yield of nitrate than with pure water is observed in the experiments on NaCl solution (1 mol/L), indicating that nitrogen compounds (such as ClN02) are formed after the uptake of N2O5 by subsequent reactions with NaCl and escape from the droplets. After correction for the known yield for the formation of CIN02, the results exhibit a slight systematic tendency for the uptake coefficient on NaCl solution to be greater than on pure water, indicating that the uptake of N2O5 by these aqueous media might be reaction-controlled. This assumption leads to a lower limit of 800 mol L-I atm-I s-l12 for the product Hk112 from a simple steady state model (where H is the Henry's law constant for N2O5 and k is the first-order hydrolysis rate constant).
Introduction The predominant permanent removal process of NO, from the atmosphere is initiated by the oxidation of NO2 to NO3 radicals by ozone, followed (via reaction with another NO2 molecule) by the formation of NzOs. Heterogeneous reactions with the aqueous content of aerosols in the troposphere or with ice surfaces in the stratosphere are thought to convert N2O5 to nitric acid rapidly. Since the gas phase reaction of N2O5 with H2O is slow,l with a rate constant lower than 2.7 X cm3 s-l, and since N2O5 is thermally stable2 with a lifetime of 0.35 min at 298 K, it plays an important role in the tropospheric night-time chemistry as a reservoir for NO3 radicals.3~~The apparent lifetime of NO3 radicals (in equilibrium with NzO5) has been observed to decrease from more than 1 h below 20% relative humidity (r.h.) to less than 10 min above 50% r.h. by Platt et al. in field measurements in California during the night.5 This apparent lifetime of NO3 can be attributed to heterogeneous loss of N2O5. The uptake of N2O5 by aerosols with, possibly, high electrolyte concentration (close to the deliquescent point) may yield compounds other than H N 0 3 . When the ozone depletion during springtime in the Antarctic stratosphere had been detected, it was soon recognized that the reaction of N205 with HC1 adsorbed on ice particles in the stratosphere can lead to ClN02, a photolytic precursor of atomic C1. Reaction probabilities of N205 on ice were studied by Tolbert et a1.6 and Leu7 and were observed to increase with the HCl content of the ice. A formation of ClNO2 from N2O5 and HCl was inferred from observations on Pyrex and Teflon surfaces by Leu.' The reaction of N2O5 with NaCl salt (as a model for sea salt) was shown to lead to ClNO2 as well,* a potential photolytic source of atomic C1 in the troposphere. The tropospheric relevance of this process for sea spray was demonstrated in a smog chamber experiment on dry and deliquescent NaCl aerosol in Teflon bags,g where yields of about 50% were
'Present address: Fraunhofer-Institute of Toxicologyand Aerosol Research,
Hannover, Germany. 8 Abstract published in Aduance ACS Absrracts, August 1, 1994.
0022-365419412098-8780$04.50/0
observed for CIN02. During the same experiments, a sticking coefficient for N2O5 of about 0.03 was determined, fairly independent of r.h. (at least between 71 and 92%). This formation process of ClNOz was estimated9 to lead to levelsof atomic Clup to4.3 X lO3cm-3 in the tropospheric marine boundary layer at 60' Northern Latitude with typical wind speeds of January from a model calculation assuming a level of 0.5 ppb NO,. With heterogeneous loss processes of N2O5 and NO3 on the walls of the Teflon bags taken into account,IO the sticking coefficient of N2O5 has been refined to be y f 2a = 0.032 f 0.003, meanwhile, at r.h. between 77 and 94% at 291 K (corresponding to an aqueous concentration of NaCl in the range 5.8-1.8 mol/L). The formation of ClNOz at high yields has been confirmed in dilute solutionsll down to 1 V mol/L, and the levels of CI were updated', to 1 X lo3 ~ m at- a~typical level of 0.1 ppb N02. Such a source of atomic CI has to be taken into account since it may contribute a factor of 2 to the source strength from the well-known reaction of O H with HCll3 and to other (still highly speculative) heterogeneous sources of atomic C1 in the troposphere,14 possibly initiated by the reaction of O3with C1-. Some evidence for the existence of photoactive precursors of atomic C1 in the marine boundary layer exists from recent field measurement^.^^ Since mass accommodation or reactivity-limited uptake may control the heterogeneous removal of NzOs from the atmosphere but also the production of ClN02 as a major precursor of atomic C1 in the troposphere, the uptake was investigated in the present study over pure water as a model for clouds and over a 1 M NaCl solution as a model for sea spray aerosol at 96.7% relative humidity. Experimental Section Production of NzOs. NzOs was produced by the following reaction sequence:
+ 0, NO, + 0, NO, + NO, 2 N 2 0 5
NO,
+
0 1994 American Chemical Society
(1)
( 2 ,-2)
The Journal of Physical Chemistry, Vol. 98, No. 35, 1994 8781
Uptake of N2O5 by Water and NaCl Solutions
1
N2°5
0.0 0.1
0
1
2000 1800 1600 1400 1200 loo0
I 800
600
Wavenumbedcm-1 Figure 1. Typical spectrum, containing N205/HNO3/03 in the range 2000-600 cm-1. The HNOl absorption at 1325 cm-* is influenced by an increasingnitrate absorption on the AgCl windowsduring the experiment. The HNOJabsorptionat 879 cm-I is influencedby an additionalabsorption band of N2O5; also, the absorption at 1712 cm-I is overlapping with the absorption band of N2O5 at 1726 cm-'. For this purpose, a flow of synthetic air, containing 600 ppm (or for higher N2O5 concentrations, 10000 ppm) NO2 (Messer Griesheim), was mixed in a tubular glass vessel with O3produced by an electric discharge generator (Fischer, Model 500) from pure02 (99.999%, Messer Griesheim). Both gas flows weredried over P2O5 (Riedel de Haen). The average reaction time in the vessel was 1 min. To make sure that N2O5 was the main product after the reaction and that NO2 and NO3 were only minor, we used O3 in excess (50-100 ppm O3 were normally contained in the N2O5 flow after reaction). HNO3 impurities were removed by a Nylon filter to less than 8% of the Nz05concentration and were therefore neglected in the later evaluation of the uptake. This technique delivered flow rates of 2 L/min with N2O5 concentrations in the range from 20 to 2000 ppm through Teflon tubing (PTFE, 4 X 20 cm in parallel, inner diameter 4 mm) to the droplet train apparatus. Analytical Methods. The gas flow containing N205 was monitored by FTIR spectroscopy (Bruker IFS 113) in the range 4000-400 cm-l in a 20 cm glass cuvette with AgCl windows at a resolution of 0.5 cm-1. Figure 1 shows a typical absorption spectrum, where 64 scans were co-added. The spectrum shows clearly at 879 cm-1 that traces of HN03are present. N2O5 was determined using the bands at 1726, 1246, and 744 cm-I. The spectra were calibrated by sampling the gas on wet denuders followed by ion chromatography in order to quantify the formation of NO3-. M The ion chromatography was performed with 1.1 X Na2CO3/2.57 X 10-3 M NaHC03 eluent at a flow rate of 1.3 mL/min using a Dionex AS4 column combined with a continuously ion suppressing system (regenerated by a 0.025 M H2SO4 solution) with conductivity detection. To avoid an overload of the column with C1- (that would otherwise interfere with the chromatographic separation of NO3-), the samples of the droplets from the NaCl experiments were diluted by a factor of 10 with eluent. Verification of the Absolute NlOs Level. For the experiments it was necessary to humidify the gas containing N2O5. Since the heterogeneous hydrolysis is expected to be fast, we verified to which amount of relative humidity N2O5 remains stable in Teflon tubing. A continuous flow of N2O5 was produced at atmospheric pressure in a manner similar to that for the droplet train experiments. Humidified air was then added 15 cm before the detection system. The detection system is based on a wettedwall flow tube technique,'O where the N205 reacts with NaCl solution and ClNO2 is produced. The yield of CIN02 in the presence of humidity divided by the yield under dry conditions represents the portion of the N2O5 surviving passage through the tube. Figure 2 shows the behavior of this ratio as a function of
I
I
t
0.6 1
I
I
10
20
I
I
30 40 r.h./%
1
50
60
Figure 2. Ratio of the yield of ClN02 in the presence. of humidity to the yield of C1N02 in the absence of humidity observed in the wetted-wall flow tube at various humidities in the inlet Teflon tubings.
relative humidity. The loss of N205 in the Teflon tube is negligible at low r.h. and becomes pronounced above 30%. The loss is less than 10% a t relative humidities below 30% (at a total flow rate of 200 mL/min). Droplet TrainTechnique. In previous publications,1618 we have presented a detailed description of the experimental setup. Here, we will only provide a brief summary of the principles of operation. The experimental apparatus involves a highly controlled train of monodispersed droplets passing through different and independent chambers at low pressure (typically between 20 and 60 Torr). These latter are (1) the droplet generation chamber, (2) a separating buffer chamber, (3) the interaction chamber (covered by Teflon film FEP 200, Dupont) where N2O5 is injected and can react with the train of droplets, (4) a second buffer chamber, and finally (5) the collection chamber where the droplets are collected in order to be analyzed by HPLC (ion chromatography). The droplets are generated by the vibrating orifice method, based on forcing water through a calibrated orifice that is excited by a piezoelectric ceramic. In this study, the diameter, d, of the droplets was in the range 80-150 pm, determined from the vibrational frequency of the orifice and the volume flow rate of the liquid. The transit time of the droplets in the interaction chamber is short; the combination of different droplet velocities with different lengths of the interaction zone provides typical gas/liquid contact times between 4 and 16 ms. The length used in the present experiments was equal to 9.5 cm. In addition to the interfacial resistance (represented by the mass accommodation coefficient, a),the rate of mass transfer can be limited by several processes (gas phase diffusion, surface saturation, and reaction in the liquid phase). This means that under our experimental conditions, where we are subject to these limitations, it will be impossible to measure the mass accommodation coefficient directly. Instead, we measure an uptake coefficient yob (with yoh d a) which represents the measured flux of molecules through the interface. This uptake coefficient is a convolution of all processes which may influence the rate of heterogeneous mass transfer of N2O5. These processes will be discussed in a following section. The trace gas density ( n 2 ) is monitored by FTIR at the entry of the flow tube (see section on analytical methods) while the quantity of gas absorbed by the droplets (nab) is derived from the aqueous concentration measured by HPLC in the collected droplets. Thus the observed uptake coefficient yo& can be c a l ~ u l a t e dvia ~ ~eq 3 from the integrated loss of N2O5 in the interaction zone:
where E is the trace gas average thermal velocity, F, is the carrier gas volume flow rate, and Nu@ is the surface area exposed by the N drodets Dresent in the interaction zone. r .
8782 The Journal of Physical Chemistry, Vol. 98, No. 35, 1994
George et al.
TABLE 1: Experimental Conditions, Number of Runs, N, and Uptake and Mass Accommodation Coefficients, y, for Nz05 Obtained on Water and on NaCl Solution (1 mol/L), and Statistical Error, u [bo1 T P (Torr) (K) (Torr) N yabf u Y f U ?’f# 1.9 2.9 4.6 5.1 6.2
262 261 213 276 211
24.1 25.3 25.2 41-65 36-62
2.1 3.0 4.4 6.2
263 268 273 278
24.1 26.2 25.4 37-55
a 6 6 10 16
a 5 5 11
On Water 0.024 f 0.001 0.030f 0.002 0.022f 0.001 0.029f 0.012 0.017 f 0,001 0.020f 0.002 o.oi1-0.012 0.016 f 0.008 o.ooa-0.013 0.013 A 0.008 On NaCl Solution 0.016f 0.001 0.022f 0.006 0.015 f 0.003 0.021 f 0.006 o.ooi5 0.001 0.008f 0.001 0.0036-0.ooa 0.007 f 0.003
.I 8
V
43”
0.039 f 0.013 0.037 f 0.012 0.014 0.003 0.014f 0.008
Corrected for the reevaporation of CIN02 in experiments on NaCl solution (see text). An important aspect of this technique is the careful control of the partial pressure of water in the flow tube since it controls the surface temperature of droplets through evaporative cooling.19 In the present study, the partial pressure of water was varied between 1.9 and 6.2 Torr, leading to surface temperatures between 262 and 277 K, respectively, for pure water. For the experiments performed with NaCl solutions, one has to consider the lowering of the saturated vapor pressure of the solution by the dissolved NaCl compared to pure water.20 In this latter case, the partial pressure was between 2.1 and 6.2 Torr, leading to temperatures between 263 and 278 K, respectively. Results The uptake of N205 by water droplets was studied as a function of temperature between 262 and 277 K. The measured uptake coefficients are listed in Table 1, where each value is an average of at least five runs. These measurements correspond to the observation of the overall kinetics of uptake, and therefore the uptakecoefficient may reflect the different limitations present during the incorporation. These latter could be due to the gas phase diffusion, surface saturation, or slow chemical reaction in the aqueous phase as discussed in different previous papers.l8J9 N2O5 is expected to react rapidly with water,21s22and thus it is unlikely that the surface of our droplets becomes saturated. Therefore, under our experimental conditions N2O5 can be considered as an infinitely soluble gas, and the experiments cannot distinguish if this corresponds to physical solvation or reactive uptake.22 The uptake of N205 by the droplets changes the surface gas phase concentration, which defines the uptake, becoming smaller than the measured bulk average concentration. The gas density is then replenished by mass transport, which is assumed to be due to molecular diffusion. Therefore, gas phase diffusion may limit the uptake. To take this into account, we have to correct our measured uptake coefficient for diffusive mass transport18 using eq 4, which yields the corrected uptake coefficient y:
7’
,
1
(4)
-.
.I 8 e
0.10 0.09 . 0.08 . 0.07 . 0.06 . 0.05 . 0.04 .
0.03
0.10 0.09 0.08 0.07 0.06 0.05
. . . .
0
H*Opvre
&
NnCl (1 molil)
’
H E:
9
.
0.02 . 0.01 ’ 0.00
0.02 . 0.01 0.00
.
%i o b 4
gas mixture and D,H20-N205 and D,N2-~205 are the binary diffusion coefficients of N205 in H20 and in N2, respectively. These latter have been calculated by a semiempirical methodz4 to be 0.17 and 0.13 cm2SKI, respectively, at 1 atm and 298 K with a temperature dependence for D,
[email protected] and D, N2-N205, An additional correction of the uptake coefficients for the nonMaxwellian velocity distribution near the droplet surfaceI8would be less than 2% and is therefore neglected here. The uptake coefficients calculated via eq 4 are listed in Table 1 and shown in Figure 3. We also studied the uptake of N205 by NaCl solution droplets at 1 mol/L. In this case, we observed a lowering of the uptake coefficient obtained by the treatment given above (see Table 1 and Figure 4). To explain this effect, one has to remember that our data evaluation relies on an indirect determination of the dissolved N205; Le., we detect the presence of N03- after the hydrolysis of N205. The lower calculated uptake coefficients reflect a loss of nitrogen from our droplets; Le., some nitrogencontaining compounds are reevaporating during the interaction. According to our earlier ClNO2 is formed and escapes back to the gas phase. Using the yield of formation of ClNO2 of (67 f 7)% from the portion of N205 taken up by the solution measured at 1 mol/L NaCl in a wetted-wall flow tube,” we can correct the measured NO3- concentrations present in the collected droplets in order to calculate the “truenuptake coefficient. Figure 5 compares the corrected uptake coefficients for NaCl droplets with the values obtained on pure water. The agreement is reasonable, confirming that ClNO2 is formed after the uptake of N z 0 5by NaCl droplets as previously identified.11326 Discussion
where d is the droplet diameter, shown to be equivalent1’ in size to the effective diameter,l8 and P i s the total pressure. D, denotes the gas phase diffusion coefficient in the mixed background gases given by23 -1= Dg
xH20 Dg H2CLN205
+
XN2 Dg N2-N205
where X H 2 0 and xN2 are the mole fractions of H20 and N2 in the
For the particular case of N2O5, the experiments performed in this study cannot distinguish if the uptake is limited by mass accommodation (characterized by the mass accommodation coefficient, CY)or limited by the reactivity of N205 in the liquid phase (characterized by a reactivity-limited uptake coefficient, yrxn).However, the measured uptake rates are consistent with previously reported studies. In Figure 6, we have plotted the data found in the literature. These latter are expressed as “uptake probabilities” since it was not always possible to distinguish the
The Journal of Physical Chemistry, Vol. 98, No. 35, 1994 8783
Uptake of N2O5 by Water and NaCl Solutions 0.10
A N.EI (1 moYn (m"
3 0.03 -5
0.02 0.01
".""
250
260
no
280 290 Temperature (K)
180
0.0036
240 260 Temperature (K)
220
280
0.0037
0.0038
1rr (K-1)
I
200
I
1
-6
300
Figure 5. Comparison between the measured uptake coefficient of N205 on aqueous droplets and NaCl solution, corrected for the reevaporation of CIN02. 0.20 1
I
300
Figure 6. Uptake probability of N z 0 5 on water, ice, HzS04 mixtures, and aerosols of various composition found in the literature. On water: ( 0 )this work, (m) Kirchner et al. (1990), (A)Van Doren (1990). On aerosols of various composition: (0)Behnke et al. (1992), (0) Mozurkewitch and Calvert (1988). On ice: (0) Hanson and Ravishankara (1991), (El) Quinlan et al. (1990), (A)Leu (1988). On H2S04 mixtures: (v)Van Doren et al. (1991), (0)Worsnop et al. (1993), (0) Hanson and Ravishankara (1991), (0)Fried et al. (1993).
reactive uptake from the mass accommodation. This figure gives an overview of the heterogeneous chemistry of N205over different surfaces: on water,21322 on ice,7,27J* on aerosols of different c o m p o ~ i t i o n s , ~and ~ ~on 2 ~various H2SO4 mixt~res;2*-3%~3 the early values by Harker and Strauss30between 220 and 240 K are below 10-4 and not included in the figure. Our measured values agree well with the lower limit of the study by Kirchner et a1.21 but exhibit a difference as large as a factor of 2 from the results of VanDoren et ~1.22obtained with a similar technique but by monitoring the loss of N2O5 in the presence of the droplets instead of the uptake in the liquid phase. This difference might be explained by a better stability of N205 in our Teflon apparatus, even in slightly humidified gases where wall reactions and losses of N2O5 were only minor. Under the least favorable experimental conditions of our droplet train experiments (23% r.h.), the total loss of N 2 0 5can be estimated to be about 0.5% from the decay rate derived from the data of Figure 2. The measured "uptake" coefficients exhibit a negative temperature dependence, as previously observed by VanDoren et a1.z2 for highly soluble gases for which the rate-limiting step is a part of the physical solvation process. In that particular case, the observed negative dependence may indicate that the rate-limiting step for the uptake of N2O5 could be the nonreactive mass accommodation. In the model developed by Jayne et al. (1992), the mass accommodation coefficient can be expressed as
where AG#,h can be regarded as the height of the Gibbs free energy barrier for the transition between the gas and solvated state. The enthalpy AHob and entropy AS,h can be derived from a plot of In{./( 1 - a))versus 1/ T , displayed in Figure 7 for our data on water in Table 1 . The slope of such a plot corresponds
Rgure7. Plotofln(o/(l -a))versus l/Taccording toeq6. Thestraight line in this figure represents the linear fit of eq 6 to our data.
to -AH,b,/R while the intercept corresponds to ASob/R. The values obtained for AHob and M o bare (-9.6 f 1.6) kcal mol-' and ( 4 3 f 6) cal mol-' K-I, respectively, where the specified error limits represent one standard deviation. Davidovits et have developed a model in which the mass accommodation is described as a continuous nucleation process where the cluster larger than a critical size (defined as a specific number W of molecules) grows by condensation and merges with the nearby liquid. The relationship between our measured AH0h and Asoh is found to be in agreement with the model and leads to a value of P = 2.4 for the critical size of clusters in the mass accommodation process. It is interesting to note that the measured uptake probabilities by H2S04 mixtures or by aerosols of various composition also exhibit, in general, a negative temperature dependence as shown in Figure 6. However, a careful examination of the results of Figure 5 shows that the uptake of N2O5 is possibly more complicated and may not be controlled by mass accommodation alone. In fact, the "uptake" coefficients measured on the NaCl solution exhibit a slight systematic tendency to be greater than those observed on pure water. This tendency could even be greater than the one depicted in Figure 5 since the production yield used to correct ourvaluesisprobablya lower limit to theactualvalue. Therefore, the uptake of N2O5 by water droplets may be reaction-limited and expected to be greater in more concentrated NaCl solutions as observed previously.II Following these conclusions and earlier work on the reactive uptake of phosgene in a liquid jet,36 the uptake of N2O5 can be described by the reactive process through the equation yrxn= 4HRT(kDa)'/'/E
(7)
where H i s the Henry's law constant for N2O5, D, is its aqueous phase diffusion coefficient, and k is the first-order rate constant for the"destruction or conversion" of N2O5. Taking into account that the measured uptake coefficients are larger than and using a typical value of 10-5 cm2 s-l for D,, one can calculate that H k ' / 2 2 800 mol L-' atm-' s-*I2 Furthermore, by assuming that the Henry's law constant of N2O5 is H 1 2 mol L-I atm-' (considered to be similar to that reported by Schwartz and White,37 for N2O4, H = 1.4 mol L-' atm-1 at 298 K), one may derive from the above lower limit that
k L lo5 s-' In that case, the reaction of N2O5 with water takes place very close to the surface, and as noted by Hanson et al.,3*the uptake coefficients measured in the present study are directly applicable for the modeling of the atmospheric fate of N2O5. In conclusion, we have measured the uptake rates of N2O5 by pure water droplets and NaCl containing droplets as a function of temperature between 262 and 278 K. The measured "uptake" coefficients are larger than 10-2; Le., the measured uptake rates
8184
The Journal of Physical Chemistry, Vol. 98, No. 35, 1994
. -
-
are sufficientlv laree for the heteroaeneous reaction of N,Oc - - to play an important role in the atmosphere. Our data confirm that some nitrogen-containing compounds are formed from the uptake of N ~ Oby S NaCl droplets and transported back to the gas phase. This reaction pathway is important since it is expected to lead to the formation Of a photolytic precursor Of the highly reactive atomic C1. Acknowledgment. This work was supported by the Ministere de la Recherche et de la Technologie and the DAAD under the framework of PROCOPE and by the CNRS and the Ministbre de I'Environnement (ATP, Phase AtmosphCrique des Cycles BiogCochimiques) and the Bundesminister fur Forschung und Technologie (Grant 07EU7671) in the framework of the EUROTRAC subproject HALIPP. References and Notes (1) Sverdrup, G. M.; Spicer, C. W.; Ward, G. F. J . Chem. Kinet. 1987, 19, 191.
(2) Wayne, R. P.; Barnes, I.; Biggs, P.; Burrows, J. P.; Canosa-Mas, C. E.; Hjorth, J.; Le Bras, G.; Moortgat, G. K.; Perner, D.; Poulet, G.; Restelli, G.; Sidebottom, H. Atmos. Environ. 1991, 25A, 1. (3) Platt, U.; Perner, D.; Schroeder, J.; Kessler, C.; Toennissen, A. J . Geophys. Res. 1981, 86, 11965. (4) Chameides, W. L. J. Geophys. Res. 1986,091, 5331. ( 5 ) Platt, U.; Perner, D.; Winer, A. M.; Biermann, H. W.; Atkinson, R.; Pitts, J. N., Jr. Environ. Sci. Technol. 1984, 18, 365. (6) Tolbert, M. A,; Rossi, M. J.;Golden,D. M. Science 1988,240,1018. (7) Leu, M. T. Geophys. Res. Letf. 1988, 15, 851. (8) Finlayson-Pitts, B. J.; Ezell, M. J.; Pitts, J. N., Jr. Nature 1989,337, 241. (9) Zetzsch, C.; Behnke, W. Ber. Bunsen-Ges. Phys. Chem. 1992,96,3, 488. (10) Behnke, W.; Kriiger, H.-U.; Scheer, V.; Zetzsch, C. J . Aerosol Sci. 1992, S23, S933. (11) Behnke, W.;Scheer,V.;Zetzsch,C.J.AerosolSci. 1993,S24,S115. (12) Behnke, W.; Zetzsch, C. Unpublished results. (13) Singh, H. B.; Kasting, J. F. Atmos. Chem. 1988, 7, 261. (14) Zetzsch, C.; Behnke, W. In Proc. NATO ARW The Tropospheric Chemistry in the Polar Regions; Niki, H., Becker, K. H., Eds.; NATO AS1 Series 17; Springer: Berlin/Heidelberg, 1993; p 291. (15) Keene, W. C.; Maben, J. R.; Pszenny, A. A. P.; Galloway, J. N. Environ.Sci. Technol. 1993,27,866. Pszenny, A. A. P.; Keene, W. C.; Jacob, D. J.; Fan, S.;Maben, J. R.; Zetwo, M. P.; Springer-Young, M.; Galloway,
George et al. J. N. Geophys. Res. Lett. 1993,20, 699. (16) Ponche, J. L.; George, C.; Mirabel, P. J . Atmos. Chem. 1993,16, 1. (17) Bongartz, A.; George, C.; Kames, J.; Mirabel, P.; Ponche, J. L.; Schurath, u, J , Amos, (-hem. 1w4, 18, 149. (18) George, C.; Lagrange, J.; Lagrange, P.; Mirabel, P.; Pallares, C.; Ponche, J. L. J. Geophys. Res. 1994,099, 1255. (19) Worsnop, D. R.; Zahniser, M. S.;Kolb, C. E.; Gardner, J. A.; Jayne, J , T.; Watson, L. R,; Van Doren, J , M.; Davidovits, p. J , phys. Chem. 1989, 93, 1159. (20) Pruppacher, H. R.; Klett, J. D. Microphysics of Clouds and Precipitation; Reidel: Dordrecht, The Netherlands, 1980. s,;(21) Schurath, Kirchner, u, J , W.; Armos, Welter, Chem. F.; Bongartz, 1990, A.; 427.Kames, J.; Schweighoefer, (22) Van Doren, J. M.; Watson, L. R.; Davidovits, P.; Worsnop, D. R.; Zahniser, M. S.;Kolb, C. E. J . Phys. Chem. 1990, 94, 3265. (23) Reid, C. R.; S h e r w d , T. K.The Properties of Gases and Liquids; McGraw-Hill O ' .: New 1986; p 520. (24) Fuller, E. N.; Schettler, P. D.;Giddings, J. C. Ind. Eng. Chem. 1966, 58, 19. (25) Behnke, W.; Zetzsch, C. J . Aerosol Sci. 1990, 21, S229. (26) Behnke, W.; Kriiger, H.-U.; Scheer, V.; Zetzsch, C. J. Aerosol Sci. 1991. 22. S609. (27) Quinlan, M. A.; Reihs, C. M.; Golden, D. M.;Tolbert, M. A. J . Phys. Chem. 1990, 94, 3255. (28) Hanson, D. R.; Ravishankara, A. R. J . Geophys. Res. 1991, 096, 17307. (29) Mozurkewich, M.; Calvert, J. G. J . Geophys. Res. 1988,093,15889. (30) Harker, A. B.;Strauss, D. R. Kineticsofthe HeterogeneousHydrolysis of Dinitrogen Pentoxide over the Temperature Range 214-263K. Fed. Aviat. Admin. Publ. FAA-EE-81-3, Rockwell Internat. Sci. Center, Thousand Oaks, CA, 1981. (31) Van Doren, J. M.; Watson, L. R.; Davidovits, P.; Worsnop, D. R.; Zahniser, M. S.; Kolb, C. E. J . Phys. Chem. 1991, 95, 1684. (32) Worsnop, D. R.; Zahniser, M. S.; Kolb, C. E.; Davidovits, P. Heterogeneous Reaction Kinetic Studies of Halogen Compounds on Sulfuric Acid Aerosols. Final report prepared for AFEAS, Contract No. CTR9014/P90-35. Aerodvne Research. Billerica. MA. 1993. '(33) Fried, A.; Henry, B. E.; Calvert, J.'G.; Mozurkewich, M. J. Geophys. Res. 1994, 99, 3517. (34) Jayne, J. T.; Duan, S. X.; Davidovits, P.; Worsnop, D. R.; Zahniser, M. S.; Kolb, C. E. J. Phys. Chem. 1991, 95, 6329. (35) Davidovits, P.; Jayne, J. T.; Duan, S. X.; Worsnop, D. R.; Zahniser, M. S.;Kolb, C. E. J . Phys. Chem. 1991, 95, 6337. (36) Manogue, W. H.; Pigford, R. L. AIChE J. 1960, 6, 494. (37) Schwartz,S. E.; White, W. H. Adu.Enuiron.Sci. Technol. 1981,12, 1. (38) Hanson, D. R.; Ravishankara, A. R.; Solomon, S. J . Geophys. Res. 1994, 99, 3615.
