Sorption of a Diverse Set of Organic Vapors To Urban Aerosols

For a list of the compounds see Table S1 in the Supporting Information (SI). ...... Ivan Coluzza , Jessie Creamean , Michel Rossi , Heike Wex , Peter ...
0 downloads 0 Views 128KB Size
Download PDF
Environ. Sci. Technol. 2005, 39, 6638-6643

Sorption of a Diverse Set of Organic Vapors To Urban Aerosols C H R I S T I N E M . R O T H , * ,† KAI-UWE GOSS,* AND R E N EÄ P . S C H W A R Z E N B A C H Swiss Federal Institute for Environmental Science and Technology (EAWAG), Ueberlandstrasse 133, Postfach 611, CH-8600 Duebendorf, Switzerland, and Swiss Federal Institute of Technology (ETH), Zurich, Switzerland

Sorption to urban aerosols is a key process in determining the transport and fate of organic pollutants in the atmosphere. The sorption properties of two urban aerosol samples have been determined using aerosol/air partition coefficients measured for a large set of diverse organic vapors. The dominant sorption process could be identified for both samples with two complementary methods: (a) by applying poly-parameter linear free energy relationships (LFERs) to the data sets, and (b) by evaluating the specific surface area, the elemental carbon (EC) content, and the organic matter (OM) content of the aerosols in combination with various sorbent-air partition coefficients from the literature. This revealed that sorption to the two urban aerosols was dominated by absorption into OM and that the diverse data set could be evaluated with an absorption model. The data further revealed that neither EC nor OM was fully available for sorption. The latter leads to the hypothesis that aerosol OM in urban samples has characteristics comparable to those of glassy polymers.

Introduction Sorption to urban aerosols crucially affects the transport and degradation fate of gaseous pollutants in the urban environment. Urban aerosols may consist of various components available for sorption, which stem from various sources, including elemental carbon (EC), primary and secondary organic matter (OM), minerals, and salts (1). External mixing of these components leads to a variety of surfaces and bulk phases that are available for sorption. In this case additive sorption behavior to the separate particles can be expected. Internal mixing, however, may form new phases, such as organic films on solid mineral particles or mixtures of salts and organics whose sorption properties are not necessarily equal to the additive sorption behavior of the components. It is still an open question whether adsorption to surfaces or absorption to the bulk phases are relevant for sorption to these mixtures. However, adsorption must be distinguished from absorption processes to describe experimental sorption data adequately and to develop correct predictive sorption models. For urban aerosols, we have shown in earlier works that adsorption to inorganic surfaces, such as minerals or salts, is too small to explain the observed sorption of nonpolar compounds (2). On the basis of composition (1), the only * Address correspondence to either author. Fax: ++41-1-82352-10. E-mail: [email protected] (K.-U.G.); [email protected] (C.M.R.). † Present address: Harvard University, Boston, Massachusetts. 6638

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005

other sorption processes that may be relevant are adsorption to EC or absorption into OM. In a companion paper, we demonstrated that adsorption to EC was the dominating sorption process for particles collected at the outlet of a diesel engine (3). However, in an aerosol sample collected in a road tunnel, both absorption into OM and adsorption to EC appeared to be relevant. The goal of this work was to examine sorption to urban aerosols, and to determine whether absorption into OM or adsorption to EC or inorganic surfaces would dominate the observed sorption. Sorption to aerosols has been investigated by many researchers, however, primarily for a few nonpolar compound classes, including PAHs and PCBs (e.g., refs 4-6). These findings do not facilitate a complete understanding of the physicochemical sorption mechanisms or the extrapolation to polar compound classes. Therefore, we determined equilibrium sorption coefficients for a large number of nonionic, organic compounds encompassing a wide range of polarities. The measured sorption data were evaluated with poly-parameter linear free energy relationships (LFER) (7-9) that take into account all relevant intermolecular interactions, i.e., van der Waals interactions, electrondonor/-acceptor interactions, and cavity formation (only in case of absorption). This allows us a reliable prediction of the sorption of compounds not measured and it helps us in identifying the dominant sorption mechanism. On one sample, measurements up to 80% RH were conducted to quantify the influence of relative humidity on equilibrium sorption. We also examined the similarity between two aerosols collected with different methods in two different regions and from different time periods.

Materials and Methods The Washington urban aerosols were purchased from NIST as Standard Reference Material 1649a (former 1649), Urban Particulate Matter. They were collected with a special bag filter system in a baghouse (10) in Washington, DC during 1976/1977. They were sieved (125 µm) and radiation sterilized. This sample was used as received. The urban aerosols from Chur were collected from a bagfilter used to filter particles from the air supply in an air conditioning system. The filter had been used for 1 year in a building at the main street of Chur, Switzerland. The filter material was a thermally stabilized synthetic fiber filter for fine particle filtration with an efficiency of 60-80% (Unifil, filter type K65-610-H). The bags of the filter were cut open at the edges and carefully laid on aluminum foil. While the filter material was laying with the inner bag side on the aluminum, it was gently moved, folded, and tapped to enhance particle settling onto the aluminum foil. The filter material stayed on the aluminum for several hours to days. The mixture thus collected on the foil consisted of aerosols, fibers, mineral dust, plant leaves, and dead insects, and was sieved two times (1 mm and 125 µm) before storing for 46 h in the freezer (-18 °C), and afterward in the refrigerator (∼7 °C). The urban aerosols were used as pure column packings in inverse gas chromatography (IGC). For a detailed description of the experimental setup and the determination of sorption coefficients from measured retention volumes see our companion paper (3). Experimental details differing from this description are given here. Experimental temperature was 15.0 ( 0.1 °C. The RH chosen were 50, 65, and 80% for the Washington aerosols, and 50% for the Chur aerosols. Carrier gas flow rates (N2) between 8 and 40 mL/min (linear velocity 20-100 cm/min) 10.1021/es0503837 CCC: $30.25

 2005 American Chemical Society Published on Web 07/21/2005

TABLE 1. Combinations of Diffusion Coefficients and Diffusion Lengths Resulting in Various Rate Constants k Used as Input Parameter for AQUASIM 2.0 Transport Calculationsa diffusion coefficient (cm2s-1) diffusion length (µm) rate constant k (s-1) retention time (s) sorption coefficient (cm3/g)

10-9 4 0.02 300 503

3 × 10-9 4 0.05 303 509

10-8 4 0.2 300 504

10-8 2 0.6 298 500

6 × 10-8 4 1.0 297 499

10-6 4 15 297 498

a Resulting apparent sorption coefficients are given. These can be compared with the true equilibrium sorption coefficient K) 500 (cm3/g) that was used in the calculations.

were used for the Washington aerosols (pressure drop along the column 50-230 Pa). For the Chur aerosols gas flow rates were 4-15 mL/min (linear velocity 10-40 cm/min, pressure drop along the column 60-270 Pa). Retention volumes were measured for 60-63 (on Washington urban aerosols at varying RH) and 70 (on urban aerosols from Chur) organic compounds (used as received, purity at least 96%, purchased from Fluka and Merck). For a list of the compounds see Table S1 in the Supporting Information (SI). The specific surface areas (SSAaerosol) of the aerosols were determined by the Brunauer-Emmet-Teller nitrogen adsorption method and were 2.2 m2/g (Washington) and 0.2 m2/g (Chur), respectively. The value for the Washington aerosols agrees well with literature values of the same sample (11). A comment on the value of SSAChur aerosol can be found in the SI. EC- and OC-contents were determined by Sunset Laboratories (Tigard, OR) using a thermal-optical method (NIOSH 5040, 5% accuracy). The Washington urban aerosols contained 2.2 ( 0.1% EC and 15.2 ( 0.8% OC, and the Chur urban aerosols contained 2.1 ( 0.1% EC and 16.0 ( 0.8% OC (w/w). EC denotes here the elemental or black carbon with respect to the nonorganic, noncarbonate, solid fraction of carbon.

Results and Discussion Evaluation of the Experimental System. The net retention volume Vi net on both aerosol samples showed repeatability within 10%. For details on peak asymmetries we refer to the SI. The evaluation of sorption constants from measured retention volumes only works for equilibrium sorption and a linear sorption isotherm. Therefore, these conditions had to be checked as is demonstrated in the following paragraphs. Kinetic Limitation. A compound has to diffuse into a bulk phase of the sorbent to access sorption sites that have no direct contact with the carrier gas. The time available for these diffusion processes depends on the carrier gas velocity and has to be long enough to allow diffusion into the whole bulk phase. Varying the flow over the range of 8-40 mL/min on the Washington aerosols for a range of nonpolar and polar compounds did not show any significant effect on the observed retention, indicating no kinetic limitation. For the Chur urban aerosols, some change could be observed by varying the flow velocity (4-15 mL/min): out of 23 (11 nonpolar, 12 polar) compounds tested, only 3 showed exactly the same retention volume. For the other compounds, a smaller retention volume was measured with a higher carrier gas flow. This resulted in a change of 25 ( 10% in Vi net when doubling the gas flow. This is only a minor effect which lies in the same range as the overall uncertainty of Vi net (10%). To reduce the influence of the slight nonequilibrium at higher flows, the lowest flow possible was used for each compound (for more details see Figure S1 in the SI). In addition, we performed calculations with the transport model AQUASIM 2.0 (12) based on a volume flow rate of 6

FIGURE 1. AQUASIM 2.0 results for calculation of diffusion for a hypothetical compound under the following conditions: column length 1 cm, cross sectional area 0.071 cm2, porosity 0.4, total aerosol density 1.4 g/cm3, injection volume 100 µL, sorption coefficient 500 cm3/g, carrier gas flow 6 mL/min, diffusion length 4 µm, diffusion coefficients 10-9 to 10-6 cm2s-1 (see Table 1). mL/min to simulate a kinetically limited sorption process and the resulting peak. The diffusion length was assumed to be 4 µm, which corresponds to a particle diameter of 8 µm as a conservative assumption of maximum urban particle size. Diffusion coefficients were varied from 10-6 cm2s-1, corresponding to diffusion in an organic liquid, to 10-9 cm2s-1, corresponding to the low end of diffusion in rubbery polymers. The rate constants k resulting from different combinations of diffusion lengths and diffusion coefficients, and the resulting retention times and sorption coefficients are shown in Table 1. Table 1 and Figure 1 show clearly that, with these conservative assumptions for our experimental conditions, the first moment of the peak and therefore Vi net do not change substantially. These calculations confirm that our experimental data are at least a good approximation of the real equilibrium data for absorption into a liquid or rubbery OM. Sorption into OM that resembles a glassy polymer where diffusion coefficients are so small (in the order of 10-16 cm2s-1) that equilibrium would neither be achieved in our experiments nor in the atmosphere, is discussed later (see Availability of Aerosol Organic Matter as Sorbent). Linear Range of the Sorption Isotherm. If sorption occurs in the nonlinear range of the sorption isotherm, then the retention volume of a given sorbate varies with varying vapor concentrations. Here, concentrations for 4 nonpolar and 5 polar compounds were varied up to an order of magnitude at 65% RH without any observable effect on the measured retention volumes on the Washington urban aerosols. On the Chur urban aerosols (50% RH), concentrations of 3 nonpolar and 3 polar compounds were varied up to an order of magnitude with no significant effect on the retention volume. The results shown here therefore correspond to the linear range of the isotherm and can be extrapolated to other concentration ranges. VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6639

TABLE 2. Potential Contribution (in %) of Various Sorption Processes to the Total Measured Sorption of Selected Compounds (at 50% RH)a max. absorption into aerosol-OM simulated by organic solvents in % compound

hexadecane

octanol

n-octane 1,1,1-trichloroethane n-propylbenzene di-n-butyl ether 3-methylbutan-2-one butan-1-ol

1700 370 930 1200 260 33

410 320 770 1000 1000 2300

n-octane 1,1,1-trichloroethane n-propylbenzene di-n-butyl ether 3-methylbutan-2-one butan-1-ol

2200 550 1200 1300 450 54

530 470 1000 1100 1700 3900

max. adsorption on inorganic material simulated by typical airborne materials in %

methanol

(NH4)2SO4

Washington 510 9 650 6 1000 14 2900 220 4200 120 6700 280

NaCl

quartz

max. adsorption possible on ECb

7 6 12 320 210 380

12 11 37 2300 1500 2100

32 11 44 310 83 88

1 1 1 30 31 54

1 1 4 210 220 300

37 15 53 310 130 130

Chur 660 960 1400 3200 7200 11400

1 1 2 20 18 40

a Contribution of OM was simulated using three solvent/air partitioning coefficients for OM/air partitioning: hexadecane, octanol, and methanol. [Ki hexadecane/air from refs 16, 17; Ki methanol/air from ref 18; Ki octanol/air from Ki octanol/water (19, 20) and Ki water/air (21); Ki octanol/air and Ki methanol/air extrapolated to 15 °C with ∆Hi vap (22); Ki hexadecane/air extrapolated to 15 °C with ∆Hi hexadecane/air (23).] The values refer to Ki solvent/air normalized to the OC content (g) of the solvent, divided by Ki aerosol/air normalized to the OC content (g) of the aerosols. Contribution of surfaces was simulated by typical inorganic airborne materials for the non-carbonaceous aerosol fraction (i.e., not identified as EC or OC) [Contribution of adsorption to other surfaces was calculated with the measured SSAaerosol (2.2 m2/g for Washington, 0.2 m2/g for Chur), the mass of non-carbonaceous material (83% for Washington aerosol, 82% for Chur aerosol) and surface parameters for 40% RH (NaCl and (NH4)2SO4) and 45% RH (quartz), taken from the literature (9, 24). At 50% RH, adsorption to these materials would be even less.]. The values refer to Ki surface/air divided by Ki aerosol/air, both normalized to the total surface area (m2) of the aerosols. b Contribution of adsorption to EC was calculated with the EC content (2.2% for Washington aerosol, 2.1% for Chur aerosol), an assumed SSAEC (50 m2/g), and the adsorption model derived from diesel soot adsorption reported in the companion paper (3).

Sorption Coefficients. The sorption coefficients determined for the two urban aerosols and the corresponding errors for 60-70 compounds covering a wide range of physicochemical properties are listed in Table S1 (SI). Absorption coefficients range over more than 3 orders of magnitude in log Ki urban aerosol/air for all data sets. The details on the error calculation can be found in SI. Prevailing Sorption Mechanisms Based on EC and OC Contents. Our urban aerosols have almost identical EC and OC contents and show very similar sorption coefficients (compare values in Table S1 and Figure S2 in SI). However, their SSAaerosol is a factor of 10 different. This can be taken as a first indication that absorption (related to OC mass) rather than adsorption (related to the surface area) is the dominant process. With the known EC content (fEC) of the samples and the equation for adsorption to EC reported in the companion paper (eq 5 in ref 3), the maximum contribution of EC adsorption can be estimated, assuming that the EC in the urban aerosols has an average SSAEC (50 m2/g, (13-15)) and the same surface properties as the diesel soot investigated in the companion study. Even under these maximum assumptions, adsorption to EC could fully explain the observed sorption only for 3 (Washington) and 16 (Chur) polar compounds (selected compounds in Table 2). For 17 compounds on Washington aerosols (8 nonpolar, 9 polar) and 31 compounds on Chur aerosols (9 nonpolar, 22 polar) 50% of observed sorption would be explained by maximal EC adsorption. As was the case for the road tunnel aerosols reported in the companion paper, maximal EC adsorption also overestimates the sorption measured for some compounds. This again indicates that only part of the EC surface was available due to internal mixing and that the estimated adsorption to EC would actually be lower than that estimated in Table 2 for all compounds. Internal mixing may lead to such a loss of sorption capacity compared to the single aerosol components. Hence, adsorption to EC might have been important for some compounds such as the polar ones but cannot have been dominating the total sorption capacity of the aerosols. 6640

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005

Furthermore, adsorption to other typical surfaces such as (NH4)2SO4, NaCl, or quartz possibly present in the noncarbonaceous material could not explain the sorption measured for nonpolar compounds, but might have been significant for polar compounds on both urban aerosols (Table 2). The OM content of both urban aerosols though is high enough to explain the observed sorption of all compounds as absorption into an organic bulk phase. Taking the OM content and solvent/air partitioning coefficients Ki solvent/air of solvents of various polarities, we can estimate the sorption that can possibly occur in that OM (Table 2). For both urban aerosols, the measured sorption could be explained for all compounds by absorption in such a hypothetical OM phase, with the exception of alcohols in a nonpolar OM substitute like hexadecane (Table 2). In fact, the observed sorption is substantially overestimated for most compounds when simulating the absorption process with organic solvents as OM substitutes. These results suggest that absorption in OM was the dominant sorption mechanism but at the same time not all of the OM was available for sorption (see Availability of Aerosol Organic Matter as Sorbent). Adsorption versus Absorption Based on Intermolecular Interaction Models. Considerations in the previous section do not give unambiguous proof as to which sorption mechanism dominates on urban aerosols. In the Appendix of the companion paper (3), we described a method based on sorption models to determine whether an adsorption or an absorption process dominates sorption of a set of organic vapors to a specific sample. Applying this method to the data sets of the Chur and the Washington urban aerosols, we find that polar compounds on both aerosols and nonpolar compounds on Chur urban aerosols are ruled by absorption into the aerosol OM (see SI for details). Nonpolar compounds on the Washington urban aerosols could be governed by an adsorptive process. However, our considerations based on the EC and OM content have already shown that adsorption to a surface cannot explain the measured sorption for nonpolar compounds (see Prevailing Sorption Mechanisms Based on EC and OC Contents).

TABLE 3. LFER Results (Eq 2) for Chur Urban Aerosols and Washington Urban Aerosolsa Chur aerosols 15 °C

R2 nb abulk bbulk cbulk dbulk constbulk a

Washington aerosols 15 °C

50% RH

50% RH

65% RH

80% RH

0.96 66 1.20 ( 0.04 0.77 ( 0.11 2.36 ( 0.19 -1.14 ( 0.13 -6.26 ( 0.11

0.93 57 1.14 ( 0.05 1.25 ( 0.17 2.64 ( 0.25 -1.05 ( 0.16 -6.08 ( 0.16

0.94 54 1.14 ( 0.05 1.44 ( 0.16 2.64 ( 0.23 -0.85 ( 0.16 -6.34 ( 0.15

0.92 57 1.16 ( 0.05 1.67 ( 0.17 2.72 ( 0.25 -1.32 ( 0.17 -5.89 ( 0.17

All model results are from fits of data sets of experimental log Ki aerosol/air, normalized to total OC content (m3/g OC).

Considering the results we reported in our companion paper (3), a clear trend appears to exist from domination of adsorption on EC for the exhaust particles (diesel soot), over a mixed sorption mechanism on road tunnel aerosols, to domination of absorption into OM for urban aerosols. Another important result is that the fitted constants from the data sets of the urban aerosols are still significantly higher than those constants that are expected for absorption into the OM content of the aerosols (see SI and Appendix in ref 3). This is in agreement with our above conclusion that not all aerosol OM was available for absorption because only the available OM should be used for the calculations. In fact, the discrepancy up to 0.7 log units (see SI) suggests that only 20% of the aerosol OM present was accessible for absorption. Nevertheless, absorption into OM was the dominating sorption mechanism. Hence, an absorption model has to be applied to describe the experimental data for both urban aerosols. Absorption Model. In analogy to the adsorption model reported in ref 8, the absorption model is based on van der Waals interactions (vdW) and electron-donor/acceptor interactions (ED/EA, i.e., hydrogen bonds) between the organic vapors and the bulk phase, but it additionally includes a cavity formation term (7):

log Ki bulk/air ) abulk‚log Ki hexadecane/air + bbulk‚Σβi + cbulk‚ΣRi + dbulk‚Vi + constbulk (1) where log Ki hexadecane/air is the hexadecane/air partition coefficient at 25 °C (see ref 8 for the temperature of the parameter) (16, 17), which is taken as a measure for the vdW property of i, ΣRi is the electron-acceptor (H-donor) and Σβi is the electron-donor (H-acceptor) of the compound i, (20, 21), Vi is the molar volume of the compound i (in 10-4 cm3/ mol). abulk (∼vdWbulk), bbulk (∼EAbulk), cbulk (∼EDbulk), and dbulk (∼cavity) are temperature-dependent descriptors of the bulk sorption properties of the sorbents. constbulk depends on the dimensions of log Ki bulk/air. The sorbent descriptors and the constant can be derived by a multiple linear regression of the experimental data set of log Ki bulk/air and the corresponding compound properties. Absorption model results are listed in Table 3. The absorption model describes the compound variability in the measured sorption data sets of these two aerosols well (R2 ) 0.92-0.96, Figure 2 for the Chur aerosols). Comparison of the two urban aerosols at 50% RH reveals that they exhibit quite similar absorption properties, which is also evident from the direct comparison of their sorption coefficients (Figure S2 in SI). This similarity is surprising because they were obtained with different sampling methods, with a 25 year difference in age and in two different cities. Also considering the large variability of sorption properties of aerosols that has been found in other work (e.g., ref 4) this similarity is unexpected. Note that all results in Tables 3 and S1 as well as the data shown in Figure 2 are normalized to the total measured OC content (g) because

b

Used in LFER.

FIGURE 2. Comparison of calculated values (eq 2, with descriptors listed in Table 3) with experimental values for Chur aerosols, 15 °C, 50% RH (R2 ) 0.96, n ) 66). of uncertainties in the conversion factor from OC to OM content (1, 25, 26). Availability of Aerosol Organic Matter as Sorbent. Several findings in our data indicate that not all aerosol OM was available for absorption in our experiments. In the section Kinetic Limitation we have discussed that this effect may not be due to kinetically hindered sorption into a liquid or rubbery organic phase with diffusion coefficients of 10-8 to 10-6 cm2s-1. However, the existence of part of the OM phase in the form of a glassy polymer with much lower diffusion coefficients would explain the observed effect. Diffusion times in such a glassy polymer matrix would be so long (months) that sorption into this matrix would also not be relevant during the average atmospheric lifetime of an aerosol. One might also hypothesize that the equilibrium sorption capacity of the organic aerosol phase simply was much smaller than any of the three OM substitutes that we used (Table 2). However, the following considerations demonstrate that no liquid organic phase exists that could exhibit the sorption behavior found in our aerosols experiments. Plotting the experimental data normalized to the volume of OM versus log Ki octanol/air (see Figure 3 for the Chur sample) shows that correlation lines for all compound classes are well below the 1:1 line, while most of them lie close to each other and exhibit slopes close to unity. The latter two observations strongly indicate similar sorption properties of the aerosol OM and octanol. However, this would unambiguously imply that the data would actually lie on the 1:1 line. By plotting various solvent/air data from literature versus log Ki octanol/air we do not find this combined effect (all compound classes close to each other with unity slopes, but well below the 1:1 line) for VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6641

FIGURE 3. Measured absorption coefficients on urban aerosols from Chur (15 °C, 50% RH, per volume OM) vs octanol/air absorption coefficients (from Ki octanol/water and Ki water/air (19-21), extrapolated to 15 °C with ∆Hi vap (22)). The 1:1 line is included in the graph. Conversion factors for aerosol OM: OM ) 1.4‚OC; GOM ) 0.9 g/cm3, estimate of organic solvents. Results in this graph do not change significantly by varying these assumptions by 1.2-1.8 and 0.9-1.2, respectively. any of the well-defined organic sorbent phases that we checked (see Figures S3-S6 in SI). Figures S3-S6 show that the more polar the solvent (e.g., dicarboxylic acids), the smaller is sorption into that solvent, the flatter are the slopes within a compound class, and the further apart lie the compound classes. The combination of features found in our work therefore can be explained only if a significant portion of the aerosol OM was not accessible to the sorbing compounds. This effect was also observed for the road tunnel aerosols examined, but to a smaller extent. One possible hypothesis could be the existence of glassy polymers as part of the OM (27, 28). Diffusion into such glassy polymers is so slow that the equilibration time would be on the order of months (29, 30). Note that even in the presence of glassy polymeric structures that decrease the overall sorption, the observed dominant sorption mechanism still remains absorption into OM, but into rubbery polymeric or liquid OM. The effect of humidity for polar compounds on the Washington aerosols could also indicate the presence of glassy polymers. At 80% RH, Vi net was higher than that at 50% RH by one-third on average, while results at 50 and 65% RH agreed well. An influence of the deliquescence point of ionic components is unlikely, as the effect was fully reversible under our experimental conditions. At higher RH, H2O could partly dissolve into glassy polymers as part of the aerosol OM. There it acts as a plasticizer turning the rigid structure of the glassy polymers into rubbery, more permeable polymers that are better accessible for organic vapors. This OM could then exhibit a higher sorption capacity (e.g., ref 31). The fact that up to now only a fraction (5-20%, e.g., 3234) of the OM of aerosols could be identified with GC-MS can be regarded as an indication of the existence of polymeric structures. Recently, polymers have actually been identified in organic aerosols (35). Since the unidentified fraction seems to be rather polar (36, 37), it is plausible that these polymers occur in a rigid (glassy) state rather than in a rubbery state. There seems to be a trend in the degree of availability of OM as well, from the road tunnel aerosols reported in the 6642

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005

companion paper, where more OM was available, to the urban aerosols, with less OM available. Further evidence for glassy polymer structures based on differential scanning calorimetry (DSC) analysis of our aerosol samples will be reported elsewhere by Gabi Schaumann (TU Berlin). It is important to note that the polymerization process itself may be time dependent and that glassy polymer structures in aerosols may only develop after weeks or months. If this was the case then our “old” aerosol samples would not have been representative for fresh aerosols that are found in the atmosphere. We therefore plan to continue our sorption studies with freshly collected aerosol samples. In fate modeling and for description of field data it is usually assumed that sorption properties and capacities for aerosol OM and octanol are identical for nonpolar and weakly monopolar compounds (5, 38-41). This assumption corresponds to the 1:1 line in Figure 3 and to the following equation (eq 10 in ref 5): log Ki particle/air (m3/µg) ) log Ki octanol/air (m3/ m3) + log ×a6OM - 11.91. Obviously this approach overestimates the sorption measured here by an order of magnitude. For selected compounds, predictions with our Chur urban aerosol model lie in the same range as the field data. However, the aerosol properties found in the field can vary significantly, as indicated by the variance of sorption data reported in the literature (e.g., 5, 6, 42). For a better understanding of sorption properties of aerosols, a broader variability of aerosols will have to be measured with organic compounds from a wide range of compound classes. Applying the adsorption and absorption models presented in this work and the companion paper (3) will help to elucidate the differences in the sorption properties of various aerosols.

Acknowledgments We thank Susan Schlatter of the Environment and Health Protection of the City of Zurich (Umwelt- und Gesundheitsschutz Zu ¨ rich) for providing the NIST 1649a standard reference material; Mr. Ja¨ger of Novintec AG (Chur) for providing the bag filter of the Chur sample; Mr. Eggstein of Unifil AG (Niederlenz, CH) for providing information on bag filters and filter users; Hermann Mo¨nch and Johanna Buschmann for help with the BET measurements; Urs Baltensperger (PSI Villigen), Craig Corrigan (University of California, San Diego), and Markus Kalberer (ETH Zu ¨ rich) for helpful discussions; and Hans Peter Arp, Kathrin Fenner, and Zach Schreiber for critical comments on an earlier version of the manuscript.

Supporting Information Available Additional experimental details and data. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change; John Wiley & Sons: New York, 1998. (2) Goss, K.-U.; Schwarzenbach, R. P. Quantification of the Effect of Humidity on the Gas/Mineral Oxide and Gas/Salt Adsorption of Organic Compounds. Environ. Sci. Technol. 1999, 33, 40734078. (3) Roth, C. M.; Goss, K.-U.; Schwarzenbach, R. P. Sorption of a diverse set of organic vapors to diesel soot and road tunnel aerosols. Environ. Sci. Technol. 2005, 39, 6632-6637. (4) Cotham, W. E.; Bidleman, T. F. Polycyclic Aromatic Hydrocarbons and Polychlorinated Biphenyls in Air at an Urban and a Rural Site near Lake Michigan. Environ. Sci. Technol. 1995, 29, 2782-2789. (5) Harner, T.; Bidleman, T. F. Octanol-Air Partition Coefficient for Describing Particle/Gas Partitoning of Aromatic Compounds in Urban Air. Environ. Sci. Technol. 1998, 32, 1494-1502. (6) Mader, B. T.; Pankow, J. F. Study of the Effects of Particle-Phase Carbon on the Gas/Particle Partitioning of Semivolatile Organic

(7) (8) (9) (10)

(11) (12) (13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22) (23) (24)

(25)

Compounds in the Atmosphere Using Controlled Field Experiments. Environ. Sci. Technol. 2002, 36, 5218-5228. Goss, K.-U.; Schwarzenbach, R. P. Linear Free Energy Relationships Used To Evaluate Equilibrium Partitioning of Organic Compounds. Environ. Sci. Technol. 2001, 35, 1-9. Roth, C. M.; Goss, K.-U.; Schwarzenbach, R. P. Adsorption of a Diverse Set of Organic Vapors on the Bulk Water Surface. J. Colloid Interface Sci. 2002, 252, 21-30. Goss, K.-U.; Schwarzenbach, R. P. Adsorption of a diverse set of organic vapors on Quartz, CaCO3 and alpha-Al2O3 at different relative humidities. J. Colloid Interface Sci. 2002, 252, 31-41. Poster, D. L.; Schantz, M. M.; Wise, S. A.; Vangel, M. G. Analysis of urban particulate standard reference materials for the determination of chlorinated organic contaminants and additional chemical and physical properties. Fresenius’ J. Anal. Chem. 1999, 363, 380-390. Sheffield, A. E.; Pankow, J. F. Specific Surface Area of Urban Atmospheric Particulate Matter in Portland, Oregon. Environ. Sci. Technol. 1994, 28, 1759-1766. Reichert, P. AQUASIM 2.0 - User manual; Swiss Federal Institute for Environmental Science and Technology (EAWAG), CH-8600 Dubendorf, Switzerland, 1998. Jonker, M. T. O.; Koelmans, A. A. Sorption of Polycyclic Aromatic Hydrocarbons and Polychlorinated Biphenyls to Soot and Sootlike Materials in the Aqueous Environment: Mechanistic Considerations. Environ. Sci. Technol. 2002, 36, 3725-3734. Bucheli, T. D.; Gustafsson, O. Quantification of the soot-water distribution coefficient of PAHs provides mechanistic basis for enhanced sorption observations. Environ. Sci. Technol. 2000, 34, 5144-5151. Goss, K.-U.; Eisenreich, S. J. Sorption of Volatile Organic Compounds to Particles from a Combustion Source at different Temperatures and Relative Humidities. Atmos. Environ. 1997, 31, 2827-2834. Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. Hydrogen bonding XVI. A new solute solvation parameter, πH2, from gas chromatographic data. J. Chromatogr. 1991, 587, 213228. Abraham, M. H. Hydrogen bonding XXVII. Solvation parameters for functionally substituted aromatic compounds and heterocyclic compounds, from gas-liquid chromatographic data. J. Chromatogr. 1993, 644, 95-139. Abraham, M. H.; Whiting, G. S.; Carr, P. W.; Ouyang, H. Hydrogen Bonding. Part 45. The Solubility of Gases and Vapours in Methanol at 298 K: an LFER Analysis. J. Chem. Soc., Perkin Trans. 2 1998, 1385-1390. Kamlet, M. J.; Doherty, R. M.; Abraham, M. H.; Marcus, V.; Taft, R. W. Linear Solvation Energy Relationships. 46. An Improved Equation for Correlation and Prediction of Octanol/Water Partition Coefficients of Organic Nonelectrolytes Including Strong Hydrogen Bond Donor Solutes. J. Phys. Chem. 1988, 92, 5244-5255. Abraham, M. H.; Chadha, H. S.; Whiting, G. S.; Mitchell, R. C. Hydrogen Bonding. 32. An Analysis of Water-Octanol and WaterAlkane Partitioning and the ∆log P Parameter of Seiler. J. Pharm. Sci. 1994, 83, 1085-1100. Abraham, M. H.; Andonian-Haftvan, J.; Whiting, G. S.; Leo, A.; Taft, R. S. Hydrogen Bonding. Part 34. The Factors that Influence the Solubility of Gases and Vapours in Water at 298 K, and a New Method for its Determination. J. Chem. Soc., Perkin Trans. 1994, 2, 1777-1791. Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 76th ed.; CRC Press: Boca Raton, FL, 1995. Abraham, M. H.; Whiting, G. S.; Fuchs, R.; Chambers, E. J. Thermodynamics of Solute Transfer from Water to Hexadecane. J. Chem. Soc., Perkin Trans. 2 1990, 291-300. Goss, K.-U.; Buschmann, J.; Schwarzenbach, R. P. Determination of the surface sorption properties of talc, different salts and clay minerals at various humidities using adsorption data of a diverse set of organic vapors. Environ. Toxicol. Chem. 2003, 22, 26672672. Russell, L. M. Aerosol organic-mass-to-organic-carbon ratio measurements. Environ. Sci. Technol. 2003, 37, 2982-2987.

(26) Turpin, B. J.; Lim, H.-J. Species Contributions to PM2.5 Mass Concentrations: Revisiting Common Assumptions for Estimating Organic Mass. Aerosol Sci. Technol. 2001, 35, 602-610. (27) Mark, J. E.; Eisenberg, A.; Graessley, W. W.; Mandelkern, L.; Samulski, E. T.; Koenig, J. L.; Wignall, G. D. Physical Properties of Polymers, 2nd ed.; American Chemical Society: Washington, DC, 1993. (28) Godovsky, Y. K. Thermophysical Properties of Polymers; SpringerVerlag: Berlin, 1992. (29) Xing, B.; Pignatello, J. J. Time-Dependent Isotherm Shape of Organic Compounds in Soil Organic Matter: Implication for Sorption Mechanism. Environ. Toxicol. Chem. 1996, 15, 12821288. (30) Xing, B.; Pignatello, J. J. Dual-Mode Sorption of Low Polarity Compounds in Glassy Poly(Vinyl Chloride) and Soil Organic Matter. Environ. Sci. Technol. 1997, 31, 792-799. (31) Surana, R.; Randall, L.; Pyne, A.; Vemuri, N. M.; Suryanarayanan, R. Determination of glass transition temperature and in situ study of the plasticizing effect of water by inverse gas chromatography. Pharm. Res. 2003, 20, 1647-1654. (32) Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. Quantification of Urban Organic Aerosols at a Molecular-LevelsIdentification, Abundance and SeasonalVariation. Atmos. Environ. Part A 1993, 27, 1309-1330. (33) Saxena, P.; Hildemann, L. M. Water-soluble organics in atmospheric particles: A critical review of the literature and application of thermodynamics to identify candidate compounds. J. Atmos. Chem. 1996, 24, 57-109. (34) Puxbaum, H.; Rendl, J.; Allabashi, R.; Otter, L.; Scholes, M. C. Mass balance of the atmospheric aerosol in a South African subtropical savanna (Nylsvley, May 1997). J. Geophys. Res., [Atmos.] 2000, 105, 20697-20706. (35) Kalberer, M.; Paulsen, D.; Sax, M.; Steinbacher, M.; Dommen, J.; Prevot, A. S. H.; Fisseha, R.; Frankevich, V.; Zenobi, R.; Baltensperger, U. First Identification of Polymers as Major Components of Atmospheric Organic Aerosols. Science 2004, 303, 1659-1662. (36) Zappoli, S.; Andracchio, A.; Fuzzi, S.; Facchini, M. C.; Gelencser, A.; Kiss, G.; Krivacsy, Z.; Molnar, A.; Meszaros, E.; Hansson, H. C.; Rosman, K.; Zebuhr, Y. Inorganic, organic and macromolecular components of fine aerosol in different areas of Europe in relation to their water solubility. Atmos. Environ. 1999, 33, 2733-2743. (37) Facchini, M. C.; Fuzzi, S.; Zappoli, S.; Andracchio, A.; Gelencser, A.; Kiss, G.; Krivacsy, Z.; Meszaros, E.; Hansson, H. C.; Alsberg, T.; Zebuhr, Y. Partitioning of the organic aerosol component between fog droplets and interstitial air. J. Geophys. Res., [Atmos.] 1999, 104, 26821-26832. (38) Wania, F.; Daly, G. L. Estimating the contribution of degradation in air and deposition to the deep sea to the global loss of PCBs. Atmos. Environ. 2002, 36, 5581-5593. (39) Lohmann, R.; Harner, T.; Thomas, G. O.; Jones, K. C. A Comparative Study of the Gas-Particle Partitioning of PCDD/ Fs, PCBs, and PAHs. Environ. Sci. Technol. 2000, 34, 49434951. (40) Shoeib, M.; Harner, T. Using Measured Octanol-Air Partition Coefficients to Explain Environmental Partitioning of Organochlorine Pesticides. Environ. Toxicol. Chem. 2002, 21, 984990. (41) Cousins, I. T.; Mackay, D. Gas-Particle Partitioning of Organic Compounds and Its Interpretation Using Relative Solubilities. Environ. Sci. Technol. 2001, 35, 643-647. (42) Leach, K.; Kamens, M. R.; Strommen, M. R.; Jang, M. Partitioning of Semivolatile Organic Compounds in the Presence of a Secondary Organic Aerosol in a Controlled Atmosphere. J. Atmos. Chem. 1999, 33, 241-264.

Received for review February 24, 2005. Revised manuscript received May 10, 2005. Accepted May 18, 2005. ES0503837

VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6643