Contrasting chemical characteristics and secondary formation of PM_2.5 from mountain and urban sites in central China

Summary

The vertical distribution of aerosol is critical to understand the effect of aerosol on climate. This work explored chemical characteristics of PM_2.5 at three urban and one mountain sites in Shiyan located in central China. Annual PM_2.5 mass concentrations at the mountain site were 70% of urban levels. The total of sulfate, nitrate, and ammonium (SNA) dominated PM_2.5, showing higher proportion (49.3%) at mountain site. SNA shows the homogeneous vertical distribution in spring, summer, and autumn with coefficient of divergences below 0.2, whereas most of components were more abundant at urban sites in winter. Higher nitrogen oxidation ratio (0.27) and ratio of secondary organic carbon to organic carbon (0.65) indicated that the aerosol in mountainous atmosphere exhibited a greater degree of aging compared to that at urban sites. These findings enhance our understanding of the vertical distribution of PM_2.5 composition and provide a scientific foundation for regional atmospheric pollution control.

Graphical abstract

Highlights

• •Vertical distribution of PM_2.5 were studied at both urban and mountain sites
• •Primary aerosol showed obvious vertical variation while secondary components were homogeneous
• •Secondary formation at mountain site was stronger than those at urban site
• •PM_2.5 at the mountain site were more aged than that at urban sites

Earth sciences; Chemical engineering; Environmental engineering

Introduction

PM_2.5 has the potential to have an effect on air quality, atmospheric visibility, human health, and global climate change.^1,2,3 Atmospheric PM comprises a complex mixture that includes sulfate, nitrate, ammonium, elemental and organic carbon, mineral dust, and trace elements. PM_2.5 comprises a range of chemicals, including main components such as metals, black carbon, and primary organic aerosols that are directly emitted from sources. Additionally, it includes secondary components like sulfate, nitrate, ammonium, and secondary organic aerosols, which are generated through intricate atmospheric chemical reactions. Despite there has been a significant enhancement in air quality across China in recent years, PM_2.5 levels continue to exceed the World Health Organization (WHO) guideline values and persist as the main pollutant in numerous Chinese cities. Consequently, examining the chemical properties of PM_2.5 and the formation mechanisms of secondary aerosols is crucial for enhancing the effectiveness of air pollution mitigation efforts in China.

Numerous field measurements have been conducted to explore the chemical characteristics of PM_2.5 in the atmosphere, carried out at urban locations,^4,5,6,7,8,9 as well as at individual high-altitude sites.^10,11 The urban sampling sites do not comprehensively capture the physicochemical characteristics of PM_2.5 at high altitudes or within the free troposphere, where PM_2.5 is significantly affected by long-range transport and has a more substantial impact on cloud formation. The significant influence of local sources in samples gathered from urban locations often results in uncertainties when quantifying the percentage of regional atmospheric transport affecting aerosol dynamics. However, the vertical characteristics of aerosols can largely offset the limitations of single-site observation and enhance our comprehension of the processes involved in the generation, development, and dissipation of atmospheric pollution.¹² The vertical variation of PM_2.5 reflects the atmosphere’s lifetime and output efficiency, offering insights into the global equilibrium characteristics of PM_2.5. The characteristics of PM_2.5, such as concentration and composition, exhibit horizontal variability and also change with altitude, interacting across different layers.¹³ Therefore, comprehending the vertical distribution of pollutants and the process of aerosol formation is crucial for interpreting the mechanisms behind aerosol and haze formation at ground level.

Remote sensing, aircraft-based and tower-based measurements are frequently employed to examine the vertical structure characteristics of aerosols^14,15; however, these methods possess inherent limitations. While remote sensing methods have been developed and used to assess the vertical distribution of aerosols, they fall short in delivering detailed and accurate information regarding these aerosols. Additionally, aircraft measurements facilitate the examination of plume transport. However, it encounters obstacles, such as the transient and irregular characteristics, which limit the comprehension of long-term variations in air quality. The challenges of capturing temporal variations in air pollutant levels during a brief measurement campaign were also demonstrated. While tower-based measurements have successfully mitigated the constraints of aircraft-based measurements in examining temporal variations in air pollution levels,¹⁶ they remain restricted in height, which prevents a comprehensive capture of aerosol characteristics within the planetary boundary layer.

Simultaneous measurements at both mountain and urban sites exhibit distinct advantages compared with above-mentioned observation method. Firstly, mountain observation platform directly obtain observational data at middle and upper levels of the boundary layer or even the lower free troposphere, which can reveal the vertical exchange process of air pollutants between the near-surface and high altitude. Furthermore, simultaneous observation at both mountain and urban sites can provide long-term continuous monitoring results, which enables a comparison of the chemical characteristics of PM_2.5 and the formation mechanisms of secondary aerosols across various sampling sites at different altitudes, thereby providing insights into the pollution characteristics on a regional scale. Liu et al. observed notable discrepancies in pollutants at both mountain and urban sites in China, underscoring the benefits of this method in capturing more detailed aerosol characteristics.¹⁷

Central China is facing a significant decline in atmospheric conditions as a result of rapid economic development.¹⁸ The repeated occurrence of haze events influenced by PM_2.5 and O₃ in recent decades has raised significant concern, particularly during the winter.¹⁹ Except for urban²⁰ and rural regions, it is essential to investigate particulate matter in mountainous areas to enhance the understanding of background pollution levels and air quality in central China. Mt. WDS is one of the prominent high mountains in central China, providing an ideal site for mountainous measurement. It is important to recognize that mountainous areas characterized by isolated ridges or parallel ridges and peaks exhibit distinct geometric structures when compared to lowland regions. This may result in significant variations in the polluted characteristics and formation mechanisms of PM_2.5 between mountainous and urban locations. Several studies have been carried out to examine the characteristics of VOCs,²¹ ultrafine particles²² and submicron aerosol²¹; however, information is scarce regarding the vertical distribution of PM_2.5 chemical composition from both urban and mountain sites within the same region.

To compare the difference of composition distribution of PM_2.5 at different altitude, simultaneous measurements of PM_2.5 were conducted both at mountain site (WDS) as well as its surrounding urban locations (Shiyan of Hubei province in the central China). This study aimed to examine the similarities and differences in the composition, concentration, and seasonal variation of major PM_2.5 components between urban and mountain locations. Additionally, it sought to identify the chemical transformation mechanisms of secondary aerosols in these distinct atmospheric environments. The anticipated outcomes aim to enhance our understanding of the chemical properties of PM_2.5 at elevated altitudes, provide essential data for improving model forecasts, and aid in the formulation of more effective air pollution mitigation strategies in the region.

Overview of the study areas

Shiyan is located in the northwest region of Hubei province in central China, positioned between the Qinling and Daba mountain ranges. The area features a standard basin topography, surrounded by mountains to the north, west, and south. Shiyan has a population of 3.14 million residents and a total of 0.89 million motor vehicles. The industrial composition of the area is primarily characterized by the petrochemical, steel, and electric power sectors.

According to the Technical Regulation for Selection of Ambient Air Quality Monitoring Stations of China,²³ three urban locations and one mountain site were selected for sampling in Shiyan (Figure 1). Three urban locations, Shizhengfu (SZF), Tie Er Chu (TEC), and Yun county (YX), are located within the urban built-up area of Shiyan. These sites are surrounded by typical residential areas and commercial districts, representing typical urban environments. The mountain site (WDS) is located at Wudangshan national air quality monitoring station on the top of Mt. Wudang at an elevation of 862 m above sea level (ASL). The WDS site lies above the planetary boundary layer during most wintertime conditions, enabling the site to largely avoid direct impacts from near-surface local emissions. Furthermore, no prominent local anthropogenic sources exist in the vicinity of this site and its air masses are mainly governed by regional-scale transport and background atmospheric processes. These characteristics render WDS well-suited to represent regional background air quality in the study area. Both the mountain (WDS) and urban sites (SZF, TEC, and YX) belong to the same regional climate and atmospheric circulation background, which supports the rationality and comparability of inter-site comparison. The detailed description of these four observation sites can be found in the STAR METHODS.

Figure 1.

The locations and terrain for each sampling site

Map showing the geographic distribution and topographic setting of the four sampling sites. Site labels, elevation color scale, scale bars, and north arrow are indicated within the figure.

Results and discussion

Overview of PM_2.5 and chemical composition

Figure 2 illustrates the changes in PM_2.5 mass over time at the four locations and the annual PM_2.5 concentration across all locations. The daily PM_2.5 levels varied between 13.06 and 251.01 μg/m³. The highest PM_2.5 mass concentration was recorded on January 18, 2017, while the lowest concentration was noted on November 23, 2017. The notable daily fluctuations in PM_2.5 mass concentration result from changes in emissions from various sources, atmospheric processes, and meteorological conditions.²⁴ The annual average PM_2.5 mass concentrations at WDS were recorded at 51.67 ± 29.90 μg/m³, which is 70% of the mass concentrations observed at urban sites (69.35 ± 44.69 μg/m³ at SZF, 71.28 ± 44.19 μg/m³ at TEC, and 69.52 ± 46.26 μg/m³ at YX). These data indicate a reduction in fine particle concentration as altitude increases. The likely reason is the reduced number of primary emission sources at higher altitudes, consistent with earlier research conducted at elevated locations. Furthermore, the wind speed at the mountain site is higher than that at the urban site, which benefits the dispersion of pollutants.²⁵

Figure 2.

Temporal variations of PM_2.5 mass concentration at four locations

Time-series plots showing daily PM_2.5 mass concentrations at four monitoring sites over the study period, with sites distinguished by color as indicated in the legend. The embedded boxplot inset summarizes the overall PM_2.5 concentration distribution for each site, with all boxplot elements defined within the figure.

The PM_2.5 chemical species are categorized into nine primary types: sulfate (SO₄²⁻), nitrate (NO₃⁻), ammonium (NH₄⁺), organic matter (OM), elemental carbon (EC), crustal materials (CMs), trace elements (TEs), kalium (K⁺), and chloronium (Cl⁻).²⁶ OM equals 1.4 multiplied by OC. CM was calculated as follows: CM = 2.20[Al] +2.49[Si] + 1.63[Ca] + 2.42[Fe] + 1.94[Ti]. TE comprises metals and various trace elements, excluding crustal elements. SO₄²⁻, NO₃⁻, NH₄⁺, K⁺, Cl⁻, and EC were analyzed directly. The unidentified matter (UM) value was calculated by subtracting the total masses of OM, EC, SO₄²⁻, NO₃⁻, NH₄⁺, CM, and TE from the PM_2.5 mass concentration determined through gravimetric methods.

The annual mean concentration and chemical composition of PM_2.5 across four locations is depicted in Figure 3. The total of SO₄²⁻, NO₃⁻, and NH₄⁺ (SNA) was the major chemical component, accounting for 45% to 49% of PM_2.5 mass according to sites, and showing a slightly higher contribution at mountain site (49%). The highest SNA contributions are found at WDS, indicating that the mountain site presents more favorable conditions for the secondary formation and aging of aerosols. OM ranked second overall, accounting for 17%–20% of the PM_2.5 mass concentration, with a slightly higher contribution observed at urban sites. CM constituted approximately 14%–18% of the PM_2.5 mass concentration, exhibiting no significant spatial variation. EC has a PM_2.5 mass contribution of 4%–6%, and is about 1.7 times higher at urban site than mountain site. This discrepancy of EC concentration may be attributed to increased vehicular emissions in the surface atmospheric environment, with a reduced impact of anthropogenic activities at the mountain site. Initial research indicates a comparatively reduced proportion of EC in particulate matter at high-altitude sites when contrasted with urban settings.^27,28 Finally, K⁺, Cl⁻, and TE are minor components of PM_2.5 in this region, ranging from 0.9% to 1.0%, from 0.6% to 1.2% and from 0.3% to 0.4% of its mass, respectively. The UM ranged from 11% to 14%, likely due to uncertainties in the multiplication factors utilized for estimating OM and FS, as well as other unidentified species and measurement uncertainties.

Figure 3.

Chemical composition of PM_2.5 at four sites

Stacked bar charts depicting (A) mass concentrations and (B) fractional contributions of major chemical components to PM_2.5 at the four monitoring sites. Each bar represents one site, with different color segments indicating individual components as defined in the figure legend.

Seasonal variation

Figure 4 presents the seasonal fluctuations in PM_2.5 mass concentration across four locations. The PM_2.5 concentrations recorded at all sites were significantly higher during the winter season compared to other times of the year. In urban locations, the PM_2.5 concentrations exhibited a variation pattern as follows: winter (122–124 μg/m³) > autumn (47–55 μg/m³) > spring (50–53 μg/m³) > summer (39–49 μg/m³). The higher levels observed during peak winter may result from increased emissions associated with fuel combustion for heating, lower temperatures (refer to Table S1), and reduced heights of the planetary boundary layer. The lowest concentration was observed in summer, attributed to concentrated precipitation during this season and optimal diffusion conditions (Table S1). At WDS, PM_2.5 concentrations peaked in winter at 79.50 μg/m³, while values in the other three seasons remained relatively similar, ranging from 40.10 to 40.91 μg/m³. It is interesting that PM_2.5 levels in summer at WDS were comparable to those in spring and autumn, which could be attributed to higher CM concentration in this season (Figure 5). This phenomenon is attributed to the elevated planetary boundary layer height and intense convective mixing in summer, which facilitates the transport of dust from low atmospheric layers to high mountain altitudes.

Figure 4.

Seasonal variation of PM_2.5 mass concentrations at four sites

Boxplots showing PM_2.5 concentrations at four sites (SZF, TEC, YX, and WDS) across four seasons. All boxplot elements (boxes, medians, whiskers, and outliers) are annotated in the figure.

Figure 5.

Seasonal variation of PM_2.5 components and composition

Stacked bar graphs presenting (A) mass concentrations and (B) mass fractions of major chemical components of PM_2.5 in four seasons at four monitoring sites.

Figure 5 demonstrates the seasonal variation of chemical composition. The concentration and contribution of NO₃⁻ across all sites exhibited a consistent seasonal variation, with the order being winter > autumn > spring > summer. The concentrations of NO₃⁻ during winter were measured at 23.54, 23.45, 23.77, and 16.22 μg/m³ at the locations SZF, TEC, YX, and WDS, respectively. These values were found to be 15 to 49 times greater than the concentrations recorded during the summer at all four sites. The peak concentration of NO₃⁻ during winter is likely associated with stable meteorological conditions, increased emissions of gaseous pollutants (like NO₂) (Table S2), and elevated nitrogen oxidation ratio (NOR), which facilitates the secondary formation of aerosols (Figure 5).²⁹ Furthermore, the elevated PM_2.5 mass concentration observed during winter provides an increased number of aerosol surfaces conducive to the heterogeneous reaction of NO₃⁻. In summer, elevated temperatures may facilitate the dissociation of ammonium nitrate particles and heighten the volatility of NH₄NO₃, resulting in the lowest concentration of NO₃⁻ during this season.^30,31 SO₄²⁻ and NH₄⁺ display distinct seasonal variations compared to NO₃⁻, following the order of winter > spring > summer > autumn. The concentrations of SO₄²⁻ and NH₄⁺ during winter were observed to be 1.8 to 3.07 times and 2.20 to 3.64 times higher, respectively, compared to those in autumn. NH₄⁺ exhibited a comparable seasonal variation to SO₄²⁻, suggesting that NH₄⁺ predominantly existed in the forms of (NH₄)₂SO₄ or NH₄HSO₄. During winter, SO₂ exhibits the highest concentration, which partially accounts for the elevated levels of SO₄²⁻ observed in this season. In spring and summer, SO₂ concentrations were lower compared to autumn; however, the SO₄²⁻ concentration during the former two seasons was significantly higher than that observed in autumn. This effect is due to the intense photochemical reactions that take place in the warmer months of spring and summer. The reactions occur due to heightened levels of oxidants (including O₃ and others) and elevated temperatures (refer to Tables S1 and S2), which then motivate the secondary formation of SO₄²⁻. As a result, there is a significant rise in the concentration of SO₄²⁻ when compared to autumn. Comparable outcomes have been noted in numerous cities across China.³² The seasonal variation of NO₃⁻ concentrations at urban locations exceeded that observed at the mountain site.

OC and EC reflected seasonal fluctuation patterns of PM_2.5 mass concentration, exhibiting maximum values in winter, followed by autumn and spring, with the lowest levels observed in summer. In winter, OC concentrations were observed to be 5.0, 4.8, 4.4, and 4.0 times greater than the levels recorded in summer at SZF, TEC, YX, and WDS, respectively. In a similar manner, the EC concentration during winter was observed to be 2.6, 2.6, 2.8, and 1.9 times greater than the levels recorded in summer. Two potential explanations clarify the observed seasonal variation. Firstly, reports indicate that emissions of carbon-containing aerosols from fuel combustion for heating purposes have increased during the winter.³³ Secondly, the meteorological conditions defined by weak wind, reduced rainfall, and shallow atmospheric boundary layer heights during autumn and winter hinder the dispersion, transport, and wet deposition of particulate matter. As a result, these factors contribute to elevated carbon concentrations in the colder season. The seasonal variation of organic carbon was more pronounced compared to elemental carbon. The seasonal variations in OC and EC levels at urban sites were significantly more pronounced compared to those observed at the mountain site. This observation suggests that the OC and EC at the mountain site are probably sourced from nearby ground-level origins.

Concentrations of CM at the mountain site exhibited distinct seasonal variations compared to those observed at urban locations. At urban sites (SZF, TEC, and YX), the concentration of CM exhibited an increasing trend in the following order: winter > spring > summer > autumn. The concentrations of CM during winter, spring, and summer were approximately 1.2–1.3, 1.1–1.5, and 1.2–1.3 times greater than those observed in autumn at the specified urban locations, respectively. During winter, there are elevated levels of crustal materials attributed to lower precipitation compared to other seasons. The concentration of CM increases in spring due to higher wind speeds, which enhance the resuspension of local dust and enable the long-distance transport of dust from the northwestern regions of China. At WDS, the concentration of CM is highest in summer at 12.04 μg/m³, followed by winter at 9.89 μg/m³ and spring at 9.49 μg/m³, with the lowest concentration occurring in autumn at 4.50 μg/m³. This pattern may be attributed to strong vertical air mixing in summer, which transports dust from ground level to WDS. The seasonal order of trace elements was observed as winter > autumn > spring > summer. The concentrations of Mn, Co, Cu, As, Mo, Cd, Sn, and Sb during winter are 2–4 times higher than those observed in summer.

The chemical composition of PM_2.5 displayed varying seasonal characteristics (Figure 5). The chemical composition of PM_2.5 across all four sites exhibited a consistent hierarchy: OM > NO₃⁻ > SO₄²⁻ > CM > NH₄⁺ > EC > TE during the winter and autumn seasons. OM and NO₃⁻ were the primary components, accounting for approximately 16%–26% and 13%–19% of PM_2.5 mass, respectively. In spring and summer, the order differed from that in winter and autumn: SO₄²⁻ > CM > OM > NH₄⁺ > NO₃⁻ > EC > TE, with SO₄²⁻ being the dominant component, followed by CM and OM. The fraction of NO₃⁻ was observed to be greater than that of SO₄²⁻ and NH₄⁺ during winter and autumn. However, in summer, the fractions of SO₄²⁻ and NH₄⁺ exceeded those of NO₃^–, even though NOx concentrations remained higher than SO₂ across all seasons. This phenomenon is likely due to the increased levels of OH and O₃ during the summer, combined with higher temperatures and greater solar radiation. The specified conditions facilitate the degradation of NO₃⁻ and the ensuing generation of SO₄²⁻ from SO₂. Recent research has documented such results across various regions in China.³²

Similarities and differences between urban and mountain sites

To analyze the variations in PM_2.5 chemical composition profiles between mountainous and urban locations, the coefficient of divergence (COD) was calculated to determine spatial differences in PM_2.5 concentrations. The COD is mathematically defined as Equation 1³⁴:

$$\text{CO}{\mathrm{D}}_{\mathrm{j},\mathrm{k}}=\sqrt{\frac{1}{\mathrm{p}}\sum _{\mathrm{i}=1}^{\mathrm{p}}{\left[\frac{\left({\mathrm{x}}_{\text{ij}}-{\mathrm{x}}_{\text{ik}}\right)}{\left({\mathrm{x}}_{\text{ij}}+{\mathrm{x}}_{\text{ik}}\right)}\right]}^2}$$ (Equation 1)

Where x_ij and x_ik denote the seasonal or annual average value for each chemical component i at sites j and k, respectively, with p representing the total amount of major chemical species. A COD value of less than 0.20 indicates a high level of uniformity in concentrations across various sites, whereas a value above 0.2 signifies the existence of heterogeneity.^35,36 It is important to note that analytical uncertainties may increase COD values, particularly for species that are present in lower concentrations.

Figure 6 illustrates a comparison of seasonal-average concentrations of major chemical components between urban and mountain sites, which is further clarified in the following discussion. PM_2.5 exhibited significant variations between urban and mountain locations during spring, autumn, and winter, with COD values exceeding 0.2. In contrast, a moderate similarity was noted in summer, characterized by relatively low COD values ranging from 0.16 to 0.21. However, when examining individual components, various chemical species in PM_2.5 exhibited significant differences in their season-average concentrations between mountain and urban locations (Figure 7), such as EC, Cl⁻ a TE across all four seasons, and CM specifically in autumn. During winter, the disparity in chemical components between mountain and urban locations was greater than in other seasons, resulting in COD values that exceeded those recorded in other times of the year. The discrepancies observed can be attributed to the varying atmospheric processes, emission sources, and meteorological parameters inherent to these two types of atmospheric environments, as well as the effects of vertical and regional transport.

Figure 6.

Mean concentrations of major components across sites in four seasons

Scatterplots showing correlation of dominant chemical constituents between WDS and the three urban sites (SZF, TEC, and YX) in (A) winter, (B) spring, (C) summer, and (D) autumn.

Figure 7.

COD values of each species between mountain and urban sites in four seasons

Bar graphs displaying seasonal COD values for (A) SO₄²⁻, (B) NH₄⁺, (C) NO₃⁻, (D) OM, (E) EC, and (F) CM. In each panel, the x axis corresponds to the four seasons; for each season, the six bars represent COD values for all pairwise combinations among the four sampling sites. Site pairs are distinguished by color as indicated in the key. Higher COD values indicate greater compositional dissimilarity. Abbreviations: SO₄²⁻, sulfate; NH₄⁺, ammonium; NO₃⁻, nitrate; OM, organic matter; EC, elemental carbon; CM, carbonaceous matter.

There was no significant difference in SO₄²⁻ concentrations between mountain and urban sites across all seasons, with COD <0.2, except for the COD in winter. However, SO₂ concentrations at the mountain site were found to be 1.12–2.25 times, 0.40–0.79 times, 0.39–0.76 times, and 0.68–2.25 times higher than those at urban sites in spring, summer, autumn, and winter, respectively. The similar SO₄²⁻ concentration observed at both mountain and urban locations may be attributed not only to the elevated oxidation ratio of SO₂ to SO₄²⁻ but also to the effects of vertical and regional transport of SO₄²⁻ at high-altitude sites. This indicates that SO₄²⁻ functions as a regional secondary aerosol and exhibits a homogeneous spatial distribution. In winter, the concentration of NO₃⁻ and NH₄⁺ at urban sites was significantly higher than that at mountain sites, with COD exceeding 0.2. This phenomenon can be attributed to stable boundary layer conditions that hinder the vertical transport of gaseous precursors, such as NO₂ and NH₃, into higher altitudes. The concentration of NO₂ at the mountain site was approximately 30%–90% of that observed at urban locations, which can partially account for the vertical variation in NO₃⁻ levels. In a manner akin to SO₄²⁻, the concentrations of NO₃⁻ and NH₄⁺ are similar at both mountain and urban locations, with COD values below 0.2 during spring, summer, and autumn. Conversely, the NO₂ concentration at mountain sites ranged from 0.27 to 0.81 times, 0.44 to 1.36 times, and 0.32 to 0.49 times those observed at urban sites during these respective seasons. This may also be attributed to the elevated oxidation ratio of NO₂ to NO₃⁻ and the effects of vertical and regional transport of NO₃⁻ at high-altitude locations during these seasons.

The concentrations of OM and EC at the mountain site were significantly lower than those observed at ground-based sites across all seasons, with a COD exceeding 0.2, except for OM in summer. This variation is primarily attributed to the lack of local combustion sources and the effective wet scavenging that occurs during vertical air mass transport. Furthermore, the average OC/EC ratio at the mountain site was 1.2 times greater than that at urban sites, mainly attributed to enhanced SOA formation in the mountain atmosphere (detailed information is provided in the section on “secondary organic carbon”).

CM exhibits a non-uniform vertical distribution across winter, summer, and autumn, with COD values exceeding 0.2. In contrast, during spring, it displays similarities with COD values below 0.2, which are associated with long-range transport from northwestern regions during this season. Ground-based sites exhibit relatively high levels of TE, ranging from 0.20 to 0.37 μg/m³, compared to WDS, which shows a level of 0.14 μg/m³.

Secondary formation of PM_2.5

The sulfur/nitrogen oxidation ratio

Gaseous pollutants undergo oxidation to transition into the particle phase through gas and aqueous heterogeneous reactions. The formation pathways of SO₄²⁻ primarily consist of gas-phase oxidation of SO₂ by OH radicals, aqueous-phase reaction pathways in cloud water involving H₂O₂ and O₃, and the NO₂ reaction pathway, which involves the oxidation of SO₂ by NO₂ in aerosol water.³⁷ The generation of secondary NO₃⁻ occurs in two stages: first, the production of gaseous HNO₃, followed by its distribution into the particulate phase. During daylight hours, gaseous HNO₃ is mainly generated via the photochemical reaction between NO₂ and O₃.³⁸ During nighttime, the primary source is the hydrolysis process of N₂O₅ on the urban particulate matter.³⁹ The reaction of gaseous HNO₃ with alkaline substances results in the formation of particulate NO₃^–. This process is affected by the presence of alkaline materials, such as NH₃, as well as by relative humidity (RH) and temperature.⁴⁰

To assess the extent of atmospheric conversion from SO₂ to SO₄²⁻ and NOx to NO₃⁻, the sulfur oxidation ratio (SOR) and NOR were computed. The SOR and NOR were determined based on the molar ratios of particulate concentration relative to the total of gaseous and particulate concentrations:

$$\text{SOR}={nSO}_4^{2-}/\left({nSO}_4^{2-}+nS{O}_2\right)$$ (Equation 2)

$$\text{NOR}={nNO}_3^{-}/\left({nNO}_3^{-}+nN{O}_2\right)$$ (Equation 3)

When nSO₄²⁻, nSO_2, nNO₃⁻ and nNO₂ are the molar concentrations of SO₂, SO₄²⁻, NO₂, and NO₃⁻, respectively. SOR and NOR values exceeding 0.10 it signifies that the atmospheric oxidation of SO₂ and NO₂ is occurring in the atmosphere, resulting in the formation of secondary SO₄²⁻ and NO₃⁻.⁴¹

Figure 8 indicates the seasonal fluctuations of SOR and NOR. The SOR and NOR at all sites exceeded 0.1, indicating that, alongside the impact of primary pollutant emissions, the development of sulfate and nitrate is closely associated with the secondary transformation of gaseous precursors. It is important to note that the SOR across all sites was significantly greater than the NOR. This observation simplifies why, even with lower SO₂ concentrations in comparison to NO₂, sulfate concentrations exceed nitrate levels at urban locations (SZF, TEC, and YX).

Figure 8.

Seasonal variation of SOR and NOR at four sites

Boxplots showing the seasonal variation of (A) SOR and (B) NOR at four sites. In each panel, data from the four sites are distinguished by color as indicated in the key. All boxplot elements (boxes, medians, whiskers, and outliers) are defined within the figure. Each box represents data from one site-season combination.

SOR and NOR exhibited distinct seasonal variations across all locations. The sequence of SOR indicates that summer (0.44–0.55) ranks higher than winter (0.28–0.51), followed by spring (0.35–0.47) and autumn (0.24–0.33). The increased SOR observed during summer across all sites can be linked to the peak temperatures and O₃ concentrations, which promote the photochemical production of OH radicals. This process may enhance the homogeneous gas-phase photochemical reactions of SO₂, subsequently leading to gas-to-particle conversion. Additionally, the elevated RH during the summer promotes the heterogeneous formation of sulfate. The NOR exhibited distinct seasonal variations, with winter (0.28–0.51) showing higher values than autumn (0.11–0.23), followed by spring (0.08–0.23), and summer (0.04–0.17), presenting the lowest values. The increased NOR during winter may be attributed to the varied formation of particulate NH₄NO₃ from gaseous HNO₃ and NH₃, which is facilitated at lower temperatures typical of the winter season. In contrast, elevated summer temperatures are expected to facilitate the dissociation of NH₄NO₃ into the gaseous state, leading to a reduction in NOR during this season.⁴²

As for the vertical variation, the annual SOR values at mountain site (0.39) was close to the value at urban sites (0.41), while annual NOR at mountain site (0.44) was higher than that at urban sites (0.37). In winter and spring, SOR values are higher at the urban site (0.34–0.49) compared to the mountain site (0.35), suggesting enhanced sulfate formation through secondary processes in the urban environment during these seasons. Conversely, in the summer and autumn months, the mountain site (0.33 and 0.55) shows elevated SOR values, indicating that sulfate formation is more pronounced at high elevations during the warmer seasons. This can be explained by similarity and differences in concentrations of gaseous precursors, chemical processes and regional transportation. In winter and spring, the SO₂ concentration at mountain site is higher than those at urban sites (Table S1), which decreased the SOR value at this site. SO₂ and SO₄²⁻ at mountain site are mainly from regional transportation. In summer and autumn, intensified photochemical activity, elevated O₃ concentrations, and a higher incidence of cloud and fog processing at mountain site would facilitates SO₂ oxidation causing higher SOR at this site in these two seasons. The NOR values for nitrate formation are consistently elevated at the mountain site throughout all seasons, suggesting a more significant contribution of secondary nitrate formation at higher elevations. The transport of NO_x and nitrate precursors from the urban area to the mountain site may be responsible for this phenomenon. At the mountain site, these compounds undergo more extensive oxidation due to cleaner atmospheric conditions and lower temperatures, which favor the formation of nitrate aerosol.

Estimation of secondary organic carbon level

OC can be obtained from both primary emissions in conjunction with EC and through secondary formation processes, including photochemical reactions, aqueous-phase production, and heterogeneous reactions involving organic precursors. The OC/EC ratio serves as an indicator for determining the relative contributions of primary and secondary sources of carbonaceous aerosols. Atmospheric aerosols primarily composed of secondary organic aerosol (SOA) exhibit a higher organic carbon to elemental carbon (OC/EC) ratio, specifically when OC/EC exceeds 2.⁴³ Our study revealed that OC/EC at WDS exhibited a comparable seasonal variation to urban sites, with the following order: winter > spring > autumn > summer (Figure 9A). The highest OC/EC ratio was observed in winter, indicating a significant contribution from secondary sources during this season. Conversely, the lowest OC/EC ratio occurred in summer, as the efficiency of precipitation in removing OC, which is highest in summer (refer to Table S2), was considerably greater than that for EC. The annual average OC/EC ratio (2.71) at mountain sites was marginally greater than that at urban sites (2.46), suggesting a more pronounced contribution of SOC at the mountain location. Previous studies have shown that a lower ratio of OC/EC is linked to fossil fuel combustion, such as coal burning (1.4)⁴⁴ and vehicle exhaust (1.08),⁴⁵ whereas a higher value is associated with biomass burning (13.7).⁴⁴ The OC/EC ratio observed at both mountain and urban locations fell within the range indicative of fossil fuel combustion, confirming that the carbonaceous species in this region were primarily influenced by fossil fuel emissions.

Figure 9.

Seasonal variation of carbonaceous materials across all sites

Boxplots showing the seasonal variation of (A) OC/EC, (B) POC, (C) SOC, and (D) SOC/OC at four sites. In each panel, data from the four sites are distinguished by color as indicated in the key. All boxplot elements (boxes, medians, whiskers, and outliers) are defined within the figure. Each box represents data from one site-season combination.

The POC and SOC were estimated by applying the (OC/EC)_min ratio methods as detailed below⁴⁶:

$$POC=EC\times {\left(OC/EC\right)}_{\mathit{min}}$$ (Equation 4)

$$SOC=OC-POC$$ (Equation 5)

where (OC/EC)_min is the minimum value of OC/EC during the whole observation campaign.^43,47 The estimation was derived by excluding data from days affected by rain, dust storms, or other occurrences that notably impact the OC/EC ratio during each season.

Figures 9B–9D presents the seasonal average concentrations of POC and SOC for the four sites. The concentrations of SOC and POC across all sites exhibited comparable seasonal variations, peaking in winter with values ranging from 9.6 to 13.42 μg/m³ for SOC and 2.90 to 7.60 μg/m³ for POC. Conversely, the lowest concentrations were recorded in summer, with SOC values between 1.11 and 1.89 μg/m³ and POC values ranging from 1.24 to 3.21 μg/m³. The elevated concentrations of POC and SOC during winter can be partially explained by the influence of lower temperatures, which enhance the condensation process of volatile organic compounds onto particulate matter. Additionally, high RH contributes to SOC formation under these colder conditions.⁴⁸ The results obtained here are in line with those reported in Beijing⁴⁹ and Zhengzhou of China, India, and some European cities,⁵⁰ but disagree with north-west Europe,⁵¹ where high SOC concentrations were observed in summer.

The SOC/OC across all sites exhibited comparable seasonal variation, ranked as follows: winter (50.8%–75.5%) > autumn (33.6%–66.4%) > spring (38.8%–60.4%) > summer (23.3%–58.6%). The rise observed during winter can be partially linked to the enhanced condensation process and elevated RH, which contribute to the formation of SOC under lower temperature conditions. In contrast, the higher temperatures experienced during summer facilitate the evaporation of semi-volatile organic compounds, while precipitation effectively removes water-soluble organic carbon from the atmosphere, thereby limiting the production of SOC.

The spatial distribution of SOC concentration varied across different seasons. During winter, the concentration of SOC at urban locations (SZF, TEC, and YX, ranging from 11.58 to 13.42 μg/m³) exceeded that at the mountain site (WDS, 9.6 μg/m³). This difference may be attributed to stable meteorological conditions and decreased solar radiation (Figure 9). During the summer season, there is no significant altitude variation in SOC levels between ground-based and mountain locations, with measurements ranging from 1.11 to 1.89 μg/m³. However, SOC/OC exhibits a higher value at mountain sites (0.65) compared to urban sites (0.48), indicating that mountain aerosols are more aged than those at ground surface.⁵² The elevated SOC/OC observed at WDS may be associated with higher biogenic emissions, leading to the formation of secondary organic aerosols through the transformation of these reactive biogenic VOCs. This highlights the intricate relationships among topography, atmospheric chemistry, and biogenic emissions that influence the composition of organic aerosols in mountainous regions.

NO₃⁻ formation mechanism

Ammonium serves as a significant component in the neutralization of particulate SO₄²⁻ and NO₃⁻. In the atmosphere, SO₄²⁻ competes with NO₃⁻ for NH₄⁺ during their formation processes. Therefore, the relationship between NO₃⁻ and NH₄⁺ at varying levels of SO₄²⁻, expressed as the molar ratios of [NO₃⁻]/[SO₄²⁻] and [NH₄⁺]/[SO₄²⁻], can provide insights into the dominant mechanisms for the formation of NO₃⁻.^53,54,55 Park et al.⁵⁴ established that the samples can be distinctly categorized into two groups based on their ammonium to sulfate molar ratio in ambient concentrations: ammonium-rich ([NH₄⁺]/[SO₄²⁻] >1.5) and ammonium-poor ([NH₄⁺]/[SO₄²⁻] ≤1.5). Under ammonium-rich conditions, a strong linear correlation between the ratios of [NO₃⁻]/[SO₄²⁻] and [NH₄⁺]/[SO₄²⁻] indicates that the HNO₃ generated through both gas oxidation and heterogeneous processes reacts according to the equation NH₃ + HNO₃ ↔ NH₄NO₃.^54,56 Under conditions with low ammonium levels, the correlation between HNO₃⁻ and other cations, including calcium carbonate, leads to the formation of NO₃^-.56 The generation of NO₃⁻ is facilitated by the hydrolysis of N₂O₅ within existing aerosols.⁵⁴

Figure 10 presents the relationship between the nitrate-to-sulfate molar ratio ([NO₃⁻]/[SO₄²⁻]) and the ammonium-to-sulfate molar ratio ([NH₄⁺]/[SO₄²⁻]) for samples collected from four sites across four different seasons. During the winter, spring, and autumn seasons, the ratio of [NO₃⁻]/[SO₄²⁻] demonstrates a strong positive correlation with the ratio of [NH₄⁺]/[SO₄²⁻], except for SZF in autumn. The coefficient of determination (R²) ranges from 0.43 to 0.86 across various sites and seasons, as detailed in Figure 10. Furthermore, the majority of [NH₄⁺]/[SO₄²⁻] values exceed 1.5, indicating that the fine nitrate aerosols primarily result from the reaction between HNO₃ and NH₃. In summer, the [NO₃⁻]/[SO₄²⁻] molar ratio exhibited either a negative correlation or no correlation with the [NH₄⁺]/[SO₄²⁻] molar ratio, despite being greater than 1.5. This suggests that NO₃⁻ may be generated through alternative chemical reactions, including interactions with fine dust particles or via the heterogeneous hydrolysis of N₂O₅ on existing sulfate aerosols.

Figure 10.

Seasonal relationship between the NO₃⁻/SO₄²⁻ molar ratio and the NH₄⁺/SO₄²⁻ molar ratio at four sites

Scatterplots showing the correlations for (A) winter, (B) spring, (C) summer, and (D) autumn. Data points from the four sites are distinguished by color as indicated in the key. Solid lines represent the ordinary least-squares regressions for each site. The vertical red dashed line at x = 1.5 indicates the stoichiometric threshold for ammonium-rich (NH₄⁺/SO₄²⁻ = 1.5).

Figure 11 illustrates that the concentration of NO₃⁻ increased in correlation with the rise of “excess-NH₄⁺,” which is defined as the surplus of ammonium beyond the necessary amount to maintain a [NH₄⁺]/[SO₄²⁻] molar ratio of 1.5. Specifically, excess-[NH₄⁺] is calculated as ([NH₄⁺]/[SO₄²⁻] −1.5) × [SO₄²⁻].⁵⁴ This indicates that the formation of NO₃⁻ primarily occurs through the homogeneous reaction of NH₃ with HNO₃ in conditions rich in NH4+. Furthermore, excess-[NH₄⁺] shows distinct seasonal variation, with maximum concentrations observed in winter (mean, 0.44) and reduced levels in summer (mean, 0.12), reflecting the trends seen in NO₃⁻. This indicates that the formation of particulate NO₃⁻ primarily takes place in environments rich in NH₄⁺. In the autumn and winter seasons, the slope of the regression line between excess-NH₄⁺ and NO₃⁻ approaches 1, which indicates a complete neutralization of excess-NH₄⁺ and NO₃⁻.^57,58 Conversely, spring and summer displayed steeper slopes (1.30 and 1.63) in the relationship between excess-NH₄⁺ and NO₃⁻ within NH₄⁺-rich samples. This suggests that excess-NH₄⁺ was insufficient to fully neutralize NO₃⁻, with the remaining NO₃⁻ being generated through different pathways. The concentration of Excess-NH₄⁺ at the mountain site during summer was measured at 0.14 μmol/m³, which is marginally higher than the 0.11 μmol/m³ recorded at urban sites. In the other three seasons, the levels were comparable across all sites.

Figure 11.

Nitrate concentration in relation to “excess ammonium” in four seasons

Scatterplots showing the correlations for (A) winter, (B) spring, (C) summer, and (D) autumn. Data points from the four sites are distinguished by color as indicated in the key. Solid lines represent the ordinary least-squares regressions for each site. The vertical red dashed line at x = 0 indicates the stoichiometric threshold for excess-NH₄⁺.

These results indicate that PM_2.5 at the mountain site exhibited reduced influence from anthropogenic pollutants originating from surface sources, and a higher presence of aged aerosol was observed at this location. The results will enhance the model, leading to improved simulations of aerosol composition at higher altitudes.

Limitations of the study

This research presents two significant limitations that require consideration in future research endeavors. The absence of meteorological parameters at the mountain site limits the ability to compare the formation mechanisms of secondary aerosols between urban and mountain locations. Secondly, our research used daily PM_2.5 filter samples and integrated data over 23 h, which limited our ability to capture rapid changes in chemical composition during brief pollution episodes or transport events. While these findings offer scientific backing for enhancing the simulation accuracy of numerical models at elevated altitudes, future efforts will necessitate measurements with higher temporal resolution.

Resource availability

Lead contact

Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Yisheng Xu (xuys@craes.org.cn).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •Data will be made available on request.
• •This study does not report the original code.
• •Any additional information required to reanalyze the data reported in this study is available from the lead contact upon request.

Acknowledgments

This work was supported by Jing-Jin-Ji Regional Integrated Environmental Improvement-National Science and Technology Major Project (2025ZD1201100) and the National key R&D Program of China (2024YFC3713500). The authors would like to express their gratitude to EditSprings (https://www.editsprings.com) for the expert linguistic services provided.

Author contributions

Conceptualization, data curation, formal analysis, acquisition, software, validation, visualization, writing – original draft, and writing – review and editing, L.R.; writing – original draft and writing – review and editing, H.L.; writing – original draft and writing – review and editing, R.Z.; acquisition, software, validation, and visualization, S.C. and Y.W.; data curation, formal analysis, and acquisition, J.L.; validation, visualization, and writing – original draft, X.Y.; writing – original draft and writing – review and editing, Y.G.; methodology, funding acquisition, and writing – review and editing, Y.X.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Deposited data
Meteorological data such as average wind speed, relative humidity, and temperature	Air automatic monitoring station in Shiyan, China	N/A
Hourly concentration data of NO, NO2,CO, O3 and SO2	Air automatic monitoring station in Shiyan, China	N/A
Software and algorithms
Microsoft EXCEL	Microsoft	https://www.microsoft.com/
Origin 2025	OriginLab Corporation	https://www.originlab.com
ArcGIS	Environmental Systems Research Institute, Inc.	https://www.esri.com

Experimental model and study participant details

This study did not involve experimental models or study participants.

Method details

Sites description

The sampling sites of SZF (E110.79°, N32.63°; roughly 20m above ground) and TEC (E110.88°, N32.57°; about 10m above ground) are situated in the central urban areas of Maojian district, Shiyan. The sampling site of YX (E110.81°, N32.84°; approximately 10 m above ground level) is situated on the rooftop of a building in the Yunyang district of Shiyan, at an elevation of about 10 m. These sites exemplify typical urban atmospheric conditions, shaped by emissions from industrial, traffic, and residential sources. They exemplify the characteristics of the urban environment. The WDS site (E111.04°, N32.39°) is situated at an altitude of 800m above sea level on the mountainside of Wudang Mountain, within the suburbs of Dangjiangkou County in Shiyan, approximately 41 km from downtown Shiyan. The WDS sites serve as the background reference for atmospheric environmental monitoring in Shiyan. The surrounding vegetation is dense and shows no signs of disturbance from traffic or industrial activities, making it a suitable indicator of the regional air quality in this area.

Field sampling

At each site, two sets of parallel low-volume samplers (MVS manufactured by Comde-Derenda GmbH, 16.7 L min⁻¹) were used concurrently for the collection of PM_2.5 samples. A set of PM_2.5 samples was prepared using 47 mm polypropylene filters for elemental analysis, while another set applied 47 mm quartz fiber filters, which were baked at 500°C to remove organic species, for the analysis of water-soluble ionic and carbonaceous materials. Calibration of the sampling devices was performed using a mass flow meter (Model 4040, TSI) both prior to and following the sampling process. The flow rates of the samplers were assessed at both the start and conclusion of the sampling process, with the variation in flow rate remaining below 10% between the initial and final sampling.

Sampling activities took place across the four seasons of 2016: winter (January 13 to February 5), spring (April 21 to May 7), summer (July 23 to August 11), and autumn (October 18 to November 28). A 23-h integrated sample was collected at each site from 10:00 a.m. Beijing Standard Time (BST) to 9:00 a.m. the following day. Furthermore, two 10-min field blank filters were employed for each season by inserting a filter into the sampler and exposing it without air intake. The field blank filters were used to reduce any artifacts resulting from the adsorption of gas-phase organic components onto the filters. During the observation period, 110 pairs of samples and 16 field blanks were collected from the four sites. After the sampling procedure, the exposed filters and field blanks were maintained at low temperatures by being stored in a refrigerator during transit from the monitoring locations to the laboratory. Subsequently, they were placed in a freezer at −18°C before analysis. This protocol was established to mitigate the loss of volatile constituents due to evaporation and to safeguard against the degradation of components, especially nitrate.

Simultaneously, real-time assessments of atmospheric pollutants were performed at four locations. SO₂ levels were recorded using the Model 43i from Thermo, CO was identified with the Model 48i, O₃ was evaluated with the Model 49i, and NO and NOx were measured using the Model 42i. Additionally, PM (PM_2.5 and PM₁₀) was tracked with the Model 5030i SHARP from Thermo. Meteorological data, such as ambient temperature (TEMP), relative humidity (RH), wind speed (WS), and precipitation (PR), were sourced from an automatic weather system managed by the local meteorological station.

Filter weighing

The PM_2.5 sampling filters were equilibrated for 48 h prior to and following sampling, maintained under strictly controlled conditions of 20 ± 3°C temperature and 50 ± 5% relative humidity (RH). Following this, the filters were weighed using a high-precision microbalance (ME5, Sartorius) that has a sensitivity of 10⁻⁵ g to determine the PM_2.5 mass concentration. Each filter was subjected to a minimum of three weight metrics, with the variations among the three readings being less than 0.04 mg. The blank filters demonstrated consistent mass concentrations, with variations of less than 5% in PM samples. The weighted filters were positioned in polystyrene Petri dishes, sealed within bags of the same material, and subsequently stored at a temperature of −20°C in a refrigerator.

Water-soluble ions analysis

A quarter section of each quartz fiber filter was employed to the analysis of water-soluble inorganic ions. The sampling filter was placed into a glass tube, to which 10 mL of deionized water was added. Following a 15-min ultrasonication at ambient temperature, the solution was transferred into a 5 mL syringe, passed through a syringe filter, and subsequently loaded into a polymeric vial equipped with a filter cap. These polymeric vials were then put into a Dionex AS-DV Autosampler and analyzed by ion chromatography (ICS-1000, DIONEX), targeting inorganic ions Na⁺, NH₄⁺, K⁺, Mg²⁺, Ca²⁺, Cl⁻, NO₃⁻ and SO₄²⁻. The method detection limits for Na⁺, NH₄⁺, K⁺, Mg²+, and Ca²⁺ were 12.45 μmol/L, 1.95 μmol/L, 3.15 μmol/L, 2.63 μmol/L, and 4.95 μmol/L, respectively. For Cl⁻, NO₃⁻, and SO₄²⁻, the limits were 7.2 μmol/L, 38.5 μmol/L, and 69.6 μmol/L, respectively. The reproducibility’s relative standard deviations for each ion were below 5%. The final concentrations of WSIIs were adjusted by deducting the blanks.

Carbonaceous materials analysis

A 0.5 cm² segment from each quartz fiber filter was analyzed for organic and elemental carbon (OC and EC). OC and EC were measured using a DRI-2001A thermal/optical carbon analyzer (CRI-2015, Desert Research Institute, USA) at the seven wavelengths (405, 445, 532, 635, 780,808 and 980 nm) in accordance with the IMPROVE (Interagency Monitoring of Protected Visual Environments) thermal/optical reflectance (TOR) protocol as stated by.⁴⁷

We used IMPROVE_A warming protocol as described by (Cao et al., 2007; Chow et al., 1993) to determine the carbon components by gradually increasing the temperature from 140°C to 840°C. The organic carbon components (OC1, OC2, OC3 and OC4 at 140. 280, 480°C and 580°C, respectively, in a He atmosphere), and elemental carbon components (EC1, EC2, and EC3 at 580, 740, and 840°C, respectively, in a 2%/98% He atmosphere) were measured. OC and EC were calculated according to the below Equations 6 and 7.

$$\text{OC}=\text{OC}1+\text{OC}2+\text{OC}3+\text{OC}4+\text{OPC}$$ (Equation 6)

$$\text{EC}=\text{EC}1+\text{EC}2+\text{EC}3-\text{OPC}$$ (Equation 7)

Where, OPC was the decomposition of some OC into EC at higher temperature, which was calculated with its absorption properties at 635 nm. Field blanks were analyzed for each batch of filters, with the resulting EC and OC concentrations falling below the method detection limits of OC (0.42 μg/cm²) and EC (0.12 μg/cm²). The concentrations of OC and EC have been adjusted based on field blanks. A replicate measurement was conducted once for every 10 samples, with relative deviations maintained at < ± 5%.

Elements analysis

Regarding metallic elements, polypropylene filter was divided into two equal parts for analysis of inorganic elements. One-half of polypropylene filter was divided into fragments and transferred into a conical flask. Subsequently, 5 mL of HF was introduced into the flask and subsequent heating at 120°C for 2h, and then dried at 130°C and finally 5 mL HCl was introduced to the flask and then heat again. The cooled solution was filtered and diluted to 15 mL for determine the concentrations of 20 particle-bound elements (Sr, Ba, Be, Na, K, V, Cr, Mn, Co, Ni, Cu, Zn, As, Mo, Cd, Sn, Sb, Tl, Pb, and Bi) by using an inductively coupled plasma mass spectrometer (ICP-MS, 7500, Agilent Technologies, USA).The other half of the filter, pre-treat with NaOH, was analyzed for Si, Al, Mg, Ti, Ca and Fe using an inductively coupled plasma optical emission spectrometer (ICP-OES, 5110, Agilent Technologies, USA). The method detection limits for trace elements ranged from 0.004 to 0.0065 μg/m³. A minimum of 10% of the samples went through evaluation with spiking with a predetermined quantity of metal to determine the recovery efficiencies. The recovery efficiency ranged from 85% to 103%.

Quantification and statistical analysis

In this study, all quantitative and statistical analysis were conducted using Microsoft Excel. Graphs were created using ArcGIS 10.8 (Figure 1), origin 2025 (Figures 2, 3, 4, 5, 7, 8, 9, 10, 11) and MATLAB 2024a (Figure 6).