- Email: [email protected]
Particle-bound PAH content in ambient air
Environmental Pollution, Vol. 96, No. 3, pp. 369-382, 1997 PII:
S0269-7491
(97)00044-4
© 1997 Elsevier Science Ltd All fights reserved. Printed in Great Britain 0269-7491/97 $17.00 + 0.00
ELSEVIER
PARTICLE-BOUND PAH CONTENT IN AMBIENT AIR Hwey-Lin Sheu, *a Wen-Jhy Lee,"* Sue J. Lin, a Guor-Cheng Fang, b Huei-Chuau Chang a and Wen-Chun You" aDepartment of Environmental Engineering, National Cheng Kung University, Tainan 70101, Taiwan hDepartment of EnvironmentalEngineeringand Health, Hungkuang Institute of Medical and Nursing Technology, Taichung43309, Taiwan (Received 26 May 1996; accepted 12 March 1997)
Abstract
dual PAHs was between 25.3 and 49.6 lain, between 27.6 and 43.9 pro, and between 19.1 and 41.9 lun, respectively. This is due to the fact that P A H dry-deposition primarily resulted from gravitational settling of the coarse particulates ( > 10 lain). © 1997 Elsevier Science Ltd
Ambient air samples from a traffic intersection, an urban site and a petrochemical-industrial site (PCI) were collected by using several dry deposition plates, two Microorifice uniform deposited impactors ( MOUDIs), one Noll Rotary Impactor (NRI) and several PS-1 (General Metal Work) samplers from March 1994 to June 1995 in southern Taiwan, to characterize the atmospheric particle-bound P A H content of these three areas. Twenty-one individual polycyclic aromatic hydrocarbons (PAHs) were analyzed primarily by using a gas chromatograph/ mass spectrometer ( GC/MS). In general, the sub-micron particles have a higher P A H content. This is due to the fact that soot from combustion sources consists primarily of fine particles and has a high P A H content. In addition, a smaller particle has a higher specific surface area and therefore may contain more organic carbon, which allows for more P A H adsorption. For a particle size range between 0.31 and 3.21am, both Urban~Traffic and PCI/ Traffic ratios of particle-bound total-PAH content have the lowest values, ranging from 0.25 to 0.28 (mean=0.26) and from 0.07 to 0.13 (mean=0.10), respectively. This indicates that, during the accumulation process, the P A H mass shifted from a particle phase to a gas phase, or the particles aggregated with lower PAHcontent particles, resulting in a reduction in particlebound P A H content. By using the particle size distribution data, the dry deposition model in this study can provide a good prediction for the P A H content of dry deposition materials. In general, lower molecular weight PAHs had a larger fraction of dry deposition flux contributed by the gas phase;for 2-ring P A H (50.4, 46.3 and 28.4%), 3-ring PAHs (15.2, 15.4 and 11.7%) and 4-ring PAHs (13.0, 3.60 and 5.01%) for the traffic intersection, urban and PCI sites, respectively. For higher molecular weight PAHs--5-ring, 6-ring and 7-ring PAHs--their cumulation fraction ( F% ) of dry deposition flux contributed by the gas phase was lower than 3.26%. At the traffic intersection, urban and PCI sites, the mass median diameter of dry deposition materials (MMDF) of indivi-
Keywords: Particle size distribution, dry content, air quality, dry deposition.
INTRODUCTION The atmosphere is a major transport pathway for the movement of semi-volatile organic compounds (SOCs) through the global environment (Eisenreich et al., 1981; Duce and Gagosian, 1982; Bidleman and Foreman, 1987). Polycyclic aromatic hydrocarbons (PAHs) and their derivatives are widespread harmful SOCs generated by incomplete combustion of organic material arising, in part, from natural combustion, such as forest fires and volcanic eruptions, but, for the most part, from anthropogenic emissions (Bjorseth and Ramdahl, 1985; Benner et al., 1989; Back et al., 1991). The extent to which humans are exposed to PAHs is a function of several parameters, including the prevailing atmospheric conditions, the concentration in ambient air, their partition between the gas and particle phases and the size distributions of the particulate fraction (Baek et al., 1991). There is a potential increase in the potential carcinogenic effect of compounds, such as benzo(a)pyrene, by the adsorption on to particulate matter of small size, which can be taken into the bronchioles and alveoli of the lungs (Pierce and Katz, 1975). It is now generally accepted that PAHs associated with small particles ( < 1/zm) tend to result from combustion and other high temperature sources, whereas large particles (> 10/zm) are likely to arise from wind action on soils, deposited dusts, and fugitive emissions from dust-producing operations. This property of PAH size distribution largely determines the degree of penetration of the respiratory system and so is related to human toxicity. It can also determine deposition velocities and atmospheric residence times.
*To whom correspondence should be addressed. Present address: Department of Environmental Engineering, Kung Shan Institute of Technology, Yon Kan 710, Tainan, Taiwan. 369
370
Hwey-Lin Sheu et al.
Windsor and Hites (1979) indicated that various fuels burned in metropolitan areas produce airborne particulate matter (soot and fly ash) on which PAHs are absorbed. These aerosols are transported by winds for distances, which are a function of the aerosols' diameter. Small aerosols (< 1/zm), which are less efficiently removed by wet and dry deposition processes, have longer atmospheric residences times, and therefore they account for most of the PAH in remote marine and lacustrine sediments. Larger airborne aerosols (> 5 #m) not only have significant gravitational settling velocities, but are also efficiently removed from the atmosphere by precipitation. Thus, large aerosols are deposited closer to their sources. The PAH mass adsorbed on to the air particles divided by the particle mass is called 'particle-bound PAH content' (#gg-~). The aromatic content, the type of aromatic fraction, and the PAH content of a fuel or lubricant have a strong effect on particle-bound PAH emission. Compared to this, the lead content of the fuel is of minor importance, and the lubricant itself has only a slight influence on particle-bound PAH emission. Lee et al. (1995) found that the particle-bound totalPAH content of the ambient air at traffic-intersections averaged 6.5 and 11.6 times higher than those of urban and rural atmospheres, respectively. This result suggests that in ambient air near traffic-sources, the condensation process prevails among young aerosols and that during the aging process, PAHs were lost from the particle phase to the gas phase (Lee et al., 1995). Total suspended particulate (TSP) concentrations have been regulated by most countries in the world. However, few countries have considered TSP standards along with toxic substances, particularly the carcinogenic PAHs in air particulates. This has probably been hindered by insufficient data of particle-bound PAHs to perform risk assessments (Lee et al., 1995). In this study, ambient air samples from a trafficintersection, an urban site and a petrochemical-industrial (PCI) site were collected from March 1994 to June 1995 in southern Taiwan. The atmospheric particlebound PAH content associated with size distribution and dry deposition were investigated and characterized. The results of this study are aimed at the estimation of PAH dry deposition for human health risk assessment.
EXPERIMENTAL SECTION
Sampling sites Traffic-intersection samples were taken at the intersection of Tong-Ning Road and Chang-Rong Road, which is 1 km away from the campus boundary of National Cheng Kung University (NCKU), Tainan, Taiwan. The height of the sampler inlet was 1.5 m above the ground. This intersection is often congested with heavy traffic, which directly contributes large amounts of PAHs to the local ambient air. At the beginning of 1995, Tainan City had a population density of 4017 capita km -2 with
an area of 176 km 2 and a population of 707 052 inhabitants. Tainan City also had a vehicle density of 3628 vehicles km -2, including 161469 sedans, 341729 motorcycles, 135 262 vans, 782 buses and 30 486 trucks. Within a 5-km radius, there is no industrial plant around this traffic sampling site. Urban-site samples were taken on the roof of a fourstorey (13 m height) building located in a mixed institutional, commercial, and residential area in the centre of Tainan City. The building is on the NCKU campus, 60 m away from a main street and approximately 900 m from a bus station. Within a 5-kin circular radius, there is no industrial plant around this urban sampling site. The sampling site of the PCI plant was in an area called Zan-Der, which is located in southern Taiwan. There were more than 32 stacks simultaneously emitting PAHs from this PCI plant. Sampling was carried out on the roof of a three-storey (10m height) building.
Dry deposition plate The dry deposition flux was measured by using a smooth surface plate with a sharp leading edge, mounted on a wind vane (Noll et al., 1989). The plate used in this study was similar to those used in wind tunnel studies (McCready, 1986). It was made of polyvinyl chloride (PVC) and was 21.5 cm long, 8.0 cm wide and 0.8 cm thick with a sharp leading edge ( < 10° angle) that was pointed into the wind. In order to verify that there were no systematic errors, six duplicate plate samples were taken side by side during all of the sampling periods. Each of the plates was covered with aluminum strips (10crux8 cm) (average tare weight = 0.261888 g) coated with approximately 5mg of silicon grease (thickness 1/zm) (No. 11025 silicon spray, Cling-surface CO., Inc., Angola, NY) to collect impacted particles and gases (132cm 2 of total exposed surface). This hydrophobic coating has a high molecular weight and low vapor pressure and is therefore suitable to measure the PAH flux. After the grease was sprayed on to the strips, the strips were baked in an oven at 65°C for 90 min to remove volatile substances. The strip was placed on the plate and held down at the edge with a 0.03-cm thick stainless steel template, which was secured at each end by acrylic slats screwed into the plate. The plate was cut so that it would slide on to a 3-cm diameter rod. Two screws were fastened through the plate to a wind vane, allowing the plate to swing freely into the wind. Each plate was separated by 55 cm (horizontally), which was shown experimentally to be sufficient to prevent sample interactions. The strips were weighed before and after sampling to determine the total mass of the particles collected. The strips were then extracted and analyzed for PAHs. Analysis of unexposed aluminum strips (blanks) showed that their PAH content was below the detection limits of this study. Micro-orifice uniform deposited impaetors (MOUDIs) During the sampling periods, atmospheric particle and PAH mass size distributions were measured with two MOUDIs and a Noll Rotary Impactor (NRI), so that a
P A H content in ambient air
complete range of particle sizes could be obtained (from 0.056 to 100/~m) for particle-bound PAH analyses. The MOUDI is a nine-stage cascade impactor with a flow rate of 301itremin -~. Available particle cut-size diameters of MOUDI are 10, 5.6, 3.2, 1.8, 1.0, 0.56, 0.31, 0.18 and 0.1 or 0.056/zm. Each stage on the MOUDI consists of an impaction plate for the nozzle plate above, and a nozzle plate for the impaction plate below (MSP Corporation, 1989). The 47-mm-diameter aluminum circular filter-strips (from Gelman Sciences), which were used for the MOUDI sampler, were first cleaned and extracted with a solvent solution (mixture of n-hexane and dichloromethane, v:v= 1:1) for 24h in a Soxhlet extractor. Silicon grease was then applied to the surface of each filterstrip using the same procedure as for the dry deposition plates. The filter-strips were then allowed to equilibrate in a dust-free desiccator for at least 24 h before weighing on a Sartorius Electrobalance Model MP8-6. The measurable weight range of this electrobalance is 0.01 mg--,42 g and is accurate to a value of 0.01 mg. The strips were weighed before and after sampling to determine the total particle mass collected for each stage. In this study, the MOUDI was operated for a maximum of 16h at the traffic intersection and 24h at the urban and PCI sites, before the filter-strips were replaced.
371
exposed surface area by its average rotational distance. The NRI is described in more detail in Noll and Fang (1986). After combining the MOUDI and NRI, the normalized particle size range for each stage, in sequence, is 0.056,,,0.166, 0.166,,,0.31, 0.31~0.52, 0.52~1.0, 1.0~1.8, 1.8~3.2, 3.2~5.6, 5.6,~10, 10~24.7, 24.7~36.5 and 36.5,,, 100/zm. In this study, the concentration at overlap size ranges measured by both MOUDI (5.6~10/zm and 10,-~18/zm) and NRI (6.5 ,-~ 11.5/zm and 11.5 -~ 24.7/zm) was presented by the averaged data.
Semi-volatile sampler Ambient-air samples for the particle and gas phases of PAHs were collected by using a standard semi-volatile sampling train (General Metal Works PS-1). The sampler with a Whatman glass fibre filter, first cleaned by heating to 450°C, was used to collect particulate and particle phase PAHs. A glass cartridge containing a 5-cm polyurethane form (PUF) plug followed by 2.5 cm of XAD-2 resin and finally a 2.5 cm PUF plug cleaned by sequential extraction was used to collect the gas phase PAHs. The glass fibre filters were weighed before and after sampling to determine the total suspended particulate (TSP) collected.
Noll Rotary Impactor (NRI)
Sampling program
The atmospheric coarse particle mass was measured with a Noll Rotary Impactor (NRI) that collects large particles, which are excluded from other conventional samplers. It is a multi-stage rotary inertial impactor that collects coarse particles by simultaneously rotating four rectangular collectors (stages) of different dimensions through the air. The stages are covered with Mylar strips coated with silicon grease (as for the MOUDI). Total collection areas are 1.2 crn 2, 3.1 cm 2, 10.3 cm 2 and 10.3cm 2 for stages A, B, C and D, respectively. The dimensions for each NRI stage are the same as those used by Holsen and Noll (1992). Mylar film with a thickness of 0.051 mm was used as the collection medium for the NRI, because it has a low enough weight to be weighed on a microbalance and is rigid enough to be handled. Silicon grease was applied to the surface of each mylar strip to collect impacted particles. The mylar strips were then allowed to equilibrate in a dust-free desiccator for at least 8h before weighing on an Electrobalance as was done for the filter-strips of MOUDI. The strips were also weighed before and after sampling to determine the total particle mass collected for each stage. In this study, the NRI operated for a maximum of 4 h at the traffic intersection and 12h at the urban and PCI sites before the mylarstrips were replaced. The NRI was operated at 320 rpm, which produced a theoretical aerodynamic cut-diameters range (assuming a particle density of 1 gcm -3) of 6.5~100, 11.5~ 100, 24.7~100 and 36.5,-,100/zm for stages A, B, C and D, respectively. The volume of air sampled by each stage of the NRI was calculated by multiplying the
Sufficient amounts of material were collected during periods of no rain in each sampling period from the dry deposition plates to determine the mean PAH content for the gas phase, and for each particle size range. The sampling information is shown in Table 1.
PAH analysis After final weighing, the samples were placed in a solvent solution (mixture of n-hexane and dichloromethane, v:v = 1:1), and extracted in a Soxhlet extractor for 24 h. The extract was then concentrated, cleaned up and reconcentrated to exactly 1.0 or 0.5 ml. Since PAHs are thermally stable (they mostly originate from combustion processes) and exhibit low polarity, they can be detected and positively identified by using gas chromatograph/mass spectrometer (GC/MS) techniques. Massselective detection is thus one of the most sensitive spectrometric detection methods for PAH analysis (Gautschi and Mandel, 1994). The identification and quantification of PAHs was accomplished by using a GC (Hewlett-Packard 5890) with a mass selectivity detector (MSD) (Hewlett-Packard 5972A). This GC/MS was controlled by a computer workstation and was equipped with a Hewlett-Packard capillary column (HP Ultra 2-50 mx0.32 mmx0.17/zm), a HP-7673A automatic sampler, injection volume of 1/zl, splitless injection at 300°C, an ion source temperature of 280°C, oven temperature from 50°C to 100°C at 20°Cmin-1; 100°C to 290°C at 3°Cmin-I; and 290°C for 40min. The primary and secondary ion numbers of PAHs were determined by using the scan mode for pure PAH standards. Then, the quantification of PAHs was
372
Hwey-Lin Sheu et al. Table 1. Sampling information
Sampling periods
Exposure time for the dry deposition plate (days)
Sampling volume (m3) MOUDI
NRI A stage
B stage
C stage
D stage
Traffic intersection
1.29 June-2 July 1994 2. 14-17 July 1994 3.23-26 August 1994
4.0 3.9 4.0
235 219 284
178 177 202
535 531 607
1606 1593 1822
1606 1593 1822
6.4 5.0 5.0
376 419 435
349 334 453
1047 1002 1359
3142 3006 4077
3142 3006 4077
5.8 6.1 5.3
483 430 364
307 371 356
921 !113 1068
2763 3339 3204
2763 3339 3204
Urban site
1.28 March-3 April 1994 2.25-30 April 1994 3. 19-25 May 1994 PCI site
I. 5-18 November 1994 2. 26 April-2 May 1995 3.24-30 June 1995
performed by using the selective ion monitoring (SIM) mode. The concentrations of the following PAHs were determined: naphthalene (Nap), acenaphthylene (AcPy), acenaphthene (Acp), fluorene (Flu), phenanthrene (PA), anthracene (Ant), fluoranthene (FL), pyrene (Pyr), cyclopenta(c,d)pyrene (CYC), benz(a)anthracene (BaA), chrysene (Chr), benzo(b)fluoranthene (BbF), benzo(k)fluoranthene (BkF), benzo(e)pyrene (BeP), benzo(a)pyrene (BaP), perylene (PER), indeno(1,2,3,-cd)pyrene (IND), dibenz(a,h)anthracene (DBA), benzo(b)chrycene (BbC), benzo(ghi)perylene (BghiP) and coronene (COR). The GC/MSD was calibrated with a diluted standard solution of sixteen PAH compounds (PAH Mixture610M from Supelco) plus five individual PAH compounds (from Merck). Analysis of serial dilutions of PAH standards showed that the limit of detection for individual PAH compounds was between 23 and 524 pg. (The limit of quantification) LOQ is defined as the limit of detection divided by the sampling volume or the exposed time for the PS-1, MOUDI and NRI sampler, and for the dry deposition plate, respectively. For the PS-1 samplers, the LOQ for individual PAHs was between l0 and 221 pgm -3, whereas those values for dry deposition plates were between 13 and 288pgm-Eday -1. For MOUDI and A, B, C, D stages of NRI, the LOQ of GC/MS for individual PAHs, in sequence, was 15.5,~365pgm -3, 34.4,,~811 pgm -3, 40.7 ~ 958 pgm -3, 13.5~ 319 pgm-3 and 4.51 ~ 106 pg m -3, respectively. Ten consecutive injections of a PAH 610M standard yielded an average relative standard deviation of the GC/MSD integration area of 6.1%, with a range of 4.1 to 9.9%. In this study, two internal standards (phenanthrene-dlo and perylene-dl2) were used to check the response factors and the recovery efficiencies for PAHs analysis. The recovery efficiencies of 21 individual PAHs and these two internal standards were determined by processing a solution containing known PAH concentrations through the same experimental procedure used for the
samples. This study showed that the recovery efficiency of 21 individual PAHs varied between 73.6 and 115% and averaged 85.9%. The recovery efficiencies of two internal standards were between 81.0 and 87.0% and were fairly constant. The blank tests for PAHs were accomplished by using the same procedure as the recovery-efficiency tests without adding the known standard solution before extraction. Analysis of field blanks, including greased strips, glass fibre filters and PUF/resin cartridges, showed no significant contamination (GC/MS integrated area below the detection limit). Analysis of duplicate experiments yielded differences in total-PAH concentration ranging from 5.6 to 11.6% and averaging 8.8% for ambient air samples. Breakthrough tests were investigated by two layers of XAD-2 cartridge. Both upper and lower layers of XAD2 resin were analyzed individually and compared for the PAH mass collected in each layer. Three breakthrough tests were investigated in this study, and no significant PAH mass was found to be collected in the lower layer of XAD-2 resin.
RESULTS AND DISCUSSION Particle size distribution of total P A H s
For the traffic intersection, the particle size distribution (dC/dlogDp versus Dp) of total PAHs was unimodal (Fig. 1A). The major peak was located at the particle size range between 0.31 and 3.2#m, which mostly belonged to the fine particle mode (<2.5#m). This result indicates that the PAH mass in the ambient air of the traffic intersection originated mainly from automobile exhausts. This not only gives some important information about the potential for health hazards, but also indicates that the whole particle size distribution of PAHs essentially refects the gas-to-particle condensation in the fine particle mode after PAHs are emitted to the atmosphere. At the urban and PCI sites, the particle size distributions (dC/dlogDp versus Dp) of total PAHs were
373
P A H content in ambient air
Total-PAHs
i
S00
I TT IS'o t
400
(A)
300
200 100
"o
g.0,
0,1
0.5
O.56
10
00
washout and/or by dry deposition. Because of the resuspension of dust from the ground by the wind and/ or travelling vehicles, some of the PAH-aerosols persist in the atmosphere primarily as these coarse particles. However, when the particle size is greater than 25/zm, the mean value of dC/dlogDp is lower than 12.1 and 3.1ngm -3, for the urban and PCI site, respectively (Fig. 1B and C), because the coarse particles have a proportionally lower surface area than the fine particles.
100
120 e~ 100 t~ I:1
k
1
St]
!
~ "o
(B)
1
20 0 0.01
0.1
0.;06
1
2.5
10
26
100
100 ~0
80
"~ ¢21
o0
"~
20
g.01
1
0.1
0.25
1
2.6
(c)
10
5§
100
Particle Diameter (/zm) Fig. 1. Particle size distribution of total PAH (dC/dlogDp) in the ambient air at the traffic intersection, urban site and PCI site. Error bar represents the standard deviation.
Size distribution of particle-bound tottfl-PAH content The particle size distributions of particle-bound totalPAH content (/.tg g-] versus Dp) are shown in Fig. 2. In general, a smaller particle has a higher total-PAH content. This is due to the fact that soot from combustion sources is primarily made up of fine particles, which carry a high PAH content. In addition, a smaller size of particle has a higher specific surface area, and a higher attachment rate for organic pollutants. In the ambient air, at the traffic intersection, the major peak in particle size distribution for total-PAH content is found in a particle size range between 0.056 and 3.2/~m, with PAH concentrations ranging from 4970 to 5940/~g g-I (average 5510/~g g- l (Fig. 2A). This result reveals that the condensation and accumulation processes of the young aerosols are primarily incorporated into particles smaller than 3.2 ~tm in the ambient air of the traffic intersection. Particles between 3.2 and 100~tm in diameter have a total-PAH content between aooo
d-"
0ooo
I
bimodal (Fig. 1B and C). The bi-peaks of Fig. 1 occur at 0.31~0.52 and 1.8~3.2/~m, and 0.166~0.31 and 5.6~ 10/zm, for the urban, and PCI, sites, respectively. The fine and coarse particle modes were differentiated at 2.5/zm. At the urban sampling site, the major peak was located at the particle size range between 0.31 and 3.2 #m, which mostly belonged to the fine particle mode (particle size less than 2.5/~m). This result indicates that the PAH mass in ambient air of the urban area originated mainly from automobile exhausts. The peak of particle size distribution of PAHs localized at a particle size < 1.0 ~tm essentially reflects the gas-to-particle condensation process in the young aerosols, whereas the peak localized at the particle size range between 1.0 and 3.2 ~tm shows the occurrence of accumulation processes in these aerosols. Particles smaller than 3.0~tm are easily transported through the upper respiratory tract into the bronchioles and alveoli of the lungs and therefore pose a direct health hazard. At the PCI site, the two peaks in the total-PAHs size distributions were located at the fine and coarse particle modes, respectively. The PAHs were originally emitted from sources in the gas phase, adsorbed on to particulates, in part resisting degradation in the environment, and then went through atmospheric transport leading to a wider distribution. PAHs adsorbed on the aged aerosols would be removed from the atmosphere by rain
7ooo
t=O
5ooo
UO
4ooo
(a)
3ooo
g
zooo
0 lOOO
vj 0
10
16
100
4ooo 35oo
0 300O 2500
(s)
2000
I ,",,,4 c:)
1500
o
500
g
4000
(..,
1000
~.01
0.1
O.SO
1
0.8
10
J6
Ioo
3SO0
0 30O0
I
2500
,,,,a
0 *ea ,Ika
Cc)
2000 1500
t,.
c)
n,
lO00 O00 0.1
0.ORS
1
0.6
Particle Diameter
10
O8
100
(~m)
Fig. 2. Particle size distribution of particle-bound total-PAH content (v.gg -1) in the ambient air at the traffic intersection, urban site and PCI site. Error bar represents the standard deviation.
374
Hwey-Lin Sheu et al.
718 and 1920/zg g- 1 (average 1130 #g g- t). The lower total-PAH content of the coarse particles indicates that the aged aerosols are primarily resuspended from road dust and thus contain a greater quantity of inorganic crustal-elements like AI, Ca, Fe and Si and, therefore, have a smaller PAH content on them. For the urban area, two groups with peaks were found in the 0.056,,~ 0.31/zm (mode I) and 0.31 ~ 100/zm (mode II) size ranges (Fig. 2B). Mode I was attributed to primary emissions from combustion sources, whereas mode II was attributed to the accumulation of secondary reaction products on primary aerosol particles. The PAH content of mode I particles ranged between 2370 and 2880 #g g-1 (average 2630 #g g-i), whereas mode II particles ranged between 539 and 1550/xg g-I (average 1090/zg g-1). Aceves and Grimalt (1993) also found that PAHs are predominantly found in the submicron fractions at urban sites, particularly the < 0.5/zm fraction. They indicated that this is likely to be characteristic of early aging stages after gas-phase emission of combustion residues to the atmosphere. The mean values of total-PAH content for the PCI site were lower than those of the urban site. At the PCI sampling site, mode I and the mode II were also found in the 0.056,-~0.31/zm and 0.31~100/zm size ranges (Fig. 2C). The PAH content of mode I particles ranged between 1440 and 1450/zgg -] (average 1445/zgg-]), whereas mode II ranged between 326 and 708/zgg -1 (average 540/zg g-l).
The Urban/Traffic (U/T) and PCI/Traffic (P/T) ratios of particle size distribution for total-PAH content The U/T and P/T ratios are defined as the mean PAH content of the ambient air at urban and PCI sites divided by that at the traffic intersection. These two ratios can serve as the mean PAH dilution factor of urban and PCI air compared with that at the traffic intersection. The major peaks of particle size distribution for U/T and P/T ratios were in the size range between 3.2 and 100#m, with values ranging from 0.60 to 1.07 (mean 0.76) and from 0.21 to 0.72 (mean 0.49), respectively (Fig. 3). However, for the particle size range between 0.31 and 3.2#m, both the U/T and P/T ratios have the lowest values, the U/T ranging from 0.25 to 0.28 (average 0.26), and the P/T ranging from 0.07 to 0.13 (average 0.10). This showed that the ambient air of the traffic intersection had a greater young-aerosol mass, primarily as submicron particles. During the transport process from the combustion source to the urban or PCI atmosphere, and in order to reach a thermodynamic equilibrium, part of the PAH mass shifted from the particle phase to the gas phase again or, during the accumulation process, it aggregated with lower PAH-content particulates, thus resulting in the dilution of particlebound PAH (Lee et al., 1995). Particle size distribution (/~g versus Dp) of individualPAH content The particle size distributions of individual PAHs are shown in Table 2. The overall particle size range covers
0.056-100#m and can be separated into the following four size ranges: 0.056~1.0, 1.0,-~2.5, 2.5--~10, and 10 ~ 100/zm. In general, the measured particle-bound individual PAH contents had their highest peaks at a particle size of less than 1.0 #m. This is probalbly due to the fact that smaller particles have a higher specific surface area and therefore may contain greater amounts of organic carbon which allow more PAH adsorption (Table 2). The mean particle-bound content for higher molecular weight PAHs--BbF, BkF, BeP, BaP, PER, IND, DBA, BbC, and BghiP--in the ambient air at the traffic intersection were 4.7 and 10.2 times higher than those at the urban and PCI site, respectively. The values of dC/dlogDp (ngm -3) for individual PAHs measured at three sampling sites are shown in Table 3. Generally, a particle size smaller than 2.5 #m has a higher value of dC/dlogDp. These results indicate that a very significant fraction of PAH mass was originally formed in the gas phase of the automobile exhaust at a high temperature. Upon entering the ambient air, these PAHs cooled, and most of the higher molecular weight PAHs were incorporated into particulate matter by adsorption and condensation processes (Lee et al., 1995).
Measurements of particle-bound PAH content of the collected dry deposition materials The dry-deposition PAH mass divided by the total mass of dry-deposition gives the particle-bound PAH content 1.1 1.0 0.9 0.8 0.7 0.6
A)
0.5 0
0.4 0.3
.< |
[.. = 0 = °~
0.2 0.1 0,0 0.01 1.1
0.1
0.25
1
;2.5
10
25
100
0.1
O.B5
1
2.5
10
85
100
1.0 0.9 0.8 0.7
0.6 0.5 m
0.4 0.3 0.2 0.1 0.0 0.01
Particle Diameter (~m) Fig. 3. The urban/traffic (U/T) ratio and PCI/traffic (P/T) ratio of particle size distribution for particle-bound total-PAH content.
375
P A H content in ambient air
¢N
t¢5
7~
=I
II
,=
,m
i °11 ~ I
7~ = ! el II
:E [] II
< i
< e~
Hwey-Lin Sheu e t
376
al.
0
7~ 0
~ o o o o o o o o o o ~ o ~ o
. . . .
0
@
0
0
0 °~ ¢xl
0 m
0
°~
0
7~
1
f
0 @
0
0
o. 0
< <
~
~ Omm
2 m~
~' ~mmO
i~
377
P A H content in ambient air 1 t'~,d 2 :v'e
3000
2,o0
I
d'~pltcatod s~mples
1
2000 I
CA)
1500
~L
1000 500
o
.g o
0
Juno-duly 1400 12oo 1000 800
I
"d o
E~ "U g
600
1 2 \ ~ \ xy \ \ N %'% \\\ \\N \\x,
"[X~
\ \ N I(X x \ ' ~
400
~,.\\ K,~
200
N \ \ KX N \ N KX N~\ < x
~ duly_
1994
ApUg~ 2
1400
I
1200
0
1000
I-
BOO
Dry deposition modelling The deposition velocities for the atmospheric particles by the dry deposition model were obtained by the following equations (Noll and Fang, 1989): Vd = //st "[- 1.12 x U* x exp(-30.36/Dp),
April
o
Q,
1904
Because of the 'synergism' of higher wind speed and larger particle MMDs, the dry deposition flux of totalparticle mass at the PCI site was much higher than at the urban site. A particulate with a larger size has less specific surface area and would, therefore, allow less PAHs to be adsorbed on it, so, for the total-particle mass collected by dry deposition, the total-PAH content in the PCI site averaged only 56% of that measured at the urban site (Fig. 4).
M~Ily 4
1994
(1)
where Vst = particle settling velocity (crn s- 1); U* = friction velocity (cm s-l); and Dp = particle diameter (/zm).
1
When Dp_>5 tzm, [c)
SO0
Vst :
1
VT, stock
(2)
400 200 NOV. 1994,
April 19g8
Juno 190fl
VT,stock =
( ,D:pg)/18u,
(3)
Sampling Periods
Fig. 4. Measured particle-bound total-PAH content (/zgg-n) of the collected dry deposition materials sampled at the traffic intersection, urban site and PCI site.
where VT,stock=terminal settling velocity (cms-~); pp=density of particle (1.0 g cm-3); g = gravitational constant (980cm s-2); and /z = absolute viscosity of air (gcm
-I s--J).
W h e n Dp < 5 tzm,
(#g g-I) of the collected dry deposition materials. These values for total-PAHs at the traffic intersection varied between 1730 and 2370/zgg -l (average 2010/xgg -1) (Fig. 4A). For the urban site, the particle-bound totalPAH content of the collected dry deposition materials varied between 993 and 1230/zgg -l (average 1200/xgg -t) (Fig. 4B), whereas for the PCI site, this ranged between 356 and 973/zgg -~ (average: 672/zgg -1) (Fig. 4C). In the ambient air of the traffic intersection, BaP (304/zgg-l), IND (229/zgg-n), PER (223/xgg-l), BbF (218/zgg -t) and DBA (201/zgg -~) had the highest concentrations in the collected dry deposition materials. The values for the remaining PAHs at the urban site varied between 2.03 and 199/zgg -l and averaged 52.2/zgg -I. In the ambient air of the urban site, Ant (136/zgg-]), IND (121/zgg-1), Flu (106/zgg-~), FL (87.4/zgg -t) and Acp (85.2/zgg -l) had the highest amounts of particle-bound PAH in the collected dry deposition materials. The values for the remaining PAHs at the urban site varied between 10.8 and 75.0/zgg -I and averaged 41.3/zgg-k At the PCI site, IND (79.3/zgg-~), DBA (62.8/zgg-~), COR ( 54.2/zg g- 1), Nap (50.6/zg g- l) and BghiP (49.3/~g g- l) were the most important particle-bound PAHs in dry deposited. The values for the remaining PAHs at the PCI site varied between 9.80 and 38.2/zgg -1 and averaged 23.5/zg g-l.
Vst :-- VT,stock x gc,
(4)
where K¢ = 1 + 2~.[1.257 + 0.400 exp(-0.55 Dp/X)]/Dp
~. =/z/[0.499P (8M/ztRT)°'5],
(5)
(6)
where K¢ = cunningham correction; ~.= mean free path (cm); P--absolute pressure (Pa); R--universal gas constant (8.314J-gmol-I°K-1); M = molecular weight (gmol-1); and T-- absolute temperature (°K). Two meteorological parameters that influence atmospheric turbulence are friction velocity (U*) and surface roughness (Z0). The relationship between these parameters for near neutral atmospheric stability condition is: U* Z - d U----k- l n - - ~ 0 ,
(7)
where U=measured mean wind speed at height Z (cms-]); Z=measured height above ground (m);
378
Hwey-Lin Sheu et al.
k = V o n Karman's constant (0.4); d=datum displacement (m); and Z0 = surface roughness height (m). Since wind was measured at only one datum displacement, the average height of the structures surrounding the sampling site of the traffic intersection, urban and PCI areas were taken as 8.0, 8.0 and 5.0 m, respectively. A datum level displacement of (6.4, 6.4 and 4.0 m) (80% of the average structure height) was subtracted from the height (13, 13 and 10m) where the wind speed was recorded for the (traffic intersection, urban and PCI sites), respectively. The roughness coefficient was estimated by the general relationship that Zo is approximately 1/30 of the average roughness height. Since the average height for the area surrounding the sampling site of (traffic intersection, urban and PCI areas) was (8.0, 8.0 and 5.0 m), the roughness heights of (0.27, 0.27 and 0.17 m) were used for calculating friction velocity, respectively. The friction velocities for wind speed measured in this study were calculated by using eqn 7. The modelled Vd,i of each impactor stage at the sampiing site of (traffic intersection, urban and PCI areas) was (0.11, 0.069 and 0.090), (0.81, 0.38 and 0.81), (6.25, 2.58 and 6.25), (14.2, 6.36 and 14.2) and (33.4, 19.9 and 33.4 cm s-1) for the particle size range of 3.2 ~ 5.6 (mean Dp=4.4#m), 5.6~10 (mean Dp=7.8/zm), 10~24.7 (mean Dp = 17.35/zm), 24.7 ~ 36.5 (mean Dp = 30.6 #m) and 36.5~ 100#m (mean Dp=68.25/zm), respectively. For each impactor stage, the variation of Vd.i for each sampling site was fairly constant, because the variation of wind speed was minor during the sampling periods. The ratio of modelled/measured (Md/Ma) dry deposition of particle-bound content
The calculated total dry-deposition flux (Ft) for total particle mass (mg m -2 day-m), total PAHs or individual PAHs (/zgm -2 day-l), was calculated from: 11
Ft = Fg + Z
Fi
(8)
i=l 11
= Cg. Vd,g'q-~f~fi. Vd, i,
bound individual-PAH content was the Ft of individual PAHs divided by the Ft of the total particle mass. The calculation procedure divided the combination of gas phase and particle distributions into 12 intervals and assigned a modelled deposition velocity to each interval. The calculated flux for each interval was then summed to calculate the total dry-deposition flux. After the calculated particle-bound PAH content of the collected dry deposition materials for total PAHs or individual PAHs was obtained, the calculated data were compared to the measured data. The Vd.g of semivolatile organics is controlled mainly by diffusion. Its magnitude is close to or smaller than the measured dry deposition velocity of naphthalene (Nap), which usually has more than 99% of mass existing in the gas phase. Here, selected a Vd,g of 0.01 cm s- i was used for the gas-phase PAHs (Sheu et al., 1996). The modelled/measured (Md/Ma) ratio of particlebound total-PAH content of dry deposition materials varied between 0.86 and 1.18 at the traffic intersection, between 0.88 and 1.22 at the urban site and between 0.77 and 0.99 at the PCI site (Table 4). All the modelled and measured dry deposition of particle-bound content data for total PAHs were within 33.8%. The Md/Ma ratio of particle-bound individual-PAH content of dry deposition materials ranged from 0.32 to 1.69, from 0.31 to 1.64 and from 0.40 to 1.54 for the traffic intersection, urban and PCI sites, respectively (Table 5). By using the particle size distribution data, in general, this dry deposition model can provide a good prediction for the particle-bound PAH content of dry deposition materials. Cumulative fraction (F%) of dry deposition of particlebound individuai-PAH content
In general, a lower molecular weight PAH had a larger fraction of dry deposition flux contributed by the gas phase. They were (50.4, 46.3 and 28.4%) for 2-ring PAH, (15.2, 15.4 and 11.7%) for 3-ring PAHs and (13.0%, 3.60% and 5.01%) for 4-ring PAHs measured at the (traffic intersection, urban and PCI site), respectively. For higher molecular weight PAHs--5-ring,
(9)
i=1
where Fg = dry-deposition flux contributed by the gasphase total PAHs or individual PAHs (/zgm-Zday-I); Fi =dry-deposition flux contributed by each impactor stage which is from stage 1 (0.056~0.166/zm) to stage 11 (36.5~100/zm), respectively; Cg=the measured ambient concentration (ng m -3) of total PAHs or individual PAHs; Vd,g= the dry deposition velocity of gasphase PAHs; Ci =measured concentration (/xgm -3) of total particle mass, total PAHs or individual PAHs in each impactor stage; Vd.i (cms-I)=calculated dry deposition velocity by using the dry deposition model (Noll and Fang, 1989) for the mean particle diameter (/zm) of each impactor stage. By using the Ft, the calculated particle-bound totalPAH content was the Ft of total PAHs divided by the Ft of the total particle mass, and the calculated particle-
Table 4. Ratios of modelled/measured (Md/Ma) particle-bound totaI-PAH content of the collected dry deposition materials sampled at the traffic intersection, the urban site and the PCI site
Sampling periods Traffic intersection June--July 1994 July 1994 August 1994 Urban site March 1994 April 1994 May 1994 PCI site November 1994 April 1995 June 1995
Ratio
A. 1.15 A. 1.12
B. 0.93 B. 1.07 B. 1.18
A. 1.04 A. 1.09 A. 0.88
B. 1.03 B. 1.22 B. 0.91
A. 0.99 A. 0.81 A. 0.77
B. 0.92 B. 0.77 B. 0.87
A. 0.86
A and B are duplicated samples.
P A H content in ambient air
379
Table 5. Ratios of modelled/measured (Mfl/Ma) purticle-bound individuaI-PAH content of the collected dry deposition materials sampled at the traffic inter~,ction (T), urban site and PCI site PAHs
Ratio
NaP AcPy Acp Flu PA Ant FL
PAHs
T
Urban
PCI
1.04 0.67 0.57 0.38 0.59 1.69 0.69
1.61 1.08 0.69 0.41 0.36 0.31 0.32
0.51 0.66 0.65 0.68 0.47 0.76 0.75
Ratio T
Pyr CYC BaA CHR BbF BkF BeP
0.51 0.71 0.47 0.87 1.51 1.83 0.48
6-ring and 7-ring P A H s - - t h e i r F % of the dry deposition flux contributed by the gas phase was lower than 3.26% (Figs 5-7). In ambient air of both urban and PCI sites, generally, the profiles of cumulative percent of modelled drydeposition flux are similar with those at the traffic intersection (Figs 5-7). The cumulative percentage of modelled dry-deposition flux (F%) based on the totalPAH mass (summation of 21 individual PAHs), more than 92.4, 91.9 and 87.5% of PAH dry deposition flux, is contributed by particulates with diameters larger than 10/zm in the ambient air of the traffic intersection, urban and PCI sites, respectively. Even though PAHs existed mainly in the gas phase and/or fine particulates ( < 2.5/zm), these results indicate that PAH dry deposition is mainly due to the coarse particulates, particularly
1 O0
l Traffic
PAHs
Urban
PCI
1.13 1.56 1.38 1.24 0.94 1.12 1.77
T
0.52 0.40 0.78 1.48 1.14 1.54 0.59
BaP PER IND DBA BbC BghiP COR
1.07 1.36 1.01 1.01 0.94 1.83 1.00
PCI
1.50 1.59 0.50 1.64 1.63 1.40 1.62
0.46 1.41 1.5l 0.70 0.61 1.24 0.66
the particulates larger than 10/zm from gravitational settling. The atmospheric dry deposition of PAHs due to the fine particulates ( < 2.5 tzm) by the initial impaction was very minor. High particle-bound PAH content of dry deposition materials resulting from the coarse particulates may undergo bioaccumulation and result in significant adverse health effects. Particle-bound individual-PAH content (/~gg-l) at the modelled MMI)v and of the measured data The mass median diameter of dry deposition materials (MMDF) was obtained from the curve of cumulative percentage of the modelled PAH dry-deposition flux in Figs 5-7 and chosen for the particle diameter at 50% of the cumulative curve (dso). In the ambient air of the traffic intersection, the MMDF was 39.1/zm for total
100 o--o--
5 0. 4%
o--o--o--, Z-ring
U ~ b f/, _n.-~
PAH
046"3g 152%
3-ring v- -
13.07.
10
Urban
]
o
o
Ratio
2-ring
.0/
PAH
O__ 0__0__0_ 0_0_0_ 0----"0"~
PAHs v-- v-- v --
v
4-ring PAHs
0--0--0--@--Q--0~
"~, o
lO
va.O0g
4-ring PAHs V/v ~ / v--v--v--v--v~
gas phase gas
/
phase
/
o .87Z
0.297*
6-ring
PAHs
~/ o~ a / " 5 - r i n g PAHs
2
///
o/ / /
•
/
J
0 147. ~o
a
7-ring
~
2
PAH
0.1
~o 0.001
0.01
0.1
1
10
0 , I 9%
~IIO'OBT~
0.1
5-ring PAHs v / '/ v__v~V / 7-ring
PAH
m/
100 ........
Particle Diameter (p,m) Fig. 5. Cumulative fraction (%) of modelled dry deposition flux for PAHs size distribution in the ambient air of traffic intersection.
0.001
i 0.01
........
Particle
i 0.1
........
Diameter
i 1
........
i 10
........
i
100
(~m)
Fig. 6. Cumulative fraction (%) of modelled dry deposition flux for PAHs size distribution in the ambient air of urban site.
Hwey-Lin Sheu et al.
380
m 0I
@
0
~
~
-
~o
~
-
-
-
~
~
~
. . . . . .
N °--
r~
~°~ @
f
@
"~,
~8
!
III
h~ °~
&
[,.
< a~
<
P A H content in ambient air
1 O0
028"'" r~
o
O--@--O--O--O--O~Q /
10
Z"
/~5.01~
"U t~
/
gag phage t~
v0"47~g
5-ring PAHs v__v__v~V ~v
r~
/
/
8
r~
./v//
0.102 0.1
0
........ 0.001
i
0.01
. . . . . . . .
6-ring PAHs 0 / O--O ~°/
i
0.1
. . . . . . .
i
1
Particle Diameter
. . . . . . . .
)
10
. . . . . . . .
i
100
(g.m)
Fig. 7. Cumulative fraction (%) of modelled dry deposition flux for PAHs size distribution in the ambient air of PCI site. PAHs and was between 25.3 and 49.6/zm for individual PAHs; the values of particle-bound total-PAH content at MMDF were 1196/xgg -1 and were between 2.11 and 288/zgg -l for individual PAHs (Table 6). In the ambient air of the urban site, the MMDF was 38.2/zm for total PAHs and was between 27.6 and 43.9/zm for individual PAHs. The values of particle-bound totalPAH and individual PAH content at MMDF were 613tzgg -l and were between 6.92 and 86.3/zgg -l, respectively (Table 6). In the ambient air of the PCI site, the MMDF was 33.8/zm for total PAHs and was between 19.1 and 41.9#m for individual PAHs. The values of particle-bound total-PAH and individual PAH content at MMDF were 486#gg -I and were between 1.73 and 103/zgg -1, respectively (Table 6). In general, most of the MMDF for individual PAHs are larger than 25.0/zm. This is due to the fact that the PAH drydeposition primarily resulted from the coarse particulates (> 10 #m) by gravitational settling.
CONCLUSION In general, smaller particles have a higher PAH content. This is due to the fact that the soot from the combustion sources is primarily a fine particulate, which carries a high PAH content. In addition, smaller particles have a higher specific surface area and a higher attachment rate for organic pollutants. It may therefore contain a greater amount of organic carbon, which allows more PAH adsorption.
381
In the particle size range between 0.31 and 3.2/zm, both U/T and P/T ratios of particle-bound total-PAH content have the lowest values. This indicates that during the accumulation process, the PAH mass is shifted from the particle phase to the gas phase or that aggregation of particulates with a lower PAH content results in a reduction in particle-bound PAH content. By using the particle size distribution data, the dry deposition model used in this study can provide a good prediction for the PAH content of the collected dry deposition materials. Lower molecular weight PAHs had a large fraction of the dry deposition flux contributed by the gas phase. For higher molecular weight PAHs (5-ring, 6-ring and 7-ring PAHs), F% of dry deposition flux contributed by the gas phase was lower than 3.3%. At the traffic intersection, urban and PCI site, the MMDF of individual PAHs was between 25.3 and 49.6#m, between 27.6 and 43.9#m and between 19.1 and 41.9/zm, respectively. In general, most of MMDF for individual PAHs are larger than 25.0ttm. The cumulative percentage of modelled dry-deposition flux (F%) based on the total-PAH mass (summation of 21 individual PAHs), more than 92.4, 91.9 and 87.5% of PAH dry deposition flux is contributed by particulates with diameters larger than 10#m in the ambient air of the traffic intersection, urban and PCI sites, respectively. These results revealed that the majority of dry-deposition PAH mass resulted primarily from the coarse particulates (> 10/zm) by gravitational settling.
ACKNOWLEDGEMENTS This research was supported by funds from the National Science Council, Taiwan, Grant No. NSC 86-2113-M006-018. The authors also thank Mr. Chun-C/aing Su, Mr. Yi-Chin Fan and Mr. How-Ran Chao for their help in sampling.
REFERENCES Aceves, M. and Grimalt, J. O. (1993) Seasonally dependent size distributions of aliphatic and polycyclic aromatic hydrocarbons in urban aerosols from densely populated areas. Environmental Science and Technology 27, 2896-2908. Baek, S. O., Field, R. A., Goldstone, M. E., Kirk, P. W., Lester, J. N. and Perry, R. (1991) A review of atmospheric polycyclic aromatic hydrocarbons: sources, fate and behavior. Water, Air, and Soil Pollution 60, 273-300. Benner, B. A., Gordon, G. E. and Wise, S. A. (1989) Mobile sources of atmospheric polycyclicaromatic hydrocarbons: a roadway tunnel study. Environmental Science and Technology 23, 1269-1278. Bidleman, T. F. and Foreman, W. T. (1987) Vapor-particle partitioning of semivolatile organic compounds. In The Chemistry of Aquatic Pollutants, ed. R. Hites and S. J. Eisenreich. ACS Advances in Chemistry Series, ACS, New York. Bjerseth, A. and Ramdahi, T. (1985) Source and emissions of PAH. In Handbook of Polycyclic Aromatic Hydrocarbons, p. 1. Marcel Dekker, New York.
382
Hwey-Lin Sheu et al.
Duce, R. A. and Gagosian, R. B. (1982) The input of atmospheric n-Ci0 to n-C30 alkanes to the ocean. Journal of Geophysics Research 87, 7192-7200. Eisenreich, S. J., Looney, B. B. and Thornton, L. D. (1981) Airborne organic contaminants in the Great Lakes ecosystem. Environmental Science and Technology 15, 3038. Gautschi, P. and Mandel, F. (1994) The detection of polynuclear aromatic hydrocarbons using the HP 5972A MSD. Hewlett-Packard Application Note, 23-5962-8614E. Holsen, T. M. and Noll, K. E. (1992) Dry deposition of atmospheric particles: application of current models of ambient data. Environmental Science and Technology 26, 1807-1815. Lee, W. J. (1991) The determination of dry deposition velocities for ambient gases and particles. Ph.D. thesis, Illinois Institute of Technology, Chicago, IL, pp. 85-138. Lee, W. J., Wang, Y. F., Lin, T. C., Chen, Y. Y., Lin, W. C., Ku, C. C. and Cheng, J. T. (1995) PAH characteristics in the ambient air of traffic-source. Science of the Total Environment 159, 185-200. McCready, D. I. (1986) Wind tunnel modeling of small particle deposition. Aerodynamics Science and Technology 5, 301-312.
MSP Corporation (I 989) Micro-orifice uniform deposit impactor. User's Instruction of MOUDL pp. 2-3. Noll, K. E. and Fang, K. Y. P. (1986) A rotary impactor for size selective sampling of atmospheric coarse particles. Proceedings of the Air Pollution Control Association 79th Annual Meeting, Minneapolis, Minnesota, Paper No. 8640.2. Noll, K. E. and Fang, K. Y. P. (1989) Development of a dry deposition model for atmospheric coarse particles. Atmospheric Environment 23, 585-594. Noll, K. E., Fang, G. C. and Kenneth, Y. P. (1989) Development of a dry deposition model for atmospheric coarse particles. Atmospheric Environment 23, 585-594. Pierce, R. C. and Katz, M. (1975) Dependency of polynuclear aromatic hydrocarbon content on size distribution of atmospheric aerosols. Environmental Science and Technology 9, 347-353. Sheu, H. U, Lee, W. J., Su, C. C., Chao, H. R. and Fan, Y. C. (1996) Dry deposition of polycyclic aromatic hydrocarbons in ambient air. Journal of Environmental Engineering ASCE 122, 1101-1109. Windsor, J. G., Jr and Hites, R. A. (1979) Polycyclic aromatic hydrocarbons in Gulf of Maine sediments and Nova Scotia soil. Geochimica Acta 43, 27-33.
S0269-7491
(97)00044-4
© 1997 Elsevier Science Ltd All fights reserved. Printed in Great Britain 0269-7491/97 $17.00 + 0.00
ELSEVIER
PARTICLE-BOUND PAH CONTENT IN AMBIENT AIR Hwey-Lin Sheu, *a Wen-Jhy Lee,"* Sue J. Lin, a Guor-Cheng Fang, b Huei-Chuau Chang a and Wen-Chun You" aDepartment of Environmental Engineering, National Cheng Kung University, Tainan 70101, Taiwan hDepartment of EnvironmentalEngineeringand Health, Hungkuang Institute of Medical and Nursing Technology, Taichung43309, Taiwan (Received 26 May 1996; accepted 12 March 1997)
Abstract
dual PAHs was between 25.3 and 49.6 lain, between 27.6 and 43.9 pro, and between 19.1 and 41.9 lun, respectively. This is due to the fact that P A H dry-deposition primarily resulted from gravitational settling of the coarse particulates ( > 10 lain). © 1997 Elsevier Science Ltd
Ambient air samples from a traffic intersection, an urban site and a petrochemical-industrial site (PCI) were collected by using several dry deposition plates, two Microorifice uniform deposited impactors ( MOUDIs), one Noll Rotary Impactor (NRI) and several PS-1 (General Metal Work) samplers from March 1994 to June 1995 in southern Taiwan, to characterize the atmospheric particle-bound P A H content of these three areas. Twenty-one individual polycyclic aromatic hydrocarbons (PAHs) were analyzed primarily by using a gas chromatograph/ mass spectrometer ( GC/MS). In general, the sub-micron particles have a higher P A H content. This is due to the fact that soot from combustion sources consists primarily of fine particles and has a high P A H content. In addition, a smaller particle has a higher specific surface area and therefore may contain more organic carbon, which allows for more P A H adsorption. For a particle size range between 0.31 and 3.21am, both Urban~Traffic and PCI/ Traffic ratios of particle-bound total-PAH content have the lowest values, ranging from 0.25 to 0.28 (mean=0.26) and from 0.07 to 0.13 (mean=0.10), respectively. This indicates that, during the accumulation process, the P A H mass shifted from a particle phase to a gas phase, or the particles aggregated with lower PAHcontent particles, resulting in a reduction in particlebound P A H content. By using the particle size distribution data, the dry deposition model in this study can provide a good prediction for the P A H content of dry deposition materials. In general, lower molecular weight PAHs had a larger fraction of dry deposition flux contributed by the gas phase;for 2-ring P A H (50.4, 46.3 and 28.4%), 3-ring PAHs (15.2, 15.4 and 11.7%) and 4-ring PAHs (13.0, 3.60 and 5.01%) for the traffic intersection, urban and PCI sites, respectively. For higher molecular weight PAHs--5-ring, 6-ring and 7-ring PAHs--their cumulation fraction ( F% ) of dry deposition flux contributed by the gas phase was lower than 3.26%. At the traffic intersection, urban and PCI sites, the mass median diameter of dry deposition materials (MMDF) of indivi-
Keywords: Particle size distribution, dry content, air quality, dry deposition.
INTRODUCTION The atmosphere is a major transport pathway for the movement of semi-volatile organic compounds (SOCs) through the global environment (Eisenreich et al., 1981; Duce and Gagosian, 1982; Bidleman and Foreman, 1987). Polycyclic aromatic hydrocarbons (PAHs) and their derivatives are widespread harmful SOCs generated by incomplete combustion of organic material arising, in part, from natural combustion, such as forest fires and volcanic eruptions, but, for the most part, from anthropogenic emissions (Bjorseth and Ramdahl, 1985; Benner et al., 1989; Back et al., 1991). The extent to which humans are exposed to PAHs is a function of several parameters, including the prevailing atmospheric conditions, the concentration in ambient air, their partition between the gas and particle phases and the size distributions of the particulate fraction (Baek et al., 1991). There is a potential increase in the potential carcinogenic effect of compounds, such as benzo(a)pyrene, by the adsorption on to particulate matter of small size, which can be taken into the bronchioles and alveoli of the lungs (Pierce and Katz, 1975). It is now generally accepted that PAHs associated with small particles ( < 1/zm) tend to result from combustion and other high temperature sources, whereas large particles (> 10/zm) are likely to arise from wind action on soils, deposited dusts, and fugitive emissions from dust-producing operations. This property of PAH size distribution largely determines the degree of penetration of the respiratory system and so is related to human toxicity. It can also determine deposition velocities and atmospheric residence times.
*To whom correspondence should be addressed. Present address: Department of Environmental Engineering, Kung Shan Institute of Technology, Yon Kan 710, Tainan, Taiwan. 369
370
Hwey-Lin Sheu et al.
Windsor and Hites (1979) indicated that various fuels burned in metropolitan areas produce airborne particulate matter (soot and fly ash) on which PAHs are absorbed. These aerosols are transported by winds for distances, which are a function of the aerosols' diameter. Small aerosols (< 1/zm), which are less efficiently removed by wet and dry deposition processes, have longer atmospheric residences times, and therefore they account for most of the PAH in remote marine and lacustrine sediments. Larger airborne aerosols (> 5 #m) not only have significant gravitational settling velocities, but are also efficiently removed from the atmosphere by precipitation. Thus, large aerosols are deposited closer to their sources. The PAH mass adsorbed on to the air particles divided by the particle mass is called 'particle-bound PAH content' (#gg-~). The aromatic content, the type of aromatic fraction, and the PAH content of a fuel or lubricant have a strong effect on particle-bound PAH emission. Compared to this, the lead content of the fuel is of minor importance, and the lubricant itself has only a slight influence on particle-bound PAH emission. Lee et al. (1995) found that the particle-bound totalPAH content of the ambient air at traffic-intersections averaged 6.5 and 11.6 times higher than those of urban and rural atmospheres, respectively. This result suggests that in ambient air near traffic-sources, the condensation process prevails among young aerosols and that during the aging process, PAHs were lost from the particle phase to the gas phase (Lee et al., 1995). Total suspended particulate (TSP) concentrations have been regulated by most countries in the world. However, few countries have considered TSP standards along with toxic substances, particularly the carcinogenic PAHs in air particulates. This has probably been hindered by insufficient data of particle-bound PAHs to perform risk assessments (Lee et al., 1995). In this study, ambient air samples from a trafficintersection, an urban site and a petrochemical-industrial (PCI) site were collected from March 1994 to June 1995 in southern Taiwan. The atmospheric particlebound PAH content associated with size distribution and dry deposition were investigated and characterized. The results of this study are aimed at the estimation of PAH dry deposition for human health risk assessment.
EXPERIMENTAL SECTION
Sampling sites Traffic-intersection samples were taken at the intersection of Tong-Ning Road and Chang-Rong Road, which is 1 km away from the campus boundary of National Cheng Kung University (NCKU), Tainan, Taiwan. The height of the sampler inlet was 1.5 m above the ground. This intersection is often congested with heavy traffic, which directly contributes large amounts of PAHs to the local ambient air. At the beginning of 1995, Tainan City had a population density of 4017 capita km -2 with
an area of 176 km 2 and a population of 707 052 inhabitants. Tainan City also had a vehicle density of 3628 vehicles km -2, including 161469 sedans, 341729 motorcycles, 135 262 vans, 782 buses and 30 486 trucks. Within a 5-km radius, there is no industrial plant around this traffic sampling site. Urban-site samples were taken on the roof of a fourstorey (13 m height) building located in a mixed institutional, commercial, and residential area in the centre of Tainan City. The building is on the NCKU campus, 60 m away from a main street and approximately 900 m from a bus station. Within a 5-kin circular radius, there is no industrial plant around this urban sampling site. The sampling site of the PCI plant was in an area called Zan-Der, which is located in southern Taiwan. There were more than 32 stacks simultaneously emitting PAHs from this PCI plant. Sampling was carried out on the roof of a three-storey (10m height) building.
Dry deposition plate The dry deposition flux was measured by using a smooth surface plate with a sharp leading edge, mounted on a wind vane (Noll et al., 1989). The plate used in this study was similar to those used in wind tunnel studies (McCready, 1986). It was made of polyvinyl chloride (PVC) and was 21.5 cm long, 8.0 cm wide and 0.8 cm thick with a sharp leading edge ( < 10° angle) that was pointed into the wind. In order to verify that there were no systematic errors, six duplicate plate samples were taken side by side during all of the sampling periods. Each of the plates was covered with aluminum strips (10crux8 cm) (average tare weight = 0.261888 g) coated with approximately 5mg of silicon grease (thickness 1/zm) (No. 11025 silicon spray, Cling-surface CO., Inc., Angola, NY) to collect impacted particles and gases (132cm 2 of total exposed surface). This hydrophobic coating has a high molecular weight and low vapor pressure and is therefore suitable to measure the PAH flux. After the grease was sprayed on to the strips, the strips were baked in an oven at 65°C for 90 min to remove volatile substances. The strip was placed on the plate and held down at the edge with a 0.03-cm thick stainless steel template, which was secured at each end by acrylic slats screwed into the plate. The plate was cut so that it would slide on to a 3-cm diameter rod. Two screws were fastened through the plate to a wind vane, allowing the plate to swing freely into the wind. Each plate was separated by 55 cm (horizontally), which was shown experimentally to be sufficient to prevent sample interactions. The strips were weighed before and after sampling to determine the total mass of the particles collected. The strips were then extracted and analyzed for PAHs. Analysis of unexposed aluminum strips (blanks) showed that their PAH content was below the detection limits of this study. Micro-orifice uniform deposited impaetors (MOUDIs) During the sampling periods, atmospheric particle and PAH mass size distributions were measured with two MOUDIs and a Noll Rotary Impactor (NRI), so that a
P A H content in ambient air
complete range of particle sizes could be obtained (from 0.056 to 100/~m) for particle-bound PAH analyses. The MOUDI is a nine-stage cascade impactor with a flow rate of 301itremin -~. Available particle cut-size diameters of MOUDI are 10, 5.6, 3.2, 1.8, 1.0, 0.56, 0.31, 0.18 and 0.1 or 0.056/zm. Each stage on the MOUDI consists of an impaction plate for the nozzle plate above, and a nozzle plate for the impaction plate below (MSP Corporation, 1989). The 47-mm-diameter aluminum circular filter-strips (from Gelman Sciences), which were used for the MOUDI sampler, were first cleaned and extracted with a solvent solution (mixture of n-hexane and dichloromethane, v:v= 1:1) for 24h in a Soxhlet extractor. Silicon grease was then applied to the surface of each filterstrip using the same procedure as for the dry deposition plates. The filter-strips were then allowed to equilibrate in a dust-free desiccator for at least 24 h before weighing on a Sartorius Electrobalance Model MP8-6. The measurable weight range of this electrobalance is 0.01 mg--,42 g and is accurate to a value of 0.01 mg. The strips were weighed before and after sampling to determine the total particle mass collected for each stage. In this study, the MOUDI was operated for a maximum of 16h at the traffic intersection and 24h at the urban and PCI sites, before the filter-strips were replaced.
371
exposed surface area by its average rotational distance. The NRI is described in more detail in Noll and Fang (1986). After combining the MOUDI and NRI, the normalized particle size range for each stage, in sequence, is 0.056,,,0.166, 0.166,,,0.31, 0.31~0.52, 0.52~1.0, 1.0~1.8, 1.8~3.2, 3.2~5.6, 5.6,~10, 10~24.7, 24.7~36.5 and 36.5,,, 100/zm. In this study, the concentration at overlap size ranges measured by both MOUDI (5.6~10/zm and 10,-~18/zm) and NRI (6.5 ,-~ 11.5/zm and 11.5 -~ 24.7/zm) was presented by the averaged data.
Semi-volatile sampler Ambient-air samples for the particle and gas phases of PAHs were collected by using a standard semi-volatile sampling train (General Metal Works PS-1). The sampler with a Whatman glass fibre filter, first cleaned by heating to 450°C, was used to collect particulate and particle phase PAHs. A glass cartridge containing a 5-cm polyurethane form (PUF) plug followed by 2.5 cm of XAD-2 resin and finally a 2.5 cm PUF plug cleaned by sequential extraction was used to collect the gas phase PAHs. The glass fibre filters were weighed before and after sampling to determine the total suspended particulate (TSP) collected.
Noll Rotary Impactor (NRI)
Sampling program
The atmospheric coarse particle mass was measured with a Noll Rotary Impactor (NRI) that collects large particles, which are excluded from other conventional samplers. It is a multi-stage rotary inertial impactor that collects coarse particles by simultaneously rotating four rectangular collectors (stages) of different dimensions through the air. The stages are covered with Mylar strips coated with silicon grease (as for the MOUDI). Total collection areas are 1.2 crn 2, 3.1 cm 2, 10.3 cm 2 and 10.3cm 2 for stages A, B, C and D, respectively. The dimensions for each NRI stage are the same as those used by Holsen and Noll (1992). Mylar film with a thickness of 0.051 mm was used as the collection medium for the NRI, because it has a low enough weight to be weighed on a microbalance and is rigid enough to be handled. Silicon grease was applied to the surface of each mylar strip to collect impacted particles. The mylar strips were then allowed to equilibrate in a dust-free desiccator for at least 8h before weighing on an Electrobalance as was done for the filter-strips of MOUDI. The strips were also weighed before and after sampling to determine the total particle mass collected for each stage. In this study, the NRI operated for a maximum of 4 h at the traffic intersection and 12h at the urban and PCI sites before the mylarstrips were replaced. The NRI was operated at 320 rpm, which produced a theoretical aerodynamic cut-diameters range (assuming a particle density of 1 gcm -3) of 6.5~100, 11.5~ 100, 24.7~100 and 36.5,-,100/zm for stages A, B, C and D, respectively. The volume of air sampled by each stage of the NRI was calculated by multiplying the
Sufficient amounts of material were collected during periods of no rain in each sampling period from the dry deposition plates to determine the mean PAH content for the gas phase, and for each particle size range. The sampling information is shown in Table 1.
PAH analysis After final weighing, the samples were placed in a solvent solution (mixture of n-hexane and dichloromethane, v:v = 1:1), and extracted in a Soxhlet extractor for 24 h. The extract was then concentrated, cleaned up and reconcentrated to exactly 1.0 or 0.5 ml. Since PAHs are thermally stable (they mostly originate from combustion processes) and exhibit low polarity, they can be detected and positively identified by using gas chromatograph/mass spectrometer (GC/MS) techniques. Massselective detection is thus one of the most sensitive spectrometric detection methods for PAH analysis (Gautschi and Mandel, 1994). The identification and quantification of PAHs was accomplished by using a GC (Hewlett-Packard 5890) with a mass selectivity detector (MSD) (Hewlett-Packard 5972A). This GC/MS was controlled by a computer workstation and was equipped with a Hewlett-Packard capillary column (HP Ultra 2-50 mx0.32 mmx0.17/zm), a HP-7673A automatic sampler, injection volume of 1/zl, splitless injection at 300°C, an ion source temperature of 280°C, oven temperature from 50°C to 100°C at 20°Cmin-1; 100°C to 290°C at 3°Cmin-I; and 290°C for 40min. The primary and secondary ion numbers of PAHs were determined by using the scan mode for pure PAH standards. Then, the quantification of PAHs was
372
Hwey-Lin Sheu et al. Table 1. Sampling information
Sampling periods
Exposure time for the dry deposition plate (days)
Sampling volume (m3) MOUDI
NRI A stage
B stage
C stage
D stage
Traffic intersection
1.29 June-2 July 1994 2. 14-17 July 1994 3.23-26 August 1994
4.0 3.9 4.0
235 219 284
178 177 202
535 531 607
1606 1593 1822
1606 1593 1822
6.4 5.0 5.0
376 419 435
349 334 453
1047 1002 1359
3142 3006 4077
3142 3006 4077
5.8 6.1 5.3
483 430 364
307 371 356
921 !113 1068
2763 3339 3204
2763 3339 3204
Urban site
1.28 March-3 April 1994 2.25-30 April 1994 3. 19-25 May 1994 PCI site
I. 5-18 November 1994 2. 26 April-2 May 1995 3.24-30 June 1995
performed by using the selective ion monitoring (SIM) mode. The concentrations of the following PAHs were determined: naphthalene (Nap), acenaphthylene (AcPy), acenaphthene (Acp), fluorene (Flu), phenanthrene (PA), anthracene (Ant), fluoranthene (FL), pyrene (Pyr), cyclopenta(c,d)pyrene (CYC), benz(a)anthracene (BaA), chrysene (Chr), benzo(b)fluoranthene (BbF), benzo(k)fluoranthene (BkF), benzo(e)pyrene (BeP), benzo(a)pyrene (BaP), perylene (PER), indeno(1,2,3,-cd)pyrene (IND), dibenz(a,h)anthracene (DBA), benzo(b)chrycene (BbC), benzo(ghi)perylene (BghiP) and coronene (COR). The GC/MSD was calibrated with a diluted standard solution of sixteen PAH compounds (PAH Mixture610M from Supelco) plus five individual PAH compounds (from Merck). Analysis of serial dilutions of PAH standards showed that the limit of detection for individual PAH compounds was between 23 and 524 pg. (The limit of quantification) LOQ is defined as the limit of detection divided by the sampling volume or the exposed time for the PS-1, MOUDI and NRI sampler, and for the dry deposition plate, respectively. For the PS-1 samplers, the LOQ for individual PAHs was between l0 and 221 pgm -3, whereas those values for dry deposition plates were between 13 and 288pgm-Eday -1. For MOUDI and A, B, C, D stages of NRI, the LOQ of GC/MS for individual PAHs, in sequence, was 15.5,~365pgm -3, 34.4,,~811 pgm -3, 40.7 ~ 958 pgm -3, 13.5~ 319 pgm-3 and 4.51 ~ 106 pg m -3, respectively. Ten consecutive injections of a PAH 610M standard yielded an average relative standard deviation of the GC/MSD integration area of 6.1%, with a range of 4.1 to 9.9%. In this study, two internal standards (phenanthrene-dlo and perylene-dl2) were used to check the response factors and the recovery efficiencies for PAHs analysis. The recovery efficiencies of 21 individual PAHs and these two internal standards were determined by processing a solution containing known PAH concentrations through the same experimental procedure used for the
samples. This study showed that the recovery efficiency of 21 individual PAHs varied between 73.6 and 115% and averaged 85.9%. The recovery efficiencies of two internal standards were between 81.0 and 87.0% and were fairly constant. The blank tests for PAHs were accomplished by using the same procedure as the recovery-efficiency tests without adding the known standard solution before extraction. Analysis of field blanks, including greased strips, glass fibre filters and PUF/resin cartridges, showed no significant contamination (GC/MS integrated area below the detection limit). Analysis of duplicate experiments yielded differences in total-PAH concentration ranging from 5.6 to 11.6% and averaging 8.8% for ambient air samples. Breakthrough tests were investigated by two layers of XAD-2 cartridge. Both upper and lower layers of XAD2 resin were analyzed individually and compared for the PAH mass collected in each layer. Three breakthrough tests were investigated in this study, and no significant PAH mass was found to be collected in the lower layer of XAD-2 resin.
RESULTS AND DISCUSSION Particle size distribution of total P A H s
For the traffic intersection, the particle size distribution (dC/dlogDp versus Dp) of total PAHs was unimodal (Fig. 1A). The major peak was located at the particle size range between 0.31 and 3.2#m, which mostly belonged to the fine particle mode (<2.5#m). This result indicates that the PAH mass in the ambient air of the traffic intersection originated mainly from automobile exhausts. This not only gives some important information about the potential for health hazards, but also indicates that the whole particle size distribution of PAHs essentially refects the gas-to-particle condensation in the fine particle mode after PAHs are emitted to the atmosphere. At the urban and PCI sites, the particle size distributions (dC/dlogDp versus Dp) of total PAHs were
373
P A H content in ambient air
Total-PAHs
i
S00
I TT IS'o t
400
(A)
300
200 100
"o
g.0,
0,1
0.5
O.56
10
00
washout and/or by dry deposition. Because of the resuspension of dust from the ground by the wind and/ or travelling vehicles, some of the PAH-aerosols persist in the atmosphere primarily as these coarse particles. However, when the particle size is greater than 25/zm, the mean value of dC/dlogDp is lower than 12.1 and 3.1ngm -3, for the urban and PCI site, respectively (Fig. 1B and C), because the coarse particles have a proportionally lower surface area than the fine particles.
100
120 e~ 100 t~ I:1
k
1
St]
!
~ "o
(B)
1
20 0 0.01
0.1
0.;06
1
2.5
10
26
100
100 ~0
80
"~ ¢21
o0
"~
20
g.01
1
0.1
0.25
1
2.6
(c)
10
5§
100
Particle Diameter (/zm) Fig. 1. Particle size distribution of total PAH (dC/dlogDp) in the ambient air at the traffic intersection, urban site and PCI site. Error bar represents the standard deviation.
Size distribution of particle-bound tottfl-PAH content The particle size distributions of particle-bound totalPAH content (/.tg g-] versus Dp) are shown in Fig. 2. In general, a smaller particle has a higher total-PAH content. This is due to the fact that soot from combustion sources is primarily made up of fine particles, which carry a high PAH content. In addition, a smaller size of particle has a higher specific surface area, and a higher attachment rate for organic pollutants. In the ambient air, at the traffic intersection, the major peak in particle size distribution for total-PAH content is found in a particle size range between 0.056 and 3.2/~m, with PAH concentrations ranging from 4970 to 5940/~g g-I (average 5510/~g g- l (Fig. 2A). This result reveals that the condensation and accumulation processes of the young aerosols are primarily incorporated into particles smaller than 3.2 ~tm in the ambient air of the traffic intersection. Particles between 3.2 and 100~tm in diameter have a total-PAH content between aooo
d-"
0ooo
I
bimodal (Fig. 1B and C). The bi-peaks of Fig. 1 occur at 0.31~0.52 and 1.8~3.2/~m, and 0.166~0.31 and 5.6~ 10/zm, for the urban, and PCI, sites, respectively. The fine and coarse particle modes were differentiated at 2.5/zm. At the urban sampling site, the major peak was located at the particle size range between 0.31 and 3.2 #m, which mostly belonged to the fine particle mode (particle size less than 2.5/~m). This result indicates that the PAH mass in ambient air of the urban area originated mainly from automobile exhausts. The peak of particle size distribution of PAHs localized at a particle size < 1.0 ~tm essentially reflects the gas-to-particle condensation process in the young aerosols, whereas the peak localized at the particle size range between 1.0 and 3.2 ~tm shows the occurrence of accumulation processes in these aerosols. Particles smaller than 3.0~tm are easily transported through the upper respiratory tract into the bronchioles and alveoli of the lungs and therefore pose a direct health hazard. At the PCI site, the two peaks in the total-PAHs size distributions were located at the fine and coarse particle modes, respectively. The PAHs were originally emitted from sources in the gas phase, adsorbed on to particulates, in part resisting degradation in the environment, and then went through atmospheric transport leading to a wider distribution. PAHs adsorbed on the aged aerosols would be removed from the atmosphere by rain
7ooo
t=O
5ooo
UO
4ooo
(a)
3ooo
g
zooo
0 lOOO
vj 0
10
16
100
4ooo 35oo
0 300O 2500
(s)
2000
I ,",,,4 c:)
1500
o
500
g
4000
(..,
1000
~.01
0.1
O.SO
1
0.8
10
J6
Ioo
3SO0
0 30O0
I
2500
,,,,a
0 *ea ,Ika
Cc)
2000 1500
t,.
c)
n,
lO00 O00 0.1
0.ORS
1
0.6
Particle Diameter
10
O8
100
(~m)
Fig. 2. Particle size distribution of particle-bound total-PAH content (v.gg -1) in the ambient air at the traffic intersection, urban site and PCI site. Error bar represents the standard deviation.
374
Hwey-Lin Sheu et al.
718 and 1920/zg g- 1 (average 1130 #g g- t). The lower total-PAH content of the coarse particles indicates that the aged aerosols are primarily resuspended from road dust and thus contain a greater quantity of inorganic crustal-elements like AI, Ca, Fe and Si and, therefore, have a smaller PAH content on them. For the urban area, two groups with peaks were found in the 0.056,,~ 0.31/zm (mode I) and 0.31 ~ 100/zm (mode II) size ranges (Fig. 2B). Mode I was attributed to primary emissions from combustion sources, whereas mode II was attributed to the accumulation of secondary reaction products on primary aerosol particles. The PAH content of mode I particles ranged between 2370 and 2880 #g g-1 (average 2630 #g g-i), whereas mode II particles ranged between 539 and 1550/xg g-I (average 1090/zg g-1). Aceves and Grimalt (1993) also found that PAHs are predominantly found in the submicron fractions at urban sites, particularly the < 0.5/zm fraction. They indicated that this is likely to be characteristic of early aging stages after gas-phase emission of combustion residues to the atmosphere. The mean values of total-PAH content for the PCI site were lower than those of the urban site. At the PCI sampling site, mode I and the mode II were also found in the 0.056,-~0.31/zm and 0.31~100/zm size ranges (Fig. 2C). The PAH content of mode I particles ranged between 1440 and 1450/zgg -] (average 1445/zgg-]), whereas mode II ranged between 326 and 708/zgg -1 (average 540/zg g-l).
The Urban/Traffic (U/T) and PCI/Traffic (P/T) ratios of particle size distribution for total-PAH content The U/T and P/T ratios are defined as the mean PAH content of the ambient air at urban and PCI sites divided by that at the traffic intersection. These two ratios can serve as the mean PAH dilution factor of urban and PCI air compared with that at the traffic intersection. The major peaks of particle size distribution for U/T and P/T ratios were in the size range between 3.2 and 100#m, with values ranging from 0.60 to 1.07 (mean 0.76) and from 0.21 to 0.72 (mean 0.49), respectively (Fig. 3). However, for the particle size range between 0.31 and 3.2#m, both the U/T and P/T ratios have the lowest values, the U/T ranging from 0.25 to 0.28 (average 0.26), and the P/T ranging from 0.07 to 0.13 (average 0.10). This showed that the ambient air of the traffic intersection had a greater young-aerosol mass, primarily as submicron particles. During the transport process from the combustion source to the urban or PCI atmosphere, and in order to reach a thermodynamic equilibrium, part of the PAH mass shifted from the particle phase to the gas phase again or, during the accumulation process, it aggregated with lower PAH-content particulates, thus resulting in the dilution of particlebound PAH (Lee et al., 1995). Particle size distribution (/~g versus Dp) of individualPAH content The particle size distributions of individual PAHs are shown in Table 2. The overall particle size range covers
0.056-100#m and can be separated into the following four size ranges: 0.056~1.0, 1.0,-~2.5, 2.5--~10, and 10 ~ 100/zm. In general, the measured particle-bound individual PAH contents had their highest peaks at a particle size of less than 1.0 #m. This is probalbly due to the fact that smaller particles have a higher specific surface area and therefore may contain greater amounts of organic carbon which allow more PAH adsorption (Table 2). The mean particle-bound content for higher molecular weight PAHs--BbF, BkF, BeP, BaP, PER, IND, DBA, BbC, and BghiP--in the ambient air at the traffic intersection were 4.7 and 10.2 times higher than those at the urban and PCI site, respectively. The values of dC/dlogDp (ngm -3) for individual PAHs measured at three sampling sites are shown in Table 3. Generally, a particle size smaller than 2.5 #m has a higher value of dC/dlogDp. These results indicate that a very significant fraction of PAH mass was originally formed in the gas phase of the automobile exhaust at a high temperature. Upon entering the ambient air, these PAHs cooled, and most of the higher molecular weight PAHs were incorporated into particulate matter by adsorption and condensation processes (Lee et al., 1995).
Measurements of particle-bound PAH content of the collected dry deposition materials The dry-deposition PAH mass divided by the total mass of dry-deposition gives the particle-bound PAH content 1.1 1.0 0.9 0.8 0.7 0.6
A)
0.5 0
0.4 0.3
.< |
[.. = 0 = °~
0.2 0.1 0,0 0.01 1.1
0.1
0.25
1
;2.5
10
25
100
0.1
O.B5
1
2.5
10
85
100
1.0 0.9 0.8 0.7
0.6 0.5 m
0.4 0.3 0.2 0.1 0.0 0.01
Particle Diameter (~m) Fig. 3. The urban/traffic (U/T) ratio and PCI/traffic (P/T) ratio of particle size distribution for particle-bound total-PAH content.
375
P A H content in ambient air
¢N
t¢5
7~
=I
II
,=
,m
i °11 ~ I
7~ = ! el II
:E [] II
< i
< e~
Hwey-Lin Sheu e t
376
al.
0
7~ 0
~ o o o o o o o o o o ~ o ~ o
. . . .
0
@
0
0
0 °~ ¢xl
0 m
0
°~
0
7~
1
f
0 @
0
0
o. 0
< <
~
~ Omm
2 m~
~' ~mmO
i~
377
P A H content in ambient air 1 t'~,d 2 :v'e
3000
2,o0
I
d'~pltcatod s~mples
1
2000 I
CA)
1500
~L
1000 500
o
.g o
0
Juno-duly 1400 12oo 1000 800
I
"d o
E~ "U g
600
1 2 \ ~ \ xy \ \ N %'% \\\ \\N \\x,
"[X~
\ \ N I(X x \ ' ~
400
~,.\\ K,~
200
N \ \ KX N \ N KX N~\ < x
~ duly_
1994
ApUg~ 2
1400
I
1200
0
1000
I-
BOO
Dry deposition modelling The deposition velocities for the atmospheric particles by the dry deposition model were obtained by the following equations (Noll and Fang, 1989): Vd = //st "[- 1.12 x U* x exp(-30.36/Dp),
April
o
Q,
1904
Because of the 'synergism' of higher wind speed and larger particle MMDs, the dry deposition flux of totalparticle mass at the PCI site was much higher than at the urban site. A particulate with a larger size has less specific surface area and would, therefore, allow less PAHs to be adsorbed on it, so, for the total-particle mass collected by dry deposition, the total-PAH content in the PCI site averaged only 56% of that measured at the urban site (Fig. 4).
M~Ily 4
1994
(1)
where Vst = particle settling velocity (crn s- 1); U* = friction velocity (cm s-l); and Dp = particle diameter (/zm).
1
When Dp_>5 tzm, [c)
SO0
Vst :
1
VT, stock
(2)
400 200 NOV. 1994,
April 19g8
Juno 190fl
VT,stock =
( ,D:pg)/18u,
(3)
Sampling Periods
Fig. 4. Measured particle-bound total-PAH content (/zgg-n) of the collected dry deposition materials sampled at the traffic intersection, urban site and PCI site.
where VT,stock=terminal settling velocity (cms-~); pp=density of particle (1.0 g cm-3); g = gravitational constant (980cm s-2); and /z = absolute viscosity of air (gcm
-I s--J).
W h e n Dp < 5 tzm,
(#g g-I) of the collected dry deposition materials. These values for total-PAHs at the traffic intersection varied between 1730 and 2370/zgg -l (average 2010/xgg -1) (Fig. 4A). For the urban site, the particle-bound totalPAH content of the collected dry deposition materials varied between 993 and 1230/zgg -l (average 1200/xgg -t) (Fig. 4B), whereas for the PCI site, this ranged between 356 and 973/zgg -~ (average: 672/zgg -1) (Fig. 4C). In the ambient air of the traffic intersection, BaP (304/zgg-l), IND (229/zgg-n), PER (223/xgg-l), BbF (218/zgg -t) and DBA (201/zgg -~) had the highest concentrations in the collected dry deposition materials. The values for the remaining PAHs at the urban site varied between 2.03 and 199/zgg -l and averaged 52.2/zgg -I. In the ambient air of the urban site, Ant (136/zgg-]), IND (121/zgg-1), Flu (106/zgg-~), FL (87.4/zgg -t) and Acp (85.2/zgg -l) had the highest amounts of particle-bound PAH in the collected dry deposition materials. The values for the remaining PAHs at the urban site varied between 10.8 and 75.0/zgg -I and averaged 41.3/zgg-k At the PCI site, IND (79.3/zgg-~), DBA (62.8/zgg-~), COR ( 54.2/zg g- 1), Nap (50.6/zg g- l) and BghiP (49.3/~g g- l) were the most important particle-bound PAHs in dry deposited. The values for the remaining PAHs at the PCI site varied between 9.80 and 38.2/zgg -1 and averaged 23.5/zg g-l.
Vst :-- VT,stock x gc,
(4)
where K¢ = 1 + 2~.[1.257 + 0.400 exp(-0.55 Dp/X)]/Dp
~. =/z/[0.499P (8M/ztRT)°'5],
(5)
(6)
where K¢ = cunningham correction; ~.= mean free path (cm); P--absolute pressure (Pa); R--universal gas constant (8.314J-gmol-I°K-1); M = molecular weight (gmol-1); and T-- absolute temperature (°K). Two meteorological parameters that influence atmospheric turbulence are friction velocity (U*) and surface roughness (Z0). The relationship between these parameters for near neutral atmospheric stability condition is: U* Z - d U----k- l n - - ~ 0 ,
(7)
where U=measured mean wind speed at height Z (cms-]); Z=measured height above ground (m);
378
Hwey-Lin Sheu et al.
k = V o n Karman's constant (0.4); d=datum displacement (m); and Z0 = surface roughness height (m). Since wind was measured at only one datum displacement, the average height of the structures surrounding the sampling site of the traffic intersection, urban and PCI areas were taken as 8.0, 8.0 and 5.0 m, respectively. A datum level displacement of (6.4, 6.4 and 4.0 m) (80% of the average structure height) was subtracted from the height (13, 13 and 10m) where the wind speed was recorded for the (traffic intersection, urban and PCI sites), respectively. The roughness coefficient was estimated by the general relationship that Zo is approximately 1/30 of the average roughness height. Since the average height for the area surrounding the sampling site of (traffic intersection, urban and PCI areas) was (8.0, 8.0 and 5.0 m), the roughness heights of (0.27, 0.27 and 0.17 m) were used for calculating friction velocity, respectively. The friction velocities for wind speed measured in this study were calculated by using eqn 7. The modelled Vd,i of each impactor stage at the sampiing site of (traffic intersection, urban and PCI areas) was (0.11, 0.069 and 0.090), (0.81, 0.38 and 0.81), (6.25, 2.58 and 6.25), (14.2, 6.36 and 14.2) and (33.4, 19.9 and 33.4 cm s-1) for the particle size range of 3.2 ~ 5.6 (mean Dp=4.4#m), 5.6~10 (mean Dp=7.8/zm), 10~24.7 (mean Dp = 17.35/zm), 24.7 ~ 36.5 (mean Dp = 30.6 #m) and 36.5~ 100#m (mean Dp=68.25/zm), respectively. For each impactor stage, the variation of Vd.i for each sampling site was fairly constant, because the variation of wind speed was minor during the sampling periods. The ratio of modelled/measured (Md/Ma) dry deposition of particle-bound content
The calculated total dry-deposition flux (Ft) for total particle mass (mg m -2 day-m), total PAHs or individual PAHs (/zgm -2 day-l), was calculated from: 11
Ft = Fg + Z
Fi
(8)
i=l 11
= Cg. Vd,g'q-~f~fi. Vd, i,
bound individual-PAH content was the Ft of individual PAHs divided by the Ft of the total particle mass. The calculation procedure divided the combination of gas phase and particle distributions into 12 intervals and assigned a modelled deposition velocity to each interval. The calculated flux for each interval was then summed to calculate the total dry-deposition flux. After the calculated particle-bound PAH content of the collected dry deposition materials for total PAHs or individual PAHs was obtained, the calculated data were compared to the measured data. The Vd.g of semivolatile organics is controlled mainly by diffusion. Its magnitude is close to or smaller than the measured dry deposition velocity of naphthalene (Nap), which usually has more than 99% of mass existing in the gas phase. Here, selected a Vd,g of 0.01 cm s- i was used for the gas-phase PAHs (Sheu et al., 1996). The modelled/measured (Md/Ma) ratio of particlebound total-PAH content of dry deposition materials varied between 0.86 and 1.18 at the traffic intersection, between 0.88 and 1.22 at the urban site and between 0.77 and 0.99 at the PCI site (Table 4). All the modelled and measured dry deposition of particle-bound content data for total PAHs were within 33.8%. The Md/Ma ratio of particle-bound individual-PAH content of dry deposition materials ranged from 0.32 to 1.69, from 0.31 to 1.64 and from 0.40 to 1.54 for the traffic intersection, urban and PCI sites, respectively (Table 5). By using the particle size distribution data, in general, this dry deposition model can provide a good prediction for the particle-bound PAH content of dry deposition materials. Cumulative fraction (F%) of dry deposition of particlebound individuai-PAH content
In general, a lower molecular weight PAH had a larger fraction of dry deposition flux contributed by the gas phase. They were (50.4, 46.3 and 28.4%) for 2-ring PAH, (15.2, 15.4 and 11.7%) for 3-ring PAHs and (13.0%, 3.60% and 5.01%) for 4-ring PAHs measured at the (traffic intersection, urban and PCI site), respectively. For higher molecular weight PAHs--5-ring,
(9)
i=1
where Fg = dry-deposition flux contributed by the gasphase total PAHs or individual PAHs (/zgm-Zday-I); Fi =dry-deposition flux contributed by each impactor stage which is from stage 1 (0.056~0.166/zm) to stage 11 (36.5~100/zm), respectively; Cg=the measured ambient concentration (ng m -3) of total PAHs or individual PAHs; Vd,g= the dry deposition velocity of gasphase PAHs; Ci =measured concentration (/xgm -3) of total particle mass, total PAHs or individual PAHs in each impactor stage; Vd.i (cms-I)=calculated dry deposition velocity by using the dry deposition model (Noll and Fang, 1989) for the mean particle diameter (/zm) of each impactor stage. By using the Ft, the calculated particle-bound totalPAH content was the Ft of total PAHs divided by the Ft of the total particle mass, and the calculated particle-
Table 4. Ratios of modelled/measured (Md/Ma) particle-bound totaI-PAH content of the collected dry deposition materials sampled at the traffic intersection, the urban site and the PCI site
Sampling periods Traffic intersection June--July 1994 July 1994 August 1994 Urban site March 1994 April 1994 May 1994 PCI site November 1994 April 1995 June 1995
Ratio
A. 1.15 A. 1.12
B. 0.93 B. 1.07 B. 1.18
A. 1.04 A. 1.09 A. 0.88
B. 1.03 B. 1.22 B. 0.91
A. 0.99 A. 0.81 A. 0.77
B. 0.92 B. 0.77 B. 0.87
A. 0.86
A and B are duplicated samples.
P A H content in ambient air
379
Table 5. Ratios of modelled/measured (Mfl/Ma) purticle-bound individuaI-PAH content of the collected dry deposition materials sampled at the traffic inter~,ction (T), urban site and PCI site PAHs
Ratio
NaP AcPy Acp Flu PA Ant FL
PAHs
T
Urban
PCI
1.04 0.67 0.57 0.38 0.59 1.69 0.69
1.61 1.08 0.69 0.41 0.36 0.31 0.32
0.51 0.66 0.65 0.68 0.47 0.76 0.75
Ratio T
Pyr CYC BaA CHR BbF BkF BeP
0.51 0.71 0.47 0.87 1.51 1.83 0.48
6-ring and 7-ring P A H s - - t h e i r F % of the dry deposition flux contributed by the gas phase was lower than 3.26% (Figs 5-7). In ambient air of both urban and PCI sites, generally, the profiles of cumulative percent of modelled drydeposition flux are similar with those at the traffic intersection (Figs 5-7). The cumulative percentage of modelled dry-deposition flux (F%) based on the totalPAH mass (summation of 21 individual PAHs), more than 92.4, 91.9 and 87.5% of PAH dry deposition flux, is contributed by particulates with diameters larger than 10/zm in the ambient air of the traffic intersection, urban and PCI sites, respectively. Even though PAHs existed mainly in the gas phase and/or fine particulates ( < 2.5/zm), these results indicate that PAH dry deposition is mainly due to the coarse particulates, particularly
1 O0
l Traffic
PAHs
Urban
PCI
1.13 1.56 1.38 1.24 0.94 1.12 1.77
T
0.52 0.40 0.78 1.48 1.14 1.54 0.59
BaP PER IND DBA BbC BghiP COR
1.07 1.36 1.01 1.01 0.94 1.83 1.00
PCI
1.50 1.59 0.50 1.64 1.63 1.40 1.62
0.46 1.41 1.5l 0.70 0.61 1.24 0.66
the particulates larger than 10/zm from gravitational settling. The atmospheric dry deposition of PAHs due to the fine particulates ( < 2.5 tzm) by the initial impaction was very minor. High particle-bound PAH content of dry deposition materials resulting from the coarse particulates may undergo bioaccumulation and result in significant adverse health effects. Particle-bound individual-PAH content (/~gg-l) at the modelled MMI)v and of the measured data The mass median diameter of dry deposition materials (MMDF) was obtained from the curve of cumulative percentage of the modelled PAH dry-deposition flux in Figs 5-7 and chosen for the particle diameter at 50% of the cumulative curve (dso). In the ambient air of the traffic intersection, the MMDF was 39.1/zm for total
100 o--o--
5 0. 4%
o--o--o--, Z-ring
U ~ b f/, _n.-~
PAH
046"3g 152%
3-ring v- -
13.07.
10
Urban
]
o
o
Ratio
2-ring
.0/
PAH
O__ 0__0__0_ 0_0_0_ 0----"0"~
PAHs v-- v-- v --
v
4-ring PAHs
0--0--0--@--Q--0~
"~, o
lO
va.O0g
4-ring PAHs V/v ~ / v--v--v--v--v~
gas phase gas
/
phase
/
o .87Z
0.297*
6-ring
PAHs
~/ o~ a / " 5 - r i n g PAHs
2
///
o/ / /
•
/
J
0 147. ~o
a
7-ring
~
2
PAH
0.1
~o 0.001
0.01
0.1
1
10
0 , I 9%
~IIO'OBT~
0.1
5-ring PAHs v / '/ v__v~V / 7-ring
PAH
m/
100 ........
Particle Diameter (p,m) Fig. 5. Cumulative fraction (%) of modelled dry deposition flux for PAHs size distribution in the ambient air of traffic intersection.
0.001
i 0.01
........
Particle
i 0.1
........
Diameter
i 1
........
i 10
........
i
100
(~m)
Fig. 6. Cumulative fraction (%) of modelled dry deposition flux for PAHs size distribution in the ambient air of urban site.
Hwey-Lin Sheu et al.
380
m 0I
@
0
~
~
-
~o
~
-
-
-
~
~
~
. . . . . .
N °--
r~
~°~ @
f
@
"~,
~8
!
III
h~ °~
&
[,.
< a~
<
P A H content in ambient air
1 O0
028"'" r~
o
O--@--O--O--O--O~Q /
10
Z"
/~5.01~
"U t~
/
gag phage t~
v0"47~g
5-ring PAHs v__v__v~V ~v
r~
/
/
8
r~
./v//
0.102 0.1
0
........ 0.001
i
0.01
. . . . . . . .
6-ring PAHs 0 / O--O ~°/
i
0.1
. . . . . . .
i
1
Particle Diameter
. . . . . . . .
)
10
. . . . . . . .
i
100
(g.m)
Fig. 7. Cumulative fraction (%) of modelled dry deposition flux for PAHs size distribution in the ambient air of PCI site. PAHs and was between 25.3 and 49.6/zm for individual PAHs; the values of particle-bound total-PAH content at MMDF were 1196/xgg -1 and were between 2.11 and 288/zgg -l for individual PAHs (Table 6). In the ambient air of the urban site, the MMDF was 38.2/zm for total PAHs and was between 27.6 and 43.9/zm for individual PAHs. The values of particle-bound totalPAH and individual PAH content at MMDF were 613tzgg -l and were between 6.92 and 86.3/zgg -l, respectively (Table 6). In the ambient air of the PCI site, the MMDF was 33.8/zm for total PAHs and was between 19.1 and 41.9#m for individual PAHs. The values of particle-bound total-PAH and individual PAH content at MMDF were 486#gg -I and were between 1.73 and 103/zgg -1, respectively (Table 6). In general, most of the MMDF for individual PAHs are larger than 25.0/zm. This is due to the fact that the PAH drydeposition primarily resulted from the coarse particulates (> 10 #m) by gravitational settling.
CONCLUSION In general, smaller particles have a higher PAH content. This is due to the fact that the soot from the combustion sources is primarily a fine particulate, which carries a high PAH content. In addition, smaller particles have a higher specific surface area and a higher attachment rate for organic pollutants. It may therefore contain a greater amount of organic carbon, which allows more PAH adsorption.
381
In the particle size range between 0.31 and 3.2/zm, both U/T and P/T ratios of particle-bound total-PAH content have the lowest values. This indicates that during the accumulation process, the PAH mass is shifted from the particle phase to the gas phase or that aggregation of particulates with a lower PAH content results in a reduction in particle-bound PAH content. By using the particle size distribution data, the dry deposition model used in this study can provide a good prediction for the PAH content of the collected dry deposition materials. Lower molecular weight PAHs had a large fraction of the dry deposition flux contributed by the gas phase. For higher molecular weight PAHs (5-ring, 6-ring and 7-ring PAHs), F% of dry deposition flux contributed by the gas phase was lower than 3.3%. At the traffic intersection, urban and PCI site, the MMDF of individual PAHs was between 25.3 and 49.6#m, between 27.6 and 43.9#m and between 19.1 and 41.9/zm, respectively. In general, most of MMDF for individual PAHs are larger than 25.0ttm. The cumulative percentage of modelled dry-deposition flux (F%) based on the total-PAH mass (summation of 21 individual PAHs), more than 92.4, 91.9 and 87.5% of PAH dry deposition flux is contributed by particulates with diameters larger than 10#m in the ambient air of the traffic intersection, urban and PCI sites, respectively. These results revealed that the majority of dry-deposition PAH mass resulted primarily from the coarse particulates (> 10/zm) by gravitational settling.
ACKNOWLEDGEMENTS This research was supported by funds from the National Science Council, Taiwan, Grant No. NSC 86-2113-M006-018. The authors also thank Mr. Chun-C/aing Su, Mr. Yi-Chin Fan and Mr. How-Ran Chao for their help in sampling.
REFERENCES Aceves, M. and Grimalt, J. O. (1993) Seasonally dependent size distributions of aliphatic and polycyclic aromatic hydrocarbons in urban aerosols from densely populated areas. Environmental Science and Technology 27, 2896-2908. Baek, S. O., Field, R. A., Goldstone, M. E., Kirk, P. W., Lester, J. N. and Perry, R. (1991) A review of atmospheric polycyclic aromatic hydrocarbons: sources, fate and behavior. Water, Air, and Soil Pollution 60, 273-300. Benner, B. A., Gordon, G. E. and Wise, S. A. (1989) Mobile sources of atmospheric polycyclicaromatic hydrocarbons: a roadway tunnel study. Environmental Science and Technology 23, 1269-1278. Bidleman, T. F. and Foreman, W. T. (1987) Vapor-particle partitioning of semivolatile organic compounds. In The Chemistry of Aquatic Pollutants, ed. R. Hites and S. J. Eisenreich. ACS Advances in Chemistry Series, ACS, New York. Bjerseth, A. and Ramdahi, T. (1985) Source and emissions of PAH. In Handbook of Polycyclic Aromatic Hydrocarbons, p. 1. Marcel Dekker, New York.
382
Hwey-Lin Sheu et al.
Duce, R. A. and Gagosian, R. B. (1982) The input of atmospheric n-Ci0 to n-C30 alkanes to the ocean. Journal of Geophysics Research 87, 7192-7200. Eisenreich, S. J., Looney, B. B. and Thornton, L. D. (1981) Airborne organic contaminants in the Great Lakes ecosystem. Environmental Science and Technology 15, 3038. Gautschi, P. and Mandel, F. (1994) The detection of polynuclear aromatic hydrocarbons using the HP 5972A MSD. Hewlett-Packard Application Note, 23-5962-8614E. Holsen, T. M. and Noll, K. E. (1992) Dry deposition of atmospheric particles: application of current models of ambient data. Environmental Science and Technology 26, 1807-1815. Lee, W. J. (1991) The determination of dry deposition velocities for ambient gases and particles. Ph.D. thesis, Illinois Institute of Technology, Chicago, IL, pp. 85-138. Lee, W. J., Wang, Y. F., Lin, T. C., Chen, Y. Y., Lin, W. C., Ku, C. C. and Cheng, J. T. (1995) PAH characteristics in the ambient air of traffic-source. Science of the Total Environment 159, 185-200. McCready, D. I. (1986) Wind tunnel modeling of small particle deposition. Aerodynamics Science and Technology 5, 301-312.
MSP Corporation (I 989) Micro-orifice uniform deposit impactor. User's Instruction of MOUDL pp. 2-3. Noll, K. E. and Fang, K. Y. P. (1986) A rotary impactor for size selective sampling of atmospheric coarse particles. Proceedings of the Air Pollution Control Association 79th Annual Meeting, Minneapolis, Minnesota, Paper No. 8640.2. Noll, K. E. and Fang, K. Y. P. (1989) Development of a dry deposition model for atmospheric coarse particles. Atmospheric Environment 23, 585-594. Noll, K. E., Fang, G. C. and Kenneth, Y. P. (1989) Development of a dry deposition model for atmospheric coarse particles. Atmospheric Environment 23, 585-594. Pierce, R. C. and Katz, M. (1975) Dependency of polynuclear aromatic hydrocarbon content on size distribution of atmospheric aerosols. Environmental Science and Technology 9, 347-353. Sheu, H. U, Lee, W. J., Su, C. C., Chao, H. R. and Fan, Y. C. (1996) Dry deposition of polycyclic aromatic hydrocarbons in ambient air. Journal of Environmental Engineering ASCE 122, 1101-1109. Windsor, J. G., Jr and Hites, R. A. (1979) Polycyclic aromatic hydrocarbons in Gulf of Maine sediments and Nova Scotia soil. Geochimica Acta 43, 27-33.