Trichloroethylene decomposition and in-situ dry sorption of Cl-products by calcium oxides prepared from hydrated limes

A comparison of CaOs produced by calcining two types of hydrated lime and calcium carbonate was made for decomposition of trichloroethylene and in-situ dry sorption of the decomposed Cl-products using a lab-scale gas flow type tubular packed bed reactor. About 20 mg of CaO sample was mixed with about 2 g of Al₂O₃ particles and packed in the reactor and allowed to react with a flowing standard gas containing 500 ppm of C₂HCl₃ (N₂ balance) at 673 and 873 K, under the condition that the reaction of CaO with C₂HCl₃ might be completed within a few hours.It was found that no thermal decomposition of C₂HCl₃ at or below 673 K was observed in a reactor packed only with Al₂O₃ particles. However, a considerable amount of decomposition of C₂HCl₃ was obtained in a reactor packed with CaO and Al₂O₃, even at 673 K. For 1 mol of CaO prepared by calcining highly reactive Ca(OH)₂ at 673 K, decomposition of 0.42 mol of C₂HCl₃ and in-situ absorption product of 0.53 mol of CaCl₂ were obtained. At 873 K, about 46% of C₂HCl₃ was thermally decomposed. The total amount of C₂HCl₃ decomposed in CaO-Al₂O₃ particle bed at 873 K became nearly twice larger than that at 673 K. For 1 mol of CaO prepared by calcining highly reactive Ca(OH)₂ at 873 K, decomposition of 0.59 mol of C₂HCl₃ and in-situ absorption product of 0.67 mol of CaCl₂ were obtained. Small amounts of C₂Cl₂, C₂Cl₄, CCl₄, etc. were detected during decomposition of C₂HCl₃ at 673 and 873 K.It was recognized that the data on decomposition of C₂HCl₃ as well as in-situ dry sorption of Cl-products in CaO particle bed were correlated with specific surface area of the CaO employed.     

The air pollution caused by the burning of fireworks during the lantern festival in Beijing

The effects of the burning of fireworks on air quality in Beijing was firstly assessed from the ambient concentrations of various air pollutants (SO₂, NO₂, PM_2.5, PM₁₀ and chemical components in the particles) during the lantern festival in 2006. Eighteen ions, 20 elements, and black carbon were measured in PM_2.5 and PM₁₀, and the levels of organic carbon could be well estimated from the concentrations of dicarboxylic acids. Primary components of Ba, K, Sr, Cl⁻, Pb, Mg and secondary components of C₅H₆O₄²⁻, C₃H₂O₄²⁻, C₂O₄²⁻, C₄H₄O₄²⁻, SO₄²⁻, NO₃⁻ were over five times higher in the lantern days than in the normal days. The firework particles were acidic and of inorganic matter mostly with less amounts of secondary components. Primary aerosols from the burning of fireworks were mainly in the fine mode, while secondary formation of acidic anions mainly took place on the coarse particles. Nitrate was mainly formed through homogeneous gas-phase reactions of NO₂, while sulfate was largely from heterogeneous catalytic transformations of SO₂. Fe could catalyze the formation of nitrate through the reaction of α-Fe₂O₃ with HNO₃, while in the formation of sulfate, Fe is not only the catalyst, but also the oxidant. A simple method using the concentration of potassium and a modified method using the ratio of Mg/Al have been developed to quantify the source contribution of fireworks. It was found that over 90% of the total mineral aerosol and 98% of Pb, 43% of total carbon, 28% of Zn, 8% of NO₃⁻, and 3% of SO₄²⁻ in PM_2.5 were from the emissions of fireworks on the lantern night.     

The chemical composition of cloud and rainwater. Results of preliminary measurements from an aircraft

Using an aircraft, preliminary investigations have been carried out into the chemical composition of clouds. The emphasis was on testing the sampling devices under several conditions. The results clearly show a considerable spatial distribution in pollutant concentrations that must have been caused by the contribution of many plume sources and the inhomogeneity of cloudwater density as well. Rather high concentrations of H₃O⁺. Cl⁻, NO⁻₃, SO²⁻₄, NH⁺₄ and H₂O₂ were measured. Except for H₂O₂, these components are scavenged very fast during travel over short distances over source regions. Moreover H₃O⁺, NO⁻₃ and SO²⁻₄ are likely to be formed by chemical reactions of gaseous precursors with oxidants in the liquid phase. This may explain the marked decrease in high background H₂O₂ concentrations (ppm range). Relatively high (molar) NO⁻₃ /SO²⁻₄ ratios mainly varying from 1 to 2 were measured over The Netherlands near source areas. The lower values found during a flight over southern Scandinavia are in agreement with data from literature. Occasionally the chloride concentration was found to be much higher than could be expected from rain network data.     

Latitudinal distributions of trace gases in and above the boundary layer

Statistical analyses of global atmospheric concentrations provide evidence that C₂Cl₄, CHCl₃ and CH₃CCl₃ (methylchloroform) are more abundant in the tropical boundary layer than above it (α ? 0.09) by 27% (±27%), 21% (?21%, +12%) and 6.4% (±6%) respectively. The air samples on which these results are based were collected by cryogenic techniques during the June 1978 project GAMETAG flights and analyzed soon afterwards by gas chromatography (EC/GC and GC/FID), thus providing latitudinal concentrations of CO, CH₄, CCl₃F, CCl₂F₂, CH₃CCl₃ and light C₂-hydrocarbons, both in and above the boundary layer. In August 1980, after further development of analytical techniques, the stored air samples were re-analyzed to establish the latitudinal distributions of CH₃I, CHCl₃, C₂Cl₄, C₂F₃Cl₃ (F-113) and CHClF₂ (F-22) in and above the boundary layer. Stability studies, spanning a year, show that the concentrations of these gases do not change in the flasks.     

Alkyl nitrate photochemistry during the tropospheric organic chemistry experiment

Alkyl nitrates (C₁–C₅) were measured at two sites (near urban and rural) in southeast England during the Tropospheric Organic Chemistry Experiment (TORCH). Methyl nitrate was the dominant species during both campaigns accounting for on average about one third of the total measured alkyl nitrates. High mixing ratios (>50 pptv) and variability of methyl nitrate were observed at the near urban site (TORCH1) that were not seen at the rural site (TORCH2) and which could not be explained by local photochemical production or direct emissions. The diurnal variation of methyl nitrate during TORCH1 showed a morning maximum that would be consistent with nighttime chemistry followed by transport to the surface by boundary layer dynamics. Similarly, elevated morning mixing ratios were also observed during TORCH2 although the magnitudes were much smaller. As a result, methyl nitrate could represent a tracer for nighttime chemistry seen at the ground the following day. At both campaigns, the dominant source of short chain alkyl nitrates and carbonyl precursor radicals (≤C₄) were from decomposition of larger compounds. The magnitude of the source increased with decreasing carbon number consistent with increasing total precursor abundance. Non-photochemical emissions of acetaldehyde and acetone could not be accounted for by automobile exhaust emissions alone and indicated that other direct sources are likely important in this environment.     

Numerical Simulation of the Thermal Destruction of Some Chlorinated C1 and C2 Hydrocarbons

We have numerically modeled the breakdown of small quantities of several chlorinated hydrocarbons (CH₃CI, CH₂CI₂, CHCI₃, CCI₄, C₂H₃CI, and C₂H₅CI) in a lean mixture of combustion products between 800 and 1480 K. This simulates the fate of poorly atomized waste in a liquid-Injection Incinerator. Kinetics calculations were performed using the CHEMKIN and SENKIN programs, with a reaction mechanism that was developed at Louisiana State University to model flat-flame burner experiments.

A 99.99-percent destruction efficiency was attained In one second at temperatures ranging from 1280 to 960 K, with CCI4 requiring the highest temperature for destruction and C₂H₅CI the lowest. For all compounds except C₂H₅CI, there was a range of temperatures at which byproducts accounted for several percent of the elemental chlorine at the outlet. The more heavily chlorinated compounds formed more byproducts even though the amount of elemental chlorine was the same in all cases. The sensitivity of results to residence time, equivalence ratio, temperature profile, and the presence of additional chlorine, was examined for the case of CHCI₃.     

The Inhibition of Photochemical Smog

Additional inhibitors for the conversion of NO to NO₂ in C3H6—NO—0₂ irradiated mixtures have been tested at 25°C. These mixtures initially contained 16 mTorr C₃H₆, 8 mTorr NO, 0.012 mTorr NO₂, additive, and enough O₂ to bring the total pressure to 100 Torr. The NO₂ pressure was monitored photometrically. In the absence of additive, the NO₂ pressure first increases with irradiation time reaching a maximum conversion at about 15 minutes. As the irradiation time increases beyond 15 min, the NO₂ pressure drops. Before adding the inhibitors, runs were done with 10 Torr of CO added, and in these runs the conversion was speeded so that the maximum in NO₂ pressure occurred at 10 min. This enhancement in conversion rate is considered to be diagnostic for the presence of HO radicals. Next 10-min irradiations were done with various amounts of hexafluorobenzene (C₆F₆), nitrobenzene (C₆H₅NO₂), or naphtha lene (C₁₀H₈) added. The NO₂ pressure was reduced to one-half its value in the absence of inhibitor with 270 mTorr C₆F₆’, 220 mTorr C₆H₅N0₂, or 4 mTorr C₁₀Hg. The C₁₀H₈ is a very efficient inhibitor, but additions of up to 1 8.5 mTorr C₁₀H₈ did not reduce the N0₂ pressure to zero. Studies of the percent conversion of NO to NO₂ vs. irradiation time were done with either 4.2 mTorr C₁₀H₈ or 40 mTorr 2,6-di-ferf-butyl-4-methylphenol (Ph) added. In the former case the peak conversion was delayed from 15 to 22 min, while in the latter case no delay occurred. However, with the Ph added, there appeared to be some reduction in the maximum value of percent conversion.     

Variability of SO2, CO,and light hydrocarbons over a megacity in Eastern India: effects of emissions and transport

The Indo-Gangetic plain (IGP) has received extensive attention of the global scientific community due to higher levels of trace gases and aerosols over this region. Satellite retrievals and model simulations show that, in particular, the eastern part IGP is highly polluted. Despite this attention, in situ measurements of trace gases are very limited over this region. This paper presents measurements of SO₂, CO, CH₄, and C₂–C₅ NMHCs during March 2012–February 2013 over Kolkata, a megacity in the eastern IGP, with a focus on processes impacting their levels. The mean SO₂ and C₂H₆ concentrations during winter and post-monsoon periods were eight and three times higher compared to pre-monsoon and monsoon. Early morning enhancements in SO₂ and several NMHCs during winter connote boundary layer effects. Daytime elevations in SO₂ during pre-monsoon and monsoon suggest impacts of photo-oxidation. Inter-species correlations and trajectory analysis evince transport of SO₂ from regional combustion sources (e.g., coal burning in power plants, industries) along the east of the Indo-Gangetic plain impacting SO₂ levels at the site. However, C₂H₂ to CO ratio over Kolkata, which are comparable to other urban regions in India, show impacts of local biofuel combustions. Further, high levels of C₃H₈ and C₄H₁₀ evince the dominance of LPG/petrochemicals over the study location. The suite of trace gases measured during this study helps to decipher between impacts of local emissions and influence of transport on their levels.     

Do aerosols act as catalysts in the OH radical initiated atmospheric oxidation of volatile organic compounds?

Smog chamber/FTIR techniques were used to study the relative reactivity of OH radicals with methanol, ethanol, phenol, C₂H₄, C₂H₂, and p-xylene in 750 Torr of air diluent at 296±2 K. Experiments were performed with, and without, 500–8000 μg m⁻³ (4000–50 000 μm² cm⁻³ surface area per volume) of NaCl, (NH₄)₂SO₄ or NH₄NO₃ aerosol. In contrast to the recent findings of Oh and Andino (Atmospheric Environment 34 (2000) 2901, 36 (2002) 149; International Journal of Chemical Kinetics 33 (2001) 422) there was no discernable effect of aerosol on the rate of loss of the organic compounds via reaction with OH radicals. Gas kinetic theory arguments cast doubt upon the findings of Oh and Andino. The available data suggest that the answer to the title question is “No”. As part of this work the rate constants for reactions of OH radicals with methanol, ethanol, and phenol in 750 Torr of air at 296 K were determined to be: k_OH+CH3OH=(8.12±0.54)×10⁻¹³, k_OH+C2H5OH=(3.47±0.32)×10⁻¹² and k_OH+phenol=(3.27±0.31)×10⁻¹¹ cm³ molecule⁻¹ s⁻¹.     

Identification of sources of pollutants in precipitation measured at the mid-Atlantic US coast using potential source contribution function (PSCF)

Potential source contribution function (PSCF) was employed to study the source receptor relationships for 14 chemical species (Mn, SO₄²⁻, Zn, Al, Fe, Cu, Cr, Ni, Cd, NO₃, NH₄⁺, K⁺, Mg²⁺,and Pb) found in precipitation collected at Lewes, Delaware. This study identified areas of the Eastern United States as possible emission source areas that could have contributed to the 14 element concentrations observed at Lewes. The identified regions in the Eastern United States generally coincide well with known emission source areas. The likely emission sources for these chemical species include oil- and coal-fired power plants, incinerators, motor vehicles, and iron and steel mills.     

Photochemical production and loss of organic acids in high Arctic aerosols during long-range transport and polar sunrise ozone depletion events

Unique daily measurements of water-soluble organics in fine (<2 μm) and coarse (>2 μm) aerosols were conducted at Alert in the Canadian Arctic in winter to spring of 1992. They yield insight into photochemical production and loss of organics during long-range transport and ozone depletion events following polar sunrise. Comprehensive analyses of α, ω-dicarboxylic acids (C₂–C₁₂), ω-oxocarboxylic acids (C₂–C₉) and α-dicarbonyls (C₂, C₃) as well as pyruvic acid and aromatic (phthalic) diacid were conducted using GC and GC/MS techniques. Oxalic (C₂) acid was generally the dominant diacid species in both fine and coarse fractions, followed by malonic (C₃) and succinic (C₄) acids. Concentrations of total diacids in the fine aerosol fraction (0.2–64 ng m⁻³) were 5–60 times higher than those in the coarse fraction (0.01–3 ng m⁻³). After polar sunrise in early-March, the total concentration of fine aerosol diacids increased by a factor of 3–5 while the coarse mode did not change significantly. From dark winter to sunlit spring, temporal changes in correlations and ratios of these water-soluble organics to vanadium and sulfate measured simultaneously suggest that atmospheric diacids and related organic compounds are largely controlled by long-range atmospheric transport of polluted air during winter, but they are significantly affected by photochemical production. The latter can occur in sunlight either during transport to the Arctic or during photochemical events associated with surface ozone depletion and bromine chemistry near Alert in spring. Conversion of gaseous precursors to particulate matter via photochemical oxidation was intensified at polar sunrise, resulting in a peak in the ratio of total diacids to V. During ozone depletion events, complex patterns are indicated in photochemical production and loss depending on the diacid compound. Unsaturated (maleic and phthalic) diacids were inversely correlated with particulate Br whereas saturated diacids (C₂–C₄) positively correlated with particulate Br. These results suggest that Br chemistry associated with ozone depletion leads to degradation of unsaturated diacids and to the production of smaller saturated diacids.     

Quantification on source/receptor relationship of primary pollutants and secondary aerosols from ground sources—Part I. Theory

A new algorithm has been derived for trajectory models to determine the transfer coefficient of each source along or adjacent to a trajectory and to calculate the concentrations of SO₂, NO_x, sulfate, nitrate, fine particulate matter (PM) and coarse PM at a receptor. The transfer coefficient t_f (s m⁻¹) is defined to be the ratio between the contributed concentration ΔC (μg m⁻³) to the receptor from a ground source and the emission rate of the source q (μg m⁻² s⁻¹) at a grid, i.e. t_f≡ΔC/q. The model is developed by combining with a backward trajectory scheme and a circuit-type's parameterization. First, the transfer coefficients of grids along or adjacent a back-trajectory are calculated. Then, the contributed concentration of each emission grid is determined by multiplying its emission rate with the transfer coefficient of the grid. Finally, the concentration at the receptor is determined by the summation of all the contributed concentrations within the domain of simulation.     

Benzene in Auto Exhaust

The gas phase thermal decomposition rates of the C₁ and C₂-substituted peroxyacyl nitrates (RC(O)OONO₂), PAN (R = CH₃), PPN (R = C₂H₅) and vinyl-PAN (R = CH₂ = CH-) have been measured at ambient temperature (288 - 299 K) and 1 atm. of air. Our results for PAN (k = A exp (-E_a/RT), log₁₀ (A, s^-1) = 16.2 ± 1.6, E_a = 26.9 ± 2.1 kcal / mol, k₂₉₈ = 3.0 × 10^?4S^?1) are consistent with literature data. Thermal decomposition rates for PPN and vinyl-PAN are similar to that for PAN, with k₂₉₈ = 3.0 × 10^?4S^?1 for PAN, 3.4 × 10^?4S^?1 for PPN and 3.0 × 10^?4S^?1 for vinyl-PAN. Implications for the atmospheric persistence of PPN and vinyl-PAN as compared to that of PAN are briefly discussed.     

Biomass consumption and CO2, CO and main hydrocarbon gas emissions in an Amazonian forest clearing fire

Biomass consumption and CO₂, CO and hydrocarbon gas emissions in an Amazonian forest clearing fire are presented and discussed. The experiment was conducted in the arc of deforestation, near the city of Alta Floresta, state of Mato Grosso, Brazil. The average carbon content of dry biomass was 48% and the estimated average moisture content of fresh biomass was 42% on wet weight basis. The fresh biomass and the amount of carbon on the ground before burning were estimated as 528 t ha^?1 and 147 t ha^?1, respectively. The overall biomass consumption for the experiment was estimated as 23.9%. A series of experiment in the same region resulted in average efficiency of 40% for areas of same size and 50% for larger areas. The lower efficiency obtained in the burn reported here occurred possibly due to rain before the experiment. Excess mixing ratios were measured for CO₂, CO, CH₄, C₂–C₃ aliphatic hydrocarbons, and PM_2.5. Excess mixing ratios of CH₄ and C₂–C₃ hydrocarbons were linearly correlated with those of CO. The average emission factors of CO₂, CO, CH₄, NMHC, and PM_2.5 were 1,599, 111.3, 9.2, 5.6, and 4.8 g kg^?1 of burned dry biomass, respectively. One hectare of burned forest released about 117,000 kg of CO₂, 8100 kg of CO, 675 kg of CH₄, 407 kg of NMHC and 354 kg of PM_2.5.     

Deodorization of Dimethyl Sulfide Using a Discharge Approach at Room Temperature

Abstract

The traditional technologies for odor removal of thiol usually create either secondary pollution for scrubbing, adsorption, and absorption processes, or sulfur (S) poisoning for catalytic incineration. This study applied a laboratory-scale radio-frequency plasma reactor to destructive percentage-grade concentrations of odorous dimethyl sulfide (CH₃SCH₃, or DMS). Odor was diminished effectively via reforming DMS into mainly carbon disulfide (CS₂) or sulfur dioxide (SO₂). The removal efficiencies of DMS elevated significantly with a lower feeding concentration of DMS or a higher applied rf power. A greater inlet oxygen (O₂)/DMS molar ratio slightly improved the removal efficiency. In an O₂-free environment, DMS was converted primarily to CS₂, methane (CH₄), acetylene (C₂H₂), ethylene (C₂H₄), and hydrogen (H₂), with traces of hydrogen sulfide (H₂S), methyl mercaptan (CH₃SH), and dimethyl disulfide. In an O₂-containing environment, the species detected were SO₂, CS₂, carbonyl sulfide, carbon dioxide (CO₂), CH₄, C₂H₄, C₂H₂, H₂, formal-dehyde, and methanol. Differences in yield of products were functions of the amounts of added O₂ and the applied power. This study provided useful information for gaining insight into the reaction pathways for the DMS dissociation and the formation of products in the plasmolysis and conversion processes.     

Exposure to hazardous air pollutants along Oba Akran road, Lagos-Nigeria

We measured toxic air pollutants along Oba Akran road in Lagos to evaluate pedestrian exposure. PM₁₀, CO, O₃, NO₂, SO₂, CH₄, noise, wind velocity and temperature were measured simultaneously with portable analyzers. Our results showed that pedestrian exposure to PM₁₀ (with an average of 274.6 μg m⁻³ for all samples) and CO (with an average of 19.27 ppm for all samples) was relatively high. CO is a traffic-related pollutant, so the influence of the local traffic emissions on CO levels is strong. The high concentration of the PM₁₀ measured at the three environments also suggests that the traffic is a major source of ultrafine particles. The overall average concentrations for the 72-day experimental period for SO₂, NO₂ and O₃ are 101.2, 62.5 and 0.32 ppb respectively, all of which are below the US national ambient air quality standards. Strong traffic impacts can be observed from the concentrations of some of these pollutants measured in these three environments. Most clear is a reflection of diesel truck traffic activity rich in black carbon concentrations. The diurnal variation of O₃ and NO₂ also showed that NO₂ was depleted by photochemically formed O₃ during the day and replenished at night as O₃ was destroyed. A multivariate statistical analysis (Principal Component Analysis, Factor Analysis) has been applied to a set of data in order to determine the contribution of different sources. It was found that the main principal components, extracted from the air pollution data, were related to gasoline combustion, oil combustion and ozone interactions.     

C2–C5 Hydrocarbon measurements in the Netherlands 1981–1991

Measurements of C₂–C₅ hydrocarbons on an hourly basis at the TNO site in Delft from 1982 to 1984 and at Moerdijk over the period 1981–1991 are presented. In combination with meteorological data (wind direction and wind speed) the Delft and Moerdijk series are evaluated to identify source categories, annual variations, background concentrations and trends. The C₂–C₅ hydrocarbon concentrations at Delft and Moerdijk are determined mainly by emission characteristics and meteorological dispersion; the dominant sources are relatively nearby and atmospheric degradation is not of much importance. Under conditions of high wind speed the concentrations measured at Moerdijk in the marine sector are close to the Atlantic background concentrations in winter and somewhat above this in summer. The continental background concentrations are higher than the marine background concentrations by a factor of almost two. The annual variation of acetylene is more pronounced than that of the other hydrocarbons, most likely due to a different seasonal variation in acetylene emissions. The annual variation of propene is smoother, indicating stronger sources in summer than in winter. This feature of propene is observed in continental as well as in marine sectors. The observations show that at Moerdijk C₂–C₄ concentrations measured in Rijnmond sector have decreased considerably since the early 1980s, corresponding with changes in emissions in that area. Averaged over all wind directions the trend of all species is downward, but for acetylene the trend is significant at a 95% confidence interval. The acetylene concentrations show an annual downward trend of 3% during the 1980s, supporting other estimates of decreasing hydrocarbon emissions from traffic over this period at the same rate.     

Vertical profiles of tropospheric gases: Chemical consequences of stratospheric intrusions

The M.R.F. Hercules aircraft was used to obtain two vertical profiles of minor atmospheric constituents in a stationary low pressure circulation over the N. Atlantic. Results for CFCl₃, CF₂Cl₂, N₂O, C₂H₂, C₂H₆, C₃H₈, C₄H₁₀ and O₃ are presented, and interpreted by means of air parcel back trajectories on isentropic surfaces. Substantial evidence was found for air of stratospheric origin in the upper troposphere on the second flight, and the data from both flights point to an enhancement in OH concentrations, and thus hydrocarbon depletion rates, following mixing between stratospheric and tropospheric air.     

Effect of ethanol addition on soot precursors emissions during benzene oxidation in a jet-stirred reactor

A constant volume reactor model (PSR) was used to investigate the effect of ethanol addition on the formation of some pollutants during benzene oxidation in a jet-stirred reactor. The blended fuels were formed by incrementally adding 4 % wt of oxygen (ethanol) to the neat benzene fuel and by keeping the inert mole fraction (nitrogen) and the equivalence ratio constants. The main objective of this work was to obtain fundamental understanding of the mechanisms through which the oxygenate compound affects soot precursor amounts. The modeling results showed that C₂H₂, C₅H₅, and C₃H₃ mole fractions decreased upon increasing the ethanol percentage in the fuel mixture.     

Determination of sources of fine particles in different ambient atmospheres in South Korea using a chemical mass balance model

Abstract

To determine the sources of particulate matter less than 2.5?μm (PM_2.5 in different ambient atmospheres (urban, roadside, industrial, and rural sites), the chemical components of PM_2.5 such as ions (Cl^-, NO₃^-, SO₄^2-, NH₄⁺, Na⁺, K⁺, Ca²⁺, and Mg²⁺), carbonaceous species, and elements (Al, As, Ba, Cd, Cu, Fe, Mn, Ni, Pb, Se, V, and Zn) were measured. The average mass concentrations of PM_2.5 at the urban, roadside, industrial, and rural sites were 31.5?±?14.8, 31.6?±?22.3, 31.4?±?16.0, and 25.8?±?12.4?μg/m³, respectively. Except for secondary ammonium sulfate and ammonium nitrate, the model results showed that the traffic source (i.e., the sum of gasoline and diesel vehicle sources) was the most dominant source of PM_2.5 (17.1%) followed by biomass burning (13.8%) at the urban site. The major primary sources of PM_2.5 were consistent with the site characteristics (diesel vehicle source at the roadside site, coal-fired plants at the industrial site, and biomass burning at the rural site). Seasonal data from the urban site suggested that ammonium sulfate and ammonium nitrate were the most dominant sources of PM_2.5 during all seasons. Further, the contribution of road dust source to PM_2.5 increased during spring and fall seasons. We conclude that the determination of the major PM_2.5 sources is useful for establishing efficient control strategies for PM_2.5 in different regions and seasons.