Source Apportionment, Environmental Impact and Health Risk Assessment of Volatile Organic Compounds in Dezhou, North China

Abstract

Volatile organic compounds (VOCs) are the main precursors of ozone (O₃) and secondary organic aerosols (SOAs), posing a significant threat to the environment and human health. In this study, the characteristics, potential for the generation of O₃ and SOAs, sources and health risks of volatile organic compounds were investigated based on annual monitoring data for 2021 in Dezhou, which is located in the northwest of Shandong Province, China, and adjacent to the Beijing–Tianjin–Hebei region. The results showed that the ambient VOC concentrations were lower in spring and summer and higher in autumn and winter. The species contributing to the O₃ formation potential (OFP) and SOA formation potential (SOAFP) varied by season, but those with high contributions were all olefins and aromatic hydrocarbons. O₃ pollution occurred frequently in summer, with a proportion of 40.22%. Focusing on summer, six sources were identified through the positive matrix factorization (PMF) model, with the petrochemical industry (26.1%) and combustion (25.6%) being the primary sources. The non-carcinogenic risk values of the involved 18 toxic VOCs were all within the safety threshold, while the carcinogenic risk of benzene was regarded as low-probability under long-term exposure. This study provides significant support for targeted control of air pollutant emissions and improvement in regional air quality.

1. Introduction

Volatile organic compounds (VOCs), as key trace pollutants in the atmosphere, profoundly influence regional air quality, the equilibrium of the climate system and human health [1,2,3,4]. On the one hand, VOCs undergo photochemical reactions with nitrogen oxides under specific solar radiation conditions, serving as core precursors for the formation of ozone (O₃) [5,6,7]. High concentrations of O₃ not only trigger respiratory diseases in humans and damage vegetation productivity but also exacerbate the greenhouse effect. On the other hand, VOCs undergo oxidation and condensation processes to form secondary organic aerosols (SOAs) [8,9]. As a significant component of fine particulate matter (PM_2.5), SOAs influence regional climate by scattering and absorbing solar radiation [10,11]. Moreover, they readily adsorb harmful substances such as heavy metals and polycyclic aromatic hydrocarbons, posing dual threats to both the ecological environment and human health [12,13].

Currently, with the advancement of industrialization and adjustments to energy consumption structures, VOC emission sources exhibit increasingly diverse and complex characteristics [14]. The contribution mechanisms, key influencing factors, and regulatory pathways of VOCs to O₃ and SOA formation have become both research hotspots and challenges in the field of atmospheric environmental science. Previous research has been carried out to investigate the VOC–NOx–O₃/VOC–SOA relationship based on observations and/or simulations [1,15,16,17]. Wang et al. examined the seasonal distribution characteristics of VOCs, their pollution sources, and their impact on O₃ by employing machine learning algorithms based on data collected in Yinchuan, a high-altitude city [15]. Zhang et al. analyzed continuous variation in O₃ precursors and meteorological parameters in the suburban area of Shanghai, identified the sources of VOCs using an observation-based model, then discovered that NOx and olefins exhibited the most significant negative and positive influences on O₃ formation, respectively [16]. Borlaza-Lacoste et al. employed an improved positive matrix factorization (PMF) method, combining atmospheric dispersion and photochemical reaction losses of VOCs, to analyze VOC observation data from the Bronx, New York, for the period 2000–2001. They investigated the role of VOC sources in the formation of O₃ and SOAs, utilizing advanced machine learning algorithms to identify synergies between sources and to quantify the impact of VOC sources on ambient O₃ concentrations [1]. Wang et al. integrated the monitoring data of O₃ and VOCs in Zhengzhou, China, in summer from 2014 to 2020 and analyzed the pollution trend of O₃, VOC components, sources and sensitivity based on an observation-based box model and identified six sources of VOCs through a PMF model. The study found that the composition of average VOC concentrations on polluted days was dominated by alkanes, with motor vehicle emissions and industrial emissions being the primary sources of VOCs [17].

In China, the Beijing–Tianjin–Hebei region and its surrounding areas are key regions for the prevention and control of complex air pollution, with the combined pollution of PM_2.5 and O₃ being a particularly pressing issue [18,19,20,21]. Dezhou City, located at the intersection of the Shandong Provincial Capital Economic Circle and the Beijing–Tianjin–Hebei region, plays a significant role in the transmission, migration and transformation of regional air pollutants and is closely linked to changes in air quality and the coordinated prevention and control of air pollution in the Beijing–Tianjin–Hebei region and surrounding areas. The industrial structure of Dezhou is relatively heavy, mainly consisting of industries such as chemical engineering, equipment manufacturing and new materials, and is highly dependent on energy. According to the data from “2021 Dezhou City Ecological Environment Quality and Pollution Emission Status”, the number of days on which ambient air quality met standards in Dezhou in 2021 was 233, representing a ‘good’ or ‘excellent’ air quality rate of 63.8%, an increase of 4% compared with the previous year. The comprehensive air quality index stood at 4.56, an improvement of 13.1% compared with the previous year. Industrial pollution sources were the primary source of air pollutants in Dezhou, followed by motor vehicle emissions and domestic sources.

Previous studies [18,19,20,21] have largely focused on the core region of the Beijing–Tianjin–Hebei area, and these studies provided a crucial theoretical basis and methodological guidance for the present study. However, there have been few reports on air pollution in Dezhou City, particularly regarding the emission characteristics and source apportionment of VOCs. Consequently, conducting research into the characteristics, distribution patterns and source apportionment of VOC pollution in Dezhou will not only provide support for the precise management and scientific control of local VOC pollution but will also further improve the basic research framework for VOC pollution in cities along the Beijing–Tianjin–Hebei air pollution transmission channel.

In this study, the characteristics of VOCs and the seasonal impacts on O₃ and SOA formation were first investigated based on the annual monitoring data of Dezhou in 2021. Then, the intrinsic correlations of VOCs with O₃ formation were further discussed, and the main sources of VOCs during summer were specially identified through the PMF model. After that, a health risk assessment was conducted on toxic VOCs. This study has significant scientific and practical value for promoting the coordinated control of O₃ and PM_2.5 in both the local area and the Beijing–Tianjin–Hebei region, as well as for fostering the continuous improvement of air quality.

2. Materials and Methods

2.1. Study Site

This study is based on the long-term online monitoring data from the Dezhou Atmospheric Environment Monitoring Supersite (DZS) (116.63° E, 37.46° N). Figure 1 shows the location of the Beijing–Tianjin–Hebei region, Dezhou and the DZS. One leading local petrochemical company and two major fiberglass manufacturers located near the DZS are also marked on the map. The transportation network around the DZS is well developed, and the DZS is situated immediately adjacent to a major urban thoroughfare.

The region has a temperate monsoon climate, with cold and dry winters and hot and rainy summers. In winter, the cold and dry weather makes it easier for temperature inversions to form, which in turn facilitates the accumulation of pollutants. The four seasons are distinct, and precipitation is unevenly distributed throughout the year. Seasonal variations in wind direction and speed in 2021 are shown in Figure 2a. Westerly winds (WSW/W) prevail in spring and winter, with relatively higher speeds (mainly 1–3 m/s), while easterly winds (E) dominate in summer and autumn, mostly at low speeds (0–2 m/s). Calm winds are most frequent in autumn (425 counts) and least frequent in spring (115 counts). Figure 2b shows the time series of temperature and relative humidity observed at DZS. The temperature exhibits a typical unimodal pattern, with the lowest values in winter (−10–10 °C), a steady increase in spring (0–25 °C), the highest and most stable values in summer (25–32 °C), and a gradual decrease in autumn (0–30 °C). Relative humidity is generally higher in summer and autumn (mostly 60–95%), while showing larger fluctuations and lower average levels in winter and spring (25–90%). A clear inverse diurnal relationship between temperature and humidity was observed throughout the year.

2.2. Measurement of VOCs

The observation of VOCs was conducted using a GC-MS/FID (8890-5977C, Agilent Technologies, Santa Clara, CA, USA) ambient air quality continuous monitoring system. The system mainly consists of a sampling unit (KORI-TT24-7-xr, Markes International, Bridgend, UK), a pre-concentration system (TT24-7xr, Markes International) and a dual-channel detection system (MS: 5977C, Aglient Technologies; FID: TRACE 1300, Thermo Fisher Scientific, Waltham, MA, USA). Samples such as ambient air and standard gases were first automatically routed at low flow rates via a sample valve into a switchable dual-channel (FID analysis channel and MS analysis channel) gas trapping system. The VOCs were trapped, dried and secondary-concentrated by a two-stage ultra-low-temperature (−160 °C) enrichment system, after which the samples were rapidly heated for desorption, before finally entering the GC-FID and GC-MS systems for quantitative analysis. The low-carbon components (C₂–C₄) were detected by a flame ionization detector and quantitatively analyzed by the external standard method. The high-carbon components were detected by a mass spectrometer detector and quantitatively analyzed by the internal standard method. The standard gas used in this research was commercial mixed standard gas (100 ppb, CNEMC Mix, Linde, Munich, Germany). The standard gas was diluted to 0.5–10 ppb to establish a seven-point monthly calibration curve, with the recovery rate of 80% of components being between 70 and 130%, and the correlation coefficient of 90% of components being above 0.99.

Instrument data quality control primarily comprised daily calibration quality control and weekly flow rate checks. A 2 ppb standard gas was introduced to perform daily quality control. The analysis concentration results of each component in the FID channel and the MS channel should be within ±15% and ±30% of the theoretical concentration. Flow rate checks were conducted once a week. When the error between the sampling volume and the set value exceeded ±11%, the gas path was inspected, and the volume was corrected to ensure that the sampling flow rate was normal.

A total of 56 VOC species, including 29 alkanes, 15 aromatic hydrocarbons, 11 olefins, and 1 alkyne, were measured to obtain the hourly variation of individual VOCs. Continuous measurements were performed from 1 January to 31 December 2021.

2.3. Measurement of Meteorological Parameters and Criteria Air Pollutants

The meteorological parameters, including temperature (T), relative humidity (RH), atmospheric pressure (P), wind direction (WD) and wind speed (WS), were sourced from the DZS and were measured using an integrated automatic weather station (Vantage Pro2, Davis Instruments). Data on common atmospheric pollutants were sourced from local national ambient air quality monitoring stations. Specifically, O₃ was measured using the ultraviolet photometric method specified in the National Environmental Protection Standard of the People’s Republic of China (HJ 590-2010), while NOx was measured using the N-(1-naphthyl) ethylene diamine dihydrochloride spectrophotometric method specified in the National Environmental Protection Standard of the People’s Republic of China (HJ 590-2009).

2.4. Data Processing and Analysis

2.4.1. Data Processing

Data quality control was carried out based on the data of 56 VOCs obtained from the DZS. The theoretical data volume for 2021 should have been 8760 sets; however, 1570 sets were actually missing due to equipment malfunctions and instrument calibrations, resulting in a missing data rate of 17.9%. Firstly, missing values were eliminated. Subsequently, data below the detection limits (MDL) were replaced with 0.5 times the detection limit. There were 7190 sets of data remaining after quality control.

Before the PMF modeling, species for which the sum of the missing ratio and the below-MDL ratio was higher than 50% and not belonging to the characteristic markers of various pollution sources were excluded. Ultimately, 38 VOC components with a few missing samples and clear tracer significance were selected for subsequent PMF modeling. The list of the selected species can be found in the Supplementary Materials (Table S1). A total of 1318 data sets were screened as input data for the source apportionment of VOCs in summer, with a missing data rate of 17.9%, a below-MDL ratio of 21.01% and a data validity rate of 78.99%.

2.4.2. Calculation of O₃ Formation Potential

The atmospheric chemical reactivity of different VOCs varies significantly, and their contributions to the formation of O₃ are also different. O₃ formation potential (OFP) represents the contribution of VOCs to O₃ generation under optimal reaction conditions. It can be calculated based on the concentration of VOCs in the ambient and the maximum incremental reactivity (MIR), thereby identifying key active components. The calculation formula is as follows:

$${\mathit{OFP}}_{\mathit{i}}={\mathit{VOC}}_{\mathit{i}}\times {\mathit{MIR}}_{\mathit{i}}$$

(1)

where OFP_i represents the O₃ formation potential of component i under the condition of maximum increment reaction (µg·m⁻³), VOC_i represents the concentration of component i in the atmosphere (µg·m⁻³), and MIR_i represents the maximum incremental reactivity of component i. MIR is derived from the reported findings of Carter [22].

2.4.3. Calculation of ·OH Reaction Rate

VOCs and ·OH can undergo a series of complex photochemical reactions, resulting in the formation of O₃; consequently, the reaction rate of ·OH is commonly used to assess the reactivity of VOCs in the study area and their contribution to daytime O₃ formation. The calculation formula is as follows:

$$R_{OHi}={VOC}_i\times K_{OHi}$$

(2)

where R_OHi represents the rate of ·OH consumption for VOC component i (s⁻¹), VOC_i represents the concentration of component i in the atmosphere (mol·cm⁻³), and K_OHi represents the ·OH reaction rate constant for the component i (cm³·mol⁻¹·s⁻¹). The K_OHi values used in this study were taken from the work of Atkinson et al. [23].

2.4.4. Calculation of Secondary Organic Aerosol Formation Potential

In this study, the aerosol generation coefficient (FAC) method was adopted to estimate the secondary organic aerosol formation potential (SOAFP). A widely used set of FAC values was proposed by Grosjean et al. [24] based on a comprehensive analysis of smog chamber experimental data and atmospheric chemical kinetics data; its applicability in open-air environments has also been confirmed by extensive research [25,26,27,28]. This method estimates SOAFP based on the VOC emission inventory or the measured ambient VOC concentrations and can reflect the relative contributions of different SOA precursors [24]. The calculation formula is as follows:

$${\mathit{SOAFP}}_{\mathit{i}}={\mathit{VOC}}_{\mathit{i},\mathit{initial}}\times {\mathit{FAC}}_{\mathit{i}}$$

(3)

$${\mathit{VOC}}_{\mathit{i},\mathit{initial}}={\mathit{VOC}}_{\mathit{i}}/(1-{\mathit{FVOCr}}_{\mathit{i}})$$

(4)

where SOAFP_i represents the secondary organic aerosol formation potential of component i (µg·m⁻³), VOC_i,initial represents the initial concentration of component i emitted by the source, VOC_i represents the concentration of component i in the atmosphere (µg·m⁻³), FAC_i represents the SOA generation coefficient of component i, and FVOCr_i represents the percentage of component i involved in the reaction. FAC and FVOCr values refer to the published literature [24,29].

2.4.5. Positive Matrix Factorization Model

The PMF model is a multi-factor analysis model. Its underlying logic is to decompose the pollutant concentration data matrix (X) into two non-negative matrices: a factor contribution matrix (G) and a factor spectrum matrix (F). By identifying the factor spectral matrix, the quantitative contribution of each factor to the sample is calculated. In this study, the PMF model was utilized to conduct source apportionment of VOCs in the atmospheric environment based on the observational data of Dezhou in the summer of 2021. Characteristic factors were extracted from the observed chemical composition data of VOCs. These factors were categorized into different pollution source types based on the identified components, and then the contribution values of different factors or pollution source types to the environment were calculated through multiple linear regression.

PMF 5.0, which was developed by the United States Environmental Protection Agency (U.S.EPA) was used to analyze the VOCs obtained from observations. During the analysis, a weighting matrix (Q) was introduced, and the value of the “objective function” Q was minimized through the weighted least squares iteration, ensuring that the data points with smaller errors were given priority in the decomposition process. The relevant calculation formula is as follows:

$${\mathit{x}}_{\mathit{ij}}={\displaystyle \sum _{\mathit{k}=1}^{\mathit{p}}}{\mathit{g}}_{\mathit{ik}}{\mathit{f}}_{\mathit{kj}} +{\mathit{e}}_{\mathit{ij}}$$

(5)

$$\mathit{Q}={{\displaystyle \sum _{\mathit{i}=1}^{\mathit{n}}}{\displaystyle \sum _{\mathit{j}=1}^{\mathit{m}}}\left({\displaystyle \frac{{\mathit{x}}_{\mathit{ij}}-{\displaystyle {\sum }_{\mathit{k}=1}^{\mathit{p}}}{\mathit{g}}_{\mathit{ik}}{\mathit{f}}_{\mathit{kj}}}{{\mathit{u}}_{\mathit{ij}}}}\right)}^2$$

(6)

$${\mathit{u}}_{\mathit{ij}}=\left\{\begin{array}{l}{\displaystyle \frac{5}{6}}\times \mathit{MDL}\\ \sqrt{(\mathit{EF}\times \mathit{Conc}.{)}^2+(0.5\times \mathit{MDL}{)}^2}\end{array}\right.$$

(7)

where x_ij represents the concentration of component j in sample i, g_ik represents the concentration of the k-th source factor in sample i, f_kj represents the mass fraction of component j in the k-th source factor, e_ij represents the residual concentration of component j in sample i, p represents the number of source factors, Q is the objective function, u_ij is the model uncertainty, MDL is the method detection limit, and EF represents the component concentration error coefficient, with a setting value of 10% in this study.

2.4.6. Health Risk Assessment

Inhalation of toxic VOCs in ambient air can pose a risk to human health [30]. This study employed the reference concentration (RfC) assessment method recommended by the U.S.EPA and the International Agency for Research on Cancer’s Integrated Risk Information System (IRIS) to estimate the hazard quotient (HQ) and risk value (Risk) for toxic VOCs [27,28,31]. It assessed the non-carcinogenic and carcinogenic risks to the general adult population in the study area from toxic VOCs in outdoor ambient air in 2021, using the following calculation formulae:

$$\mathit{EC}={\displaystyle \frac{\mathit{CA}\times \mathit{ET}\times \mathit{EF}\times \mathit{ED}}{\mathit{AT}}}$$

(8)

$$\mathit{HQ}={\displaystyle \frac{\mathit{EC}}{\mathit{RfC}\times 1000}}$$

(9)

$$\mathit{HI}=\sum \mathit{HQ}$$

(10)

$$\mathit{Risk}=\mathit{EC}\times \mathit{IUR}$$

(11)

In the formula, EC represents the exposure concentration (μg·m⁻³), CA represents the mass concentration of toxic VOCs (μg·m⁻³), ET represents the exposure time (2.73 h for spring, 3.65 h for summer, 2.73 h for autumn and 1.83 h for winter, according to the “Highlights of the Chinese Exposure Factors Handbook (Adults)” [32]), EF represents the exposure frequency (365 d·a⁻¹ for outdoor atmospheric environment exposure), ED represents the duration of exposure (10 a, 20 a, 30 a, 40 a, 50 a, 60 a and 70 a were taken, respectively, in this study), AT represents the average time (669,789.6 h, the number of hours corresponding to 76.46 a) [32], HQ is the hazard entropy value, RfC is the reference concentration for inhalation of VOCs (mg·m⁻³), HI is the hazard index, Risk is the carcinogenic risk value for species i under exposure, and IUR is the unit of inhalable carcinogenic risk ((μg·m⁻³)⁻¹). The RfC and IUR values were obtained from the U.S.EPA IRIS [33], the Agency for Toxic Substances and Disease Registry (ATSDR) [34], and the Office of Environmental Health Hazard Assessment (OEHHA) [35].

3. Results and Discussion

3.1. Distribution Characteristics of VOCs

The monthly variation and seasonal distribution of TVOCs are shown in Figure 3. In Figure 3a, the upper and lower boundaries of the box represent the upper quartile and lower quartile of the data, respectively; the horizontal line within the box represents the median; and the length of the box is the interquartile range, which reflects the degree of central tendency of the data. The whiskers represent the range between the minimum and maximum values of the non-outlier data. Among the twelve months, the volume fraction of TVOCs is the lowest in July (14.21 × 10⁻⁹) and the highest in January (50.46 × 10⁻⁹). Data fluctuations are more significant in the autumn and winter months than in the spring and summer months. In Figure 3b, the numbers marked for each season represent the sum of means for each month within that season. As shown, TVOC volume fractions are lower in spring and summer than those in autumn and winter. March (25.05 × 10⁻⁹) has the highest TVOC contribution to spring, June (23.97 × 10⁻⁹) to summer, November (36.22 × 10⁻⁹) to autumn, and January (50.46 × 10⁻⁹) to winter. Ranked from low to high, the monthly average values for each season are as follows: summer (19.44 × 10⁻⁹), spring (20.81 × 10⁻⁹), autumn (28.56 × 10⁻⁹), and winter (40.21 × 10⁻⁹). The top 10 VOC components by concentration are shown in Figure 3c, comprising five alkanes (ethane, propane, n-butane, iso-pentane and iso-butane), two olefins (ethylene and propylene), one alkyne (acetylene) and two aromatic hydrocarbons (toluene and benzene). These 10 components account for 81% of the TVOC concentration. The seasonal distribution of VOCs by category is shown in Figure 3d. Among the different categories of VOCs, alkanes accounted for the highest proportion in all four seasons (58.29% in spring, 53.76% in summer, 49.57% in autumn, and 51.77% in winter), followed by olefins (18.64% in spring, 18.83% in summer, 28.19% in autumn, and 27.73% in winter). Consequently, the two categories of VOCs that accounted for the highest proportion of ambient VOCs in Dezhou in 2021 were alkanes and olefins.

3.2. The Impact of VOCs on O₃ and SOA Formation

To assess the impacts of different VOCs on O₃ and SOAs, the OFP and SOAFP of VOCs were calculated, respectively. Figure 4a shows the top 10 components in each season. It was shown that the total contribution values and key VOCs to O₃ and SOAs varied obviously by season.

For OFP, the total value of the top 10 species in spring was 100.47 μg·m⁻³, with the three highest contributing species being ethylene (34.90 μg·m⁻³), m,p-xylene (18.99 μg·m⁻³) and toluene (11.81 μg·m⁻³). In summer, the total value of the top 10 species was 138.11 μg·m⁻³, among which isoprene (34.15 μg·m⁻³), ethylene (28.80 μg·m⁻³) and m,p-xylene (23.12 μg·m⁻³) were the top three contributors. In autumn, the total value was 207.85 μg·m⁻³, and the top three species were ethylene (85.90 μg·m⁻³), m,p-xylene (33.32 μg·m⁻³) and isoprene (19.36 μg·m⁻³). Lastly, in winter, the total value was 233.66 μg·m⁻³, with ethylene (111.24 μg·m⁻³), m,p-xylene (31.38 μg·m⁻³) and propylene (25.65 μg·m⁻³) ranking the highest contributions. In summary, the total value of the top 10 species in each season is ranked as follows: winter > autumn > summer > spring. Ethylene ranks first in winter, autumn and spring, with an OFP value significantly higher than that of other species. Isoprene ranks first in summer, with an OFP value slightly higher than that of ethylene, which ranks second. Meanwhile, combining the contribution rates of different types of VOC components to OFP in Figure 4b, it could be determined that olefins and aromatic hydrocarbons are the key active components for O₃ generation. It should be noted that the OFP value represents the possibility of O₃ generation. Therefore, although the total value of OFP is the highest in winter, it does not represent the actual level of O₃ pollution. The O₃ formation also depends on meteorological conditions such as sunlight intensity, temperature, humidity and dispersion conditions [19,20].

For SOAFP, the key species mainly included toluene, m,p-xylene, o-xylene, benzene and ethylbenzene. The total value of the top 10 species varied significantly with the seasons in the following order: autumn (104.61 μg·m⁻³) > winter (102.04 μg·m⁻³) > summer (89.41 μg·m⁻³) > spring (60.68 μg·m⁻³). The top three species are the same in summer and autumn (m,p-xylene, toluene and o-xylene), and the top three species are the same in spring and winter (toluene, m,p-xylene and o-xylene). In combination with the results in Figure 4b, it can be determined that aromatic hydrocarbons are the key active components for SOA generation.

The seasonal variations in VOC concentration, OFP, and SOAFP were closely linked to meteorological conditions, and further affected regional air quality. Autumn had the highest calm wind frequency and the worst dispersion conditions, while winter also presented weak wind and frequent temperature inversion; both seasons facilitated VOC accumulation, resulting in top-ranked VOC concentration and OFP, which greatly increased the potential risk of O₃ pollution. Due to moderate temperature and sufficient precursors, autumn showed the highest SOAFP and became the critical period for SOA pollution. By contrast, strong photochemical reactions in summer depleted VOCs substantially, thus lowering OFP and SOAFP. With favorable dispersion conditions and insufficient photochemical reactivity, spring had the lowest OFP and SOAFP, corresponding to the optimal air quality among the four seasons.

3.3. ·OH Reaction Rate

Figure 5 shows the top ten VOCs ranked by ·OH reaction rate (R_OH) across the four seasons. In spring, ethylene was the most reactive (R_OH = 1.18 s⁻¹), followed by propylene, styrene, and m,p-xylene. In summer, isoprene dominated (R_OH = 4.73 s⁻¹), far exceeding other species. Ethylene and isoprene were the top contributors in autumn, both with R_OH > 2.5. In winter, ethylene showed the highest reactivity (R_OH = 3.75 s⁻¹), followed by isoprene and propylene. Among the top ten species of R_OH and OFP in spring, summer, autumn and winter, the number of species that existed in both lists was seven, eight, eight and eight, respectively, showing a high degree of similarity. Olefins (ethylene, propylene, and isoprene) consistently dominated ·OH reactivity across all seasons, while isoprene peaked in summer and ethylene in winter. This result was consistent with the relatively high contribution rate of olefins to OFP in each season. Furthermore, the ranking of the total value of R_OH in the four seasons was: winter (12.46 s⁻¹) > autumn (11.39 s⁻¹) > summer (11.34 s⁻¹) > spring (5.04 s⁻¹), which was also consistent with the seasonal ranking of the total OFP values.

3.4. Atmospheric Photochemical Sensitivity Analysis

Confirming whether O₃ formation is in the VOC-limited or NOx-limited zone is crucial for the management of precursor emissions. The ratio of VOCs to NOx can be used to determine which precursor control zone O₃ is in. When the VOC-to-NOx ratio exceeds 8, O₃ formation is governed by the NOx-limited zone; when the VOC-to-NOx ratio is between 4 and 8, it is in a mixed control zone for both VOCs and NOx; when the VOC-to-NOx ratio is less than 4, reducing ambient VOC concentrations can effectively suppress O₃ formation [36].

Figure 6 showed the results of atmospheric photochemical sensitivity in the study area during each season. Each point represented the daily average volume fraction of TVOCs, the daily average volume fraction of NOx and the maximum daily 8 h average (MDA8) O₃ concentration. According to the concentration limits for Grade II areas set out in the Chinese National Ambient Air Quality Standards (GB 3095-2026) [37], days are defined as O₃-pollution days if the MDA8 concentration of O₃ exceeds 160 µg/m³. Conversely, if the concentration is below that value, it is defined as an O₃-clean day in this study. As shown, the ratios of VOCs to NOx in both spring and autumn were all less than 4. The number of O₃-pollution days in spring and autumn is nine and five, respectively. The ratios of VOCs to NOx in summer and winter were mostly less than 4. A small number of points were located in the mixed control zone and the NOx control zone. The number of O₃-pollution days in summer and winter was 37 and 0, respectively. This indicated that in 2021, summer was the season in which O₃ pollution occurred most frequently, with the proportion of O₃-pollution days being 40.22%. However, it should be noted that the generation of O₃ is a highly nonlinear process; the result here is only presented as a preliminary auxiliary indicator. Combined with the analysis results of R_OH in Section 3.3, the ranking of the total values of R_OH in the four seasons was: winter (12.46 s⁻¹) > autumn (11.39 s⁻¹) > summer (11.34 s⁻¹) > spring (5.04 s⁻¹). The lower R_OH values in summer and spring corresponded to the frequent O₃ pollution. A lower R_OH value indicated a higher proportion of VOCs undergoing photochemical reactions to transform into other pollutants such as O₃.

The VOC components on O₃-clean days and pollution days in summer were also analyzed, and the top ten species with OFP were identified. The results are shown in Figure 7. In the summer of 2021, the VOCs in ambient air in Dezhou on O₃-clean days consisted mainly of alkanes, accounting for 53.15%, followed by alkynes and olefins, accounting for 17.17% and 16.81%, respectively, while aromatic hydrocarbons accounted for a proportion of 12.87%. On O₃-pollution days, alkanes were the predominant component, accounting for 54.39%, followed by olefins and alkynes, accounting for 20.91% and 12.50%, while aromatic hydrocarbons accounted for a proportion of 12.21%. In terms of the proportion of VOCs by category, there was no significant difference between O₃-clean days and O₃-pollution days.

The hybrid single-particle Lagrangian integrated trajectory (HYSPLIT) model was further employed to analyze the airflow trajectories on typical O₃-clean days and O₃-pollution days in the city during the summer, and the results are shown in Figure S1. The results indicated that O₃-pollution days were dominated by slow-moving air masses transported from the southwest, which carried abundant O₃ precursors and favored pollutant accumulation, while O₃-clean days were controlled by fast-moving, clean southeasterly marine air masses. The difference in transport pathways provided a reasonable explanation, to some extent, for the similar VOC category proportions observed on both types of days. Specifically, while the VOC category proportions remained stable, the southwestward long-range transport on O₃-pollution days introduced additional reactive precursors, which, together with poor diffusion conditions, ultimately led to elevated O₃ levels.

As shown in Figure 7b, on O₃-clean days, the top three species by OFP were isoprene, m,p-xylene and ethylene, whereas on O₃-pollution days, ethylene rose to the top, followed by isoprene and m,p-xylene. This result confirmed the finding shown in Figure 4b that olefins were the species with the highest contribution to OFP in summer and further demonstrated that ethylene was the key active precursor to summer O₃ pollution. In the industrial structure of Dezhou, the petrochemical industry is a traditional pillar. And as ethylene serves as a tracer for petrochemical enterprises [17], the results suggest that petrochemical-related ethylene emissions may contribute to summer O₃ formation. To reduce the occurrence of O₃ pollution, it might be necessary to strengthen controls on petrochemical emissions.

3.5. Source Apportionment of VOCs Through the PMF Model

Since O₃ pollution occurred frequently in the study area during summer, the sources of VOCs in the ambient air during summer were specifically analyzed through the PMF model based on the selected 38 VOC components. Upon inputting the initial concentration data and uncertainties, the TVOC signal-to-noise ratio (S/N) was found to be greater than 8.2. Following multiple model fittings, the results were found to be relatively stable when six factors were selected. The model’s robustness parameter (Q_robust) was greater than the true parameter (Q_ture), and there was no correlation between the factors. The fitting equation between the model’s predicted values and the observed values is y = 0.84203x + 2.78091, with R² being 0.80011 and SE being 4.91378. The residual values for most species fall within the range of −3 to 3, indicating that the credibility of the model’s calculation results was relatively high. The PMF source apportionment results are shown in Figure 8.

For Factor #1, aromatic hydrocarbons contribute the most: o-xylene (85.98%), m-ethyltoluene (83.12%), m,p-xylene (72.40%), 1,2,4-tri-m-benzene (61.04%), ethylbenzene (54.16%) and toluene (51.31%). The above six aromatic hydrocarbons are the main components of organic solvents and are commonly used as solvents for coatings, paints and cleaning agents [38,39]. On the basis of the regional industrial structure, the local high-end equipment manufacturing and biopharmaceutical industries utilizing paints and organic solvents in the production processes, Factor #1 was considered to be a solvent emission source with a total contribution rate of 11.9%. For Factor #2, ethylene (86.22%), propylene (43.30%) and ethane (35.89%) were characteristic pollutants. Ethylene is typically derived from petrochemical emissions, as well as small amounts of propylene and ethane [17]. The petrochemical industry is the traditional pillar industry in Dezhou, and the Dezhou Canal Hengsheng Chemical Industry Park is located about 12 km northwest of the DZS. Therefore, Factor #2 may primarily correspond to the petrochemical industry source. The total contribution rate of Factor #2 was 26.1%. In Factor #3, styrene (86.38%) and n-octane (37.54%) ranked highest in terms of contribution. The fiberglass-reinforced plastic (FRP) manufacturing enterprises are affiliated enterprises in the high-end equipment manufacturing industrial cluster of Dezhou. About 2.4 km northeast and 8.4 km northwest of the monitoring station, there are two large FRP manufacturing enterprises, with styrene serving as a key component in the production system. Thus, Factor #3 could be identified as the fiberglass industry source, with a total contribution rate of 10.5%.

The main contributing components of Factor #4 were 3-methylpentane (81.01%), methylcyclohexane (73.83%), cyclohexane (66.24%) and n-hexane (62.50%), as well as n-pentane (44.40%), 2-methylpentane (42.66%) and iso-pentane (37.53%). As reported, C₅–C₆ alkanes are characteristic pollutants emitted by vehicles [40]. 3-methylpentane is an important tracer for gasoline vehicle exhaust [41], 2-methylpentane is a marker of traffic emissions, n-hexane may originate from diesel evaporative emissions [42], while n-pentane and isopentane are typical tracers of vehicle exhaust emissions [43]. Therefore, Factor #4 was identified as a source of emissions from vehicles, with a contribution of 16.2%.

In Factor #5, acetylene (70.53%), methylcyclopentane (69.95%) and 1-butene (51.63%) contributed the most. Among these, acetylene serves as an important tracer for combustion sources [44], methylcyclopentane is a common component in petroleum products and exhibits characteristics of incomplete combustion products [45], while 1-butene is a characteristic component of liquefied petroleum gas combustion [46]. Hence, Factor #5 was regarded as a fossil fuel combustion source, with a total contribution of 25.6%. The contribution rate of Factor #6 was 9.7%, and the characteristic pollutant was isoprene, which was regarded as a tracer of plant-based emissions [15,47]. Considering that the sampling points are surrounded by well-landscaped areas and there are multiple parks with high vegetation coverage, Factor #6 was regarded as a natural source.

In conclusion, through the PMF model, a total of six factors of VOCs in the summer ambient air of the study area were identified, namely solvent emission source (11.9%), petrochemical industry source (26.1%), fiberglass industry source (10.5%), vehicle emission source (16.2%), fossil fuel combustion source (25.6%) and plant emission source (9.7%). The sources that contribute significantly to the VOCs in the study area during summer were the petrochemical industry source and the combustion source.

3.6. Health Risk Assessment of Toxic VOCs

Health risk assessment of toxic VOCs in Dezhou in 2021 was conducted for the general population using the methodology described in Section 2.4.6. The RfC and IUR values were obtained from the U.S.EPA IRIS [33], ATSDR [34], and OEHHA [35]. Ultimately, non-carcinogenic risk assessment could be carried out for 18 VOC species based on the available RfCs, including 6 alkanes, 2 olefins and 10 aromatic hydrocarbons. According to the recommendations of the U.S.EPA and the presented research [48,49], non-carcinogenic risks may be present when HQ > 1. The non-carcinogenic risk assessment results of ambient VOCs in Dezhou during the summer in 2021 are shown in

Table 1. Assessment of non-carcinogenic risks of ambient VOCs in Dezhou (summer of 2021).

Component	RfC /(mg·m−3)−1	HQ Under Different ED
		10 a	20 a	30 a	40 a	50 a	60 a	70 a
Propylene	3	6.57 × 10−6	1.31 × 10−5	1.97 × 10−5	2.63 × 10−5	3.29 × 10−5	3.94 × 10−5	4.60 × 10−5
n-Pentane	1	4.54 × 10−5	9.08 × 10−5	1.36 × 10−4	1.82 × 10−4	2.27 × 10−4	2.72 × 10−4	3.18 × 10−4
n-Hexane	0.7	3.51 × 10−5	7.01 × 10−5	1.05 × 10−4	1.40 × 10−4	1.75 × 10−4	2.10 × 10−4	2.46 × 10−4
Benzene	0.03	1.65 × 10−3	3.30 × 10−3	4.95 × 10−3	6.59 × 10−3	8.24 × 10−3	9.89 × 10−3	1.15 × 10−2
Cyclohexane	6	1.30 × 10−6	2.60 × 10−6	3.90 × 10−6	5.19 × 10−6	6.49 × 10−6	7.79 × 10−6	9.09 × 10−6
n-Heptane	0.4	4.23 × 10−5	8.45 × 10−5	1.27 × 10−4	1.69 × 10−4	2.11 × 10−4	2.54 × 10−4	2.96 × 10−4
Methylcyclohexane	3	3.59 × 10−6	7.18 × 10−6	1.08 × 10−5	1.44 × 10−5	1.80 × 10−5	2.15 × 10−5	2.51 × 10−5
iso-Propyl benzene	0.4	1.08 × 10−5	2.15 × 10−5	3.23 × 10−5	4.31 × 10−5	5.38 × 10−5	6.46 × 10−5	7.54 × 10−5
n-Propyl benzene	1	9.09 × 10−6	1.82 × 10−5	2.73 × 10−5	3.63 × 10−5	4.54 × 10−5	5.45 × 10−5	6.36 × 10−5
Toluene	5	1.26 × 10−5	2.52 × 10−5	3.78 × 10−5	5.04 × 10−5	6.30 × 10−5	7.56 × 10−5	8.82 × 10−5
Ethylben zene	1	2.68 × 10−5	5.36 × 10−5	8.03 × 10−5	1.07 × 10−4	1.34 × 10−4	1.61 × 10−4	1.87 × 10−4
m,p-Xylene	0.1	5.90 × 10−4	1.18 × 10−3	1.77 × 10−3	2.36 × 10−3	2.95 × 10−3	3.54 × 10−3	4.13 × 10−3
Styrene	1	2.60 × 10−5	5.20 × 10−5	7.80 × 10−5	1.04 × 10−4	1.30 × 10−4	1.56 × 10−4	1.82 × 10−4
o-Xylene	0.1	2.52 × 10−4	5.05 × 10−4	7.57 × 10−4	1.01 × 10−3	1.26 × 10−3	1.51 × 10−3	1.77 × 10−3
1,2,3-Tri-m-benzene	0.06	1.03 × 10−4	2.07 × 10−4	3.10 × 10−4	4.13 × 10−4	5.17 × 10−4	6.20 × 10−4	7.23 × 10−4
1,2,4-Tri-m-benzene	0.06	1.45 × 10−4	2.90 × 10−4	4.36 × 10−4	5.81 × 10−4	7.26 × 10−4	8.71 × 10−4	1.02 × 10−3
1,3,5-Tri-m-benzene	0.06	1.20 × 10−4	2.39 × 10−4	3.59 × 10−4	4.79 × 10−4	5.98 × 10−4	7.18 × 10−4	8.38 × 10−4
n-Nonane	0.02	3.75 × 10−4	7.49 × 10−4	1.12 × 10−3	1.50 × 10−3	1.87 × 10−3	2.25 × 10−3	2.62 × 10−3
Categories		HI
Alkanes		5.02 × 10−4	1.00 × 10−3	1.51 × 10−3	2.01 × 10−3	2.51 × 10−3	3.01 × 10−3	3.52 × 10−3
Olefins		3.26 × 10−5	6.51 × 10−5	9.77 × 10−5	1.30 × 10−4	1.63 × 10−4	1.95 × 10−4	2.28 × 10−4
Aromatic hydrocarbons		2.92 × 10−3	5.84 × 10−3	8.75 × 10−3	1.17 × 10−2	1.46 × 10−2	1.75 × 10−2	2.04 × 10−2
Total		3.45 × 10−3	6.91 × 10−3	1.04 × 10−2	1.38 × 10−2	1.73 × 10−2	2.07 × 10−2	2.42 × 10−2

, taking into account the influence of ED. As calculated, the HQ values increased with the increase in ED. However, even when ED reached 70 years, the HQ values of each species, the HI values of each type of VOC, and the total HI value were all less than 1. The HI value of alkanes was 3.52 × 10⁻³, that of olefins was 2.28 × 10⁻⁴, and that of aromatic hydrocarbons was 2.04 × 10⁻², with a total HI of 2.42 × 10⁻². Assessments for other seasons also showed similar results, which are presented in the Supplementary Materials (Tables S2–S4). It should be noted that, since ethylene, ethane, propane and other species were not included in the assessment due to the lack of RfC parameters, the non-carcinogenic risks obtained in this study were significantly underestimated. Overall, this non-carcinogenic risk assessment was the first step in understanding the health risks of VOCs in Dezhou’s atmospheric environment to public health.

The U.S.EPA IRIS classifies potential carcinogens according to their carcinogenicity. Benzene is defined as a “Known Human Carcinogen”, meaning it is a known human carcinogen for all routes of exposure based upon convincing human evidence as well as supporting evidence from animal studies. Ethylbenzene is described as “not classifiable as to human carcinogenicity” due to lack of animal bioassays and human studies, but it has been included in the Toxic Substances Control Act (TSCA) high-priority risk assessment substances (to be initiated in 2024) [50]. Styrene is also not included in its official “carcinogen list”, but it is classified as a “possible human carcinogen” in the TSCA risk assessment (based on sufficient animal experimental evidence and limited human evidence). Furthermore, the International Agency for Research on Cancer (IARC) classifies benzene as a Group 1 carcinogen, styrene as Group 2A and ethylbenzene as Group 2B carcinogens [51]. Therefore, this study focused on the carcinogenic risk assessment of benzene, ethylbenzene and styrene.

The carcinogenic risk assessment of ambient VOCs in Dezhou in 2021 was conducted, and the influence of ED on the evaluation results was also considered.

Table 2. Assessment of carcinogenic risks of ambient VOCs in Dezhou (summer of 2021).

Component	Benzene	Ethylbenzene	Styrene
Concentration/μg·m−3	2.49	1.35	1.31
IUR/(μg·m−3)−1	7.80 × 10−6	1.10 × 10−6	5.00 × 10−7
ED/a	Risk
10	3.86 × 10−7	2.95 × 10−8	1.30 × 10−8
20	7.71 × 10−7	5.89 × 10−8	2.60 × 10−8
30	1.16 × 10−6	8.84 × 10−8	3.90 × 10−8
40	1.54 × 10−6	1.18 × 10−7	5.20 × 10−8
50	1.93 × 10−6	1.47 × 10−7	6.50 × 10−8
60	2.31 × 10−6	1.77 × 10−7	7.80 × 10−8
70	2.70 × 10−6	2.06 × 10−7	9.10 × 10−8

shows the results for the summer. The risk values of benzene, ethylbenzene and styrene increased with the increase in ED. According to the classification of carcinogenic risk levels [52], negligible risk (risk < 1 × 10⁻⁶), low-probability risk (1 × 10⁻⁶ < risk < 1 × 10⁻⁵), high-probability risk (1 × 10⁻⁵ < risk < 1 × 10⁻⁴), and definite risk (risk > 1 × 10⁻⁴), the carcinogenic risk of benzene in the atmosphere during summer in Dezhou in 2021 was negligible when the ED was 10 years and 20 years, and the risk level was low-probability when the ED reached 30, 40, 50, 60 and 70 years. For ethylbenzene and styrene, carcinogenic risks under different EDs were all negligible. Assessments for other seasons also showed similar results, which were presented in the Supplementary Materials (Tables S5–S7). This indicated that benzene in the ambient air was harmful to people exposed over the long term, and that there was a potential carcinogenic risk. Benzene is primarily derived from the use of solvents and the combustion process. It is therefore necessary to strengthen targeted measures to control the emission of benzene, thereby safeguarding the health of local residents.

4. Conclusions

Based on the annual monitoring data of ambient VOCs in Dezhou in 2021, the characterization of VOCs, seasonal impacts on O₃ and SOA formation, the main sources of VOCs during summer, as well as health risk assessment were investigated. The main conclusions are as follows.

(1) The ambient VOC concentrations exhibited significant seasonal variations, with lower concentrations in spring and summer and higher concentrations in winter and autumn.

(2) The species that contributed the most to OFP were olefins and aromatic hydrocarbons. The top three components in each season were as follows: spring—ethylene, m,p-xylene, and toluene; summer—isoprene, ethylene, and m,p-xylene; autumn—ethylene, m,p-xylene, and isoprene; winter—ethylene, m,p-xylene, and propylene. And for SOAs, the contribution of aromatic hydrocarbons was overwhelming. The highest contributors in spring and winter were toluene, m,p-xylene and o-xylene; in summer and autumn, m,p-xylene, toluene and o-xylene contributed the most.

(3) O₃ pollution occurred frequently in summer in the study area, accounting for 40.22% of the total number of days in summer. Ethylene was the key active precursor of O₃ on pollution days. The emissions from petrochemical enterprises should be given priority for control.

(4) The results of PMF analysis indicated that there were six main sources of VOCs in summer: solvent emission source (11.9%), petrochemical industry source (26.1%), fiberglass industry source (10.5%), vehicle emission source (16.2%), fossil fuel combustion source (25.6%) and plant emission source (9.7%). Consequently, in summer, the focus should be on controlling emissions from the petrochemical industry and combustion sources.

(5) The results of the health risk assessment for 18 toxic VOCs indicated that the non-carcinogenic risk value of each species was within the safety threshold. However, benzene, which was primarily derived from solvent use and combustion processes, was associated with a low probability of carcinogenic risk. It was therefore necessary to step up targeted measures to control benzene pollution in order to safeguard the health of local residents.