1 Introduction

With the rapid development of the social economy and continuously increasing energy consumption, air pollution has become a serious problem in China in recent years (Chan and Yao 2008; Lin et al. 2008; Rose et al. 2010; Yang et al. 2009; Zhang et al. 2012). Beijing has a population of 16 million within an area of 16,800 km2, making it one of the largest and most densely populated cities in China. Beijing is located at the northwestern border of the North China Plain and is bound by mountains on the north, east, and west. Many heavily populated industrialized cities are located near Beijing to the southwest and southeast (Xu et al. 2011). Unfavorable geographical conditions and rapid growth in traffic emissions and regional pollutant emissions have made Beijing one of the most seriously polluted cities in China. The latest research (Yang et al. 2015) showed that about 28 % ~ 36 % of PM2.5 in Beijing was transported from the exterior region. Among the local emissions, the main sources of PM2.5 were vehicle (31.1 %), coal consumption (22.4 %), industrial production (18.1 %), dust (14.3 %) and others (14.1 %), respectively.

Particulate matter (PM), especially PM2.5 (fine particles with aerodynamic diameters less than 2.5 μm), plays important roles in atmospheric visibility reduction, acid deposition, and climate change (Yan et al. 2008; Garland et al. 2009; Ma et al. 2011). PM2.5 also has significant adverse health effects. Exposure to high concentrations of PM2.5 has been found to result in increased hospitalizations and higher mortality rates (Michaels and Kleinman 2000; Dockery 2001; Schwartz et al. 2002; Zhang et al. 2010; Zheng et al. 2015). PM2.5 is a complex mixture of elemental carbon (EC), organic carbon (OC), nitrate (NO3 −), sulfate (SO4 2−), ammonium (NH4 +), mineral dust, sea salt, and heavy metals (Hueglin et al. 2005). OC and EC affect the extinction coefficient for light scattering and absorption, which can contribute up to 30 % of the total extinction (Chan et al. 1999). The strong absorption capacity of EC has a significant impact on aerosol radiative forcing, which can lead to global warming (Menon et al. 2002; Ramanathan et al. 2001; Bond et al. 2013). Water-soluble ions such as SO4 2−, NO3 −, and NH4 + are hydrophilic and can promote the formation of cloud condensation nuclei, which significantly influence visibility and climate (Yao et al. 2002; Wang et al. 2005).

Since 2000, the number of studies of PM2.5 has gradually increased in various large city agglomerations in China such as the Pearl River Delta (Cao et al. 2004; Lai et al. 2007; Hagler et al. 2006; Duan et al. 2007; Hu et al. 2008; Liu et al. 2008), the Yangtze River Delta (Ho et al. 2011; Huang et al. 2012; Yang et al. 2005; Ding et al. 2013; Wang et al. 2008; Li et al. 2011a, b), and the Beijing–Tianjin–Hebei region (Wang et al. 2015a; Wang et al. 2015b). In Beijing, several studies have examined the general characteristics of PM2.5 chemical components and discussed their seasonal variations (He et al. 2001; Sun et al. 2004; Song et al. 2006). Also in Beijing, multiple studies have focused on correlations among PM2.5 components and the formation of secondary particles (Yao et al. 2002; Dan et al. 2004; Huang et al. 2006; Wang et al. 2009; Pathak et al. 2009; Ianniello et al. 2011). There have been several studies aimed at revealing the health effects of PM2.5 (Zhang et al. 2000; Guo et al. 2009; Kipen et al. 2010; Wu et al. 2010). Additionally, the aerosol number concentrations (Wu et al. 2007, 2008; Yue et al. 2009, 2011; Shen et al. 2011; Zhang et al. 2011; Gao et al. 2012), size distributions of aerosol chemical species (Yao et al. 2003; Cheng et al. 2009; Guo et al. 2010; Li et al. 2012), and aerosol optical characteristics and mixing states (Cheng et al. 2009; Deng et al. 2011a, b; Ma et al. 2012; Chen et al. 2012) have been discussed.

In this study, SO2, NO2, PM2.5, and chemical components of PM2.5 were measured at four sites in Beijing over four seasons. The main objectives of this paper were to (1) characterize the concentration levels and spatiotemporal variations in chemical components of PM2.5 in Beijing; (2) investigate the state of OC secondary pollution in Beijing; and (3) discuss the formation mechanisms of SO4 2− and NO3 − in urban Beijing.

2 Sampling and analysis

2.1 Sampling sites and occasions

Four sites, Dingling (DL), Chegongzhuang (CG), Fangshan (FS), and Yufa (YF), in different areas of Beijing were selected for this study, as shown in Fig. 1. DL is a background site in Beijing, which is located in the northern region of Beijing. CG is an urban environmental monitoring site in Beijing, which is located between 2nd ring and 3rd ring in Beijing. FS is a suburban environmental monitoring site in Beijing, which is located in the northeast of Fangshan district. YF is a regional transport pollution monitoring site in Beijing, which is located near the southern boundary in Beijing. Thus, the four sites can be a good representative of Beijing. Mass concentrations of gaseous pollutants and PM2.5 were measured by automatic monitoring instruments in the field (PM2.5 concentration data was only available in 2013). Chemical components of PM2.5 were manually sampled and analyzed in the laboratory. PNS 16 T-3.1/PNS 16 T-6.1 (Derenda, Germany) four-channel atmospheric particulate matter samplers with a flow of 16.67 L/min and Teflon membranes (Whatman, UK) were used to sample chemical components of PM2.5. Meteorological elements were also monitored at the CG site. The sampling period for the DL, CG, and FS sites was from August 2012 to July 2013. The sampling period for the YF site was from November 2012 to July 2013. A period of 5–7 continuous days was selected each month for sampling, and within that period, each site was monitored for 24 h.

Fig. 1
figure 1

Locations of the four monitoring sites

Full size image

2.2 Monitoring instruments and analysis methods

SO2 was monitored using a Thermo Fisher 43i pulsed UV fluorescence analyzer with a minimum detection limit of 0.5 × 10−9 (volume fraction) and accuracy of 1.0 × 10−9 (volume fraction). NO x was monitored using a Thermo Fisher 42C chemiluminescence NO–NO2–NO x analyzer with a minimum detection limit of 0.05 × 10−9 (volume fraction). PM2.5 was monitored with a Thermo Fisher 1405F analyzer with a tapered element oscillating microbalance (TEOM) monitoring mechanism with a minimum detection limit of 1.0 μg·m−3. OC and EC were analyzed with an RT-4 analyzer with a minimum detection limit of 0.2 μg·m−3 (Sunset Lab, USA). The analysis of OC and EC did not require any pre-treatment processes. Samples were stored under cryogenic refrigeration until use, and were then brought to room temperature before measurements. In the process of measurement, some OC tended to occur pyrolysis and be converted into EC at about 300 °C, thus, an optical detector was implied to solve this problem. A red laser was chose to be the source of the optical detector, and the strong absorption capacity of EC was applied to correct the errors caused by pyrolysis. In the end, a methane gas with known quantities was applied to calibrate the instrument (NIOSH 1996 and 1999). Water-soluble cations and anions were monitored using an ICS-2000 and an ICS-3000 ion chromatographic analyzer (Dionex), respectively. Sample filters and blank filters were digested in a clean Teflon-TFM digestion tank. Saturated boric acid solution (20 mL) was added to neutralize the excess F−. After another microwave digestion, the solution was brought to a 50-mL constant volume. Selenium was monitored with an AFS-9230 atomic fluorescence spectrometer (China), and other inorganic elements were monitored using an Intrepid II-XDL spectrometer (Thermo Fisher, USA). Ground meteorological elements were observed with a WXT520 meteorological instrument (Vaisala, Netherlands). Before each chemical component measurement, standard samples from the Institute for Environmental Reference Materials of the Ministry of Environmental Protection were applied to calibrate the instruments. A multi-point calibration was done every week. Parallel samples contributed at least 10 % of the total number of samples.

3 Results and discussion

3.1 Annual average concentrations of chemical components

After the laboratory analyses, 15 chemical components with relatively higher concentrations in PM2.5 were selected for further analysis, annual average concentrations of which are shown in Table 1. Overall, OC, SO4 2−, NO3 −, and NH4 + were dominant among 15 components, the annual average concentrations of which in the four sites were 22.62 ± 21.86, 19.39 ± 21.06, 18.89 ± 19.82, and 13.20 ± 12.80 μg·m−3, respectively. For most components, the lowest concentrations were observed at the DL site, reflecting the characteristics of the site’s urban background. The sum of the annual average concentrations of the 15 selected chemical components at the DL site was 64.61 μg·m−3. The sum of the annual average concentrations of the 15 selected chemical components at the YF site was 116.32 μg·m−3, 1.8 times that of the DL site. Overall, the concentrations of PM2.5 chemical components were higher in the south and lower in the north. The concentration of OC was the highest among the 15 components in DL, FS, and YF, accounting for 24.2 %, 24.4 %, and 30.7 % of the summed concentrations, respectively. NO3 − had the highest concentration among the 15 components in CG, accounting for 23.3 % of the summed concentrations. The CG site was close to Chegongzhuang West Road and Capital Stadium Road, where NO2 emitted by motor vehicles is converted into NO3 − through atmospheric chemical reactions (Xie et al. 2000; Fu et al. 2001; Carslaw 2005).

Table 1 Annual average concentrations of 15 chemical components of PM2.5 at the four sites (μg·m−3)
Full size table

Figure 2 shows the average concentrations of the 15 chemical components over four seasons. Metal elements accounted for a relatively small proportion of PM2.5, and their concentrations were higher in winter and lower in summer. The most abundant metal element was Si, concentration of which were higher in spring and summer, this may have been caused by sand and dust storms that occur in Beijing in the spring and summer (Zhang et al. 1993). Concentrations of the water-soluble ions K+ and Cl− were highest in winter and lowest in summer. The sampling period in February, 2013 was a few days after the Spring Festival in China. The discharge of fireworks has a significant effect on K+ (Cheng et al. 2013, 2014; Chang et al. 2011). Also, the burning of fossil fuels in winter facilitates the formation of organic and inorganic substances containing Cl−. Concentrations of the water-soluble ions NH4 + and SO4 2− were highest in winter and lowest in autumn. Coal combustion in winter increases the emission of NH3, and lower temperatures favor the formation of particulate inorganic nitrogen salts from NH3, NO x , and HNO3 (O’Dowd et al. 2000; Schlünzen and Meyer 2007; Elminir 2005; Watson et al. 1994). Coal combustion in winter also increases the emission of SO2, and the haze that frequently occurs in winter in Beijing promotes the formation of SO4 2− from SO2 (Sun et al. 2006; Galindo et al. 2008). The concentration of NO3 −, another water-soluble ion, was highest in spring and lowest in autumn. Studies have reported that the dust that invades Beijing carries particulate inorganic nitrogen components (Zhang and Iwasaka 1999). Additionally, coarse particles in dust, such as CaCO3, provide a reaction medium for NO3 −, NH4 +, Na+, and Ca+, facilitating the formation of NO3 −.

Fig. 2
figure 2

Seasonal variations in the concentrations of 15 chemical components at four monitoring sites

Full size image

Figure 3 shows the annual average concentrations of OC, EC, and water-soluble ions observed in this study and previous studies from 2001 to 2010 (Tan et al. 2009; Wang et al. 2005; Ho et al. 2006; Shon et al. 2012; Pathak et al. 2009; Zhao et al. 2013; Yang et al. 2011; Hagler et al. 2006; Kim et al. 2007; Qin et al. 2006). Concentrations of OC, NO3 −, and SO4 2− in this study were significantly higher than monitoring results for Beijing, Lanzhou, Hong Kong, Shanghai, New York, and Seoul from the period 2001–2006, whereas they were similar to monitoring results for Shijiazhuang and Tianjin from 2010. The average concentration of NH4 + was higher than all previously reported levels, as shown in Fig. 3 and Table 2.

Fig. 3
figure 3

Comparison of concentrations of OC, EC, and water-soluble ions from this study and previous studies

Full size image
Table 2 Comparison of concentration of NH4 + with previous monitoring results
Full size table

3.2 OC and EC

Both OC and EC concentrations were higher in winter and lower in summer, as shown in Fig. 4. The average concentration of OC over the four sites was 13.49 ± 6.00 μg·m−3 in summer, whereas in winter, it was almost three times higher at 39.30 ± 29.00 μg·m−3. The average concentration of EC over the four sites was 2.12 ± 0.92 μg·m−3 in summer and 5.22 ± 3.89 μg·m−3 in winter, 2.5 times higher. It can be concluded that coal combustion in winter significantly influenced OC and EC. In addition, the low boundary height caused by temperature inversion was another factor that influenced OC and EC. Among the four sites, the concentrations of EC were highest at the FS site in summer, autumn, and winter. The FS site was close to traffic trunks located outside the Fifth Ring Road, and heavy diesel vehicles were allowed on the road, increasing EC emissions.

Fig. 4
figure 4

Seasonal concentrations of OC and EC at the four monitoring sites

Full size image

The OC/EC ratio was applied to discern transformation characteristics of carbonaceous aerosols and to identify secondary sources of particulate matter. Most EC is produced from the incomplete combustion of carbonaceous fuels. Therefore, EC is considered a tracer of direct anthropogenic emissions (Strader et al. 1999). OC comes from both primary organic carbon (POC) that is directly emitted from pollution sources and secondary organic carbon (SOC) that is formed through photochemical reactions (Cabada et al. 2004). Studies have shown that when the OC/EC ratio is >2, SOC is formed in the atmosphere (Chow et al. 1996; Castro et al. 1999; Cao et al. 2009). In addition to photochemical reactions, biomass burning emits large quantities of OC, but relatively little EC, also resulting in high OC/EC ratios (Zhang et al. 2007).

The OC/EC ratio remained high throughout the entire year, as shown in Fig. 5. The average OC/EC ratios for the four seasons were 4.9, 7.0, 8.1, and 8.4, in the spring, summer, autumn, and winter, respectively. The OC/EC ratio was relatively low in the spring, and there were small differences among the four sites. The OC/EC ratio increased in the summer due to increased sunlight intensity and temperature. In summer, among the four sites, both OC and EC concentrations were lowest at the DL site, whereas the OC/EC ratio was the highest. This is likely due to the fact that EC concentrations for many of the samples from DL were <1.0 μg·m−3 in the summer, resulting in high OC/EC ratios. The monitoring results for DL were similar to those from a regional background site in Europe (Novakov et al. 2005).

Fig. 5
figure 5

Seasonal OC/EC ratios at the four monitoring sites

Full size image

3.3 Formation of SO4 2− and NO3 −

3.3.1 [NO3 −]/[SO4 2−] ratios

The [NO3 −]/[SO4 2−] ratio was applied to characterize the relative contributions of stationary sources and mobile sources (Arimoto et al. 1996). The [NO3 −]/[SO4 2−] ratios were lower in summer and winter and higher in spring and autumn, as shown in Fig. 6. Coal combustion in winter increased SO2 emissions. High temperatures, high humidity, and strong solar radiation in summer favor the formation of SO4 2−. The distribution of NO3 − across the gas phase, particle phase, and different sizes of particulate matter is related to temperature, humidity, SO4 2− concentration, and the concentration of crustal elements in the particulate matter. The combination of these potential factors causes seasonal variations in the [NO3 −]/[SO4 2−] ratio. The annual average [NO3 −]/[SO4 2−] ratios in DL, CG, FS, and YF were 1.01, 1.25, 1.08, and 1.12, respectively, which were significantly higher than previous studies in Beijing, indicating that the contribution ratio of mobile source increased in recent years in Beijing. At the CG site, the contribution of mobile source was higher than other sites due to its special location (Table 3).

Fig. 6
figure 6

Seasonal [NO3 −]/[SO4 2−] ratios at the four monitoring sites

Full size image
Table 3 [NO3 −]/[SO4 2−] ratios from previous studies
Full size table

3.3.2 Formation mechanisms of SO4 2− and NO3 −

The main sources of SO4 2− and NO3 − are photochemical reactions of SO2 and NO2, respectively. Therefore, concentrations of SO4 2− and NO3 − are related to the oxidation efficiency of SO2 and NO2. The sulfur and nitrogen oxidation ratios (SOR and NOR, respectively) were applied to characterize the conversion efficiency of SO2 and NO2 (Sun et al. 2006; Sun et al. 2014; Sun et al. 2015), as follows:

$$ SOR=\left[{\mathrm{SO}}_4^{2-}\right]/\left(\left[{\mathrm{SO}}_4^{2-}\right]+\left[{\mathrm{SO}}_2\right]\right)\kern0.5em NOR=\left[{\mathrm{NO}}_3^{-}\right]/\left(\left[{\mathrm{NO}}_3^{-}\right]+\left[{\mathrm{NO}}_2\right]\right) $$

Previous studies have demonstrated that the SOR is lower than 0.1 in pollutants that are discharged directly, such as vehicle exhaust (Pierson et al. 1979). Truex et al. (1980) and Ohta and Okita (1990) reported that when the SOR is higher than 0.1, SO2 participates in photochemical reactions through which SO4 2− is formed. The SOR values were higher than 0.1 throughout all four seasons, reflecting obvious photochemical reactions of SO2 in Beijing, as shown in Fig. 7. Among the four seasons, the SOR was significantly higher in the summer, likely due to the fact that higher temperatures are favorable for the transformation from SO2 to SO4 2− (Quan et al. 2008; McMurry P and Wilson J 1983). The NORs were significantly lower than the SORs. The NOR was highest in the spring at the CG site, but it was highest in summer at the other three sites. The annual average SORs were 0.433, 0.383, 0.438, and 0.460 at the DL, CG, FS, and YF sites, respectively. The annual average NORs were 0.269, 0.186, 0.223, and 0.305 at the DL, CG, FS, and YF sites, respectively. Among the four sites, the NO3 − concentration was highest at the CG site; however, the NOR was the lowest.

Fig. 7
figure 7

Seasonal SORs and NORs at the four monitoring sites

Full size image

Many mechanisms have been reported for the formation of SO4 2− from SO2, such as gas-phase reactions of SO2 and OH radicals, aqueous transformation processes (metal-catalyzed oxidation or H2O2/O3 oxidation), and in-cloud processes (Ziegler 1979; Meng and Seinfeld 1994; Xiu et al. 2004). The formation of SO4 2− and NO3 − is related to many factors, including the concentrations of precursors, concentrations of oxidant in gas- and liquid-phases, characteristics of preexisting aerosol particles, and meteorological conditions. The gas-phase conversion of SO2 to SO4 2− initiated by OH radicals is strongly a function of temperature (Seinfeld 1986; Radojević 1992), whereas heterogeneous reactions are closely related to humidity and particulate matter load (Liang and Jacobson 1999). Currently, studies of heterogeneous reactions on particle surfaces are in the preliminary stages compared with studies of homogeneous reactions. The Chinese scientist Zhu and his research group (Zhu et al. 2011; Shang et al. 2010; Xu et al. 2010; Ye et al. 2010; Tang et al. 2010; Li et al. 2011a, b) have conducted related studies on atmospheric heterogeneous reaction processes on particle surfaces, confirming the importance of heterogeneous reactions in the formation of secondary particles of PM2.5 in Beijing.

Meteorological data were only available at the CG site. The Spearman correlation coefficients between the SOR and NOR and the temperature and relative humidity (RH) were calculated (Table 4) to suggest formation mechanisms of SO4 2− and NO3 −, following the evaluation protocols of Yao et al. (2002) and Wang et al. (2009). In spring and summer, the SOR and temperature had a strong positive correlation, whereas the SOR and RH had no significant correlation, suggesting that a possible oxidation mechanism of SO2 to SO4 2− was the local gas-phase oxidation of SO2 by OH radicals, followed by condensation or absorption into the particle phase. The correlations of SOR with temperature and RH are close in autumn and winter. Thus, the formation mechanisms of SO4 2− included both slow gas-phase oxidation and heterogeneous reactions in the autumn and winter in urban Beijing. The NOR and temperature had no significant correlation during any season. Meanwhile the NOR and RH had strong positive correlations in spring, autumn, and winter, suggesting that heterogeneous reactions played major roles in the formation of NO3 − throughout the year in urban Beijing. Additionally, studies have shown that NO3 − is formed by gas-phase oxidation when NH4 + is in excess of SO4 2− (i.e., [NH4 +]/[SO4 2−] > 1.5; Pathak et al. 2004, 2009). However, the [NH4 +]/[SO4 2−] ratios at the CG site were 0.73, 0.62, 1.26, and 0.67 in the spring, summer, autumn, and winter, respectively, suggesting that NH4 + was not in excess of SO4 2−. Thus, gas-phase oxidation cannot explain the high NO3 − concentrations observed in urban Beijing. However, the analysis is based on the overall situation of the whole year. In some heavy air pollution process, the rapid increase of concentration of PM2.5 was related to more complex chemical reaction mechanism.

Table 4 Correlation coefficients of the SOR and NOR with temperature and RH at the CG site
Full size table

Laboratory and field experiments have shown that under conditions of high NO x concentrations, the heterogeneous hydrolysis reaction of N2O5 is one of the main sources of NO3 − in particulate matter (Anttila et al. 2006; Martinez et al. 2000; Hu and Abbatt 1997; McLaren et al. 2004; Hallquist et al. 2003). Two factors are favorable for the heterogeneous hydrolysis reaction of N2O5 to NO3 − on particle surfaces of existing SO4 2−. PM2.5 in Beijing has a large surface area and strong hygroscopicity resulting from the high concentration of PM2.5 and water-soluble ions in PM2.5. The formation and heterogeneous hydrolysis reaction of N2O5 include the R1–R3 reactions (Mentel et al. 1999; Underwood et al. 2001; Seisel et al. 2005; Bauer et al. 2004; Bian and Zender 2003). R1 is the speed-control reaction. HNO3 formed by R3 has high reactivity and participates in chemical reactions (R4–R7) to form secondary particulate matter such as surface nitrates (Hanisch and Crowley 2001; Goodman et al. 2001). In our monitoring results, the NORs were lower than the SORs; this may be related to the different formation mechanisms of SO4 2− and NO3 −, a question that warrants further study.

$$ {\mathrm{NO}}_2+{\mathrm{O}}_3\to {\mathrm{NO}}_3+{\mathrm{O}}_2 $$
(R1)
$$ {\mathrm{N}\mathrm{O}}_2+{\mathrm{N}\mathrm{O}}_3\to {\mathrm{N}}_2{\mathrm{O}}_5 $$
(R2)
$$ {\mathrm{N}}_2{\mathrm{O}}_5\left(\mathrm{g}\right)+{\mathrm{H}}_2\mathrm{O}(aq.)\to 2{HNO}_3(aq.) $$
(R3)
$$ CaO+2{HNO}_3\to Ca{\left({\mathrm{NO}}_3\right)}_2+{\mathrm{H}}_2\mathrm{O} $$
(R4)
$$ MgO+2{HNO}_3\to \mathrm{Mg}{\left({\mathrm{NO}}_3\right)}_2+{\mathrm{H}}_2\mathrm{O} $$
(R5)
$$ {\mathrm{CaCO}}_3+2{HNO}_3\to Ca{\left({\mathrm{NO}}_3\right)}_2+{CO}_2+{\mathrm{H}}_2\mathrm{O} $$
(R6)
$$ {\mathrm{MgCO}}_3+2{HNO}_3\to \mathrm{Mg}{\left({\mathrm{NO}}_3\right)}_2+{CO}_2+{\mathrm{H}}_2\mathrm{O} $$
(R7)

4 Conclusions

OC, SO4 2−, NO3 −, and NH4 + were dominant among 15 components, the annual average concentrations of which in the four sites were 22.62 ± 21.86, 19.39 ± 21.06, 18.89 ± 19.82, and 13.20 ± 12.80 μg·m−3, respectively. Concentration of OC and EC in winter were significantly higher than those in summer, indicating that coal combustion greatly influenced OC and EC. The annual average [NO3 −]/[SO4 2−] ratios in DL, CG, FS, and YF were 1.01, 1.25, 1.08, and 1.12, respectively, which were significantly higher than previous studies in Beijing, indicating that the contribution ratio of mobile source increased in recent years in Beijing. Correlation analyses between temperature and RH, and SOR and NOR showed that gas-phase oxidation reactions were a major mechanism of SO4 2− formation in the spring and summer in urban Beijing, whereas slow gas-phase oxidation reactions and heterogeneous reactions were both important in autumn and winter. Heterogeneous reactions were the main formation mechanism of NO3 − year round in urban Beijing.