Scholarly article on topic 'Physiochemical properties of carbonaceous aerosol from agricultural residue burning: Density, volatility, and hygroscopicity'

Physiochemical properties of carbonaceous aerosol from agricultural residue burning: Density, volatility, and hygroscopicity Academic research paper on "Earth and related environmental sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Atmospheric Environment
Keywords
{"Smoke particle" / Density / Volatility / Hygroscopicity / APM / V/H-TDMA}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — Chunlin Li, Yunjie Hu, Jianmin Chen, Zhen Ma, Xingnan Ye, et al.

Abstract Size-resolved effective density, mixing state, and hygroscopicity of smoke particles from five kinds of agricultural residues burning were characterized using an aerosol chamber system, including a volatility/hygroscopic tandem differential mobility analyzer (V/H-TDMA) combined with an aerosol particle mass analyzer (APM). To profile relationship between the thermodynamic properties and chemical compositions, smoke PM1.0 and PM2.5 were also measured for the water soluble inorganics, mineral elements, and carbonaceous materials like organic carbon (OC) and elemental carbon (EC). Smoke particle has a density of 1.1–1.4 g cm−3, and hygroscopicity parameter (κ) derived from hygroscopic growth factor (GF) of the particles ranges from 0.20 to 0.35. Size- and fuel type-dependence of density and κ are obvious. The integrated effective densities (ρ) and hygroscopicity parameters (κ) both scale with alkali species, which could be parameterized as a function of organic and inorganic mass fraction (f org & f inorg ) in smoke PM1.0 and PM2.5: ρ − 1 = f i n o r g · ρ i n o r g − 1 + f o r g · ρ o r g − 1 and κ = f i n o r g · κ i n o r g + f o r g · κ o r g . The extrapolated values of ρ inorg and ρ org are 2.13 and 1.14 g cm−3 in smoke PM1.0, while the characteristic κ values of organic and inorganic components are about 0.087 and 0.734, which are similar to the bulk density and κ calculated from predefined chemical species and also consistent with those values observed in ambient air. Volatility of smoke particle was quantified as volume fraction remaining (VFR) and mass fraction remaining (MFR). The gradient temperature of V-TDMA was set to be consistent with the splitting temperature in the OC-EC measurement (OC1 and OC2 separated at 150 and 250 °C). Combing the thermogram data and chemical composition of smoke PM1.0, the densities of organic matter (OM1 and OM2 correspond to OC1 and OC2) are estimated as 0.61–0.90 and 0.86–1.13 g cm−3, and the ratios of OM1/OC1 and OM2/OC2 are 1.07 and 1.29 on average, indicating more volatile organic materials have less density and lower OM/OC ratios in the external mixed smoke particles.

Academic research paper on topic "Physiochemical properties of carbonaceous aerosol from agricultural residue burning: Density, volatility, and hygroscopicity"

Contents lists available at ScienceDirect

Atmospheric Environment

journal homepage: www.elsevier.com/locate/atmosenv

Physiochemical properties of carbonaceous aerosol from agricultural residue burning: Density, volatility, and hygroscopicity

Chunlin Li a, Yunjie Hu a, Jianmin Chen a' *, Zhen Ma a, Xingnan Ye a, Xin Yang a, Lin Wang a, Xinming Wang b, Abdelwahid Mellouki c

a Shanghai Key Laboratory of Atmospheric Particle Pollution and Prevention (LAP3), Fudan Tyndall Center, Department of Environmental Science & Engineering, Fudan University, Shanghai, 200433, China

b State Key Lab of Organ Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou, 510640, China c Institut de Combustion, Aerothermique, Reactivite et Environnement, CNRS, 45071, Orleans cedex 02, France

CrossMark

HIGHLIGHTS

GRAPHICAL ABSTRACT

> Effective densities, volatility, and hygroscopicity of smoke particle were measured using V/H-TDMA-APM system.

Smoke particles were produced from filed burning simulation of crop straws with aerosol chamber system. Density, volatility, and hygroscopicity of smoke particles were measured using V/H-TDMA-APM system. Integrated density and hygroscopici-ty of smoke particles scale with the inorganic mass fraction. More volatile organic materials have less density and lower OM/OC ratios in the external mixed smoke particles.

ARTICLE INFO

Article history: Received 21 March 2016 Received in revised form 24 May 2016 Accepted 26 May 2016 Available online 27 May 2016

Keywords:

Smoke particle

Density

Volatility

Hygroscopicity

V/H-TDMA

ABSTRACT

Size-resolved effective density, mixing state, and hygroscopicity of smoke particles from five kinds of agricultural residues burning were characterized using an aerosol chamber system, including a volatility/ hygroscopic tandem differential mobility analyzer (V/H-TDMA) combined with an aerosol particle mass analyzer (APM). To profile relationship between the thermodynamic properties and chemical compositions, smoke PM10 and PM25 were also measured for the water soluble inorganics, mineral elements, and carbonaceous materials like organic carbon (OC) and elemental carbon (EC). Smoke particle has a

density of 1.1—1.4 g cmo3, and hygroscopicity parameter (k) derived from hygroscopic growth factor (GF) of the particles ranges from 0.20 to 0.35. Size- and fuel type-dependence of density and k are obvious. The integrated effective densities (p) and hygroscopicity parameters (k) both scale with alkali species, which could be parameterized as a function of organic and inorganic mass fraction (forg & finorg) in smoke PM1.0 and PM2.5: po1 = finorg■p°n1org + forg-porg and k = finorg-kinorg + forg$korg. The extrapolated values of pinorg and porg are 2.13 and 1.14 g cm 3 in smoke PM10, while the characteristic k values of organic and inorganic components are about 0.087 and 0.734, which are similar to the bulk density and k calculated from predefined chemical species and also consistent with those values observed in ambient air. Volatility of smoke particle was quantified as volume fraction remaining (VFR) and mass fraction remaining

* Corresponding author. E-mail address: jmchen@fudan.edu.cn (J. Chen).

http://dx.doi.org/10.1016Zj.atmosenv.2016.05.052

1352-2310/© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

(MFR). The gradient temperature of V-TDMA was set to be consistent with the splitting temperature in the OC-EC measurement (OC1 and OC2 separated at 150 and 250 °C). Combing the thermogram data and chemical composition of smoke PM10, the densities of organic matter (OM1 and OM2 correspond to OC1 and OC2) are estimated as 0.61-0.90 and 0.86-1.13 g cm~3, and the ratios of OM1 /OC1 and OM2/OC2 are 1.07 and 1.29 on average, indicating more volatile organic materials have less density and lower OM/OC ratios in the external mixed smoke particles.

© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Combustion of biomass has caught extensive concern among the scientific community as a major source of carbonaceous aerosol regionally and globally (Ram et al., 2011; Saikawa et al., 2009; Tian et al., 2008). These smoke aerosols have potential CCN (cloud condensation nuclei) activity, chemical reactivity, and inhalation toxicity that present great effect on the climate changes and public health (Bölling et al., 2009; Huang et al., 2014; Janssen et al., 2011; Rosenfeld, 2006; Shindell et al., 2012).

Biomass burning is an important global source of CCN, which eventually influence the hydrologic cycle, precipitation, and forcing balance (Ramanathan et al., 2001 ; Spracklen et al., 2011 ). However, highly variable CCN activity of source-dependent smoke particles challenges the climate modelling. And it was believed the CCN activity of smoke particles was response to the organic contents and also SOA (secondary organic aerosol) production upon photochemical aging, while organic compositions tightly relate to the source, burning condition, size, and extent of oxidation, which make the predication of CCN activity even complicated (Novakov et al., 1996; Engelhart et al., 2012; Giordano et al., 2013). Based on field investigation using aerosol mass spectrometer (AMS) and CCN counter, Gunthe et al. simplified the influence factors and established a parameterized function that linearly related the hygro-scopicity parameter and integrated inorganic and organic fractions of certain sized ambient aerosol (Gunthe et al., 2011; Rose et al., 2011), but the derived characteristic hygroscopicity parameter values of inorganic and organic fractions still need to be proved by detailed chemical compositions.

Smoke particle density as a function of individual compound density affect the estimation of bulk hygroscopicity parameter and refractive index of OC and EC content (Petters et al., 2007; Schmid et al., 2009; Schkolnik et al., 2007). Besides, density bonds the aerodynamic diameter and mobility diameter, and it also plays a vital role in the mass concentration conversion and mass closure calculation (Beddows et al., 2010; Khlystov et al., 2004). Thus, value of density applied in the research will influence the result of online instruments like aerosol time-of-flight mass spectrometer (ATOFMS), AMS, and aerodynamic particle sizer (APS) in particle mass spectra analysis and size distribution measurements. However, few study ever reported the densities of smoke particles, and the limited densities of associated aerosol from APM measurement, from size conversion extrapolation based on APS-SMPS (scanning mobility particle sizer) data, and from bulk density calculation according to chemical composition all vary widely. And most study applied hypotheses-driven density values range from 1.0 to 1.7 g cm~3 without considering size difference or atmospheric aging effect (Engelhart et al., 2012; Gunthe et al., 2011; Reid et al., 2005; Schmid et al., 2007; Rissler et al., 2006).

Hygroscopicity and density are both result of chemical composition and mixing state of particles. Volatility or thermostability analysis helps get insight into the mixing sate and aging degree of carbonaceous aerosol (Pratt et al., 2009). Besides, volatility is a key

property in the phase partition and secondary organic aerosol formation of the organic components (Grieshop et al., 2009; Tritscher et al., 2011). Volatility of aerosol has been commonly applied in thermo-optical method based OC-EC measurement and organic chemical compound analysis with thermal-denuder mass spectrometer (Pratt et al., 2009; Seinfeld et al., 2012). However, seldom research ever conducted to investigate the volatility of fresh smoke particles (Grieshop et al., 2009; Pratt et al., 2009), and it is fundamental to study the photochemical oxidation and atmospheric aging of biomass burning aerosol.

Herein, V/H-TDMA-APM system was coupled to an aerosol chamber to provide size- and fuel type-resolved hygroscopicity, effective density, and volatility of fresh smoke particles. The relationship among hygroscopicity, density, and chemical compositions of smoke particles were elaborated using Kohler theory analysis. Combining V-TDMA-APM data and OC-EC fractions of smoke particles, density of OM and OM/OC factors were estimated.

2. Experiment and method

2.1. Smoke particle production and aerosol chamber system

Physical and chemical properties of fresh smoke particles were characterized by injecting the emissions from crop residues burning into the dark aerosol chamber system (Fig. 1). The chamber is capsule-like stainless tank of 4.5 m3 with 0.3 mm Teflon coating at the inner face, details can be found elsewhere (Zhang et al., 2008,

Fig. 1. Schematic graph of experimental setup. a) Flow chart of experiment, b) Detailed setup of thermodynamic properties measurement of aerosol including hygroscopicity, volatility, and density.

2011). Before each experiment, the chamber was scrubbed using aqueous ethanol (50 vol%), followed by continuous flushing with dry, HEPA-filtrated air (~300 L/min) for 3 h, then it was oxidized by high concentration ozone (~3 ppm) for 12 h, finally the chamber was purged and vacuumized for use.

The biomass fuels included wheat, corn, rice, soybean, and cotton residues which represent major agricultural residues in China. 10.0 g of each fuel was cut up and dehydrated at 100 °C for 24 h, and then it was burned in a sealed combustion stove with HEPA-filtrated air supply and with only one exhaust vent connecting to the chamber. The emissions were aspirated into the chamber till room pressure. Temperature and relative humidity (RH) in the chamber were maintained within 19.5-23.5 °C and 55%-65% RH, which were tracked using a hygroclip monitor (model IM-4, Rotronic). Modified combustion efficiency (MCE) was monitored with CO and CO2 measuring using a gas-chromatograph (GC, Model 930, Shanghai Hai Xin Gas Chromatograph Co., LTD). 4 tests for each agricultural fuel burning were repeated. At each experiment, filter sampling, MCE, and physiochemical properties measurements were conducted. A nitrogen dilution system (dilution ratio, nitrogen: sample flow z 5:1) was fixed ahead ofV/H-TDMA-APM system before the density, hygroscopicity, and volatility measurements. MCE for all the burnings was 0.89-0.96, indicating flaming phase dominated, which were comparable with that of field burnings (Ferek et al., 1998; Li et al., 2007; Reid et al., 2005).

2.2. Chemical composition

PM1.0 and PM2.5 samples were collected onto 90 mm quartz filter (Tissuquartz, Pall Corp., USA) from the aerosol chamber using a high-volume Particle Sampler (HY-100, Qingdao Hengyuan S.T. Development Co., Ltd) operating at 100 L min-1 for 5 min. The quartz filters were prebaked for 8 h at 450 °C to eliminate contamination. Before and after sampling, the filters were weighted using a balance (Sartorius BP211D) with an accuracy of 10 mg, and the balance was treated in an electronic desiccator (40% RH, 22 °C) for 24 h before its use. After weighting, the loaded filters were stored at -20 °C in a refrigerator for further analysis.

1/4 of each filter was extracted with 25 mL MiliQ-water for water soluble inorganic species analysis. Water soluble ions (F-, Cl-, NO-, NO-, SO2-, Na+, NHj, K+, Ca2+, Mg2+) were measured using Ion Chromatography (IC, Model 850 Professional IC, Metrohm, USA) consisting of a separation column (Metrosep A Supp 7250/4.0 for anion and Metrosep C-4150/4.0 for cation). Another 1/ 4 of each filter was acid digested (high concentrated HF: HNO3: HClO4, with a mixture of 3:1:1 mL, kept at 170 °C for 4 h in high-pressure Teflon digestion vessel), then diluted by MiliQ-water to be 25 mL for element measurement. 6 USEPA (Environmental Protection Agency of the United States) Environment priority controlled elements (As, Zn, Pb, Cd, Cr, and Ni) and two other metals (V and Al) were quantified using Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES, Atom Scan, 2000; JarroU-Ash, USA).

Carbonaceous materials (OC and EC) were measured using a thermal-optical transmittance OC/EC analyzer (Sunset laboratory Inc., Forest Grove, OR) with modified NIOSH 5040 protocol and produced four OC fractions (OC1, OC2, OC3, and OC4 at 150, 250, 450, and 550 °C respectively), OP fraction (a pyrolyzed carbonaceous component determined when transmitted laser returned to its original intensity after the sample was exposed to oxygen), and three EC fractions (EC1, EC2, and EC3 at 550, 700, and 800 °C, respectively). OC is defined as OC1+OC2+OC3+OC4+OP and EC is defined as EC1+EC2+EC3-OP (Chow et al., 1993; Seinfeld et al.,

2012). The quality of the data above was guaranteed by standard materials calibration, recovery rate, and operational blank

correction.

2.3. Size distribution

Size distribution was measured using a wide-range particle sizer (WPS, Model 1000XP, TSI Inc., USA), which has been described elsewhere (Zhang et al., 2011; Li et al., 2015). WPS integrates the functions of scanning mobility sizer (SMPS) and laser particle spectrometer (LPS), SMPS part classifies particles with mobility diameters from 10 nm to 500 nm, and LPS part scan aerodynamic particles from 350 nm to 10 mm, the effective density and refractive index were set as 1.3 g cm-3 and 1.45. Before experiment, WPS was calibrated with polystyrene latex sphere particles (PSLs, 0.04, 0.08, 0.22,0.70, and 1.50 mm, 1.05 g cm-3, Duke Scientific, USA) generated from an atomizer (Model 3076, TSI Inc., USA) that was purged with high purity N2.

2.4. Effective density

Effective density of dehydrated fresh smoke particle in mobility diameter of 100-400 nm was characterized using the TDMA-APM system. The performance of TDMA has been described in details before (Hu et al., 2011; Zhang et al., 2011). An APM (Model 3601, Kanomax Inc.) is combined to classify aerosol particles according to the mass-to-charge ratio, and see the classification principle in the work of Tajima et al. (2011). Briefly, smoke particle from the chamber is neutralized with 210Po, and subsequently dehydrated through diffusion dyer and a nafion tube, then size-selected by the first DMA (DMA1, Model 3081, TSI Inc., USA). Afterwards, the monodispersed particles are introduced into the APM and a CPC (Model 3771, TSI Inc.) at the downstream. Finally, mass distribution is obtained by voltage scanning and particle counting. The effective density can be calculated by the following equation:

pi = V =

pn3 6ni

Where pi and m; are size-specific density and mass of particle at mobility diameter Di.

Before measurement, TDMA-APM system was calibrated using PSLs (40, 80, and 220 nm, 1.05 g cm-3, Duke Scientific) and (NH4)2SO4 particles (material density of 1.77 g cm-3). The average effective densities of PSLs and (NH4)2SO4 particle were measured as 1.07 ± 0.01 and 1.73 ± 0.03 g cm-3, in good agreement with material densities.

2.5. Hygroscopicity

Hygroscopicity of 50-300 nm smoke particles at 80% RH was studied using H-TDMA, which measures the new size distribution of monodispersed particles after hygroscopic growth (Ye et al., 2013; Hu et al., 2011). Briefly, dehydrated and charged particles is first selected by DMA1, then introduced into the humidifier (multitube Nafion humidifier, Model PD-70T-24ss, Perma Pure Inc., USA) at defined RH to gain a new diameter when water vapor condense on or be absorbed, the new size distribution will be finally measured by DMA2 (Model 3081L,TSI Inc., USA) and CPC. And hygroscopic growth factor (GF) of particle is defined as the ratio of the humidified diameter at a certain RH to its initial dried diameter, mathematically described below:

GFi(RH) =

Where Di is initial dried particle diameter selected from DMA1, DHi

is new diameter of humidified particle.

In this study, approximately 2.5 s is allowed for the particle to be exposed in the humidifier, which is sufficient to get hygroscopic equilibrium (Hu et al., 2011). H-TDMA operates at a sampling flow and sheath flow of 0.3 and 3.0 L min-1,122 s is needed for each test loop including up and down scanning of DMA2-CPC. The function of DMAs is checked with PSLs (40, 80, and 220 nm, Duke Scientific, USA) periodically, and the accuracy of RH controller and RH sensor is calibrated by measuring deliquescence RH (DRH) and subsequent hygroscopic growth of (NH4)2SO4 particle prior to the test (Ye et al., 2009).

growth/shrinking factor probability distribution function (PDF), thus an inversion algorithm of multimodal implementation developed by Gysel et al. (2009) was applied to best fit the size distribution of DMAs output to derive the proper peak value of the inversed distribution, details of the method referring to our previous work (Yin et al., 2015; Ye et al., 2013).

3. Results and discussions

2.6. Volatility

Measurement of volatility is alike that of hygroscopicity. The conditioned aerosol flow from DMA1 was switched to a heater rather than the humidifier to pyrolyze the particles. The heater in V-TDMA system is custom-built and consists of an inner linear stainless tube (DN8, 30 cm) surrounded by heating ceramic jacket. The temperature sensor fixed in the center of the inner tube controls the temperature in a range from 25 to 350 °C(±1.5 °C). Particle may shrink as the contained volatile and thermal unstable species evaporate or decomposite during passing through the heater. A charcoal denuder and 1.0 m brass tube are equipped at the downstream of heater to remove the gaseous evaporated materials and cool down the aerosol before entering DMA2 and CPC. The calculated flow retention time (RT) in the heater is ~3.0 s at a sampling flow rate of 0.3 L min-1, which is comparable to many studies (Jonsson et al., 2007; Meyer et al., 2009). Long RT will ensure the evaporation equilibrium of the particle studied. Volatility of particles can be described as shrink ratio (named as shrinking factor, SF) of diameter at predefined temperature:

SF(T)=DDf

Where D; is initial diameter of dehydrated particle, DV is shrinked diameter of particle exposed at temperature T.

Apart from SF, volume fraction remaining (VFR) is commonly used to represent aerosol volatility:

Dvi Di

To further analyze the density of the remained and evaporated fraction of particle, APM was also equipped at downstream of the heater to measure mass distribution of the pyrolyzed particle (mVi). Combining the initial mass determined by TDMA-APM system, mass fraction remaining (MFR) can also deduced by the equation below:

Volatility of smoke particle from 100 to 400 nm was measured at temperature from 50 to 250 °C with temperature gradient of 50 °C. And before the measurement, (NH4)2SO4 particle was used to test the function of V-TDMA, the result was displayed in Fig. S1 in supporting information (SI).

To the measurement of hygroscopicity and volatility in term of GF and SF of smoke particles, DMA1 selects size-defined particles in a quasi-monodispersed distribution, followed by humidification or thermo-volatilization, DMA2 discretely steps through voltage to measure the new size distribution, due to the finite width of the size increments of the DMAs, the measured distribution function (MDF) is a skewed and smoothed integral transform of the actual

3.1. Chemical profiles

Agricultural residue burnings release a large amount of ultrafine particles with peak diameter ranges from 100 to 300 nm (Fig. S2), which has been confirmed by many studies (Reid et al., 2005; Zauscher et al., 2013; Li et al., 2015). From chemical profiles in Fig. 2, the compositions of the emitted smoke particles are overwhelmingly dominated by organic components, which make up over 70 wt% of PM10 and PM25 (detailed chemical compositions and emission factors in Table S1-S3, SI). Organic matter (OM) was converted from OC by multiplying a factor of 1.3 to account for the non-carbon mass (Chow et al., 1993; Li et al., 2007). Organic composition and OC/EC ratio of smoke particles are fuel type-, burning condition-, extent of oxidation-, and also particle size-dependent (Andreae et al., 2001; Grieshop et al., 2009; Reid et al., 2005). Smoldering produces more organic materials with relative high OC/EC ratio. Under the flaming combustion condition in this study (MCE = 0.89—0.96), organic mass proportions and OC/EC ratios show small variations in agricultural residues and samplings of PM2.5 and PM1.0. Average OC/EC ratio and OM mass fraction are 3.58 and 57.6 wt% in smoke PM10, and 3.82 and 60.2 wt% in PM2.5, which are comparable with previous studies as summarized in Table 1. The EC was further classified as char-EC and soot-EC based on different thermodynamic/optical properties (Reid et al., 2005). Han et al. (2007, 2009) ever differentiated char-EC from EC measurement as EC1-OP, and defined soot-EC as EC2+EC3. As OC/EC is sensitive to mass transfer of organic aerosol including SOA formation and evaporation/condensation of organic material during atmospheric transmission (Grieshop et al., 2009; Cao et al., 2005; Cubison et al., 2011; Jolleys et al., 2012; Tritscher et al., 2011), Char-EC/soot-EC ratio can be a more effective indicator for source identification of carbonaceous aerosol (Han et al., 2009). Herein, char-EC/soot-EC ratios of fuel-specific emissions in PM2.5 and PM1.0 were calculated and given in Table 1. Average char-EC/soot-EC ratio is 7.82 and 6.29 in smoke PM2.5 and PM1.0, confirming soot particles is mainly within micrometer (China et al., 2014; Wilson et al., 2013).

Smoke particles emitted from crop residues comprise 16—33 wt % water soluble inorganics, which also vary with fuel types and PM fraction. Smoke PM1.0 comprises more alkali species with average mass fraction of 24.6 wt%, and it is 24.6 wt% in PM2 5. K+ and Cl3 are the primary ions. Mean charge ratios of Cl3: K+ are 1.46 and 1.50 in PM1.o and PM2 5. Field and laboratory studies both found the ratio of Cl3: K+ decreased with aging of smoke particles once released in the atmosphere, as chloride was displaced by sulfate and nitrate (Li et al., 2003; Li et al., 2015). Charge ratios of major cations (K+ + NHj) to anions (Cl~ + SO^3 + NO") are 1.03 and 0.99 in PM1.0 and PM2.5 on average. 8 quantified mineral elements account 0.1 —0.9 wt% of smoke particles, and the mineral elements reside more in PM2 5 (details in Table S1-S2, SI). As is the second largest metallic emission followed by Cr, Ni, Zn from agricultural residue burning, which present potential carcinogenic risk and impose a long term burden on biogeochemical cycling (Duan et al., 2013; Wu et al., 2011).

Wheat straw Corn straw Rice straw Cotton residue soybean residue

■ OM ■ EC »None ■ THM «S042- «CI- ■ N03- «NH4+ - K+ ■ OWSI

Fig. 2. Chemical profiles of smoke PM25 and PM10 from 5 types agricultural residue burnings. (OM, organic matter = 1.3 x OC; OWSI, other water soluble inorganic species include F-, NO-, Na+, Mg2+, and Ca2+; None means mass difference from filter sampling weight and measured chemical materials, which can be attributed to uncertainty in OM conversion from OC).

Table 1

Summary of fuel-specific carbonaceous emissions (mean value ± standard deviation).

Fuel MCEa Emission factorb OC (g/kg) OC:EC Char-EC: Soot-EC Method Reference

Wheat straw 0.91 2.8 4.2 8.7 TOT OC/EC analyzer This study

Corn straw 0.93 2.4 3.1 6.0

Rice straw 0.89 6.9 3.2 4.2

Cotton residue 0.93 7.4 6.2 9.0

Soybean residue 0.96 1.5 2.5 11.2

Average 0.92 4.2 3.8 7.8

Yellow pine 0.88 - 2.2 - AMS Grieshop et al., 2009

Wheat residue 0.94 1.9 5.4 - TOT OC/EC analyzer Dhammapala et al., 2007

Crop residues 0.89 2.1 2.7 - TOT OC/EC analyzer Turn et al., 1997

Wheat residue 0.92 2.7 5.5 - TOT OC/EC analyzer Li et al., 2007

Grass 0.93 7.0 11.2 - TOT OC/EC analyzer Aurell et al., 2015

Crop residues 0.92 4.0 7.1 - TOT OC/EC analyzer Hayashi et al., 2014

Wood >0.90 - 0.5-2.0 -24.0 TOT OC/EC analyzer Arora et al., 2015

a MCE = Modified Combustion Efficiency (Reid et al., 2005). b Emission factor of OC in smoke PM2.5.

3.2. Effective density

Effective density of size-classified particle was derived from Gaussian model fitting of each density scan from DMA-APM-CPC, the peak mode was selected as effective density (fitting method in Fig. S3, SI). However, bimodal density distributions were commonly observed, and the less mode peaks around 1.3 g cm~3 was believed to be effective one, while the larger mode over 2.5 g cm~3 should be artifact data generated from multiple charged particles (Yin et al., 2015). Size-resolved effective density of smoke particle was calculated and presented in Fig. 3a, the densities range from 1.1 to 1.4 g cm~3, and size- and fuel issue-dependence are obvious. To cotton, rice, and wheat residue burning particles, dry effective density increases with particle size, while it present inverse trends for the rest residues burning particle. Density is the result of chemical composition, mixing state, and also morphology of particles (Reid et al., 2005; Katrib et al., 2005; Lewis et al., 2009). To eliminate the size effect, ensemble density of agricultural residue burning smoke can be recalculated by the equation below:

YÂWPi TÂW

The calculated ensemble density of cotton, soybean, corn, rice, and wheat residue burning smoke is 1.317 ± 0.013, 1.194 ± 0.016, 1.346 ± 0.015,1.301 ± 0.035, and 1.286 ± 0.033 g cm-3, respectively. The densities are comparable to that from Levin et al. they estimated bulk densities of various biomass burning aerosol to be 1.22-1.92 g cm-3 based on chemical compositions (Levin et al., 2010). Effective density of smoke particle is similar to the density of urban aerosol, implicating the difficulty in discriminating biomass burning source from ambient density measurement (Yin et al., 2015). Instant mass concentration of smoke particle in the aerosol chamber can be derived from size distribution in combination of ensemble density, combining chamber volume and biomass fuel that burned, PM emission factors of agricultural residue burning can also be estimated (Fig. S4, SI). Based on size distribution and effective density, the recalculated emission factors of PM1.0 from cotton, soybean, corn, rice, and wheat residue burning are 11.884 ± 0.587, 3.433 ± 0.092, 6.601 ± 0.147,12.987 ± 0.403, and 6.164 ± 0.198 g kg-1, respectively. The results are consistent with that from filter sampling weight method (SI), implying that emission factor and density measurement are reliable in this study.

Form Fig. 3b, ensemble density scales with inorganic mass proportion in smoke particles, the function of p-1 against inorganic mass fraction in PM1.o and PM2.5 can be rearranged as below:

where p is ensemble density of smoke particles, Ni is number concentration of smoke particle at mobility diameter Di.

0.20 0.25 0.30 Inorganic mass fraction

Fig. 3. Smoke particle density analysis. a) Size-resolved effective density of particle from 100 to 400 nm, b) Ensemble mean densities of different agricultural residue burning particles as a function of inorganic mass fraction in PM10 and PM2 5 (error bar means standard deviation).

1 finorg forg p Pinorg Porg

PM10 - = 0.470 X finorg + 0.878 x forg — ^ +114 (7)

pinorg and porg are bulk effective density of particulate inorganics and organics.

Therefore, bulk density of inorganics and organics can be deduced. In smoke PM10, pinorg and porg are 2.13 ± 0.44 g cm-3 and 1.14 ± 0.13 g cm-3, while in PM25, pinorg and porg are 2.42 ± 0.53 g cm-3 and 1.13 ± 0.13 g cm-3. It is reasonable that pinorg is larger in PM2.5 than in PM1.0, as more mineral elements with larger densities reside in PM2.5. Of the inorganic fraction in PM1.0, average equivalent molar charge ratio of SO4-:Cl-:K+:NH;f is about 0.18:0.79:0.56:0.44 from chemical analysis, and the defined KCl—K2SO4—NH4Cl-(NH4)2SO4 mixture have a bulk density of 1.83—1.91 g cm-3estimated from densities of individual species that were summarized in Table 2. Bring mineral elements and other alkali species (Na+, Ca2+, Mg2+, NO-, and F-) into the analytical system, extrapolated pinorg may deposit in the calculated bulk density of inorganic components. Integrated porg has no significant difference in PM1.0 and PM2.5. Organics in smoke aerosol are mixture of BC and OM. Although BC has a material density of ~2.0 g cm-3, the effective density of BC is much smaller and varies with source, particle size, microstructure, mixing state, and the degree of aging (China et al., 2014; Pagels et al., 2009; Rissler et al., 2014; Zhang et al., 2008). The density of fresh porous soot was reported to be 0.1—1.1 g cm-3, driving/engine load and atmospheric aging will alter the fractal dimension to make it more compact and dense (China et al., 2014). Schmid et al. assessed density of EC and OM in smoke particle to be 1.8 and 1.4 g cm-3 by constraining these densities to best fit their optically and chemically derived refractive indices (RI), while Schkolnik et al. applied density of EC and OM in smoke aerosol to be 1.8 and 0.9 g cm-3 to best fit their RI (Schkolnik et al., 2007; Schmid et al., 2009). The density of OM also varies from source, size, extend of oxidation (O:C ratio), chemical composition, and morphology (Grieshop et al., 2009; Katrib et al., 2005). Density of OM in fresh biomass burning smoke has a vast range from 0.8 to 1.5 g cm-3, depending on burning phase and fuel issues, and the suggested value is 1.1—1.3 g/cm3 (Cross et al., 2007; Chen et al., 2009). From chemical analysis above, average mass ratio of OC/EC in smoke particle is 3.6 in PM1.o and 3.8 in PM2 5, take density of EC as 1.8 g cm-3, then density of OM can be calculated by splitting the bulk density of porg, and it is 1.06 ± 0.14 g cm-3 in PM10 and 1.05 ± 0.14 g cm-3 in PM2.5, in agreement with the result of many studies (Schkolnik et al., 2007; Cross et al., 2007; Chen et al., 2009).

PM2.5 p = 0.414 x finorg + 0.886 x forg

finorg + forg — 1

finorg forg

2.42 + 1.13

forg is mass fraction of organics in smoke aerosol including organic matter (OM) and elemental carbon (EC). finorg is the rest of mass fraction in smoke particles including water-soluble inorganic salts, trace mineral metals, and some other materials accounting for the mass difference between filter weighting and chemical measurement.

As ensemble density of aerosol is response to bulk densities of the contained inorganics and organics, which can be calculated as below:

3.3. Volatility and thermostability

Fig. 4a illustrates ensemble VFR, MFR, and effective density of the nonvolatile residues as a function of temperature (25—250 °C) for all 5 types agricultural residue burning smoke particles (detailed size-resolved VFR and effective density at specific temperature in SI, Figs. S5 and S6). Integrated VFR and MFR were calculated via equations below:

EnNi x D3i

pnNi x D3

PnNi x mvi

EnNi x mi

Density of thermostabilized particles at fixed temperature T can be estimated according to equation (13):

Table 2 Thermodynamic parameter of individual species (room temperature, 1 atm).

Species Weighta Density (g cm 3)a DRH (%)b Pyrolysis Temp (oC)a,b kc

KCl 74.55 1.99 84.2 - 0.3 >300 >0.65

K2SO4 174.24 2.66 97.0 >300 N/A

KNO3 101.10 2.11 92.0 N/A N/A

NH4Cl 53.49 1.53 80.0 <150 1.75

(NH4)2SO4 132.14 1.77 79.9 0.5 150-280 0.53

NH4NO3 80.04 1.72 61.8 <150 0.67

NaCl 58.5 2.17 75.3 0.1 >300 1.27

Na2SO4 142.06 2.68 84.2 0.4 >300 0.68

NaNO3 85.00 2.26 74.3 0.4 N/A 0.74

CaCl2 110.98 2.15 30.0 N/A 0.44

MgCl2 95.21 2.36 33.0 N/A 0.44

HOOC-COOH 90.04 1.63 97.8 <150 0.48

CH3COOH 60.05 1.05 N/A <75 N/A

CH3SO3H 96.10 1.48 N/A 171 N/A

DRH (deliquescence relative humidity), N/A means not available from literature. a Chemical Index Database (http://www.chemicalbook.com).

b Tritscher et al., 2011; Freney et al., 2009; Park et al., 2009; Seinfeld et al., 2012; Tiitta et al., 2010; E-AIM model III (http://www.aim.env.uea.ac.uk/aim/aim.php). c Park et al., 2009; Freney et al., 2009; Petters et al., 2007; Tang et al., 1993.

Pt = p0 x

(13) OMi/OCi

where po is initial density at room temperature 25 °C.

A gradual decline of VFR and MFR for smoke particle is observed. There is no significant difference in thermostability except the cotton residue burning particle with larger volatility. After exposed at 250 °C, over 50 vol% and 40 wt% of smoke particles lost via evaporation or decomposition. On the contrast, densities of refractory particles increase substantially with temperature increase, implicating the particulate remainders have larger density and lower volatility. According to chemical composition and thermal properties of individual materials in Fig. 1 and Table 2, NH4NO3, NH4Cl, and (NH4)2SO4 will decomposite at 250 ° C, while thermostable materials like KCl, EC, mineral elements, and some other refractory organics may remain in the shrinked smoke particles. In field research, 250-300 °C is critical temperature to distinguish volatile or non-volatile materials of the aerosol, and the retained non-volatile particulate material at such temperature is considered to be mostly soot in polluted urban air (Wehner et al., 2002, 2009).

Error function was empirically used to fit the mean VFR, MFR, and density changes of all the smokes, the functions were only valid within 250 ° C. It was set 1.0 to be the initial value to VFR and MFR, and 1.3 g cm-3 as initial value to density changes. They have the following equations with T (unit: oC) as the heater temperature:

VFR = 0.986 — 0.873 x erf

MFR = 0.992 — 0.767 x erf

T - 25

T - 25 494.3

p = 1.282 + 1.073 x erf

t - 25 3774

These empirical functions simply represent the observations in the most suitable way, which help gain the information on the complex chemical nature of smoke particles. The gradient temperature (150 and 250 °C) in volatility analysis was intentionally set to be consistent with the splitting temperature of OC1 and OC2, thus densities and OM/OC factors of OC1 and OC2 can be furtherly estimated based on equations (17) and (18):

DMFRi - £ mfk

POMi =

DMFRi - £mfk

DVFRi - £ ml

Where OMj is organic matter corresponded to OQ at temperature gradient i, AMFRj is changes of MFR within temperature gradient i, Em fk is summed mass fractions of individual species that will evaporate or decomposite at temperature gradient i, fOCi is mass fraction of OC in PM1.0 at temperature gradient i.

Mass proportions of OC1 and OC2 in smoke PM10 were displayed in Fig. 4b, OC1 is 18.1 wt.% of PM1.0 on average, and OC2 is 10.9 wt.%. Cotton residue burning particles comprise more non-refractory OM, especially the more volatile OC1 and OC2 (~40 wt.%), which may justify the larger volatility. According to equivalent charge analysis of primary inorganic salts in smoke PM10, mean charge ratio of Cl-: K+ is 1.46, and (NH++K+): (SO24-+Cl-+NO3-) is 1.03. Therefore, the inorganics can be simplified treated as a mixture of KCl-NH4Cl-(NH4)2SO4-NH4NO3. At 150 °C, NH4Cl and NH4NO3 will decomposite, when temperature increase to 250 °C, only KCl is left. Take the smoke particle as spherical and non-porous one, and regardless the morphological changes during pyrolysis, phasing out the inorganic fractions lost due to thermal destruction, the remaining decrease in the mass is contributed by OM evaporation. The results of OM/OC factor and OM density are displayed in Fig. 4b. A similar trend of the factors and densities is observed to vary with smoke particle source and volatility of organics. The more volatile organics have lower OM/OC factors and less OM densities. OM1/OC1 is 1.07 ± 0.08 and OM2/ OC2 is 1.29 ± 0.10, corresponded density of OM1 and OM2 is

0.61-0.90 g cm-3 and 0.86-1.13 g cm-3. OM/OC factor is used to offset the hydrogen, oxygen, and other minor species in the organic materials, which plays a vital role in organic aerosol concentration assessment (Seinfeld et al., 2012). However, the factor of biomass burning particles was seldom reported (Reid et al., 2005; Turpin et al., 2001). The factor from ambient aerosol and specific or-ganics characterization was recommended as 1.2-1.4, and 1.3 was commonly applied (Chow et al., 1993; Li et al., 2007). The factor can reflect saturation degree and aging level of organic materials. The low value of OM1/OC1 may indicate more unsaturated hydrocarbon in OM1.

Fig. 4. Thermostability of smoke particles from agricultural residue burning. a) Thermograms, average normalized volume fraction remaining (VFR), mass fraction remaining (MFR), and effective density against the heater temperature of agri-fire smoke particles. b) Density and mass fraction of separated organics (OC1 and OC2) in PM10 that evaporate at about 150 °C and 250 °C during OC-EC measurement (error bar means standard deviation).

On the other hand, restriction from thermo-denuder design may underestimate the value of OM/OC. Study ever compared the

Fig. 5. Hygroscopicity analysis of smoke particles. a) Hygroscopicity of size-specific smoke particles from 100 to 400 nm b) Hygroscopicity parameter of size-resolved smoke particles from 100 to 400 nm c) Ensemble averaged hygroscopicity parameter

evaporation efficiency of organic aerosol with regard to RT and T effect (Riipinen et al., 2010; Tritscher et al., 2011). If the evaporation is kinetically limited, relative short RT can hardly allow the carbonaceous particle to get evaporation equilibrium at defined temperature, i.e. shorter RT leads to higher VFR-MFR and vice versa. Furthermore, the real VFR at equilibrium and measured VFR can be treated as a hysteretic curve of temperature (T), the gap between the critical and measured values is larger at low T, while the difference can be neglected at relative high temperature. However, seldom study ever defined the critical time allowed to get the evaporation equilibrium, and 3.0 s in our study as RT was comparable to many studies (Meyer et al., 2009; Jonsson et al., 2007; Paulsen et al., 2005). If it was not sufficient, VFR and MFR at 150 °C would be overestimated, resulting in less AVFR1 and AMFR1. This may partly explain the low value of OM1/OC1. Meanwhile, AVFR2 and AMFR2 at 250 °C could be overestimated, leading to higher OM2/OC2 factor.

3.4. Hygroscopicity

Fig. 5a depicts the size-resolved hygroscopic growth factors (GFs) of smoke particles at 80% RH. Fresh smoke particles are hygroscopic, which enable them to be CCN (Martins et al., 2009; Engelhart et al., 2012). Size- and fuel type-dependence of GFs are also obvious, and size effect on GFs can be eliminated via equation (19):

And calculated ensemble GFs of five type agricultural residue burning smokes are 1.25 ± 0.05,1.21 ± 0.04,1.32 ± 0.04,1.24 ± 0.04, and 1.28 ± 0.07 for cotton, soybean, corn, rice, and wheat residue burning smokes, respectively. Herein, hygroscopicity parameter k was used to quantitatively assess the CCN activity and water uptake properties of smoke particles (Rose et al., 2011; Dusek et al., 2010; Ye et al., 2013). According to k-Kohler theory, k value is calculated using following equations:

S = aw x exp

4Ss=aMw RTpwD

GF3 - 1

Where S is saturation ratio (RH, %), aw is activity of water in solution, pw (density of water, 1.0 g cm-3) and Mw (molecular weight of water, 18 g mol-1) are parameter of water, ss/a is surface tension of the solution/air interface, T (thermodynamic temperature, -298.15 K) and R (universal gas constant, 8.314 J mol-1 K-1) are thermodynamic parameter. D is the mobility diameter of droplet.

Apply 0.072 J m-2 as ss/a at 298.15 K from literature (Petters et al., 2007), the results of size-resolved k are presented in Fig. 5b. The corresponding hygroscopicity parameter k showed a size dependence of smoke particle, which decreases with increasing size, indicating more hygroscopic or CCN potential of smaller size at diameter from 50 to 300 nm. Study ever confirmed size is more important than chemical composition to enable particle to be CCN (Dusek et al., 2006, 2010; Rose et al., 2011). Smoke particles have k values in a range from 0.20 to 0.40, which can be classified in less- or more-hygroscopic level, exhibiting potential

k of different agricultural residue burning particles as a function of inorganic mass fraction in PM1.0 and PM2.5 (error bar means standard deviation).

effect on hydropogical cycle and climate changes (Petters et al., 2009). Mean hygroscopicity parameter can be calculated via equation (22):

ZnN x ki YÂNi

Mean values of k are 0.23 ± 0.03, 0.19 ± 0.02, 0.32 ± 0.03, 0.24 ± 0.02, and 0.26 ± 0.04 for cotton, soybean, corn, rice, and wheat residue burning smoke, respectively. The results agree well with previous study as summarized in Table 3, k of biomass burning aerosol ranges from 0.02 to 0.80 (Dusek et al., 2011; Petters et al., 2009). Engelhart et al. reported that photochemical production of SOA decreased CCN activation of smoke aerosol, and k of different fuel burning particles eventually converged to a similar value of about 0.2 (Engelhart et al., 2012).

For a multicomponent system (internal mixing of organics and inorganics) at equilibrium, equation (23) can be rearranged based on ZSR model (Zdanovskii, Stokes, and Robinson), bulk hygro-scopicity parameter k is integration of k for individual species weighted by volume fraction (Clegg et al., 2004):

where ei is volume fraction of individual species i in particle, Kj is hygroscopicity parameter for individual species i.

The fuel-dependence variation of mean hygroscopicity parameter k could be parameterized as a function of organic and inorganic fractions based on equations (24) and (25), as the result showed in Fig. 5c.

k/PMi.0 k = 0.738 x fin0rg + 0.087 x forg (24)

k/PM2.5 k = 0.846 x finorg + 0.068 x forg (25)

In field study, size-resolved chemical compositions of non-refractory submicron particles were measured using AMS, and effective density of particle was assumed to be 1.6—1.7 g cm~3 to convert the aerodynamic diameter to the mobility one, then AMS results were plotted with k directly as: k = finorg x kinorg + forg x korg (Rose et al., 2011; Gunthe et al., 2009, 2011). The approximate average korg of about 0.1 has been reported by many studies including ambient investigation and laboratory simulation (Dusek et al., 2010; Giordano et al., 2013; Gunthe et al., 2011; Rose et al., 2011), and the value deposits in the characteristic range for individual organic component from zero of absolute insoluble species

like soot to ~0.5 of oxalic acid (Zhang et al., 2008; Petters et al., 2007). In this work, the values of korg are 0.087 ± 0.033 in smoke PM1.0 and 0.068 ± 0.031 in smoke PM2.5, agree well with previous study. Extrapolated kinorg is 0.738 ± 0.280 in smoke PM1.0 and 0.846 ± 0.286 in PM2.5, while the value from documents is 0.6—0.7 (Dusek et al., 2010; Gunthe et al., 2009, 2011). The deviation can be explained by the chemical difference in the research subject, 0.6—0.7 is for ambient aerosol including biomass burning particles, in which ammonium, sulfate and nitrate are the major inorganic materials due to atmospheric mixing and aging, thus the value deposits into the k of NH4NO3 and (NH4)2SO4 from Table 2, but in the fresh smoke particles, chlorides and potassium contribute primarily to the alkaline species, KCl and NH4Cl both have larger k values. The vast uncertainties in smoke korg and kinorg assessments lie mainly in uncertainty of off-line chemical analysis, 10% deviation in smoke compositions lead to over 40% error in korg and kinorg estimations, besides, k averaged from characteristics of 50—300 nm particles may not fit that of PM25 considering the significant size related chemical difference in PM1.0 and PM2.5. Nevertheless, korg and kinorg of smoke PM1.0 are reasonable in actual study.

By definition, the mixing rule of the k values of multicompo-nents refers to volume fractions (Petters et al., 2007). From chemical measurement, average mass ratio of primary inorganic ions Cl3 :SO4~ :NO3 :NH+ :K+ is 0.41:0.13:0.02:0.12:0.32 in smoke PM1.0 for 5 types of agricultural residues burning, the chemical speciation can be trated as KCl—NH4Cl—NH4NO3-(NH4)2SO4 or NH4Cl—K-Cl—KNO3—K2SO4 mixture for simplicity, based on equation (23) and physical parameter of individual salt from Table 2, bulk kinorg can be estimated. As k of potassium-salt can hardly found from literature, k for KCl was estimated to be ~0.65 (Freney et al., 2009; Park et al., 2009), the values for KNO3 and K2SO4 were cited as that of corresponded sodium-salt, and computed bulk kinorg of these mixtures is 0.66—0.72, which is in line with the result from extrapolation of fit line between k and inorganic mass fraction in smoke PM1.0.

4. Conclusion

Effective density of fresh smoke particle varies from 1.1 to 1.4 g cm~3, and corresponded hygroscopicity parameter k ranges from 0.2 to 0.35, the size- and fuel type-dependence are distinct. Ensemble densities and k values both scale with inorganics proportion of smoke particles, thus, the parameterized functions of density and k against organic and inorganic mass fraction in PM1.o and PM2.5 was established. The extrapolated pinorg and Kinorg are 2.13 g cm~3 and 0.734 in smoke PM1.o, which agree well with the bulk values calculated from inorganic chemical compositions. The derived porg and Korg are 1.14 g cm~3 and 0.087. Therefore, density

Table 3

Summary of hygroscopicity parameter (k) of biomass burning aerosol.

Fuel MCE k Method Reference

Wheat straw 0.91 0.26 LS, HTDMA (80% RH) This work

Corn straw 0.93 0.32

Rice straw 0.89 0.24

Cotton residue 0.93 0.23

Soybean residue 0.96 0.19

Average 0.92 0.25

chamise, manzanita >0.98 0.10-0.24 LS, Scanning Mobility CCN Analysis Giordano et al., 2013

24 biomass fuels >0.9 0.06-0.70 LS, HTDMA (89% RH), CCN measurement Petters et al., 2009

Wheat residue >0.9 0.27 LS, HTDMA (80% RH) Li et al., 2015

12 biomass fuels >0.9 0.06-0.60 LS, Scanning Mobility CCN Analysis Engelhart et al., 2012

oak, musasa >0.9 0.05-0.20 LS, Scanning Mobility CCN Analysis Dusek et al., 2011

3 biomass fuels >0.9 0.2-0.7 LS, HTDMA (95% RH) Lewis et al., 2009

Agricultural fire - 0.1-0.6 FS, Scanning Mobility CCN Analysis Rose et al., 2011

Amazon fire - 0.03-0.07 FS, HTDMA (90% RH) Rissler et al., 2006

Note: LS means laboratory simulation; FS means field study.

and k of agricultural residue burning particle can be estimated by the empirical functions: p-1 = 0.470 x finorg + 0.878 x forg and k = 0.738 x finorg + 0.087 x forg.

Fuel-specific VFR and MFR decrease with increasing heating temperature, while density of the non-refractory materials increases. The gradient temperature of V-TDMA was set consistent with the splitting temperature of OC1 and OC2 in OC-EC measurement (150 and 250 °C correspond to OC1 and OC2). Based on VFR, MFR, and detailed chemical composition in PM1.0, densities of OM so as ratios of OM/OC at these two defined temperatures were estimated. Average density of OM1 and OM2 are 0.79 and 1.05 g cm-3. Average OM1/OC1 and OM2/OC2 factors are 1.07 and 1.29. It is obvious that more volatile or thermo-unstable organics have less density and lower OM/OC factor in the external mixed smoke particles.

Acknowledgement

This work is supported by National Natural Science Foundation of China (No. 21190053, 21177025, 21177025), Cyrus Tang Foundation (No. CTF-FD2014001), Ministry of Science and Technology of China (No. 2014BAC22B01), and Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDB05010200), and the IRSES-EU Marie Curie Actions project "AMIS" (PIRSES-GA-2011-295132).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atmosenv.2016.05.052.

References

Andreae, M.O., Merlet, P., 2001. Emission of trace gases and aerosols from biomass

burning. Glob. Biogeochem. Cycles 15, 955—966. Arora, P., Jain, S., 2015. Estimation of organic and elemental carbon emitted from wood burning in traditional and improved cookstoves using controlled cooking test. Environ. Sci. Technol. 49, 3958—3965. Aurell, J., Gullett, B.K., Tabor, D., 2015. Emissions from southeastern U.S. Grasslands and pine savannas: comparison of aerial and ground field measurements with laboratory burns. Atmos. Environ. 111,170—178. Beddows, D.C.S., Dall'Osto, M., Harrison, R.M., 2010. An enhanced procedure for the merging of atmospheric particle size distribution data measured using electrical mobility and time-of-flight analysers. Aerosol Sci. Tech. 44, 930—938. Bolling, A.K., Pagels, J., Yttri, K.E., Barregard, L., Sallsten, G., Schwarze, P.E., Boman, C., 2009. Health effects of residential wood smoke particles: the importance of combustion conditions and physicochemical particle properties. Part. Fibre Toxicol. 29 http://dx.doi.org/10.1186/1743-8977-6-29. Cao, J.J., Wu, F., Chow, J.C., Lee, S.C., Li, Y., Chen, S.W., An, Z.S., Fung, K.K., Watson, J.G., Zhu, C.S., 2005. Characterization and source apportionment of atmospheric organic and elemental carbon during fall and winter of 2003 in Xi'an, China. Atmos. Chem. Phys. 5, 3127—3137. Chen, Q., Farmer, D.K., Schneider, J., Zorn, S.R., Heald, C.L., Karl, T.G., Guenther, A., Allan, J.D., Robinson, N., Coe, H., 2009. Mass spectral characterization of submicron biogenic organic particles in the Amazon Basin. Geophys. Res. Lett. 36 http://dx.doi.org/10.1029/2009GL039880. China, S., Salvadori, N., Mazzoleni, C., 2014. Effect of traffic and driving characteristics on morphology of atmospheric soot particles at freeway on-ramps. Environ. Sci. Technol. 48, 3128—3135. Chow, J.C., Watson, J.G., Pritchett, L.C., Pierson, W.R., Frazier, C.A., Purcell, R.G., 1993. The DRI thermal/optical reflectance carbon analysis system: description, evaluation, and applications in U.S. air quality studies. Atmos. Environ. 27A, 1185—1201.

Clegg, S.L., Seinfeld, J.H., 2004. Improvement of the Zdanovskii-Stokes-Robinson model for mixtures containing solutes of different charge types. J. Phys. Chem. Atmos. 108, 1008—1017. Cross, E.S., Slowik, J.G., Davidovits, P., Allan, J.D., Worsnop, D.R., Jayne, J.T., Lewis, D.K., Canagaratna, M., Onasch, T.B., 2007. Laboratory and ambient particle density determinations using light scattering in conjunction with aerosol mass spectrometry. Aerosol Sci. Tech. 41, 343—359. Cubison, M.J., Ortega, A.M., Hayes, P.L., Farmer, D.K., Day, D., Lechner, M.J., Brune, W.H., Apel, E., Diskin, G.S., Fisher, J.A., 2011. Effects of aging on organic aerosol from open biomass burning smoke in aircraft and laboratory studies. Atmos. Chem. Phys. 11,12049—12064. Dhammapala, R., Claiborn, C., Jimenez, J., Corkill, J., Gullett, B., Simpson, C.,

Paulsen, M., 2007. Emission factors of PAHs, methoxyphenols, levoglucosan, elemental carbon and organic carbon from simulated wheat and Kentucky bluegrass stubble burns. Atmos. Environ. 41, 2660—2669.

Duan, J., Tan, J., 2013. Atmospheric heavy metals and Arsenic in China: situation, sources and control policies. Atmos. Environ. 74, 93—101.

Dusek, U., Frank, G.P., Hildebrandt, L., Curtius, J., Schneider, J., Walter, S., Chand, D., Drewnick, F., Hings, S., Jung, D., 2006. Size matters more than chemistry for cloud-nucleating ability of aerosol particles. Science 312,1375—1378.

Dusek, U., Frank, G.P., Curtius, J., Drewnick, F., Schneider, J., Kurten, A., Rose, D., Andreae, M.O., Borrmann, S., Poschl, U., 2010. Enhanced organic mass fraction and decreased hygroscopicity of cloud condensation nuclei (CCN) during new particle formation events. Geophys. Res. Lett. 37 http://dx.doi.org/10.1029/ 2009GL040930.

Dusek, U., Frank, G.P., Massling, A., Zeromskiene, K., linuma, Y., Schmid, O., Helas, G., Hennig, T., Wiedensohler, A., Andreae, M.O., 2011. Water uptake by biomass burning aerosol at sub- and supersaturated conditions: closure studies and implications for the role of organics. Atmos. Chem. Phys. 11, 9519—9532.

Engelhart, G.J., Hennigan, C.J., Miracolo, M.A., Robinson, A.L., Pandis, S.N., 2012. Cloud condensation nuclei activity of fresh primary and aged biomass burning aerosol. Atmos. Chem. Phys. 12, 7285—7293.

Ferek, R.J., Reid, J.S., Hobbs, P.V., Blake, D.R., Liousse, C., 1998. Emission factors of hydrocarbons, halocarbons, trace gases and particles from biomass burning in Brazil. J. Geophys. Res. 103, 32107—32118.

Freney, E.J., Martin, S.T., Buseck, P.R., 2009. Deliquescence and efflorescence of potassium salts relevant to biomass-burning aerosol particles. Aerosol Sci. Tech-nol. 43, 799—80 .

Giordano, M.R., Short, D.Z., Hosseini, S., Lichtenberg, W., Asa-Awuku, A.A., 2013. Changes in droplet surface tension affect the observed hygroscopicity of photochemically aged biomass burning aerosol. Environ. Sci. Technol. 47, 10980—10986.

Grieshop, A.P., Logue, J.M., Donahue, N.M., Robinson, A.L., 2009. Laboratory investigation of photochemical oxidation of organic aerosol from wood fires 1: measurement and simulation of organic aerosol evolution. Atmos. Chem. Phys. 9,1263—1277.

Gunthe, S.S., King, S.M., Rose, D., Chen, Q., Roldin, P., Farmer, D.K., Jimenez, J.L., Artaxo, P., Andreae, M.O., Martin, S.T., Poschl, U., 2009. Cloud condensation nuclei in pristine tropical rainforest air of Amazonia: size-resolved measurements and modeling of atmospheric aerosol composition and CCN activity. Atmos. Chem. Phys. 9, 7551—7575.

Gunthe, S.S., Rose, D., Su, H., Garland, R.M., Achtert, P., Nowak, A., Wiedensohler, A., Kuwata, M., Takegawa, N., Kondo, Y., Hu, M., Shao, M., Zhu, T., Andreae, M.O., Poschl, U., 2011. Cloud condensation nuclei (CCN) from fresh and aged air pollution in the megacity region of Beijing. Atmos. Chem. Phys. 11, 11023—11039.

Gysel, M., McFiggans, G.B., Coe, H., 2009. lnversion of tandem differential mobility analyser (TDMA) measurements. J. Aerosol Sci. 40, 134—151.

Jonsson, A.M., Hallquist, M., Saathoff, H., 2007. Volatility of secondary organic aerosols from the ozone initiated oxidation of alpha-pinene and limonene. J. Aerosol Sci. 38, 843—852.

Han, Y., Cao, J., Chow, J.C., Watson, J.G., An, Z., Jin, Z., Fung, K., Liu, S., 2007. Evaluation of the thermal/optical reflectance method for discrimination between char- and soot-EC. Chemosphere 69, 569—574.

Han, Y.M., Lee, S.C., Cao, J.J., Ho, K.F., An, Z.S., 2009. Spatial distribution and seasonal variation of char-EC and soot-EC in the atmosphere over China. Atmos. Environ. 43, 6066—6073.

Hayashi, K., Ono, K., Kajiura, M., Sudo, S., Yonemura, S., Fushimi, A., Saitoh, K., Fujitani, Y., Tanabe, K., 2014. Trace gas and particle emissions from open burning of three cereal crop residues: increase in residue moistness enhances emissions of carbon monoxide, methane, and particulate organic carbon. Atmos. Environ. 95, 36—44.

Hu, D., Chen, J., Ye, X., Li, L., Yang, X., 2011. Hygroscopicity and evaporation of ammonium chloride and ammonium nitrate: relative humidity and size effects on the growth factor. Atmos. Environ. 45, 2349—2355.

Huang, R., Zhang, Y., Bozzetti, C., Ho, K., Cao, J., Han, Y., Daellenbach, K.R., Slowik, J.G., Platt, S.M., Canonaco, F., Zotter, P., Wolf, R., Pieber, S.M., Bruns, E.A., Crippa, M., Ciarelli, G., Piazzalunga, A., Schwikowski, M., Abbaszade, G., Schnelle-Kreis, J., Zimmermann, R., An, Z., Szidat, S., Baltensperger, U., Haddad, I.E., Prévôt, A.S.H., 2014. High secondary aerosol contribution to particulate pollution during haze events in China. Nature 514, 218—222.

Janssen, N.A.H., Hoek, G., Simic-Lawson, M., Fischer, P., van Bree, L., Ten Brink, H., Keuken, M., Atkinson, R.W., Anderson, H.R., Brunekreef, B., Cassee, F.R., 2011. Black carbon as an additional indicator of the adverse health effects of airborne particles compared with PM10 and PM2.5. Environ. Health Persp. 119, 1691—1699.

Jolleys, M.D., Coe, H., McFiggans, G., Capes, G., Allan, J.D., Crosier, J., Williams, P.l., Allen, G., Bower, K.N., Jimenez, J.L., 2012. Characterizing the aging of biomass burning organic aerosol by use of mixing ratios: a meta-analysis of four regions. Environ. Sci. Technol. 46, 13093—13102.

Katrib, Y., Martin, S.T., Rudich, Y., Davidovits, P., Jayne, J.T., Worsnop, D.R., 2005. Density changes of aerosol particles as a result of chemical reaction. Atmos. Chem. Phys. 5, 275—29 .

Khlystov, A., Stanier, C., Pandis, S.N., 2004. An algorithm for combining electrical mobility and aerodynamic size distributions data when measuring ambient aerosol special issue of aerosol science and technology on findings from the fine particulate matter supersites program. Aerosol Sci. Technol. 38, 229—238.

Levin, E.J.T., McMeeking, G.R., Carrico, C.M., Mack, L.E., Kreidenweis, S.M., Wold, C.E., Moosmüller, H., Arnott, W.P., Hao, W.M., Collett, J.L., Malm, W.C., 2010. Biomass burning smoke aerosol properties measured during fire laboratory at missoula experiments (FLAME). J. Geophys. Res. 115 http://dx.doi.org/10.1029/ 2009JD013601.

Lewis, K.A., Arnott, W.P., Moosmuller, H., Chakrabarty, R.K., Carrico, C.M., Kreidenweis, S.M., Day, D.E., Malm, W.C., Laskin, A., Jimenez, J.L., Ulbrich, I.M., Huffman, J.A., Onasch, T.B., Trimborn, A., Liu, L., Mishchenko, M.I., 2009. Reduction in biomass burning aerosol light absorption upon humidification: roles of inorganically-induced hygroscopicity, particle collapse, and photo-acoustic heat and mass transfer. Atmos. Chem. Phys. 9, 8949—8966.

Li, J., Posfai, M., Hobbs, P.V., Buseck, P.R., 2003. Individual aerosol particles from biomass burning in southern Africa: 2, Compositions and aging of inorganic particles. J. Geophys. Res. Atmos. 108 http://dx.doi.org/10.1029/2002JD002310.

Li, X., Wang, S., Duan, L., Hao, J., Li, C., Chen, Y., Yang, L., 2007. Particulate and trace gas emissions from open burning of wheat straw and corn stover in China. Environ. Sci. Technol. 41, 6052—6058.

Li, C., Ma, Z., Chen, J., Wang, X., Ye, X., Wang, L., Yang, X., Kan, H., Donaldson, D.J., Mellouki, A., 2015. Evolution of biomass burning smoke particles in the dark. Atmos. Environ. 120, 244—252.

Martins, J.A., Gon^alves, F.L.T., Morales, C.A., Fisch, G.F., Pinheiro, F.G.M., Leal Júnior, J.B.V., Oliveira, C.J., Silva, E.M., Oliveira, J.C.P., Costa, A.A., Silva Dias, M.A.F., 2009. Cloud condensation nuclei from biomass burning during the Amazonian dry-to-wet transition season. Meteorol. Atmos. Phys. 104, 83—93.

Meyer, N.K., Duplissy, J., Gysel, M., Metzger, A., Dommen, J., Weingartner, E., Alfarra, M.R., Prevot, A.S.H., Fletcher, C., Good, N., McFiggans, G., Jonsson, A.M., 2009. Analysis of the hygroscopic and volatile properties of ammonium sulphate seeded and unseeded SOA particles. Atmos. Chem. Phys. 9, 721—732.

Novakov, T., Corrigan, C.E., 1996. Cloud condensation nucleus activity of the organic component of biomass smoke particles. Geophys. Res. Lett. 23, 2141—2144.

Pagels, J., Khalizov, A.F., McMurry, P.H., Zhang, R.Y., 2009. Processing of soot by controlled sulphuric acid and water condensation-mass and mobility relationship. Aerosol Sci. Technol. 43, 629—640.

Park, K., Kim, J., Miller, A.L., 2009. A study on effects of size and structure on hy-groscopicity of nanoparticles using a tandem differential mobility analyzer and TEM. J. Nanopart. Res. 11,175—183.

Paulsen, D., Dommen, J., Kalberer, M., Prevot, A.S.H., Richter, R., Sax, M., Steinbacher, M., Weingartner, E., Baltensperger, U., 2005. Secondary organic aerosol formation by irradiation of 1,3,5-trimethylbenzene-NOx-H2O in a new reaction chamber for atmospheric chemistry and physics. Environ. Sci. Technol. 39, 2668—2678.

Petters, M.D., Kreidenweis, S.M., 2007. A single parameter representation of hygroscopic growth and cloud condensation nucleus activity. Atmos. Chem. Phys. 7,1961 —1971.

Petters, M.D., Carrico, C.M., Kreidenweis, S.M., Prenni, A.J., DeMott, P.J., Collett, J.L., Moosmueller, H., 2009. Cloud condensation nucleation activity of biomass burning aerosol. J. Geophys. Res. 114 http://dx.doi.org/10.1029/2009JD012353.

Pratt, K.A., Prather, K.A., 2009. Real-time, single-particle volatility, size, and chemical composition measurements of aged urban aerosols. Environ. Sci. Technol. 43, 8276—8282.

Ram, K., Sarin, M.M., Tripathi, S.N., 2011. Temporal trends in atmospheric PM25, PM10, elemental carbon, organic carbon, water-soluble organic carbon, and optical properties: impact of biomass burning emissions in the Indo-Gangetic plain. Environ. Sci. Technol. 46, 686—695.

Ramanathan, V., Crutzen, P.J., Kiehl, J.T., Rosenfeld, D., 2001. Aerosols, climate, and the hydrological cycle. Science 294, 2119—2124.

Reid, J.S., Eck, T.F., Christopher, S.A., Koppmann, R., Dubovik, O., Eleuterio, D.P., Holben, B.N., Reid, E.A., Zhang, J., 2005a. A review of biomass burning emissions part III: intensive optical properties of biomass burning particles. Atmos. Chem. Phys. 5, 827—849.

Reid, J.S., Koppmann, R., Eck, T.F., Eleuterio, D.P., 2005b. A review of biomass burning emissions part II: intensive physical properties of biomass burning particles. Atmos. Chem. Phys. 5, 799—825.

Riipinen, I., Pierce, J.R., Donahue, N.M., Pandis, S.N., 2010. Equilibration time scales of organic aerosol inside thermodenuders: evaporation kinetics versus thermodynamics. Atmos. Environ. 44, 597—607.

Rissler, J., Vestin, A., Swietlicki, E., Fisch, G., Zhou, J., Artaxo, P., Andreae, M.O., 2006. Size distribution and hygroscopic properties of aerosol particles from dry-season biomass burning in Amazonia. Atmos. Chem. Phys. 6, 471—491.

Rissler, J., Nordin, E.Z., Eriksson, A.C., Nilsson, P.T., Frosch, M., Sporre, M.K., Wierzbicka, A., Svenningsson, B., Londahl, J., Messing, M., Sjogren, S., Hemmingsen, J.G., Loft, S., Pagels, J.H., Swietlicki, E., 2014. Effective density and mixing state of aerosol particles in a near-traffic urban environment. Environ. Sci. Technol. 48, 6300—6308.

Rose, D., Gunthe, S.S., Su, H., Garland, R.M., Yang, H., Berghof, M., Cheng, Y.F., Wehner, B., Achtert, P., Nowak, A., Wiedensohler, A., Takegawa, N., Kondo, Y., Hu, M., Zhang, Y., Andreae, M.O., Pooschl, U., 2011. Cloud condensation nuclei in polluted air and biomass burning smoke near the mega-city Guangzhou, ChinaPart 2: size-resolved aerosol chemical composition, diurnal cycles, and externally mixed weakly CCN-active soot particles. Atmos. Chem. Phys. 11, 2817—2836.

Rosenfeld, D., 2006. Atmosphere: aerosols, clouds, and climate. Science 312, 1323—1324.

Saikawa, E., Naik, V., Horowitz, L.W., Liu, J., Mauzerall, D.L., 2009. Present and

potential future contributions of sulfate, black and organic carbon aerosols from China to global air quality, premature mortality and radiative forcing. Atmos. Environ. 43, 2814-2822.

Schkolnik, G., Chand, D., Hoffer, A., Andreae, M.O., Erlick, C., Swietlicki, E., Rudich, Y., 2007. Constraining the density and complex refractive index of elemental and organic carbon in biomass burning aerosol using optical and chemical measurements. Atmos. Environ. 41, 1107-1118.

Schmid, O., Karg, E., Hagen, D.E., Whitefield, P.D., Ferron, G.A., 2007. On the effective density of non-spherical particles as derived from combined measurements of aerodynamic and mobility equivalent size. J. Aerosol Sci. 38, 431-443.

Schmid, O., Chand, D., Karg, E., Guyon, P., Frank, G.P., Swietlicki, E., Andreae, M.O., 2009. Derivation of the density and refractive index of organic matter and elemental carbon from closure between physical and chemical aerosol properties. Environ. Sci. Technol. 43,1166-1172.

Seinfeld, J.H., Pandis, S.N., 2012. Atmospheric Chemistry and Physics: from Air Pollution to Climate Change. John Wiley & Sons.

Shindell, D., Kuylenstierna, J.C.I., Vignati, E., van Dingenen, R., Amann, M., Klimont, Z., Anenberg, S.C., Muller, N., Janssens, M.G., Raes, F., Schwartz, J., Faluvegi, G., Pozzoli, L., Kupiainen, K., Hoglund, I.L., Emberson, L., Streets, D., Ramanathan, V., Hicks, K., Oanh, N.T.K., Milly, G., Williams, M., Demkine, V., Fowler, D., 2012. Simultaneously mitigating near-term climate change and improving human health and food security. Science 335, 183-189.

Spracklen, D.V., Carslaw, K.S., Poschl, U., Rap, A., Forster, P.M., 2011. Global cloud condensation nuclei influenced by carbonaceous combustion aerosol. Atmos. Chem. Phys. 11, 9067-9087.

Tajima, N., Fukushima, N., Ehara, K., Sakurai, H., 2011. Mass range and optimized operation of the aerosol particle mass analyzer. Aerosol Sci. Technol. 45, 196-214.

Tang, I.N., Munkelwitz, H.R., 1993. Composition and temperature dependence of the deliquescence properties of hygroscopic aerosols. Atmos. Environ. 27, 467-473.

Tian, D., Hu, Y., Wang, Y., Boylan, J.W., Zheng, M., Russell, A.G., 2008. Assessment of biomass burning emissions and their impacts on urban and regional PM25: a Georgia case study. Environ. Sci. Technol. 43, 299-305.

Tiitta, P., Miettinen, P., Vaattovaara, P., Joutsensaari, J., Petâja, T., Virtanen, A., Raatikainen, T., Aalto, P., Portin, H., Romakkaniemi, S., 2010. Roadside aerosol study using hygroscopic, organic and volatility TDMAs: characterization and mixing state. Atmos. Environ. 44, 976-986.

Tritscher, T., Dommen, J., DeCarlo, P.F., Gysel, M., Barmet, P.B., Praplan, A.P., Weingartner, E., Prévôt, A.S.H., Riipinen, I., Donahue, N.M., Baltensperger, U., 2011. Volatility and hygroscopicity of aging secondary organic aerosol in a smog chamber. Atmos. Chem. Phys. 11,11477-11496.

Turn, S.Q., Jenkins, B.M., Chow, J.C., Pritchett, L.C., Campbell, D., Cahill, T., Whalen, S.A., 1997. Elemental characterization of particulate matter emitted from biomass burning' Wind tunnel derived source profiles. J. Geophys. Res. 102, 3683-3699.

Turpin, B.J., Lim, H., 2001. Species contributions to PM25 mass concentrations: revisiting common assumptions for estimating organic mass. Aerosol Sci. Technol. 35, 602-610.

Wehner, B., Philippin, S., Wiedensohler, A., 2002. Design and calibration of a ther-modenuder with an improved heating unit to measure the size-dependent volatile fraction of aerosol particles. J. Aerosol Sci. 33, 1087-1093.

Wehner, B., Berghof, M., Cheng, Y.F., Achtert, P., Birmili, W., Nowak, A., Wiedensohler, A., Garland, R.M., Poschl, U., Hu, M., Zhu, T., 2009. Mixing state of nonvolatile aerosol particle fractions and comparison with light absorption in the polluted Beijing region. J. Geophys. Res. 114 http://dx.doi.org/10.1029/ 2008JD010923.

Wilson, J.M., Baeza-Romero, M.T., Jones, J.M., Pourkashanian, M., Williams, A., LeaLangton, A.R., Ross, A.B., Bartle, K.D., 2013. Soot formation from the combustion of biomass pyrolysis products and a hydrocarbon fuel, n-Decane: an aerosol time of flight mass spectrometer (ATOFMS) Study. Energy Fuel 27,1668-1678.

Wu, C., Liu, L.J.S., Cullen, A., Westberg, H., Williamson, J., 2011. Spatial-temporal and cancer risk assessment of selected hazardous air pollutants in Seattle. Environ. Int. 37,11-17.

Ye, X., Chen, T., Hu, D., Yang, X., Chen, J., Zhang, R., Khakuziv, A., Wang, L., Prospero, A., 2009. A multifunctional HTDMA system with a robust temperature control. Adv. Atmos. Sci. 26, 1235-1240.

Ye, X., Tang, C., Yin, Z., Chen, J., Ma, Z., Kong, L., Yang, X., Gao, W., Geng, F., 2013. Hygroscopic growth of urban aerosol particles during the 2009 MirageShanghai Campaign. Atmos. Environ. 64, 263-269.

Yin, Z., Ye, X., Jiang, S., Tao, Y., Shi, Y., Yang, X., Chen, J., 2015. Size-resolved effective density of urban aerosols in Shanghai. Atmos. Environ. 100,133-140.

Zauscher, M.D., Wang, Y., Moore, M.J.K., Gaston, C.J., Prather, K.A., 2013. Air quality impact and physicochemical aging of biomass burning aerosols during the 2007 San Diego wildfires. Environ. Sci. Technol. 47, 7633-7643.

Zhang, H., Ye, X., Cheng, T., Chen, J., Yang, X., Wang, L., Zhang, R., 2008a. A laboratory study of agricultural crop residue combustion in China: emission factors and emission inventory. Atmos. Environ. 42, 8432-844 .

Zhang, R., Khalizov, A.F., Pagels, J., Zhang, D., Xue, H., McMurry, P.H., 2008b. Variability in morphology, hygroscopicity, and optical properties of soot aerosols during atmospheric processing. Proc. Natl. Acad. Sci. 105,10291-10296.

Zhang, H., Hu, D., Chen, J., Ye, X., Wang, S.X., Hao, J.M., Wang, L., Zhang, R., An, Z., 2011. Particle size distribution and polycyclic aromatic hydrocarbons emissions from agricultural crop residue burning. Environ. Sci. Technol. 45, 5477-5482.