Qinkai Li1,2, Zhou Yang3, Xiaodong Li 1,2, Shiyuan Ding1,2, Feng Du4

Institute of Surface-Earth System Science, Tianjin University, Tianjin 300072, China
Tianjin Key Laboratory of Earth Critical Zone Science and Sustainable Development in Bohai Rim, Tianjin University, Tianjin 300072, China
College of Tourism and Geographical Sciences, Tongren University, Tongren 565300, China
State Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of Technology, Chengdu 610059, China

Received: July 29, 2019
Revised: November 9, 2019
Accepted: November 10, 2019
Download Citation: ||https://doi.org/10.4209/aaqr.2019.07.0368  

  • Download: PDF

Cite this article:

Li, Q., Yang, Z., Li, X., Ding, S. and Du, F. (2019). Seasonal Characteristics of Sulfate and Nitrate in Size-segregated Particles in Ammonia-poor and -rich Atmospheres in Chengdu, Southwest China. Aerosol Air Qual. Res. 19: 2697-2706. https://doi.org/10.4209/aaqr.2019.07.0368


  • An NH3-poor atmosphere was observed at Chengdu during 2012–2013.
  • NO3 linearly related with NH4+/SO42– in both NH3-poor and -rich atmosphere.
  • NO3 formed more efficiently than SO42– during cold and NH3-rich atmosphere in winter.


In order to determine the seasonal characteristics of water-soluble inorganic ions (WSIIs) in aerosols in urban atmospheres, size-segregated particulate matter (PM) samples were collected over a one-year period from February 2012 to January 2013 in a typical urban location, Chengdu in Southwest China, using an Andersen cascade impactor sampler. The PM mass concentrations, particularly the fine fraction, peaked during winter, and the WSIIs were more enriched in the fine fraction (21.7%) than the coarse fraction (9.2%). The sums of the equivalent ratios of cations (Na+, NH4+, K+, Mg2+, and Ca2+) to anions (SO42–, NO3, Cl, and F) indicated that the fine particles (0.86) were more acidic than the coarse ones (1.60). The average NH4+/SO42– molar ratio (A/S) in the fine fraction (1.79) was much higher during winter than the other three seasons (< 1.5), implying a generally NH3-poor atmosphere in Chengdu; hence, the NO3 in the fine particles was principally formed through homogeneous reactions involving ammonia and nitric acid during winter, whereas it was heterogeneously formed during the other three seasons. Significant positive correlations were observed between the A/S and NO3 molar concentrations during spring and winter; therefore, the formation of particle-phase NO3 may be accelerated by increased A/S in both NH3-poor and -rich atmospheres. Moreover, the A/S and NO3/SO42– molar ratios displayed negative and positive correlations during spring and winter, respectively, suggesting that the variation in atmospheric NH4+ (or NH3) during winter affected the formation of NO3 more strongly than that of SO42–, whereas more SO42– than NO3 was formed in the NH3-poor atmosphere during spring, when most of the NO3 in the aerosols would be expected to form via heterogeneous reactions.

Keywords: Size-segregated particles; Sulfate; Nitrate; Ammonium; Chengdu.


Atmospheric particulate matter (PM) is one of the most important components that could affect the air quality, climate change and public health (IPCC, 2013; WHO, 2013). Over last few decades, researches have been carried out in many studies worldwide, focusing on the sources and secondary formation and transformation mechanisms of PM and its effects on human health and the earth radiation balance (Seinfeld and Pandis, 1998; Dockery et al., 1993; Huang et al., 2014; Wang et al., 2016; Yao et al., 2018). Atmospheric PM can be derived from both anthropogenic and natural sources, and can also be formed in the atmosphere from secondary processes (oxidation of gaseous species), and the particle size ranges from a few nanometers to several tens of micrometers (Zhang et al., 2015; Wang et al., 2017a) that plays a critical role in their environmental and human health effects (Kulmala et al., 2004; West et al., 2016; Tian et al., 2018; Klimont et al., 2017).

The water-soluble inorganic ions (WSIIs) including anions (SO42–, NO3, Cl and F) and cations (Ca2+, Mg2+, Na+, K+ and NH4+), often account for a major fraction of atmospheric PM mass, play a vital role in scattering incoming solar radiation and altering cloud properties and thus enhance the indirect radiative forcing (Seinfeld and Pandis, 1998; Zhao et al., 2011). Recent studies have indicated that the WSII fraction in PM varies not only with time and space but also with the particle size, which could provide key information for interpreting the sources and formation mechanisms of the PM (Zhang et al., 2008; Huang et al., 2016; Wu et al., 2017). To a large extent, WSIIs determines the particle acidity and hygroscopicity, which could control the rates of heterogeneous chemical reactions (He et al., 2012; Hennigan et al., 2015; Wang et al., 2016; Weber et al., 2016; Tian et al., 2018). In general, the SO42–, NO3 and NH4+ (together called SNA) are dominant WSIIs in fine-mode PM and exist primarily in the forms of (NH4)2SO4, NH4HSO4 and NH4NO3 and expected to be formed through the neutralization reactions between acids (i.e., H2SO4 and HNO3) and NH3 in gas phase (Yao et al., 2002; Lin et al., 2006; Lin and Cheng, 2007; Wu et al., 2018). As NH3 could neutralize the H2SO4 to (NH4)2SO4 or NH4HSO4 (Wang et al., 2005; Vieira-Filho et al., 2016), the NH4+/SO42– (A/S) molar ratio of 1.5 can be considered as an indicator in determining the formation pathways of NO3 in the atmosphere (Pathak et al., 2004, 2009). In “excess NH4+” (i.e., A/S > 1.5, NH3-rich) environment, NO3 is supposed to be formed homogeneously whereas in “deficient NH4+” (i.e., A/S < 1.5, NH3-poor) environment, it tends to be formed via heterogeneous reactions, the NH3 is insufficient to neutralize the HNO3 (He et al., 2012; Behera et al., 2013; Tian et al., 2018). However, the current available studies on A/S values focused mainly on fine-mode particles, since the NH4+ could be of less importance on coarse particles (Wu et al., 2017; Jiang et al., 2019). Recently, atmospheric NH3 has become a public concern for its potential in facilitating the formation of secondary aerosol, but it is yet to be regulated in many regions around the world. Therefore, to better understand the formation of SO42– and NO3 in both fine and coarse PM, long-term observations of WSIIs in PM is necessary, despite the recent progress, including their seasonal characteristics, which could help in identifying the PM sources and elucidating the transformation mechanisms of heavy haze formation over megacities in China (Wang et al., 2017b; Zou et al., 2018).

Chengdu, the capital city of Sichuan Province, is located at the western part of the Sichuan Basin. Due to the blocked terrain and relatively stagnant weather in Sichuan Basin, the pollutants emitted from local sources are retained for longer time compared to the other areas of China, and frequently results heavy air pollution (haze) events at Chengdu (Ning et al., 2018). Besides, the increase in energy consumption due to rapid industrialization and urbanization over the past few decades should be significantly contributing to the severe urban haze in Chengdu. However, the studies on characterization of PM and the sources of haze in Chengdu are limited to short-term observations (e.g., the dust and biomass burning) and model simulations (Chen and Xie, 2013; Tao et al., 2013; Tian et al., 2013; Wang et al., 2013b; Tao et al., 2014; Chen et al., 2015; Shi et al., 2017). In this study, we focused on identification of PM sources through the measurements of major WSIIs in size-segregated particles collected over a one-year period in Chengdu, Southwest China. We discuss the seasonal characteristics and formation mechanisms of aerosol SO42– and NO3 in the NH3-poor and -rich atmosphere at Chengdu.


Study Location

Chengdu is situated in Sichuan Basin with the Qinghai-Tibet Plateau on the west, Qinling Mountains and the Loess Plateau on the north, Hunan and Hubei Province on the east, and the Yunnan-Guizhou Plateau on the south. The climate of Sichuan Basin is characterized by subtropical monsoon and temperate oceanic climate (Chen and Xie, 2014; Li et al., 2015). The weather is categorized into four distinct seasons with the temperature ranging from 5°C in winter to 26°C in summer. Average annual precipitation is 918 mm with high abundance and frequency in summer to autumn (i.e., from June to November) period than in winter to spring (i.e., from December to May).

As one of the most crowded regions in Sichuan Basin, Chengdu has a permanent population of more than 14 million, and the population density is about 964 people per square kilometer. Like many other megacities in China, Chengdu is growing at an unprecedented pace with an urbanization rate of 60.2% in 2012, and it had the eighth largest gross domestic production in China. According to the statistical yearbook of Chengdu for 2013 (Chengdu Bureau of Statistics), the total vehicle number exceeded 3.0 million by the end of 2012, which is the second highest number among all Chinese cities. The industrial areas are mostly located in the northern suburbs of city, about 20 km away from the sampling site. The total energy consumption of Chengdu increased from 24 million tons of standard coal equivalent to 44.4 million tons of standard coal equivalent between 2005 and 2012, according the statistics from the World Resources Institute (http://www.wri.org).


Size-segregated atmospheric particle sampling was performed using an eight-stage Andersen cascade impactor air sampler (AN-200; SIBATA, Japan) in Chengdu, Southwest China. The sampler was placed on the rooftop of a building (15 m above ground level) in Chengdu University of Technology (CDUT) campus (30°40ʹ40²N, 104°08ʹ30²E), located at Chenghua district in the urban northeastern part of Chengdu. The cut points of the sampler were set as 11, 7.0, 4.7, 3.3, 2.1, 1.1, 0.65 and 0.43 µm, and the samples were collected from 15 February 2012 to 21 January 2013 at flow rate of 28.3 L min–1 with a frequency of 4–8 days for each sample, in order to ensure adequate particle load for reliable chemical analysis. Before sampling, the quartz fiber filters (Ø80 mm, Tissuquartz 2500QAT-UP; Pallflex) were preheated in a muffle furnace at 450°C for 6 h, and wrapped in aluminum foil individually after sampling, then sealed in airtight polyethylene bags and stored at –18°C until analysis. In this study, a total of 108 filter samples were obtained, and field blanks were sampled without sucking air for 10 min and treated the same way as particle samples.

Chemical and Statistical Analyses

Filters were gravimetrically weighed before and after the sampling, after 24 h equilibration at constant temperature (20–23°C) and relative humidity (35–45%), using a Sartorius MC5 electronic microbalance with a sensitivity of ±1 µg (Sartorius; Göttingen, Germany). The weight difference between before and after sampling was considered as particle weight and the mass concentration was calculated by dividing it with the sample air volume.

Half of the sample filters was cut into pieces, soaked in 15 mL deionized water (18.2 MΩ; Milli-Q, USA) and ultrasonically extracted three times (i.e., 10 min each). After the extraction, the solutions were then filtered through syringe filters (0.22 µm; Millipore) and stored in high-density polyethylene bottles for subsequent WSII analysis. Anions (F, Cl, NO3 and SO42–) and cations (Na+, K+, Ca2+ and Mg2+) were measured using ion chromatography (ICS-90; Dionex, USA) and inductively coupled plasma optical emission spectrometry (ICP-OES; MPX, USA), respectively, and the NH4+ was measured with a SKALAR continuous flow analyzer (San++ Automated Wet Chemistry Analyzer; Holland). The instrument detection limits are 0.02 mg L–1 for anions and 0.005 mg L–1 for cations and NH4+ (Xiao and Liu, 2004; Yang et al., 2015). Standard reference materials produced by the National Institute of Metrology in China were used for quality assurance, and the measurement uncertainties for different ion concentrations in this study were typically less than 5% based on replicate analysis of standards. No targeted ions were detected in the field blank.

To compare the results obtained in size-segregated particles among the four seasons, Student’s t-test was performed by assuming equal variance in two populations at a 95% confidence level with SPSS Statistics (version 19.0; IBM Corp., Armonk, NY, USA). Correlation analysis was also performed among concentration data and molar data across the study period. Statistically significant difference was set at p values less than 0.05 unless otherwise stated.


Aerosol Particle Mass Concentrations

Since the Anderson impactor sampler does not have a 2.5 µm cut point, the diameter of 2.1 µm is defined as the boundary for fine and coarse particles in this study (Fig. 1). The annual mean concentrations of PM2.1 and PM11 are 125.9 ± 56.1 µg m–3 and 224.5 ± 83.6 µg m–3, respectively. They were significantly higher than the annual PM2.5 and PM10 limits of both Chinese National Ambient Air Quality Standards (35 and 70 µg m–3; Grade II of GB3095-2012) and the WHO air quality guidelines (10 and 20 µg m–3; WHO, 2006). Nevertheless, these values are comparable with other studies synchronously conducted at Chengdu (120.4 and 195.5 µg m–3), but higher than those reported at Wuhan (89.6 and 134.9 µg m–3), Shanghai (103.1 and 149.2 µg m–3), and Guangzhou (97.5 and 144.4 µg m–3) and lower than those reported at Xi’an (140.9 and 257.8 µg m–3) over the parallel periods (Wang et al., 2006; Shi et al., 2017; Wang et al., 2013a; Wang et al., 2015; Wang et al., 2017c), indicating that the PM loading is high at Chengdu.

Fig. 1. Time series of PM2.1 and PM2.1-11 mass concentrations and PM2.1/PM11 mass ratios.Fig. 1. Time series of PM2.1 and PM2.1-11 mass concentrations and PM2.1/PM11 mass ratios.

The PM11 mass concentration peaked in March (378.4 µg m–3) followed by February (371.4 µg m–3). However, the PM2.1/PM11 ratio was significantly higher in February (0.70) than that in March (0.54) (Fig. 1), despite the fact that the PM11 mass concentration was comparable, implying that the sources and/or formation pathways of the PM were different in these two months (Tao et al., 2013; Wang et al., 2013b; Chen et al., 2015). An increase trend of the PM2.1/PM11 ratio was observed during the campaign from spring to winter, except for two anomalous high values in May and July, which might be caused by the crop residue burning around Chengdu (Tao et al., 2013; Chen and Xie, 2014). Seasonally, the PM2.1/PM11 ratio in winter (0.65 ± 0.04) is significantly higher than that in the other three seasons, which might be caused by the enhanced emission from anthropogenic activities (e.g., bulk coal and biomass burning for civil heating) and poor air diffusion condition due to low temperature and wind speed, which promote the fine particles’ formation and accumulation (Table 1) (Ning et al., 2018).

Table 1. Meteorological records during each sampling month.

Size Distribution of WSIIs

Total WSIIs of PM2.1 ranged from 10.5 to 68.9 µg m–3, with an annual average of 27.6 ± 19.8 µg m–3, and accounted for 20.7% of the PM2.1 mass, whereas the annual average of WSIIs in PM2.1-11 was 9.6 ± 7.6 µg m–3 (range: 4.4–28.6 µg m–3) and accounted for about 9.2% of PM2.1-11, which indicates that the WSIIs are highly enriched in fine particles. Furthermore, the total WSIIs in PM2.1 was the highest in winter (40.5 ± 24.8 µg m–3), followed by spring (39.6 ± 22.9 µg m–3), summer (17.4 ± 7.5 µg m–3) and autumn (12.7 ± 0.9 µg m–3), while the WSIIs in PM2.1-11 was higher in spring (16.9 ± 10.2 µg m–3), followed by winter (10.0 ± 9.4 µg m–3), summer (6.3 ± 1.8 µg m–3) and autumn (5.1 ± 0.4 µg m–3). It is noteworthy that the total WSIIs in both fine and coarse particles reduced sharply in summer and autumn at Chengdu that can be attributed largely to high rainfall during monsoon seasons (Xiao and Liu, 2004; Pan et al., 2017). The content of WSIIs in fine fraction in winter is comparable to that in spring, whereas in coarse fraction, it was significantly higher in spring than that in winter, suggesting their enhanced contribution from biomass burning and/or dust storms in spring (Tao et al., 2013; Wang et al., 2013b; Chen et al., 2015; Tian et al., 2016; Shi et al., 2017). The abundances of WSIIs in PM2.1 were: SO42– > NO3 > NH4+ > Cl > Mg2+ > K+ > Ca2+ > Na+ > F, while in PM2.1-11 they were: SO42– > Ca2+ > NO3 > Cl > NH4+ > Mg2+ > K+ > Na+ > F (Table 2), indicating that SO42– and NO3 are dominant species in PM at Chengdu. The discrepancies in the order of major ions’ abundances in fine and coarse fractions might have been driven by the differences in their formation pathways (Li and Shao, 2009; Liu et al., 2015; Li et al., 2018). Generally, SO42–, NO3, and NH4+ are formed by chemical reactions between the gas-phase precursors such as NOx, SO2, and NH3 in the atmosphere, however, the relatively high abundances of SO42–, Ca2+, and NO3 and low abundance of NH4+ in the coarse fraction indicate that they were mostly formed through heterogeneous reactions of the gas-phase precursors on pre-existing particles and/or derived from soil dust (Lin and Cheng, 2007; Tao et al., 2013; Chen et al., 2015; Shi et al., 2017).

Table 2. The mean and S.D. of WSIIs’ mass concentrations in PM2.1 and PM2.1-11.

Size distributions of major ions in each season are plotted as dC/dlogDp as a function of Dp (diameter of particles) and showed in Fig. 2. SO42– showed a unimodal distribution with a peak at 0.65–1.1 µm. SO42– can be directly emitted into the atmosphere from primary sources like burning processes and can also be produced in the atmosphere by homogeneous oxidation reactions of reduced sulfur species (e.g., SO2 or DMS) with OH radical, H2O2 and ozone (Seinfeld and Pandis, 1998; Guo et al., 2010). On the other hand, it can also be formed by heterogeneous reactions including the aqueous-phase reactions (e.g., cloud processes) (Guo et al., 2010; Huang et al., 2018). In spring, in addition to a major peak at 0.65–1.1 µm, SO42– showed two minor peaks at 2.1–4.7 and 7.0–11 µm, which indicate that it should have been significantly produced by both homogeneous and heterogeneous reactions in Chengdu (Zhuang et al., 1999). Interestingly, NO3 showed unimodal distribution in warm (i.e., spring and summer) and bimodal distribution in cold (autumn and winter) periods (Fig. 2(b)). Such differences imply that NO3 sources and/or formation processes might be different during warm and cold periods in Chengdu. NO3 can be formed more favorably through gas-to-particle conversion at lower temperature, and it might also under the influence of sand storms occurred during dry seasons. In addition, contrary to SO42–, NO3 showed a larger peak in winter than that in spring. In fact, ambient temperature and relative humidity are quite different during spring and winter, which might have greatly impacted their transformation mechanisms (Huang et al., 2018). Aerosol NH4+ is generally formed from the alkaline gas (i.e., NH3) by reacting with acidic species (e.g., H2SO4, HNO3 and HCl). It was largely enriched in fine-mode particles and peaked at 0.65–1.1 µm, consistent with SO42– and NO3 (Fig. 2(c)), suggesting that NH4NO3, (NH4)2SO4 and/or NH4HSO4 are the main forms of these ions in Chengdu PM (Wang et al., 2013b; Huang et al., 2016). The seasonal distributions of NH4+ could mainly be affected by the meteorological conditions. For example, ambient temperature can impact both the release of NH3 and NH4+ phase portioning, and wet deposition of NH4+ should be significant in summer and autumn due to high rainfall, whereas in winter, the conversion and accumulation of NH4+ might be exacerbated due to low temperature and wind speed and frequent fog events (Pan et al., 2017; Ning et al., 2018). 

Fig. 2. The distributions of major WSIIs over four seasons, which showed dC/dlogDp value as a function of Dp (diameter of particles). The included WSIIs are: (a) SO42–, (b) NO3–, (c) NH4+, (d) K+, (e) Ca2+, and (f) F–.Fig
. 2. The distributions of major WSIIs over four seasons, which showed dC/dlogDp value as a function of Dp (diameter of particles). The included WSIIs are: (a) SO42–, (b) NO3, (c) NH4+, (d) K+, (e) Ca2+, and (f) F.

K+ showed bimodal distribution with a large peak in fine particles and exhibited the highest abundance in spring (Fig. 2(d)), probably due to enhanced biomass burning emissions. Generally, the particle K+ is regarded as a good tracer for biomass burning emission into the atmosphere. Chen and Xie (2014) reported high concentrations of fine-mode K+ during biomass burning episode occurred in spring at urban Chengdu. While during other seasons, it can be inferred that the fumes of biomass burning by domestic use (e.g., daily cooking) at rural areas are responsible for the fine-mode K+. Ca2+ showed a bimodal distribution with the peaks in coarse-mode particles (3.3–4.7 and 7.0–11.0 µm) in each and every season (Fig. 2(e)), and showed extremely high loading in spring, which can be attributed for local fugitive dust due to the strong winds (Table 1) as well as long-range transported dust sands from arid or semi-arid areas of North China (Tao et al., 2013). Resuspension of soil dust, marine aerosols, and material from volcanic eruptions are the major natural pathways of particle F entering the atmosphere, while aluminum smelting, coal combustion and brick manufacturing are its main anthropogenic sources (Carpenter, 1969; Kalinić et al., 1997). A recent study also showed that biomass burning could be one of the major sources of fine particle F (Jayarathne et al., 2014). Interestingly, the size distribution of F was similar to that of Ca2+, i.e., peaked at 3.3–4.7 and 7.0–11.0 µm in spring (Fig. 2(f)). Such size distribution indicates that they were probably derived from the common source, such as the local fugitive cement dust, and the gaseous hydrogen fluoride emitted from combustion sources might heterogeneously react with coarse-mode Ca2+. However, it is difficult to provide a definite conclusion from this study due to limited data, a subject of future research.

Particle Acidity in Urban Chengdu

Particle acidity is an important parameter that could influence the concentration, chemical composition, and toxicity of PM (Weber et al., 2016). Nevertheless, direct measurement of particle acidity is difficult because of its low water content and nanoscale particle size, thus proxy methods and parameters, such as the in situ acidity, strong acidity and ion-balanced acidity are used for its estimation (Zhang et al., 2008; He et al., 2012; Kerminen et al., 2001; Tian et al., 2018). Recently, Tian et al. (2018) reported that charge equivalent ratio between measured cations to anions (RCE/AE) is a good measure of acidity in fine-mode aerosol, because the temporal trend between RCE/AE and in situ pH was observed to be similar. He et al. (2012) also used the equivalent charge ratio and showed similar seasonal variation in in situ PM2.5 acidity and RCE/AE, and the field observations have showed a general higher particle acidity in South China than North China. Moreover, an investigation conducted in four major cities of China by calculating the equivalent ratio of NH4+ to the sum of SO42– plus NO3 during summertime indicated that the particle acidity was mostly dependent on the environmental SNA abundance (Pathak et al., 2009).

In this study, we estimated the particle acidity by calculating RCE/AE in both sums of fine- and coarse-mode particles, and the results indicated that the fine particles are more acidic (RCE/AE < 1) than coarse particles (RCE/AE > 1) in Chengdu (Fig. 3). Since SO42– and NO3 are dominant ions in PM2.1, it is likely that the ambient NH3 was insufficient to completely neutralize the H2SO4 and HNO3 acids, despite the significantly high loading of NH4+. While in coarse particles, the SO42–, Ca2+ and NO3 accounted for 76% of the total WSIIs, and the Ca2+ was excessive to integrate with the SO42– and NO3 (He et al., 2012; Chen et al., 2015; Huang et al., 2018). However, two anomalous high RCE/AE values (i.e., > 1) of fine particles were observed in May and July of 2012, when the PM2.1/PM11 ratios were high (Fig. 1), which can be ascribed to the influence from biomass burning episodes that contribute the large amount of cations (e.g., NH4+, K+, and Na+). The decrease trend of RCE/AE for fine particles from spring to winter was likely associated with the more intensive emission and conversion of anthropogenic acidic species (e.g., SO2 and NOx) (Kerminen et al., 2001; Tian et al., 2018).

Fig. 3. The RCE/AE and [NO3–]/[SO42–] mass ratios in (a and d) different months and (b and e) seasons, and (c and f) size ranges, respectively (red for coarse-mode particles; black for fine-mode particles).Fig. 3. The RCE/AE and [NO3]/[SO42–] mass ratios in (a and d) different months and (b and e) seasons, and (c and f) size ranges, respectively (red for coarse-mode particles; black for fine-mode particles).

Implications for SO42– and NO3 Pollution

The A/S ratios in PM were found to be 1.79, 1.33, 1.19 and 1.14 in winter, spring, summer and autumn, respectively, at Chengdu. According to the threshold value of A/S (1.5) suggested by Pathak et al. (2004, 2009), it is clear that the “NH3-rich” and “NH3-poor” atmospheres prevailed in winter and all other three seasons, respectively. Hence, the NO3 in fine-mode particles at Chengdu should have been formed through heterogeneous reactions during spring to autumn. Whereas in winter, it might have been generated by homogeneous gas-phase reactions of HNO3 with NH3, because the rate may be much higher for the formation of NH4NO3 rather than the neutralization rate of H2SO4 by ambient NH3 (Huang et al., 2018; Tian et al., 2018). Additionally, the size distribution of NO3 in Fig. 2(b) may be partially interpreted by the seasonal A/S values obtained, such as the minor coarse-mode peaks could be associated with Ca(NO3)2 or Mg(NO3)2, which were heterogeneously formed on pre-existed soil particles (Zhuang et al., 1999; Li et al., 2018).

We also calculated the [NO3]/[SO42–] mass ratios in both fine and coarse fractions in each season. Their averages were generally less than 1 (Figs. 3(d)–3(e)), indicating higher contribution from stationary sources emissions than that from mobile sources at Chengdu (Yao et al., 2002). Average coarse-mode [NO3]/[SO42–] in warm seasons (i.e., 0.84 and 0.54 during spring and summer, respectively) were significantly higher than that of fine mode (0.43 and 0.27, respectively), but it was the opposite in cold seasons (0.59 and 0.68 in fine mode and 0.58 and 0.43 in coarse mode in autumn and winter, respectively). Based on hourly measurements of acidic gases, ammonia and secondary inorganic aerosols in a tropical urban atmosphere, Behera et al. (2013) reported that the conversion of SO2 into SO42– and HNO3 into NO3 was sensitive to the changes in temperature and relative humidity, respectively. However, due to the existence of thermodynamic equilibrium between precursor gases and fine particle NH4+ salts (e.g., NH4NO3), the conversion of HNO3 into NO3 may also be sensitive to the changes in temperature, but the Ca(NO3)2 or Mg(NO3)2 in coarse mode were supposed to be less sensitive to the variations in the temperature (Zhuang et al., 1999; Zhang et al., 2008; He et al., 2012). Moreover, the formation of aerosol NO3 could also be influenced by the ambient NH3 content (Pathak et al., 2004, 2009; Huang et al., 2018). Therefore, we infer that the size and seasonal distributions of aerosol [NO3]/[SO42–] at Chengdu were driven by the conversion of gaseous precursors and meteorological conditions (Zhuang et al., 1999; Guo et al., 2010; Huang et al., 2016). Besides, the annual averages of fine-mode [NO3]/[SO42–] were lower than that of coarse-mode particles (Fig. 3(f)), indicating the high abundance of coarse-mode NO3 and fine-mode SO42 resulted due to the prevailing NH3-poor atmosphere at Chengdu (Pathak et al., 2009).

The scatter plots of the molar concentration of NO3 and NO3/SO42– (N/S) molar ratios as a function of A/S ratios in each season are illustrated in Fig. 4. Correlation between A/S and NO3 found to be positive and significant in spring (p < 0.01) and winter (p < 0.01). It implies that the formation of NO3 is increased with the increase of A/S. However, the molar concentrations of NO3 were comparable in each and every season, suggesting that NH3 might have played a critical role in altering the heterogeneous and homogeneous formation process of aerosol NO3, depending on NH3-poor and -rich atmosphere, respectively (Pathak et al., 2009; Huang et al., 2018). This result is consistent with the results obtained by Weber et al. (2016) that the aerosol NH4NO3 concentrations increase with decreasing SO42– in the steady-state NH3 atmosphere of southeastern United States. Whereas the linear relations between A/S and N/S molar ratio were positive and negative in winter (0.01 < p < 0.05) and spring (p < 0.001), respectively (Fig. 4(b)). These relations together with the significant positive correlations between A/S and NO3 molar concentration in winter and spring (Fig. 4(a)) imply that the formation of NO3 was more influenced than that of SO42– formation with the increase in A/S during winter, when the formation of (NH4)2SO4 come to its limited regimes relative to the NH3 content, yet the HNO3(g) become available for NH3 to form the aerosol NO3 predominantly (Huang et al., 2018; Tian et al., 2018). In addition, the previous studies also indicated that the low temperature is favorable for the formation of NO3 but not for the SO42– (Guo et al., 2010; Huang et al., 2016;). Thus, the formation of NO3 could be more efficient than SO42– during cold and NH3-rich atmosphere in winter. 

Fig. 4. Scatter plots of (a) NO3– molar concentration and (b) NO3–/SO42– molar ratios as a function of NH4+/SO42– molar ratios over the four sampling seasons (n = 24 for each season).Fig. 4. Scatter plots of (a) NO3 molar concentration and (b) NO3/SO42– molar ratios as a function of NH4+/SO42– molar ratios over the four sampling seasons (n = 24 for each season).

In contrast, aerosol SO42– might have more efficiently formed from homogeneous gas-phase reactions of H2SO4 and NH3 than aerosol NO3 with the increase in A/S in spring in NH3-poor and dry atmosphere (Table 1). It is likely because the kinetic rate of neutralization of H2SO4 by NH3 could be significantly higher than forming aerosol NO3 in NH3-poor atmosphere (Huang et al., 2018), whereas the NO3 formation could occur from the hydrolysis of N2O5 under high relative humidity (Pathak et al., 2009). Furthermore, the low relative humidity in spring might have promoted the heterogeneous reaction of HNO3 on soil dust particles, one of the formation pathways of aerosol NO3, at Chengdu (Pathak et al., 2011). According to the lower A/S (< 1.5) and higher RCE/AE (> 1) values, the nonvolatile cations from minerals could have played a vital role in forming the particles in spring at Chengdu, and this can also be evidenced by the PM2.1/PM11 ratio shows in Fig. 1. In Fig. 3(e), as the coarse-mode N/S is significantly higher than fine-mode particles in spring, the NO3 were supposed to be formed on the minerals, through heterogeneous reactions, and mainly in the forms of Mg(NO3)2 or Ca(NO3)2. Otherwise, only several minor peaks for SO42– are shown in coarse mode, while the fine-mode SO42– is likely to be formed more efficiently than NO3 in spring with the increase in A/S (Fig. 4). On the other hand, the correlations are insignificant in summer and winter, and this might principally be ascribed to the high rainfall, which greatly impacts the WSII loading in the atmosphere (Pan et al., 2017).


The characteristics of WSIIs in size-segregated particles in urban Chengdu were investigated over a one-year period. The results showed that the WSIIs were more enriched in the fine particles than the coarse ones, and NH4+, SO42–, and NO3, the dominant WSIIs in the fine fraction, were mostly generated by the conversion of anthropogenic gaseous precursors, especially during the cold seasons. The calculated equivalent ratios (RCE/AE) indicated that the fine particles were more acidic (< 1) than the coarse particles (> 1). Furthermore, the NH4+/SO42– (A/S) molar ratio in the fine-mode PM suggested that the atmosphere was NH3-poor during spring, summer, and autumn but NH3-rich during winter; therefore, the NO3 in the fine fraction was formed homogeneously during winter and heterogeneously during the other three seasons. Based on the positive correlations between the A/S and NO3 molar concentrations observed during spring and winter, the increase in NH3 or NH4+ may enhance the formation of NO3 in aerosols, but certain meteorological conditions, such as high temperatures and concentrated rainfall, significantly disrupt this relationship. Finally, we infer from the seasonal correlations between the A/S and NO3/SO42– molar ratios in addition to the temporal variation in the NO3/SO42– molar ratio that the NH3-rich atmosphere during winter favors the formation of NO3 rather than SO42–, whereas the NH3-poor atmosphere during spring favors the formation of the latter, as the heterogeneous formation of the former during this season is less efficient.


This work was supported by the National Natural Science Foundation of China (Grant No. 41173022 and 41773006) and Natural Science Foundation of Tianjin (Grant No. 17JCZDJC39400). We also appreciate Ph.D. Jing Yu and Long Lv for helping field particulate matter sampling.

Share this article with your colleagues 


Subscribe to our Newsletter 

Aerosol and Air Quality Research has published over 2,000 peer-reviewed articles. Enter your email address to receive latest updates and research articles to your inbox every second week.

77st percentile
Powered by
   SCImago Journal & Country Rank

2022 Impact Factor: 4.0
5-Year Impact Factor: 3.4

Aerosol and Air Quality Research partners with Publons

CLOCKSS system has permission to ingest, preserve, and serve this Archival Unit
CLOCKSS system has permission to ingest, preserve, and serve this Archival Unit

Aerosol and Air Quality Research (AAQR) is an independently-run non-profit journal that promotes submissions of high-quality research and strives to be one of the leading aerosol and air quality open-access journals in the world. We use cookies on this website to personalize content to improve your user experience and analyze our traffic. By using this site you agree to its use of cookies.