Journal Archive

A visual spot test is a traditional manual method for the detecting of ionic components (e.g., sulfate, nitrate, and chloride) on individual particles (Mamane and Pueschel, 1980; Ayers, 1977; Bigg et al., 1974). This method has apparently not been widely used for particle analysis probably because of difficulties in preparing films, in controlling reactions, in standardizing of the reaction, and in time consuming of the reading. However, this method is still useful to identify the internal mixing state of individual particles with gaseous and secondary aerosol components.

In this study, in order to identify the nitrate, sulfate, mixture of nitrate/sulfate, chloride, and mixture of chloride/nitrate on individual particles collected on AD event, two different dual thin-films were prepared. The dual thin-films (20 nm thickness) of BaCl₂/nitron (C₂₀H₁₆N₄) and AgF/nitron were prepared by dual simultaneous vacuum deposition of the reagents onto polycarbonate films. And then, two kinds of dual thin-films (BaCl₂/nitron and AgF/nitron) were placed on each stage of the impactor. Details of the preparation of two different dual thin-films can be found in elsewhere (Yamasaki et al., 2003).

Byunsan peninsula of Korea was selected as the sampling site of size-classified particles (Fig. 1). Because of its closeness to China, this site, marked as an empty star in Fig. 1, is the well suited area to measure AD.

Fig. 1. 
Maps showing sampling location (an empty star) of Byunsan Peninsula (an empty square).

Fig. 2 shows the flows of particle sampling and chemical analysis of individual and bulk particles.

Fig. 2. 
Flows of sampling and chemical identification of individual and bulk particles.

For the sampling of aerosols as a function of their size (8-step classification between 0.01 μm and 11 μm), an Andersen air sampler (Tokyo Dylec Co., AN-200) was operated at the ground based site (35.37N; 126.27 E, 30 meters above the sea), indicated by an empty star in Fig. 1.

Quartz filters were put on all stage of Andersen air sampler. Meanwhile, two different dual thin-films (BaCl₂/nitron and AgF/nitron) were arranged on the cutoff diameter of five-size steps, i.e., the stage with 7.65 μm, the stage with 5.07 μm, the stage with 3.45 μm, the stage with 2.05 μm, and the stage with 1.17 μm. In addition, for the purpose of identifying the elemental components in individual particles a polycarbonate filter was put on the stage with cutoff diameter of 3.45 μm.

Our sampling equipment mounting various filter media was intensively operated during AD event (Mar. 30, 2004) and non-AD day (May 26, 2004), respectively. In order to avoid the pile-up of particles, which obscures the chemical determination of particle-toparticle variations, the duration of sampling was adjusted to 20 minutes and 60 minutes at each stage of sampler. After sample collection, every sample was placed in a clean sterilized petridish, they were then sealed with Teflon tape and wrapped with aluminum foil. Every sample was placed in a cold storage bag during air transportation.

On AD and non-AD field campaign days the ranges of wind speed were 2.1-7.8 m s^-1 and 1.2-4.3 m s^-1, respectively and it was generally blowing from the west. The temperatures of AD and non-AD days were around 3.0-12.9°C and 16.7-27.2°C, respectively.

Fig. 3 shows a schematic diagram of the beam scanning and data acquisition system of micro-PIXE employing double X-ray detectors. The HP-Ge X-ray detector has moderate energy resolution and poor detection efficiency below 2 keV because of its Ge L-shell absorption edge and backscattering proton absorber (60 μm thick polypropylene). The Si(Li) X-ray detector was set at a symmetrical position with the HP-Ge X-ray detector with respect to the beam axis. The energy resolution of the Si(Li) detector is excellent, and this higher energy resolution results in a better signal-to-background ratio and smaller peak overlapping, especially in the low-energy regions. The detector window is 8 μm-thick Be and attached with an annular type absorber (100 μm-thick Mylar) with a center hole (3 mm in diameter). This Si(Li) detector can finally provide a fairly good detection efficiency for X-rays below 2 keV. Details of this double X-ray detectors system for PIXE Analysis have been described elsewhere (Sakai et al., 2005).

Fig. 3. 
A schematic diagram of the beam scanning and data acquisition system of micro-PIXE system consisting of the advanced HP-Ge : Si(Li) X-ray detectors.

The concentration of carbon was determined from the particles deposited on quartz filters using the TOR^® (DRI) Method. Two 0.64 cm² punches were taken from the quartz filter set on each stage of Andersen sampler and placed in the analyzer. Organic carbon (OC) was defined as all carbon that evolves from the sample without added oxygen when heated up to 550°C. Two additional temperature steps of 700°C and 800°C are made. Elemental carbon (EC) was defined as all carbon that evolves from the sample when heated up to 800°C in 2% oxygen in 98% helium atmosphere after the OC was removed. This TOR^® method is a well-accepted technique in which the sample is progressively pyrolyzed with continuous detection of evolved carbon. Chow et al. (1993) gave a full detail of TOR^® method.

In order to determine the source region of aerosols at the receptor, i.e., Byunsan peninsula, the atmospheric backward dispersion model was applied. Fig. 4 displays the backward aerosol dispersion simulated by the NOAA Air Resources Laboratory (ARL) HYSPLIT dispersion-trajectory model “backwards”. A detailed model description of HYSPLIT was given by Rolph (2003).

Fig. 4. 
Backward trajectories at 20 UTC 30 Mar., 2004 drawn by the NOAA HYSPLIT model. Source point (Byunsan) and label interval are 35.37°N; 126.27°E and 6 hours, respectively.

According to this model results, the air parcels of AD storm on Mar. 30, 2004 was coming from the desert and loess regions in eastern China and southern Mongolia through urban Beijing metropolis by the prevailing winds originated from northwest. And it was extended to the Korean Peninsula after passing through the Yellow Sea.

In order to promote the reactions between the dual thin-film and the aged ionic components, the dual thin-film loading particles has to be exposed to the adequate solvent vapor atmosphere (Mamane and Pueschel, 1980; Ayers, 1977; Bigg et al., 1974).

In the present study, the grids of the dual thin films of BaCl₂/nitron and AgF/nitron were exposed to the vapors of ethanol and prophanol/ethanol under room temperature for 24 h. After leaving of thin-films from alcohol environment, the morphologies of spot were photographed by a SEM (HITACHI, S-3000H).

Sulfate and nitrate on and/or in individual particles can be identified by the appearance of the Liesegang ring (i.e., the product of sulfate and barium chloride reaction) and the needle-like bundle (i.e., the product of nitrate and nitron reaction), respectively. Meanwhile, the particles containing chloride and nitrate collected on the dual thin-film of AgF/nitron form AgCl ring and needle crystal, respectively.

Fig. 5 illustrates a BaSO₄ ring formed on BaCl₂ thin-film, a NO₃ needle crystal formed on nitron thin-film, and the mixture of ring and needle formed on the dual thin-film of BaCl₂/nitron. The mixed AgCl ring and needle crystal was also formed on the dual thin-film of AgF/nitron.

Fig. 5. 
BaSO₄ ring formed on BaCl₂ thin film (left upper), NO₃ needle crystal formed on nitron thin film (left down), the mixed ring (sulfate) and needle (nitrate) (right upper) formed on the dual thin film of BaCl₂/nitron, and the mixed AgCl ring (chloride) and needle crystal (nitrate) (right down) formed on the dual thin film of AgF/nitron.

Fig. 6 shows the variation of relative number fraction (the number of each ionic species/300×100) of each ionic composition and their mixtures identified from the size-classified individual particles collected during AD event. A total of 1,500 individual particles (300 particles from each fraction of 5-step particle size) deposited on the dual thin-films placed on each stage of Andersen sampler was identified. The needle shape crystals formed by nitrate on both the dual thin-films of BaCl₂/nitron and AgF/nitron were averaged. The strong particle size-dependent variations were found for the content of ionic compositions and their mixtures. Sulfate showing the maximum number fraction was principally enriched in the particles with big size (PM_10.85-7.65). It seems reasonable to say that the enrichment of sulfate in coarse fraction was originated from the heterogeneous mixing with gaseous SO₂ as well as the condensation of H₂SO₄ (or fine sulfate particle) onto dust particles. While on the other hand, a large fraction of sulfate in PM_3.45-2.05 was probably due to the cohesion between fine particles which were formed by a gas-to-particle conversion in the atmosphere or the condensation of gases vaporized in the combustion process.

Fig. 6. 
Relative number fraction of each composition type identified from the size-classified particles collected during AD event at the Byunsan Peninsula by two kinds of dual thin-films. A total of 300 individual particles were identified at each particle size for the two different films.

Throughout the one-year observations of size distribution characteristics of major inorganic ions in aerosols at a coastal receptor site in Hong Kong, Bian et al. (2014) reported that the dominant presence of sulfate was found in the submicron fraction and its mass fraction in ~0.8 μm was 78-86.9%. They also suggested that nitrate was closely associated with the coarsemode dust particles as a result of its formation process through the reactions of acidic HNO₃ gas with alkaline components.

Owing to particle sampling considering the limitation of our morphological spot test, unfortunately, a full explanation of ionic components in the whole size range was not realized in this study.

Through an atmospheric chemistry model study, Xiao et al. (1997) reported that the chemical conversion of SO₂ to sulfate on the surface of dust particles might contribute as much as 20-40% of total sulfate in AD particles.

We cannot completely deny the possibility of the presence of mineral CaSO₄ contained in coarse mode AD particles. However, even if it existed in the nonaged AD particles, it had almost a negligible impact on our results because the soluble fraction of sulfur is barely existed in the original desert sand of AD (Ma et al., 2008).

Ishizaka et al. (2009) reported that the major Pb components for the AD particles collected in Japan were PbO, PbSO₄, PbCl₂, and PbCO₃, with molar percentages of 44%, 30%, 21%, and 5%, respectively. This indicates that the Liesegang rings appeared from this PbSO₄ contained in AD particles.

The average number fraction of coarse particles (>2.05 μm) containing chloride was 16.2%. This was a little bit higher than that (12.3) of fine particles (PM_2.05-1.17). The coarse fraction chloride was probably originated from sea-salts and the mixture of sea-salts and mineral particles. Elsewhere, as a possible source of the coarse fraction chloride, it can be thought that the artificially emitted gaseous Cl into urban atmosphere such as coal-fired power plants, automobiles, and municipal incinerators or the fine particles containing Cl (e.g., NH₄Cl) were absorbed or coalesced with coarse fraction dust particles during long-range transport.

Meanwhile, the nitrate containing particles did not show a strong particle-size depending and they accounted for only 4.3-5.3% through whole particle size. This result was found to be relatively low compared to that of Zhang and Iwasaka (1999). They reported that nitrate was found in some AD particles collected in Beijing, China and the average frequency of nitrate-containing AD particles was 10.8%.

The well-known heterogeneous reactions between gaseous nitrate and coarse particles are below chloride loss and carbonate loss processes.

NaCl_{(aq, s)}+HNO_3(g) →NaNO_{3(aq, s)}+HCl_(g)CaCO_3(s)+2HNO_3(g) → Ca(NO₃)_2(s)+H₂O+CO_2(g)

However, there are many unanswered questions concerning heterogeneous reactions to form aqueous droplets of Ca(NO₃)₂ from Ca-rich dust particles under atmospheric conditions, largely owing to a lacking of information on the amount of atmospheric gases and the amount of water vapor needed on particle surface as reactant in the heterogeneous reactions.

The mixtures of nitrate/chloride and nitrate/sulfate showed a very small number fraction (i.e., 0.3-1.7% and 1.3-1.7%, respectively). They have negligible number fractions relative to other ions.

Fig. 7 shows the relative frequency of each composition type identified from the particles collected on a non-AD day. With the exception of nitrate, the frequencies of sulfate and chloride containing particles were much smaller in non-AD period than on AD event. An exceptionally high difference in coarse fraction between two periods suggests that sulfate and chloride were intensely formed on dust particles during their transport from source areas to the Byunsan Peninsula. On the other hand, the frequency of particles involving nitrate increased in a non-AD day in comparison with AD event.

Fig. 7. 
Relative number fraction of each composition type identified from the size-classified particles collected during non-AD period at the Byunsan Peninsula by two kinds of dual thin-films. A total of 300 individual particles were identified at each particle size.

The fossil fuel source dominates the NO_x budget in industrialized areas, while the soil NO emissions control NO_x budgets in remote and rural areas (Delmas et al., 1997). From the study of distinguishment between the dominant sources of NO_x emissions: anthropogenic (fossil fuel and biofuel) emission and biomass burning and soil emissions, Sheel et al. (2010) suggested that higher temperature leads to more soil emission of NO_x. Therefore, it is also speculated that the high loading of nitrate at Byunsan in the end of May (the maximum temperature on 26 May was 27.2°C) could be due to more local emission of NO_x such as soil emission.

The fact that the internally mixed chloride with nitrate was identified only on AD event is good evidence that the increasing of the number fraction for chloride containing particles on AD event was derived from the aging of dust particles with chloride during long-range transport.

The aging of AD particles by secondary pollutants is discussed above. Not only these, but many kinds of hazardous metallic components generated from small and large scaled coal combustion facilities including power station contribute to the chemical transformation of AD particles.

Although the standard desert samples of China loess contain an infinitesimal quantity of zinc (59-79 μg g^-1) (Sino-Japan Friendship Centre for Environmental Protection), zinc is a good measure of the coal combustion, automobile, and refuse incineration. This has also practical importance in evaluating the health effect of AD particles. The study on the health effects of the metals bound to AD particles conducted by Hong et al. (2010) suggested that the exposure of the AD particles containing metals reduces children’s pulmonary function.

In this study, an attempt was made to verify the internal mixing of AD particles with zinc based on the visual interpretation of micro-PIXE elemental maps. There are few data regarding the source profile of zinc released to the atmosphere. Although zinc is removed from atmosphere by dry deposition and wet scavenging, however, artificial zinc particles with small sizes and low densities can suspend in the atmosphere and then travel a long distance from emission sources.

Fig. 8 illustrates the elemental maps of silica and zinc and their overlapped elemental mask drawn on and/or in individual coarse (PM_5.07-3.45) particles collected during AD event. The encircled particles on the overlapped elemental mask indicate the internal mixing of AD particles with zinc. As shown in Fig. 8, a large number of AD particles were internally mixed with zinc.

Fig. 8. 
Elemental maps of silica and zinc (left) on individual coarse fraction (PM_5.07-3.45) particles collected during AD event drawn by the micro-PIXE analytical system (scanning area of microbeam: 450×450 μm²) and their overlapped elemental mask (right).

However, since, as mentioned above, a tiny amount of zinc was detected from the natural and the standard desert samples of China loess, it is necessary to make clear about the sources of the zinc internally mixed with AD particles. To solve this, we investigate correlationships between the micro-PIXE net counts of zinc and those of cobalt. There was a strong correlation between the two with 0.96 R score. Cobalt was not detected from both the natural Asian desert sand (Ma et al., 2008) and the standard desert samples of China loess (Sino-Japan Friendship Centre for Environmental Protection). It could therefore be suggested that zinc contained in numerous particles were originated from artificial sources.

Primary anthropogenic sources of zinc include fossil fuel and waste combustion, and vehicular and aircraft exhausts (Barceloux and Barceloux, 1999). The internal mixture of AD particle with zinc containing particle was probably produced by several mechanisms, such as Brownian coagulation, impaction by differential sedimentation, and coalescence of cloud droplets containing AD particles with those containing zinc particles.

Size-resolved concentrations for EC, OC, and TC concentrations are given in Fig. 9. As shown in Fig. 9, a strong bimodal size distribution of EC and OC was found. Both EC and OC concentrations have peaks at 0.01-0.43 μm particle aerodynamic diameter.

Fig. 9. 
EC and OC in size-resolved PM collected during AD event.

Wang et al. (2010) measured the size distributions of EC and OC collected in Jiading District, Shanghai. According to their study, the size distributions of both EC and OC showed a bi-modal with peaks in the particles with size of <0.49 μm at one of the most urbanized cities in China. Compared with OC, EC was preferably enriched in this particle fraction. The similar size distribution of carbonaceous compositions was also reported in the former tunnel study (Allen et al., 2001). Allen et al. (2001) reported that the aerosol mass concentration mainly composed of EC and OC showed a peak at 0.1-0.2 μm particle aerodynamic diameter collected in the heavy traffic tunnel located in the San Francisco Bay area.

EC concentration in submicron size fraction of PM measured in this study can be compared to that of measured at Gosan background site in Korea during the Asian Pacific Regional Aerosol Characterization Experiment (ACE-Asia) intensive field campaign. EC concentration in PM₁ in this study was measured as 4.59 μg m^-3. Meanwhile, daily mass concentrations of EC at Gosan during non-AD and AD were 0.01-0.9 μg m^-3 and 0.9-1.6 μg m^-3, respectively (Chuang et al., 2003). EC concentration at the Byunsan site of this study was about 4 times higher than that measured at Gosan during AD event. This indicates that the EC concentration of Byunsan site of this study was greatly influenced by other sources. In general, the primary sources of carbonaceous components are vehicle exhaust and industrial emissions. Especially, EC concentration is highest in heavy traffic areas.

One of the reliable causes for high EC concentration at Byunsan site was might be the Saemangeum Seawall Project that was being conducted in our field measurement period. This project completed in April 2006 was the world’s longest man-made dike, measuring 33 kilometers and 400 square kilometers of farmland and a freshwater reservoir (Jajuminbo, 2010). It was might be expected that the enormous volume of mobilized equipments including heavy dump trucks and dredging equipments emitted a huge amount of carbon into the local atmosphere.