Effects of Early Pollution Control Measures on Secondary Species of PM2.5 in Jiaozuo, China.

Various measures for reducing air pollution have been promulgated since 2013 in China. To investigate the synergistic results of emission control and meteorological environment, PM2.5 samples collected from October 2013 to July 2016 and November 2018 to October 2019 in Jiaozuo city were analyzed for their compositions, secondary species (Ss) variations, and factors changing for Ss formation. The results showed that the concentrations of sulfate, nitrate, ammonium, and secondary organic aerosols (SOAs) generally decreased over the same seasonal period during these years. In addition, the concentrations and proportions of each Ss increased with the increase in the PM2.5 level in these years, implying that although PM2.5 levels have been reduced by various control policies, Ss formation would remain the major contributor to PM elevations. The enhanced effects of gas-phase reactions on intensification of sulfate, SOA, and PM were observed in 2018-2019, which was consistent with the elevation of nitrate and SOA at PM levels of >150 μg/m3. Only sulfate in all PM levels sharply decreased after 2015, showing the fine effect of coal-related pollution control and the importance of collaborative control of NO x , volatile organic compounds, and organic aerosol emissions with SO2 emissions in the future.

Introduction

Outdoor fine particles have been of high concern in China since 2013 due to their heavy contamination status[1,2] and the serious impacts on environmental quality,[3,4] physical health,[5−7] and climate changes.[4,8,9] In the meantime, the government promulgated various measures to reduce pollution and improve the air quality, such as the “Clean Air Action Plan (2013–2017)”. Previous studies have found that the concentrations and compositions of particles are based on their sources,[10,11] especially human activities. Thus, it is significant to understand the variations of PM2.5 components during the implementation of policies to evaluate the improvement of air quality and the development of subsequent mitigation strategies.

An extensive literature indicating that secondary species (Ss), that is, secondary inorganic aerosol (SIA) of sulfates, nitrates, and ammonium and secondary organic aerosol (SOA), are the major chemical constituents in PM2.5.[12,13] The haze formation is usually accompanied by a Ss increase in particles,[12−14] such as the January 2013 winter haze, driven by sulfate enhancement caused by heterogeneous oxidation.[15] Primary emissions,[16,17] levels of PM gas precursors,[18] meteorological conditions, regional transport,[19−21] and the efficiency of secondary formation[13,22] synergistically act on the formation of particles.[11,23−25] Therefore, to demonstrate the synergistic results of source emission control and meteorological factors, the PM2.5 composition variations and temporal patterns of Ss were analyzed based on the PM2.5 samples collected from 2013 to 2016 and 2018 to 2019 in Jiaozuo.

Jiaozuo is located in the northwestern Henan province, one of the regions seriously polluted with particles in China[2] (http://106.37.208.233:20035/). Its terrain inclines from the northwest to southeast, south of the Taihang mountains, which is the intersection of the south side and the east side of the Taihang Mountains, and is not conducive to pollution dispersion and the circulation of air with frequent stable meteorological conditions.[8,26] Thermoelectricity and cement are major industries, and coal energy is the main energy consumed in Jiaozuo. It showed 16.96 million tons of raw coal consumption in 2016, increased by 11.5% compared to that in 2014, based on the statistical yearbook of Jiaozuo (http://www.jztjj.gov.cn/jzww/tjsj/tjnj/A080304index_1.htm). However, the Jiaozuo government issued various measures to improve air quality from 2013 to 2016. The main contents, promulgation time, and the phased implementation effects of the clean air measures issued by Jiaozuo government are listed in Tables S1 and S2. According to the statistics of the Jiaozuo Ecological Environment Bureau, the air quality has been improved in Jiaozuo, with the annual mean concentrations of PM2.5 decreasing from 86 in 2015 to 66 μg/m3 in 2017. In addition to local emissions, the particulate level at this site is also comprehensively affected by the north–south and east–west airflows, which can reflect the overall level of PM decline in the region. By analyzing the implementation of the pollution control policy and the process of PM quality improvement, more effective control measures can be provided for the future.

Results

Concentrations of PM2.5 and Its Species

The average concentration of PM2.5 was 101.5 ± 49.8 μg/m3 during the study period, with great variations from 28.0 to 272.7 μg/m3, which exceeded the Chinese daily grade II guideline of 75 μg/m3 on 54.8% of the selected days. As the predominant species of PM2.5, sulfate (SO42–), nitrate (NO3–), ammonium (NH4+), organic carbon (OC), elemental carbon (EC), and crustal elemental oxides (CEO) accounted for 17.9, 15.9, 10.5, 14.3, 5.1, and 6.1%, respectively. According to the PM2.5 reconstruction in Figure S2b, SIAs, SOAs, primary organic aerosols (POAs), EC, and CEO accounted for 44.3, 14.0, 8.9, 5.1, and 6.1%, respectively. Ss contributed 58.3% to PM2.5 on the whole. The reconstructed and measured PM2.5 concentration showed good correlation, with a slope of 1.25 (Pearson’s r 0.99, p < 0.05).

Temporal Variations of Ss

The seasonal concentrations of conventional pollutants (NO2, SO2, and O3), PM2.5, and its major species during the sampling period are shown in Figures , S3, and S4. The concentrations of PM2.5 presented clearly seasonal and inter-annual variations. The highest seasonal mass concentrations of PM2.5 appeared in winter, followed by autumn, spring, and summer. The average winter concentrations of PM2.5 were 1.0–1.3, 1.4–3.0, and 1.5–2.7 times those of other seasons in 2013–2014 (autumn 2013 to summer 2014), 2015–2016 (autumn to summer), and 2018–2019 (winter to autumn), respectively. Meanwhile, it showed a decreasing trend for the seasonal concentrations of PM2.5 in the same period. The PM2.5 concentrations decreased by 27.9, 53.0, and 52.1% in spring, summer, and autumn from 2014 to 2019, respectively, and by 31.7% in winter from 2015 to 2018. For autumn, there was no decreasing trend of seasonal PM2.5 concentrations due to the lack of data from 2016 to 2018. However, the mean concentrations in autumn were significantly lower in 2019 than that in 2015.

Figure 1

Seasonal variations of (a) concentrations and (b) proportions for major components in PM2.5 [including SO42–, NO3–, NH4+, SOA, POA, EC, CEO, and trace elements (TE)] during the sampling period. Relevant data for winter 2016 and winter 2017 are from ref (27).

The seasonal variations of concentrations of its Ss were generally consistent with that of PM2.5. It generally showed low in summer and high in winter in these years. The proportion of overall Ss was ranked in the order winter, autumn, summer, and spring. Previous studies have found that stable conditions, usually with the average wind speed (WS) <1.5 m/s, occur frequently (∼60%) in North and Central China,[8,26] especially in winter. Stable conditions are unfavorable to pollution dispersion and could largely be responsible for the local buildup of air pollutants. Moreover, there is additional emission from residential coal heating in winter, which is also attributed to the high pollution of Ss.

The concentrations of Ss generally decreased over the same seasonal period during these years. It showed a significant dropping trend in spring, summer, and autumn from 2014 to 2019 and in winter from 2015 to 2018, with a reduction of 48.8, 59.6, 44.9, and 44.8% for SO42–, 47.7, 73.8, 70.6, and 31.3% for NO3–, 28.5, 56.6, 54.8, and 34.0% for NH4+, 55.8, 55.7, 57.4, and 37.7% for SOA, respectively. The effective control and reduction of SO2 (Figure S3b) may cause the sulfate in PM2.5 to decrease significantly and the heterogeneous reactions against SO2 and NO in the atmosphere to reduce. Although the possession of civil vehicles increased from 299,042 in the end of 2013[28] to 555,911 in the end of 2018,[29] the strict implementation of traffic restrictions made no significant increase in NO2 in Jiaozuo. Two reasons made the significant descent of NO3– and SO42– in synchronization. Ammonia can react with SO2 and NO2 in the atmosphere to form ammonium bisulfate, ammonium sulfate, and ammonium nitrate. The equivalent charge between SO42– and NH4+, NO3– and NH4+, and SO42– + NO3– and NH4+ (Figure S5b) was all significantly correlated at 0.01 levels (sig. p value = 0.000 < 0.05, t-test), indicating that ammonium salts probably existed as the combination of NH4+ with NO3– and SO42–. The equivalent charge ratios of NH4+ to SO42– + NO3– in 2015–2016 and 2018–2019 were higher than that in earlier years, which was consistent with the decrease of sulfate in these two years and highlighted the dominance of ammonium sulfate in comparison with ammonium bisulfate. Thus, although there were steady emissions from ammonia sources and few specific measures to control ammonia emissions, as concentrations of SO42– and NO3– decreased in PM2.5, so did that of NH4+.

Discussion

Comparisons of Control Effect with Other Cities

Compared with the same season in other years, high levels of the concentrations of PM2.5 and its components in autumn and winter were all observed in 2015. It indicated that growth dominated the concentrations of PM2.5 before 2015. After that, each seasonal mean concentration of PM2.5 and Ss began to show a downward trend. On carrying out the control measures related to coal combustion and emissions, such as controlling total coal consumption, desulfurization upgrade of coal-fired boilers, and shutdown of small coal-fired boilers (Table S2), the concentrations of SO2 in each season showed a downward trend from winter 2013 (Figure S3b). The annual variations of the seasonal concentrations of NO2 and O3 were different from that of SO2. The mean O3 concentration of each season in the same period decreased first (from autumn 2013 to summer 2015) and then increased. Each seasonal concentration of NO2 showed a slight decrease since 2015. Overall, the conditions of SO2 and PM2.5 in an outdoor environment in Jiaozuo have been getting better with the implementation of various policies and regulations on air pollution control, while that of NO2 and O3 have not.

According to the Henan energy balance sheet (actual quantity),[30−33] the residential consumption of coal in the Henan province increased by 19.2% from 2013 to 2015, and the residential consumption of raw coal accounted for 21.0% of the total final consumption in 2015. In particular, coal consumption in rural areas, which rarely have pollution control measures, accounts for more than 70.0% of residential consumption until the end of 2015. Previous studies have found that the particles emitted by household coal combustion are mainly carbonaceous.[34] Therefore, there were still higher concentrations of SOA and POA in PM2.5. Furthermore, the volume of SO2, NO, and soot emission from urban life all increased from 2013 to 2015.[28,35] For Jiaozuo, less than 20% of city residents realized centralized heating before 2014 (https://www.henan.gov.cn/2014/08-12/533515.html), and the policy of replacing residential coal with gas and electricity was also in its infancy until winter 2015 (Tables S1 and S2). Decentralized coal combustion remained widespread. Additionally, by the end of winter 2015, measures resulting in the reduction of NO2 emissions, such as denitrification upgrading in power and cement industry and elimination of the yellow-label and old vehicles, were still in the process and had not been completed. While the number of civil vehicles sharply increased from the year of 2013[28] to 2015[35] (increase by 21.7%), the concentrations of NO2 in autumn and wintertime increased by 28.0 and 23.9% from 2013 to 2015, respectively. These reasons caused high levels of Ss in PM2.5 before 2016. Moreover, NO2 can catalyze the transformation of SO2 and promote the formation of sulfate, nitrite, and ammonium.[14] The stable meteorological conditions in winter[8,26] could significantly intensify secondary formations of precursors.

Distinct to the buildup of PM2.5 components in winter 2015, a significant decrease has been observed since 2016. With further implementation of specific control measures involving vehicles and coal-related industries (Tables S1 and S2), the precursors and primary particles showed a remarkable decrease (Figures S3 and S4), which leads to the apparent decline of PM2.5 and Ss since 2016.

Considering that the joint prevention and control measures were adopted in Beijing–Tianjin–Hebei and surrounding areas before 2017, the study selected provincial cities surrounding Jiaozuo, including Beijing, Tianjin, Shijiazhuang, Taiyuan, Jinan, and Zhengzhou, to compare the temporal variations of air pollutants (PM2.5, NO2, SO2, and O3) and the control results between Jiaozuo and these cities (Figure S6). The control measures implemented in these cities are similar to those in Jiaozuo, which differ only slightly in timing and the degree of target requirements. It showed a similar temporal variation on these pollutants in the whole region. Although the concentrations of PM2.5 and SO2 varied from city to city in each season, the inter-annual variation was similar among these cities in each season, showing a decreasing trend. Also, the decreasing trend of SO2 was very significant in all cities. Except for Jinan, no obvious inter-annual variation of NO2 concentration was observed, indicating that NO2 emission reduction measures have not achieved significant results in other cities. With the implementation of the policy, the seasonal O3 concentration in all cities showed an increasing trend.

The concentration changes of Ss in PM are not exactly the same in surrounding cities after control measure implementation, but the general decreasing trend is similar for sulfate, such as a decrease of sulfate concentrations by 42.4% from 2013 to 2017 in Handan,[36] by 11.0% from winter 2013 to winter 2016 in Beijing,[16] and by 41.6% from winter 2016 to winter 2017 in cities along the Taihang Mountains.[27] The variation trends of nitrate, ammonium, and SOA in PM2.5 were not consistent in each city, and they may increase in some cities.[16,36] However, the concentrations of these three Ss decreased to different degrees in most surrounding cities,[27] which was consistent with the variations in Jiaozuo.

Species Contributions at Different Particle Levels

The concentrations and proportions of major components at different PM2.5 levels in Jiaozuo were reconstructed and analyzed (Figure ). It was divided into four levels based on the PM2.5 concentration:[24,37] <75, 75–150, 150–250, and >250 μg/m3. It can be clearly seen that the concentrations of all major species show an increasing trend with the increase in PM2.5 concentration. SO42–, NO3–, SOA, and NH4+ were the most obviously increasing components with the increase of PM2.5 pollution level, followed by POA, EC, CEO, and TE. The concentrations of SIA were 21.6, 49.9, 95.2, and 131.4 μg/m3 at the level of PM2.5 <75, 75–150, 150–250, and >250 μg/m3, which was 3.0, 3.2, 3.8, and 3.2 times that of SOA, respectively.

Figure 2

(a) Average concentrations and (b) fractions of major species (SO42–, NO3–, NH4+, SOA, POA, EC, CEO, and TE) at different PM2.5 levels. The value of error bars in (b) was 1SD for species fraction.

The trend of species proportion in PM2.5, however, was not exactly consistent with its concentration variations as PM level increased. The proportion of POA, EC, CEO, and TE in PM2.5 decreased with increasing pollution, while that of SO42–, NO3–, and SOA increased. Although the proportion of NH4+ fluctuated, it was higher at the level of PM2.5 > 150 μg/m3 than that at PM2.5 < 150 μg/m3. The increment of SO42– proportion was in agreement with the enhanced sulfate formation during China’s severe winter haze episode in January 2013.[15] The OC/EC was 2.5, 2.8, 3.5, and 4.3 from low to high PM levels, respectively [when OC/EC > 2, there could exist secondary organic carbon (SOC)[38,39]], which is consistent with the SOA reconstructions at different PM2.5 levels (Figure ). Overall, the mass fractions of primary species in PM2.5 decreased from 25.1 to 12.6% when the PM2.5 concentrations increased from <75 to >250 μg/m3 level, while the fractions of Ss increased from 52.4 to 66.2%. Additionally, the proportions of SIA in PM2.5 were 2.1, 2.6, 3.3, and 5.2 times that of SOA under the four levels, respectively. It could be ascertained that secondary matter was the major contributor to PM2.5 elevations, and the secondary inorganic species dominated the atmospheric PM2.5 in heavy polluted weather.

The variation trend of concentrations and proportions of each species with the increase of PM2.5 levels were also applicable in 2013–2014, 2014–2015, 2015–2016 (autumn to summer), and 2018–2019, while the concentrations and proportions of particle species at the same PM2.5 level were different between these years (Figure ). The concentrations of PM2.5, primary species, and Ss decreased gradually during 2013–2014 to 2018–2019 at PM2.5 levels of <75 and 75–150 μg/m3. However, these concentrations were first increased and then decreased for Ss and slowly increased for primary species at PM2.5 > 150 μg/m3 level during 2013–2014 to 2018–2019. Meanwhile, the proportions of Ss decreased basically during 2013–2014 to 2018–2019 at each PM2.5 level, although they were the highest in 2015–2016 at PM2.5 > 150 μg/m3 level. The proportions of Ss were 53.3, 60.1, and 60.6%, from low to high PM2.5 levels in 2013–2014, 51.2, 56.2, and 66.5% in 2015–2016, and 51.8, 49.5, and 58.5% in 2018–2019, respectively. There was a high proportion of Ss at each PM2.5 level, especially at the high level, which indicated that the secondary formation was the major contributor to PM2.5 elevations for a long time and would remain so.

Figure 3

(a) Concentrations and (b) mass fractions of major species in 2013–2014, 2014–2015, 2015–2016, and 2018–2019 at different PM2.5 levels. The data of winter 2014 (2014–2015) were missing.

Similarly, the concentrations and proportions of SO42–, NO3–, NH4+, and SOA varied considerably at the same PM2.5 level compared those years. At PM levels of <75 μg/m3, it showed ranges of 8.3–11.7 μg/m3 for SO42–, 5.3–8.0 μg/m3 for NO3–, 4.8–6.6 μg/m3 for NH4+, and 6.1–8.9 μg/m3 for SOA, respectively. Although the average concentration of each Ss fluctuated slightly, their concentrations decreased year by year. At PM levels of 75–150 μg/m3, the concentrations of SO42– decreased the most from 24.1 μg/m3 in 2013–2014 to 14.0 μg/m3 in 2018–2019, followed by SOA, NH4+, and NO3–. It was decreased by 3.5 μg/m3 for SOA, 2.2 μg/m3 for NH4+, and 1.6 μg/m3 for NO3–. Obvious decreasing of sulfate concentrations in particles at the same PM level indicated the remarkable effect of controlling measures related to coal.

At PM levels of >150 μg/m3, the concentrations of each Ss were the highest due to the high PM2.5 concentrations in 2015–2016. Hence, their proportions were mainly compared during these years. Compared to that in 2013–2014, the proportions of SO42– in PM2.5 sharply decreased by 8.3%, and its concentrations also decreased by 11.4 μg/m3 in 2018–2019. NH4+ was slightly decreased by 0.8% in proportion (0.1 μg/m3) from 2013–2014 to 2018–2019. The variations of SO42– and NH4+ were consistent with that at PM levels of 75–150 μg/m3. However, SOA and NO3– were observed to be increasing by 6.1% (11.4 μg/m3) and by 0.7% (0.8 μg/m3) from 2013–2014 to 2018–2019, respectively. Several reasons for this are as follows. First, the substantial reduction of atmospheric SO2 caused a decrease of SO42– formation. Moreover, when the proportion of SO42– in PM2.5 is lower than 20%, the inhibitory effect of sulfate on the formation of nitrate[37] may be weakened, which further promotes the formation of nitrate in turn. Second, there was still abundant NO2 to form NO3– in 2018–2019 compared with that in 2013–2014 (Figure S3d). The increase in the number of vehicles led to the increase in NO2 emissions in recent years. The ratio of NO3–/SO42– is usually used to reflect the relative contribution of mobile and stationary sources to particles.[40,41] In this study, it was 0.8 and 0.6 in 2013–2014, 1.0 and 0.8 in 2015–2016, and 1.2 and 1.0 in 2018–2019 at PM levels of 75–150 and >150 μg/m3, respectively. It suggested that the stationary sources were the major contributors to particles of higher concentrations at the early stage of sampling. While the NO3–/SO42– ratios generally increased in 2015–2016 and 2018–2019 at higher PM levels, indicating that the relative contributions of mobile sources to form PM2.5 increased. Meanwhile, because of the cleaner national six standards for vehicles that have not been implemented before 2019, the increase of vehicles would inevitably lead to the increase of primary carbon components and organic gaseous precursor emissions. However, only at PM levels of >150 μg/m3, the concentrations and proportions of SOA increased from 2013–2014 to 2018–2019. Also, PM concentrations more than 150 μg/m3 were all observed in winter. Therefore, this accumulation of SOA was the result of the large volume of precursors produced by vehicle exhaust and coal burning and then superimposed with meteorological effects. It also indicated that precursors accelerate the formation of secondary organics in a statically stable environment. On the whole, in addition to controlling emissions from coal-fired sources, it is also necessary to collaboratively control the emission of vehicle exhausts to improve air quality.

Factors Changing Sulfate and Nitrate Formation

Sulfate and nitrate can be either directly emitted from primary emissions,[11,25] or generated by gas-phase reactions, aqueous reactions in cloud or fog droplets, and heterogeneous processes associated with aerosols.[42] For example, Li et al. found 33% of sulfate was from primary emissions in winter 2014 in Xi’an.[25] Sulfur oxidation ratio (SOR) and nitrogen oxidation ratio (NOR) are usually used to characterize the oxidized degree of SO2 and NO, respectively, and to identify the formation processes of sulfate and nitrate. SOR is the molar ratio of [SO42–] to [SO42– + SO2], and NOR is the molar ratio of [NO3–] to [NO3– + NO2]. Sulfate mainly comes from primary emissions when the SOR < 0.1, while it mainly comes from secondary transformations when the SOR > 0.1.[43] The larger the SOR value is, the greater the photochemical reaction takes place, and the more obvious secondary transformations are. Similar to the SOR, the larger the NOR value is, the more obvious NO conversions are.

From PM2.5 levels of <75 to >150 μg/m3, the SOR values were 0.15, 0.20, and 0.22 in 2013–2014, 0.14, 0.21, and 0.27 in 2015–2016, and 0.39, 0.35, and 0.40 in 2018–2019, respectively. The NOR values were 0.12, 0.24, and 0.25 in 2013–2014, 0.11, 0.17, and 0.31 in 2015–2016, and 0.13, 0.22, and 0.27 in 2018–2019 from low to high PM levels, respectively. It can be seen that the sulfate and nitrate in PM2.5 at each PM level in those years was mainly derived from the secondary particles produced by the photochemical reaction or aqueous/heterogeneous processes. Except for SOR in 2018–2019, the value of SOR and NOR at high PM levels were both higher than those at low levels in those years, indicating the higher oxidation rate of SO2 and NO2 at high PM levels.

The chemical process producing sulfate and nitrate was based on meteorological conditions and reaction medium. There are several pathways for the oxidation of SO2 and NO2.[42] SO2 can be oxidized into sulfate via the oxidation by OH radical in the gas phase, and by ozone, hydrogen peroxide, organic peroxides, oxygen, and nitrogen dioxide in bulk aqueous phase or under high-relative humidity (RH) conditions. Homogeneous gas-phase oxidation of NO2 by OH and O3 and the heterogeneous reactions of N2O5 are the two important pathways for nitrate formation. To reveal the changes of factors affecting the formation of sulfate and nitrate, SOR and NOR were paired with T, RH, and O3 concentrations in Figures and S7.

Figure 4

SOR, NOR, and SOC/EC were paired with the concentrations of RH and O3 during the sampling period. The data of winter 2014 were missing. (a) RH vs SOR, (b) RH vs NOR, (c) RH vs SOC/EC, (d) O3 vs SOR, (e) O3 vs NOR, and (f) O3 vs SOC/EC.

SOR and NOR showed weak correlation with temperature (T) (Figure S7), indicating that temperature had little influence on SOR and NOR; similar results were also found in Handan by Zhao et al.[36] and in Beijing by Zhang et al.[13] Moreover, there were no obvious correlations between SOR and O3, and it only showed an observable trend of decrease at O3 < 25 μg/m3 for NOR (Figure ), while SOR and NOR both performed an obvious increase in these years with the increase in RH. It has been found that sulfate particles visibly grow when RH is about ∼60–70%.[14] In this study, SOR and NOR were elevated from 0.18 ± 0.09 and 0.19 ± 0.08 at RH < 60% to 0.33 ± 0.04 and 0.37 ± 0.08 at RH > 60% in 2013–2014, from 0.16 ± 0.09 and 0.14 ± 0.09 at RH < 60% to 0.23 ± 0.09 and 0.23 ± 0.10 at RH > 60% in 2015–2016, and from 0.34 ± 0.09 and 0.16 ± 0.06 at RH < 60% to 0.44 ± 0.09 and 0.17 ± 0.06 at RH > 60% in 2018–2019, respectively. This result suggested that the increases of RH could promote the formation of nitrate and sulfate. The dissolution of NO2 and SO2 enhances with increasing RH, which is the foundation of the subsequent transformation in the aqueous/heterogeneous phase oxidation process. Previous studies[18,44,45] found that low RH (usually <40%) favors homogeneous gas-phase reactions, and high RH favors heterogeneous gas–particle reactions. It is worth noting that O3 concentration tended to decrease with the increase of RH (Figure S8). The environment of strong solar radiation and low RH favors photochemistry and suppresses heterogeneous reactions. As the RH increases and the O3 concentration decreases, the formation of sulfate and nitrate from heterogeneous processes gradually increases. It suggested that aqueous/heterogeneous chemical processes played an important role in the formation of SO42– and NO3–.

The results of the reaction process are also absolutely affected by the content of gas precursors. The influence of the atmospheric precursors on the chemical constituents in PM2.5 is notable,[14,46] especially the increase of SO2, which can not only increase SO42– in the particulate matter, but also promote the NO3– content to some extent. For example, because of the low concentrated gas supply rate and large-scale retail coal cooking and winter heating, the concentration of SO2 in Jiaozuo was significantly higher than that in Beijing (Figure S10). NO3– in PM2.5 in Jiaozuo was slightly higher than that in Beijing (+3.1 μg/m3) in winter 2016–2017,[16] even with the lower NO2 (74.6 μg/m3 in Jiaozuo and 81.7 μg/m3 in Beijing). The ratio of (NO3–/NO2)Jiaozuo/(NO3–/NO2)Beijing was 1.2, showing that NO3– was more easily formed in Jiaozuo due to the high SO2 concentration. Moreover, SOR and NOR were close before summer 2016 in Jiaozuo, with the values of 0.20 and 0.21 in 2013–2014 and 0.18 and 0.17 in 2015–2016. Additionally, there were comparable concentrations of SO2 and NO2 before summer 2016. With the different decreasing rates of SO2 and NO2, SOR increased to 0.37 and NOR decreased to 0.17 in 2018–2019, indicating that to some extent, the increase of the volume-mixing ratio of NO2 in the atmosphere could effectively promote the gas to particle conversion of SO2. It was consistent with the conclusion drawn in previous studies that a high concentration of NO2 promotes sulfate formation by heterogeneous and aqueous oxidation of SO2.[14,46−48] According to the conceptual model proposed by Li et al.,[24] the relative contributions were reduced for photochemical reactions and increased for heterogeneous processes with the increase in RH. In the process, it simultaneously showed ozone decrease, the positive feedback of particles and enhanced SIA formation. It was mentioned that SOR and NOR increased with the elevation of PM level in the previous section. O3 concentrations were found to be decreasing with the increase of particle (Figure S9) and SIA concentrations in 2018–2019. RH corresponding to the initial significant increase of sulfate (∼40–80%) was observed to be decreasing with the increase of volume-mixing ratio of NO2 and NH3 by Wang et al.[14] However, no matter at high or low RH level, SOR was greatly elevated and was higher than NOR in 2018–2019 compared with other years, whereas NOR showed a small change. Furthermore, the concentrations of sulfate and nitrate merely increased by 3.50 and 2.12 from 16.28 and 12.16 μg/m3 at RH <60% to 19.78 and 14.28 μg/m3 at RH >60% in 2018–2019, respectively, and the increments were markedly smaller than that in other years. Therefore, the efficient formation of sulfate in 2018–2019 may depend not only on the relative enhancement of heterogeneous reactions but also partly on the intensification of the gas-phase oxidation of SO2.

Incremental Contributions of Organic Matter to Particles

The OC/EC ratio can be used to characterize the emission sources and the conversion characteristics of carbon aerosols, depending on both the proximity of the emissions and the relative weight of road traffic and biomass burning. The daily OC/EC ratios were in the range of 1.4–4.2 with an average of 2.5, 2.4, and 3.5 in 2013–2014, 1.3–4.8 with an average of 2.3, 2.7, and 3.6 in 2015–2016, 1.3–5.6 with an average of 2.8, 3.5, and 4.8 in 2018–2019 for PM levels ranging from low to high, respectively. Previous studies have found that the OC/EC ratio of road traffic emissions varies between 1.4 and 5 for gasoline catalyst vehicles,[49−53] from 0.3 to 1 for diesel vehicles,[49,51−54] from 2.5 to 10.5 for coal smoke,[55] from 4.3 to 7.7 for kitchen emissions,[56] and larger values (3–70) for biomass burning emissions.[57−60] It indicated that despite the difference in the OC/EC ratio between these years, the carbon components in these years were still mainly derived from fossil fuel combustion and vehicle emission.

SOC/EC can represent the chemical reaction rate, which is similar to OC/EC. The highest seasonal SOC/EC ratio generally appeared in wintertime and the lowest in summertime, which indicates the favorable conditions of winter to form SOC and was in agreement with the seasonal change of OC proportion. The seasonal variations of EC, primary organic carbon (POC), and SOC were shown in Figures and S11. The concentration of carbonaceous species in autumn (19.7 μg/m3) and winter (26.4 μg/m3) was obviously higher than that in spring (19.3 μg/m3) and summer (12.7 μg/m3), the same as that of SOC. A large number of small coal-fired boilers and dispersed domestic coal burning still existed before 2018. It led to increased emissions of volatile or semi-volatile organic compounds (VOCs) and organic gases. The special climatic conditions in winter, such as low temperature and low mixing layers, are not conducive to the diffusion of carbonaceous gas and pollutants, allowing them to remain in the atmosphere for long periods and to fully produce photochemical reactions. Therefore, it resulted in the content of SOC in winter increasing significantly, reaching 1.5 times that in autumn and 2.9 times that in summer.

The SOC/EC ratio increased with PM levels in each year, and it also increased year by year at the same PM level. It changed from 1.5, 1.5, and 2.0 in 2013–2014 to 1.7, 2.0, and 2.8 in 2018–2019 at the low to high PM levels in order, indicating that SOA was easily formed under severe pollution, and the favorable conditions to enhance SOA formation was continuously intensified. It is commonly believed that SOA can be formed by gas–particle partitioning of organic matter (OM) and heterogeneous reactions of oxygenated organics.[42] Gas-phase reactions are dominated by absorptive partition of low-volatility and semi-volatile oxidation products associated with VOC emissions.[61] Particle-phase reactions mainly include hydration reactions, acid-catalyzed reactions, and reactions with basic species.[42] The meteorological conditions under different SOC/EC in these years were compared in Figures and S7. SOC/EC did not correlate well with RH and temperature in these years, indicating that the changes of temperature and RH may have little difference on the formation of SOA. However, there was a relatively increasing trend and decreasing trend for SOC/EC with the increase of RH and temperature, respectively. It showed that increasing temperature promoted the volatilization of SOA from the particle phase to gas phase, and the hydroscopic property of particles also plays a role in the formation of SOA. The ratio of SOC/EC showed a decreasing trend at O3 < 60 μg/m3 and an increasing trend at O3 > 60 μg/m3 with the increase of O3 concentrations. There was no obvious difference between these years, while O3 in 2019 showed an increasing trend compared with other years. Low temperature, high O3 concentrations, and high RH tended to enhance secondary OM in 2018–2019, that is, gas phase reactions seemed to be dominant in formation of SOA after 2018.

Conclusions

Various measures have been taken in many areas of China to prevent and control haze since 2013. To evaluate the initial effects on air quality after measure implementation, the variations of outdoor PM2.5 and its components from October 2013 to July 2016 and from November 2018 to October 2019 in Jiaozuo city were analyzed. The concentrations of PM2.5 and Ss generally decreased over the same seasonal period during these years. The total proportions of Ss decreased during 2013–2014 to 2018–2019 at each PM2.5 level, but there were still high proportions and increases with the increase of PM levels, indicating that the secondary formation was the major contributor to PM2.5 elevations. Moreover, due to similar measures, the inter-annual variations of seasonal concentrations of PM2.5, SO2, NO2, and O3 in Jiaozuo were consistent with that in surrounding cities, showing a decreasing trend for PM2.5 and SO2, an increasing trend for O3, and fluctuations for NO2.

Sulfate and ammonium in the three PM levels have decreased by different degrees after 2015, but nitrate and SOA have increased significantly at high PM levels of >150 μg/m3. Nitrate and SOA were observed as having an increasing effect on particle enhancement, showing that the collaborative emission controls of NO, VOCs, and organic aerosols should be strengthened in Jiaozuo. The enhanced effects of gas-phase reactions on the intensification of sulfate, SOA, and PM were also observed in 2018–2019, which was consistent with the elevation of nitrate and SOA. In addition, the significant decrease of sulfate after 2015 was mainly attributed to the sharp reduction of SO2, indicating that compared with other pollution control measures, coal-related control measures played an important role in reducing PM2.5 first.

Methods

Sample Collection

The sampling site was located on the rooftop of the school of resources and environment of Henan Polytechnic University in Jiaozuo, Henan province (Figure ). The inlet was approximately 15 m above ground level, 1.5 km northeast of the high-tech district government site, a state-controlled air quality monitoring station. There were no obvious obstructions and pollution sources within 5 km of and around the sampling site. So, we assume that the composition of particles at this site represents the general situation of particulate matter in general in Jiaozuo. Sample collection and weighing process strictly referred to “technical specifications for gravimetric measurement methods for PM2.5 in ambient air” (HJ 656-2013). The 24 h integrated PM2.5 samples (8:00–7:30, UTC+8) were collected by prebaked (450 °C, 4 h) quartz membranes (MK 360, Munktell, Sweden, 8 × 10 in.) with a TFIA-2 high-volume air sampler (Staplex, USA) at flow rates of 1.05 m3/min. A total of 289 valid samples of PM2.5 in different seasons were monitored from October 2013 to July 2016 and from the end of November 2018 to mid-October 2019. The hourly average concentrations of PM10, NO2, SO2, CO, and O3 were based on the data from the high-tech district government site (http://222.143.24.250:100/flex/index.html). In addition, the hourly meteorological parameters such as temperature (T), WS, wind direction, RH, precipitation, and pressure were also recorded concurrently.

Figure 5

Location of the sampling site. On the left is the map of China, showing the position of Jiaozuo in the Henan province (red-filled). On the right is the location map showing the sampling site (red dot) and the high-tech district government site (blue dot) in Jiaozuo.

Composition Determination

The concentrations of main species in selected PM2.5 samples, that is, major inorganic elements, TEs, water soluble inorganic ions (WSII), OC, and EC were carried out. Inorganic elements, including Na, Mg, Al, Si, S, Cl, K, Ca, and Fe, were determined by X-ray fluorescence spectrometry (ARL Quant’X EDXRF, Thermo Scientific Inc., USA).[62] The minimum detection limits (MDL) were within the range of 0.13–0.85 μg/cm2 for these elements. Mercury (Hg) was measured by direct mercury analyzer from Milestone (DMA-80, Milestone, Italy).[63] The instrument’s detection limit was 0.001 ng. Other TEs of Ni, Br, Cr, As, Ba, Co, Cu, Pb, Sb, Se, V, Zn, Mn, and Cd were analyzed by inductively coupled plasma/mass spectrometry (820-MS, Varian, USA), according to the determination of metals in ambient particulate matter (HJ 657-2013). The MDLs were within the range of 0.02–8 μg/L for these TEs. Samples were digested before the test. The digestion process of PM2.5 on the filters using microwaves was also performed according to HJ 657-2013. Water-soluble cations and anions were measured, including Ca2+, Mg2+, K+, Na+, NH4+, SO42–, NO3–, Cl–, and F–. The detailed operation for measuring WSII is as follows. Part of the sample was cut and placed in 10 mL of ultra-pure deionized water (>18.2 MΩ cm) for ultrasonic extraction for 30 min. Then, the extract was filtered through a lure syringe filter (0.45 μm). The ions were determined by ion chromatography (ICS-3000, Dionex Co., USA) after the filtration solution was fixed to 15 mL. MDLs were within the range of 1–11 μg/L for these WSII. OC and EC were measured by an OC/EC analyzer (DRI model 2001A, Atmoslytic Inc., USA), with the thermal reflection method (TOR, IMPROVE A protocol). The MDLs were 0.01 μg/cm2 for EC and 0.39 μg/cm2 for OC.

Field blanks and parallel samples were tested for composition determination. Field blanks were collected every tenth sample. Parallel experiments (n = 3 set for each parallel test) were conducted with reagent blanks, filter blanks, and 10% of the test samples. The reagent blanks and filter blanks were determined to ensure the blank value did not interfere with the determination of samples. Analytical precision was calculated as the relative standard deviation (RSD) from duplicates. The measured results showed that the RSD results were less than 15% for all parallel tests. The RSD ranged within 0.7–4.8% for major element analysis by EDXRF, 0.3–3.3% for mercury analysis by DMA, 0.9–12.6% for TEs analysis by MS, and 0.5–1.3% for carbon analysis.

Mass Reconstruction

In order to assess the chemical composition in PM2.5 thoroughly, reconstructed mass (RM-PM2.5) was conducted on the basis of different PM2.5 pollution levels by considering SIA, SOA, POA, EC, CEO, and remaining TE.

SIA is the sum of SO42–, NO3–, and NH4+. OM was estimated as 1.6 times OC, which is based on the OM/OC ratios of 1.6 ± 0.2 for urban aerosols[64] and 1.64 ± 0.18 in PM2.5 for Northern Chinese cities[65] obtained in previous investigations. Furthermore, OC can be divided into POC and SOC according to its sources and formation pathway. So, POA and SOA were estimated through the following formula: POA = 1.6 × POC = 1.6 × ECmeas × (OC/EC)pri and SOA = 1.6 × SOC = 1.6 × (OCmeas – POC). POC can be estimated using the minimum OC/EC ratio[38,39,66] because (OC/EC)min is almost equivalent to the ratio of the local primary source affecting the measured concentrations when assuming only combustion sources of OC. CEO were calculated by the IMPROVE “soil” formula,[67] applying a factor of 1.16 for all oxide multipliers to account for unmeasured compounds, that is, CEO = 2.20 × Al + 2.49 × Si + 1.63 × Ca + 2.42 × Fe + 1.94 × Ti [assuming oxides of Al2O3, SiO2, CaO, Fe2O3, FeO (in equal amounts), K2O (assuming that soil K is 0.6 × Fe), and TiO2]. TEs were calculated as a sum of remaining elements (excluding S and geological elements).[68−70] Because TEs only account for a small fraction of PM2.5 mass, variations in the assumptions regarding metal oxides or multipliers did not contribute to large variations in RM-PM2.5. As a consequence, potential molecular forms were ignored for the remaining elements.