Environ Sci Pollut Res DOI 10.1007/s11356-015-5355-0
RESEARCH ARTICLE
Characterization and risk assessment of polycyclic aromatic hydrocarbons (PAHs) in urban atmospheric Particulate of Tehran, Iran Mohammad Hoseini 1 & Masud Yunesian 1,2 & Ramin Nabizadeh 1,2 & Kamyar Yaghmaeian 1,3 & Reza Ahmadkhaniha 4 & Noushin Rastkari 2 & Saeid Parmy 1 & Sasan Faridi 1 & Ata Rafiee 1 & Kazem Naddafi 1,2
Received: 30 May 2015 / Accepted: 1 September 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract In this study, atmospheric concentrations of particulate-bound polycyclic aromatic hydrocarbons (PAHs) in Tehran megacity were determined to investigate the concentration, distribution, and sources of PAHs in PM10. The health risk from exposure to airborne BaPeq through inhalation pathway was also assessed. Toxic equivalency factors (TEFs) approach was used for quantitative risk estimate, and incremental lifetime cancer risk (ILCR) was calculated. PM10 samples were collected at ten sampling locations during the summer 2013 and winter 2014 by using two independent methods of field sampling. The PM10 concentration in winter (89.55±15.56 μg m−3) was 1.19 times higher than that in summer (75.42±14.93 μg m−3). Sixteen PAHs were measured
with the total average concentrations of PAHs ranged from 56.98±15.91 to 110.35±57.31 ng m−3 in summer and from 125.87±79.02 to 171.25±73.94 ng m−3 in winter which were much higher than concentrations measured in most similar studies conducted around the world. Molecular diagnostic ratios were used to identify PAH emission sources. The results indicated that gasoline-driven vehicles are the major sources of PAHs in the study area. Risk analysis showed that the mean and 90 % probability estimated inhalation ILCRs were 7.85 × 10−6 and 16.78 × 10−6, respectively. Results of a sensitivity analysis indicated that BaP concentration and cancer slope factor (CSF) contributed most to effect on ILCR mean.
Responsible editor: Hongwen Sun * Kazem Naddafi
[email protected]
Sasan Faridi
[email protected]
Mohammad Hoseini
[email protected]
Ata Rafiee
[email protected]
Masud Yunesian
[email protected] Ramin Nabizadeh
[email protected] Kamyar Yaghmaeian
[email protected]
1
Department of Environmental Health Engineering, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran
2
Center for Air Pollution Research (CAPR), Institute for Environmental Research (IER), Tehran University of Medical Sciences, Tehran, Iran
3
Center for Water Quality Research, Institute of Environmental Research (IER), Tehran University of Medical Sciences, Tehran, Iran
4
Department of Human Ecology, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran
Reza Ahmadkhaniha
[email protected] Noushin Rastkari
[email protected] Saeid Parmy
[email protected]
Environ Sci Pollut Res
Keywords Air pollution . Polycyclic aromatic hydrocarbons . PAHs . Health risk assessment . Monte Carlo simulation . Sensitivity analysis
Introduction Polycyclic aromatic hydrocarbons (PAHs) are a large group of complex organic chemicals, mostly formed during the incomplete burning of coal, oil and gas, or other organic substances (Wang et al. 2011; Wu et al. 2014). PAHs, which contain only carbon and hydrogen and constituted by two or more benzene rings in linear, angular or cluster structures, are major toxic and ubiquitous environmental contaminants and mainly found in fine particulate matter in the urban atmosphere due to increasing vehicular traffic (Jamhari et al. 2014; Liaud et al. 2015). Due to their well-characterized potential for human exposure and documented toxicity, 16 PAH species, some of which are considered possible or probable human carcinogens, have been listed by the US Environmental Protection Agency (US EPA) as priority pollutants (Jamhari et al. 2014). Because of their toxicity, mutagenicity, carcinogenicity, and endocrine-disrupting potential and their widespread presence which results in exposure in the general population, PAHs have recently received considerable attention (Armstrong et al. 2004; Li et al. 2008; Okuda et al. 2010). Several PAHs, including benzo(a)antheracene, benzo(b)fluoranthene, benzo(k)fluoranthene, and benzo(a)pyrene, associated with particulate matter have been shown to be indirect-acting mutagens (Chen and Liao 2006). Among them, benzo[a]pyrene (BaP) is a well-known human carcinogen and has always been selected as an indicator of carcinogenic PAHs because concentrations of other PAHs in the urban atmosphere are highly intercorrelated (Norramit et al. 2005). The PAH emission profile in an urban area depends on the processes producing the PAHs (Manoli et al. 2004). The higher concentration of PAHs in air of urban area due to motor vehicles was reported by many studies (Bahry et al. 2009; Okuda et al. 2002); however, several other studies indicated the contribution of other sources such as biomass burning and coal combustion toward the amount of PAHs in ambient air (Fang et al. 2010; Omar et al. 2007; Wan et al. 2006). PAH diagnostic ratios may be used to identify contributions of different emission sources (Tobiszewski and Namieśnik 2012). Many researchers have measured and reported the atmospheric concentrations of PAHs in different geographical locations around the world, e.g., Santiago, Chile (del Rosario Sienra et al. 2005), Athens, Greece (Mantis et al. 2005), Dalian, China (Wan et al. 2006),
Bursa, Turkey (Esen et al. 2008), Delhi, India (Sarkar and Khillare 2013), Kuala Lumpur, Malaysia (Jamhari et al. 2014), etc. There are, however, limited comprehensive reports on concentration, distribution, source identification, and potential health risk assessment of PAHs in Tehran, Iran. The aim of this study was to investigate PAHs abundance and speciation in particulate matter (PM10) collected from ten sites in Tehran, Iran, to characterize potential sources of PAHs, and to estimate the health risk of PAHs in aerosols.
Materials and methods Study area and sampling sites Tehran, the capital of Iran and one of the most crowded area in Iran and in the Middle East, is located in a fairly restricted basin on the southern foothills of the Alborz Mountains (35° 34–35° 50′ N and 51° 08–51° 37′ E). The city has an area of 730 km2 with 11 million habitants, with a density of ∼11,000 persons km−2 (Kamani et al. 2014). Annual average temperature is between −7.4 and 38.7 °C with the mean temperature of 18 °C. The average yearly rainfall varies from 245 to 316 mm. The urban area is surrounded on the north, northwest, east, and southeast by mountains which range from 1000 to 4000 m above sea level. Air circulation is therefore limited and during the cold season from October to March; occasional temperature inversions make the pollution situation worse (Naddafi et al. 2012). Samples of urban PM10 were collected at 10 of the 21 monitoring stations belonging to Tehran municipality, technically operated by the subsidiary Air Quality Control Company (AQCC) (Fig. 1) at which criteria pollutants (e.g., O3, NOx, CO, SO 2 , PM 10 , and PM 2.5 ) were routinely measured (Goudarzi et al. 2009). The PM10 measurements are carried out with beta-attenuation monitors (Environnement S.A., type MP 101 M or Horiba models) equipped with size selective air intakes by collecting PM on fiberglass tape and measuring the amount of radiation that a sample absorbs when exposed to a radioactive source. The BAMs operate at 16.7 L min−1, and particle deposits are focused on 16-mm diameter spots. The tapes automatically advance once every hour. So, 24 spots are formed each day on the tapes. The distances between BAM deposit spots formed each day are 3.2 cm which were taken and used to determine blank values for subtraction. The AQCC records the time and date of placing tapes in the BAMs, but it does not date-stamp the tape. The tapes were collected after sampling duration, which was about 3 weeks, and transferred to laboratory for analysis. The tapes were unrolled in the laboratory, and the spots were counted from the end to find the regions of the predetermined sampling dates as described by Watson et al. (2012). Beside, a SKC
Environ Sci Pollut Res
Fig. 1 Location map of the study area to show the sampling sites
Flite 2 Air Sampling Pump (SKC, USA) equipped with a sampling head and PM10 size-selective inlet was used in parallel for sampling at four stations (Table 1) to compare the results obtained by collecting AQCC monitor (AQCCMs) tapes with the standard sampling procedures for assessing the interchangeability of two field sampling methods.
were removed from the sampler, folded with the particulate matter inside, wrapped with aluminum foils, transported to the laboratory, and stored in a freezer (−18 °C) until the analysis in order to prevent analyte decomposition and volatilization.
Sample extraction and analysis Sample collection Ambient air samples were during the summer 2013 and winter 2014. The SKC sampling pump used in this study was equipped with a PTFE filter (47-mm ID, 0.5-μm pore size) with an air flow rate of 16.6 l/min. The duration of sampling was a consecutive 24 h on a biweekly base (19/07/2013, 1/08/ 2013, 14/08/2013, 27/08/2013, 15/01/2014, 28/01/2014, 10/ 02/2014, and 24/02/2014), and a total of 112 samples were collected during the study period. Field blanks for every sampling period were taken in order to consider the background contamination. For each sampling period, three blank filters were exposed to the air during the whole sampling period and then were treated with the same method as sample during storage and chemical analysis. None of the target PAHs were detected in blank samples. After sampling, the loaded filters
Before the extraction and analysis, filters were removed from the freezer and stored in a desiccator for 24 h. The extraction procedure for filters acquired from SKC sampler was modified from Omar et al. (2006) and Jamhari et al. (2014). Briefly, half of each filter was cut into small pieces. Regarding the filters acquired from AQCCMs, after unrolling the tapes in a laminar flow hood in the laboratory, all deposit spots were removed from the tapes, and each 24 spots belonging to one sampling day were stored in an individual vial and assigned an ID number (Watson et al. 2012). After preparing, the filters were extracted using ultrasonic agitation (30-min sonication time) using 10 mL of a dichloromethane (DCM)/methanol mixture (3:1 v/v) as solvent. The process was repeated two to three times, and the extracts were combined. The combined extract was syringe-filtered through 0.22-μm PTFE filters (Jet Biofil) for the removal of any remaining insoluble particles.
Table 1
Description of four stations sampled by SKC Air Sampling Pump
Stations
Description
Poonak Sharif University Tarbiat Modarres Tehransar
Located in a public park without direct influence of anthropogenic emission sources. Located in backyard of a university away from traffic pollution sources. Near an educational building, about 50 m away from the next major street. Sited in an urban background location on the premises of an administrative building.
Environ Sci Pollut Res
The solution was evaporated and finally reduced to 1 mL using a gentle stream of nitrogen (5–10 psi, gradually increasing during evaporation) and a water bath (40 °C) and kept refrigerated in GC vials until analysis. Identification and quantification of 16 priority PAHs including naphthalene (Nap), acenaphtylene (Acy), acenaphthene (Ace), fluorene (Flu), phenanthrene (Phe), anthracene (Ant), anthracene (Fl), pyrene (Py), benzo(a)anthracene (BaA), chrysene (Chry), benzo(b)fluoranthene (BbF), benzo(k)fluoranthene (BkF), benzo(a)pyrene (BaP), indeno(1,2,3-cd)pyrene (IcP), dibenz(ah)anthracene (DBahA), and benzo(ghi)perylene (BghiP) were carried out using a gas chromatography (GC, Agilent 7890N, Agilent Co.) with mass spectrometry (MS, Agilent 5975C, Agilent Co.). A 30 m×0.25 μm×0.5 μm fused silica (DB-5MS) capillary column (Agilent Co., USA) in selected ion monitoring (SIM) mode was used. The injection volume was 3.0 μL in the splitless mode at 290 °C, and helium with a purity of 99.99 % was used as a carrier gas at constant flow rate of 1 mL/min. The column temperature program was as follows: initial temperature of 60 °C, increased at a rate of 10 °C/min to 100 °C, then held for 1 min, raised to 285 °C at 4 °C/min, and held for 15 min. Quality control Quality control was conducted by preparing and analyzing both field and laboratory blank samples. The analytical blank values for each target compounds were significantly lower than those of samples. Recovery efficiency was determined by spiking a predetermined amount of the standard mixture of 16 PAHs onto the filters and performing the same analytical methods. The recoveries ranged from 71 to 104 and 60 to 114 % for PTFE filters and fiberglass tapes, respectively. Appropriate corrections were made to the measured concentrations. The limit of detection (LOD) and limit of quantification (LOQ) were calculated as the concentrations equivalent to three and ten times the noise of the quantifier ion for a blank sample. The LOD and LOQ of each PAHs are reported in Table 2. All glassware used during extraction steps was ultrasound-cleaned, soaked in 20 % nitric acid (overnight soaking), and rinsed with distilled water prior to being baked in an oven (5 h at 180 °C) to volatilize and remove any organic contaminants. Cancer risk assessment Numerous organizations and expert bodies have offered advice on the risk assessment of PAHs through different exposure pathways. In this study, toxic equivalency factors (TEFs) approach was used for quantitative risk estimate from inhalation pathway. For cancer risk assessment, 16 target PAHs were ranked according to their cancer potency relative to BaP. BaP
Table 2 Limit of detection (LOD) and limit of quantification (LOQ) for different PAH spices Compounds
LOD (ng m−3)
LOQ (ng m−3)
Naphthalene
0.021
0.071
Acenaphtylene Acenaphthene
0.010 0.062
0.034 0.247
Fluorene
0.052
0.173
Phenanthrene
0.080
0.272
Anthracene Anthracene
0.031 0.031
0.103 0.103
Pyrene
0.083
0.290
Benzo(a)anthracene Chrysene
0.073 0.104
0.256 0.347
Benzo(b)fluoranthene Benzo(k)fluoranthene
0.104 0.104
0.347 0.347
Benzo(a)pyrene
0.094
0.313
Dibenz(ah)anthracene Benzo(ghi)perylene
0.021 0.073
0.071 0.256
Indeno(1,2,3-cd)pyrene
0.062
0.247
equivalents (BaP-eq) were calculated by multiplying each individual PAH concentration with its corresponding TEF (Ramírez et al. 2011). Risks of additional cancers were estimated by applying the incremental lifetime cancer risk (ILCR) model. The ILCR was calculated by multiplying the lifetime average daily dose (LADD) by the BaP slope factor. The lifetime was split up into three periods including infant (0– 1 year), children (2–18 years), and adults (19–70 years). The overall LADD was calculated by summing the LADD values of above three age groups. The equations used for estimating LADD and ICLR were as follows: C IR E F ED BW AT ( 1 ) BW 3 ILCR ¼ LADD CS F cf 70
LADD ¼
ð1Þ ð2Þ
where C is the concentration of BaPeq in air (ng m−3) which was obtained by converting concentrations of PAHs according to toxic equivalents of BaP using the toxic equivalency factors, IR is the air inhalation rate (m3 day−1), EF is the exposure frequency (day year−1), ED is the exposure duration (years), BW is the body weight (kg), AT is the averaging lifetime for carcinogens (days), CSF is the cancer slope factor (mg kg−1 day−1), and cf is the conversion factor (10−6) (Moya et al. 2011). In risk assessment, addressing uncertainties is necessary in order to avoiding inaccurate or biased estimates (Qu et al. 2015). In order to quantify this uncertainty, the Monte Carlo simulation technique was implemented in this study. Table 3 shows the selected parameters used for uncertainty analysis.
Environ Sci Pollut Res Table 3
Risk parameters for different age groups used for Monte Carlo simulation
Parameters
Symbol
Age Body weight
BW
Units
Infants
Children
Year
0–1
2–18
19–70
kg
9.1±1.25 5.36
29.70±5.62 11.41
71.05±13.60 15.73
3
−1
Adults
References
(Burmaster and Crouch 1997) (Moya et al. 2011)
Inhalation rate
IR
m day
Exposure frequency
EF
Days year−1
350
350
350
Exposure duration
ED
Year
0–1
0–17
0–52
(Moya et al. 2011)
Averaging time
AT
Days
25,550
25,550
25,550
(Moya et al. 2011)
Assessing agreement between two independent methods of field sampling The average concentrations of low molecular weight (LMW) PAHs including Nap, Acy, Ace, Flu, Phe, and Ant and high molecular weight (HMW) PAHs (other PAHs including Fl, Py, BaA, Chry, BbF, BkF, BaP, IcP, DBahA, and BghiP) measured on PM10 samples collected by SKC pump were 76.44± 26.24 ng m−3and 65.13±27.61 ng m−3, respectively. The corresponding concentrations measured by collecting AQCCM fiberglass tapes were 59.01±27.62 and 56.90±32.83 ng m−3, respectively. Figure 2 shows scatter plots of PAH concentrations measured on PM10 samples collected by SKC vs. AQCCMs. Figure 3 shows the difference between concentrations of 16 PAH species collected by SKC pump and AQCC monitors for LMW and HMW PAHs. As shown in this figure, a higher difference between concentrations was observed for LMW PAHs. This higher difference can be due to volatilization of LMW PAHs during the long time between sampling and collection of filter tapes which was 3 to 4 weeks. The intraclass correlation coefficient (ICC) was used as a statistic for assessing agreement between two independent
a PAHs concentration collected by AQCC (ng m-3)
Results and discussion
methods of field measurement. The ICC values for LMW, HMW, and total PAHs are shown in Table 4. There are no universally standard ICC values representing adequate agreement, but the following values have been suggested to aid interpretation: ICC 0.80 ‘almost perfect agreement’ (Fleiss 2011; Montgomery et al. 2002). As shown in Table 4, ICC for HMW was 0.868 indicating almost perfect agreement. For LMW and total PAHs, there were substantial agreements between two methods of field measurement. These high ICC values indicate that the sampling campaign
25
20 y = 1.0691x + 1.4489 R² = 0.8843
15
10
5
0
0
5
10
15
20
25
PAHs concentration collected by SKC pump (ng m-3)
b PAHs concentration collected by AQCC (ng m-3)
The CSFs of B[a]P for inhalation pathway are obtained from the open published literature. Chen and Liao (2006) have averaged and log-transformed the different CSF values estimated by Collins et al. (1991) appropriately to a lognormal distribution with a geometric mean of 3.14(mg kg−1 day)−1 and a geometric standard deviation of 1.80. In this study, it also was estimated that residents in age-specific group are exposed for 350 days a year during their life span. To ensure the stability of results, simulations were run for 5000 iterations with each parameter sampled independently (Chen and Liao 2006). Sensitivity analysis was performed to determine which probability density functions have the greatest effect on the risk estimates. The Monte Carlo simulation and sensitivity analysis were implemented using @Risk software (version 6.0; Palisade Corporation; NY, USA).
25
20 y = 1.1074x + 1.9859 R² = 0.7572
15
10
5
0
0
5
10
15
20
25
PAHs concentration collected by SKC pump (ng m-3)
Fig. 2 Scatter plots of HMW (a) and LMW (b) PAHs concentrations measured on PM10 samples collected by SKC vs. AQCCMs
Environ Sci Pollut Res
Outlier Value Q3+1.5×IQR Q3 Interquartile Range (IQR) Q1 Q1-1.5×IQR
Fig. 3 Plot of difference between concentrations of LMW and HMW PAHs collected by SKC pump and AQCC monitors
by using SKC pumps, which is an expensive and timeconsuming procedure, can be replaced by collecting fiberglass tapes from AQCCMs, especially when the sampling is intended only for measuring HMW PAHs. PM10 and PAHs concentration The average concentrations of 16 PAH species, total PAHs, and PM10 in different stations are shown in Table 5. The highest 24-h concentration of PM10 (135 μg m−3) was measured in Fath station, and the lowest was in Poonak station (47 μg m−3). The average PM10 concentration over all ten sampling sites in winter (89.55±15.56 μg m−3) was 1.19 times higher than that in summer (75.42±14.93 μg m−3). As represented in Table 5, the average concentrations of the total PM10-bound PAHs (sum of the 16 PAH species) varied at different sampled stations and ranged from 97.49 ng m−3 at Poonak station to 137.14 ng m−3 at Shad Abad station. The differences in concentration of PAHs among the sampling sites could be due to characteristics of sampling sites. The Shad Abad station is located adjacent to a major street with high traffic volume. It has been well known that vehicle exhausts especially gasoline and diesel vehicles are the major source of PAHs. On the other hand, Poonak station is situated in a public park area far away from traffic and other anthropogenic emission sources. A significant correlation was observed between PM10 and total PAHs (r=0.82, p
