Chemical Analysis of Particulate Matter in the Harvest Period in an Agricultural Region of Eastern China

PM2.5 samples were collected for August 13–22 (non-harvest period, NHP) and for October 21–31 (harvest period, HP) in 2014 from an agricultural region of Eastern China. The samples were subsequently analysed to determine mass concentrations and fractions of elements, water-soluble ions and carbon components. Online datasets (SO2, NO2, O3, CO, PM10 and PM2.5) and meteorological conditions were synchronously monitored. The average mass concentrations of PM2.5 during the HP and NHP were respectively 108.3 and 62.6 μg m. Compared with the mass concentrations of organic carbon (OC), Cl, NO3 and K during the NHP, those during the HP were significantly increased. Moreover, the mass fractions of OC, elemental carbon (EC), Cl and K during the HP were respectively 1.6, 1.3, 3.2 and 1.3 times of those during the NHP. SO4, NO3, and OC were the major chemical components in PM2.5 during the HP, indicating that biomass burning and secondary transformation may be two major sources of PM2.5 during the HP. The K/Cl value in PM2.5 during the HP was lower than 1, indicating that maize straws were the crop residues in the study area. Although the sulphur and nitrogen oxidation ratios during the HP were lower than during the NHP, the effects of the secondary transformation on particles cannot be ignored during the HP. Biomass burning yielded a 58% OC concentration during the HP.


INTRODUCTION
With the increasing demand for more favourable air quality in recent years, PM 2.5 has drawn considerable attention (Hu et al., 2011;Zheng et al., 2014;Sun et al., 2016); PM 2.5 is a type of particulate matter with an aerodynamic diameter of less than 2.5 µm.As an atmospheric pollutant, PM 2.5 can cause visibility degradation, affect surface solar radiation and damage to human health (Dockery et al., 1993;Watson, 2002;Yuan et al., 2006;Duan et al., 2013;Zhang et al., 2013).
In China, opening burning of crop residue has been reported as the principal source of PM 2.5 , which also causes heavy regional haze pollution (Huang et al., 2013;Cheng et al., 2014).More than 1.8 billion hectares of farmland in China is used for food production (Luo et al., 2016), and large quantities of crop residue are burned in farmlands during the harvest period (HP) each year (Andreae and Merlet, 2001;Cao et al., 2008;Jain et al., 2014).In China, the harvest period includes two periods, summer period (from late May to early June) and autumn period (from late October to early November) (Huang et al., 2013;Zhang et al., 2017).This human activity has increased the emissions of harmful gases and PM 2.5 , which aggravates air pollution, affects the ground radiation balance, and damages human health during the harvest season.Crop residue burning is a type of biomass burning, Street et al. (2003) estimated that Asia burned 730 Tg of biomass in 2000, among which, 250 Tg biomass was released by crop residue burning.Therefore, crop residue burning is a critical source of PM 2.5 , and it is essential that the effects of crop residue burning on air quality and the characteristics of pollutants during the HP are determined.
Studies have determined the source apportionment of particulates in numerous large cities, such as Beijing (Liu et al., 2014), Shanghai (Hu et al., 2014), and Guangzhou (Cui et al., 2015), and biomass burning has markedly contributed to PM 2.5 and increased the mass concentrations of K + , Cl -and carbon components (Li et al., 2007;Lin et al., 2010;Zhang et al., 2015).Farmlands in the aforementioned developed cities are mainly located in the suburbs or surrounding cities, therefore, the effects of crop residue burning on air quality are small in these cities.Hence, we must select an agricultural city which burns crop residues during the harvest season to determine how crop residue burning affects atmospheric quality.
As the Ministry of Environmental Protection of the People's Republic of China reported, for late October in 2014, the Shandong province ranked third through the country according the numbers of straw burning fire and fifth according to the intensity of straw burning fire.We chose Heze City as the study area, which is a developing city located in the southwest of the Shandong province (34°39′-35°52′N, 114°45′-116°25′E).It is a predominantly agricultural city whose farmlands accounts for 68.3% of the city's total area.October is the maize harvest period for Heze, and large quantities of maize residues are burned during the HP.We selected this city as the study area and analysed the following factors: (1) pollutant characteristics during the HP, (2) chemical component characteristics of PM 2.5 during the HP, (3) the differences in the chemical components between HP and the non-harvest period (NHP), and (4) the effects of crop residue burning on the atmosphere.

Sampling
Fig. 1 and Table 1 show the location and surroundings of the sampling sites.PM 2.5 samples were collected for October to 21-31 in 2014 during the HP, and for August 13-22 in 2014 during a NHP for comparison.Gaseous pollutant concentrations and meteorological conditions were monitored synchronously at the sampling stations.The difference in the pollution sources for HP and NHP depend on agricultural emission conditions, these include crop residue burning emissions for HP and biogenic emissions for NHP.During the NHP, the main sources of particles are combustion, industrial emissions, automobile exhaust and flowing dust.For HP, crop residue burning emissions is the main source of PM 2.5 , the contribution of other sources (e.g., combustion, industrial emissions, automobile exhaust and flowing dust) to PM 2.5 is low.
Two types of filter were used to collect particulate matters: Teflon filters (90 mm diameter) were used to analyse elements, and quartz filters (90 mm diameter) were used to analyse ions and carbon.The pre-treatment for filters was as follows: the Teflon and quartz filters were respectively baked in ovens at 60°C and 400-500°C for at least two hours; after baked, all blank filters were kept in silica gel desiccators for at least three days before being weighed.An over one hundred thousandth (1/100000) scale (Mettler Toledo AX205) was used to weigh the filters before and after sampling.Two pre-calibrated samplers (TH-150, Wuhan Tianhong Intelligence Instrument Facility, China) collected PM 2.5 samples at each site via equipped Teflon  Quality assurance and quality control (QA/QC) in the sampling process of PM 2.5 mainly included: (1) The flow of each sampler was calibrated every day to eliminate system errors; (2) all of the samplers at each site were started at the same time every day; (3) all of the filters were ensured to maintain integrity at every step of sampling; (4) the fraction of the parallel samplers aimed to be 10% of the total, and the relative standard deviations of the parallel samples were equal to or less than 20%.

Chemical Analysis
After sampling and weighing, each sample was analysed for chemical components including elements (Na, Mg, Al, Si, K, Ca, Ti, V, Cr, Fe, Mn, Zn, Ni, Cu and Pb), organic carbon (OC), elemental carbon (EC), and water-soluble inorganic ions (Na + , Mg 2+ , Ca 2+ , K + , NH 4 + , F -, Cl -, NO 3 and SO 4 2-).Elements were analysed using inductively coupled plasma mass spectrometry [ICP 9000(N+M), USA].A quarter of each Teflon filter was cut into fragments and treated with a mixture of 4 mL HNO 3 , 2 mL HCl and 1 mL H 2 O 2 .The sample and acid mixtures were heated with an electric stove and evaporated until 3 mL of residual remained.When cooled to room temperature, the sample was transferred to a test tube and diluted with deionised water.Si and Al were analysed using an alkali solution was used, and the treatment procedure was the same.Reagent blanks were tested, and the test results were under the detection limits.Two blank samples was analysed for every 20 samples to ensure relative standard deviation of element contents between the two blank samples was less than 20%.The limits of detection for Na, Mg, Al, Si, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn and Pb were 0.0002, 0.00005, 0.0002, 0.005, 0.004, 0.0005, 0.005, 0.003, 0.004, 0.0005, 0.002, 0.009, 0.002, 0.005 and 0.03 µg mL -1 .
OC and EC were determined by an OC/EC analyser (Atmoslytic Inc.DRI2001A, USA) using a punch of each quartz filter of area 0.558 cm 2 .An IMPROVE thermal/optical transmittance method was applied to measure the concentrations of OC and EC.The carbon analysis process contained seven steps of heating programs, including 140°C (OC1), 280°C (OC2), 480°C (OC3), 580°C (OC4), 580°C (EC1), 740°C (EC2) and 840°C (EC3).When each sample was analysed, we could determine the concentrations of OC1, OC2, OC3, OC4, EC1, EC2, EC3 and pyrolysis carbon (OPC).IMPROVE defines OC and EC as follows: The respective detection limits of OC and EC were 0.45 and 0.06 µg cm -2 .A repeat sample was analysed for every 10 samples to ensure that the instrument precision error was less than 2%.The analyser was calibrated every day before and after the analysis.
The water-soluble inorganic ions were analysed via ionic chromatography (IC, Dionex ICS-900, USA).A quarter of the quartz filter was extracted into 5 mL with deionized water in an ultrasonic bath (GT sonic, GT-2120QTS, China) for 15 min, with a frequency of 40Hz.Then, 1 mL of supernatant solution was extracted using a syringe equipped with a disposable filter head (with 0.22 µm pore size).Finally, the supernatant solution was injected into the ion chromatograph.Field blanks were tested to calibrate the concentrations of ionic species, and the test results were under the detection limits.The standard solutions were detected three times prior to analysis, and low relative standard deviations were observed.The respective detection limits of Na + , Mg 2+ , Ca 2+ , K + , NH 4 + , F -, Cl -, NO 3 -and SO 4 2-were 0.004, 0.006, 0.007, 0.007, 0.017, 0.009, 0.01, 0.07 and 0.05 µg m -3 .

Concentrations of Gaseous Pollutants and Particulate Matters
Table 2 shows the mass concentrations of the six pollutants and meteorological conditions during the NHP and HP.The relative humidity and wind speed showed minor differences between the NHP and HP, and temperature during the HP was 8.9°C lower than during the NHP (Table 2).Temperature during the NHP and HP were above 16°C, NHP and HP were in a warmer season in a year, usually diffusion condition in a warmer season is better.
The mass concentration of SO 2 during the HP was 48.7 µg m -3 , which was 2.3 times that during the NHP.The mass concentrations of NO 2 and CO during the HP were respectively 53.7 µg m -3 and 1.7 mg m -3 , which were 1.6 and 1.3-fold higher than those during the NHP.The respective concentrations of PM 10 and PM 2.5 during the HP were 181.5 and 108.3 µg m -3 , which were 1.9 and 1.7-fold higher than during the NHP.The mass concentration of O 3 was markedly higher during the NHP (84.7 µg m -3 ) because of the stronger sunlight and higher temperatures.This may  (Khoder, 2002;Liu et al., 2011).Briefly, the mass concentrations of particulate matter and gaseous pollutants markedly increased during the HP, whereas the mass concentration of O 3 decreased.The present findings are in concordance with the studies conducted by Andreae and Merlet (2001) and Cheng et al. (2013), who reported that biomass burning, including crop residue burning, contributed to large quantities of particles and trace gases.

Chemical Components Characteristics in PM 2.5
The mass concentrations of major chemical components in PM 2.5 during the NHP are shown in Fig. (2(a)).Among the major chemical components, the mass concentration of SO 4 2-was highest (23.8 µg m -3 ).NO 3 -, OC, crustal elements (Al, Si and Ca), NH 4 + , and EC were present in moderate concentrations, with respective mass concentrations of 8.7, 7.1, 6.9, 4.8 and 3.3 µg m -3 .Among the water-soluble ions, the respective mass concentrations of Cl -and K + were lower, at 1.1 and 0.9 µg m -3 .
The mass concentrations of the major chemical components in PM 2.5 during the HP are shown in Fig. 2(b).The major components were SO 4 2-, NO 3 -and OC, at respective mass concentrations of 21.5, 20.8, and 18.4 µg m -3 .The respective mass concentrations of NH 4 + , EC, and Cl -were moderate, at 8.7, 6.7, and 5.7 µg m -3 .The mass concentration crustal elements were similar to those of EC, which was approximately 3.5-fold higher than that of K + .Compared with the mass concentrations of OC, Cl -, and NO 3 -during the NHP, those during the HP were significantly increased.Although the mass concentrations of K + during the NHP and HP were lower than 2 µg m -3 , the mass concentration of K + during the HP was 2.1-fold higher than during the NHP.
Fig. 3(a) shows the mass fractions of different chemical components detected during the NHP.The mass fraction of SO 4 2-was highest (40.4%), followed by that of NO 3 -(14.6%).SO 4 2-and NO 3 -are respectively markers of secondary SO 4 2and secondary NO 3 - (Song et al., 2006;Lestari et al., 2009), and their high mass fractions indicate that the secondary transformations of SO 2 and NO x are crucial sources of PM 2.5 during the NHP.Considering carbon components, the mass fractions of OC and EC were respectively 12.1% and 5.5%.The mass fraction of crustal elements was 11.6%, and NH 4 + (mass fraction, 8.0%) ranked third among the water-soluble ions.The mass fractions of Cl -(1.9%) and K + (1.6%) were similar. 2-and NO 3 -exceeded 20%, and the mass fraction of OC was 19.6%, these components accounted for more than half of the detected chemical components.NH 4 + (mass fraction, 9.3%) ranked third among the water-soluble ions.Furthermore, the mass fraction of the marker for coal combustion and wheat straw burning (Cl -) (Zheng et al., 2005a;Song et al., 2006;Li et al., 2007) was 6.0%.The mass fraction of K + was 2.1%, K + was the tracer of biomass burning (Cheng et al., 2013;Sopittaporn et al., 2013).The mass fractions of both EC and the crustal elements were approximately 7%.Therefore, biomass burning and the secondary transformation of SO 2 and NO x may be concluded as being two major sources of PM 2.5 during the HP in Heze City.Compared with the mass fractions of OC, EC, Cl -, and K + during the NHP, those during the HP were markedly increased, and were respectively 1.6, 1.3, 3.2, and 1.3-fold higher than those during the NHP.
Crop residues have varying emission characteristics;  Table 3 shows the percentages of OC, EC, K + and Cl -in PM 2.5 of different kinds of crop residues.We can conclude that the K + /Cl -values vary in crop residues.The K + /Cl - values for maize straws were lower than 1 (Turn et al., 1997;Li et al., 2007), but that for wheat straws was higher than 1 (Turn et al., 1997;Hays et al., 2008).This value ranged 0.6-1.0 for rice straws (Turn et al., 1997;Hays et al., 2008).The K + /Cl -values for sugarcane straw were similar to those for rice straws, in the range 0.6-1.0(Turn et al., 1997).In our study, the K + /Cl -value of PM 2.5 was 0.4, which is lower than 1, and maize harvesting in Heze City was performed in October.Therefore, maize straws were the crop residues in Heze City during the HP.

Secondary Inorganic Ions
As previously mentioned, we concluded that SO 4 2-and NO 3 -play crucial roles in PM 2.5 , and they are commonly called secondary inorganic ions.Therefore, it is vital to focus on secondary inorganic ions.
Secondary inorganic ions (SO 4 2-, NO 3 -, and NH 4 + ) are crucial components of fine particle matter (Mather et al., 2003;Eliana et al., 2014) and play vital roles in visibility reduction, global radiation budgets, regional haze pollution, and human health (Haywood et al., 2008;Liu et al., 2008;Wang et al., 2012).Ammonia vapour can react with acidic gas or condense on the surface of acidic particles and accumulate in droplets, subsequently generating particulate NH 4 + (Hong et al., 1999).The correlation between NH 4 + and SO 4 2-in PM 2.5 can suggest their presence in particulate matter (Wang et al., 2015).
Fig. 4 shows that when the molar ratio of NH 4 + to SO 4 2for PM 2.5 was 2, NH 4 + and SO 4 2-existed in the form of (NH 4 ) 2 SO 4 (Possanzini et al., 1999;Deng et al., 2010).When the molar ratio of NH 4 + to SO 4 2-for PM 2.5 was 1, they existed in the form NH 4 HSO 4 (Deng et al., 2010).The average molar concentration ratio of NH 4 + to SO 4 2-for PM 2.5 during the HP was 1.8, which was close to 2, indicating that NH 4 + and SO 4 2-mainly existed in the form (NH 4 ) 2 SO 4 in PM 2.5 .Moreover, the average molar concentration ratio of NH 4 + to SO 4 2-for PM 2.5 during the NHP was 1.0, which was close to 1, indicating that NH 4 + could not sufficiently neutralize SO 4 2-during the NHP and may exist in the atmosphere in the form of NH 4 HSO 4 .
As the primary water-soluble ions in atmospheric particulate matter, SO 4 2-and NO 3 -are mainly formed by SO 2 and NO x , respectively, through a series of photochemical reactions (Duan et al., 2003).We used the sulphur oxidation ratio (SOR) and nitrogen oxidation ratio (NOR) to represent the respective transformation ratios of SO 2 and NO 2 to estimate the respective transformation of SO 2 and NO x to SO 4 2-and NO 3 -.Higher SOR and NOR were associated with a stronger oxidation capacity in the atmosphere, and more gaseous pollutants transformed into SO 4 2-and NO 3 in the particulate matter (Grosjean and Seinfeld, 1989;  Wen et al., 2007).The formulae for SOR and NOR are as follows (Grosjean and Friedlander, 1975;Kadowaki, 1986;Ohta and Okita, 1990): where [SO 4 2-] and [SO 2 ] respectively represent the mass concentrations (µg S m -3 ) of SO 4 2-and SO 2 , and [NO 3 -] and [NO 2 ] respectively represent the mass concentrations (µg N m -3 ) of NO 3 -and NO 2 (Kadowaki, 1986;Grosjean and Friedlander, 1975;Ohta and Okita, 1990).
The SORs were respectively 0.41 and 0.24 during the NHP and HP.The NORs were respectively 0.15 and 0.22 during the NHP and HP.The NHP met the condition of a strong photochemical oxidation reaction (SOR > 0.25 and NOR > 0.1) (Khoder, 2002), indicating that most SO 4 2and NO 3 -were formed through the photochemical oxidation of SO 2 and NO 2 , respectively.The SOR during the NHP was higher than that during the HP, indicating that the more crucial role of secondary SO 4 2-during the NHP.NOR during the HP was slightly higher than that during the NHP; this was mainly affected by temperature because NH 4 NO 3 decomposes in high temperature conditions, which results in decreased NO 3 -in particulate matter (Liu et al., 2011).The SOR and NOR were lower during the HP than during the NHP; however, the effect of the secondary transformation on particles during the HP cannot be ignored.

Carbonaceous Aerosols Estimation from Biomass Burning
Crop residue burning can be expected to greatly contribute to carbonaceous aerosols during the HP, and the contribution can be estimated from the relationship between K + and OC, which are crucial components of biomass burning.K + has other sources (including sea salt, combustion, and industrial emissions); regarding reducing the effects of other sources of K + , Pachon et al. (2013) developed a method for estimating K + from biomass burning on the basis of regression analysis between K + and other species that share sources with K + , except biomass burning.The formula for estimating K + from biomass burning is as follows (Pachon et al., 2013): where K + biomass burning represents the concentration of K + ion from biomass burning, K + represents the K + ion concentration, and [Fe] represents the Fe concentration (Pachon et al., 2013).
K + biomass burning during the HP was 1.72 µg m -3 , which 3.0-fold higher than that during the NHP, suggesting a higher contribution of biomass burning during the HP.Fig. 5 shows the correlation between K + biomass burning and OC during the NHP and HP, and their respective correlation factors (Person's R) were 0.56 and 0.74.The strong correlation during the HP indicated biomass burning to be a vital source of OC.

CONCLUSIONS
In this study, PM 2.5 samples were collected, and online datasets (SO 2 , NO 2 , O 3 , CO, PM 10 and PM 2.5 ) were monitor for August 13-22 and October 21-31 in 2014.The chemical components in the samples were analysed to determine differences in pollutant characteristics between the NHP and HP.The mass concentrations of particles and gaseous pollutants, except for O 3 , were higher during the HP than during the NHP.
Compared with the mass concentrations of OC, Cl -, NO 3 -and K + during the NHP, those during the NHP were significantly increased, and the mass fractions of OC, EC, Cl -and K + during the HP were respectively 1.6, 1.3, 3.2, and 1.3-fold higher than those during the NHP.SO 4 2-, NO 3 -, and OC were the major chemical components of PM 2.5 during the HP.Biomass burning and secondary transformation may be concluded as being two major sources of PM 2.5 during the HP.Crop residues have varying K + /Cl - values, and the proportion of K + was lower than that of Cl - in PM 2.5 from crop residue burning, indicating the presence of maize straw crop residues in the study area.
NH 4 + and SO 4 2-respectively existed in (NH 4 ) 2 SO 4 and NH 4 HSO 4 forms, during the HP and NHP.Although the SORs and NORs were lower during the HP (respectively 0.24 and 0.15) than during the NHP (respectively 0.41 and 0.22), the effects of secondary transformations on particles during the HP cannot be ignored.Biomass burning yielded a 58% OC contribution during the HP.

Fig. 3
(b) shows the mass fractions of different chemical components detected during the HP.The mass fractions of SO 4

Fig. 2 .
Fig. 2. Mass concentrations of major chemical components in PM 2.5 during the NHP and HP (heavy metals: As, Cd, Cr, Cu, Fe, Mn, Ni, Pb, V, and Zn; crustal elements: Al, Si, and Ca).

Fig. 3 .
Fig. 3. Mass fractions of different chemical components in the detected components during the NHP and HP (heavy metals: As, Cd, Cr, Cu, Fe, Mn, Ni, Pb, V, and Zn; crustal elements: Al, Si, and Ca).

Fig. 5 .
Fig. 5. Correlation between OC and K + biomass burning during the NHP and HP.

Table 1 .
Location and surroundings of the sampling sites.

Table 2 .
Mass concentrations of pollutants and meteorological conditions in Heze City during the study period.RH, relative humidity; WS, wind speed.also account for the mass concentrations of SO 2 and NO 2 being lower during the NHP than during the HP.During the NHP, strong sunlight facilitated photochemical reactions, and more SO 2 and NO 2 were respectively transformed into SO 4 2-and NO 3 -

Table 3 .
Chemical composition of PM 2.5 emitted from different crop residue fires (% PM 2.5 ).