Source Diagnostics of Polycyclic Aromatic Hydrocarbons Based on

Oct 29, 2005 - transport of the PAH compounds in a multimedia environment. To examine such ... applied to account for the ratio changes in a multimedi...
0 downloads 0 Views 153KB Size
Download PDF
Environ. Sci. Technol. 2005, 39, 9109-9114

Source Diagnostics of Polycyclic Aromatic Hydrocarbons Based on Species Ratios: A Multimedia Approach X. L. ZHANG, S. TAO,* W. X. LIU, Y. YANG, Q. ZUO, AND S. Z. LIU Laboratory for Earth Surface Processes, College of Environmental Sciences, Peking University, Beijing 100871, China

Often, the sources of polycyclic aromatic hydrocarbons (PAHs) in environmental media can be identified by comparing the ratios of concentrations of selected pairs of PAH congeners in the source emissions to the ratios in the contaminated environmental media. However, these ratios can be altered significantly due to differences in the transport of the PAH compounds in a multimedia environment. To examine such changes, a fugacity model was applied to PAH ratios in a model environment. A linear relationship between the rate of emission and the bulk media concentration was identified for each PAH compound in an environmental medium at steady state and was quantified by a receptor-to-source ratio (RRS). It was demonstrated that the RRS values of the two congeners usually differ significantly. Consequentially, PAH ratios changed remarkably from the source emissions to various environmental media. A site-specific rectification factor (RF) was defined as the ratio of the two RRS values of the paired congeners for a specific PAH ratio in a given medium, which can be applied to account for the ratio changes in a multimedia environment. The PAH ratio changes were further verified with the surface soil data collected from Tianjin, China, and the observed changes of PAH ratios were compared favorably with the model predictions. The sensitivity analysis revealed that PAH ratios of the low molecular weight compounds were less stable. The most influential parameters controlling PAH ratios were those pertaining to dry precipitation, surface-to-air diffusion, degradation in air and water, and exchange between water and sediment.

Introduction Polycyclic aromatic hydrocarbons (PAHs) are a widespread class of toxic or carcinogenic pollutants, produced mainly through the incomplete combustion of both fossil fuels and biomasses (1, 2). Due to the wide variety of emission sources for PAHs and the relatively high levels of these contaminants in various environmental media, PAH contamination is of a particular concern for both environmental scientists and policy makers in China (3, 4). Being able to apportion PAHs to different sources is a critical step toward their risk assessment and management. Both source- and receptor-oriented approaches could be used to evaluate source contributions. The receptor-oriented * Corresponding author phone/fax: (86)-10-62751938; e-mail: [email protected]. 10.1021/es0513741 CCC: $30.25 Published on Web 10/29/2005

 2005 American Chemical Society

approach usually infers the contribution from various sources by determining the best-fit to a linear combination of equations for the emission sources needed to reconstruct the measured composition of a sample (5) or by using multivariate analysis (6). In this regard, the simplest profile is the ratio of two PAH congeners. Parent PAH ratios widely used as tracers of PAH emission sources include anthracene/ phenanthrene (ANT/PHE), fluoranthene/pyrene (FLA/PYR), benz[a]anthracene/chrysene (BaA/CHR), benzo[b]fluoranthene/benzo[k]fluoranthene (BbF/BkF), indeno(l,2,3-cd)pyrene/benzo(g,h,i)perylene (IcdP/BghiP), and benzo[a]pyrene/benzo(g,h,i)perylene (BaP/BghiP) (7-14). When PAH ratios are used to determine the source of an emission, it is assumed that the paired compounds are diluted to a similar extent during transport, and consequently, the ratios remain constant en route from sources to receptors. For this reason, ratio calculations are usually restricted to PAHs with a given molecular mass to minimize confounding factors, such as differences in volatility, water solubility, affinity to organic carbon, etc. (15-17). However, this assumption is not always valid because, more often than not, the physicochemical properties of the paired PAH species are not identical (18-22). Consequently, changes in diagnostic ratios from sources to receptors are almost unavoidable. For example, PAH ratios in the atmosphere often depart from those observed in source emissions (23, 24). Hwang et al. found that ANT degraded faster than PHE, leading to a reduced ANT/PHE ratio in pine needles (25). Bucheli et al. indicated that chemical transformation of PAHs to any significant extent in the environment may influence the diagnostic capabilities of PAH ratios (13). If PAH ratios are used to apportion contamination to sources, they should be applied with caution, checked for consistency, and corrected or rectified for the changes caused by transformation (13). Attempts have been made to account for the inconsistency in the PAH ratios due to the different rates of transport and transformation (23, 26-28). Venkataraman and Friedlander estimated relative decay factors for 10 PAH compounds for Los Angeles area (27). Schauer et al. proposed a coefficient of fractionation, representing losses due to gravitational settling, chemical transformation, or evaporation of PAHs, for source apportionment of airborne particles (23). These factors, however, were often experimentally derived and focused only on decay within one specific bulk medium. When the diagnostic ratios are to be used in media other than air, the differences in the transport rates of PAHs across the media interfaces as well as the differences in degradation rates in the target media must be considered within a multimedia framework. Among a variety of multimedia models available, the fugacity model has been applied to quantitatively depict the fate of persistent organic pollutants in the environment (29-31). The model can also be used to assess overall persistence and long-range transport potential of organic pollutants (32). The application of the fugacity model can help quantify the differences in transport of individual PAH compounds and improve our ability to identify sources of PAH contamination using PAH ratios. The hypotheses to be tested in this study were two-fold: (1) the pairs of PAH compounds used to calculate a diagnostic ratio should have different rates of transport and transformation in a multimedia environment causing a significant change in their diagnostic ratio; (2) the changes in the diagnostic ratio can be rectified with a multimedia model under the specific site conditions. To test the hypotheses, a fugacity model was applied to the change in the PAH ratio VOL. 39, NO. 23, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

9109

FIGURE 1. Concentrations of PHE and ANT in air particles, suspended solids, soil, and sediment as the function of their emission rates. The slopes of the plots were defined as receptor-to-source ratios (RRS). in a model environment. The bulk media studied include air, water, soil, and sediment. For air and water, only the solid fractions were considered. The major influences affecting the diagnostic ratio were evaluated, and the uncertainty of the model results was assessed. The model was verified by comparing model calculations with the actual changes in ratio of PAHs as contaminants moved from the sources of the emissions to the surface soil in Tianjin, China.

Methodology Modeling PAH Ratio Changes in a Multimedia Environment. Six PAH ratios were investigated that involved 11 PAH congeners (ANT, PHE, FLA, PYR, BaA, CHR, BbF, BkF, IcdP, BaP, IcdP and BghiP). The quantitative relationships between the emission and the concentrations in the bulk media were characterized for each of the 11 congeners individually using a level-III fugacity model based on the approach of Mackay and Paterson (30). An attempt was made to quantify possible changes in the PAH ratios from the sources to the bulk media including PAH ratios on particles in air, suspended solids, soil, and sediment. The numerical model simulation was performed using MATLAB v.6.5 (33). All site-specific parameters were appropriate to the conditions in Tianijn, China except that PAH concentrations in the advective air and water flows were set to null so that the source to be characterized was limited to a simulated emission to air. The details of the model framework and parameters are described in the Supporting Information and can also be found in a previous study (34). Field Verification of PAH Ratio Changes. In a previous study, PAHs emissions from major sources in Tianjin, China were compiled. These sources include the coking industry, domestic and industrial coal combustion, automobile emissions, and burning of biomass (35). The PAH ratios in the emissions from the major sources were available. Also, during an extensive survey in the same area, 188 surface soil samples were collected for PAH measurement in an approximately uniform grid with a size of 8 × 8 km2 (36). To verify the changes in PAH ratios from the emission sources to the surface soil, the observed PAH ratios in the soil were compared with those at the sources with or without a multimedia rectification. Multimedia fate modeling was conducted using the emission as the sole input for such a rectification. Identification of the Influential Parameters. There were more than 70 parameters and many transformation and transport processes in the multimedia model. The identification of the most influential parameters and processes responsible for changes in the PAH ratio will not only help understand the interactions of these parameters and processes but will also help simplify the assessment process. The most influential parameters were identified and evaluated with a sensitivity analysis. The calculated PAH ratios in various bulk media were tested against all input parameters except for the gas constant by running the model with each individual parameter multiplied by 19 factors ranging from 0.1 to 10.0. A coefficient of sensitivity (CS) was defined as the relative change of a model estimate, PAH concentration ratio 9110

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 23, 2005

in this case, versus parameter change (37). To further investigate the specific processes responsible for such a change, the sensitivity analysis was also conducted on a flux basis by calculating CS values in terms of PAH flux ratios. Coupled with the multimedia model framework, the most influential processes represented by the transfer rate coefficients (D values) were identified and used to interpret the changes in PAH concentration ratios in the multimedia environment. Model Uncertainty Analysis. Monte Carlo simulation was employed to propagate collective variance of the inputs through the model for assessment of the overall uncertainty in the predictions. A normal distribution was assumed for all input parameters except for the gas constant. The dispersion of each parameter was assigned based on the normal distribution (37). The simulation was repeated 10 000 times, with new values randomly selected for all parameters from their probability distributions using the “randn” builtin function in MATLAB (33). With the use of the output of the repeated runs, statistics of dispersion of the calculated PAH ratios were used for evaluating model uncertainty.

Results and Discussion Relationship between the Emission and the Bulk Media Concentrations. To carry out the fugacity modeling, a set of mass balance equations was solved to calculate fugacities, which were then used to calculate concentrations in the environmental media. It was assumed that the calculated fugacities and the media concentrations (mol/m3) that derived from the fugacities were proportional to the emission (mol/h) for a given PAH compound in a particular medium at steady state. Such linear relationships are illustrated using PHE and ANT as examples in Figure 1. The concentrations of PHE and ANT on particles in air, in suspended solids, and in soil and sediment were calculated using the multimedia model and plotted against their total emissions. It should be pointed out that the slopes are strong functions of several parameters, which vary from one region to another. The results shown in Figure 1 are specific to the case study of Tianjin. The linear relationship shown in Figure 1 can be simply described by the slope constant that was defined as a receptor-to-source ratio (RRS, h/m3), which represents the concentration increment (mol/m3) in a given medium per unit emission (mol/h) and can be used to describe quantitatively the relationship between the emission rate and the media concentration. In addition to the linearity between the emissions and bulk media concentrations of both ANT and PHE, the RRS values of the two compounds were significantly different in air, water and soil (Figure 1). The RRS values of ANT were about twice as high as those of PHE for air particles and soil but were lower than those of PHE in suspended solids and sediment. It was demonstrated using a fate model that the flux of PHE from air to soil was the predominate across-media transport process in Tianjin (34). Such a strong connection between soil and air could explain the similarity in the relationship between PHE-ANT in the

FIGURE 2. Log-scaled RRS of PHE, ANT, FLA, PYR, BaA, CHR, BbF, BkF, BaP, IcdP, and BghiP in air particles, suspended solids, soil, and sediment derived from a steady-state multimedia fate modeling.

TABLE 1. RF Values of the Six PAH Ratios in Air Particles, Suspended Solids, Soil, and Sediment Derived from the Multimedia Modeling bulk media

ANT/PHE

FLA/PYR

BaA/CHR

BbF/BkF

BaP/BghiP

IcdP/BghiP

∑(RFi - 1)2

air particles suspended solids soil sediment ∑(RFi - 1)2

2.16 0.83 1.77 0.97 1.97

1.68 2.26 1.54 2.00 3.34

0.59 0.60 0.46 0.47 0.90

0.81 0.75 0.41 0.37 0.84

0.60 0.57 0.31 0.28 1.34

0.92 0.89 0.72 0.69 0.19

2.18 2.04 2.08 2.29

two media. The differences between the absolute RRS values of the two species in air particles and soil can be explained by significant differences in Henry’s law constants and vapor pressure between PHE (KH ) 3.05 Pa‚m3/mol, Ps ) 1.98 × 10-2 Pa) and ANT (KH ) 19.3 Pa‚m3/mol, Ps ) 4.10 × 10-3 Pa) (18). The difference between the absolute RRS values in suspended solids compared to those in other media was also a result of the relatively higher vapor pressure of ANT compared to that of PHE. As will be demonstrated later, vapor pressure was one of the most influential parameters dictating the ANT/PHE ratio in suspended solids. The two lines for PHE and ANT in sediment almost overlap (Figure 1), indicating the similarity of the RRS values in sediment. Calculated RRS values for all 11 PAH compounds are presented in Figure 2. The RRS values for a given PAH compound varied over several orders of magnitude among the four bulk media, ranging from 10-9 (soil) to 10-5 (air particles) for PHE and ANT, and from 10-5 (suspended solids, soil, and sediment) to 10-2 (air particles) for IcdP and BghiP, respectively. There is a clear trend of increasing values of RRS from the lower molecular weight (LMW) species to the higher molecular weight (HMW) species in the same bulk media, suggesting a relatively higher accumulation of the HMW compounds. Also, the RRS values for the HMW PAH species in the different environmental media are more similar than those for the LMW species. Change in PAH Ratios from Source to Receptors. The RRS difference between two paired PAH congeners suggests possible difference in transport rates in the multimedia environment and potential change of their ratio from sources to receptors. When there is a significant change in the ratio, the basic assumption for source diagnostics based on PAH ratios is in question. If, however, such differences can be quantified, the ratio can be corrected or rectified, and source identification using PAH ratios is still valid. For this purpose, a rectification factor (RF) is defined as the ratio of two RRS values for a specific diagnostic ratio in a given medium, which is equivalent to the concentration ratio of the paired PAH congeners in that medium at equal emission rates. For example, a soil rectification factor of ANT/PHE describing the change of the ratio from sources to soil can be written as RFsoil-ANT/PHE, which equals to RRS-soil-ANT/RRS-soil-PHE. The value of an RF depends on the RRS values of the two PAH congeners in the ratio. A notable deviation of RF from unity indicates the significant differences in the transport rates of the two species. The RRS values of individual PAH compounds, and subsequently RFs of the PAH ratios, were functions of

a variety of parameters defining the multimedia environment in concern. For the model environment studied here, the calculated RFs for the six ratios in the four bulk media are listed in Table 1. The ∑(RFi - 1)2 values are also shown for all the ratios (the last row) as well as all the bulk media (the last column) for an overall evaluation of the ratio changes across the media and the ratios. Under the model environment conditions, the RF values ranged from 0.28 (RFsediment-BaP/BghiP) to 2.26 (RFsuspended solids-FLA/PYR) (Table 1), implying considerable difference in the transport rates of the paired PAH congeners. The RFFLA/PYR values in all media and the RFANT/PHE values in air and soil were greater than 1, while the RF values of the other four ratios were all less than 1. Among the six ratios studied, IcdP/BghiP is the best diagnostic ratio for source identification in various media with the RF values from 0.69 to 0.92 and mean ∑(RFi - 1)2 of 0.19. On the other hand, ∑(RFi - 1)2 of FLA/PYR was as high as 3.34 with the RF values from 1.54 to 2.26, suggesting that there are significant differences in the behaviors of FLA and PYR in the multimedia environment. On average, the ratio deviations in the four bulk media were similar with the ∑(RFi - 1)2 values from 2.04 to 2.29. The only source of PAHs in the model was emissions directly into the atmosphere. This does not necessarily mean the PAH ratio in air (particles) would be the same as the PAH ratio in the source. Instead, significant changes of PAH ratios in air did occur for ANT/PHE, BaA/CHR, BbF/BkF, and FLA/PYR, indicating that these diagnostic ratios should be used with caution for source identification, even when contaminants are sampled from particles in air. The explanation of such changes in air is particularly due to the differences in vapor/particulates partitioning between PAH compounds. In summary, PAH ratios may have undergone significant changes from sources to receptors, and therefore it is necessary to account for such changes when the ratios are used for source diagnostics. The exact changes, represented by the RF values, are obviously dependent on the situation in each case. PAH Ratios in Tianjin’s Soil. On the basis of the PAH contents in the surface soil (36) and the total annual emission rates (35) of Tianjin (a composite of the sources), PAH ratios in the soil were compared with the PAH ratios in a composite of all the source emissions in the Tianjin area. For each of the six pairs of PAH species, the mean and standard deviation values of the PAH ratio in soil were calculated from the field data. A multimedia simulation was also conducted to predict the soil ratios (Figure 3). There are significant differences in VOL. 39, NO. 23, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

9111

FIGURE 3. Comparison of PAH ratios between the emission source and the surface soil in Tianjin. Both multimedia fate modeling and the field survey were conducted to derive the ratios in the soil. Standard deviations of the measured ratios are also presented. all PAH ratios between the emission source and the soil measurements, confirming the differences in transport rates of the paired PAH congeners in the multimedia environment. Since the model environment for simulation was established based on the physical settings of Tianjin, the observed distribution patterns of PAH ratios in Tianjin’s surface soil were identical to those indicated by the RF values (Table 1). When the RF values are larger than 1 (ANT/PHE and FLA/PYR, predicted > source), PAH ratios in the soil are always larger than those at the emission source. For other ratios with the RF values less than 1 (predicted < source), soil ratios are smaller than those at the source. Although some of the predicted and measured PAH ratios were notably different (for example, the calculated and the measured BbF/BkF values were 1.80 and 1.09, and the calculated and the measured BaP/BghiP values were 0.40 and 1.06, respectively), the relative change of the PAH ratios between the source and the soil were well predicted by the multimedia approach (PAH ratios for ANT/PHE and FLA/PYR increase while the other ratios decrease). Influential Parameters Affecting the PAH Ratio Change. Although the significant changes of PAH ratios from sources to receptors were demonstrated in general, the extent of the changes for a given ratio in a particular medium was casedependent. For the six PAH ratios investigated, there was a general trend of decreasing number of influential parameters from LMW to HMW PAH species. In Table 2, the influential parameters on the ratio of ANT/PHE are tabulated for each separate environmental medium. Table 3 offers the results for IcdP/BghiP. As an example, the ratio of ANT/PHE in soil was sensitive to X13, F, Ps, and other parameters, while no sensitive

parameter could be identified for this ratio in air particles. Since the transport of PAHs in to and out of sediment had to go through a water phase, the influential parameters for water and sediment were identical and are listed together. Only those parameters with the absolute values of CS larger than 0.5 are listed. A complete list of the influential parameters for all ratios can be found in the Supporting Information. On the basis of the model framework adopted in the current study, the key processes closely related to some of the influential parameters are also listed in Tables 2 and 3. For instance, transfer rate coefficients of air-to-surface dry precipitation (D12p and D13p) are functions of volume fractions of air particles (X13) and dry deposition velocity (Kp). The soil-to-air diffusion rate coefficient (D31d) is a function of diffusion path lengths in soil (L3), molecular diffusivities in air (B1), and air molecular transfer coefficient over soil (k13). The effects of the processes listed were also confirmed by the results of flux sensitivity analysis. For instance, for ANT and PHE, their flux ratios of air-to-soil dry/wet precipitations and soil-to-air diffusion were also sensitive to the associated parameters. Although other rate coefficients can also be expressed as the functions of these influential parameters, the results of the flux sensitivity analysis did not show a significant effect of these parameters on the flux ratios of these rate coefficients. Therefore, they are not included in the list. Although the air-to-soil diffusion rate coefficient (D13d) is also a function of L3, B1, and k13, the ratio of air-to-soil diffusion fluxes (FluxANT/FluxPHE), for instance, was not sensitive to any parameters listed in Table 2. The ANT/PHE ratio and other LMW PAH ratios were under stronger influence of more parameters than the IcdP/BghiP ratio and other HMW PAH ratios. This implies that HMW PAH ratios are more stable diagnostic indicators than LMW ones in the multimedia environment. For ANT/PHE and other LMW PAH ratios, the higher sensitivity to influential parameters was primarily due to their higher mobility, especially their higher volatility. No influential parameter was identified for ANT/PHE in air particles. The high RF value (2.16) (Table 1) appeared to be relatively robust. In contrast, ANT/PHE ratios in all other bulk media were sensitive to a range of parameters. Among the parameters affecting ANT/PHE in water, soil, and sediment, the most influential were those closely related to air-to-surface transport of particles (X13, Kp, Kw, and Sc), surface-to-air transport of gas phase (KH, T, k12, k21, k13, L3, B1, and B2) and degradation in water (h2 and km2).

TABLE 2. Most Influential Parameters and Processes for RFANT/PHE media

influential parameter

CS

influential process

air particles

none

soil

X13 volume fraction of air solid F fugacity ratio Ps vapor pressure Kp dry deposition velocity L3 diffusion path length in soil B1 molecular diffusivity in air k13 air molecular transfer coefficient over soil Kw rain rate Sc rain-scavenging rate

3.75 3.75 -3.60 3.18 2.77 -2.21 -1.26 0.60 0.60

D13p air-to-soil dry precipitation D13w air-to-soil wet precipitation D31d soil-to-air diffusion

suspended solids and sediment

k12 air molecular transfer coefficient over water T temperature H Henry’s law constant X13 volume fraction of air solid F fugacity ratio k21 water molecular transfer coefficient over air Ps vapor pressure Kp dry deposition velocity h2 water depth km2 degradation rate in water

-6.01 3.53 -3.43 2.68 2.68 2.58 2.29 -2.62 0.63 0.63

D12p air-to-water dry precipitation D12w air-to-water wet precipitation D21d water-to-air diffusion

9112

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 23, 2005

none

D40m degradation in water

TABLE 3. Most Influential Parameters and Processes for RFIcdP/BghiP CS

influential process

air particles

media

h1 air thickness km1 degradation rate in air Kp dry deposition velocity

influential parameter

-1.07 -1.07 0.91

D10m degradation in air D12p air-to-water dry precipitation D13p air-to-soil dry precipitation

soil

h1 air thickness km1 degradation rate in air Kp dry deposition velocity

-1.07 -1.07 0.91

D13p air-to-soil dry precipitation

suspended solids and sediment

h1 air thickness km1 degradation rate in air Kp dry deposition velocity A2 interface area of air/water A3 interface area of air/soil

-1.07 -1.07 0.91 0.55 -0.55

D12p air-to-water dry precipitation D21d water-to-air diffusion D24s sedimentation from water D24d water-to-sediment diffusion D42d sediment-to-water diffusion

FIGURE 4. Comparison between the calculated (total) and the observed (shaded parts) standard deviations of the log-transformed PAH ratios. The observed data for the last two ratios in air are not available. For the IcdP/BghiP ratio and other HMW ratios, the influential parameters for the four bulk media were similar to one another. Three parameters, including dry deposition velocity (Kp), air thickness (h1), and degradation rate in air (km1), appeared to be equally important to all bulk media. Apparently, air-to-surface precipitation and degradation in air were the most important processes affecting the change on the ratio of IcdP/BghiP. In addition, processes of exchange between water column and bottom sediment also influenced this ratio in water and sediment. The PAH ratios were sensitive not only to the compoundspecific parameters such as degradation rates and Henry’s law constants, which directly affect the environmental behaviors of the paired congeners, but also to area-specific parameters, such as dry deposition velocity and water depth. It implies that the RF values calculated in this study (Table 1) should not be directly used in other cases without accounting for the case-specific conditions. Uncertainties of the Calculated PAH Ratios. The frequency distributions of individual PAH concentrations and PAH ratios were derived using a Monte Carlo simulation with 10 000 runs. All the concentrations and ratios followed a typical log-normal distribution. The same log-normal distribution of PAH ratios was also observed in the field survey, which consisted of 188 surface soil samples. The distribution of the ratios derived from the Monte Carlo simulation represents the combined effect of both uncertainty and variability. Uncertainty is that inherent in the parameters used in the simulation, whereas variability is mainly due to spatial and temporal variations. For uncertainty analysis, it is always desirable that the relative contributions of the variability and the true uncertainty of the model estimates be distinguished (38). In this study, the total variations were compared with the field-observed distribution. Since most area-specific parameters used for the simulation were similar to those in Tianjin, field-observed PAH concentrations in air particles, suspended solids, soil, and sediment from the area were used to generate the observed PAH ratios (4, 36, 39). Figure 4 is a stacked bar chart showing the standard deviations derived from the Monte Carlo simulation (total bar heights) and the field survey (bottom parts of the bars), based on the log-transformed

data. All six ratios in the four bulk media are presented. The two exceptions are BaP/BghiP and IcdP/BghiP in air, as no field data were available. The calculated standard deviations varied from 0.36 to 0.69, suggesting that the total uncertainty was about onethird to two-thirds of an order of magnitude, and the total uncertainties of the six ratios followed a similar pattern: air < water < soil < sediment. There was a general trend of decreasing uncertainties from LMW to HMW PAH compounds. The total heights and the shaded parts of the bars represent the calculated total uncertainties and the observed variabilities in log-scale, respectively. On average, the variabilities accounted for 41% of the total standard deviations, varying from 35% for sediment to 51% for water.

Acknowledgments The funding of this research was provided by National Basic Research Program (2003CB415004), the National Science Foundation of China (Grant 40332015/40021101), and the Ministry of Education. We are particularly grateful to Dr. John Wilson for reviewing the manuscript.

Supporting Information Available Detailed information for the multimedia fugacity modeling and emission estimation and the calculated RRS values for all PAHs investigated. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Mastral, A. M.; Callen, M. S. A review on polycyclic aromatic hydrocarbon (PAH) emission from energy generation. Environ. Sci. Technol. 2000, 34, 3051-3057. (2) Pozzoli, L.; Gilardoni, S.; Perrone, M. G.; De Gennaro, G.; De Rienzo, M.; Vione, D. Polycyclic aromatic hydrocarbons in the atmosphere: Monitoring, sources, sinks and fate. I: Monitoring and sources. Ann. Chim. 2004, 94, 17-32. (3) Duan, Y. H.; Tao, S.; Wang, X. J.; Li, B. G.; Xu, F. L.; Liu, W. X.; Cao, J. Source and spatial distribution of PAHs in Tianjin’s topsoil. Acta Pedol. Sinica., in press. (4) Shi, Z.; Tao, S.; Pan, B.; Fan, W.; He, X. C.; Zuo, Q.; Wu, S. P.; Li, B. G.; Cao, J.; Liu, W. X.; Xu, F. L.; Wang, X. J.; Shen, W. R.; Qing, B. P.; Sun, R. Contamination of polycyclic aromatic VOL. 39, NO. 23, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

9113

(5) (6)

(7) (8) (9)

(10)

(11)

(12)

(13) (14)

(15)

(16) (17)

(18) (19) (20)

(21) (22)

hydrocarbons in rivers in Tianjin, China. Environ. Pollut. 2005, 134, 97-111. Watson, J. G. Overview of receptor model principles. J. Air Pollut. Control Assoc. 1984, 34, 619-624. Simcik, M. F.; Eisenreich, S. J.; Lioy, P. J. Source apportionment and source/sink relationships of PAHs in the coastal atmosphere of Chicago and Lake Michigan. Atmos. Environ. 1999, 33, 50715079. Simoneit, B. R. Application of molecular marker analysis to vehicle exhaust for source reconciliations. Int. J. Environ. Anal. Chem. 1985, 22, 203-233. Lipiatou, E.; Saliot, A. Fluxes and transport of anthropogenic and natural polycyclic aromatic hydrocarbons in the western Mediterranean Sea. Mar. Chem. 1991, 32, 51-71. Benner, B. A.; Wise, S. A.; Currie, L. A.; Klouda, G. A.; Klinedinst, D. B.; Zweidinger, R. B.; Stevens, R. K.; Lewis, C. W. Distinguishing the contributions of residential wood combustion and mobile source emissions using relative concentrations of dimethylphenanthrene isomers. Environ. Sci. Technol. 1995, 29, 23822389. Yunker, M. B.; Macdonald, R. W.; Goyette, D.; Paton, D. W.; Fowler, B. R.; Sullivan, D.; Boyd, J. Natural and anthropogenic inputs of hydrocarbons to the Strait of Georgia. Sci. Total Environ. 1999, 225, 181-209. Budzinski, H.; Jones, I.; Bellocq, J.; Pie´rard, C.; Garrigues, P. Evaluation of sediment contamination by polycyclic aromatic hydrocarbons in the Gironde estuary. Mar. Chem. 1997, 58, 85-97. Dickhut, R. M.; Canuel, E. A.; Gustafson, K. E.; Liu, K.; Arzayus, K. M.; Walker, S. E.; Edgecombe, G.; Gaylor, M. O.; Macdonald, E. H. Automotive sources of carcinogenic polycyclic aromatic hydrocarbons associated with particulate matter in the Chesapeake Bay region. Environ. Sci. Technol. 2000, 34, 4635-4640. Bucheli, T. D.; Blum, F.; Desaules, A.; Gustafsson, O. Polycyclic aromatic hydrocarbons, black carbon, and molecular markers in soils of Switzerland. Chemosphere 2004, 56, 1061-1076. Castellano, A. V.; Cancio, J. L.; Aleman, P. S.; Rodriguez, J. S. Polycyclic aromatic hydrocarbons in ambient air particles in the city of Las Palmas de Gran Canaria. Environ. Int. 2003, 29, 475-480. Readman, J. W.; Mantoura, R. F.; Rhead, M. M. A record of polycyclic aromatic hydrocarbon (PAH) pollution obtained from accreting sediments of the Tamar estuary, UK: evidence for nonequilibrium behaviour of PAH. Sci. Total Environ. 1987, 66, 73-94. McVeety, B. D.; Hites, R. A. Atmospheric deposition of polycyclic aromatic hydrocarbons to water surfaces: a mass balance approach. Atmos. Environ. 1988, 22, 511-536. Yunker, M. B.; Macdonald, R. W.; Vingarzan, R.; Mitchell, R. H.; Goyette, D.; Sylvestre, S. PAHs in the Fraser river basin: a critical appraisal of PAH ratios as indicators of PAH source and composition. Org. Geochem. 2002, 33, 489-515. Mackay, D.; Shiu, W.; Ma, K. Illustrated Handbook of PhysicalChemical Properties and Environmental Fate for Organic Chemicals. Vol. II; Lewis Publishers: Boca Raton, FL, 1992. Masclet, P.; Mouvier, G.; Nikolaou, K. Relative decay index and sources of polycyclic aromatic hydrocarbons. Atmos. Environ. 1986, 20, 439-446. Kamens, R. M.; Guo, Z.; Fulcher, J. N.; Bell, D. A. Influence of humidity, sunlight and temperature on the daytime decay of polyaromatic hydrocarbons on atmospheric soot particles. Environ. Sci. Technol. 1988, 22, 103-108. Behymer, T. D.; Hites, R. A. Photolysis of polycyclic aromatic hydrocarbons adsorbed on fly ash. Environ. Sci. Technol. 1988, 22, 1311-1319. Li, A.; Jang, J. K.; Scheff, P. A. Application of EPA CMB8.2 model for source apportionment of sediment PAHs in Lake Calumet,

9114

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 23, 2005

Chicago. Environ. Sci. Technol. 2003, 37, 2958-2965. (23) Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R. Source apportionment of airborne particulate matter using organic compounds as tracers. Atmos. Environ. 1996, 30, 3838-3855. (24) Fraser, M. P.; Cass, G. R.; Simoneit, B. R.; Rasmussen, R. A. Air quality model evaluation data for organics C6-C22 nonpolar and semipolar aromatic compounds. Environ. Sci. Technol. 1998, 32, 1760-1770. (25) Hwang, H. M.; Wade, T. L.; Sericano, J. L. Concentrations and source characterization of polycyclic aromatic hydrocarbons in pine needles from Korea, Mexico, and United States. Atmos. Environ. 2003, 37, 2259-2267. (26) Kalberer, M.; Henne, S.; Prevot, A. S. H.; Steinbacher, M. Vertical transport and degradation of polycyclic aromatic hydrocarbons in an Alpine Valley. Atmos. Environ. 2005, 38, 6447-6456. (27) Venkataraman, C.; Friedlander, S. K. Source resolution of fine particulate PAHs using a receptor model modified for reactivity. J. Air Waste Manage. Assoc. 1994, 44, 1103-1108. (28) Su, M.; Christensen, E. R.; Karls, J. F.; Kosuru, S.; Imamoglu, I. Apportionment of polycyclic aromatic hydrocarbon sources in lower Fox River, USA, sediments by a chemical mass balance model. Environ. Toxicol. Chem. 2000, 19, 1481-1490. (29) Mackay, D. Finding fugacity feasible. Environ. Sci. Technol. 1979, 13, 1218-1223. (30) Mackay, D.; Paterson, S. Evaluating the multimedia fate of organic chemicals: A level III fugacity model. Environ. Sci. Technol. 1991, 25, 427-436. (31) Wania, F.; Mclachlan, M. S. Estimating the influence of forests on the overall fate of semivolatile organic compounds using a multimedia fate model. Environ. Sci. Technol. 2001, 35, 582590. (32) Fenner, K.; Scheringer, M.; MacLeod, M.; Matthies, M.; McKone, T.; Stroebe, M; Beyer, A.; Bonnell, M.; Le Gall, A. C.; Klasmeier, J.; Mackay, D.; Van de Meent, D.; Pennington, D.; Scharenberg, B.; Suzuki, N.; Wania, F. Comparing estimates of persistence and long-range transport potential among multimedia models. Environ. Sci. Technol. 2005, 39, 1932-1942. (33) MathWorks. Using MATLAB version 6; The MathWorks, Inc.: Natick, MA, 2002. (34) Tao, S.; Cao, H. Y.; Liu, W. X.; Li, B. G.; Cao, J.; Xu, F. L.; Wang, X. J.; Coveney, R., Jr.; Shen, W. R.; Qing, B. P.; Sun, R. Fate modeling of phenanthrene with regional variation in Tianjin, China. Environ. Sci. Technol. 2003, 37, 2453-2459. (35) Xu, S. S.; Liu, W. X.; Tao, S. Emission estimation of PAHs in China. J. AgrosEnviron. Sci. 2005, 24, 476-479. (36) Wang, X. J.; Zheng, Y.; Liu, R. M.; Li, B. G.; Cao, J.; Tao, S. Kriging and PAH pollution assessment in the topsoil of Tianjin area. Environ. Contam. Toxicol. 2003, 71, 189-195. (37) Cao, H. Y.; Tao, S.; Coveney, R.; Cao, J.; Li, B. G.; Xu, F. L.; Liu, W. X.; Wang, X. J.; Shen, W. R.; Qing, B. P.; Sun, R. A multimedia fate model for hexachlorocyclohexane(γ-HCH) in Tianjin. China. Environ. Pollut. 2004, 38, 2126-2132. (38) McKone, T. E. Alternative modeling approaches for contaminant fate in soils: uncertainty, variability, and reliability. Reliab. Eng. Syst. Saf. 1996, 54, 165-181. (39) Wu, S. P.; Tao, S.; Liu, W. X. Particle size distributions of polycyclic aromatic hydrocarbons in rural and urban atmosphere of Tianjin, China. Chemosphere, in press.

Received for review July 15, 2005. Revised manuscript received September 26, 2005. Accepted September 30, 2005. ES0513741