Health Risk of Ambient PM10-bound PAHs at Bus Stops in Spring and Autumn in Tianjin, China

A study was conducted to measure ambient concentrations of PM10 and PM10-bound PAHs to estimate health risks due to exposure to PAHs during wait times at bus stops in September 2012 and March 2014. Samples were collected by personal exposure monitors at Balitai and Haiguangsi bus stops in Tianjin, China. The equivalent concentration of benzo[a]pyrene (BaPeq) was used to estimate the health risks of PAHs in PM10 inhaled by passengers waiting at these bus stops. The results showed the average PM10 level was higher in autumn (non-heating season) compared to spring (heating season) (307 ± 67 μg m vs. 226 ± 100 μg m). When averaged over the two bus stops, concentrations of total PAHs in PM10 were much higher during spring compared to autumn (417 vs. 193 ng m), while BaPeq was slightly lower during spring (29.7 vs. 32.8 ng m). The incremental lifetime cancer risks (ILCR) at the two bus stops, Balitai and Haiguangsi, in spring were 9.0 × 10 and 4.5 × 10, and in autumn were 1.1 × 10 and 7.7 × 10, respectively. All these risk values were lower than the acceptable risk range of 10–10 approved by US Environmental Protection Agency.


INTRODUCTION
Polycyclic aromatic hydrocarbons (PAHs) are a group of persistent organic pollutants that composed of two or more benzene rings (Cheruyiot et al., 2015).They are originated from both nature and human life, while incomplete combustion of carbonaceous organic materials is the main anthropogenic source of PAHs (ATSDR, 1995;Ravindra et al., 2008).PAHs are also semi-volatile organic compounds (SVOCs), which could be partitioned between the gas and particulate phases.Low molecular weight or 2-3 rings PAHs usually account for over 70% of gas-phase PAH concentrations, whilst those with four or more rings are found most in particle phase (Table 1) (Tasdemir et al., 2007;Ravindra et al., 2008).This study focuses on health risk assessment of the PM-bound PAHs because it has been shown that heavier PAHs (four or more rings) have a much greater carcinogenic and mutagenic potential risk than lighter ones (Erel et al., 2007;Urbancok et al., 2017;Zhang et al., 2017).
Once released into the atmosphere, most carcinogenic PAHs are predominantly associated with particulate matter (PM).Due to their persistence, bioaccumulation, carcinogenic, and mutagenic effects on human health, atmospheric pollution of PM-bound PAHs is recently an environment issue of major concern (USEPA, 2003;Petch et al., 2009;Kim et al., 2013;Dimitriou et al., 2018).The concentrations of PAHs are closely dependent upon the size of PM.The particles with a diameter less than 10 µm (PM 10 ) can deposit and accumulate in the respiratory system and can result in a potential threat to human health (Lodovici et al., 2003;Wiriya et al., 2013).Previous studies have indicated that vehicular emission is one of the major contributors to PAH emissions in urban areas (Tuominen et al., 1988;Shen et al., 2013;Hanedar et al., 2014;Xue et al., 2014).Since PAHs were mainly infiltrated from the outdoor environment through fuel, coal and gasoline burning (Cheruiyot et al., 2015), more attention should be paid for the environment of transportation.
In urban settings, public transport is a significant portion of the traffic system, where a large amount of people gathered every day.Passengers are exposed to motor vehicle emission PM and PAHs when waiting and taking the bus, especially during rush hours.PM levels were higher at road intersections with a lot of bus stops (Han and Naeher, 2006).
Many experiments have been conducted to evaluate the health risks of PM-bound PAHs when people exposed to various microenvironments such as hotel, home and roadside (Wu et al., 2014;Fan et al., 2017;Li et al., 2017).Nisbet et al., 1992; b Tasdemir et al., 2007.A risk assessment study in Tianjin reported the lifetime occupational risk of traffic policemen ranging 10 -6 to 10 -3 (Hu et al., 2007), which was above the acceptable risk of PAHs set by USEPA (1990).While few studies have focused on the health risk of PAHs for numerous of passengers in the bus stop.
To more thoroughly understand the effect of PM-bound PAHs to the passengers, adequate information on exposure concentrations to PAHs in traffic microenvironments is essential for risk assessments and management.In this study, PM 10 samples were collected at two bus stops Balitai and Haiguangsi during the autumn and spring in urban area of Tianjin.Experiment works were distributed in non-heating season and heating season (Nov.15-Mar.15 for the next year) separately to distinguish the influence of heating.The concentrations of PM 10 and PAHs were presented.The concentration of multi-component PAHs at these bus stops were converted into benzo(a)pyrene equivalent (BaP eq ) concentrations to estimate their exposure levels and health risks.

Sampling Sites
PM 10 samples were collected in Tianjin, China in September 2012 and in March 2014.Tianjin is a typical metropolis with a population of approximately 14 million (National Statistics Bureau of China, 2013).PM has been the major ambient air pollutant in Tianjin for a long time; suspended dust, coal burning, and vehicle emission are the major sources of PM pollution (Xue et al., 2014).
In this study, two bus stops, Balitai and Haiguangsi, both on Weijin Road, were chosen for PM monitoring.Weijin Road is an arterial road in Nankai District with a high traffic density (Fig. 1).The aspect ratios of this street are 0.5 to Balitai and 0.3 to Haiguangsi, which indicates that the sampling street is not a street canyon and the dispersion conditions are relatively good (Vardoulakis et al., 2003).Meanwhile, Tianjin Environment Monitoring Center (TEMC) which is located 3 km southwest to this study area was involved to afford the daily (24 hrs) concentration.

Sample Collection
PM 10 samples were collected at Balitai and Haiguangsi bus stops.The sampling took place in September 2012 (autumn, from September 23 to September 28 at both stops) and March 2014 (spring, from March 1 to March 7 at Balitai stop, and from March 8 to March 14 at Haiguangsi stop).Sampling times were from 7:00 to 9:30 and 16:30 to 19:00 local time, including both morning and afternoon rush hours on weekdays and weekends.A total of 24 daily rushhour PM 10 samples (12 samples in autumn and 12 in spring) were collected.
Each of the 24 samples consisted of one quartz filter and one Teflon filter by Model 200 personal environmental monitors (PEMTM, SKC.Co, USA) attached to a pole at the roadside of each bus stop, in the height of 1.5 m above the ground.The quartz filters were analyzed for PAHs, while Teflon filters were weighed to obtain the mass concentration of PM 10 .The samplers operated at a constant flow rate of 4 L min -1 .During the 5-hour sampling period, field technicians recorded the commuters' wait times per half-hour at each bus stop.The number of buses stopping at or passing by the bus stops was also monitored.
Prior to sampling, quartz filters were pretreated in the oven for 3 h at 800°C to volatilize all the organics.The quartz and Teflon filters were packed with aluminum foil to avoid light and were conditioned at a constant temperature and constant humidity for 48 hrs before weighted.The filters were weighed before and after sampling using a balance to determine the collected mass of the total PM.After weighed, the filters were repacked with aluminum foil and put in a refrigerator until extraction.All filters were extracted within 2 weeks of sampling.It should be noted that artifacts in sampling of ambient particulate PAHs have been reported (Cheruiyot et al., 2015(Cheruiyot et al., , 2016)).It was estimated that particulate PAHs collected by conventional samplers could be average 20% lower than those collected by a denuder device to remove ozone (Balducci et al., 2017).

Quality Control
All the operations were in strict quality control.The experimental tools like tweezers and sampling heads were sterilized by alcohol.Before sampling, quartz filters were pretreated in the oven for 3 h at 800°C to volatilize all the organics.Model 200 personal environmental monitor and sampling pump were calibrated by mini-BUCK calibrator every day to ensure the accuracy of flow.During the analysis of PAHs, standard solution was used to test the efficiency of the equipment and method.Extraction recoveries of these 16 priority PAHs were in a reasonable range.

Risk Assessment
In this study, inhalation of airborne particle-bound PAHs was considered for human health risk assessment.The total PAH was defined as the sum of all species above detection limits in each sample.The toxicity equivalency factor (TEF) was used to convert concentrations of carcinogenic PAHs (Table 1) to an equivalent concentration of BaP (BaP eq ) for assessing the cancer risks posed by PAHs exposure (USEPA, 1993;Hu et al., 2007;Yang et al., 2017).The BaP eq concentration (ng m -3 ) of each sample was calculated using Eq.(1): where C i and TEF i are the concentrations (ng m -3 ) and TEF corresponding to species i, as listed in Table 1.
A USEPA model was employed to assess the incremental lifetime cancer risk (ILCR) of adult passengers due to exposure to PAHs during wait times, as in Eq. (2) (USEPA, 1989): where CSF is the inhalation cancer slope factor of BaP (kg day -1 mg -1 ); BaP eq is the concentration (mg m -3 ) calculated in Eq.(1); IR is the inhalation rate (m 3 h -1 ); ET is the daily exposure duration (h d -1 ); EF is exposure frequency (d yr -1 ); ED is the lifetime exposure duration (yr); BW is body weight (kg); AT is the averaging time (d).
In this study, the daily exposure duration is the summation of average AM and PM wait times.The cancer slope factor of BaP is 3.1 mg -1 •kg•d from USEPA (1991).The inhalation rate of adults standing were 0.60 m 3 h -1 for male and 0.48 m 3 h -1 for female (USEPA, 1997), thus an average value of 0.54 m 3 h -1 was used in this study.The following factors were adapted from USEPA (1989): exposure frequency of 365 day yr -1 , lifetime exposure duration of 30 years, adult body weight of 60 kg, and averaging time of 70 years.
Two-sample T-test was used to examine the differences in mean concentrations between the two bus stops during the two seasons.Pearson correlation analysis was conducted to investigate the association among PAH species from the same sample.

PM 10 Concentrations
The PM 10 concentrations of samples collected at the two bus stops during autumn (non-heating season) and spring (heating season) and daily concentrations in Nankai area in spring are shown in Fig. 2. The average concentration (± SD) of PM 10 in Haiguangsi bus stop was higher than that in Balitai bus stop (335 ± 60 vs. 281 ± 68 µg m -3 ) during autumn.There were more bus lines through the Haiguangsi bus stop (18 vs. 16 lines) and more buses during rush hours (180 h -1 at Haiguangsi vs. 120 h -1 at Balitai bus stop).
PM 10 concentrations in autumn ranged from 226 to 403 µg m -3 with an average value ( ± SD) of 307 ± 67 µg m -3 , while that in spring were statistically lower (ranged from 113 to 410 µg m -3 , average of 226 ± 100 µg m -3 ).During spring, the daily rush-hour (5 hrs) PM 10 concentrations at each bus stop were higher than the daily (24 hrs) concentration at TEMC using a continuous monitor, except on March 12.The average concentration of PM 10 in the spring at bus stops was much higher than that at the TEMC (226 ± 100 vs. 153 ± 69 µg m -3 ).

PAHs Concentrations
Eight and twelve PAHs were detected in autumn and spring, respectively.The concentrations of PAHs in different seasons at the two bus stops are shown in Fig. 3. Higher total PAH levels were observed during spring (417 ± 231 vs. 193 ± 76.7 ng m -3 , p < 0.05).At both stops, the average values of total PAHs were higher during spring than those of autumn (spring: 445 ± 333 ng m -3 and 388 ± 64.7 ng m -3 at Balitai and Haiguangsi bus stop, respectively; autumn: 219 ± 98.5 ng m -3 and 171 ± 52.7 ng m -3 at Balitai and Haiguangsi bus stop).
In both seasons, NaP had the lowest concentrations, while Ace and DahA were not detected.Phe was the most abundant compound accounted for 40.8% in spring, while BbFA was the highest with a percentage of 18.4% in autumn.Among the six PHAs detected only in spring, Ant, Flua and Pyr are tracers of coal combustion (Simcik et al., 1999), Flu is mainly from biomass burning (Simcik et al., 1999), Phe is produced by combustion of natural gas and vehicle exhaust (Simcik et al., 1999), and NaP comes from coal and fuel combustion (Jia and Batterman, 2010).BbFA and IcdP were detected only in autumn, they are source marks of gasoline engine emissions (Ravindra et al., 2008).The non-detection of BbFA and IcdP in spring needs further investigation.Only six compounds were detected in both seasons, they are mainly from combustion of natural gas (BaA and Chr) and wood combustion (BaP) (Rogge et al., 1993), diesel emission (BghiP) (Harrison et al., 1996), coal and fuel combustion (NaP) and vehicle emission (BkFA) (Li et al., 1993).However, the seasonal variation of these six compounds was relatively small.That may because our sampling sites are close to the bus stops, the constant flow of vehicle results a relatively stable level of these six PAHs.
The ratios of individual PAHs had been employed as diagnostic ratios to identify the origin of PAHs (Ravindra et al., 2008).Table 2 showed the diagnostic ratios of some PAHs detected in this experiment, which indicated that sources of PAHs in the two bus stops in autumn were mostly vehicle emission, while those in spring (heating season) were mainly vehicle emission and coal burning.
Our autumn results were similar to the scenario in Beijing observed by Wu et al. (2014).Five PAHs were observed with high concentrations in both studies, with four out of five being the same in the two cities (Chr, B(b+k)FA, IcdP and BaP).

Correlation between the Two Sampling Sites
Correlation coefficients among PAHs at each bus stop are provided in Tables S1-S4 (Supplementary Information).The results are similar between the two bus stops.During spring, strong positive correlations (r > 0.81, p < 0.05) were found between group (Acy and Flu) and group (Phe  and Ant) at both stops; among Flua, BaA, Chr, Pyr and BaP at Balitai bus stop, and between group (BaA, Chr) and group (Pyr and BaP) at Haiguangsi.In autumn, all 8 detected PAHs were positively correlated with each other at Haiguangsi stop, while at Balitai PAHs with strong crosscorrelation were reduced to 6 (excluding BkFA and NaP).Jin et al. (2014) also reported strong correlations between 3-ring isomers (Phe and Ant) and among 4 and 5-rings isomers (BaP, BbFA, BkFA).Strong positive correlations among species suggest those compounds have the same sources, which was consistent with the research findings in Jin et al. (2014).
Statistics of waiting durations of passengers at the two bus stops during each season is presented in Table 3.During autumn, the differences in the mean waiting durations between morning and evening rush-hours were less than  2% at both stops.In spring, the differences were up to 12%.The mean daily (AM + PM) waiting durations were slightly shorter in spring.The lifetime cancer risks of the passengers waiting for buses were calculated by Eq. ( 2), using the mean and 95% confidence intervals of BaP eq concentrations at each bus stop in each season.The mean health risks were 9.0 × 10 -8 and 4.5 × 10 -8 in spring, at Balitai and Haiguangsi bus stops respectively, and the risks were slightly higher (1.1 × 10 -7 and 7.7 × 10 -8 ) in autumn.
The exposure risks at the two bus stops were both lower than the acceptable risk range (10 -6 -10 -4 ) set by the USEPA (1990).

DISSCUSION
Exposure levels and health risks in this and other studies are summarized in Table 4. PM 10 concentrations at the bus stops in Tianjin are twice higher than that in traffic environment of downtown Ulsan (79.4 µg m -3 ) where the sampling sites were also near the roads (Vu et al., 2011).As expected, the 5-hr rush hour PM 10 -bound PAHs concentrations in our study were much higher than the 24hr PM 10 -bound PAHs levels in Beijing-Tianjin-Hebei District (spring: 417 ± 231 ng m -3 vs. 159.4± 26.9 ng m -3 ; autumn: 193 ± 77 ng m -3 vs.78.1 ± 66.2 ng m -3 ) (Wang et al., 2015).Higher concentrations of PM 10 and PM 10 -bound PAHs at bus stop than those at background sites were primarily due to traffic emissions.
The average winter/summer ratio of PAHs concentrations was approximately 7 during the non-Olympic period in Wu et al. (2014).The average spring/autumn ratio of PAHs concentrations was 2 in our study, that was comparable to the range of 1 to 5 at Beijing, Tianjin and Hebei Districts in four seasons from Wang et al. (2015).The ratios indicated obvious seasonal variation of PAHs concentrations.The order of PAHs concentrations in different season is usually as follows: winter > spring > autumn > summer, and the influencing factors include emission sources, temperature, humidity, wind speed, etc.
In comparison with other studies conducted in Tianjin, the upper bound concentration of total PAHs in this study is half of the exposure concentrations for traffic policemen in 2005 and the upper bound BaP eq is also 50% of that of traffic policemen (Hu et al., 2007), 13-16% of that of the traffic assistants (Qin et al., 2011;Xue et al., 2014), but an order of magnitude higher than that in downtown Ulsan and in Istanbul near a road (Hanedar et al., 2014).
The average risks due to exposure to ambient PAHs found in our study are three orders of magnitude lower than the occupation risk of traffic policemen in Hu's study (2007), four orders of magnitude lower than the risk of traffic assistants by Xue et al. (2014), and six orders of magnitude lower than the risk of traffic assistants reported by Qin et al. (2011), all four studies were conducted in Tianjin.This is because of the much longer exposure duration by traffic policemen or traffic assistants than waiting passengers (8 h day -1 vs. 0.2 h day -1 ) and much higher exposure concentrations of BaP eq (policemen 82.4 ng m -3 , traffic assistants 251 ng m -3 , passengers up to 40 ng m -3 ).

CONCLUSION
PM 10 samples were collected in spring and autumn at Balitai and Haiguangsi bus stops in Tianjin.Results showed that the PM 10 concentrations were slightly lower in spring than autumn (226 vs. 307 µg m -3 ), while total PAHs concentrations were much higher in spring compared to autumn (417 vs. 195 ng m -3 ).The difference in the PAH concentrations was largely attributable to the increased emissions from heating sources including coal and biomass combustion in colder months.However, slightly lower BaP eq levels were observed in spring (29.7 vs. 32.8ng m -3 ) because of the reduction of BaP concentrations.The lifetime cancer risk of BaP eq at the two bus stops in the two seasons ranged from 2.5 × 10 -8 to 1.7 × 10 -7 , lower than the limit of acceptable risk range approved by USEPA (10 -6 -10 -4 ) (USEPA, 1990).The positive correlations among PAH species at each site suggest common source, however that is more pronounce in autumn.
Previous studies mainly focused on the health risks of PAHs concentrations during winter and summer.However, spring and autumn make half of the year in which people live and transport, and this study aims to fill the seasonal gap.
Limitations of our study include a relatively small sample size, and a short sampling period.Moreover, exposure to PM 2.5 -bound PAHs could be more relevant from human health point of view.Future risk assessment studies should include gas-phase PAHs and PM 2.5 -bound PAHs.In addition, future studies need to compare health risk due to exposure of PAHs at bus stops with the risks in other microenvironments, such as inside buses.

Fig. 2 .
Fig.2.PM 10 concentrations during spring and autumn at two bus stops and daily concentration at Tianjin Environment Monitoring Center (TEMC).

Fig. 4 .
Fig. 4. BaP eq concentration in different seasons at two bus stops.

Table 1 .
Toxicity equivalency factor (TEF) and gas-particle partition of sixteen EPA listed PAHs.

Table 2 .
Diagnosis ratios of individual PAHs.

Table 3 .
Waiting durations by bus stop and by season.