Source Apportionment and Lung Cancer Risk Assessment of PM2.5-Bound Polycyclic Aromatic Hydrocarbons in an Industrial City in Northwestern China


 In this study, we investigated the occurrence of particulate matter (PM2.5)-bound polycyclic aromatic hydrocarbons (PAHs) in the atmosphere in Shihezi, China and estimated their incremental lifetime cancer risk (ILCR). In addition, the potential sources of PM2.5-bound PAHs were statistically estimated using principal component analysis and positive matrix factorization. The backward trajectory was employed to determine the potential source area using the HYSPLIT software. Concentrations of PM2.5 and 16 PM2.5-bound PAHs ranged from 4.32 μg/m3 to 114.67 μg/m3 and 6.26 ng/m3 to 114.79 ng/m3, respectively, with the highest PAH concentrations occurring in the winter. Three principal components were analyzed by PCA, namely a combination of coal combustion sources and vehicle emissions, fugitive dust, and industrial emissions. Five sources were analyzed by positive matrix factorization (PMF) including coal combustion, vehicle emissions, fugitive dust, industrial emissions, and natural gas emissions. Risk assessment analysis showed that source-specific lung cancer risk assessment was in the 10-6–10-5 range, which constitutes a cancer risk that exceeds the guideline safety value (10-6) according to the U.S. Environmental Protection Agency. It is necessary to take measures to reduce the concentration of PM2.5-bound PAHs in order to reduce human health hazards.


Introduction
The increase in numbers of motor vehicles (Li et  The composition and toxicity of PAHs vary greatly depending on their sources. Some studies (Taghvaee et al. 2018, Wang et al. 2020a) used source-speci c cancer risk assessment to verify that different sources pose different hazards to the human body. Therefore, it is necessary to fully understand the sources of PAHs and the toxicity caused by these different sources.
PAHs are signi cant organic pollutants in ambient air (Manisalidis et al. 2020, Yadav et al. 2020, Zhou et al. 2022, and mainly exist in the vapor and particle phases. They are common contaminants that can pollute areas not only near their sources, but also at receptor sites far away from their sources (Lin et al. 2020), as they can be transported long distances (Shrivastava et al. 2017). In addition, long-term and lowlevel exposure to PAHs may cause chronic health effects such as lung cancer, poor fertility outcomes, intestinal damage , and DNA damage (Wang et al. 2020b). Epidemiological studies have indicated that chronic exposure to PAHs is associated with decreased intelligence(Blazkova et al. 2020), maternal demoralization (Rundle et al. 2012), asthma (Hu et al.), and obesity (Rundle et al. 2012). Stading et al indicated that PAHs in ambient air cause more lung cancer cases.
Shihezi, an important city in Xinjiang, is located along the Belt and Road. The topography of the Shihezi Reclamation Area from south to north consists of the Tianshan mountainous area, piedmont hilly area, piedmont slope plain, ood alluvial plain, and eolian desert area (Chen et al. 2019). Furthermore, special geographical environments and meteorological factors have caused the wind speed to decrease (Huang et al. 2021), which is not conducive to the diffusion of pollutants, especially particulate matter. In 2020, the air quality of Shihezi had 69.04% standard compliance, 11.23% light pollution, 5.47% moderate pollution, and 12.06% severe pollution, and 2.19% severe pollution, with the main pollutants being inhalable particulate matter.
The main objectives of this study were to (1) analyze the concentration level and distribution characteristics of 16 PM 2.5 -bound PAHs in Shihezi from October 2020 through September 2021, (2) perform source-speci c risk assessment of PM 2.5 -bound PAHs by employing the PCA and the PMF models to distinguish the potential source pro le of local emissions, and (3) evaluate the carcinogenic risk via inhalation exposure to PAHs for different groups of people.

Sampling location description and collection
Shihezi, an inland city located in the north of Xinjiang, is a rapidly developing city in northwestern China.
Based on the distribution of urban functional areas, three sample points (S1, S2, and S3) were set up in three locations, as shown in Figure 1. The sampling locations were selected based on the comprehensive situation, including transport, convenience, and security. All sampling sites were 5 m above the ground, with no surrounding obstacles.
Particle sampling was carried out by employing low-volume air samplers with a ow rate of 16.7 L/min (manufactured by US BGI) to collect 23-h PM 2.5 , using quartz lters (47 mm diameter, Whatman) from October 2020 through October 2021. Prior to sampling, the lters were pre-baked in a mu e furnace at 700°C for 5 h to remove organic compounds. The lter blanks were then processed in the same manner as the sample. All sampled lters were stored at −15°C until further analysis.

PAHs extraction and analysis
An accelerated solvent extraction (ASE) instrument was used to chemically extract 16 particle-bound PAHs. Half of the quartz lter was cut into small pieces and placed in the extraction tank. Extractions were performed under 100 bar, the hold time was 5 min, and the number of cycles was 2. After the sample was concentrated, we quanti ed and separated PM 2.5 -bound PAHs using ultra performance liquid chromatography (UPLC). Detailed information from the 16 PAHs and instrument conditions are shown in Table S1 and Table S2.

Quality control and quality assurance
In this study, all operations and analysis processes strictly complied with quality assurance and quality control (QA/QC) procedures. First, we ensured the authenticity and accuracy of the samples during the sampling process. The ow rate of the particulate matter sampler was calibrated before and after sampling in each season and at each stage to ensure accurate ow. Second, we ensured quality control during the analysis process. Whole blanks, trip blanks, and laboratory blanks were analyzed using the same method used for sample analysis. Sixteen PAH standards were added to the blanks to calculate the recoveries, which were approximately 87.2-106.5%. The method detection limits (MDLs) were calculated as three times the standard deviation of the mean blank concentrations. In addition, the concentration of PAHs in the blank sample was below the detection limit of the method. The abbreviations of 16 PAHs were shown in Table S1. Rregression equation, R 2 , detection limit and average recovery rate were shown in Table S2.
where Y is the matrix of the principal component score, PT is a transposed matrix of principal component loadings, E is the matrix of residues, and X is the original PAH data matrix.

Positive Matrix Factorization Model
where C i represents the concentration of the i-th PAH, TEFi, and MEFi, respectively, and indicate the toxic equivalency factor and mutagenic equivalent factor, respectively.
Lifetime carcinogenic risk based on the USEPA standard models was used to assess the carcinogenic risk of PAH exposure in humans. First, each exposure route was calculated, then the carcinogenic slope factor of each exposure route was multiplied to obtain the carcinogenic risk of each route, and nally the total carcinogenic risk was totaled (Wang et al. 2019). The calculation formula is as follows:

5
where Does Ingestion , Does Inhalation , and Does Dermal are the exposure doses (mg/kg/D) by ingestion, inhalation, and absorption, respectively. SF Oral , SF Inhalation , and SF Dermal represent the carcinogenic slope factors of ingestion, inhalation, and absorption, respectively.

Spatial variation of PAHs concentrations
The spatial variation of ΣPAHs mass concentrations is shown in Table S1. , which showed that winter conditions typically lead to higher PAH concentrations. The seasonal pattern of PAHs was attributed to changes in the emission sources and meteorological conditions. Temperature is an important meteorological condition that affects the levels of pollutants during the heating season. On one hand, temperature affects the migration and conversion of PAHs in the atmosphere (Dron et al. 2021); on the other hand, temperature greatly affects the heating demand, which indirectly affects the emission source. Hence, the lower the temperature, the higher the coal and natural gas emissions.
High molecular weight (HMW, 4-6 ring) PAHs are considered to be more toxic. However, the toxicity of low molecular weight (LMW, 2-3 ring) compounds cannot be ignored. In this study, both HMW and LMW PAHs were found in PM 2.5 . As shown in Figure 3, the concentrations of 4-6 ring PAHs were 81.2% and 81.5% in the heating and non-heating seasons, respectively. 4-ring PAHs accounted for the highest proportion in the heating season, and 5-ring PAHs dominated during the non-heating season.

Correlation analysis
The  Figure S2 illustrates the correlations between the measured and model-predicted total PAH mass concentrations. Based on this, the PMF model was able to perfectly predict the total PAH mass concentrations, indicating that this model is suitable for PAH source apportionment.

Source apportionment of PAHs by PMF
Factor 1 was mainly de ned by naphthalene (Nap), Pyr, and benzo(a)anthracene (BaA), which contributed 22.9% to the total PAH concentration, interpreted as natural gas emissions, as shown in Figure 5.
Previous studies have shown that Pyr and BaA are typical tracers of natural gas (Li et al. 2019b). Xinjiang is rich in natural gas resources, which have been used as fuel for automobiles and households. Shihezi City has achieved full coverage of natural gas installations, which means that every household uses natural gas. In addition, the use of LNG in vehicles is trending because of low prices. Generally, the highload low-ring aromatic hydrocarbons also verify that the combustion of LNG produces PAHs with lower and intermediate volatility (Mehmood et al. 2020). The lowest contribution to this source occurs in winter, but there is no seasonal change in tra c ow near the sampling point. This may be due to the poor performance of the car engine exhaust control system at low temperatures (Pham et al. 2013, Zhou et al. 2005 (Jamhari et al. 2014) showed that these substances are direct markers of coal combustion. The contribution of coal-red sources to PAHs is less than that of automobile sources and natural gas sources, due to the local government's achievements in coal substitution. However, the contribution of coal sources to PAHs in winter has signi cantly increased because coal is still the primary fuel source in this area, which is used in power plants and central heating system (Zhang et al. 2021). In addition, the contribution of coal sources in spring was also relatively high, which is related to the six-month heating season in Shihezi. Figure 6 shows the source direction and proportion of backward trajectory clustering of the four seasons during the sampling period. Shihezi, the largest province in China, is located in the north of Xinjiang.

Backward trajectory analysis
Therefore, most of the trajectories came from Xinjiang, and a small portion of the trajectories came from countries bordering Xinjiang. In winter, the air mass of cluster 1 came from the northeast of Xinjiang, accounting for 6.45%, which is a short-distance transportation. Both cluster 2 and cluster 4 were from the western part of Xinjiang. The former accounted for 22.85%, and the latter accounted for 10.89%. Cluster 3 was the air mass with the largest proportion (41.26%), which belongs to the local transportation around Shihezi. This indicates that winter heating caused this result. Additionally, the source contribution of around Urumqi cannot be ignored, and the air mass from this part is cluster 5 in the gure. Cluster 6 originated in Kazakhstan and reached Shihezi after long-distance transportation. In spring, clusters 1 and 3, accounting for 71.5% of the total air masses, came from short-distance transportation in the northeast and southwest directions of Shihezi, respectively. Clusters 2 and 6 both came from long-distance transportation in Kazakhstan, accounting for 11.69% and 4.44%, respectively. In summer, clusters 1, 3, and 5 were all from Kazakhstan, accounting for 23.19%, 14.03%, and 5.97%, respectively. Trajectory 2 comes from Karamay, an oil eld city, which contributes to the high polycyclic aromatic hydrocarbons in Shihezi. Figure 6 clearly shows that the sources were roughly the same in summer and autumn. In autumn, cluster 1 had the highest proportion (27.55%), followed by cluster 3 (24.19%). Both originated from short-distance transportation. Clusters 4 and 6 were from the Ili Kazakh Autonomous Prefecture and Altay Region, respectively. The other two trajectories were from Kazakhstan.
The gure 7 shows the WPSCF analysis result of Shihezi's backward trajectory. A larger WPSCF value indicates that this area has a greater impact on the local PAHs. In winter, there were many concentrated areas where the CWT value was greater than 100, mainly in the surrounding areas of Shihezi. This may be due to temperature inversion in winter, which is not conducive to the diffusion of pollutants. In addition, the CWT value in autumn was roughly the same as the distribution in winter, and the potential source was located near Shihezi, which may have been caused by biomass burning in the surrounding area in autumn. In contrast, there were relatively few areas with a CWT value greater than 60 in summer, and they were chie y distributed in Kazakhstan, which is far from Shihezi. The number of severely affected areas in spring was signi cantly lower than that in winter and autumn.

Source-speci c lung cancer risk assessment
The health risk assessment for children and adults was conducted using a Monte Carlo simulation, and the BaP eq and ILCR results are illustrated in Figure 8. The mean BaP eq concentrations in winter, spring, summer and autumn were 10.97, 6.85, 5.52, 3.98 ng/m 3 , respectively, with an annual mean concentration of 6.83 ng/m 3 . Table S3  were as low as 2.7 ± 3.0 ng/m 3 and 3.85×10 −2 ng/m 3 , respectively, both being lower than the BaP eq level observed in Shihezi. As shown in Figure 8, the ILCR value of PAH exposures in the four seasons ranged from 2.61×10 −6 to 9.91×10 −6 for child and adult, with all values being higher than the threshold value (10 −6 ), suggesting a potential cancer risk. The total risk followed the order of winter > spring > autumn > summer for both children and adults, which was consistent with the seasonal distribution order of PAHs. The ILCR for adults was higher than that for children, which may be related to an unhealthy lifestyle, frequent exposure, and higher body weight of adults. Figure 9 presents the lung cancer risk for each of the PAH emission sources, simulated by the PMF model. Results showed that the contributions of every source to ILCR in winter, except for the contribution of vehicles emissions to children, exceeded the prescribed threshold value (10 −6 ). Coal combustion and fugitive dust were the major sources of PAHs, with signi cantly higher cancer risks than the others. In summer, vehicle emissions account for the largest proportion of cancer risk, with ILCR values of 1.42×10 −6 for children and 1.98×10 −6 for adults, both higher than the threshold value. In summer and autumn, the contribution of most sources to ILCR was less than the threshold, whether for children or adults, suggesting a negligible cancer risk. However, the sum of their contributions to ILCR was greater than the threshold, indicating a potential cancer risk. The above results indicated that strengthening the comprehensive management of coal combustion source, vehicle emissions, fugitive dust, industrial emissions, and nature gas emissions is important for public health.

Conclusion
Few studies have been conducted on industrial cities with unique climate and geographical conditions in the arid and semi-arid regions of northwest China, especially Xinjiang. During the development process, these cities will inevitably cause damage to the environment, thereby threatening the health of local residents. Therefore, it is necessary to monitor the highly toxic pollutants and use a variety of methods to trace their potential sources. This study uses principal component analysis and the PMF model to identify PM 2.5 -bound PAH sources, which will be of help in policy formulation.
The main objectives of this research was to carry out the concentration distribution, analyze source apportionment based on multiple methods (PCA, PMF, and backward trajectory), and assess sourcespeci c lung cancer risk of ambient PM 2.5 -bound polycyclic aromatic hydrocarbons (PAHs) in Shihezi. Validation and investigation. All authors read and approved the nal manuscript.

Date availability
The data and materials presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions.

Ethics declarations
Ethics approval Not applicable.

Consent to participate
Not applicable.

Consent for publication
All authors reviewed and approved the manuscript for publication.
Chemosphere   Sources pro les of PAHs identi ed from the PMF model for annual PAHs Figure 6 Page 19/20 The backward trajectories cluster of PM 2.5 -bound PAHs in Shihezi during each season.

Figure 7
Potential sources of PM 2.5 -bound form the CWT model.

Figure 8
BaP eq concentration and ILCRs from PAH inhalation exposure in four seasons.

Figure 9
The ILCR and potential contributions of different emission sources in the four seasons.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.