Characteristics of Carbonaceous Aerosol in a Typical Industrial City — Nanjing in Yangtze River Delta , China : Size Distributions , Seasonal Variations , and Sources

In order to investigate the size distributions and seasonal variations of carbonaceous aerosols (organic carbon (OC) and elemental carbon (EC)), the carbonaceous species were collected in Nanjing, a typical industrial city located in Yangtze River Delta, China, in the summer, autumn, and winter of 2013 and spring of 2014, and then analyzed by using a 9-stage Anderson-type aerosol sampler and DRI Model 2001A Thermal/Optical Carbon Analyzer. OC, EC, secondary organic carbon (SOC), and primary organic carbon (POC) exhibited obvious seasonal variations, with the highest levels in winter (39.1±14.0, 5.7 ± 2.1, 23.6 ± 11.7, and 14.1 ± 5.7 μg·m−3, respectively) and the lowest levels in summer (20.6 ± 6.7, 3.3 ± 2.0, 12.2 ± 3.8 and 8.4 ± 4.1 μg·m−3, respectively), and were mainly centralized in PM1.1 in four seasons. The concentrations of OC in PM1.1 varied in the order of winter > autumn > spring > summer, while EC ranked in the order of autumn > winter > summer > spring. In the PM1.1–2.1 and PM2.1–10, the concentrations of OC and EC decreased in the sequence of winter > spring > autumn > summer. The size spectra of OC, EC, and SOC had bimodal distributions in four seasons, except for EC with four peaks in summer. The size spectra of POC varied greatly with seasons, exhibiting bimodal distribution in winter, trimodal distribution in spring and summer, and four peaks in autumn. The OC/EC ratios were 7.0, 6.3, 7.6, and 6.9 in spring, summer, autumn and winter, respectively, which demonstrated the abundance of secondary organic aerosols in Nanjing. The sources of carbonaceous aerosol varied significantly with seasons, and were dominated by vehicle exhaust, and coal and biomass burning in PM2.1, and dominated by dust, and coal and biomass burning in PM2.1–10.


Introduction
Carbonaceous aerosol fractions, accounting for 10%-50% of particle matter (PM) concentrations, play a key role in visibility, human health, Earth's radiation balance, and cultural heritage [1][2][3][4][5][6][7][8][9][10].Carbonaceous aerosols are generally divided into organic carbon (OC), elemental carbon (EC), and inorganic carbonate carbon (CC).The OC are normally generated from primary and secondary sources, while EC originate mainly from incomplete fuel combustions of industrial boilers/kilns and  1).The main meteorological factor changes were recorded by an automatic weather station.The size-segregated aerosol particles were continuously collected for 23 h.After excluding the invalid data, 135 samples were obtained during the observation period, there were 36, 44, 30, and 25 samples in spring, summer, autumn, and winter, respectively.Every sample contained data about OC and EC particles in nine size segments in PM10.The YRD region has a subtropical monsoon climate, which generally consists of a southeast wind, high temperature, and abundant precipitation in summer.The average temperature and relative humidity (RH) in summer were 26.4 ± 4.5 °C and 67.3% ± 16.2%, respectively (Table 1).Table 1 also shows that the RH in each of the four seasons was higher than 50%, and that the wind speeds were high in spring and summer, and low in autumn and winter.

Instrumentation
The sampling observations were carried out by a 9-stage Anderson-type aerosol sampler (Anderson 2000 Inc., Peachtree City, GA, USA) with size ranges of <0.43, 0.43-0.65,0.65-1.1,1.1-2.1,2.1-3.3,3.3-4.7,4.7-5.8,5.8-9.0, and 9.0-10.0μm for carbonaceous aerosols.The flow rate required by the Anderson-type aerosol sampler is 28.3 L/min.The sampler was operated with 81 mm quartz fiber filters (Whatman, Clifton, UK), and the filters were baked at 800 °C for 4 h before sampling in order to minimize the potential background effect of organic carbon.    .The main meteorological factor changes were recorded by an automatic weather station.The size-segregated aerosol particles were continuously collected for 23 h.After excluding the invalid data, 135 samples were obtained during the observation period, there were 36, 44, 30, and 25 samples in spring, summer, autumn, and winter, respectively.Every sample contained data about OC and EC particles in nine size segments in PM 10 .The YRD region has a subtropical monsoon climate, which generally consists of a southeast wind, high temperature, and abundant precipitation in summer.The average temperature and relative humidity (RH) in summer were 26.4 ± 4.5 • C and 67.3% ± 16.2%, respectively (Table 1).Table 1 also shows that the RH in each of the four seasons was higher than 50%, and that the wind speeds were high in spring and summer, and low in autumn and winter.

Instrumentation
The sampling observations were carried out by a 9-stage Anderson-type aerosol sampler (Anderson 2000 Inc., Peachtree City, GA, USA) with size ranges of <0.43, 0.43-0.65,0.65-1.1,1.1-2.1,2.1-3.3,3.3-4.7,4.7-5.8,5.8-9.0, and 9.0-10.0µm for carbonaceous aerosols.The flow rate required by the Anderson-type aerosol sampler is 28.3 L/min.The sampler was operated with 81 mm quartz fiber filters (Whatman, Clifton, UK), and the filters were baked at 800 • C for 4 h before sampling in order to minimize the potential background effect of organic carbon.
The OC and EC collected by quartz filter were then analyzed with a DRI Model 2001A Thermal/Optical Carbon Analyzer (Desert Research Institute, Reno, NV, USA), using the IMPROVE_A (Interagency Monitoring of Protected Visual Environments) protocol [47] to measure the carbon fractions.The filter was heated stepwise to 140 • C, 280 • C, 480 • C, and 580 • C in a pure helium environment to determine OC1, OC2, OC3, and OC4, respectively.Consequently, the filter was heated to 580 • C, 740 • C, and 840 • C in 2% O 2 /98% He to determine EC1, EC2, and EC3, respectively.One sample was selected at random from every 10 samples to carry out a duplicate sample analysis.The errors in the measurement presented here were less than 10% for total carbon (TC).Field blanks were used to determine any possible background contamination; the OC and EC concentrations in the field blanks were less than 1% of the sample levels.
For calibration and quality control, measurement with pre-fired filter blank, standard sucrose solution, and replicate analysis were performed.Field blank samples were collected in each season and their concentration levels of OC and EC were 0.5-0.7 and 0-0.03 µg•cm −2 , less than 3% and 1% of the samples, respectively.The samples were corrected by each of the field blank samples.The detect limit for OC and EC was 0.45 µg•cm −2 and 0.06 µg•cm −2 , respectively.More information about the instruments (e.g., precision and calibration) and the methods of data quality control were reported by Cao et al. [11], Chow et al. [4,47], Han et al. [48], and Li et al. [22,26].

Air Mass Backward Trajectories
Air mass backward trajectories for 24 h of each day were simulated using the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model, which was developed by the National Oceanic and Atmospheric Administration (NOAA) Air Resources Laboratory (ARL).The backward trajectories were calculated at 12:00 (LST, local time) and at the height of 500 m above the observation sites.The National Weather Service's National Centers for Environmental Prediction (NCEP) Global Data Assimilation System (GDAS) archive was used for meteorological input data.The GDAS data has a horizontal resolution of 1.0 • × 1.0 • .

Seasonal Variations of the Carbonaceous Aerosol
Figure 2a exhibits that the seasonal variations of OC ranked in the order of winter (39.1 µg•m −3 ) > autumn (30.5 µg•m −3 ) > spring (25.2 µg•m −3 ) >summer (20.6 µg•m −3 ) in Nanjing, which was similar to that of EC, which had the average concentrations of 5.7, 4.0, 3.6, and 3.3 µg•m −3 in winter, autumn, spring, and summer, respectively (see Figure 2b).This phenomenon might be attributed to the monsoon effects, where the air masses in summer were clean due to winds predominantly occurring from the southeastern ocean area or southwestern mountain area (Figure 3).In winter, the air masses were mainly from the inland area and carried pollutants transported by the dominant northwest wind (Figure 3d).The East Asian monsoon arrived in spring and autumn (see Figure 3), resulting in jumbled air masses mostly from ocean and inland, which corresponded to centered pollutant concentrations.The concentrations of OC and EC across the four seasons were mainly centralized in fine particles (Figure 2), with 59.1%-69.2%and 51.2%-71.9%, of the respective total mass concentrations in PM2.1, and 43.6%-54.3%and 34.5%-58.7%,respectively, in PM1.1.The clear enrichment of carbonaceous composition in smaller particle size, particularly for EC, indicates the important roles of certain combustion sources of anthropogenic origin (e.g., diesel vehicles and biomass burning), from which OC and EC were confirmed to be concentrated in fine particles [22,49].The concentrations of OC and EC across the four seasons were mainly centralized in fine particles (Figure 2), with 59.1%-69.2%and 51.2%-71.9%, of the respective total mass concentrations in PM2.1, and 43.6%-54.3%and 34.5%-58.7%,respectively, in PM1.1.The clear enrichment of carbonaceous composition in smaller particle size, particularly for EC, indicates the important roles of certain combustion sources of anthropogenic origin (e.g., diesel vehicles and biomass burning), from which OC and EC were confirmed to be concentrated in fine particles [22,49].The concentrations of OC and EC across the four seasons were mainly centralized in fine particles (Figure 2), with 59.1%-69.2%and 51.2%-71.9%, of the respective total mass concentrations in PM 2.1 , and 43.6%-54.3%and 34.5%-58.7%,respectively, in PM 1.1 .The clear enrichment of carbonaceous composition in smaller particle size, particularly for EC, indicates the important roles of certain combustion sources of anthropogenic origin (e.g., diesel vehicles and biomass burning), from which OC and EC were confirmed to be concentrated in fine particles [22,49].
The maximum values of OC were within the range of 0.65-1.1 µm in spring, autumn, and winter, accounting for 18.9%, 19.9%, and 21.6% of the total concentration, respectively.In summer, the maximum value of OC was within the range of 0-0.43µm, accounting for 19.8% of the total concentration.The maximum value of EC was within the range of 1.1-2.1 µm in spring and winter and 0-0.43 µm in summer and autumn, accounting for 20.3%, 16.8%, 32.7%, and 23.0% of the total concentration, respectively.
Table 2 shows that the seasonal variations of OC and EC were quite different in PM 1.1 , while they were comparable in PM 1.1-2.1 and PM 2.1-10 .The concentrations of OC and EC in PM 1.1 varied in the order of winter > autumn > spring >summer, and autumn > winter > summer > spring, respectively.In the PM 1.1-2.1 and PM 2.1-10 , the concentrations of OC and EC varied identically in the order of winter >spring > autumn > summer.
2.0 ± 1.4 0.9 ± 0. The concentrations of OC in PM 1.1-2.1 were almost identical in spring and autumn, as a result of frequent biomass burning processes in the two seasons.It was observed that the biomass burning mainly affected aerosol chemical composition at 1.1-2.1 µm [50,51].The concentrations of OC in PM 1.1-2.1 was 7.4 µg•m −3 in winter and was 3.1 times higher than summer.This phenomenon was mainly because of a stable boundary layer and few precipitation events in winter, additionally, haze events occurred frequently as well.
The proportions of EC in PM 1.1 varied in the order of summer (58.7%) > autumn (55.2%) > spring (40.2%) > winter (34.5%), and spring (20.3%) > winter (16.8%) > autumn (12.5%) = summer (12.5%) in PM 1.1-2.1 .EC particles are mainly from primary source emissions; in summer and autumn, the air masses were clean due to winds predominantly occurring from the southeast ocean area and because there were sufficient numbers of precipitation events, thereby leading to a low level of coarse EC particles.In winter, the air masses were mainly from heavy pollution districts of Shanxi, Henan, An'hui Province, and Jing-jin-ji and brought coarse EC particles as a result of the dominant northwest winds.The concentration of EC in PM 2.1-10 in winter was 3.0 times higher than summer, mainly because of a stable boundary layer and more coal emissions in winter.

Size Distributions of OC and EC
Figure 4a shows that the size spectra of OC had bimodal distributions in the four seasons.The highest peak appears in the size range of 0.43-0.65 µm, except for spring, and the second peak is located at 4.7-5.8µm.It should be pointed out that the highest peak in spring was observed to shift to a larger size (i.e., 0.65-1.1 µm).
carriers for carbonaceous constituents [53].In addition, biogenic aerosols, such as algae, pollen, vegetation debris, viruses, and microorganisms may be responsible for the second peak as well [28].
EC also presents bimodal distributions with the first peak at 1.1-2.1 μm, 0.65-1.1 μm, and 1.1-2.1 μm in spring, autumn, and winter (see Figure 4b).The second peak is consistently found at 4.7-5.8μm.Fine EC particles mainly come from vehicle exhaust, industrial emissions, and coal and biomass combustion [2,20,54].The EC peak at 4.7-5.8μm may be related to the re-suspension of EC-containing soil/dust particles [25].However, the size spectra of EC had four peaks in summer (Figure 4b), which were observed at 0.5, 1.6, 4.0, and 7.4 μm, respectively.The reason is not clear, and may be associated with subway construction at Ningliu road during the observation period.

The Characteristics of SOC and POC
The roles of secondary organic aerosol (SOC) in haze, visibility, climate, and heath have been recognized for decades, but few research studies have been carried out [11,26,31,55].As shown in Table 3, most ratios of the minimum value of OC/EC concentrations in the four seasons are above 2.0, indicating that the SOC particles were considered to be formed in the observation period.For the control of particulate pollution, it is very crucial to quantify the contributions of the primary and secondary organic carbon to carbonaceous aerosol.Although there is no simple direct analytical technique to attain the evaluation of SOC formation in ambient aerosols, several indirect methodologies have been applied [11,13,18,19,56].According to Castro et al. [13], the production of SOC can be calculated from the following equation:  3).The peak at 0.43-0.65 µm for OC was normally formed by the condensation and coagulation growth process of fine particles (gas-to-particle) [6,8,52].The larger peak at 4.7-5.8µm appeared to be related to the dust particles, which provide surfaces for the uptake of gas species and serve as carriers for carbonaceous constituents [53].In addition, biogenic aerosols, such as algae, pollen, vegetation debris, viruses, and microorganisms may be responsible for the second peak as well [28].
EC also presents bimodal distributions with the first peak at 1.1-2.1 µm, 0.65-1.1 µm, and 1.1-2.1 µm in spring, autumn, and winter (see Figure 4b).The second peak is consistently found at 4.7-5.8µm.Fine EC particles mainly come from vehicle exhaust, industrial emissions, and coal and biomass combustion [2,20,54].The EC peak at 4.7-5.8µm may be related to the re-suspension of EC-containing soil/dust particles [25].However, the size spectra of EC had four peaks in summer (Figure 4b), which were observed at 0.5, 1.6, 4.0, and 7.4 µm, respectively.The reason is not clear, and may be associated with subway construction at Ningliu road during the observation period.

The Characteristics of SOC and POC
The roles of secondary organic aerosol (SOC) in haze, visibility, climate, and heath have been recognized for decades, but few research studies have been carried out [11,26,31,55].As shown in Table 3, most ratios of the minimum value of OC/EC concentrations in the four seasons are above 2.0, indicating that the SOC particles were considered to be formed in the observation period.For the control of particulate pollution, it is very crucial to quantify the contributions of the primary and secondary organic carbon to carbonaceous aerosol.Although there is no simple direct analytical technique to attain the evaluation of SOC formation in ambient aerosols, several indirect methodologies have been applied [11,13,18,19,56].According to Castro et al. [13], the production of SOC can be calculated from the following equation: where SOC is the secondary OC, and (OC/EC) min is the minimum ratio observed.We calculated the SOC concentration in nine size segments in each of the four seasons based on the values of (OC/EC) min (as shown in Table 3).The concentrations of SOC in PM 1.1 and PM 2.1-10 varied identically in the order of winter > autumn > spring > summer, and varied in the order of winter > autumn > summer > spring in PM 1.1-2.1 (Table 2).The proportions of SOC in PM 1.1 , PM 1.1-2.1 , and PM 2.1-10 relative to the total concentration were 44.1%-57.1%,10.4%-17.7%, and 28.9%-43.6%,respectively.The formation mechanism of SOC is not clear at present, although it is generally considered to be from gas phase reactions followed by gas-to-particle conversion of products (i.e., through nucleation, condensation, and/or sorption), perhaps followed by aerosol phase reactions [13,55].
The concentrations of POC in PM 1.1 were ranked in the order of winter > autumn > summer > spring, and in the order of winter > spring > autumn> summer in both PM 1.1-2.1 and PM 2.1-10 (Table 2).POC in PM 1.1 was the highest in winter (6.4 µg•m −3 ) and were the same in the rest of the three seasons (4.1-4.9 µg•m −3 ).POC in PM 1.1-2.1 varied greatly with the seasons, and had the highest level in winter (3.0 µg•m −3 ) and lowest level in summer (0.9 µg•m −3 ).The concentrations of POC in PM 2.1-10 were almost identical across all of the seasons, with the slight exception in that the concentration in winter (4.7 µg•m −3 ) was slightly higher than in the other three seasons (3.0-3.5 µg•m −3 ).POC particles are mainly from direct emissions of combustion processes, and mostly centralized in fine particles.POC in coarse particles are observed to be mainly from the emission of plant spores and pollen, vegetation debris, tire rubber, and soil organic matter from wind erosion, as well as some non-combustion industrial activities [31,57].

Relationships between OC and EC
The relationships between OC and EC can provide valuable information for the source origins of carbonaceous aerosols [9,27].A strong correlation indicates common source origins and transport processes.Figure 5a shows that the correlation coefficients of OC with EC were the highest in summer (0.73), followed by spring (0.54) and autumn (0.52), and the lowest in winter (0.25).The strong correlations between OC and EC in spring, summer, and autumn samples suggests that they were derived largely from similar sources, which were different to those in winter.Meanwhile, the correlation coefficients between OC and EC varied greatly with seasons in PM 1.1 , PM 1.1-2.1 , and PM 2.1-10 .In PM 1.1 , the correlations were observed to be large in spring (0.60) and summer (0.76), and low in autumn (0.44) and winter (0.32) according to Figure 5b.In PM 1.1-2.1 , the correlation coefficients were high in spring (0.60), summer (0.70), and autumn (0.64), and low in winter (0.36) (Figure 5c).In PM 2.1-10 , the correlation coefficients were high in all seasons (0.54-0.87).It can be seen that source emissions of OC and EC in fine particles had large seasonal variations, whereas they had common sources for coarse particles.

Implications of the OC/EC Ratios
The ratio of OC/EC is an important diagnostic index that has been used to reflect the type and source strength of carbonaceous aerosols [10].The ratios can be mainly affected by emission sources, secondary organic aerosol (SOA) formation, and different removal rates by deposition [58].OC/EC ratios exceeding 2.0 have been used to indicate the presence of secondary organic aerosols [2,9]; 1.0-4.2indicate diesel and gasoline vehicle exhaust emissions [35,36], 2.4-14.5 indicates biomass burning emissions [59], 2.5-10.5 indicates coal combustion emissions [15], and OC/EC ratios of 32.9-81.6 indicate cooking emissions [34].Watson et al. [60] reported the OC/EC ratio of 13.1 for dust emissions.
Figure 6 reveals that the ratio of OC/EC exceeded 2 in all of the seasons, which demonstrated the abundance of SOC in Nanjing.The spectra of OC/EC had bimodal distributions in spring and summer.The two peaks were located at 0.43-1.1 μm and 2.1-5.8μm in spring, with values of 18.0 and 14.0, respectively.Therefore, the carbonaceous species in fine particles may originate from coal and biomass burning, and from coal and biomass burning and dust emissions in coarse particles.Additionally, long-range transport from the northern area was another important source of carbonaceous aerosols in Nanjing.The values of OC/EC in <2.1 μm were 4.2-9.4 in summer, suggesting that their sources were from vehicle exhaust, and coal and biomass burning emissions.The value of OC/EC in the coarse particles varied greatly insizes in summer, with the highest peak value of 26.9 at 4.7-5.8μm, and the lowest value of 8.8 at 9-10 μm, suggesting that the sources were from dust, and coal and biomass burning.
The spectra of OC/EC had trimodal distributions in autumn and winter (Figure 6).The peak at

Implications of the OC/EC Ratios
The ratio of OC/EC is an important diagnostic index that has been used to reflect the type and source strength of carbonaceous aerosols [10].The ratios can be mainly affected by emission sources, secondary organic aerosol (SOA) formation, and different removal rates by deposition [58].OC/EC ratios exceeding 2.0 have been used to indicate the presence of secondary organic aerosols [2,9]; 1.0-4.2indicate diesel and gasoline vehicle exhaust emissions [35,36], 2.4-14.5 indicates biomass burning emissions [59], 2.5-10.5 indicates coal combustion emissions [15], and OC/EC ratios of 32.9-81.6 indicate cooking emissions [34].Watson et al. [60] reported the OC/EC ratio of 13.1 for dust emissions.
Figure 6 reveals that the ratio of OC/EC exceeded 2 in all of the seasons, which demonstrated the abundance of SOC in Nanjing.The spectra of OC/EC had bimodal distributions in spring and summer.The two peaks were located at 0.43-1.1 µm and 2.1-5.8µm in spring, with values of 18.0 and 14.0, respectively.Therefore, the carbonaceous species in fine particles may originate from coal and biomass burning, and from coal and biomass burning and dust emissions in coarse particles.Additionally, long-range transport from the northern area was another important source of carbonaceous aerosols in Nanjing.The values of OC/EC in <2.1 µm were 4.2-9.4 in summer, suggesting that their sources were from vehicle exhaust, and coal and biomass burning emissions.The value of OC/EC in the coarse particles varied greatly insizes in summer, with the highest peak value of 26.9 at 4.7-5.8µm, and the lowest value of 8.8 at 9-10 µm, suggesting that the sources were from dust, and coal and biomass burning.

Conclusions
In order to investigate the size distributions and seasonal variations of carbonaceous aerosols (OC and EC), the carbonaceous species were observed for Nanjing, a typical industrial city located in Yangtze River Delta, China.OC, EC, SOC, and POC exhibited obvious seasonal variations, with the highest levels in winter (39.1 ± 14.0, 5.7 ± 2.1, 23.6 ± 11.7, and 14.1 ± 5.7 μg•m −3 ) and the lowest levels in summer (20.6 ± 6.7, 3.3 ± 2.0, 12.2 ± 3.8, and 8.4 ± 4.1 μg•m −3 ).The OC, EC, SOC, and POC across the four seasons were mainly centralized in PM1.1, with the proportions in the range of 34.5%-58.7%.OC concentrations at 0.65-1.1 μm were the highest in spring, autumn, and winter, and the lowest at 0-0.43 μm in summer.The EC maximum values were located at 1.1-2.1 μm in spring and winter, and at 0-0.43 μm in summer and autumn, respectively.The concentrations of OC and EC in PM1.1 varied in the order of winter > autumn > spring > summer, and autumn > winter > summer > spring, respectively.In the PM1.1-2.1 and PM2.1-10, the concentrations of OC and EC ranked uniformly in the order of winter > spring > autumn > summer.
The size spectra of OC, EC, and SOC had bimodal distributions acrossthe four seasons, except for EC, with four peaks in summer.However, the size distributions of OC, EC, and SOC in fine particles are a little complex in the four seasons.The size spectra of POC varied greatly with seasons, exhibiting bimodal distribution in winter, trimodal distribution in spring and summer, and four peaks in autumn.
The correlation coefficients between OC and EC were the highest in summer (0.73), followed by spring (0.54) and autumn (0.52), and the lowest in winter (0.25), indicating their common emission sources in spring, summer, and autumn.The correlation coefficients of OC with EC varied greatly with seasons in PM1.1, PM1.1-2.1 and PM2.1-10.Strong correlations between OC and EC were observed in spring (0.60) and summer (0.76) in PM1.1, in spring (0.60), summer (0.70), and autumn (0.64) in PM1.1-2.1, and in all seasons (0.54-0.87) in PM2.1-10.The ratio of OC/EC was 7.0, 6.3, 7.6, and 6.9 in spring, summer, autumn, and winter, respectively, which demonstrated the abundance of secondary organic aerosols in Nanjing.The sources of carbonaceous aerosols exhibited significant seasonal The spectra of OC/EC had trimodal distributions in autumn and winter (Figure 6).The peak at 0.43-1.1 µm had the highest value of 26.9-33.1 in autumn, suggesting cooking emission sources.The value of OC/EC in <0.43 µm was 5.7 for vehicle exhaust emissions; 13.0-13.8at 1.1-4.7 µm for dust and coal and biomass burning.The peak values at 4.7-5.8µm and 9.0-10 µm were 19.1 and 16.2, mainly from biomass burning.In winter, the values of OC/EC were 5.3-6.4 in <0.65 µm, indicating vehicle exhaust and coal combustion; 12.0-20.8at 0.65-2.1 µm, indicating coal and biomass burning, and 2.1-7.3 at 2.1-10 µm, indicating vehicle exhaust and coal combustion.

Conclusions
In order to investigate the size distributions and seasonal variations of carbonaceous aerosols (OC and EC), the carbonaceous species were observed for Nanjing, a typical industrial city located in Yangtze River Delta, China.OC, EC, SOC, and POC exhibited obvious seasonal variations, with the highest levels in winter (39.1 ± 14.0, 5.7 ± 2.1, 23.6 ± 11.7, and 14.1 ± 5.7 µg•m −3 ) and the lowest levels in summer (20.6 ± 6.7, 3.3 ± 2.0, 12.2 ± 3.8, and 8.4 ± 4.1 µg•m −3 ).The OC, EC, SOC, and POC across the four seasons were mainly centralized in PM 1.1 , with the proportions in the range of 34.5%-58.7%.OC concentrations at 0.65-1.1 µm were the highest in spring, autumn, and winter, and the lowest at 0-0.43 µm in summer.The EC maximum values were located at 1.1-2.1 µm in spring and winter, and at 0-0.43 µm in summer and autumn, respectively.The concentrations of OC and EC in PM 1.1 varied in the order of winter > autumn > spring > summer, and autumn > winter > summer > spring, respectively.In the PM 1.1-2.1 and PM 2.1-10 , the concentrations of OC and EC ranked uniformly in the order of winter > spring > autumn > summer.
The size spectra of OC, EC, and SOC had bimodal distributions acrossthe four seasons, except for EC, with four peaks in summer.However, the size distributions of OC, EC, and SOC in fine particles are a little complex in the four seasons.The size spectra of POC varied greatly with seasons, exhibiting bimodal distribution in winter, trimodal distribution in spring and summer, and four peaks in autumn.
The correlation coefficients between OC and EC were the highest in summer (0.73), followed by spring (0.54) and autumn (0.52), and the lowest in winter (0.25), indicating their common emission sources in spring, summer, and autumn.The correlation coefficients of OC with EC varied greatly with seasons in PM 1.1 , PM 1.1-2.1 and PM 2.1-10 .Strong correlations between OC and EC were observed in spring (0.60) and summer (0.76) in PM 1.1 , in spring (0.60), summer (0.70), and autumn (0.64) in PM 1.1-2.1 , and in all seasons (0.54-0.87) in PM 2.1-10 .The ratio of OC/EC was 7.0, 6.3, 7.6, and 6.9 in spring, summer, autumn, and winter, respectively, which demonstrated the abundance of secondary organic aerosols in Nanjing.The sources of carbonaceous aerosols exhibited significant seasonal variations, which were dominated by vehicle exhaust, and coal and biomass burning in PM 2.1 , and by dust, and coal and biomass burning in PM 2.1-10 .

Figure 1 .
Figure 1.Location of sampling site in Nanjing.

Figure 1 .
Figure 1.Location of sampling site in Nanjing.

Figure 3 .
Figure 3. Daily backward air-mass trajectories reaching the Nanjing site ending at 500 m a.s.l. and 12:00 LT in four seasons (a: spring, b: summer, c: autumn and d: winter).

Figure 3 .
Figure 3. Daily backward air-mass trajectories reaching the Nanjing site ending at 500 m a.s.l. and 12:00 LT in four seasons (a: spring, b: summer, c: autumn and d: winter).

Figure 3 .
Figure 3. Daily backward air-mass trajectories reaching the Nanjing site ending at 500 m a.s.l. and 12:00 LT in four seasons (a: spring, b: summer, c: autumn and d: winter).

Figure 4 .
Figure 4. Spectral distributions of OC, EC, SOC, and POC mass concentration in four seasons (a: OC, b: EC, c: SOC and d: POC).
ratio observed.We calculated the SOC concentration in nine size segments in each of the four seasons based on the values of

Figure 4 .
Figure 4. Spectral distributions of OC, EC, SOC, and POC mass concentration in four seasons (a: OC, b: EC, c: SOC and d: POC).

Figure 5 .
Figure 5.The relationships between OC and EC in each of the four seasons at Nanjing (a: total, b: PM 1.1 , c: PM 1.1-2.1 and d: PM 2.1-10 ).

Figure 6 .
Figure 6.Spectral distributions of OC/EC in four seasons.

Table 1 .
Observation period and major meteorological elements in Nanjing.

Table 1 .
Observation period and major meteorological elements in Nanjing.

Table 3 .
The minimum value of OC/EC concentrations in each of the four seasons.