Introduction

Carbonaceous aerosol is among the major constituents of atmospheric particle (Saylor et al. 2006; Schwier et al. 2010). Particulate carbonaceous species have triggered more crucial environmental problems of particular concern, having a close connection with human health issues for its longer lifetime in the atmosphere and easy penetration into human lung (Künzi et al. 2013). Carbonaceous particles consist of a complex mixture of chemical compounds and are usually divided into two major fractions. Elemental carbon (EC) derives from the incomplete combustion of carboniferous materials, such as coal, gasoline, and diesel, and the burning of biomass. Particulate organic carbon (OC) can be emitted from primary sources, especially incomplete combustion of fossil fuels, or formed in the oxidation process of volatile organic compounds followed by condensing on preexisting aerosol (Jiang et al. 2012; Kroll and Seinfeld 2008).

Carbonaceous aerosol is not only a pollutant but also an important driver of global change. It has long been known that EC is referred to as absorbing component while OC is associated with optical scattering until absorbing organic components have been recognized (Andreae and Gelencsér 2006; Sun et al. 2007; Yang et al. 2009). Absorption of solar radiation by carbonaceous aerosols may heat the atmosphere, thereby altering the vertical temperature profile, while scattering of solar radiation may lead to a net cooling of the atmosphere/ocean system. Besides directly by scattering or absorbing light, carbonaceous particles also affect the Earth’s radiative balance indirectly by acting as cloud condensation nuclei (CCN), thereby influencing the albedo and lifetime of clouds. Moreover, the influence of carbonaceous particles exerted on climate was also complicated by their changing mixing status (Huang and Yu 2008; Leaitch et al. 2010; Yu et al. 2010). The evaluation of the health and climate impacts of carbonaceous aerosols is burdened with relatively large uncertainties due to very limited understanding of the regional and global distributions of carbonaceous aerosol.

In China, rapid economic growth and urbanization have been changing the physical, chemical, and optical characteristics of atmospheric environment (Wang et al. 2011a, b). Moreover, deserts over northern China are major emission sources of Asian dust. Thus, the aerosols over northern China in spring comprise complex compositions including anthropogenic and natural particles in a highly mixed condition. Estimates of health and climate effects of EC and OC in China are more uncertain in such a huge territory and high density population country, which makes quantifying the weather-climate effects of carbonaceous aerosol more difficult. The crux of understanding the distribution and climate effect of carbonaceous aerosol in China is the lack of a consistent measurement network in China with good spatiotemporal coverage like the Interagency Monitoring of Protected Visual Environments (IMPROVE) network in USA and EMEP in Europe. Carbonaceous aerosols in China have been studied at an increasing number of locations; however, these studies were mainly based on bulk PM2.5 or PM10 and focus on the rapid economic growth and high population density (Li and Bai 2009; Zhou et al. 2012). So far, only two papers reported carbonaceous aerosols simultaneously at multiple locations across China in winter and summer and over an entire year, respectively (Cao et al. 2007; Zhang et al. 2008). They both concluded that high levels of OC and EC, and higher OC/EC ratios in China. To our best knowledge, there is no systematic atmosphere watch network which covers different ecosystems and provides size distribution information of carbonaceous particle in China although the wealth of measurements should be compiled to present us aerosol formation, growth, and removal which determine its health and climate effects. In the present study, OC and EC data collected in a spring season over northern China have been analyzed. The purpose of these measurements was to provide more useful data to reduce the uncertainties in quantification of carbonaceous aerosol health effect in northern China during spring.

Materials and methods

Site descriptions and aerosol sampling

The aerosol sampling was conducted across northern China from March to May, 2012. The locations of the observation network shown in Fig. 1 are summarized in details in Table 1. Size-segregated particles were collected at the sites using eight-stage air samplers (Andersen, USA) at an airflow rate of 28.3 L/min for 48 h from 9:00 a.m. Monday to 9:00 a.m. Wednesday bi-weekly, with cutoff points as 0.4, 0.7, 1.1, 2.1, 3.3, 4.7, 5.8, and 9.0 μm. The filtration media were 81-mm quartz microfiber filters (Munktell Filter AB, Swedish) which were cleaned by heating at 800 °C for 3 h before using.

Fig. 1
figure 1

Locations of sites over northern China

Full size image
Table 1 Site and sampling descriptions
Full size table

Elemental carbon and organic carbon measurements

A thermal/optical carbon aerosol analyzer (DRI Model 2001A, Desert Research Institute, USA) was used to analyze EC and OC content. The temperature program for the thermal analysis followed the Interagency Monitoring of Protected Visual Environments (IMPROVE_A) protocol (Chow et al. 2004, 2007). A punch aliquot (0.50 cm2) of a quartz fiber filter sample was heated stepwise in an oven at 140 °C (OC1), 280 °C (OC2), 480 °C (OC3), and 580 °C (OC4) in a pure helium atmosphere for OC volatilization and 580 °C (EC1), 740 °C (EC2), and 840 °C (EC3) in a 2 % oxygen-contained helium atmosphere for EC oxidation. At each stage, the CO2 formed was catalytically converted to CH4 by a MnO2 catalyst, and the resulting CH4 was then measured using a flame ionization detector (FID). The analyzer was calibrated using a standard mixture of CH4 and CO2 before and after sample analysis. The charring OC is little in my samples, so charring correction by optical correction was not made in this study. Field blank filters were also analyzed, and the sample results were subtracted by the average blank concentrations, which were 1.8 ± 0.4 and 0.1 ± 0.1 μg m−3 for OC and EC, respectively.

Results and discussion

Spatial variations of concentration of OC and EC

Among all the sites, Taiyuan had the highest OC and EC concentrations, which is predominantly attributable to heavy coal consumption of industrial complexes from the power plants and chemical industrial plants (Meng et al. 2007). OC and EC in PM9.0 in Taiyuan were 46.6 ± 5.3 and 11.6 ± 3.9, respectively. Xianghe is a distant suburb near Beijing City, and OC and EC at this station were higher than those in Beijing. Xianghe is heavily polluted due to both local and distant sources of fossil fuel combustion, vehicular and industrial emissions, and occasional dust and biomass fires (Yang et al. 2009). High average OC and EC were also found at the agricultural site (Yucheng, Shandong). This site stands for rural areas in China which has high income levels of peasants and rural population densities. Open biomass burning such as agricultural residues in Shandong Province may contribute to the high levels of EC and OC at this agricultural site (Cao et al. 2006). Agricultural field burning appeared in part of Shandong and Hebei Provinces could exert influence even on Beijing (Duan et al. 2004). In fact, biomass burning is ubiquitous in rural areas in China. As Zhang et al. ( 2012) reported, biomass burning and anthropogenic emissions exerted influence on carbonaceous aerosols in desert area in northeastern China.

Very high OC and EC concentration levels were found in the North China Plain and Northeast China Plain. Many large cities located in these regions, such as Beijing, Taiyuan, and Shenyang, had high OC and EC concentrations. Some rural areas in the above plains also had high OC and EC concentrations, such as Yucheng and Xianghe. This matched with the BC emission inventory very much, in which very high emission densities can be found in the North China Plain, Northeast Plain, and Shanxi Highland (Wang et al. 2012). Also, relatively high BC emissions can be found in almost all large cities and populated rural areas, such as those in the Provinces of Shandong, and Shanxi, where rural solid fuel burning and coke ovens are the major sources (Wang et al. 2012).

Sites in Northwest China (including Dunhuang, Fukang, Shapotou, and Cele) had lower OC and EC concentration levels than those in North China and northeastern China. However, compared to other sites in desert regions around the world, the EC and OC concentrations were higher in this study. For instance, EC and OM in PM2.5 in Nagarkot during spring were 1.5 and 23 μg/m3, respectively (Carrico et al. 2003). PM2.5 EC and OC in a desert region in northern Chile between December 17, 2007 and January 20, 2008 were 1.084 and 3.728 μg/m3, respectively (Jorquera and Barraza 2013). In Israel and Jordan, the annual average results in PM2.5 were as follows: Eilat, 0.82 μg EC/m3 and 3.3 μg OC/m3; Aqabah, 0.92 μg EC/m3 and 3.7 μg OC/m3; and Rachma, 0.45 μg EC/m3 and 2.2 μg OC/m3 (Schneidemesser et al. 2010).

According to the EC tracer method, primary organic carbon (POC) and secondary organic carbon (SOC) were calculated (Table 2) (Turpin et al. 1991; Cabada et al. 2004). The averaged fractions of SOC showed that the contributions from SOC to OC loadings were in the range of 30.7 % (in Xianghe) and 80.9 % (in Fukang). Although SOC estimation is burdened with substantial uncertainty using EC tracer method, it still suggested an important role of secondary organic aerosol (SOA) formation in China, especially in rural areas. In the urban regions, such as Shenyang and Taiyuan, POC had high percentages (32.1 ± 17.1 and 39.2 ± 21.3 %, respectively), indicating high OC and EC emissions. In Xianghe, SOC accounted for 30.7 ± 13.3 % of OC. This figure was lower than those of urban sites and may reflect local anthropogenic emissions in rural area.

Table 2 Averaged daily concentrations (μg/m3) for EC, POC, and SOC in PM9 in the spring
Full size table

Size distributions of OC and EC

Size distributions of EC and OC could reflect the emission sources, formation, and growth mechanism of aerosol particles. Also, the information of OC and EC may be helpful in understanding the role particulate carbon plays in urban, regional, and global climate change.

Due to absence of cutoff size of 2.5 μm in Andersen eight-stage air samplers, fine and coarse particles are defined as particles with sizes <2.1 and >2.1 μm, respectively. Figure 2 presents size distributions of OC and EC at all the sites during spring periods. Except for the Cele site, EC at all the sites showed a bimodal distribution with primary mode in the size range of 0.4–1.1 μm and secondary mode in the size of 4.7–5.8 μm and stood for fine and coarse modes, respectively.

Fig. 2
figure 2figure 2

Particle size distributions (dM/dlogDp) of OC and EC at all the observation sites

Full size image

The primary peak in fine mode EC predominantly indicated combustion processes, such as vehicular exhaust, coal combustion, and biomass burning. As mentioned above, fossil source may be relatively more important for urban areas and while biomass burning may have great influence on OC and EC in rural areas (Chen et al. 2013). Another peak in the coarse modes existed in the size range of 4.7–5.8 μm, albeit with a low mass fraction. Since the RH was below 40 %, the peak was not attributable to hygroscopic reaction which leads to growth in particle sizes. So the peak could have been caused by four reasons: (1) During impactor sampling, coarse EC as reaction interface, where many constituents are processed though heterogeneous reaction, produced all kinds of secondary products, such as SOC, sulfate, and natrate. Thus, internal mixture absorbed much more EC; (2) during impactor sampling, EC in fine modes was in adherence to coarse modes; (3) re-suspension of EC-containing soil/dust particles and tire abrasion in road dusts (Glaser et al. 2005). Han et al. (2009) reported that road dust corresponded well with the human emission sources, showing high EC levels; (4) in China, lower combustion efficiency is another reason for coarse soot.

Except for Cele, OC at all the sites also showed a bimodal distribution with primary mode in the size range of 0.4–1.1 μm and secondary mode in the size of 4.7–5.8 μm, respectively. OC showed a large peak in many sites, such as Beijing, Shenyang, and Xianghe. This may indicate the effect of dust on the size distribution of OC. Yan et al. (2012) found a similar phenomenon in Beijing during spring 2004.

OC and EC were concentrated in fine mode. Furthermore, EC was more enriched in fine mode. For instance, At the Yantai, Yucheng, and Xinglong sites, where the fine peaks were higher than the coarse ones, PM2.1 EC accounted for 77 %, 73 %, and 71 % of PM9.0 EC, respectively. The percentages for OC were 58 %, 59 %, and 64 %, respectively. This probably reflects the generally greater proportional contribution of coarse particles to PM mass at the urban sites due to particle re-suspension, industrial activity, etc. Furthermore, The percentages of OC and EC concentrated in fine mode were higher in Northwest China than in North China and northeastern China. This phenomenon may reflect the effect of dust on carbonaceous aerosols. Situated in a desert with considerable sources of dust around, it does not seem surprising that there would be significant coarse mode contributions at times. Spring is well known as an enhanced period of dust storms in that area of the world, and those storms can have very long range impacts. Therefore, there is a  possibility that OC deposited on the desert dust may contribute to the  results as well as the possibilities that OC is attached during dust transport or SOA is formed by reaction of VOCs on dust particles.

The mass median aerodynamic diameter (MMAD) values of OC and EC in the fine and coarse mode and total particle size range are presented in Table 3. The MMAD values of OC and EC in the fine mode ranges from 0.65 to 0.85 μm and from 0.51 to 0.85 μm, respectively. These size ranges were larger than typically emitted by primary sources, such as diesel, gasoline-powered cars, coal combustion, and biomass burning, which existed in the range of 0.1–0.52 μm (Hays et al. 2004; Hildemann et al. 1991; Kleeman et al. 2000; McElroy et al. 1982). This phenomenon indicated aged carbonaceous aerosols in this study.

Table 3 Mass median aerodynamic diameter (MMAD) values of EC and OC (μm)
Full size table

Among all the sampling sites, MMAD values of EC and OC in the fine mode were smaller in the urban air and higher in the rural/remote sites. This may be due to the differences in their sources and atmospheric processes such as in-cloud processing during long-range atmospheric transport. In densely populated rural area in northern-central China where coal is in abundant supply, small domestic stoves are used for cooking and heating. Usually, coal combustion from these stoves has insufficient efficiencies and introduces larger EC particles. Another explanation is that when aerosol particles containing EC are transported from high concentration urban areas to the rural sites, EC gets aged with extra OC or other ion coatings.

Relationship between OC and EC and OC/EC ratios

EC is directly emitted from combustion emission sources, so it has often been used as an indicator of POC. Therefore, relationship between OC and EC and OC/EC ratios can be used to indicate the emission sources of carbonaceous aerosol. Higher R 2 means stronger similarity of the emission sources of OC and EC.

Urban areas in China showed stronger correlations than rural/remote sites; for instance, strong correlations between OC and EC in PM2.5 were observed in spring (R 2 = 0.67) and winter (R 2 = 0.84) in Shanghai (Hou et al. 2011). The correlation between OC and EC in PM2.5 was significant (R 2 = 0.90 and 0.66 in autumn and winter, respectively) in Xi’an (Cao et al. 2005). These stronger correlations can be probably explained by similar dominant emissions from traffic and industry sources in urban and surrounding rural areas in the three regions. Contrarily, the correlation between OC and EC in PM2.5 in Daihai (a rural site) was moderate (R 2 = 0.41 in spring) (Han et al. 2008).

Correlations between OC and EC in northern China are depicted in Fig. 3 to examine the similarity between the OC and EC emission sources. Level of significance for the r 2 values was 0.05 for all the sites.

Fig. 3
figure 3figure 3

Correlations between EC and OC at the sites over northern China

Full size image

When the full data set of North China was separated into several data populations by particle sizes, the correlation coefficients would rise slightly. Despite the different locations of the six monitoring stations, results in this study showed that the correlations between OC and EC were very similar among the urban sites. This similarity may be explained by the fact that this region is highly urbanized and this area is with strong influences of similar pollutant sources.

OC/EC ratio from biomass fire is 8.0 in spring according to the inventory of black carbon and organic carbon emissions from China (Cao et al. 2008). The average OC/EC ratios were 12.0 for long-range transport, 2.7 for coal combustion, and 1.1 for motor vehicles (Saarikoski et al. 2008; Watson et al. 2001) in field experimental measurements. Therefore, OC/EC ratios from coal combustion and motor were much lower than those from biomass burning and long-range transport which accompany decreasing EC and OC loading (Zhang et al. 2008).

The scatter plot between OC and EC concentrations in fine particle showed better correlation than in the coarse mode in northern China, especially at sites in North China, Shenyang, Dunhuang, and Fukang sites, implying common source contributions to EC and OC in fine mode due to anthropogenic activities. The weak correlation in northwestern China indicated that carbonaceous aerosols in this region have more varying source contributions. In North China, where population, vehicles, and industries are concentrated, coal combustion and motor may be the main source of carbonaceous aerosol. In Northeast and Northwest China populated rural areas, biomass burning may play a dominant role.

The OC/EC ratios in PM1.1, PM2.1, and PM9.0 in this study were similar with those reported in the rural sites, but higher than those measured at the urban sites (Zhang et al. 2008). One reasonable explanation is that these sites were influenced by open biomass burning near them in spring. For instance, biomass burning is a significant repetitive pollution factor in Beijing (Duan et al. 2004). It is interesting that the OC/EC ratios in the fine particles were lower than those in the coarse particles with the aerodynamic diameter between 2.1 and 5.8 μm. A higher and increasing OC/EC ratio usually accompanies decreasing EC loading in the coarse particles. One explanation is that during long-range transport, the EC and OC concentrations are diluted with cleaner air. With the addition of semi-volatile organics (primary or secondary) onto the preexisting particles, the aged aerosol particles contain more OC, resulting in an increased OC/EC ratio in the coarse particles. OC/EC ratios at the Xinglong site were lower than those of Changbai Mountain and Cele for PM1.1, PM2.1, and PM2.1-9.0, respectively. For instance, in submicron, OC/EC ratio was 3.9 at the Xinglong site, and the OC/EC ratios were 5.1 and 6.7 at the Changban Mountain and Cele sites, respectively. This may due to urbanization and industrialization in North China. In PM2.1-9.0, OC/EC ratios at the Changbai Mountain and Cele sites were much higher than that of Xinglong. This may give some indication of biomass burning in the remote regions.

The relative amounts of optically absorbing carbon and optically scattering carbon control the direct radiative climate forcing effect of carbonaceous aerosol. Besides EC, a part of organic carbon (OC) strongly absorbs in the UV and visible wavelengths but has weak absorption in the near-infrared (IR) (Andreae and Gelencsér 2006; Sun et al. 2007; Yang et al. 2009). So caution needs to be exerted when assessing the impacts of carbonaceous aerosols in northern China on regional and global climate.

Conclusions

This study analyzed the spatial variations, size distributions, and characteristics of sources of carbonaceous aerosols on the basis of data over northern China during spring 2012. Both urban and rural areas of China showed higher concentrations of carbonaceous aerosols than clear sites in the world. A notable contrast between the urban and rural/regional background areas showed anthropogenic carbonaceous particle pollutants in the densely populated and urbanized area. Among all the sites except Cele, OC and EC presented a bimodal distribution with two peaks at the sizes of 0.4–1.1 and >4.7 μm. OC and EC were mainly enriched in fine modes. Based on the calculation of the MMAD values for OC and EC, carbonaceous aerosols in the fine mode in the urban areas are mostly presented in smaller sizes than those in the rural//regional background areas. Such spatial variations demonstrated the differences in the sources and the atmospheric processes of OC and EC in different areas. Good correlations between OC and EC in all the cities (5 in North China and 1 in northeast China) may suggest the impact anthropogenic emissions on carbonaceous aerosols in the above regions. In the fine mode, lower OC/EC ratios and stronger correlations between OC and EC mainly arise from anthropogenic emissions, especially motor exhaust and coal combustion. In coarse mode, poor correlations between OC and EC and higher OC/EC ratios may indicate biomass burning and aerosol aging during spring. In North China, where population, vehicles, and industries are concentrated, coal combustion and motor may be the main source of carbonaceous aerosol. In Northeast and Northwest China populated rural areas, biomass burning may play a dominant role.