Secondary PM2.5 in Zhengzhou, China: Chemical Species Based on Three Years of Observations

The chemical properties and secondary components of PM2.5 were investigated in the city of Zhengzhou, China. Watersoluble ionic species (Na, NH4, K, Mg, Ca, F, Cl, NO3 and SO4) contents, carbonaceous components (organic carbon (OC) and elemental carbon (EC)) in PM2.5 were measured for three years. The EC tracer method was used to estimate the secondary organic carbon (SOC) content, and the Interagency Monitoring of Protected Visual Environments formula was used to estimate light extinction due to the chemical composition of PM2.5. The annual mean concentrations of PM2.5 were 186, 180 and 218 μg m in 2011, 2012 and 2013, respectively. These concentrations were 5–6 times greater than the National Ambient Air Quality Standards of China (annual value of 35 μg m) and indicated the presence of severe PM2.5 pollution in Zhengzhou. Particulate organic matter (OM) contributed the most (18–26%) to the annual average PM2.5, followed by SO4 (14–19%), NO3 (10–11%), NH4 (8–9%) and EC (3%). From 2011 to 2013, the contributions of OM and SO4 increased by 8% and 3%, respectively. The higher sulfur oxidation ratio indicated the formation of significant amounts of secondary inorganic aerosols (SIA), particularly during the summer and spring. Obvious SOC enrichment occurred during the winter and autumn. In addition, SIA and secondary organic aerosols accounted for 26–50% and 4–21% of the PM2.5 by mass, respectively. An investigation of the secondary species revealed that secondary aerosols played a dominant role in the total PM2.5 mass and the decrease in visibility. The secondary aerosols ((NH4)2SO4 + NH4NO3 + SOC) accounted for 80% of bext. The main secondary aerosols that led to poor visibility in Zhengzhou were (NH4)2SO4 and NH4NO3.


INTRODUCTION
Since 2011, the frequency of haze in urban areas throughout the country has increased.Consequently, haze has become a serious problem in China, with persistent haze pollution events in China occurring in 2013.Action plans for controlling particulate pollution and decreasing PM 2.5 (atmospheric particulate matter with an aerodynamic diameter of less than 2.5 µm) emission limits have been developed by the Chinese State Council (CSC, 2013).However, due to rapid industrialization and urbanization, China's current coal-dominated energy structure has not changed (Wang et al., 2006;Zhao, 2014).Moreover, the number of vehicles in China continues to increase, which is increasing the amount of fine particulate matter in the atmosphere and decreasing atmospheric visibility.Similar to other megacities, Zhengzhou, which is the capital of Henan Province in central China, experiences serious particulate pollution problems and poor visibility (DEPH, 2013;EMSC, 2013).The primary pollutant in Zhengzhou during the daytime is PM 2.5 , which is responsible for 77% of polluted days (DEPH, 2014).Thus, haze occurs frequently in Zhengzhou.
PM 2.5 adversely affects human health (Valavanidis et al., 2008;Anderson et al., 2012) and visibility (Xu et al., 2013).Furthermore, PM 2.5 contributes to global climate change by scattering light and increasing the amount of particles that can act as nuclei for cloud condensation (Chung and Seinfeld, 2002).PM 2.5 is a multi-component system that originates from both natural and anthropogenic sources (Alves et al., 2000), and mainly consists of primary aerosols that are directly emitted from sources and secondary aerosols that are formed in the atmosphere.Secondary aerosols are mainly generated through a series of chemical reactions and physical processes that involve nitrogen oxides (NO x ), sulfur dioxide (SO 2 ), ammonia (NH 3 ) and several volatile organic compounds (VOCs) that may react with ozone (O 3 ), hydroxyl radicals (•OH) and other reactive molecules to form sulfate (SO 4 2-), nitrate (NO 3 -), ammonium (NH 4 + ), secondary inorganic aerosols (SIA) and secondary organic carbon (SOC) (Liousse et al., 1996;Squizzato et al., 2013).
Aerosol measurements (Tao et al., 2009;Deng et al., 2013) have revealed that secondary chemical species strongly contribute to the overall concentration of fine particles in the atmosphere (approximately 50%).Therefore, secondary fractions are important for controlling PM 2.5 pollution.Previous studies (Cabada et al., 2002;Tao et al., 2009;Behera and Sharma, 2010) have shown that secondary organic aerosols (SOA) are an important component of PM 2.5 .The contributions of SOA to the total organic carbon (OC) concentration vary from nearly zero during the winter to 50% during the summer in Pittsburgh, Pennsylvania.In addition, SOC accounts for 29% of the OC concentration in PM 2.5 in Guangzhou, China, and the SOA contents of PM 2.5 during the winter and summer in Kanpur, India, are 18% and 12%, respectively.These aerosol species (SIA and SOA) are a significant cause of visibility degradation (Kim et al., 2006;Tao et al., 2009;Deng et al., 2013) and vary between different seasons and regions.In addition, the Interagency Monitoring of Protected Visual Environments (IMPROVE) formula has been used to estimate the contributions of the chemical compositions of PM 2.5 to light extinction, which helped determine the main factors of poor visibility (US EPA, 1999; Kim et al., 2006;Hand et al., 2011).
To better understand the secondary aerosol pollution conditions of PM 2.5 , which have not been studied in Zhengzhou, the contributions of secondary species to the overall mass of PM 2.5 and their impacts on visibility were investigated.The PM 2.5 pollution levels and chemical properties (in terms of water-soluble ionic species and carbonaceous components) were determined in this study based on three years of measurements.Although many earlier studies have used short-term data, long-term data can more accurately reflect local pollution characteristics.In addition, the air quality trends shown in this study are helpful for understanding the effectiveness of pollution control strategies.Furthermore, this analysis will provide useful information for regulatory agencies and for creating strategies to control PM 2.5 in the atmosphere.

Sampling and Mass Measurement
Zhengzhou, the capital of Henan Province, is located in central China and has a temperate continental monsoon climate.In Zhengzhou, the average temperature is approximately 14.4°C, and the average annual precipitation is approximately 640 mm.Approximately 70% of the annual precipitation in Zhengzhou occurs from June to September.Southerly winds prevail during the summer, and northerly winds prevail during the winter.Due to rapid economic development, Zhengzhou has expanded to an area of 373 square kilometers and has a total population of over 4 million (an annual natural growth of approximately 59 thousand people).
PM 2.5 sample collections were conducted using quartz microfiber filters (20.3 × 25.4 cm, PALL, USA) and a highvolume sampler (TE-6070D, Tisch Environmental, USA) from April 2011 to December 2013 at Zhengzhou University (Fig. 1).A TE-6070D sampler, which was equipped with a single-stage high-volume cascade impactor (TE-231, Tisch Environmental, USA), was used to remove particles larger than 10 µm.At least 15 samples were collected each season (winter, spring, summer and autumn) using a sampling duration of 22 h (9:00 AM-7:00 AM).The mass concentration of PM 2.5 was determined using the gravimetric method.The filters were conditioned at 20 ± 5°C before and after sampling at a relative humidity of 50 ± 5% for at least 48 h before weighing on a microbalance (Mettler Toledo XS205, Switzerland) with a precision of 0.01 mg.Next, the samples were stored in a freezer (-18°C) until analysis.

Analysis of Water-Soluble Ions
Ion concentrations (Na + , NH 4 + , K + , Mg 2+ , Ca 2+ , F -, Cl -, NO 3 and SO 4 2-) were determined using an ion chromatograph (ICS-90, Dionex, USA).A circular punch with a known cross-sectional area was used to randomly intercept 2-6 pieces of the filter (including the middle and edge) in the sample zone.Furthermore, deionized water or an organic solvent was used to clean the punch before use.Two pieces of the samples were subjected to extractions using 25 mL of ultra-pure Milli-Q water (specific resistance: 18.2 MΩ cm).The samples were ultrasonicated in a water bath at < 30°C for 30 min before filtering through 0.22 µm filters (Menbrana, German).The eluent used for detecting anions was composed of 8.0 mM Na 2 CO 3 and 1.0 mM NaHCO 3 , and a flow rate of 0.8 mL min -1 was used.The eluent used for detecting cations was 20 mM methane sulfonic acid (CH 3 SO 3 H), and a flow rate of 1.0 mL min -1 was used.All ions were identified based on their respective retention times.

Analysis of Carbon Species
A piece of filter was removed from the 20.3 × 25.4 cm quartz filter and used for OC and elemental carbon (EC) analyses, which were conducted using an OC/EC analyzer (Sunset Laboratory, USA) and the thermal/optical transmission (TOT) method (Chow et al., 2001).The detection limit was 0.2 µg m -3 , which was calculated as three times the standard deviation (SD) of the seven blank replicates.All analyses were conducted according to the NIOSH TOT protocol (low temperature in a helium atmosphere and high temperature in a 2% oxygen/98% helium atmosphere).A He-Ne laser (633 nm) was used to monitor sample transmission and correct for EC during OC pyrolysis at high temperatures.

PM 2.5 Mass Concentration
Overall, 66, 52 and 55 PM 2.5 samples were collected during 2011, 2012 and 2013, respectively, from the sampling site at Zhengzhou University, China.The PM 2.5 concentrations ranged from 60 to 548 µg m -3 , 55 to 565 µg m -3 and 56 to 698 µg m -3 in 2011, 2012 and 2013, respectively, with annual mean concentrations of 186, 180 and 218 µg m -3 .Compared with the National Ambient Air Quality Standards of the USA (annual value of 15 µg m -3 ) and the air quality guidelines of the World Health Organization (annual value of 10 µg m -3 ), our results clearly indicated severe PM 2.5 pollutions.In addition, the 3˗year results (Fig. 2) showed that at least 90% of the daily PM 2.5 concentrations exceeded the proposed standards in China (75 µg m -3 daily) that will be implemented on January 1, 2016.
The PM 2.5 compositions and concentrations are presented in Table 1.The annual averages of PM 2.5 and its components have steadily increased over time.PM 2.5 results from primary anthropogenic emissions and the secondary transformation of gas pollutants in the atmosphere.Direct emissions mainly result from combustion of fossil fuels (coal, gasoline and diesel), biomass (straw and wood), and waste.The main gas pollutants that are converted into PM 2.5 include SO 2 , NO x , NH 3 and VOCs.Other anthropogenic sources include road dust, construction dust, industrial dust, kitchen smoke, etc. Zhengzhou is currently undergoing rapid economic development and urbanization.According to the Henan Statistical Yearbook of 2011-2013, the total energy consumption, number of vehicles, and population of Zhengzhou from 2011 to 2013 grew dramatically, with the greatest increase occurring in 2013 (Table 2).Increases in both PM 2.5 and individual species are associated with increases in the number of vehicles and the total energy consumption in Zhengzhou.In addition, with the implementation of urban village reconstruction, construction dust has become a main source of PM 2.5 emissions.Strong seasonal variations in PM 2.5 were observed, with the highest concentrations occurring during the winter and the lowest concentrations occurring during the summer.The average PM 2.5 concentrations during the winters of 2011, 2012 and 2013 were 297 ± 160, 234 ± 12 and 337 ± 168 µg m -3 , respectively, with corresponding summer values of 120 ± 40,   2013 (n = 15) 337 ± 168 1.6 ± 0.6 31 ± 18 4.5 ± 1.9 0.5 ± 0.2 4.7 ± 1.9 1.5 ± 0.8 17.5 ± 10.3 39 ± 20 56 ± 39 54 ± 28 8 ± 3 Annual 2011 (n = 66) 186 ± 100 0.9 ± 0.9 15 ± 10 2.1 ± 2.0 0.3 ± 0.2 2.1 ± 2.1 0.5 ± 0.6 5.0 ± 5.9   84 ± 21 and 128 ± 47 µg m -3 . Each species (except SO 4   2-) exhibited consistent PM 2.5 trends, with the highest PM 2.5 levels occurring during the winter and the lowest levels occurring during the summer.The SO 4 2-concentrations in PM 2.5 were lowest in the autumn.
The results shown in Table 1 indicated that SO 4 2-, OC, NO 3 -, NH 4 + and EC were the main species of the analyzed components in PM 2.5 .The mass contributions of these chemical species to PM 2.5 were calculated for the annual samples (Fig. 3).The particulate organic matter (OM) shown in Fig. 3 was obtained by multiplying OC by 1.6 (Turpin and Lim, 2001).The other ions, including Na + , K + , Mg 2+ and F -, were individual species with relatively low concentrations.The non-apportioned portion was obtained by subtracting the OM, EC and nine water-soluble ions from the total mass of PM 2.5 .
As observed from the annual average (Fig. 3), OM contributed the most (18-26%) to PM 2.5 , followed by SO 4 2-(14-19%), NO 3 -(10-11%), NH 4 + (8-9%) and EC (3%).From 2011 to 2013, the contributions of each of these species increased, except for the contribution of EC which did not change.Among these species, the contributions of OM increased by 8% and the contributions of SO 4 2increased by 3%.The contributions of the non-apportioned species decreased from 40% (2011) to 25% (2013), which indicated that the proportions of OM, SO 4 2-, NO 3 -and NH 4 + in PM 2.5 increased.Regarding the non-apportioned portion, the results of our study were comparable with those observed in a previous study conducted in Beijing, in which approximately 35%-48% of PM 2.5 was non-apportioned (Hu et al., 2015).Several scientific studies have shown that quartz filters, which show a higher retention of organics and lower release of ammonium salts (Schaap et al., 2004;Wittmaack and Keck, 2004;Vecchi et al., 2009), tend to absorb more water vapor due to their hydrophilic nature and wettability (Zdziennicka et al., 2009).This water vapor would contribute to the non-apportioned mass.In addition, Hu (Hu et al., 2015) inferred that approximately 15%-32% of PM 2.5 in the quartz filters was from water vapor.
In terms of the annual trends, the mass concentration of the non-apportioned part and its contribution to the total PM mass decreased each year.According to our previous study (Geng et al., 2013), the contribution of dust to PM 2.5 was 26%.If the non-apportioned portion represents only one source, we believe that it is most likely resuspended dust (including soil dust, construction dust and road dust).However, we did not measure metal concentrations in this study, which is an area that warrants future study.
Accordingly, we attribute the non-apportioned portions in this study to metal oxides, water and undefined chemical species (refractory or insoluble).
Fig. 4 shows the percentages of the chemical species in the seasonally sampled PM 2.5 .Among the spring samples, the contributions of OM, EC, NH 4 + , Cl -, Ca 2+ and the other ions changed very little over the years studied (1% Overall, OM comprised a larger percentage of the PM 2.5 during the autumn and winter (22-30%).The maximum contributions of Cl -to PM 2.5 occurred during the winter and reached 5%.Coal combustion releases significant amounts of chlorine (McCulloch et al., 1999); thus, fine particles are enriched in OC and Cl -in areas that use coal-fire heat during the winter.
The maximum contributions of SO 4 2-(25-29%) and NH 4 + (11-13%) to PM 2.5 occurred during the summer, while the maximum contribution of NO 3 -occurred during the autumn (11-14%) and differed from the contributions of SO 4 2-and NH 4 + .This observation may be attributed to the volatility of ammonium nitrate, which is the main chemical form of nitrate and can evaporate when subjected to the relatively high temperatures that occur during the summer (Wang et al., 2005).The different seasonal trends observed in the three major secondary species resulted from the differences in their formation mechanisms and the seasonal variations of the meteorological conditions in Zhengzhou.The contributions of Ca 2+ (the crustal ion) to the PM 2.5 were 2-4% greater during the spring than during the other seasons because sandstorms frequently occurred during the spring.

Ionic Components and Carbonaceous Species Ionic Components
To describe the magnitude of the transformation of atmospheric SO 2 to SO 4 2-and NO 2 to NO 3 -, the sulfur oxidation ratio (SOR) and nitrogen oxidation ratio (NOR) were calculated (Wang et al., 2005;Gao et al., 2011).The SOR and NOR are defined as follows: where n-SO 4 2- , n-SO 2 , n-NO 3 -and n-NO 2 are the molar concentrations of SO 4 2-, SO 2 , NO 3 -and NO 2 , respectively.The concentrations of SO 2 and NO 2 were obtained from air quality monitoring stations operated by the Environmental Protection Bureau of Zhengzhou, which is located four kilometers from the sampling site.
The seasonal average conversion ratios for sulfate and nitrate were calculated for PM 2.5 in Zhengzhou (Table 3).The sulfate conversion fractions were greater during the summer (0.4 ± 0.2) and spring (0.4 ± 0.1) and were relatively constant (0.2 ± 0.1) during the autumn and winter.The mean values of NOR were similar during all four seasons (0.2 ± 0.1).Previous studies have reported SOR values of < 0.10 in the flue gas of oil combustion boilers (Kircher et al., 1977) and < 0.027 in vehicle exhaust (Pierson et al., 1979;Truex et al., 1980).Coal-fired boilers are commonly used in China.According to a study conducted by Wang (Wang et al., 2008), the SOR in the flue gas of coal-fired power plants is much less than 0.10.Therefore, 0.10 is regarded as a conservative (i.e., maximum) SOR value for primary pollutants.Thus, atmospheric SORs > 0.10 suggest the occurrence of photochemical oxidation of SO 2 .
As shown in Table 3, significant amounts of secondary SO 4 2-were generated during all four seasons, and conditions that are more favorable for formation occurred during the summer and spring.In this study, the seasonal variations of SOR (or NOR) are not obvious.
The equivalent ratio of ]) in PM 2.5 (Squizzato et al., 2013;Voutsa et al., 2014) was calculated in Table 4.These values were >1.0, except during the spring, which indicated that excess NH 4 was present.The mass ratio of NO 3 -/SO 4 2-was used as an indicator of the relative importance of mobile vs. stationary sources in this study (Yao et al., 2002;Wang et al., 2006;Gao et al., 2011).According to Yao et al. (2002), the estimated ratios of NO x to SO x in gasoline, diesel fuel and coal combustion emissions are 13:1, 8:1 and 1:2, respectively.In addition, high NO 3 -/SO 4 2-mass ratios result from a predominance of mobile sources over stationary sources.High NO 3 -/SO 4 2mass ratios have been reported in southern California, with values of 2 in downtown Los Angeles and 5 in Rubidoux due to low coal use (Kim et al., 2000).As shown in Table 4, the mean NO 3 -/SO 4 2-mass ratios for the seasonal samples in this study were approximately 0.3-1.1.Therefore, stationary source emissions were more important for contributing fine particles in the study area.

EC and OC
The seasonal EC and OC concentrations are summarized in Table 1.The OC concentrations were highest during the : 54 ± 28 µg m -3 ), followed by autumn, spring and summer, probably due to the combined effects of enhanced emissions from coal combustion for residential heating and stable atmospheric conditions during the winter (Duarte et al., 2008).The mean OC concentrations during the summer of 2013 (15 ± 9 µg m -3 ) were much higher than those during the summer of 2011 (9 ± 4 µg m -3 ) and 2012 (6 ± 2 µg m -3 ).
The EC concentrations showed no strong seasonal variations, except for slightly higher concentrations during the winter.
The annual mean concentrations of OC in PM 2.5 were 22 ± 19, 24 ± 17 and 33 ± 25 µg m -3 in 2011, 2012 and 2013, respectively.The annual mean EC values were the same for all three years (6 ± 3 µg m -3 ), as shown in Table 1.The combined effects of greater energy consumption and improved combustion technology resulted in higher OC levels and stable EC levels.
Each of the source materials resulted in a different OC/EC ratio when burned.The reported OC/EC ratios of typical emission sources are: 1.0-4.2 for diesel-and gasolinepowered vehicular exhaust (Schauer et al., 1999(Schauer et al., , 2002)); 2.5-10.5 for residential coal smoke (Chen et al., 2006); and 7.7 for cereal straw burning (Zhang et al., 2007).The OC/EC ratios reported by Koch et al. (2007) using the NASA Goddard Global Climate Model were < 1 for fossil fuel burning and > 5 for biomass burning.
In this study, the OC/EC ratios were 3.3 and 2.6 during the spring and summer, respectively, which were in the OC/EC ratio ranges of petrochemical fuel (< 1) and residential coalfired emissions (2.5-10.5).The OC/EC ratios were 5.4 and 6 during the autumn and winter, respectively, which corresponded to the OC/EC ratios of coal (2.5-10.5)and biomass (> 5) burning.Therefore, we concluded that fossil fuels (petrol, gasoline and coal) may be dominant during the spring and summer and that coal smoke and biomass burning were the main sources of OC and EC emissions during the autumn and winter.
In our previous study, the sources of PM 2.5 in Zhengzhou included dust (26%), secondary aerosols (24%), coal combustion (23%), biomass burning/oil combustion/ incineration (13%), vehicular emissions (10%) and industrial emissions (4%) (Geng et al., 2013).In 2011, coal consumption in Zhengzhou accounted for 78% of the total energy consumption, and oil accounted for 9% (Zhou, 2013).No statistics were available regarding biomass energy.The results of this study are consistent with those of our study and indicate the presence of a coal-dominated energy structure in Zhengzhou.Biomass energy sources, such as straw and wood, are traditional non-commercial energy sources that are dominant in rural areas.These energy sources are particularly used in households for cooking and domestic heating during the winter through direct combustion.Straw is an important component of energy consumption.

SOC
Directly quantifying primary and secondary organic components in PM 2.5 using chemical analysis is difficult.Several studies have estimated SOC using an EC tracer method that considers the minimum ratio of OC/EC as a constant mixture of primary OC and EC (Castro et al., 1999;Cabada et al., 2002Cabada et al., , 2004;;Duarte et al., 2008).
In 2011 and 2012, the maximum SOC to OC ratios occurred during the autumn (56 ± 22% and 60 ± 8%, respectively), and the minimum ratios (39 ± 19% and 35 ± 19%, respectively) occurred during the summer.In 2013, the lowest SOC to OC ratio occurred during the spring (47 ± 8%) and increased to > 62% during the autumn and winter.This result clearly indicated that the OC was dominated by SOC.On average, the contributions of SOC to PM 2.5 were 3-5, 3-10, 6-12 and 8-11% during the spring, summer, autumn and winter, respectively.This seasonal pattern (i.e., the contribution being highest in winter) was consistent with the results for Xiamen, as reported by Niu et al. (2012); however, these results were different from those obtained by Castro et al. (1999) and Cao et al. (2007), who reported that minimum SOC production occurred during the winter.SOC was formed from the atmospheric oxidation of volatile organic compounds (VOCs) and subsequent gas-to-particle conversion processes.In the study by Niu et al. (2012), it was concluded that low humidity, suitable temperatures (4-20°C) and low sunlight were advantageous for the formation of SOC.Temperature had a dramatic effect on the formation of secondary organic aerosols; the formation of SOA was approximately 2.5-6 times greater at 5°C (the lowest temperature studied) than at > 27°C (Warren et al., 2009).The climate of Zhengzhou is dry, with an annual average temperature of 14°C (-1 to 9°C in winter) and moderate sunlight, which may be favorable for the formation of SOC.
Meanwhile, greater emissions from residential heating (from mid-November to mid-March) could significantly contribute to the formation of the SOC fraction.The largest anthropogenic source of VOCs in Henan was stationary combustion (mainly residential coal combustion and biomass burning) (Fan et al., 2012).Zhengzhou is the provincial capital of Henan province and is surrounded by several rural areas.Biomass energy sources, such as straw and wood, have long been used as traditional non-commercial energy sources and dominate the energy consumption in rural areas.As the rural economy improved and lifestyles changed, the proportion of non-commercial energy decreased, stabilizing at approximately 50% after 2000 (Shi et al., 2010).However, straw consumption remained stable, and straw remains an important energy source.
Overall, the enrichment of SOC during autumn and winter is due to an increase in the emissions from residential coal and biomass combustion and the low-temperature transformation of VOCs and their condensation on particulate matter.

Secondary Components
The SIA fraction and the secondary particulate matter (SPM) fraction are summarized in Table 5.The highest SIA occurred during the winter (125 ± 76 µg m -3 ), and the lowest occurred during the summer (41 ± 16 µg m -3 ).A similar seasonal pattern was observed for SPM, with concentrations of 43 ± 16 µg m -3 during the summer and 164 ± 98 µg m -3 during the winter.
To understand the secondary aerosol pollution conditions, the contributions of secondary species to the overall PM 2.5 mass are shown in Fig. 5. On average, the contributions of SIA (NH 4 + , SO 4 2− and NO 3
Regarding the annual average, SIA accounted for 35%, 33% and 38% of the PM 2.5 during 2011, 2012 and 2013, respectively, with corresponding SOA contributions of 9%, 12% and 18%, respectively.Secondary aerosols accounted for 35-41%, 54-60%, 36-56% and 39-53% of the PM 2.5 mass during the spring, summer, autumn and winter, respectively, with the greatest contributions occurring during the summer of 2013 and the lowest contributions occurring during the spring of 2011.Thus, secondary aerosols, which comprise a large fraction of PM 2.5 , are important.

Visibility Degradation Impacts
The light extinction coefficient (b ext ) was estimated using the following formula proposed by the IMPROVE program and by applying the PM 2.5 chemical composition measurements to reconstruct b ext (IMPROVE, a): The specific dry scattering efficiency is 3 m 2 g -1 for ammonium sulfate and ammonium nitrate, 4 m 2 g -1 for organic carbon, 1 m 2 g -1 for soil, and 0.6 m 2 g -1 for coarse mass.In addition, the specific absorption efficiency for LAC is 10 m 2 g -1 .The relative humidity scattering enhancement factor, f(RH), was defined as 2 in this study based on the IMPROVE monitoring network (IMPROVE, b) and the annual average relative humidity of Zhengzhou (67%).
In our previous study, crust and trace elements only comprised 2-3% of the PM 2.5 in Zhengzhou (Geng et al., 2103).This study focused on the major chemical components of PM 2.5 and omitted the fine soil component, which was not included in the calculation.In this study, the factor used to convert OC to OM to account for the hydrogen, oxygen, and nitrogen present in the OM was changed from 1.4 to 1.6 (Turpin and Lim, 2001).Furthermore, the organic mass was divided into two fractions: POC (primary organic carbon) and SOC ([OMC] = 1.6{[POC] + [SOC]}).
Fig. 6 provides the seasonal and annual extinction coefficients from 2011to 2013.Generally, the annual mean extinction coefficients were 581, 566 and 847 Mm -1 for Higher values of the extinction coefficient indicate lower visibility.This result also reflects the important impacts of PM 2.5 on atmospheric visibility.In addition, the b ext value at the Zhengzhou site was much higher than that obtained from the PM 2.5 reconstructed aerosol light extinction coefficient (b ext_aer ) in the IMPROVE regions over 2005-2008.These values were reported as 5.73 Mm -1 in the Great Basin in January, 127.26 Mm -1 in Appalachia, USA, in August (rural) and 246.09Mm -1 in urban Fresno CA, USA, in December (Hand et al., 2011).Therefore, the visibility in the Zhengzhou region was very poor.Fig. 7 shows the seasonal and annual percentages of the contributions of the chemical species to the light extinction coefficient.As shown in Fig. 7, (NH 4 ) 2 SO 4 contributed the most (37-42%) to b ext according to the annual average results, followed by NH 4 NO 3 (23-25%), SOM (11-19%), POM (9-13%) and EC (8-11%).The percentages of SOM increased by 8% from 2011 to 2013, while those of the other species decreased in 2013.The secondary aerosols ((NH 4 ) 2 SO 4 + NH 4 NO 3 + SOM) accounted for 76-83% of the total b ext , with sulfate and nitrate ((NH 4 ) 2 SO 4 + NH 4 NO 3 ) accounting for 61-67%.These results implied that sulfate was the largest contributor to visibility degradation.However, the secondary species, particularly SIA, played a dominant role in the degradation of visibility in Zhengzhou.The major species that influenced the visibility depended on the region.The light extinction contributions that were reported for the spring of 2007 in Guangzhou, China (Tao et al., 2009), were as follows: (NH 4 ) 2 SO 4 (40 ± 6%) > OM (22 ± 4%) ≈ EC (22 ± 4%) > NH 4 NO 3 (16 ± 4%).These contributions were different from those in Zhengzhou.The significant contributions of (NH 4 ) 2 SO 4 + NH 4 NO 3 to the degradation of visibility resulted from the relatively high percentage of SIA in PM 2.5 .
In terms of seasonal percentages, (NH 4 ) 2 SO 4 contributed the most to PM 2.5 during the summer.NH 4 NO 3 and SOM had the highest contributions during the autumn, and these species exhibited seasonal patterns that were similar to those of the mass-based contributions to PM 2.5 .Overall, secondary aerosols ((NH 4 ) 2 SO 4 + NH 4 NO 3 + SOM) accounted for 80% of b ext , and primary aerosols (POM + EC) accounted for 20%.Clearly, secondary aerosols played a dominant role in the degradation of visibility in Zhengzhou.The same conclusion has been reported for other cities in the Pearl River Delta in China (Deng et al., 2013), where the extinction contribution of secondary aerosols was 60% in the dry environment (f(RH) ≈ 1.0) and increased to 75% and 82% at RH = 80% (f(RH) ≈ 3.0) and 95% (f(RH) ≈ 6.0), respectively.These values are closely related to the high f(RH) values that are used to calculate b ext in the Pearl River Delta.
Even when considering all of the involved uncertainties (e.g., the variable RH, empirical equations and magnitudes of the coefficients), the secondary species still played an important role in PM 2.5 .NOx and VOCs emissions drive photochemical reactions and their associated oxidants.Secondary fine particles are formed from photochemical and other reactions that involve precursor gases, such as SO 2 , NO x , NH 3 and VOCs.Most cities lack routine VOC measurements, and control of NH 3 and VOCs is lacking.Therefore, reducing secondary species by strengthening the control of SIA and SOA precursors, such as NO x , SO 2 , NH 3 and VOCs, will play an important role in controlling China's PM 2.5 levels and reducing their influences on the environment.

SUMMARY AND CONCLUSIONS
The annual mean concentrations of PM 2.5 were 186, 180 and 218 µg m -3 in 2011, 2012 and 2013, respectively.The PM 2.5 levels in Zhengzhou were 5-6 times greater than the values listed by the National Ambient Air Quality Standard of China (annual value of 35 µg m -3 ).Thus, severe PM 2.5 pollution occurs in Zhengzhou.The general increases observed in PM 2.5 , particularly in SIA and SOC, are associated with increases in the number of vehicles and the total energy consumption in Zhengzhou.Regulatory agencies should strictly enforce air pollution control measures and identify effective measures for reducing primary and secondary PM 2.5 to control air pollution.
As indicated by the NO 3 -/SO 4 2-mass ratio, stationary source emissions remain an important contributor of fine particles in Zhengzhou.Obvious SOC enrichment can be observed during the winter and autumn, most likely due to biomass burning and increased coal combustion during the autumn and winter.
An investigation of secondary species revealed that secondary aerosols played a dominant role in the total mass of PM 2.5 and the degradation of visibility.SIA accounted for 26-50% of the PM 2.5 mass, while SOM accounted for 4-21%.In addition, secondary aerosols ((NH 4 ) 2 SO 4 + NH 4 NO 3 + SOC) accounted for 80% of the b ext , and (NH 4 ) 2 SO 4 and NH 4 NO 3 were the main factors related to poor visibility in Zhengzhou.
To achieve PM 2.5 pollution control targets, measures must be taken to control NO x , SO 2 , NH 3 and VOC emissions.

Fig. 5 .
Fig. 5. Annual and seasonal contributions of the secondary components to the PM 2.5 mass concentrations during 2011-2013.

Fig. 7 .
Fig. 7. Annual and seasonal extinction contributions of the individual components to PM 2.5 .

Table 2 .
Civil vehicles, total energy consumption and population of Zhengzhou from 2011 to 2013.

Table 3 .
Calculated sulfur and nitrogen oxidation ratios (SOR and NOR) of PM 2.5 .

Table 5 .
Seasonal OC/EC ratios, SOC and SIA fraction contributions to PM 2.5