Insight into the size-resolved markers and eco-health significance of microplastics from typical sources in northwest China

Research on atmospheric microplastics (MPs) from typical sources is limited, constraining the targeted management of pollution. Here, the source profiles of eight types of common MPs and three classes of plasticizers (i.e., phthalates, benzothiazole and its derivatives, and bisphenol A) emitted from five living sources, including Plastic Burning (PB), Fruit-bag Burning, Road Traffic (RT), Agricultural Film, and Livestock Breeding (LB), were determined in PM_2.5 (particulate matter with aerodynamic diameters ≤ 2.5 µm) and PM₁₀ (≤ 10 µm) in the Guanzhong Plain, northern China. PB exhibits high proportions of poly(methyl methacrylate) (PMMA) and 2-hydroxy benzothiazole (HOBT), with PMMA being more abundant in PM_2.5–10 (aerodynamic diameters between 2.5 and 10 µm). FB exhibits a higher proportion of di-n-octyl phthalate (DnOP) in PM_2.5–10 than in PM_2.5. RT shows a distinguishable profile with high abundances of rubber. The abundance of 2-benzothiazolyl-N-morpholinosulfide (OBS) in PM_2.5–10 was twice that in PM_2.5 for RT. Polystyrene (PS) is the most abundant MP in AF. LB shows the distinguishing feature of benzothiazoles, especially OBS and N-cyclohexyl-2-benzothiazolesulfenamide (CBS). The eco-health risk assessments reveal that combustion-derived MPs (Plastic Burning and Fruit-bag Burning) indicate the highest ecological risk (Level III). Elevated hazard indices to human health were observed in LB and PB, primarily attributed to bis(2-ethylhexyl) phthalate (DEHP). Notably, PMMA, polyethylene terephthalate (PET), polyethylene (PE), bisphenol A (BPA), and phthalates (PAEs) emerged as key drivers of oxidative stress of PMs. This study advances the understanding of atmospheric MPs, offering critical insights for source tracking and risk assessment to mitigate their eco-health effects.

1 Introduction

Global plastic production has increased exponentially after 1990, resulting in serious environmental contamination (Geyer et al., 2017; Klein et al., 2023). Waste plastics have accumulated in the environment and have been degraded into plastic debris under the influences of UV-radiation, mechanical abrasion, and temperature changes (Peeken et al., 2018; Akhbarizadeh et al., 2021). Microplastics (MPs) are plastic particles with 1 µm–5 mm in size (Can-Guven, 2021). Research on MP pollution initially focused on aquatic and terrestrial ecosystems, but recent years have seen growing attention to atmospheric MP pollution (Allen et al., 2020). Understanding the sources of atmospheric MPs can assist in developing efficient MP management strategies.

Common atmospheric MP sources include waste incineration, agricultural activities, and road traffic (Panko et al., 2013; Luo et al., 2022; Yang et al., 2024; Chen et al., 2024). Incineration activities can lead to the fragmentation of plastics, accelerating the release of MPs (Luo et al., 2022, 2024b). Yang et al. (2021) have estimated that a metric ton of plastic can potentially produce 360 to 102 000 MPs, primarily composed of polypropylene (PP) and polystyrene (PS). In addition to industrial incineration processes, open burning activities in daily life also contribute significantly to atmospheric MPs. Due to the relatively limited facilities and means of waste disposal, residents in rural areas often resort to open burning when disposing of plastic waste (Pathak et al., 2023). In addition, given the flammability of plastics, residents also tend to use plastics as igniters or even burn them directly when using stoves for cooking or heating, which is an important household source of MPs. Agricultural activities are also a significant contributor to atmospheric MPs (Jin et al., 2022; Yuan et al., 2025). The large consumption of plastic film combined with a short life cycle results in a number of films being left in farming soil, which then transform into MPs via degradation or fragmentation (Brahney et al., 2021; Wang et al., 2022a; Aini et al., 2023). Agricultural activities (e.g., plowing and harvesting), by increasing soil disturbance, may cause the resuspension of MPs into the atmosphere (Jin et al., 2022; Lakhiar et al., 2024). Furthermore, tire and road wear microplastics (TRWMPs), produced from the interaction between tires and the road surface, are a significant source of atmospheric MPs (Panko et al., 2013; Liu et al., 2023; Xu et al., 2024b). Evangeliou et al. (2020) have estimated that annual total global tire wear particle emissions were 2907 kt yr⁻¹, with 29 and 288 kt yr⁻¹ for PM_2.5 (particulate matter with aerodynamic diameters ≤ 2.5 µm) and PM₁₀ (≤ 10 µm), respectively. Liu et al. (2023) have shown that rubbers were the dominant compounds of TRWMPs in PM_2.5 in tunnels, including natural rubber (NR), styrene-butadiene rubber (SBR), and butadiene rubber (BR) polymers. Current research on atmospheric MP sources focuses on industrial emissions and natural processes but neglects air pollution sources closely related to daily life activities. Given that sources from daily human activities significantly affect human health, this study pays particular attention to such sources.

Plasticizers are widely used in the production of plastics in order to achieve the desired material properties (Demir and Ulutan, 2013). Since plasticizers are not chemically bound to the plastic products, they can easily diffuse into the surrounding environment during their lifetime (Demir and Ulutan, 2013; Yadav et al., 2017). Phthalate esters (PAEs), benzothiazoles (BTs), and bisphenol A (BPA) are the most common plastic additives that are ubiquitous in the environment and pose potential health risks. PAEs are the most widely used plasticizers globally, dominating the plastic additive market. He et al. (2020) demonstrated that during 2007–2017, the annual global production of PAEs increased from 2.7×10⁶ to 6×10⁶ t. China is recognized as the largest importer of PAEs worldwide (Cui et al., 2025). BTs are extensively used in automotive tires and agrochemicals. High concentrations of BTs were discovered in street runoff, suggesting that these tire material-related compounds can persist in the environment (Zhang et al., 2018). Exposure to BTs may result in central nervous system depression, liver and kidney damage, dermatitis, and pulmonary irritation (Ginsberg et al., 2011). BPA, as a common industrial chemical component in many products, has steadily increased over the last 50 years (Corrales et al., 2015). The growth of global production has consistently ranged between 0 % and 5 % annually (Corrales et al., 2015). PAEs and BPA, considered endocrine disruptors, have been demonstrated to impair reproductive function and development in laboratory animals (Wang et al., 2019).

Previous studies have investigated the emission characteristics of plasticizers from various sources. Simoneit et al. (2005) illustrated that the major plasticizers detected in PMs from open burning of plastics were dibutyl phthalate (DBP), diethylhexyl adipate (DEHA), and diethylhexyl phthalate (DEHP). Zeng et al. (2020) reported that phthalate concentrations in greenhouse air were higher than those in ambient air. Liu et al. (2023) found that phthalates were the most dominant plasticizer components in tunnel PM_2.5, accounting for 64.8 % of the detected plasticizers. Zhang et al. (2018) demonstrated that tire material-related compounds, benzothiazole (BT), and 2-hydroxybenzothiazole (2-OH-BT) were the major compounds in both tire and road dust samples. The majority of existing studies on atmospheric MPs and plasticizers have focused on analyzing the emission characteristics of individual sources and have lacked a comprehensive and comparative analysis of the MP emission profiles of various sources.

MPs and plasticizers can remain suspended and spread to other areas when they are emitted from sources into the air (Gasperi et al., 2018). Airborne MPs can easily enter the human body directly via respiration compared to other environmental exposure pathways, posing a serious health concern (Liao et al., 2021; Luo and Guo, 2025). Recent studies suggest that these inhaled pollutants can promote reactive oxygen species (ROS) generation (Wang et al., 2024). Oxidative potential is a metric reflecting the ability of inhaled pollutants to produce ROS, serving as a critical indicator of PM toxicity (Jiang et al., 2019; Bates et al., 2019; Luo et al., 2024c). ROS overproduction acts as a central driver of oxidative stress, which can damage biomolecules and disrupt cellular functions (Bates et al., 2019; Jiang et al., 2019). Previous studies have demonstrated that metals and organic compounds can affect the oxidative potential of PMs (Ghanem et al., 2021; Luo et al., 2023). However, most studies on MPs and plasticizers have focused on their environmental occurrence rather than systematic health risk assessments from atmospheric pollution sources.

The Guanzhong Plain, located in the central part of Shaanxi Province in northwestern China, inevitably consumes a large number of plastics with a developed economy and a large population (Chen et al., 2022; Wang et al., 2022b; Xu et al., 2024a). The environmental conditions of strong wind and ultraviolet radiation in this area exacerbate the problem of atmospheric MP pollution (Liu et al., 2017). There is a notable absence of systematic comparative analyses examining the emission profiles across various emission sources, which is the key to controlling MP pollution. The aims of this research are to (i) characterize the distributions of MPs and plasticizers in dual-size PMs (PM_2.5, PM_2.5–10) from typical MP sources (anthropogenic sources from daily life) in the Guanzhong Plain, (ii) obtain MP and plasticizer tracers for the five typical MP sources, and (iii) evaluate the health risks of MPs and plasticizers in PM_2.5 and PM₁₀. This study could provide valuable scientific support for the development of targeted pollution control strategies, as well as sustainable improvement of the regional environment and public health protection.

2 Methods

2.1 Sample collection and gravimetric method

During January and February 2024, PM_2.5 and PM₁₀ samples were collected simultaneously from five distinct sources in three key cities of the Guanzhong Plain: Xi'an, Tongchuan, and Xianyang (Fig. S1 in the Supplement). The selected sources included Plastic Burning, Fruit-bag Burning, Road Traffic, Agricultural Film, and Livestock Breeding. It should be noted that plastics from the Plastic Burning source, including plastic bags, bottles, disposable tableware, foam boxes, and other plastic daily necessities, are incinerated. Fruit bags are typically lightweight thin films, which are designed for single use and are often discarded after a short period, differing from common household plastic waste. These bags are usually made from low-density polyethylene and nylon, which are known for their flexibility and transparency (Ali et al., 2021; Yang et al., 2022). The Guanzhong Plain is an important fruit production base in China, with the highest consumption of fruit bags. Local residents often use the above-mentioned plastic products to ignite solid fuels for indoor heating or cooking. The reason we distinguish between Plastic Burning and Fruit-bag Burning rather than classifying them as a single combustion source is that the wax layer in fruit bags cannot be separated from the plastic. This is a distinctive source in the Guanzhong region (this is also quite common in fruit-producing areas in northern China), and local residents typically burn fruit bags directly without separating the wax. Table 1 provides a summary of the essential details for each source.

The specimens were gathered using pre-fired quartz-fiber filters (QM/A, PALL, Ann Arbor, MI, USA) with a diameter of 47 mm, which had been subjected to a temperature of 800 °C for 3 h (Wang et al., 2022b). MiniVOL samplers (Airmetrics, Springfield, OR, USA), which are inertial impactors, were employed for collection, operating at a steady flow rate of 5 L min⁻¹ (Wang et al., 2022b) (Fig. S1). Sampling periods for each source ranged from 2 to 24 h, depending on the emission amount. In Agricultural Film and Livestock Breeding, the sampler was set at about 1.5 m in height, corresponding to the human breathing height. For Plastic Burning, Fruit-bag Burning, and Road Traffic sources, the sampling heights were related to the height of the chimney and flyover, about 3–4 m above the ground. The field blank of each type of source was synchronously collected with active sampling. Unused filters (the same batch as sampling filters) were loaded into identical sampling devices, which were placed adjacent to operational samplers for the entire duration of one sampling event.

Table 1 Basic sampling information of target emission sources.

Filters were transferred using stainless steel tweezers into pre-labeled clean glass cassettes after collection and frozen at −20 °C until chemical analyses. An electronic microbalance (±1 µg sensitivity, ME 5-F, Sartorius, Germany), with an anti-static instrument, was used to weigh the filters before and after sampling (Wang et al., 2022b). Cotton lab coats and nitrile gloves were utilized during sampling, while the use of plastic materials was minimized (Bogdanowicz et al., 2021).

2.2 Chemical analysis

This study quantified eight types of microplastics and three classes of plasticizers (phthalates, benzothiazole and its derivatives, and bisphenol A) in PM_2.5 and PM₁₀ samples.

2.2.1 Microplastics (MPs)

To quantify the contents of the MPs, a setup was employed in which a Curie-point pyrolyzer (JHS-3, Japan Analytical Industry Co., Ltd.) was connected to a gas chromatography–mass spectrometry (GC/MS) system (7890GC/5975MS, Agilent Technology, USA) (Liu et al., 2023). The pyrolysates of polyethylene (PE), polypropylene (PP), polystyrene (PS), polyethylene terephthalate (PET), poly(methyl methacrylate) (PMMA), natural rubber (NR), styrene-butadiene rubber (SBR), and butadiene rubber (BR) were identified using mass spectrum fragments, retention times, and target product intensities compared to plastic standards. All standards except rubbers (99 %, JSR Corporation) were purchased from DuPont (≥ 98 %, USA). The quantified markers for the pyrolyzed compounds are shown in Table S1. Details regarding the preparation of samples and instrument configurations (Sun et al., 2022) are available in Sect. S1.

2.2.2 Phthalates (PAEs)

An in-port thermal tube (78 mm long, 4 mm I.D., 6.35 mm O.D., Agilent Technology, USA) coupled with a GC/MS system (7890GC/5975MS, Agilent Technology, USA) was utilized to analyze phthalates, including dimethyl phthalate (DMP), diethyl phthalate (DEP), di-n-butyl phthalate (DBP), butyl benzyl phthalate (BBP), bis(2-ethylhexyl)phthalate (DEHP), and di-n-octyl phthalate (DnOP). All PAEs were purchased from Sigma-Aldrich (≥ 98 %, Steinheim, Germany). Aliquots of the filters (1.578 cm²) were diced into smaller fragments, augmented with the internal standard Chrysene-d12 (96 %, LGC Standards Limited, United States), and then inserted into thermal tubes (78 mm long, 4 mm I.D., 6.35 mm O.D., Agilent Technology, USA) for analyses (Liu et al., 2023). The sample tube was inserted directly into a GC injection port set to an initial temperature of 50 °C (Wang et al., 2016). Detailed sample preparation and analytical procedures (Ho et al., 2019; Liu et al., 2023) are provided in Sect. S2.

2.2.3 Benzothiazole and its derivatives (BTs)

Nine types of benzothiazole-related compounds (97 %, Thermo Fisher Scientific Co., Ltd., Waltham, MA, United States) were quantified, including benzothiazole (BT), 2-hydroxy benzothiazole (HOBT), 2-mercaptobenzothiazole (MBT), 2-aminobenzothiazole (2-NH2-BT), 2-(methylthio)benzothiazole (MTBT), 2-(4-morpholinyl)benzothiazole (24MoBT), N-cyclohexyl-2-benzothiazolamine (NCBA), 2-benzothiazolyl-N-morpholinosulfide (OBS), and N-cyclohexyl-2-benzothiazolesulfenamide (CBS). The appropriate filter sample was cut (1.578 cm²) and spiked with an internal standard (IS) of benzothiazole-d4 (95 %, LGC Standard Limited, United States). After a series of extraction and concentration procedures (Sect. S2), the target analytes were washed out with 5 mL of methanol (HPLC grade, Fisher Chemical, USA). Before analysis, the eluates were dried to 1 mL under a stream of nitrogen (Zhang et al., 2018). Target analytes were separated using an ultra-performance liquid chromatography system (UPLC; ACQUITY, Waters, USA) and subsequently identified with a triple quadrupole mass spectrometer (ESI-MS/MS; Xevo TQ-S, Waters, USA). Analytical details (Zhang et al., 2018) are provided in Sect. S2.

2.2.4 Bisphenol A (BPA)

Quantification of total BPA and separation from the matrix components were carried out by LC–fluorescence detection (García-Prieto et al., 2008). The mobile phase was composed of acetonitrile and water (Jian-Ke et al., 2011). The BPA standard was obtained from Sigma-Aldrich (USA). Moreover, all employed solvents and diluents were of HPLC grade and filtered through 0.45 µm membranes. The extracted sample was separated by a PerkinElmer Brownlee™ HRes Biphenyl 1.9 µm, 50 × 2.1 mm column with an isocratic elution program of water : acetonitrile (6 : 4) at 0.5 mL min⁻¹ for 4 min. The target analyte was measured using a fluorescence detector at excitation and emission wavelengths of 275 and 313 nm, respectively (García-Prieto et al., 2008). BPA levels were quantified based on measured peak areas (García-Prieto et al., 2008).

2.3 Quality assurance/quality control (QA/QC)

The flow rates of all samplers were calibrated using a mass flowmeter (Model 4140, TSI, Shoreview, MN, USA) before and after each sampling cycle. All quartz filters used in this study were preheated at 800 °C for 3 h to remove any potential contaminants and then cooled before use. To minimize experimental error, sampling was conducted in duplicate for each particle size of each source. For the chemical measurements, one in every 10 samples was reanalyzed for quality assurance purposes, and the standard deviation errors of replicate trials were within 10 % for the pyrolysis analyses. Calibration curves were established using reference standards. The linearities of the standard calibration curves were >0.987. The standard deviations of the pyrolyzed standards were within 94.1 % to 98.3 %. Background contamination (Table S3) was monitored by processing operational blanks (unexposed filters) simultaneously with field samples.

2.4 Oxidative potential determination with DTT assay

Four 0.526 cm² punches per sample from different sources were individually dissolved in 5 mL methanol (HPLC grade, Fisher Chemical, USA) in an amber centrifuge tube and ultrasonically extracted for 2 h. The PM extract was used for the subsequent analysis.

The Dithiothreitol (DTT) consumption in this study was quantified following the methodology established by Luo et al. (2024c). 4 mL of sample extract was combined with 1 mL of 1 mM DTT solution (≥ 98 %, Meryer; pH 7.4 buffer), yielding a final concentration of 200 µM. At each time point (0, 5, 15, 30, 45, and 60 min), 0.5 mL of the DTT reaction mixture was added to the amber centrifuge tube preloaded with 0.5 mL of trichloroacetic acid (1 %, $w/v$) to terminate the reaction. Subsequently, 25 µL of 10 mM 5,5^′-dithiobis-(2-nitrobenzoic acid) (DTNB, ≥ 98 %, Meryer) and 1 mL of 1 M Tris–HCl buffer were added to each tube. The solutions (200 µL) were transferred to 96-well plates, and absorbance was measured at 412 nm using a microplate reader (Flex Station 3 Multi-Mode, Molecular Devices). The volume-normalized DTT consumption rates for each sample were calculated from absorbance measurements taken at predetermined time points (nmol min⁻¹ m⁻³).

To ensure the accuracy of the results, the entire experiment was performed under dark conditions. Prior to sample analysis, a standard curve was generated by measuring the absorbance of 11 DTT concentration gradients within the range of 0 to 450 µmol L⁻¹, achieving a correlation coefficient (R²) of 0.9997. Pure methanol solution was used as a blank control, which was processed and measured in the same manner as the samples. The DTT consumption rate of each sample was corrected using the DTT consumption rate of the blank. Each batch of samples and methanol blanks was measured in duplicate to verify experimental reproducibility. The linear fitting R² for DTT consumption rates was consistently greater than 0.9, and the coefficient of variation (standard deviation) for parallel experiments was less than 15 %.

2.5 Risk assessment model

To evaluate the potential ecological risks, the hazard indices of various MPs were estimated using Eq. (1) (Xu et al., 2018; Wang et al., 2021a). The risk index (H) was calculated by multiplying the proportion (P_n) of each polymer identified in MPs by its respective hazard score (S_n) (Lithner et al., 2011).

$$\begin{array}{}\text{(1)}& H=\mathrm{\Sigma }{P}_n\times S_n\end{array}$$

The average daily exposure dose (ADD) via respiratory inhalation was calculated using Eq. (2), as defined by the U.S. Environmental Protection Agency (U.S. EPA, 1989; Liu et al., 2023). The non-carcinogenic and carcinogenic health risks of MPs and plasticizers were quantified using the hazard quotient (HQ) (Eq. 3) and incremental lifetime cancer risk (ILCR) (Eq. 4), respectively.

$$\begin{array}{}\text{(2)}& {\displaystyle }\mathrm{ADD}={\displaystyle \frac{C\times \mathrm{ET}\times \mathrm{IR}\times \mathrm{EF}\times \mathrm{ED}}{\mathrm{AT}\times \mathrm{BW}}}\text{(3)}& {\displaystyle }\mathrm{HQ}={\displaystyle \frac{\mathrm{ADD}}{\mathrm{RfD}}}\text{(4)}& {\displaystyle }\mathrm{ILCR}=\mathrm{ADD}\times \mathrm{SF},\end{array}$$

where C represents the measured mass concentration of MPs and plasticizers from five sources. Exposure parameters included ET (exposure time, 0.5 h d⁻¹ for combustion sources (Plastic Burning and Fruit-bag Burning), 1.5 h d⁻¹ for others), IR (inhalation rate, 20 m³ d⁻¹), EF (exposure frequency, 120 d yr⁻¹ for Plastic Burning and Fruit-bag Burning, 350 d yr⁻¹ for others), ED (exposure duration, 30 years), AT (average exposure time, ED × 365 d yr⁻¹ × 24 h), and BW (adult body weight, 70 kg) (Liu et al., 2023). Reference dose (RfD) and slope factor (SF) were obtained from the Integrated Risk Information System of U.S. EPA (https://www.epa.gov/risk/risk-assessment-guidance-superfund-volume-i-human, last access: 25 March 1991) and Ma et al. (2020), as detailed in Table S7.

2.6 Data analysis and statistical method

Data entry and organization were conducted using Excel 2016 (Microsoft Corporation, Redmond, WA, USA), while one-way analysis of variance (ANOVA) was performed with SPSS 26.0 (IBM, Armonk, NY, USA). Spearman correlation analysis was used to assess the relationships of MPs and plasticizers with ROS. Additionally, all data are presented as mean ± standard deviation, with significant differences denoted by P<0.05.

The Source–Pathway–Receptor (SPR) model serves as a key tool for illustrating how environmental pollutants travel from their origins, navigate various pathways, and ultimately reach potential receptors (Waldschläger et al., 2020).

3 Results and discussion

3.1 Concentrations of microplastics and plasticizers

For the convenience of comparison, we subtracted the concentrations of MPs and plasticizers in PM_2.5 from PM₁₀ in this study to obtain their concentrations in coarser particulate matter (PM_2.5–10). The total concentrations of MPs and plasticizers in PM_2.5 and PM_2.5–10 from five different sources are presented in Fig. 1. MPs were more enriched in PM_2.5–10 in the Fruit-bag Burning source (59 % of PM₁₀), while they were higher in PM_2.5 for the remaining four sources (Plastic Burning, Road Traffic, Agricultural Film, and Livestock Breeding). The fruit bags are coated with a wax layer to enhance the waterproofing and durability of the material. The presence of this wax layer may affect particle formation during combustion, contributing to the creation of larger agglomerates and thus a higher proportion of coarse particles. Notably, MPs in Plastic Burning and Livestock Breeding constituted a comparable proportion in both fine and coarse fractions, both close to 50 %. The variable order of MP concentrations in the five sources in PM_2.5–10 was roughly consistent with that of PM_2.5, showing Plastic Burning > Fruit-bag Burning > Livestock Breeding > Agricultural Film > Road Traffic. The average concentration of MPs ranged from 77.7 ± 25.3 (Road Traffic) to 1906 ± 587 (Plastic Burning) ng m⁻³ in the fine fraction and from 41.5 ± 11.7 (Road Traffic) to 1634 ± 20.3 (Plastic Burning) ng m⁻³ in the coarse fraction of PMs, as summarized in Table S2. The highest MP concentrations in fine and coarse PMs in the Plastic Burning source were both ∼ 5 times higher than the averages in other sources. One possible explanation for this is that plastic waste can be fragmented into MPs during the process of combustion (Yang et al., 2021; Luo et al., 2024a). Another important pathway for elevated MPs from the Plastic Burning source is the resuspension of bottom ash (Yang et al., 2021).

Figure 1 Average concentrations of MPs (a) and plasticizers (b) in PM_2.5 and PM_2.5–10 from five sources (PB: Plastic Burning, FB: Fruit-bag Burning, RT: Road Traffic, AF: Agricultural Film, LB: Livestock Breeding).

The total concentrations of the plasticizers in the samples were 1 order of magnitude higher than those of MPs (Table S2). The mass concentrations of plasticizers were higher in PM_2.5 than in PM_2.5–10 for Fruit-bag Burning, Road Traffic, Agricultural Film, and Livestock Breeding sources, especially with the value of 80 % in fine particles from Road Traffic. Both MPs and plasticizers in Road Traffic were more abundant in PM_2.5, which enhanced the potential for long-range transport and respiratory penetration. Therefore, even though the emission concentrations from the Road Traffic source were lower, the potential environmental and health risks posed by road traffic cannot be overlooked. Conversely, plasticizers in the Plastic Burning source were abundant in PM_2.5–10 (59 %). The highest concentration values of plasticizers in this study were also observed in Plastic Burning (15.6 ± 5.61 µg m⁻³ in PM_2.5, 22.3 ± 1.68 µg m⁻³ in PM_2.5–10), followed by Fruit-bag Burning (4.53 ± 0.39 µg m⁻³ in PM_2.5, 2.75 ± 0.65 µg m⁻³ in PM_2.5–10). This is because plastic products contain many additives to enhance their performance (Do et al., 2022). Many additives are not covalently bound to the polymer matrix, resulting in the liberation of plastic additives during crushing and combustion (Do et al., 2022; Billings et al., 2023). Furthermore, Livestock Breeding exhibited a higher emission for plasticizers compared with other non-combustion sources, but still much lower than the combustion sources (Road Traffic and Agricultural Film; P < 0.05 in PM_2.5), with the values of 2.17 ± 1.05 and 1.16 ± 0.88 µg m⁻³, respectively, for PM_2.5 and PM_2.5–10. The lack of an effective plastic recycling and disposal system under traditional retail farming may exacerbate the release of plasticizers.

3.2 Chemical composition of microplastics

The proportions of MPs identified in PM_2.5 and PM_2.5–10 for the five sources are presented in Fig. 2. The composition of MPs from five sources varied greatly, but there was no significant size distribution difference within the same source. Road Traffic exhibited distinctive features compared with the other four rural sources, with high proportions of both BR + SBR and NR in PM_2.5 (46.2 ± 3.31 % and 33.3 ± 2.65 % of MPs) and PM_2.5–10 (50.7 % ± 2.94 and 18.6 % ± 0.79 of MPs), which are the basic materials of tire treads. In previous studies, BR + SBR was observed to be the predominant MP in light-duty vehicle tires in the tunnel PM_2.5, and conversely, NR is extensively used in tire treads for trucks (Liu et al., 2023). The Road Traffic sample collection in this work was conducted at the downtown flyover in urban Xi'an, where light-duty cars are the dominant vehicle type, explaining the high proportion of BR + SBR compared with NR both in PM_2.5 and PM_2.5–10.

Figure 2 Chemical composition of microplastics in PM_2.5–10 and PM_2.5 from the five sources (PB: Plastic Burning, FB: Fruit-bag Burning, RT: Road Traffic, AF: Agricultural Film, LB: Livestock Breeding).

The MP compositions of Plastic Burning, Fruit-bag Burning, Agricultural Film, and Livestock Breeding were relatively similar. PET was the most common polymer type in MPs (Fig. 2), which is widely used in the production of textiles. The Plastic Burning source inevitably included a certain amount of waste textiles, inducing the release of PET (Yang et al., 2021). Moreover, PET is widely applied in packaging and agriculture due to its advantageous properties, such as good strength, durability, elasticity, and clarity (Liu et al., 2019; Lu et al., 2024). These materials may break into MPs due to wear and tear, subsequently discharging into the agricultural and breeding environment. Moreover, the highest proportion of PS was found in the Agricultural Film source. Agricultural facilities made of PS (e.g., lamp-chimneys, electrical devices) in greenhouses may influence the MP composition of the Agricultural Film source (Qi et al., 2023).

3.3 Chemical composition of plasticizers

PAEs were the most prevalent (>90 %) among the three plasticizers in the five sources in this study. PAEs have been the most widely used plasticizer, and the global production is expected to reach 500×10⁶ t by 2050 (Huang et al., 2021; Billings et al., 2024). The levels of the total PAEs ranged from 468 ± 175 ng m⁻³ (Agricultural Film) to 15 640 ± 5609 ng m⁻³ (Plastic Burning) in PM_2.5 and 115 ± 54.4 ng m⁻³ (Road Traffic) to 22 274 ± 1680 ng m⁻³ (Plastic Burning) in PM_2.5–10 (Table S2). The percentages of BTs and BPA among the three detected plasticizer types were below 2 %. The highest concentrations of PAEs, BTs, and BPA still appeared in PB among five sources. The concentrations of BPA in Plastic Burning and Fruit-bag Burning sources were 1 order of magnitude higher than those in Road Traffic and Agricultural Film (Table S2). The results indicated that Plastic Burning was the primary emission source of atmospheric plasticizers, in agreement with prior research (Zhen et al., 2019; Chandra and Chakraborty, 2023).

Compared to other sources, the Road Traffic source demonstrated a higher concentration of BTs (34.8 ± 13.0 ng m⁻³ for PM_2.5, P<0.05; 12.9 ± 7.28 ng m⁻³ for PM_2.5–10) (Table S2). This may be related to the widespread use of BTs in tire manufacturing, and these additives are released into the air during friction between tires and the road surface (Liu et al., 2023). At the same time, some tire rubber substances were also involved in the Plastic Burning source of this study. Livestock Breeding exhibited the highest emission of BPA among non-combustion sources, with values of 50.9 ± 27.1 and 38.6 ± 22.2 ng m⁻³, respectively, for PM_2.5 and PM_2.5–10, higher than those in Road Traffic (4.43 ± 1.45 and 7.8 ± 0.9 ng m⁻³, P<0.05) and Agricultural Film (1.4 ± 0.71 and 4.29 ± 6.68 ng m⁻³, P<0.05), partly due to the migration of BPA from animal feed plastic packaging into the air (Wang et al., 2021b, c). Furthermore, BTs, PAEs, and BPA from sources other than Plastic Burning were prevalent in PM_2.5 relative to PM_2.5–10, contrary to the results reported by Nunez et al. (2020). This discrepancy may be attributed to differences in pollution sources. Nunez et al. (2020) demonstrated that port industrial activities (e.g., cargo handling and industrial emissions) predominantly generated coarse PMs, resulting in higher concentrations of plasticizers in this fraction. In contrast, high temperatures in the Plastic Burning source promoted the formation of fine particles, with a larger surface area that enhanced the adsorption of plasticizers.

3.3.1 Compositions and distributions of PAEs

DnOP was the most abundant PAE species across Plastic Burning, Fruit-bag Burning, Road Traffic, and Agricultural Film. For the Fruit-bag Burning source, DnOP was significantly more prevalent in PM_2.5–10, accounting for 51 ± 12 % of the total PAEs, compared to 36 ± 1.8 % in PM_2.5. Conversely, DnOP was more abundant in the fine (59 ± 1.0 %) fraction of PMs than the coarse (44 ± 0.4 %) fraction in Road Traffic. As a common plasticizer, DnOP possesses a high molecular weight and low volatility, increasing its persistence in the environment. In addition to DnOP, DEHP and BBP were also identified as major components in the five sources. DEHP was the second most abundant PAE component in Road Traffic (23 ± 0.5 % and 30 ± 0.2 % of PAEs in PM_2.5 and PM_2.5–10), as it has a high consumption in the plasticizer market, especially in the automobile industry (Zhen et al., 2019). The lowest percentage of DEHP in Agricultural Film in both PM_2.5 and PM_2.5–10 (13 ± 0.1 %, 12 ± 0.3 %) among the five sources was a significant characteristic of Agricultural Film. BBP was the most abundant PAE in PM_2.5–10 in Livestock Breeding, and the proportion was higher in coarse (40 ± 14 %) than fine (28 ± 1.6 %) PMs. Moreover, as shown in Fig. 3, the proportions of the sum of DMP, DEP, and DBP were below 30 % in both PM_2.5 and PM_2.5–10 and were even below 15 % in Fruit-bag Burning and Road Traffic. The proportion of DEP (12 ± 5.1 % and 15 ± 0.4 % in PM_2.5 and PM_2.5–10, respectively) was the highest in Plastic Burning compared to other sources, which could be used as a source marker (Fig. 3a).

Figure 3 Mass proportions of phthalates (PAEs) (a, b, c, d, e) and benzothiazole and its derivatives (BTs) (f, g, h, i, j) in PM_2.5–10 (outer ring) and PM_2.5 (inner ring) from the five typical sources (PB: Plastic Burning, FB: Fruit-bag Burning, RT: Road Traffic, AF: Agricultural Film, LB: Livestock Breeding).

3.3.2 Compositions and distributions of BTs

The distribution patterns of BTs in the five typical MP sources in PM_2.5 and PM_2.5–10 were more different than those of PAEs. The compositions of Plastic Burning and Agricultural Film were quite similar, which may prove that rural households use discarded agricultural film for heating or cooking during indoor fuel combustion. HOBT was the most abundant compound in Plastic Burning and Agricultural Film, with values of 42 ± 1.5 % and 31 ± 0.0 %, respectively, for PM_2.5, and 28 ± 2.5 % and 35 ± 5.2 % for PM_2.5–10 (Fig. 3f, i). Furthermore, MBT was more prominent than other species for Fruit-bag Burning, with the values of more than 20 %. The abundances of OBS in PM_2.5–10 (36 ± 0.1 %) were higher than those in PM_2.5 (18 ± 1.8 %) for Road Traffic. Some previous studies have implied that OBS is mainly used in tire manufacture (Liao et al., 2018; Liu et al., 2023). BT in Road Traffic was more predominant in PM_2.5 (27 ± 1.0 %) compared to PM_2.5–10 (10 ± 1.0 %), aligning with the prevalence of MPs and plasticizers in PM_2.5 in Road Traffic. A high concentration of BT in tire debris was reported from Sweden, demonstrating that tire wear is the main cause of road traffic pollution (Avagyan et al., 2014). OBS + CBS accounted for more than 70 % of BTs only in the Livestock Breeding source, which was significantly higher than those in other sources and could be used as a source marker.

3.4 Source profiles of MPs and plasticizers

The source profiles of MPs, BTs, PAEs, and BPA in PM₁₀ and PM_2.5 emitted from the five emission sources are shown in Fig. 4. The distribution patterns of each chemical species exhibited insignificant differences between PM_2.5 and PM₁₀. DnOP emerged as the predominant contributor across all sources, with Plastic Burning being the most significant, representing 1.4 ± 0.33 % and 2.9 ± 0.06 % of PM_2.5 and PM₁₀ mass concentrations, respectively. The profiles of the combustion sources (Plastic Burning and Fruit-bag Burning) were more similar. However, PMMA exhibited a higher proportion in Plastic Burning (0.085 ± 0.033 % and 0.23 ± 0.01 % in PM_2.5 and PM₁₀, respectively) compared to Fruit-bag Burning (0.023 ± 0.001 % and 0.041 ± 0.004 %). In addition, HOBT, the most abundant BT derivative in the current study, accounted for 0.024 ± 0.015 % in PM_2.5 and 0.037 ± 0.003 % in PM₁₀ in Plastic Burning but less than 0.001 % in Fruit-bag Burning. For non-combustion sources, Road Traffic was significantly influenced by tire wear particles, which were characterized by high abundances of tire-related materials, such as NR (0.047 ± 0.005 %) and BR + SBR (0.072 ± 0.008 %). Liu et al. (2023) revealed that NR and other rubber particles were emitted at high levels in tunnel traffic, emerging as the dominant microplastics in traffic-dominated environments. Moreover, 24MoBT constituted the highest percentage of BTs in the Road Traffic source, which can also be used as an indicator of vehicle emissions plasticizer. PS constituted a higher proportion in PMs than PET in Agricultural Film, with 0.18 ± 0.04 % and 0.14 ± 0.02 %, respectively, for PM_2.5, and 0.16 ± 0.19 % and 0.15 ± 0.15 % for PM₁₀. This is in line with the findings of Liu et al. (2019), who documented that PS was the predominant MP. OBS and CBS were the prevalent BT compounds in Livestock Breeding, and BPA played a more significant contribution to PM₁₀ than PM_2.5.

Figure 4 Source profiles of microplastics and plasticizers in PM_2.5 and PM₁₀ (black arrows indicate source markers).

3.5 Eco-health significance

In this section, a comprehensive eco-health risk evaluation system was established to provide scientific support for estimating the hazards of MPs and plasticizers from different sources. (1) The transport pathways of MPs and plasticizers from Plastic Burning, Road Traffic, and Agricultural Film sources were analyzed to clarify the exposure routes from “source” to “receptor” (Fig. 5). (2) The ecological and health risks of MPs and plasticizers were assessed through different evaluation metrics (H, HI, ILCR, and oxidation potential).

3.5.1 Transport pathways of MPs and plasticizers

As shown in Fig. 5, plastic combustion emitted MPs and attached plasticizers into the ambient air (Velis and Cook, 2021); the residual in the bottom ash can break into MPs via wind abrasion, then be re-suspended in the air or deposited onto surrounding soil or into water with a risk of entering the food chain (Yang et al., 2021; Velis and Cook, 2021; Pathak et al., 2024). Small microplastics (micro-rubber) from Road Traffic are emitted as airborne fine particles or trapped in the road surface, from where they can enter the water by surface runoff, migrating and transforming across different environmental media (Kole et al., 2017). Under ultraviolet degradation and wind erosion, agricultural films can release MPs and plasticizers into the air directly, while larger particles are deposited in farmland (Song et al., 2017). Disturbed by agricultural activities and wind, MPs created by residual films in the soil may be re-suspended in the air (Brahney et al., 2021; Jin et al., 2022). These pathways are all possible routes for MPs to enter the human body, and controlling these pathways can reduce exposure levels.

Figure 5 Source–Pathway–Receptor model associated with three different MP and plasticizer sources.

3.5.2 Risk assessment of MPs

Based on the ecological risks of MPs for different sources (Table S6), Plastic Burning and Fruit-bag Burning were categorized as Level III (high risk). This may be attributed to the fact that PMMA, a compound with a high hazard score, accounted for a higher proportion of MPs emitted from combustion sources. In contrast, Road Traffic, Agricultural Film, and Livestock Breeding sources, with lower hazard scores, were categorized as Level II (lower risk).

In this study, the health risks of MPs and plasticizers in PM_2.5 and PM₁₀ from five sources were analyzed as well. The total non-carcinogenic risk (HI) ranged from 1.36 × 10⁻⁴ (Agricultural Film) to 5.20 × 10⁻⁴ (Livestock Breeding) in PM_2.5 and 2.01 × 10⁻⁴ (Road Traffic) to 8.96 × 10⁻⁴ (Livestock Breeding) in PM₁₀, inconsistent with the mass concentration ranking of MPs and plasticizers in various sources. All HI values of each source were significantly lower than the international safety threshold (HI = 1). The highest HI was observed in Livestock Breeding, followed by Plastic Burning, with values of 4.49 × 10⁻⁴ and 8.73 × 10⁻⁴ for PM_2.5 and PM₁₀, respectively. Figure S2 illustrates the contributions of different compounds to HI. PAEs contributed most significantly to HI, accounting for more than 60 % in most sources, especially in Livestock Breeding (93.6 % and 92.9 % for PM_2.5 and PM₁₀, respectively). Among all compounds DEHP, one of the PAEs, displayed the highest non-carcinogenic risk (Fig. S2). In Road Traffic, BT and MBT exhibited higher HI than in other sources, with BT accounting for 20.3 % (PM_2.5) and 18.1 % (PM₁₀), followed by MBT (6.5 % and 6.7 %, respectively). Moreover, PS in the Agricultural Film source exhibited prominent HI values, with proportions of 27.6 % and 26.3 % for PM_2.5 and PM₁₀, respectively, 10–77 times higher than those in other sources. These findings emphasize the need to focus on PAEs in Livestock Breeding, BT and MBT in Road Traffic, and PS in Agricultural Film as a priority for MP pollution control, aiming to minimize associated human non-carcinogenic risks.

ILCR for the three carcinogenic compounds (BT, BBP, and DEHP) was calculated in this study (Guyton et al., 2009; Ma et al., 2020; Liu et al., 2023). The ILCR values for each compound varied between 7.03 × 10⁻¹⁶ and 1.77 × 10⁻⁷, which were all below the safety threshold (10⁻⁶). However, this cannot be taken lightly, as there are many types of environmental pollutants and their carcinogenic risks are additive and cumulative. Compared to other sources, Livestock Breeding had the highest total ILCR values (∑ILCR) (1.01 × 10⁻⁷ in PM_2.5 and 1.8 × 10⁻⁷ in PM₁₀), although the mass concentration of MPs and plasticizers in this source was not the highest. Combined with the HI results, we can see that Livestock Breeding emitted higher concentrations of toxic MPs and plasticizers, thereby increasing the human health risk. Comparison of the carcinogenic risks of different compounds showed that DEHP accounted for more than 97 % of ∑ILCR in each source, which is the compound that needs to be controlled the most in this study.

3.5.3 Effect of MPs and plasticizers on oxidative potential

Figure S3 demonstrates the oxidative potential capacity of PM_2.5 and PM₁₀ from five sources. Overall, PM_2.5 exhibited a generally higher level of oxidative potential than PM₁₀. The larger specific surface area of PM_2.5 can enhance its reactivity with DTT and facilitate ROS production (Boogaard et al., 2012; Feng et al., 2016; Chirizzi et al., 2017). Moreover, the presence of certain components in PM_2.5–10 may actually weaken the ability of PM_2.5 components to induce ROS production (Boogaard et al., 2012; Chirizzi et al., 2017). This suggests a completely different mechanism for the generation of ROS between coarse and fine particles. Therefore, the results show that PM₁₀ has lower oxidative potential than PM_2.5 for most sources in this study (Plastic Burning, Fruit-bag Burning, Agricultural Film, and Livestock Breeding), which requires further research in the future.

PM_2.5 from Fruit-bag Burning exhibited the highest oxidative potential with a value of 77.0 ± 59.8 nmol min⁻¹ m⁻³, while its PM₁₀ DTT value was only 18.32 ± 8.27 nmol min⁻¹ m⁻³, indicating that the oxidative potential of Fruit-bag Burning was mainly driven by PM_2.5. For the Plastic Burning source, the DTT values of PM_2.5 and PM₁₀ were 58.6 ± 21.2 and 28.0 ± 23.7 nmol min⁻¹ m⁻³, respectively, both at relatively high levels. In contrast, PM_2.5 from road sources showed a low oxidative potential (0.75 ± 0.09 nmol min⁻¹ m⁻³), and Road Traffic was the only source with a higher oxidative potential for PM₁₀ than PM_2.5. This is likely attributed to the unique characteristics of road dust, which is rich in coarse particles (Boogaard et al., 2012; Pant et al., 2015; Shirmohammadi et al., 2017). Road dust contains a high concentration of metal compounds, which catalyze the DTT consumption (Shirmohammadi et al., 2017). Future research should focus on the size dependency of oxidative potential for different sources, as it has significant implications for health impacts.

To investigate the impact of MPs and plasticizers on oxidative potential, Spearman correlation analysis was employed to assess the relationships between these compounds and DTT. As shown in Fig. 6, PMMA (R=0.77, P<0.01), PET (R=0.72, P<0.05), and PE (R=0.72, P<0.05) showed a positive correlation with DTT, indicating that these components significantly enhance the oxidative potential of PMs. Additionally, all PAE species exhibited significant positive correlations with DTT (R = 0.70–0.77, P<0.05), especially DnOP (R=0.77, P<0.01) and DEHP (R=0.76, P<0.05). The significant association of BPA with DTT (R=0.70, P<0.05) further corroborated its established impact on oxidative damage (Zhang et al., 2022). However, among BTs, only NCBA showed a weak correlation with DTT (R=0.65, P<0.05). These findings suggest that BTs may contribute minimally to oxidative stress, whereas PMMA, PET, PE, BPA, and PAEs are the main drivers of oxidative potential.

Figure 6 Correlations between DTT, MPs, and plasticizers (^* P<0.05; ${}^{**}$ P<0.01).

4 Conclusion

In this study, the five typical plastic emission sources in the Guanzhong Plain, China, were selected to investigate the characteristics of MPs and plasticizers in PM_2.5 and PM_2.5–10. The concentration levels of MPs and plasticizers in combustion sources (Plastic Burning and Fruit-bag Burning) were higher than those in non-combustion sources (Road Traffic, Agricultural Film, and Livestock Breeding), highlighting the necessity of tightening plastic combustion regulations to address atmospheric MP pollution. Most detected MPs and plasticizers were more abundant in PM_2.5 than PM_2.5–10 for most sources. The Plastic Burning source was recognized for high loadings of HOBT, PMMA, and DEP. Fruit-bag Burning exhibited high abundances of DnOP, which were higher in PM_2.5–10 than in PM_2.5. Since tire wear particles are one of the main sources of Road Traffic MPs, rubber compositions (NR, BR + SBR) accounted for the highest proportions. Agricultural Film was mainly characterized by a high abundance of PS. The high proportions of OBS and CBS distinguished Livestock Breeding from the other sources, and there are still many unknown aspects of Livestock Breeding sources that require further research attention. This study develops a complete eco-health risk assessment system, identifying combustion sources (Plastic Burning and Fruit-bag Burning) as high-risk ecological emitters, Livestock Breeding as a high health risk contributor, and DEHP as a key health-damaging pollutant due to its combined non-carcinogenic risk, carcinogenic risk, and oxidative potential effects.

Our results could contribute to providing a scientific foundation for accurately identifying the sources and risks of atmospheric MPs and developing efficient management strategies. The single-shot pyrolysis protocol used in detecting MPs in this study enables quasi-instantaneous and homogenous pyrolysis (Seeley and Lynch, 2023). However, the complex products of the combustion process may lead to interference from non-polymeric components during single pyrolysis. Therefore, future work should systematically compare different pyrolysis approaches in order to improve the accuracy of MP detection for complex samples. Additionally, future studies should expand the range of assessed MPs and plasticizers, integrate multiple ecological health assessment methods to further refine the health risk assessment system, and deepen the understanding of the environmental and health hazards of MPs.