Chemical Composition of PM 2 . 5 Based on Two-Year Measurements at an Urban Site in Beijing

In this study, the daily PM2.5 atmospheric aerosols were collected on quartz and PTFE filters simultaneously from January 2008 to December 2009. Organic carbon (OC) and elemental carbon (EC), water-soluble ions including SO4, NO3, Cl, NH4, K, Ca, and Mg were analyzed for the samples. The annual average mass concentration of PM2.5 for PTFE was 79 μg m. The OC, SO4, NO3, NH4 accounted for about 75 % of PM2.5, and secondary organic carbon (SOC) was estimated about 50% of OC. Monthly averages of SO4, NO3, NH4, K were maximum in June and minimum in October, but EC, Cl displayed the highest in December and the lowest in July, which is related to the coal-combustion emission from heating. Three groups with high, medium and low PM2.5 concentrations were categorized. SO4, NO3, OC and NH4 accounted for 24%, 20%, 21% and 11% of sum of defined components in the high concentration days. In the medium and low concentrations, OC occupied large fractions of defined components. Mass closure was obtained for PTFE, but not for quartz. The PM2.5 mass concentrations on quartz filters were about 50 μg m higher than that on PTFE. The concentrations of water soluble ions on quartz filters were about 60–70% of that on PTFE filters. About 15–30% of PM2.5 was considered as the contribution of water vapor, the artifact of water vapor on quartz filter should be noted in later research works. PM2.5 displayed neutral during the year of 2008 and appeared acidic at the next year according to the calculation of cations/anion, concentration of hydrogen and acidic purity. Carbonaceous aerosols occupied same fractions in neutral and acidic aerosols. While sulfate and nitrate contributed 32% and 21% to PM2.5 for acidic aerosols, and 22%, 17% of PM2.5 from sulfate and nitrate for neutral aerosols.


INTRODUCTION
Atmospheric aerosols are largely responsible for air quality deterioration, visibility reduction (Watson, 2002;Molina and Molina, 2004) and has an adverse effect upon human health ( (Pope III et al., 2002).Particulate Matter (PM) with the diameter less than 10 µm (PM 10 ) and 2.5 µm (PM 2.5 ) are widely used to investigate the chemical, physical and optical properties of atmospheric aerosols.Organic matters (Cao et al., 2004) and water-soluble ions are major constituents (Putaud et al., 2004;Yang et al., 2011), which influence the absorption and scattering properties, resulting in hazy days.Moreover, acidic aerosols cause harm to human health and contribute to acid deposition and the ecological system damage (Bouwman et al., 2002).Beijing, as the capital and the center of politics, economics and culture in China, has experienced effects of serious PM pollution such as haze.Although the government has devoted itself to improving air quality and numerous studies have been conducted, Beijing is still suffering with serious hazy days nowadays, some essential questions remain unknown (Zhang et al., 2013).Therefore, comprehensive investigation of PM 2.5 and obtain the mass concentration of these main components are the key method to track the emission sources, understand their formation process and supply the references of reducing the pollutions.
A number of studies have dedicated to the characterization of particulate matter in recent years based on filters measurement and analysis in Beijing (He et al., 2001;Dan et al., 2004;Sun et al., 2004;Wang et al., 2005;Duan et al., 2006;Wu and Wang, 2007;Zhang et al., 2007;Pathak et al., 2009;Cao et al., 2012;Wang et al., 2013;Zhao et al., 2013;Hu et al., 2014).These studies reported a lot of valuable results on PM and chemical composition in different years which make us understand aerosols properties better, but most of them focus on the characterization of aerosol at short-term measurements representing one or two seasons.Seldom researches covered more than one-year measurements.For limited samples and short monitoring time would make deliberate source identification and trend estimation difficult (Wang et al., 2013), a set of two-year monitoring data would be more representative to evaluate the pollution situation and sources identification.
In this paper, we report the results of a two-year study carried out in Beijing using MiniVol sampling instruments to collect PM 2.5 on both quartz and PTFE filters during [2008][2009].First of all, the differences of PM 2.5 mass concentration and water-soluble ions between quartz and PTFE filters are surveyed.The temporal variation of mass concentrations and their chemical components are investigated.Based on the national air quality standards for PM 2.5 (35 µg m -3 for first grade, and 75 µg m -3 for second grade), three groups of days with high, medium and low PM 2.5 concentrations were categorized, the chemical components related with polluted formation, source emissions and mass closures are discussed.Finally, the mass concentrations of hydrogen ions and acidic purity, the aerosol acidities are estimated via the ratio of cations to anions.These results are valuable for descripting the general level of PM 2.5 , understanding the chemical composition in different polluted cases relating to local sources and atmospheric transformation processes.It also supplies a reference for the formulation of effective air quality management strategies.

Measurement Site and Sampling
During this research project, three side-by-side MiniVol™ air samplers (Airmetrics, Oregon USA) were placed at the top of a 9 story building (39°56′N, 116°24′E, 35 m a.s.l.) in the campus of the China Meteorological Administration (CMA) of Beijing, which is believed to represent the typical urban environment in Beijing (Zhang et al., 2011).The atmospheric aerosols were collected on filtration media, operating for 24 h from 09:00 a.m. to 09:00 a.m.(Beijing time) at the following day, with ambient air flow rate of 5 L min -1 between 1 January 2008 and 31 December 2009.The 47mm Whatman quartz microfiber filters (QMA) were baked at 800°C for 3 h to remove the potential OC interferences before use PTFE filters with 2 µm pore size and diameter of 46.2 mm with the supported PP ring (Whatman Inc., Clifton, NJ, USA) were also used.From 1 January 2008 to 30 April 2009 (period 1), aerosols were collected every day except rainy day, and sampled every Tuesdays and Thursdays from 1 May 2009 afterwards (period 2).All of these filters were weighted before and after the ambient sampling to obtain the mass concentration of PM 2.5 .

Data Analysis
The filters were weighted with an electronic microbalance (Sartorius-ME5, ± 1 μg) after equilibrating the filter under constant temperature (25 ± 1°C) and relative humidity (30 ± 2%) for over 24 h.Quartz samples were used for organic carbon (OC) and elemental carbon (EC) analysis by means of a DRI 2000A carbon analyzer (Desert Research Institute, Reno, NV, USA) following the IMPROVE protocol (Chow et al., 2001).Both quartz and PTFE filters were analyzed by ion chromatography using ICS 3000 (Dionix Corp., Sunnyvale, CA, USA), water-soluble ions were extracted from one fourth of each filter with 10 mL ultraclean water.All the samples were extracted at an ice-water ultrasonic bath at 0°C for 1.5 hours to avoid the losses of ions.The solution was filtered by a syringe filter and then analyzed the concentration of cations (Ca 2+ , Mg 2+ , K + , Na + , NH 4 + ) and anion (SO 4 2-, NO 3 -, Cl -,) by ICS 3000 (Zhang et al., 2012b).In order to estimate the primary OC (POC) and secondary OC (SOC), the minimum OC/EC ratio method was used with the following equation (Turpin and Huntzicker, 1995;Castro et al., 1999): SOC = OC -EC × (OC/EC) min (Cao et al., 2004).In this work, the observed minimum OC/EC ratio 2.13 was used for SOC calculation.
To better understand the acidity or neutralization status of PM, the concentration of hydrogen ion (H + ) and acid purity (f) in each sample were calculated using all measured ions.The concentration of hydrogen ion (H + ) in each sample is calculated using all measured ions (Schwab et al., 2004;Hu et al., 2014).
The brackets represent concentrations in molar units.This equation assumes that the unmeasured ions (except [H + ]) are negligible compared to the measured ions.The acid purity is assumed as the ratio of [H + ] to the sum of all cation concentrations according to the formula f (Stevens et al., 1980).

Comparison of PM 2.5 on Quartz and PTFE FILTERS
The scatter plot of PM 2.5 mass concentrations on Quartz filters verse on PTFE filters was shown in Fig. 1.PM 2.5 mass concentration on Quartz and PTFE filters shows an acceptable correlation with the correlation coefficient (r 2 = 0.73), the fitting curve slope of 0.99, and the intercept is 48.8 µg m -3 , suggesting the mass concentration on the quartz filters is much higher than that on the PTFE filters.The Fig. 1.Scatter plot of mass concentration of PM 2.5 between quartz and PTFE filters, the uncertainty in the regression slope is at the 68% confidence interval.) r 2 = 0.73 Slope = 0.99 ± 0.03 Intercept =48.8 ± 2.9 Line 1:1 hydrophobic nature of PTFE filters prevents transfer water and water molecules remain linked to the dust particle until the end of the sampling (Zdziennicka et al., 2009), but quartz filters are more sensitive of absorbing organics vapor and water vapor would be the main reason.So the data reported by quartz filters may overestimate the concentration of PM 2.5 .On the purpose of surveying the collection efficiencies of different water-soluble ions on quartz filters, regarding the ions concentration of PTFE as reference, the relationships for PM 2.5 chemical species between the quartz and PTFE filters are plotted in Fig. 2.There are moderate correlations in which the squares of correlation coefficient for SO 4 2-, NH 4 + , NO 3 -, Cl -and K + are 0.76, 0.74, 0.55, 0.69 and 0.78, respectively.However, for Ca 2+ , Mg 2+ , the correlations are not good, and there is even no relationship for Na + in this campaign.The concentrations of chemical species on quartz filters are general lower than on PTFE filters, which is 0.77, 0.66, 0.76 and 0.56 of that on PTFE filters for sulfate, nitrate, ammonium and chloride respectively.These differences between quartz and PTFE filter samples could be caused different recovery/extraction efficiency, and mass size distributions of aerosol.Moreover, the sampling flow rate and specification of quartz or PTFE filters may influence the collection efficiency of Quartz or PTFE filters for aerosol.Wake (Wake et al., 1994) suggested that deposition of HNO 3 didn't influence the nitrate concentration for 47mm PTFE filters.Volatilization loss of NH 4 NO 3 from the filter has been found during sampling, especially during the warm months and during the warmest periods of the day (Chow et al., 2005), which might also influence the concentration of nitrate and ammonium in quartz filters.
According to the comparison between quartz and PTFE filters, PTFE filters are more optimize to report the concentration of PM 2.5 and water-soluble ions, and quartz filters are suitable to obtain the concentration of OC and EC.Here, we combined the data of PM 2.5 , water-soluble ions from PTFE and the data of OC, EC from quartz filters to discuss the characterization of PM 2.5 .

Temporal Variation of PM 2.5 and Their Compositions
The concentration time series of daily PM 2.5 mass and of soluble species (SO 4 2-, NO 3 -, NH 4 + , Cl -, K + , Na + , Ca 2+ and Mg 2+ ) based on PTFE filters, and that of OC, EC, and  calculated SOC, POC based on quartz filers are plotted in Fig. 3.The average mass concentrations for PM 2.5 and their chemical composition during period 1 are summarized in Table 1.The average mass concentration of PM 2.5 was 77.9 µg m -3 , varied from 0.8 µg m -3 to 372 µg m -3 , which is about twice of the annual PM 2.5 national air quality standards in China (35 µg m -3 ).The mass loadings of OC, SO 4 2-, NO 3 -, and NH 4 + also varied from several µg m -3 to hundred µg m -3 with the average level of 17.3 µg m -3 , 18.4 µg m -3 , 14.6 µg m -3 and 8.3 µg m -3 respectively.Table 2 summarized the published dataset of PM 2.5 and their chemical compositions from 2001 to 2010 in Beijing.It reveals that the annual PM 2.5 decrease from 150 µg m -3 during 2001-2003 (Wang et al., 2005) to 78 µg m -3 during 2008-2009, and then increase to 123 µg m -3 in 2009 (Zhao et al., 2013) and 98 µg m -3 in 2010-2011 (Wang et al., 2013).Amongst the studies, the concentration of PM 2.5 for this work is the lowest, which is about half of that during 2001-2003, and about 40% lower than in 2009.As the host city of the 29 th Olympic Game in 2008, many measures of    improving air quality were carried out before, during and after the Game in Beijing (Zhang et al., 2009), which resulted in the reduced concentrations of particulate matter in this study.Seasonally, it is surveyed that the highest PM 2.5 always occurred in winter and lowest in spring and fall for the published works, but no obvious increasing or decreasing trends were found within the ten years.Multiple factors such as different monitoring periods for concerned studies, changes of source emissions, photochemical processes and the variation of meteorological conditions and so on could have effects on the aerosol concentration from year to year.
As for the carbonaceous aerosols, the average OC and EC in PM 2.5 was 17.3 µg m -3 and 3.9 µg m -3 respectively.The OC concentration is same as the previous results in 2009 (Zhao et al., 2013), while the EC is lower than that in 2009 due to the strict traffic constrain measures.In Beijing, the coal combustion and diesel traffic are considered as main contributions to the primary organic aerosol with 35% and 25% (Zhang et al., 2012a).As precursors of SOC, Volatile organic compounds (VOCs) mainly come from industrial (13%), housing waste disposal (12%), combustions of gasoline (14%) and diesel (12%) (Zhang et al., 2012a).The OC concentrations during 2009 (Zhao et al., 2013) display about twice of those in 2004 (Dan et al., 2004).With the increase number of vehicles in Beijing (http://www.bjstats.gov.cn), the contribution of vehicles to the OC could not be ignored.
In terms of inorganic species, sulfate, nitrate and ammonium were the major ions, accounting for 90% of ions and 55% of PM (Table 1).The average mass concentrations of sulfate, nitrate and ammonium in PM 2.5 was 18 µg m -3 , 15 µg m -3 , 8 µg m -3 respectively (Table 2).They are similar as that reported by Wang et al. (2005) and Zhao et al. (2013).In summer, it is higher than other studies, but lower than others in winter.As the main precursor of sulfate, it was reported that 27% and 26% of SO 2 from industrial coal and coke combustion (Zhang et al., 2012a).Beijing stated that during the 11 th five-year (2006-2010) plan, both coal and natural gas have been developed for heating in the city, and modification of desulfurization and dedusting for boilers have been accelerated.Small coal-burning boilers have been replaced by electric boilers (Zhang et al., 2012c).These measures would work on reducing the local emission of SO 2 and result in lower sulfate in winter.But in summer, photochemical reaction, meteorological situation and regional transportation would attribute to the high concentration for this study.

Monthly Variations of Chemical Components in PM 2.5
Monthly cycles of chemical components concentration including OC, EC, POC, SOC and water soluble species of PM 2.5 are plotted in Fig. 4. The result shows the inorganic species such as sulfate, nitrate, ammonium and potassium appear higher levels in summer and lower levels in winter.However, OC, EC and chloride present higher levels in winter and lower levels in summer.In summer, high temperature and intensive radiation are beneficial for secondary aerosol formation from its precursors by gas-particle transformation, resulting in the high concentrations of these secondary inorganic species.On the other hand, higher concentration of nitrate would relate with the artifact of HNO 3 from the evaporation of nitrate ammonium under the high temperature.Many earlier studies have also observed higher inorganic species concentrations in summer at Xi'an and Beijing (Duan et al., 2006;Cao et al., 2012;Hu et al., 2014).The high loadings of OC, EC and chloride in winter are likely related with vehicles, combustion emissions and the low temperature.Coal combustion for heating in winter contributes more to the primary organic matter than other seasons.In addition, low temperature benefits the condensation of VOCs on the existed particles and enhances the gasparticle transformation process of organic matters.The fact that chloride levels are higher in winter and lower in summer may have been caused by higher emission from combustion in winter (Sun et al., 2013) and higher extent of chloride depletion and evaporation of ammonium chloride during hot seasons.In terms of SOC, higher concentrations were found in winter, which could be explained to some extent by that semi-volatile organics prefer to stay in particle phase due to low temperature.
Potassium is accepted as a marker of biomass burning (Cao et al., 2006).In our study, the potassium concentration showed the higher concentrations in June, which are in good agreement with the wheat harvest season in North China Plain (Qu et al., 2012).A previous study mentioned that the field open burning of wheat straws in the North China Plain during May-June 2006 influenced urban areas such as Beijing (Li et al., 2007).

Characterization of Chemical Components during Different Polluted Cases
In previous section, dramatic variations for mass concentration of PM and their chemical components were observed in our study.The differences on day by day basis are linked with sources emission, pollutant formation process, human activities, meteorological situation etc.To clarify the contributions of chemical components to these different polluted cases, three types of polluted cases were categorized based on the first grade (daily average < 35 µg m -3 ) and second grade (daily average < 75 µg m -3 ) of national air quality standard of PM 2.5 .High concentrations were selected from days with PM 2.5 concentration higher than 75 µg m -3 , and medium concentrations were days with PM 2.5 concentration higher than 35 µg m -3 and less than 75 µg m -3 .The days that PM 2.5 mass concentration less than 35 µg m -3 were defined as low concentrations.Table 3 summarized the average mass concentration of chemical components in the three types of polluted cases.
It is concluded that, during the study period, 74% of days were influenced by high concentration and 21% were at medium concentration, while only 5% days were at low concentration.The average mass concentrations of PM 2.5 in low, medium and high were 17 µg m -3 , 31 µg m -3 and 95 µg m -3 respectively.
The contributions of chemical components to PM 2.5 under high, medium and low concentration days are plotted in Fig. 5(a).The pie charts demonstrate that the unidentified parts of PM 2.5 on PTFE filters under high, medium and low concentrations was 14%, 16% and 21% respectively.In this study, OC, EC and water-soluble ions were analyzed, and to our knowledge, the unidentified parts should include oxygen in metal oxides from mineral aerosols, water vapor and some non-C atoms organic matters (including H, O, N atoms, etc.) excluding organic carbon.It was reported that mineral elements contribute less (0.5%) to PM 2.5 (Zhang et al., 2012b).The ratio of organic matter (OM) to organic carbon (OM: OC) could vary from 1.4 to 2.0 in Beijing (Xing et al., 2013).Provided that OM equal to twice of OC, 13%, 19% and 20% of PM 2.5 could be defined as non-C atoms organic matters.Based on this calculation, mass closures were obtained for high and low concentrations and 3% was overestimated for medium concentrations may come from the analysis errors.
For quartz filters were also used to report the mass concentration of PM 2.5 , water-soluble ions, EC and OC by some measurements, in order to evaluate the impact of water vapor on mass closures for quartz filters, the mass closures of PM 2.5 using all the data (mass concentration of PM 2.5 , water-soluble ions and EC OC) from quartz filters under high, medium and low concentrations were also calculated and plotted in Fig. 5(b).The result revealed that  about 35%, 48% and 36% of PM 2.5 was unidentified under high, medium and low concentration days.The hydrophilic nature of quartz filters and their wettability (Zdziennicka et al., 2009) facilitate the transfer of water molecules from the particles to the filters (Perrino et al., 2013), and the water vapor would contribute to the undefined mass.Comparing the undefined parts of quartz and PTFE filters, it infers that about 21%, 32% and 15% of PM 2.5 in quartz filters from water vapor during different polluted cases.And this number should be noted for the measurements using quartz filters.
In addition to the undefined parts in the filters, the chemical components displayed different performances in the three types of polluted cases.In the high concentrations, sulfate, nitrate, OC and ammonium ranked as the four highest components in the three types of cases, and the respective mass concentrations of these four chemical components were 23 µg m -3 , 19 µg m -3 , 20 µg m -3 and 11 µg m -3 .Ignoring the undefined parts, the total mass concentration of defined components (ΣDC) was 82 µg m -3 , and sulfate, nitrate, OC and ammonium account for 24%, 20%, 21% and 11% of ΣDC in the high concentration days.In the medium and low concentrations, the percentage of inorganic components to ΣDC decreases, while the OC increases, suggesting their different roles in the different types of cases.For PM 2.5 , similar results show that inorganic components play more important roles in polluted cases, while the role of OC increases in clean cases.

Aerosol Acidity, Hydrogen Ion (H + ) and Acidity Purity (f) in PM 2.5
Aerosol acidity influences aerosol hygroscopicity (Khlystov et al., 2005), and secondary aerosol formation.The ratio of total anions to total cations in PM is commonly used to evaluate the aerosol acidity (Kerminen et al., 2001;Zhang et al., 2008).If the cations concentration is significantly (25% or more) lower than the anions concentration, then a particle is considered highly acidic.If the ratio of cation to anion is 0.75, roughly 50% of SO 4 2-is considered to be in the form of bisulfate (HSO 4 -) and another 50% in the form of (NH 4 ) 2 SO 4 .Another method evaluating the extent of aerosol neutralization is to use the concentration of hydrogen ions (H + ) and acidity purity (f) (Hu et al., 2014).
The scatter plot between cations (NH 4 + , Na + , K + , Mg 2+ , Ca 2+ ) and anions (SO 4 2-, NO 3 -, Cl -) for PM 2.5 is shown in Fig. 6.There are two groups with different ratios of cations to anions, one follows line 1:1 and another tracks line 1:2, suggesting the samples for group 1 are neutral and that for group 2 are highly acidic.Two pie charts for chemical components of group 1 and group 2 are plotted in Fig. 6.Sulfate and nitrate contribute 31% and 28% to the PM 2.5 in group 2, but 22% and 17% in group 1.No obvious differences for ammonium and carbonaceous aerosols are observed between two groups.That means the acidic compounds in group 2 play significant roles on the acidity of PM 2.5 .
The daily values of the estimated hydrogen [H + ] and acid purity (f) in PM 2.5 are shown in Figs.7(a) and 7(b).During period 2 (1 May 2009-end), the concentration of [H + ] was higher than that period 1 (1 January 2008-30 April 2009), and the average concentration of [H + ] during period 1 and period 2 was 0.2 ± 0.3 µmol m -3 and 0.75 ± 0.5 µmol m -3 respectively.This [H + ] value in period 1 is comparable to the results from the semi-continuous measurement of Ambient Ion Monitor (AIM) (URG 9000D Series, USA) at the same site in Beijing (Hu et al., 2014), an urban site in Shanghai (Pathak et al., 2009).The [H + ] is nearly 2 times of that at Mount Tai, and almost ten times of those at some sites in the US (Zhang et al., 2007;Murray et al., 2009), South Korea (Lee et al., 1999) and Japan (Shimohara et al., 2001).Higher [H + ] couple with the lower ratio of cation to anion for period 2 may be related with sampling time, increasing emission and some meteorological situation comparing with more strict controlling measures for period covered Olympic Games in 2008.
On average, the calculated acidity purity (f) for period 1 and period 2 was 0.2 and 0.5 respectively.The different acidity purity suggest that sulfate may exist as a mix of (NH 4 ) 2 SO 4 and NH 4 HSO 4 in most bulk aerosols during different periods.It was presumed that when f = 0.5, the bulk aerosol existed as a mixture of chemical species including (NH 4 ) 2 SO 4 , NH 4 HSO 4 and H 2 SO 4 , and exhibits the properties of NH 4 HSO 4 .In period 1, PM 2.5 could be considered as letovicites with chemical formula (NH 4 ) 3 H(SO 4 ) 2 (Ziemba et al., 2007), and in period 2, PM 2.5 would be exist as NH 4 HSO 4 and H 2 SO 4 for the higher acidic environment.Fig. 6.Scatter plot of cations versus anions and chemical components in two groups, the uncertainty in the regression slope is at the 68% confidence interval.

SUMMARY AND CONCLUSION
In this study, PM 2.5 collected on quartz and PTFE filters were simultaneously by two side-by-side MiniVol sampling instruments during 2008-2009 in Beijing. in Beijing.Reasonable correlations between quartz and PTFE filters were found from the comparison of mass concentration and their chemical compositions for two types of filters.The water vapor and organic vapor in the atmosphere result in higher PM 2.5 concentration for Quartz filter.The concentrations of different ions on quartz filters were general lower than in PTFE filters, which was 0.77, 0.66, 0.76 and 0.56 of in PTFE filters for sulfate, nitrate, ammonium and chloride respectively.
The mass concentrations of PM 2.5 , their chemical components, monthly variation and their acidities are discussed.It was shown that PM 2.5 varied dramatically day by day throughout the entire study by about 3 orders.The average PM 2.5 was 79 µg m -3 which is about twice of national air quality standard of PM 2.5 .Water-soluble ions and OC were the major components, accounting for 75% of PM 2.5 .About 50% of OC were formed by secondary formation process.Seasonally, it was surveyed that the highest PM 2.5 always occurred in winter and lowest in spring and fall for the published works, but no obvious increasing or decreasing trends were found within the ten years.The OC concentrations display about twice of those in 2004 and the contribution of vehicles to the OC could not be ignored.Decrease trend of sulfate in winter was observed from the summary of previous works on PM 2.5 in Beijing, while higher concentration of sulfate, nitrate and carbonaceous aerosol were found in summer.
Significant monthly cycles of chemical components concentration were found.Inorganics species such as sulfate, nitrate, ammonium and potassium appear higher in summer and lower in winter, but OC, EC and chloride were higher in winter and lower in summer.
High, medium and low concentration cases were categorized based on the first grade of national air quality standard of PM 2.5 .It was shown that 74% of days influenced by high concentration and 21% were at medium concentration, while only 5% of points throughout the entire study were in low concentration.Mass closures for PTFE filters were obtained during different polluted cases, the non-C atom organic matter and water vapor influenced the undefined parts of quartz filters.The ratio of cations to anion was employed to evaluate the acidity of PM.Samples displayed neutral for period 1 and acidic for period 2. The average concentration of [H + ] in period 1 and period 2 was 0.2 ± 0.3 µmol m -3 and 0.75 ± 0.5 µmol m -3 respectively, and the acidic purity was 0.2 and 0.5.The controlling measures during the Olympic Games in 2008 and the meteorological situation result in the different acidity of PM 2.5 .

Fig. 2 .
Fig. 2.Comparison of water-soluble ions in quartz and PTFE filters, the uncertainties in the regression slopes are at the 68% confidence interval.

Fig. 3 .
Fig. 3. Time series of mass concentration of PM 2.5 and their chemical components.

Fig. 4 .
Fig. 4. Monthly variations of PM 2.5 and their chemical components.

Fig. 5 .
Fig. 5. Percentages of different chemical components in the three polluted cases a) for PTFE filters and b) for quartz filters.

Fig. 7 .
Fig. 7. Time series of H + and f in PM 2.5 .

Table 1 .
Summary of chemical components in PM 2.5 quartz filter from 1 January 2008 to 30 April 2009 (period 1).

Table 2 .
Summary for studies on PM 2.5 from this work and other research works.
BJ urban: sites in urban area of Beijing; BNU: Beijing normal university; CGZ: Chegongzhaung; CAMS: Chinese Academy of meteorological Sciences.

Table 3 .
Average mass concentrations of chemical components in high, medium and low PM 2.5 concentration cases.