Size-resolved aerosol chemical analysis of extreme haze pollution events during early 2013 in urban Beijing, China

Journal of Hazardous Materials 279 (2014) 452–460

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Size-resolved aerosol chemical analysis of extreme haze pollution events during early 2013 in urban Beijing, China Shili Tian, Yuepeng Pan ∗ , Zirui Liu, Tianxue Wen, Yuesi Wang ∗ State Key Laboratory of Atmospheric Boundary Layer Physics and Atmospheric Chemistry (LAPC), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China

h i g h l i g h t s • • • • •

Anthropogenic species substantially accumulated in both fine and coarse particles. Secondary organic carbon in PM1.1 decreased from clear to haze days. The mass peak shifted to larger particles from clear to haze days. The NO3 − /SO4 2− ratio decreased with enhanced haze pollution. Both mobile local and stationary regional sources were vital for haze formation.

a r t i c l e

i n f o

Article history: Received 9 April 2014 Received in revised form 13 June 2014 Accepted 7 July 2014 Available online 21 July 2014 Keywords: Haze pollution Size distributions Aerosol chemistry Mass closure Urban Beijing

a b s t r a c t Using size-resolved filter sampling and chemical characterization, high concentrations of water-soluble ions, carbonaceous species and heavy metals were found in both fine (PM2.1 ) and coarse (PM2.1–9 ) particles in Beijing during haze events in early 2013. Even on clear days, average mass concentration of submicron particles (PM1.1 ) was several times higher than that previously measured in most of abroad urban areas. A high concentration of particulate matter on haze days weakens the incident solar radiation, which reduces the generation rate of secondary organic carbon in PM1.1 . We show that the peak mass concentration of particles shifted from 0.43–0.65 ␮m on clear days to 0.65–1.1 ␮m on lightly polluted days and to 1.1–2.1 ␮m on heavily polluted days. The peak shifts were also found for the following species: organic carbon, elemental carbon, NH4 + , SO4 2− , NO3 − , K, Cu, Zn, Cd and Pb. Our findings demonstrate that secondary inorganic aerosols (36%) and organic matter (26%) dominated the fine particle mass on heavily polluted days, while their contribution reduced to 29% and 18%, respectively, on clear days. Besides fine particles, anthropogenic chemical species also substantially accumulated in the coarse mode, which suggests that particles with aerodynamic diameter larger than 2.1 ␮m cannot be neglected during severe haze events. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Haze is defined as a weather phenomenon featuring a high concentration of fine particles that leads to a visibility of less than 10 km at a relative humidity (RH) lower than 90% [1–3]. This severe air pollution can not only decrease air quality and endanger public health but can also alter the radiation budget and the hydrological cycle [4,5]. As a consequence, the persistence and severity of haze events have resulted in widespread public concern. In recent years, China has faced serious haze pollution due to vast energy consumption and a rapid increase in the number of vehicles. However, further research is necessary to understand the characteristics, formation and evolution of large-area and long-lasting haze pollution events across China.

∗ Corresponding authors. Tel.: +86 01082020530; fax: +86 01062362389. E-mail addresses: [email protected] (Y. Pan), [email protected] (Y. Wang). http://dx.doi.org/10.1016/j.jhazmat.2014.07.023 0304-3894/© 2014 Elsevier B.V. All rights reserved.

In January 2013, several of the most severe haze events on record swept across most of east-central cities and covered a quarter of the total land area in China. These severe wintertime haze events included substantial regional particulate matter (PM) pollution according to both satellite- and surface-based aerosol optical depth (AOD) observations and model simulations [6–8]. In addition, high concentrations of important trace gases, such as carbon monoxide (CO), sulfur dioxide (SO2 ) and ammonia (NH3 ), and ammonium sulfate aerosols ((NH4 )2 SO4 )) were found using IASI (Infrared Atmospheric Sounding Interferometer) satellite measurements [9]. Based on surface measurements, we previously found that extremely high nitric oxide (NOx ) concentrations played either a direct or indirect role in the rapid secondary transformation of coal-burning SO2 into sulfate aerosols during haze episodes [10]. In addition to sulfate and nitrate, high concentrations of ammonium and organic carbon were also observed in Beijing and Shanghai during these severe regional haze events (e.g., Jan. 12–13 and Jan. 29–30, 2013) [11–13]. However, most of these studies focused only on inorganic and organic chemical compounds in single particle fractions (e.g., PM1 or PM2.5 ). To date, the size-resolved organic carbon, elemental carbon and water-soluble ion chemistry for these haze episodes remains unclear, especially regarding water-soluble organic carbon and toxic

S. Tian et al. / Journal of Hazardous Materials 279 (2014) 452–460 metals. Knowledge of the size distribution of various chemical species is important for understanding the physical and chemical atmospheric processes that affect aerosol properties during haze episodes [14]. Haze pollution is characterized by elevated levels of fine particles, which have severe adverse effects on human health, visibility and climate forcing due to their chemical compositions and longer lifetimes in the atmosphere [15]. Coarse particle superposition is also an important factor for the reduced visibility during haze pollution [10]. Besides, the deposition of acids and toxic species through coarse particles causes adverse effects in ecosystems [16]. Such size-resolved clemical investigations are especially needed in megacities, such as Beijing, that have been subjected to periodic haze pollution in recent years. In this study, two cascade impactors simultaneously collected size-segregated airborne particles in urban Beijing, where prolonged haze episodes occurred in January and February of 2013. To our knowledge, these are the only measurements obtained in China during these haze periods to characterize the size distributions of chemical species (i.e., water-soluble ions, carbonaceous species and elements). The unique datasets may improve our knowledge of aerosol properties during haze pollution episodes in Beijing and thereby provide baseline information for future measurement and modeling studies.

2. Materials and methods 2.1. Sampling site The sampling site was on the roof of a building at the Institute of Atmospheric Physics (IAP). The building is approximately 15 m above the ground, located near a residential area in North Beijing (39◦ 58 N, 116◦ 22 E) and situated between the 3rd and 4th ring roads (Fig. S1). The site is approximately 1 km from the 3rd ring road, 200 m west of G6 Highway (which runs north-south) and 50 m south of Beitucheng West Road (which runs east-west). Daily average traffic volumes for 3rd and 4th ring roads were around 10 million/day [17]. There were no industrial sources of atmospheric pollutants during the study period. The experimental campaign was conducted between Jan. 1 and Feb. 28, 2013.

453

wind direction, were observed using an automatic meteorological observation instrument (Milos520, Vaisala, Finland). 2.3. Chemical analyses A quarter of each quartz filter was cut and extracted using 25 ml of deionized water (Millipore, 18.2 M) and an ultrasonic bath at room temperature for 30 min. The extraction liquid was filtered using 0.22 ␮m filters and subsequently analyzed to determine the Na+ , NH4 + , K+ , Mg2+ , Ca2+ , Cl− , NO3 − and SO4 2− concentrations with an ion chromatograph (DIONEX, ICS-90, USA). Water-soluble organic carbon (WSOC) concentrations were analyzed using a Multi N/C 3000 Analyzer (Analytik Jena AG, Germany). After another quarter of each quartz filter was cut, a thermal/optical carbon aerosol analyzer (DRI Model 2001A, Desert Research Institute, USA) was used to examine the organic carbon (OC) and elemental carbon (EC) loaded on the samples. A quarter of the cellulose membrane was digested in a mixture of concentrated HNO3 (6 ml), HCl (2 ml) and HF (0.2 ml) using a closed vessel microwave digestion system (MARS5, CEM Corporation, Matthews, NC, USA). The concentrations of 14 trace elements (TEs) (Na, Mg, Al, K, Ca, V, Cr, Mn, Fe, Cu, Zn, As, Cd and Pb) in the digests were determined using Agilent 7500 ce inductively coupled plasma mass spectrometry (ICP-MS, Agilent Technologies, Tokyo, Japan). The extraction and analysis methods of all of the chemical species have been reported in previous studies [19–22]. Most of the information about the instruments (e.g., detection limit, precision and calibration) and the results of data quality control work can also be found in these cited papers. 3. Results and discussion 3.1. Mass concentrations of particles

2.2. Sample collection To obtain comprehensive chemical characteristics of the sizesegregated airborne particles, two nine-stage samplers (Andersen Series 20-800, USA) were used to simultaneously collect particles at a flow rate of 28.3 l min−1 with cutoff points of 0.43, 0.65, 1.1, 2.1, 3.3, 4.7, 5.8 and 9.0 ␮m. Quartz fiber filters (for water-soluble ions and carbonaceous species analysis) and cellulose membranes (for element analysis) were used as substrates in the Andersen samplers; both substrates had a diameter of 81 mm. Each set of the size-segregated samples was continuously collected for 24 or 48 h. Regular aerosol sampling was conducted for 48 h from 10:00 LT on Monday to 10:00 LT on Wednesday every week. When an air pollution episode occurred, we attempted to densify the sampling to capture the formation and dissipation processes of the haze pollution. Therefore, a relative short sampling time last for 24 h was performed. The quartz fiber filters were wrapped with aluminum foil and pre-heated at 800 ◦ C for 2 h to remove all organic material. The filters were subsequently conditioned in a dryer (temperature: 25 ◦ C; humidity: 10%) for 72 h before weighing. The filters were weighed before and after sampling on a microbalance with a balance sensitivity of ±0.01 mg. After re-weighing the samples, the exposed filters were divided into different portions (see below) with clean tools and prepared for the chemical analysis. Each filter was weighed three times before and after sampling, and the average value was used. Blanks and duplicate sample analyses were performed for approximately 10% of the samples. Blank filters were processed simultaneously with the field samples. In addition, the PM1 , PM2.5 and PM10 concentrations were measured with RP1400 instruments, which provide continuous direct mass measurements of particles through a tapered element oscillating microbalance (TEOM) [18]. Meanwhile, the meteorological parameters, including temperature, humidity, wind speed and

3.1.1. Online measurement of particulate matter The hourly concentrations of PM1 , PM2.5 and PM10 in Beijing, along with the meteorological parameters (i.e., temperature, RH, wind speed and direction), from Jan. 1 to Feb. 28, 2013, are depicted in Fig. 1. Surprisingly, there were only 4 days in which the daily average PM10 concentration was less than the Chinese National Ambient Air Quality Standard (GB3095-2012) of 50 ␮g m−3 (NAAQS, Grade I) (Table 1). This finding suggests that 27 days in January exhibited serious particulate pollution. In addition, 17 days had daily average PM10 values that exceeded the NAAQS Grade II level of 150 ␮g m−3 , which is harmful to human health [23]. Besides PM10 , the pollution from PM2.5 and PM1 was also severe in January. For example, the highest instantaneous 5-min

Fig. 1. Temporal variations of (a) ambient temperature and relative humidity; (b) wind direction and wind speed; (c) PM1 , PM2.5 and PM10 concentrations from Jan. 1 to Feb. 28, 2013.

454

S. Tian et al. / Journal of Hazardous Materials 279 (2014) 452–460

Table 1 Summary of concentrations of particles during the observation period from Jan. 1 to Feb. 28, 2013 in urban Beijing.

Jan Feb

PM10 (days/proportion)

PM2.5 (days/proportion)

>150 ␮g m−3

<50 ␮g m−3

>150 ␮g m−3

>75 ␮g m−3

<35 ␮g m−3

17/55% 12/43%

4/13% 3/11%

11/35% 8/29%

20/65% 15/54%

5/16% 6/21%

PM2.5 concentration reached 770 ␮g m−3 at 20:48 on Jan. 12, 2013. Moreover, the highest instantaneous PM2.5 concentration reached 1000 ␮g m−3 in some heavily polluted areas of Beijing [24]. To our knowledge, this is the first time that this level of PM2.5 pollution has been measured in Beijing. In general, there were 20 days with PM2.5 concentrations above the NAAQS of 75 ␮g m−3 . It increased to 26 days when compared with the U.S. EPA standard (daily average value of 35 ␮g m−3 ). Overall, the air quality in February was slightly better than in January. However, the number of days that the PM2.5 concentrations exceeded the NAAQS and EPA standards was 15 and 22, respectively. Additionally, the number of days with a PM10 concentration above the NAAQS Grade II and Grade I levels was 12 and 25, respectively. As shown in Fig. 1, the temporal variations in the PM1 were similar to those of PM2.5 during most of the observation period. Moreover, the PM1 contributed to 61% of the concentration of PM2.5 , which indicates that the fine particle pollution was primarily due to submicron aerosols in Beijing. 3.1.2. Filter measurement of particulate matter During the experiment period, 24 sets of size-resolved filter samples were collected and subjected to chemical analyses (Fig. 2). As shown, most of these samples were collected during January. Note that the heaviest haze pollution was not anticipated, and it is difficult to forecast the start or end times of the haze. Thus, some of the days with the heaviest PM2.5 concentrations were not sampled. However, the samples are still representative of the haze pollution considering that the concentration levels and temporal variations of PM2.1 and PM9 of the selected samples were consistent with those of PM2.5 and PM10 during January and February, as discussed in Section 3.1.3. Fortunately, several typical haze pollution episodes from Jan. 24 to Feb. 1 were continuously sampled during the study period. To examine the variations in mass concentrations and chemical compositions during the haze evolution process, the sampling days were divided into three categories based on the PM2.5 concentrations: heavily polluted days (HP, PM2.5 > 150 ␮g m−3 ), lightly polluted days (LP, 75 ␮g m−3 < PM2.5 < 150 ␮g m−3 ) and clear days

Fig. 2. Size-resolved mass concentration distributions that are marked as HP, LP and C denote heavily polluted days, lightly polluted days and clear days, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

(C, PM2.5 < 75 ␮g m−3 ). Based on meteorological data collected at Beijing Capital International Airport (http://english. wunderground.com), the visibilities on the HP, LP and clear days were <5, 5–10 and >10 km, respectively. From the total sizeresolved aerosol sample set, 9, 9 and 4 sets were collected for HP, LP and clear days, respectively (denoted in Fig. 2). In addition to the haze episodes, high PM pollution also occurred during fireworks (i.e., the Lantern Festival on Feb. 24, 2013) and during a dust event (Feb. 28, 2013). These two typical pollution days were also sampled for chemical analysis; however, these events are not discussed in this study. 3.1.3. Fine and coarse mode particle concentrations Table 2 describes the average, minimum and maximum concentrations of the size-resolved mass and chemical compositions on HP, LP and clear days. Because the sampler does not have cut-off sizes of 1.0, 2.5 and 10 ␮m, submicron, fine and coarse mode particles were defined as particles with sizes <1.1, <2.1 and 2.1–9.0 ␮m, respectively. There was a significantly linear correlation between PM2.1 and PM2.5 (R2 = 0.89, p < 0.05). A significantly linear correlation between PM9 and PM10 (R2 = 0.87, p < 0.05) was also observed. The PM1.1 concentration on HP days was 126.9 ␮g m−3 , which was approximately 1.5 times the concentration that was found in Shanghai (78.9 ␮g m−3 ) during a heavily polluted haze event in 2012 [25]. The average PM1.1 concentration on clear days was 36.3 ␮g m−3 , which was higher than that in most of abroad urban areas (e.g., it was approximately twice the concentration that reported in urban Barcelona and Athens [26,27]). The average PM2.1 concentration on HP days was 207.8 ␮g m−3 , which was approximately two and four times the average concentrations on LP (94.2 ␮g m−3 ) and clear days (52.0 ␮g m−3 ), respectively. Similar to PM2.1 , the average PM2.1–9 concentration (107.2 ␮g m−3 ) on HP days was also approximately twice the average concentration on LP days (52.9 ␮g m−3 ) and four times the average concentration on clear days (26.0 ␮g m−3 ). Therefore, it is evident that both fine and coarse particles substantially accumulated during the haze pollution period. 3.2. Chemical concentrations 3.2.1. Overall description As shown in Table 2, the average concentrations of OC, WSOC, EC, Na+ , NH4 + , K+ , Mg2+ , Ca2+ , Cl− , NO3 − and SO4 2− on HP days ranged from 0.3 to 39.1 ␮g m−3 for PM2.1 and from 0.4 to 15.9 ␮g m−3 for PM2.1–9 . On HP days, concentrations of these species in both the fine and coarse modes were higher than on LP and clear days, which indicates that the accumulation of these species could be responsible for the haze pollution formation [28,29]. In addition, the concentrations of OC, EC, SO4 2− , NO3 − and NH4 + were higher than those found in most haze pollution studies in Beijing and other Chinese cities [14,30]. Furthermore, the maximum concentrations of TEs on HP days were 2.3 ␮g m−3 for K (PM2.1 ) and 2.5 ␮g m−3 for Ca (PM2.1–9 ). With the exception of Al and Na, the TE concentrations in both the fine and coarse modes on HP days were greater than the concentrations on clear days. Due to the relatively stable meteorological conditions and the lower boundary layer height on haze days, it was difficult for these species to diffuse [31,32]. 3.2.2. Concentration enhancement ratios To focus our discussion on the roles of individual chemical species in the formation of haze pollution, concentration enhancement ratios from clear to HP days (HP/C) and from LP to HP days (HP/LP) were examined. All of the previously mentioned species were divided into three categories based on HP/C ratios. The HP/C mass concentration ratio was approximately 4 for both fine and

Table 2 Concentrations of different chemical compositions in size-resolved particles during HP, LP and C days (for Mn, Cu, Cr, AS, V and Cd in unit ng m-3 and for others in unit ␮g m−3 ). Heavily polluted days PM2.1–9

PM1.1

Clear days PM2.1

PM2.1–9

PM1.1

PM2.1

PM2.1–9

Mean ± SD

Min–Max

Mean ± SD

Min–Max

Mean ± SD

Min–Max

Mean ± SD

Min–Max

Mean ± SD

Min–Max

Mean ± SD

Min–Max

Mean ± SD

Min–Max

Mean ± SD

Min–Max

Mean ± SD

Min–Max

126.9 ± 55.8 23.4 ± 6.2 13.9 ± 4.5 4.2 ± 2.1 0.4 ± 0.3 14.1 ± 4.1 2.2 ± 3.4 0.2 ± 0.3 1.3 ± 2.0 3.3 ± 2.6 15.7 ± 3.3 16.1 ± 5.8 1.4 ± 0.5 0.9 ± 0.1 0.7 ± 0.3 0.4 ± 0.5 0.3 ± 0.1 0.3 ± 0.2 0.1 ± 0.1 0.1 ± 0.02 49.1 ± 24.4 41.2 ± 20.3 36.0 ± 4.0 34.7 ± 29.5 9.3 ± 5.2 3.9 ± 2.1

59.1–216.0 14.4–32.6 8.5–20.2 1.9–7.7 0.1–1.0 9.8–20.6 0.4–11.1 0.01–0.9 0.1–6.1 1.3–9.4 11.5–22.3 10.0–26.5 0.9–2.0 0.8–1.0 0.4–1.0 0.03–1.0 0.1–0.4 0.1–0.4 0.003–0.3 0.1–0.13 32.0–77.1 18.0–55.7 31.6–39.4 7.4–66.0 3.7–13.9 1.4–5.3

207.8 ± 73.2 39.1 ± 12.1 21.9 ± 8.5 5.5 ± 2.0 0.6 ± 0.4 24.1 ± 8.6 3.3 ± 4.3 0.3 ± 0.4 1.5 ± 2.2 5.0 ± 3.5 26.1 ± 8.4 33.3 ± 15.0 2.3 ± 0.6 1.4 ± 0.2 1.1 ± 0.3 0.9 ± 0.5 0.4 ± 0.1 0.5 ± 0.3 0.3 ± 0.1 0.2 ± 0.05 83.2 ± 22.6 66.9 ± 31.3 48.7 ± 8.5 59.8 ± 55.5 15.0 ± 4.7 6.9 ± 3.6

135.0–343.0 22.5–61.0 8.5–35.2 2.9–8.7 0.3–1.3 15.1–37.7 0.8–14.5 0.01–1.1 0.1–6.9 1.9–12.4 16.4–42.0 16.8–58.9 1.7–2.9 1.3–1.5 0.8–1.4 0.5–1.5 0.3–0.5 0.2–0.7 0.1–0.3 0.16–0.23 69.4–109.3 34.2–96.6 43.0–58.5 13.7–121.4 9.5–18.2 3.4–10.5

107.2 ± 48.0 13.7 ± 5.7 7.7 ± 2.7 1.8 ± 0.9 0.7 ± 0.4 10.4 ± 8.6 0.7 ± 0.7 0.4 ± 0.2 2.4 ± 1.1 1.8 ± 0.8 8.6 ± 4.7 15.9 ± 11.7 0.7 ± 0.4 3.6 ± 0.6 0.9 ± 0.5 2.4 ± 1.3 0.1 ± 0.03 0.3 ± 0.1 1.3 ± 0.5 0.7 ± 0.1 51.9 ± 20.4 30.3 ± 16.6 29.6 ± 17.6 21.0 ± 15.2 11.7 ± 5.0 2.3 ± 0.7

43.1–200.5 5.9–23.0 4.4–11.8 0.7–3.6 0.1–1.3 2.4–25.1 0.1–2.3 0.2–0.7 0.9–4.1 0.8–3.5 3.7–17.1 4.7–37.2 0.3–1.1 3.2–4.1 0.5–1.4 1.4–3.9 0.1–0.2 0.1–0.3 0.7–1.6 0.6–0.7 37.0–75.1 11.2–40.5 9.4–41.6 7.8–37.7 5.9–15.1 1.7–3.0

63.9 ± 23.1 15.7 ± 3.0 7.7 ± 1.9 2.1 ± 0.7 0.1 ± 0.1 7.9 ± 2.2 1.3 ± 1.3 0.1 ± 0.1 0.4 ± 0.4 1.9 ± 0.9 8.0 ± 2.4 8.6 ± 2.5

39.6–117.2 12.9–20.7 6.3–9.0 1.3–3.4 0.01–0.4 4.6–11.4 0.4–4.6 0.02–0.3 0.02–1.1 0.5–3.5 3.8–11.5 4.3–11.4

94.2 ± 29.2 21.7 ± 4.3 10.2 ± 2.7 3.0 ± 1.2 0.2 ± 0.2 11.0 ± 3.5 1.8 ± 1.6 0.2 ± 0.1 0.5 ± 0.6 2.5 ± 1.1 10.9 ± 3.8 12.9 ± 4.8

53.2–153.5 16.3–29.3 8.3–12.0 1.6–5.3 0.01–0.5 6.0–15.1 0.5–5.8 0.1–0.4 0.04–1.6 0.5–4.3 5.2–17.2 6.2–18.9

52.9 ± 26.5 5.4 ± 2.2 2.9 ± 1.3 0.6 ± 0.5 0.5 ± 0.4 1.8 ± 0.8 0.3 ± 0.3 0.3 ± 0.2 1.7 ± 1.3 0.9 ± 0.8 2.1 ± 1.3 2.8 ± 1.4

12.8–90.6 2.3–9.0 2.0–3.9 0.1–1.7 0.03–1.5 0.9–3.1 0.1–1.1 0.1–0.6 0.1–4.2 0.1–2.4 0.6–4.2 1.3–5.2

37.1 ± 19.5 7.9 ± 4.1 3.4 ± 1.6 0.9 ± 0.7 0.2 ± 0.4 2.8 ± 2.5 0.4 ± 0.4 0.1 ± 0.1 1.1 ± 0.5 0.7 ± 0.1 3.2 ± 1.6 2.8 ± 2.3 0.4 ± 0.2 1.1 ± 1.1 0.4 ± 0.01 0.3 ± 0.02 0.1 ± 0.1 0.1 ± 0.02 0.3 ± 0.4 0.1 ± 0.1 14.2 ± 5.9 8.6 ± 5.8 16.1 ± 7.9 3.0 ± 3.1 4.6 ± 0.7 0.5 ± 0.3

22.6–65.6 2.2–11.4 2.2–4.6 0.1–1.6 0.03–0.7 0.8–6.2 0.1–0.9 0.003–0.2 0.8–1.4 0.6–0.7 1.4–4.3 0.4–5.6 0.2–0.5 0.3–1.9 0.4–0.4 0.3–0.3 0.01–0.1 0.05–0.1 0.04–0.5 0.02–0.2 10.0–18.4 4.5–12.8 10.5–21.6 0.8–5.2 4.1–5.1 0.3–0.8

52.0 ± 28.1 9.5 ± 5.1 4.4 ± 2.9 1.2 ± 0.6 0.3 ± 0.5 4.9 ± 5.4 0.6 ± 0.5 0.1 ± 0.1 0.9 ± 0.9 1.0 ± 0.1 4.9 ± 3.0 5.0 ± 5.2 0.5 ± 0.3 1.1 ± 1.1 0.5 ± 0.0003 0.5 ± 0.1 0.1 ± 0.1 0.1 ± 0.03 0.4 ± 0.3 0.2 ± 0.2 24.9 ± 13.1 11.8 ± 7.6 16.3 ± 8.1 4.7 ± 5.5 4.6 ± 0.7 0.9 ± 0.6

22.6–88.0 2.3–13.9 3.0–7.1 0.3–1.7 0.03–0.9 1.1–12.6 0.1–1.2 0.01–0.3 0.01–1.8 0.9–1.1 1.8–7.8 0.5–12.2 0.3–0.7 0.3–1.9 0.5–0.5 0.4–0.6 0.02–0.2 0.06–0.1 0.1–0.6 0.1–0.3 15.7–34.2 6.4–17.2 10.5–22.0 0.8–8.6 4.1–5.1 0.4–1.3

26.0 ± 7.1 5.0 ± 3.2 1.8 ± 0.7 0.7 ± 0.6 0.5 ± 0.4 1.4 ± 1.1 0.2 ± 0.3 0.1 ± 0.1 1.4 ± 1.0 0.8 ± 0.3 1.2 ± 0.3 1.5 ± 1.6 0.3 ± 0.1 1.8 ± 2.1 1.3 ± 0.6 0.9 ± 0.2 0.01 ± 0.004 0.02 ± 0.01 0.7 ± 0.2 0.4 ± 0.1 16.2 ± 1.1 15.8 ± 9.4 19.4 ± 16.8 6.0 ± 5.0 6.1 ± 3.8 0.1 ± 0.03

18.8–33.1 1.7–9.4 1.3–2.3 0.2–1.5 0.1–1.0 0.5–2.9 0.01–0.7 0.01–0.2 0.6–2.9 0.4–1.0 1.0–1.6 0.2–3.8 0.2–0.4 0.3–3.3 0.9–1.7 0.7–1.1 0.005–0.01 0.01–0.02 0.6–0.9 0.3–0.4 15.4–16.9 9.2–22.5 7.5–31.3 2.5–9.6 3.4–8.8 0.1–0.14

S. Tian et al. / Journal of Hazardous Materials 279 (2014) 452–460

Mass OC WSOC EC Na+ NH4 + K+ Mg2+ Ca2+ Cl− NO3 − SO4 2− K Ca Na Fe Pb Zn Al Mg Mn Cu Cr As V Cd

Lightly polluted days PM2.1

PM1.1

Average concentrations of trace elements on heavily polluted days were the concentration of that during Jan. 28–31, 2013, and average concentrations of trace elements on clear days were the concentration of that from Jan. 31 to Feb. 2, 2013.

455

456

S. Tian et al. / Journal of Hazardous Materials 279 (2014) 452–460

coarse mode particles. Note that not all of the samples were performed for TEs analysis. Data of TEs on HP and clear days were the mean concentration during Jan. 28–31 and from Jan. 31 to Feb. 2, 2013, respectively. Simultaneously, the HP/C mass concentration ratio for the days with TE data was also approximately 4. Thus, this ratio was defined as the threshold at which the target species substantially accumulated or not. 3.2.2.1. Type I: SO4 2− , NO3 − , NH4 + , WSOC, Pb, Zn and Cd. Species belonging to the first category exhibited HP/C ratios in both the fine and coarse modes that were greater than 4. Among these species, SO4 2− , NO3 − and NH4 + had relatively high concentrations in the fine particle mode, i.e., 33.3, 26.1 and 24.1 ␮g m−3 , respectively, on HP days; these concentrations were 2.6, 2.4 and 2.2 times the concentrations observed on LP days, respectively, and 6.7, 5.3 and 5.0 times the concentrations on clear days, respectively. SO4 2− , NO3 − and NH4 + primarily originated from secondary pollution particles that were produced by the transformation of their SO2 , nitrogen dioxide (NO2 ) and NH3 precursors [33]. The pronounced increase in the concentration of these species from clear to haze days was primarily due to the faster heterogeneous formation of secondary aerosols [2,14]. For coarse mode particles, the average concentrations of SO4 2− , NO3 − and NH4 + were 15.9, 8.6 and 10.4 ␮g m−3 , respectively, on HP days. In addition, the HP/LP (5.7, 4.1 and 5.9) and HP/C (10.5, 7.2 and 7.1) ratios for SO4 2− , NO3 − and NH4 + , respectively, were higher than those of fine mode particles, which implies that SO4 2− , NO3 − and NH4 + in coarse mode particles also substantially accumulated on haze pollution days. Under high RH conditions, coarse mineral aerosols can react with acidic gas and form a liquid surface membrane that pushes more sulfate and nitrate onto the surface [34]. Simultaneously, the acidic liquid surface largely promotes the formation of secondary aerosols by increasing the absorption rate of the precursor material and increasing the reaction rate [35]. Furthermore, the NO3 − /SO4 2− mass ratio and NO2 /SO2 mixing ratio were used as indicators to examine the relative importance of mobile vs. stationary sources of sulfate and nitrate in the atmosphere [36,37]. In this study the average NO3 − /SO4 2− mass ratios decreased from clear (0.99 and 0.79, respectively) to LP (0.85 and 0.74, respectively) and to HP (0.78 and 0.54, respectively) days for both fine and coarse particles. These values were much higher than that measured during the 2001–2003 winters (0.49) [38], which indicates that the contribution of mobile sources increased in urban Beijing. However, the average NO2 /SO2 ratios slightly decreased from clear to LP days but remarkably increased from LP to HP days (Fig. S3(a)). This finding may suggest that NO2 mainly comes from local sources but SO2 may be partially transported into Beijing from surrounding areas [28]. Local sources played a more important role in the formation of particles due to the relatively stable meteorological conditions and low boundary layer heights on haze days. The ratios of NO3 − /SO4 2− and NO2 /SO2 showed different trends from LP to HP days. Although the enhancement ratios of the gas precursor NO2 from LP to HP days were larger than those of SO2 , the average NO3 − /SO4 2− mass ratios decreased due to the higher sulfate oxidation ratios than nitrate oxidation ratios at higher RH conditions on hazy days [39]. Most of important, it was recently found that high concentration of NOx promotes the conversion of SO2 to sulfate on heavy pollution days [40]. WSOC is an important fraction of OC due to its ability to alter the hygroscopic properties of aerosols. Moreover, light absorption by WSOC has been shown to be higher in Beijing than in other regions [41]. In this study, the average WSOC concentration was 21.9 ␮g m−3 in the fine particle mode on HP days, which was 2.2 and 4.3 times the concentrations on LP and clear days, respectively. Furthermore, this concentration was approximately three times

higher than that for PM2.5 in Shanghai [42] and nine times higher than that in Guangzhou [43]. The HP/C ratio of WSOC was much higher than the previous observations in Xi’an [30]. This result indicates that WSOC accumulated in both the fine and coarse modes during haze pollution events. The ratios of WSOC/OC for both fine and coarse particles increased from clear (0.46 and 0.49, respectively) to LP (0.47 and 0.55, respectively) and to HP (0.56 and 0.59, respectively) days. High concentrations and high proportions of hygroscopic chemical compositions, i.e., WSOC, SO4 2− , NO3 − and NH4 + , are beneficial for the aggravation of haze pollution. 3.2.2.2. Type II: OC, EC, Cl− , K+ , Cu, K and As. The second category was associated with species that exhibited HP/C ratios greater than 4 in the fine particle mode but less than 4 in the coarse particle mode. Previous studies have shown that EC, Cl− and K+ primarily originate from fuel and biomass burning particles [44,45] and accumulate in the fine particle mode during haze pollution events. The concentrations of EC, Cl− and K+ in the fine particle mode were 5.5, 5.0 and 3.3 ␮g m−3 , respectively, on HP days, which were 4.7, 5.0 and 5.6 times the concentrations on clear days, respectively. Similar to SO4 2− , NO3 − and NH4 + , OC had a relatively high concentration in the fine particle mode, i.e., 39.1 ␮g m−3 on HP days. The HP/C ratio for OC was 4.1, which was lower than that for EC (4.7). The OC/EC ratios for PM2.1 decreased gradually from 9.0 on clear days to 7.6 on LP days and to 7.1 on HP days. The decrease in OC/EC suggests that the contribution of secondary organic carbons (SOC) to OC decreased [46]. This finding is inconsistent with previous studies that have shown a larger contribution of SOC to PM2.5 during haze periods than on non-haze days [47]. The contribution of SOC was previously found to be more substantial due to the high RH during haze days [28]. However, in this study, a lower observed contribution of SOC on HP days may be explained as follows. A high concentration of particulate matter on haze days weakens the incident solar radiation; therefore, the atmospheric oxidization capability of organic matter is considerably weakened, and the SOC generation rate is reduced. As a consequence, the accumulation of OC from clear to HP days is weaker than that of EC, which is primarily due to the incomplete combustion of fuel emissions and is not affected by radiation. This conclusion is supported by similar findings observed for PM2.5 in Shanghai during the same period [12]. A closer examination of the size distribution suggests that the OC/EC ratios for PM1.1–2.1 increased gradually from 5.1 on clear days to 6.3 on LP days and to 11.9 on HP days. However, the OC/EC ratios for PM1.1 decreased gradually from 9.0 on clear days to 7.6 on LP days and to 5.6 on HP days. This finding suggests that the slower SOC generation rate in fine particles on HP days primarily occurred at diameters <1.1 ␮m instead of 1.1–2.1 ␮m. 3.2.2.3. Type III: Ca2+ , Mg2+ , Na+ , Na, Cr, Ca, Mg, Al, V, Fe and Mn. The species belonging to the third category exhibited HP/C ratios in both the fine and coarse particle modes that were less than 4. On HP days, the concentrations of Ca2+ , Mg2+ and Na+ in the fine particle mode were 1.5, 0.3 and 0.6 ␮g m−3 , which were 1.6, 1.8 and 3.0 times the concentrations observed on clear days, respectively. In addition, the HP/LP ratios for Ca2+ , Mg2+ and Na+ and the HP/C ratios for Ca2+ and Na+ in the coarse particle mode were lower than the ratios in the fine particle mode. Most of these species, which primarily originate from natural sources that may include soil dust or construction dust, had difficulty ascending in the atmosphere on hazy days due to the stable weather conditions and high RH [48]. Most of the TEs in the first two categories substantially accumulated in the fine or both the fine and coarse modes and were primarily attributed to anthropogenic emissions. However, the majority of TEs in the third category were attributed to natural sources and accumulated in the coarse particle mode.

S. Tian et al. / Journal of Hazardous Materials 279 (2014) 452–460

Concentration(dM/dlogDp)/μg m

-3

400 HP LP Clear

300

200

100

0 0.1

1

10 Dp/μm

Fig. 3. Size-resolved particle mass distributions.

3.3. Size distributions 3.3.1. Particle mass size distributions The average mass size distributions of aerosols on HP, LP and clear days from Jan. 1 to Feb. 28, 2013 are shown in Fig. 3 and are considered to exhibit unimodal, bimodal and trimodal distributions, respectively. The fine modes exhibited maxima at 1.1–2.1 ␮m on HP days, 0.65–1.1 ␮m on LP days and 0.43–0.65 ␮m on clear days. The coarse modes exhibited maxima at 3.3–4.7 ␮m on LP days and 4.7–5.8 ␮m on clear days. As shown in Fig. 3, the amplitude of the fine mode was larger on HP days than on LP and clear days, which suggests that the complexity of the aerosol components was largely influenced by pollution emissions on HP days. 3.3.2. Size distributions of chemical species The size distributions of chemical species on HP days can be divided into four categories. The average concentrations of the main components in different size fractions are presented in Fig. 4.

457

3.3.2.1. Type I: SO4 2− , NO3 − , NH4 + , WSOC, K, Mn, Cu, As, Cd and Pb. The species belonging to the first category were abundant in the fine size mode. SO4 2− , NH4 + and WSOC exhibited maxima at 1.1–2.1 ␮m, whereas the other species exhibited maxima at 0.65–1.1 ␮m. The results in Section 3.2 show that SO4 2− , NO3 − , NH4 + , WSOC, Pb and Cd exhibited similar concentration variations from clear to HP days. These species appear to represent coal and motor vehicles sources [21]. 3.3.2.2. Type II: OC, EC, Cl− and K+ . The chemical species in the second category were primarily concentrated in the fine mode from 1.1 to 2.1 ␮m, with a minor coarse mode component from 4.7 to 5.8 ␮m. The results in Section 3.2 show that OC, EC, Cl− and K+ exhibited similar concentration variations. These species may represent biomass burning sources because Cl− and K+ are good biomass burning tracers [49]. 3.3.2.3. Type III: Ca2+ , Mg2+ , Na+ , Na, V, Cr and Zn. The chemical compositions in this category exhibited typical bimodal distributions, and the amplitude of the fine mode was consistent with the coarse mode amplitude. The size distributions of Ca2+ and Mg2+ were different than those observed in previous haze pollution events in Beijing, which showed that the distribution of these species primarily peaked in the coarse mode [29]. This difference might be due to the high acidity of fine particles on HP days, which improves the solubility of calcium and magnesium [50]. Ca2+ , Mg2+ , Na+ , Na, Cr and V also exhibited similar concentration variations and may represent dust sources, including road surface dust and dust from long-range transport. 3.3.2.4. Type IV: Mg, Ca, Al and Fe. The species in the fourth category were primarily concentrated in the coarse mode from 3.3 to 5.8 ␮m, which likely represent natural sources such as soil dust or mechanical abrasion processes. 3.3.3. Peak shifts of the mass and chemical species From the mass size distributions, it can be concluded that the fine mode maximum shifted from 0.43–0.65 ␮m on clear days to

Fig. 4. Size distributions of typical chemical species in different categories.

458

S. Tian et al. / Journal of Hazardous Materials 279 (2014) 452–460 Table 3 Calculation methods of main components used in mass closure study. Calculation methods CM OM HM SS SIA UM

CM = SiO2 + Al2 O3 + CaO + Fe2 O3 + K2 O + Na2 O + MgO = 1.89[Al] + 1.66[Mg]n + 1.21[K] + 1.40[Ca]n + 1.43[Fe]n + 1.35[Na-ss-Na+ ] + 2.14[Si] OM = 1.4OC HM = Cu + Pb + Zn + Cd + As + Cr + V + Mn SS = [Na+ ] + [SS-Cl− ] + [SS-Mg2+ ] + [SS-Ca2+ ] + [SS-K+ ] + [SSSO4 2− ] = 3.246[Na+ ] SIA = [NSS-SO4 2− ] + [NO3 − ] + [NH4 + ] UM = PM-CM-OM-SS-SIA-EC-HM

The calculation method for secondary inorganic aerosol and sea salt can be found in [49]. Moreover, [NSS-SO4 2− ] refers to water-soluble SO4 2− apart from sea salt; [SS-Na+ ] refers to water-soluble Na+ in sea salt. The calculation method for crustal materials can be found in [50]; [Sin ], [Fen ], [Can ] and [Mgn ] were calculated based on the ratio to Al in the crustal materials. Fig. 5. Scatter plots of the RH and PM1.1 /PM2.1 ratios.

0.65–1.1 ␮m on LP days and to 1.1–2.1 ␮m on HP days. In addition, PM1.1 /PM2.1 decreased from 0.71 on clear days to 0.68 on LP days and to 0.61 on HP days, which provides further evidence of the transformation from submicron particles to fine particles during haze pollution events. A similar size distribution difference was observed between haze days and non-haze days in Guangzhou [14]. Moreover, OC, EC, SO4 2− , NO3 − and NH4 + exhibited a similar size distribution shift with particle mass, which implies that submicron particles were transformed into fine particles during haze pollution events. This transformation likely resulted from the hygroscopic growth of submicron particles and the formation of secondary particles [51]. The first reason for the size distribution shift was the high RH conditions on hazy days. The liquid water contents in the PM1.1–2.1 fraction on Jan. 24, 25 and 29, which represent clear, LP and HP days, respectively, were calculated based on the Extend Aerosol Thermodynamics Model (E-AIM, Model II) [52]. The liquid water content in PM1.1–2.1 increased from clear days to LP and to HP days, which indicates that submicron particles experienced hygroscopic growth during the haze pollution. By examining the corresponding RH for each aerosol sample, it was found that the PM1.1 /PM2.1 mass concentration ratio decreased in humid conditions (Fig. 5), e.g., on the heavily polluted haze days during Jan. 29–31, 2013. Similar to the mass concentrations, the PM1.1 /PM2.1 ratios for SO4 2− , NO3 − , NH4 + and OC, decreased as the RH increased, implying that the humidity affects the size distribution variations. For haze pollution that is associated with high RH, the aqueous phase on the aerosol surface provides a means for the rapid heterogeneous gas–liquid conversion of gaseous precursors to produce secondary inorganic aerosols in PM1.1–2.1 [51]. The formation of secondary aerosols was the second reason for the size distribution shift. The heterogeneous formation of secondary aerosols is associated with the abundance of gas precursors, such as SO2 , NO2 and volatile organic compounds (VOCs), and the liquid water content in aerosols [51]. Fig. S3(a) suggests that the precursors (SO2 , NO2 ) increased from clear to LP and to HP days. Previous studies have shown that the VOC concentrations are typically high in the urban area of Beijing [53]. The gas precursors are so abundant that the aerosol water content acts as the limiting factor for the heterogeneous uptake of gas precursors on the aerosol surface. Therefore, the heterogeneous formation of secondary aerosols is enhanced as the RH increases. Fig. S3(b) shows the apparent sulfate oxidation ratios (defined as SOR = SO4 2− /(SO4 2− + SO2 ), in molar) and nitrate oxidation ratios (defined as NOR = NO3 − /(NO3 − + NO2 ), in molar) in PM1.1–2.1 . SOR (R2 = 0.57, p < 0.05) and NOR (R2 = 0.41, p < 0.05) were found to be well correlated with RH. Moreover, as discussed in Section 3.2.2, the generation ratio of SOC in PM1.1–2.1 also increased from clear to HP days as the RH increased.

Secondary SO4 2− in the droplet mode (0.56–1.8 ␮m) was associated with the heterogeneous oxidation of SO2 under catalysis by Fe and Mn, which were emitted from local steel smelting and coal combustion [51]. In this study, the HP/C ratios in PM1.1–2.1 for Fe were higher than the ratios for Ca, Mg and Al, which indicates that the abundance of Fe with sizes of 1.1–2.1 ␮m was primarily emitted from anthropogenic sources instead of ground dust. In addition, K, Cu, Zn, Cd and Pb also exhibited similar size distribution variations with particle mass. This finding may be a result of these species being emitted from the same source or having similar reaction pathways as particles [54]. 3.4. Mass closure studies 3.4.1. Mass reconstruction method In order to discuss the relative contributions of the major chemical components in aerosols and assess the different sources of aerosol loading in the air, mass closure studies were adopted. Two kinds of particle mass concentration were considered. One was determined gravimetrically and the other was reconstructed by a suite of analyzed chemical components. Mass closure studies allow us to assess the differences in emission sources and processes controlling the aerosol composition. The chemical species were divided into the following six categories for the chemical mass closure analysis: secondary inorganic aerosol (SIA), organic matter (OM), crustal materials (CM), heavy metals (HM), elemental carbon (EC) and sea salt (SS). The calculation methods of the main components are shown in Table 3. 3.4.2. Contribution from different chemical components Fig. 6 shows the contributions of different chemical components to both fine and coarse mode particles on HP and clear days. SIA and OM dominated the fine particulate mass. Moreover, OM and SIA accounted for 62.6% of the total PM2.1 mass on HP days. The contribution of SIA to the total PM2.1 mass (36.2%) was greater than that of OM (26.4%), which implies that both primary and secondary particles had a significant contribution to the fine particle mass. The contribution of SIA to the fine particle mass was much higher on HP days than on clear days and higher than previous findings in Beijing during the winter of 2009 [55], which further suggests that more secondary particles are generated on HP days. CM, HM, EC and SS individually accounted for comparably small fractions of the total PM2.1 mass on both HP and clear days (in total 7.5% and 17.2%, respectively). CM alone accounted for 3.2% of the total PM2.1 mass on HP days and 10.6% on clear days. The higher percentage on clear days most likely resulted from a higher contribution of road surface dust, which is difficult to form on haze days due to stable weather conditions and high RH. A high contribution of CM was observed in PM2.1–9 , accounting for 30.2% of the total PM2.1–9 mass on clear days. The contributions

S. Tian et al. / Journal of Hazardous Materials 279 (2014) 452–460

459

Fig. 6. Contributions of different components to the total PM2.1 and PM2.1–9 masses on both HP and clear days.

of SIA and OM also increased substantially in the coarse mode particles. Unlike clear days, when the coarse particle mass was dominated by CM, the coarse particle mass on HP days was dominated by SIA, CM and OM. The contributions of CM, OM and SIA to the total PM2.1–9 mass were 14.8%, 17.9% and 28.8%, respectively. This finding suggests that anthropogenic chemical species also accumulated substantially on coarse mode particles during haze pollution events; moreover, these species can affect the ecological environment and further threaten human health through the deposition of coarse particles. Therefore, the mitigation of particles with a diameter larger than 2.1 ␮m cannot be neglected. Although EC and HM only accounted for approximately 1.7% and 0.5% of the total PM2.1–9 mass, respectively, on HP days, these species represented primarily toxic anthropogenic chemical species. As discussed in Section 3.2.2, the enhancement ratios of EC and heavy metals in coarse particles from clear to HP days were 3.8 and 5.6, respectively. This result further suggests the importance of controlling particles with aerodynamic diameter larger than 2.1 ␮m. 4. Conclusions Size-segregated airborne particles in Beijing during several of the heaviest haze periods from Jan. 1 to Feb. 28, 2013 were collected using nine-stage samplers. The mass concentrations, water-soluble inorganic ions, organic carbon, elemental carbon, water-soluble organic carbon and trace elements were analyzed. The major results and conclusions are as follows:

from submicron particles to fine particles during haze pollution events most likely resulted from the hygroscopic growth of submicron particles and the formation of secondary PM1.1–2.1 particles. (3) The mass closure analysis indicated that SIA and OM dominated the fine particulate mass. The contributions of SIA (36.2%) and OM (26.4%) were much larger on HP days than on clear days. Moreover, the contributions of CM, OM and SIA to the total PM2.1–9 mass were 14.8%, 17.9% and 28.8%, respectively, suggesting that anthropogenic chemical species accumulated substantially in the coarse mode during the haze pollution events. Furthermore, our findings demonstrate that the mitigation of particles with aerodynamic diameter greater than 2.1 ␮m cannot be neglected. Acknowledgments This study supported by the “Strategic Priority Research Program” of the Chinese Academy of Sciences (XDB05020000 and XDA05100100) and the National Natural Science Foundation of China (No.: 41230642 and 41021004). We gratefully acknowledge Lan Zhang in Capital Normal University for the trace elements analysis and the staff in Institute of Atmospheric Physics for instrument maintenance, especially for Yue Wang providing the data of NO2 and SO2 used in the supporting materials. Appendix A. Supplementary data

(1) High concentrations of water-soluble ions, carbonaceous species and heavy metals were found in both fine and coarse mode particles on haze days. Even on clear days, the mass concentration of submicron particles in Beijing was several times higher than concentrations previously observed in megacities in developed countries. With enhanced haze pollution, the OC/EC ratio in PM1.1 decreased, however, the OC/EC ratio in PM1.1-2.1 increased, which indicates the relative importance of solar radiation and RH on the SOC production. (2) The fine mode maxima shifted from diameters of 0.43–0.65 ␮m on clear days to 0.65–1.1 ␮m on LP days and to 1.1–2.1 ␮m on HP days. Such size shifts were also found for OC, EC, NH4 + , SO4 2− , NO3 − , K, Cu, Zn, Cd and Pb. The transformation of mass

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhazmat. 2014.07.023. These data include Google maps of the area around the sampling site in this article. References [1] Y. Sun, G. Zhuang, A. Tang, Y. Wang, Z. An, Chemical characteristics of PM2.5 and PM10 in haze-fog episodes in Beijing, Environ. Sci. Technol. 40 (2006) 3148–3155. [2] J.H. Tan, J.C. Duan, D.H. Chen, X.H. Wang, S.J. Guo, X.H. Bi, G.Y. Sheng, K.B. He, J.M. Fu, Chemical characteristics of haze during summer and winter in Guangzhou, Atmos. Res. 94 (2009) 238–245.

460

S. Tian et al. / Journal of Hazardous Materials 279 (2014) 452–460

[3] X. Zhuang, Y. Wang, H. He, J. Liu, X. Wang, T. Zhu, M. Ge, J. Zhou, G. Tang, J. Ma, Haze insights and mitigation in China: an overview, J. Environ. Sci. 26 (2014) 2–12. [4] E. Kang, J. Han, M. Lee, G. Lee, J.C. Kim, Chemical characteristics of size-resolved aerosols from Asian dust and haze episode in Seoul Metropolitan City, Atmos. Res. 127 (2013) 34–46. [5] B.K. Lee, G.H. Park, Characteristics of heavy metals in airborne particulate matter on misty and clear days, J. Hazard. Mater. 184 (2010) 406–416. [6] M. Tao, L. Chen, Z. Wang, P. Ma, J. Tao, S. Jia, A study of urban pollution and haze clouds over northern China during the dusty season based on satellite and surface observations, Atmos. Environ. 82 (2014) 183–192. [7] H. Che, X. Xia, J. Zhu, Z. Li, O. Dubovik, B. Holben, P. Goloub, H. Chen, V. Estelles, E. Cuevas-Agulló, L. Blarel, H. Wang, H. Zhao, X. Zhang, Y. Wang, J. Sun, R. Tao, G. Shi, Column aerosol optical properties and aerosol radiative forcing during a serious haze-fog month over North China Plain in 2013 based on ground-based sunphotometer measurements, Atmos. Chem. Phys. 14 (2014) 2125–2138. [8] L.T. Wang, Z. Wei, J. Yang, Y. Zhang, F.F. Zhang, J. Su, C.C. Meng, Q. Zhang, The 2013 severe haze over the southern Hebei, China: model evaluation, source apportionment, and policy implications, Atmos. Chem. Phys. 14 (2014) 3151–3173. [9] A. Boynard, C. Clerbaux, L. Clarisse, S. Safieddine, M. Pommier, M. Van Damme, S. Bauduin, C. Oudot, J. Hadji-Lazaro, D. Hurtmans, P.-F. Coheur, First simultaneous space measurements of atmospheric pollutants in the boundary layer from IASI: a case study in the North China Plain, Geophys. Res. Lett. 41 (2014) 645–651. [10] Y. Wang, L. Yao, L. Wang, Z. Liu, D. Ji, G. Tang, J. Zhang, Y. Sun, B. Hu, J. Xin, Mechanism for the formation of the January 2013 heavy haze pollution episode over central and eastern China, Sci. China D 57 (2014) 14–25. [11] Q. Liu, Y. Liu, Z. Yang, T. Zhang, Daily variations of chemical properties in airborne particulate matter during a high pollution winter episode in Beijing, Acta Sci. Circumst. 34 (2014) 12–18 (in Chinese). [12] M. Zhou, C. Chen, L. Qiao, C. Lou, The chemical characteristics of particulate matters in Shanghai during heavy air pollution episode in Central and Eastern China in January 2013, Acta Sci. Circumst. 33 (2013) 3118–3126 (in Chinese). [13] J.K. Zhang, Y. Sun, Z.R. Liu, D.S. Ji, B. Hu, Q. Liu, Y.S. Wang, Characterization of submicron aerosols during a month of serious pollution in Beijing, 2013, Atmos. Chem. Phys. 14 (2014) 2887–2903. [14] G. Zhang, X. Bi, L.Y. Chan, X. Wang, G. Sheng, J. Fu, Size-segregated chemical characteristics of aerosol during haze in an urban area of the Pearl River Delta region, China, Urban Clim. 4 (2013) 74–84. [15] L. Künzi, P. Mertes, S. Schneider, N. Jeannet, C. Menzi, J. Dommen, U. Baltensperger, A.S.H. Prévôt, M. Salathe, M. Kalberer, Responses of lung cells to realistic exposure of primary and aged carbonaceous aerosols, Atmos. Environ. 68 (2013) 143–150. [16] Y.P. Pan, Y.S. Wang, G.Q. Tang, D. Wu, Spatial distribution and temporal variations of atmospheric sulfur deposition in Northern China: insights into the potential acidfication risks, Atmos. Chem. Phys. 13 (2013) 1675–1688. [17] H. Cai, S. Xie, Traffic-related air pollution modeling during the 2008 Beijing Olympic Games: the effects of an odd-even day traffic restriction scheme, Sci. Total. Environ. 409 (2011) 1935–1948. [18] D. Ji, Y. Wang, L. Wang, L. Chen, B. Hu, G. Tang, J. Xin, T. Song, T. Wen, Y. Sun, Y. Pan, Z. Liu, Analysis of heavy pollution episodes in selected cities of northern China, Atmos. Environ. 50 (2012) 338–348. [19] X. Li, L. Wang, Y. Wang, T. Wen, Y. Yang, Y. Zhao, Y. Wang, Chemical composition and size distribution of airborne particulate matters in Beijing during the 2008 Olympics, Atmos. Environ. 50 (2012) 278–286. [20] Y. Pan, Y. Wang, J. Xin, G. Tang, T. Song, Y. Wang, X. Li, F. Wu, Study on dissolved organic carbon in precipitation in Northern China, Atmos. Environ. 44 (2010) 2350–2357. [21] X. Li, L. Wang, D. Ji, T. Wen, Y. Pan, Y. Sun, Y. Wang, Characterization of the sizesegregated water-soluble inorganic ions in the Jing-Jin-Ji urban agglomeration: spatial/temporal variability, size distribution and sources, Atmos. Environ. 77 (2013) 250–259. [22] Y. Pan, Y. Wang, Y. Sun, S. Tian, M. Cheng, Size-resolved aerosol trace elements at a rural mountainous site in Northern China: importance of regional transport, Sci. Total. Environ. 461–462 (2013) 761–771. [23] X. An, Q. Hou, N. Li, S. Zhai, Assessment of human exposure level to PM10 in China, Atmos. Environ. 70 (2013) 376–386. [24] J. Quan, X. Tie, Q. Zhang, Q. Liu, X. Li, Y. Gao, D. Zhao, Characteristics of heavy aerosol pollution during the 2012-2013 winter in Beijing, China, Atmos. Environ. 88 (2014) 83–89. [25] Y. Shi, J. Chen, D. Hu, L. Wang, X. Yang, X. Wang, Airborne submicron particulate (PM1 ) pollution in Shanghai, China: chemical variability, formation/dissociation of associated semi-volatile components and the impacts on visibility, Sci. Total. Environ. 473–474 (2014) 199–206. [26] N. Pérez, J. Pey, X. Querol, A. Alastuey, J.M. López, M. Viana, Partitioning of major and trace components in PM10 –PM2.5 –PM1 at an urban site in Southern Europe, Atmos. Environ. 42 (2008) 1677–1691. [27] S. Pateraki, D.N. Asimakopoulos, A. Bougiatioti, T. Maggos, C. Vasilakos, N. Mihalopoulos, Assessment of PM2.5 and PM1 chemical profile in a multipleimpacted Mediterranean urban area: origin, sources and meteorological dependence, Sci. Total. Environ. 479–480 (2014) 210–220. [28] B. Hou, G. Zhuang, R. Zhang, T. Liu, Z. Guo, Y. Chen, The implication of carbonaceous aerosol to the formation of haze: revealed from the characteristics and sources of OC/EC over a mega-city in China, J. Hazard. Mater. 190 (2011) 529–536.

[29] Z. Sun, Y. Mu, Y. Liu, L. Shao, A comparison study on airborne particles during haze days and non-haze days in Beijing, Sci. Total. Environ. 456–457 (2013) 1–8. [30] C. Cheng, G. Wang, B. Zhou, J. Meng, J. Li, J. Cao, S. Xiao, Comparison of dicarboxylic acids and related compounds in aerosol samples collected in Xi’an, China during haze and clean periods, Atmos. Environ. 81 (2013) 443–449. [31] X. Zhang, Y. Wang, W. Lin, Y. Zhang, X. Zhang, P. Zhao, Y. Yang, J. Wang, Q. Hou, H. Che, Changes of atmospheric composition and optical properties over Beijing 2008 Olympic monitoring campaign, B. Am. Meteorol. Soc. 90 (2009) 1633–1651. [32] J. Tan, J. Duan, K. He, Y. Ma, F. Duan, Y. Chen, J. Fu, Chemical characteristics of PM2.5 during a typical haze episode in Guangzhou, J. Environ. Sci. 21 (2009) 774–781. [33] Y. Wang, G. Zhuang, X. Zhang, K. Huang, C. Xu, A. Tang, J. Chen, Z. An, The ion chemistry, seasonal cycle, and sources of PM2.5 and TSP aerosol in Shanghai, Atmos. Environ. 40 (2006) 2935–2952. [34] W. Li, P. Li, G. Sun, S. Zhou, Q. Yuan, W. Wang, Cloud residues and interstitial aerosols from non-precipitating clouds over an industrial and urban area in northern China, Atmos. Environ. 45 (2011) 2488–2495. [35] Z. Liu, L.Y. Wu, T.H. Wang, M.F. Ge, W.G. Wang, Uptake of methacrolein into aqueous solutions of sulfuric acid and hydrogen peroxide, J. Phys. Chem. A 116 (2011) 437–442. [36] R. Arimoto, R. Duce, D. Savoie, J. Prospero, R. Talbot, J. Cullen, U. Tomza, N. Lewis, B. Ray, Relationships among aerosol constituents from Asia and the North Pacific during PEM-West A, J. Geophys. Res. 101 (1996) 2011–2023. [37] X. Yao, C.K. Chan, M. Fang, S. Cadle, T. Chan, P. Mulawa, K. He, B. Ye, The water-soluble ionic composition of PM2.5 in Shanghai and Beijing, China, Atmos. Environ. 36 (2002) 4223–4234. [38] Y. Wang, G. Zhuang, A. Tang, H. Yuan, Y. Sun, S. Chen, A. Zheng, The ion chemistry and the source of PM2.5 aerosol in Beijing, Atmos. Environ. 39 (2005) 3771–3784. [39] Y. Sun, Q. Jiang, Z. Wang, P. Fu, J. Li, T. Yang, Y. Yin, Investigation of the sources and evolution processes of severe haze pollution in Beijing in January 2013, J. Geophys. Res. D 119 (2014) 4380–4398. [40] H. He, Y. Wang, Q. Ma, J. Ma, B. Chu, D. Ji, G. Tang, C. Liu, H. Zhang, J. Hao, Mineral dust and NOx promote the conversion of SO2 to sulfate in heavy pollution days, Sci. Rep. 4 (2014), http://dx.doi.org/10.1038/srep04172. [41] Z. Du, K. He, Y. Cheng, F. Duan, Y. Ma, J. Liu, X. Zhang, M. Zheng, R. Weber, A year long study of water-soluble organic carbon in Beijing II: light absorption properties, Atmos. Environ. 89 (2014) 235–241. [42] J. Feng, M. Li, P. Zhang, S. Gong, M. Zhong, M. Wu, M. Zheng, C. Chen, H. Wang, S. Lou, Investigation of the sources and seasonal variations of secondary organic aerosols in PM2.5 in Shanghai with organic tracers, Atmos. Environ. 79 (2013) 614–622. [43] H. Huang, K.F. Ho, S.C. Lee, P.K. Tsang, S.S.H. Ho, C.W. Zou, S.C. Zou, J.J. Cao, H.M. Xu, Characteristics of carbonaceous aerosol in PM2.5 : Pearl Delta River Region, China, Atmos. Res. 104–105 (2012) 227–236. [44] J.E. Dibb, R.W. Talbot, S.I. Whitlow, M.C. Shipham, J. Winterle, J. McConnell, R. Bales, Biomass burning signatures in the atmosphere and snow at Summit, Greenland: an event on 5 August 1994, Atmos. Environ. 30 (1996) 553– 561. [45] M.S. Bae, J.J. Schauer, J.R. Turner, P.K. Hopke, Seasonal variations of elemental carbon in urban aerosols as measured by two common thermal-optical carbon methods, Sci. Total. Environ. 407 (2009) 5176–5183. [46] J. Cao, S. Lee, J.C. Chow, J.G. Watson, K. Ho, R. Zhang, Z. Jin, Z. Shen, G. Chen, Y. Kang, Spatial and seasonal distributions of carbonaceous aerosols over China, J. Geophys. Res. D 112 (2007), http://dx.doi.org/10.1029/2006JD008205. [47] F. Zhang, L. Xu, J. Chen, X. Chen, Z. Niu, T. Lei, C. Li, J. Zhao, Chemical characteristics of PM2.5 during haze episodes in the urban of Fuzhou, China, Particuology 11 (2013) 264–272. [48] Y. Wang, G. Zhuang, Y. Sun, Z. An, The variation of characteristics and formation mechanisms of aerosols in dust, haze, and clear days in Beijing, Atmos. Environ. 40 (2006) 6579–6591. [49] H. Du, L. Kong, T. Cheng, J. Chen, J. Du, L. Li, X. Xia, C. Leng, G. Huang, Insights into summertime haze pollution events over Shanghai based on online watersoluble ionic composition of aerosols, Atmos. Environ. 45 (2011) 5131–5137. [50] B. Kunwar, K. Kawamura, One-year observations of carbonaceous and nitrogenous components and major ions in the aerosols from subtropical Okinawa Island, an outflow region of Asian dusts, Atmos. Chem. Phys. 14 (2014) 1819–1836. [51] X. Wang, W. Wang, L. Yang, X. Gao, W. Nie, Y. Yu, P. Xu, Y. Zhou, Z. Wang, The secondary formation of inorganic aerosols in the droplet mode through heterogeneous aqueous reactions under haze conditions, Atmos. Environ. 63 (2012) 68–76. [52] S.L. Clegg, P. Brimblecombe, A.S. Wexler, Thermodynamic model of the system H+ -NH4 + -SO4 2− -NO3 − -H2 O at tropospheric temperatures, J. Phys. Chem. A 102 (1998) 2137–2154. [53] Y. Wang, X. Ren, D. Ji, J. Zhang, J. Sun, F. Wu, Characterization of volatile organic compounds in the urban area of Beijing from 2000 to 2007, J. Environ. Sci. 24 (2012) 95–101. [54] J. Duan, J. Tan, J. Hao, F. Chai, Size distribution, characteristics and sources of heavy metals in haze episod in Beijing, J. Environ. Sci. 26 (2014) 189– 196. [55] R. Zhang, J. Jing, J. Tao, S.C. Hsu, G. Wang, J. Cao, C.S.L. Lee, L. Zhu, Z. Chen, Y. Zhao, Z. Shen, Chemical characterization and source apportionment of PM2.5 in Beijing: seasonal perspective, Atmos. Chem. Phys. 13 (2013) 7053–7074.