Structural analysis of aerosol particles by microscopic observation using a time-of-flight secondary ion mass spectrometer



[1] The chemical composition and structure of fine aerosol particles with diameters of less than 1 µm were analyzed in the spring of 2011 at Fukue Island, Japan, using an aerosol quadrupole mass spectrometer (Q-AMS, Aerodyne Research, Inc.) and a time-of-flight secondary ion mass spectrometer (TOF-SIMS). The Q-AMS results were similar to those of previous studies conducted at the same location, suggesting that the composition we observed is typical of this site. Based on the TOF-SIMS results, we classified the fine aerosol particles into three types: Type A, in which sulfate was covered with organic matter (OM); Type B, in which soil-containing particles with a diameter of 0.5 µm were associated with sulfate and OM; and Type C, in which black carbon (BC) and sulfate aggregates were associated with OM. During the observation period, the relative abundances of Type A, B, and C particles were 55%, 20%, and 25%, respectively. The structure, chemical composition, and the proportion of each type of particles provide information for a more representative particle model in radiative models. The relative abundance of Type C, i.e., BC-containing particles, is quite different to that in Tokyo, suggesting that Type C could be an indicator of transboundary air pollution, in this case from mainland China.

1 Introduction

[2] There has been very rapid development in East Asian countries, and as a result, the emission of atmospheric pollutants has increased [Ohara et al., 2007]. The major species emitted into the atmosphere are sulfur dioxide (SO2), nitrogen oxides (NOx), and volatile organic compounds (VOC). The emissions of these species across Asia in 2000 were estimated to be about 40, 25, and 40 Mt y−1, respectively. With the reference scenario, it has been predicted that in 2020, these emissions will be 50, 35, and 80 Mt y−1, respectively. Among the Asian countries, emissions of SO2, NOx, and VOC from China are largest and are predicted to be 27, 16, and 35 Mt y−1, respectively, in 2020 [Ohara et al., 2007]. A recent study has shown that from 2000 to 2006, total SO2 emissions in China increased by 53%, i.e., from 22 to 33 Mt y−1, although the total SO2 emissions in China have tended to decrease since 2006 [Lu et al., 2010]. The annual NOx emissions from coal-fired plants in China were estimated to be 8.1 Mt for 2005 and 9.6 Mt for 2007, which was validated using NO2 data from the improved ozone monitoring instrument and the nested-grid Goddard Earth Observing System–Chem model [Wang et al., 2012]. Overall, emissions in East Asia are currently considered to be high.

[3] Owing to prevailing westerly winds in the winter-spring season, gas and aerosol are transported from the Asian continent to the Pacific Ocean. There have been many studies concerning the detection of gas and aerosol outflow from mainland China, such as PEACAMPOT [Hatakeyama et al., 2001], PEM-West B [Arimoto et al., 1997], ACE-Asia [Huebert et al., 2003], TRACE-P [TRACE-P Science Team, 2003], PEACE [Parrish et al., 2004], Atmospheric Particulate Environment Change Studies [Takami et al., 2005], and W-Pass [Furutani et al., 2011]. In previous studies, we measured the chemical composition of aerosol using a filter sampling method and an aerosol mass spectrometer (Q-AMS) at the Cape Hedo station in Okinawa and at the Fukue station in Nagasaki, Japan [Murano et al., 2000; Shimohara et al., 2001; Takami et al., 2005, 2006, 2007; Takiguchi et al., 2008]. In these studies, it was found that the sulfate concentration was high when the air mass was transported from the Chinese coastal region (between Shanghai and the Shangdon Peninsula) to the East China Sea area by a high-pressure system, after the frontal system there in the East China Sea area had moved eastward. The size distributions of sulfate, ammonium, nitrate, and organics measured using Q-AMS had a similar peak mode diameter (mode diameter Dva = 400–500 nm depending on the observation site), indicating that they were internally mixed. The organics were highly oxidized (Q-AMS f44 = 0.15–0.2, where f44 denotes the ratio of mass 44 to organics signal) when observed on the remote islands, Fukue and Okinawa, and that aerosol processing and ageing had occurred during transport from the Chinese continent to Japan.

[4] The mixing state and chemical transformation of aerosol particles influence direct radiative forcing (DRF) [Adachi et al., 2010] and regional air pollution. The direct effect is a change in radiative forcing through their absorption and scattering of solar radiation. The calculated DRF values of the particles can differ by almost a factor of 2 [Bond et al., 2006], since the optical properties depend on the shape and position of black carbon (BC) in the host materials [Adachi et al., 2010; Cappa et al., 2012; Sedlacek et al., 2012].

[5] Individual aerosol particles have been analyzed by electron probe microanalysis [e.g., Tomiyasu et al., 1996], scanning electron microscopy (SEM) [e.g., Ault et al., 2012], and transmission electron microscopy (TEM) [Adachi et al., 2010; Ueda et al., 2011; Fu et al., 2012]. TEM is commonly used to investigate the structure of fine aerosol particles. Adachi et al. [2010] used electron tomography analyses to propose the structure and formation of BC and its host materials for 46 “standard particles” collected in Mexico City. Ueda et al. [2011] used TEM after extraction of water-soluble species to investigate the morphological features of aerosol particles containing soot. They found that the major components were sulfate and organic matter (OM) and that there were five types of soot structure, including aggregates, dome-shaped particles, particles with a satellite structure, single spheroid or coccoid particles, and clusters of spheroid or coccoid units. Fu et al. [2012] reported the morphology, composition, and mixing state of carbonaceous particles collected in urban Shanghai. They investigated organic aerosols such as polymeric organic compounds, soot, tar ball, and biogenic particles. Soot was common to all samples, and the trace elements K, Si, Ca, Fe, Mn, Zn, and S were found in the soot particles, indicating that the soot had both biogenic and anthropogenic origin and had experienced deposition in atmosphere. The relative abundances of soot, sulfate, and Si for approximately 200 particle samples were reported to be 22%, 43%, and 14% in the air pollution case. Shiraiwa et al. [2008] used a single-particle soot photometer (SP2) at Fukue Island, Nagasaki, Japan, to investigate the nature of BC coating. They classified the air mass using back trajectory analysis and found that the maximum shell/core diameter ratio (R) was 1.6 for particles with diameters of 200 nm, and BC was mainly associated with sulfate and OM, as measured using Q-AMS. Recently, Sedlacek et al. [2012] have also used an SP2 technique to explore the structure of particles containing BC. They found that a core/shell model does not provide an accurate description for some particles. With these methods, however, structure and chemical composition cannot be measured simultaneously.

[6] Time-of-flight secondary ion mass spectrometry (TOF-SIMS) has recently been introduced to study the structure of aerosol particles [Sakamoto et al., 2008]. The TOF-SIMS apparatus, in which a gallium focused ion beam (FIB) is used as the primary ion beam, can analyze the materials located inside the particles by the combination of FIB milling and repeated elemental mapping (Figure 1). The apparatus used here has a high spatial resolution (approximately 40 nm, which, to the best of our knowledge, is currently the finest resolution), and therefore, fine particles smaller than 1 µm can be analyzed. In addition, SIMS can simultaneously analyze the chemical composition.

Figure 1.

Schematic of the particle analysis by TOF-SIMS. (left) The analytical procedure for determining the internal structure of solid-state particles using FIB milling. (right) The analytical procedure for determining the internal structure of viscous-state particles using repeated FIB elemental mapping.

[7] The objectives of this study are to elucidate the structure and chemical composition of fine aerosol particles with diameters of 1 µm observed in the East Asian region.

2 Materials and Methods

2.1 Sampling Location

[8] Sampling was performed at the National Institute of Environmental Studies Fukue Atmospheric Observation site (Fukue site: 128.7°E, 32.8°N; Figure 2). Fukue Island is located to the west of Kyushu Island, Japan, and in the northern part of the East China Sea. It is a suitable location to monitor the outflow from the Chinese continent in the winter-spring season, since the seasonal monsoon from the continent to the Pacific Ocean prevails in this area. The population of the island is about 40,000, and there is no major industrial area. The sampling site is located in the northwestern part of the island, about 20 km away from the downtown area; it is on the upstream side of the major wind direction in the winter-spring season. Therefore, our measurements were focused on the outflow from the continent. The area around the sampling site is sparsely populated, and low anthropogenic emissions are expected. Previous studies have shown that the continental outflow can be observed frequently in the spring season [Takami et al., 2005; Shiraiwa et al., 2008; Kaneyasu et al., 2011].

Figure 2.

Location of Fukue Island.

2.2 Bulk Sampling by Q-AMS

[9] The chemical composition of the aerosol particles was measured and analyzed using a quadrupole-type aerosol mass spectrometer (Q-AMS, Aerodyne Research, Inc.). For this observation, Q-AMS was used in order to show that the individual sampling was carried out under typical weather and transport conditions for this region and period. A detailed description of Q-AMS can be found elsewhere [Jayne et al., 2000; Allan et al., 2003a, 2003b; Jimenez et al., 2003; Canagaratna et al., 2007; Takami et al., 2005, 2007]. Briefly, ambient air is introduced through an aerodynamic lens, which separates gaseous species and particles and forms the particle beam. The particle beam hits the vaporizer set at a temperature of 873 K (600°C), where nonrefractory species are vaporized. Vaporized molecules are ionized by electron impact ionization at 70 eV, and the ions are then analyzed by the quadrupole mass spectrometer. Sulfate, nitrate, ammonium, and chloride contents were calculated from the fragment signals of the mass spectra. Organics were calculated by subtracting the known inorganic and gaseous species, such as sulfate, nitrate, nitrogen, oxygen, and argon, from the total mass. The particle mass was calibrated using ammonium nitrate particles. The ionization efficiency used here was 3.21 × 10−6, and the relative ionization efficiency for NH4+ was 3.2. The collection efficiency for all the measured species was 0.5. The AMS data were analyzed using the standard AMS analysis software [Allan et al., 2004].

[10] The sampling inlet was set about 3 m above the ground. A PM2.5 cut cyclone (University Research Glass) was attached at the inlet to remove coarse particles. A stainless steel tube with an outer diameter of half an inch was used for the main sampling line. Taking into account isokinetic sampling, the main sampling line was connected to the Q-AMS inlet tube (outer diameter, 1/8 inch).

2.3 Individual Particle Analysis by TOF-SIMS

[11] A custom-made single-stage impactor [Yamaguchi and Sakamoto, 2008] was used for particle sampling. The impactor collected particles smaller than 10 µm in diameter. Silicon wafers with an area of 4 mm × 4 mm were used as collection substrates. Each sampling was performed for 10 min at a suction rate of 1.5 L min−1. The silicon wafers loaded with the aerosols were introduced into the TOF-SIMS apparatus at room temperature and without any pretreatment. Individual aerosol particles have been previously analyzed with the apparatus to determine their chemical reactions and sources, but a quantitative analysis cannot be performed owing to the different sensitivities for different compounds [Sakamoto et al., 2012; Mayama et al., 2012].

[12] The apparatus has two beams: the primary gallium FIB and an electron beam (EB) (Figure 3). The FIB was used as a milling tool to section the particles, and the observation of the particle surface was performed using the FIB-induced secondary electron image and elemental mapping in FIB-pulsed irradiation mode. The FIB was rastered at an irradiation energy of 30 keV to obtain the FIB-induced secondary electron image. The EB was used to search for particles to be analyzed and to observe the FIB milling as an SEM image. The base pressure of the specimen chamber was around 3.0 × 10−6 Pa. The high-resolution SEM image of the particles was obtained using a field-emission scanning electron microscope (JSM6700F, JEOL).

Figure 3.

Schematic of TOF-SIMS apparatus.

[13] The individual analysis of fine aerosol particles was performed using a different procedure because the particles were sputtered by the FIB irradiation. First, the dense areas of fine particles were selected on the silicon wafer by observation with SEM. Then, we selected the area where particles were well separated to avoid collision and aggregation of particles on the substrate. Elemental mapping (128 × 128 pixels) was performed with as little sputtering as possible on an area of a few tens of square micrometers containing many particles. The elemental mapping (64 × 64 pixels) was successively performed in negative and positive secondary ion modes, in that order, for a few selected fine particles in that area until the internal structure of the particles had been elucidated or the particles had been completely sputtered by the repetition of this cycle. Such analyses were performed for more than 500 fine particles in randomly selected areas. The FIB pulse duration was 300 ns, and 50–100 shots per pixel were performed. The FIB current was about 300 pA in direct current mode. The time for one cycle of elemental mapping (64 × 64 pixels, 100 shots per pixel) was less than 1 min. The information depth of one cycle of elemental mapping was typically about 20 nm based on the above experimental conditions.

3 Results

3.1 Bulk Measurement Results

[14] Fine aerosol particles were collected at Fukue Island from 5 to 26 April 2011. The time series of the chemical species measured using Q-AMS is shown in Figure 4. High mass concentrations of sulfate were observed on 11 and 16 April, and relatively high organic mass concentrations were observed on 16, 19–21, and 23–24 April. The maximum total mass concentration of 70 µg m−3 was recorded on 16 April. No precipitation was observed from 16 to 20 April. The average mass concentrations and standard deviations (expressed in ±1σ) of the ammonium, nitrate, sulfate, chloride, and organic contents from 5 to 26 April were 2.8 ± 1.9, 1.0 ± 1.3, 7.1 ± 3.8, 0.1 ± 0.1, and 7.6 ± 2.9 µg m−3, respectively, and their fractions were 15%, 5%, 38%, 1%, and 42%, respectively (Table 1). The molar ratio of ammonium to sulfate was 2.08 ± 0.40, indicating that the aerosol particles were neutralized [Takami et al., 2005, 2007]. The mass concentrations of these species in spring of 2003 at Fukue were 1.6, 0.6, 4.8, 0.1, and 5.0 µg m−3, respectively, and their mass fractions were 13%, 5%, 40%, 1%, and 42%, respectively [Takami et al., 2005]. Our mass concentrations were higher than those observed in 2003, although the fraction of each component was similar.

Figure 4.

Variation of mass concentration observed using Q-AMS at Fukue.

Table 1. Average Mass Concentration and Mass Fractions of Chemical Components Observed From 16 to 21 April 2011 Using Q-AMS at Fukue
Average mass (µg m−3)
Standard deviation (1σ)
Average fraction0.150.050.380.010.42 
Standard deviation (1σ) 

[15] High concentrations of sulfate (30 µg m−3) were observed on 16 April, while both sulfate and organics were high on 20–21 April. The weather charts provided by the Japan Meteorological Agency showed that the high-pressure system (HPS) stayed in the Chinese coastal region on 16 and 19–20 April. When the HPS was in the Chinese coastal region, air masses were transported along the edge of the HPS in a clockwise direction, and high mass concentrations were observed in the western part of Japan [Takami et al., 2007; Kaneyasu et al., 2011]. This is a typical weather condition in this region in spring.

[16] Back trajectories were calculated using the NOAA Hybrid Single-Particle Lagrangian Integrated Trajectory model [Draxler and Rolph, 2012; Rolph, 2012]. Back trajectories showed that the air mass was transported from the Shangdon Peninsula, China, on 16 April and from the Korean Peninsula on 20 April. This agrees well with previous observations that high sulfate concentrations were observed for air masses coming from China and that relatively high organic contents were observed for air masses from Korea [Topping et al., 2004; Takami et al., 2005]. Thus, different air masses have different chemical compositions.

[17] The chemically resolved size distributions of the Q-AMS species are shown in Figure 5. The peak mode of the aerodynamic diameter (Dva) was centered at 600 nm, and the shape of each component was similar, indicating that the chemical components of aerosol particles were internally mixed; the organic mode diameter peaked at slightly lower Dva, indicating that organic and sulfate were not totally internally mixed. The large Dva is typical of aged particles, consistent with long-range transport, whereas smaller Dva values (~100 nm) are typically indicative of freshly emitted particles such as those associated with vehicle exhaust in urban areas [Alfarra et al., 2004].

Figure 5.

Averaged size distribution of chemical components observed from 16 to 21 April 2011 using Q-AMS at Fukue.

[18] Figure 6 shows the triangle plot of f44 against f43 from 16 to 21 April when organics were high [Ng et al., 2010]. f44 (f43) is the ratio of the m/z = 44 (43) signal to the organics signal. The m/z = 43 signal represents ions of the aldehyde (CH2CHO+) and alkyl (CH2CH2CH3+) functional groups. The m/z = 44 signal represents ions of the carboxyl (COO+) functional group. The plot shows that most of the points are scattered around the central positions of f43 = 0.05 and f44 = 0.15, indicating that the observed aerosol particles mainly contained low-volatile oxygenated organic aerosol (LV-OOA), and experienced photochemical aging processes during long-range transport [Ng et al., 2010].

Figure 6.

f44 plotted against f43 for observations made from 16 to 21 April 2011 using Q-AMS at Fukue.

[19] The results of our observations and analysis using Q-AMS show that the conditions during the sampling period for the individual particle analysis were typical for this region.

3.2 Individual Particle Analysis Results From TOF-SIMS Measurements

[20] Figure 7 shows the FIB-induced secondary electron image of fine particles after FIB presputtering. The particles identified were coarse particles such as dust (yellow sand), sea salts, and fine particles such as small soil, sulfate, nitrate, BC, and organic compounds [Mayama et al., 2012, 2013].

Figure 7.

FIB-induced secondary electron image of fine viscous particles with a diameter of about 1 µm after FIB presputtering. Type A: particles in the viscous state from the surface to the interior. Type B: particles with a solid material of diameter ~500 nm inside the particle. Type C: particles with a solid material of diameter ~100 nm at the edge of the sulfate particle.

[21] From the TOF-SIMS results, we classified particles into three types based on the residual materials inside the particle after sputtering. In Type A, no solid core was detected from the surface to the interior. In Type B, a solid material with a diameter of about 500 nm was detected inside the particle. In Type C, a solid material with a diameter of about 100 nm was detected at the edge of the particle.

[22] Elemental mapping of these particles over a wide area is shown in Figure 8, and the elemental mappings of individual particles in three areas (I to III) are shown in Figures 9-11, respectively. Most particles in Figures 9-11 were completely sputtered after only one cycle of elemental mapping. They were classified as Type A and consisted of sodium sulfate, sodium nitrate, potassium, and ammonium surrounded by carbon. The carbon was identified as OM from the mass spectrum [Mayama et al., 2013]. In area II, Type B particles were observed (Figure 10). The remaining solid with a diameter of about 500 nm inside the particle consisted of silicon, potassium, and calcium oxide. Based on the composition, the solid core was identified as a soil (mineral dust). Type B had a core/shell structure. In area III, Type C particles were observed (Figure 11). The remaining solid, with a diameter of about 100 nm at the edge of the sulfate and nitrate part, consisted of carbon. The carbon was identified as BC because the mass spectrum was consistent with that of graphite (Figure 12). The high-resolution secondary electron image of the Type C particles is shown in Figure 13. Aggregations of soot particles of a few tens of nanometers were observed at the edge of the sulfate particles. The structure of Type C particles is schematically depicted in Figure 11. Since neither isolated BC nor isolated soil particles were observed, they were considered to have mixed with sulfate, nitrate, OM, etc., in the atmosphere.

Figure 8.

Total ion image and elemental map in negative analysis mode of fine particles over a wide area.

Figure 9.

Surface elemental maps for individual particles in area I in Figure 8 and a schematic explanation of these maps.

Figure 10.

Elemental maps of the surface and interior of individual particles in area II in Figure 8 and a schematic explanation of these maps.

Figure 11.

Elemental maps in the negative analysis mode of the surface and interior of individual particles in area III in Figure 8 and a schematic explanation of these maps.

Figure 12.

Mass spectra of BC and graphite from TOF-SIMS measurements.

Figure 13.

Secondary electron image of Type C particle. OM: organic matter, BC: black carbon.

[23] Figure 14 shows schematics of each structure type. All particles consist of sulfate and nitrate, typically with a diameter of less than 1 µm, surrounded by OM. From the depth of OM on the silicon wafer, the thickness of the OM could be estimated to be less than 100 nm.

Figure 14.

Schematics of the structures of fine particles of Types A–C.

[24] The relative abundances of each type of fine particle collected in Fukue and in the Tokyo Metropolitan area are listed in Table 2 along with the number counts, measured by 548. The ratio of particles without solids (Type A) and those containing solids (Types B and C) was about 6:4 in both areas. Type C particles collected in Tokyo accounted for about 5% of the total and in Fukue for 25%.

Table 2. Fraction of Each Type of Fine Particle With a Diameter of About 1 µm Collected in Fukue and Metropolitan Tokyo
 Type A (%)Type B (%)Type C (%)Particle Number

4 Discussion

4.1 Potential Measurement Bias

[25] It is important to discuss the potential measurement bias in our observations. There are two major sources of bias that could influence the determination of the particle morphology and chemical composition: the sampling method and the TOF-SIMS analysis.

4.1.1 Biases From the Sampling Method

[26] The particle morphology, the chemical composition, and the location of BC and soil in the particle could be influenced during the sampling procedure. The first bias is changes in morphology, which probably occur by impaction when ambient particles are collected on the Si substrate. Although solid particles such as dry sea salt retain their shape, liquid particles (droplets) change shape. The second bias is the change of chemical composition caused by collision and/or aggregation of particles on the substrate. In order to avoid analyzing these particles, we chose areas where particles were well separated. We observed neither isolated BC nor isolated soil particles. Therefore, BC and soil were considered to be mixed with sulfate and OM in the atmosphere. The third bias is the change of the location of BC and soil within the particles mixed with sulfate and OM, i.e., Type B and C particles. Since all particle types were found to be flat on the substrate, these particles are considered to be liquid particles, and therefore, the morphology was changed by impaction. On the other hand, since the locations of BC and soil were found to be different in Type B and C particles, respectively, this difference is considered to infer mixing in the atmosphere.

4.1.2 Biases From the TOF-SIMS Analysis

[27] In the TOF-SIMS analysis, the high-vacuum system, dry condition, and ion beam bombardment are factors that could influence the particle, similar to in a TEM analysis. Fu et al. [2012] described two major influences on particles in a BC observation using TEM.

  1. [28] Evaporation and loss of semivolatile compounds such as water, ammonium nitrate, and semivolatile organic compounds due to the strong vacuum and electron beam exposure [Niemi et al., 2006]. However, it is thought that, except for volatile species, the components observed in the TEM analysis should reflect the original composition.

  2. [29] Morphological change of some particles from their shape prior to collection as a result of the dry conditions in the TEM analysis chamber [Li et al., 2003]. Although some particles were observed to be dry in the TEM chamber, in the atmosphere, they may have been droplets owing to the high humidity. Ammonium sulfate particles crystallized in the dry conditions in the TEM chamber.

[30] We must take into account effects similar to these in the interpretation of our results. In summary, morphology was influenced by the impaction. However, the flat shape of particles observed by the TOF-SIMS analysis indicates that they were probably in the liquid phase in the atmosphere. Chemical composition was influenced by the TOF-SIMS observation. Therefore, discussion was based on the detected components, which are solid materials of BC and soil and low-volatile compounds of sulfate and low-volatile organics, because these compounds were contained in the particles in the atmosphere.

4.2 Chemical Composition and Mixing State of Aerosol Particles

[31] We have analyzed the chemical composition of aerosol particles using both Q-AMS and TOF-SIMS. The results for the individual particles show that sulfate and organics are the more abundant species. This agrees with the Q-AMS bulk results, indicating that typical aerosol particles were analyzed in the individual particle analyses.

[32] We have observed the three types of aerosol particles mentioned above (Types A–C). Type A consists of sulfate and OM. (Although TOF-SIMS observed both sulfate and nitrate, but to simplify the explanation, we only refer to sulfate in this section.) Type A particles were completely evaporated after ion beam irradiation, meaning that Type A has no core/shell structure. It contains neither soil nor BC.

[33] Type A particles observed by TOF-SIMS had a flat and round shape. This is possibly due to impaction on the substrate, indicating that Type A particles were a liquid (viscous) particle in the atmosphere. Since the TOF-SIMS analysis was performed under high-vacuum and dry conditions, volatile compounds such as water and semivolatile organic compounds were lost by evaporation. The observed compounds were low-volatility compounds such as sulfate and LV-OOA, which is supported by the fact that organics measured by Q-AMS showed a high f44 value.

[34] When the particles were transported from the Chinese continent to Japan, they passed over large bodies of water, i.e., the East China Sea. Since both sulfate and LV-OOA are hydrophilic, Type A particles were considered to be in the liquid phase (and likely to have deliquesced) in the atmosphere during transport.

[35] Since aerosol particles were transported under various relative humidity (RH) conditions, phase separation may occur within the particle. Ciobanu et al. [2009] observed the phase separation of micrometer-sized particles consisting of ammonium sulfate (AS) and poly (ethylene glycol)-400 (PEG-400) using optical microscopy and a micro-Raman spectroscopic method. When the organic to inorganic (AS) ratio was 1:1 or 1:2, phase separation occurred immediately below a relative humidity of 90%, with both AS and PEG-400 remaining in the liquid phase. The Q-AMS results showed that the chemical composition of the aerosol particles observed in Fukue consisted mainly of aged organics and sulfate, and the mass ratio of organics and sulfate ranged from 1:1 to 1:2. The relative humidity in the spring season was in the range 40%–70% from ground observations in Fukue. Under these conditions, it is possible for phase separation to occur in Type A particles and thus could explain that the observation of OM being found on the outer surface was probably due to the phase separation of AS and organics in the aerosol particles.

[36] As for Type B, there is soil in the center, and sulfate and OM cover the outside of the particle. This type of particle has a typical core/shell-type structure. The size of the soil core is about 500 nm. Since it is smaller than that of particles observed in a yellow sand storm [Arao and Ishizaka, 1986; Carmichael et al., 1996], the origin of the soil particle is considered not to be desert sand but roads, fields, ground, etc.

[37] As for Type C, BC consists of nanosize primary BC particles and forms aggregates, as shown in Figure 13. The size of the BC aggregates is about 100 nm, which is typical for ambient sampling [Shiraiwa et al., 2008]. Interestingly, BC is never found in the center but at the edge and/or on the surface of the sulfate particle. This is in contrast to the Type B particles, in which soil was associated with sulfate. In Shanghai, sulfur (S) was found in all soot samples, and it was considered that S-containing compounds such as sulfate were deposited on soot particles in the atmosphere [Fu et al., 2012]. As mentioned above, Type A particles were considered to be in the liquid phase in the atmosphere, and phase separation took place under dry conditions. Since BC is hydrophobic, BC may be segregated by phase separation in the mixture of sulfate and OM particles, while soil may stay in sulfate since it is hydrophilic. This means that BC may be at the off-center position in the Type C particle in the atmosphere when the RH is below 90%, which is a typical condition in the winter-spring season in the East China Sea region.

4.3 Relative Abundances

[38] The relative abundance of observed particles collected in Fukue and in the Tokyo Metropolitan area are listed in Table 2 along with the number of particle counts (n = 548 for Fukue and n = 226 for Tokyo). For Type A, the relative abundance was 55% in Fukue and 60% in Tokyo. For Type B, it was 35% in Tokyo and 20% in Fukue, indicating that both Type A and B particles are ubiquitous. The value for Type C in Tokyo is less than 5%, which is much lower than that of the particles collected in Fukue (25%). Since diesel vehicles have been restricted in the Tokyo Metropolitan area, the fraction of BC-containing particles is expected to be low. In contrast, since emissions of BC in China are still high [Ohara et al., 2007], and soot was found in all samples in Shanghai [Fu et al., 2012], we expected to observe a higher fraction of BC-containing particles on Fukue Island, where transboundary air pollution has an influence. We suggest that the proportion of BC-containing particles could be used as one of the indicators of transboundary transport of aerosol particles found in Japan.

4.4 Optical Properties and Simulation Model of Aerosol Particles

[39] In Chin et al. [2009], the single scattering albedo (SSA) values for sulfate, BC, organic carbon (OC), sea salt (coarse), and dust (soil) were calculated using refractive index, particle size, and relative humidity. In the near-UV (300–400 nm) and visible regions, the SSAs of sulfate and sea salt are both unity; the SSAs of OC are 0.8–0.95 and unity, respectively; the SSAs of BC (uncoated) with a diameter of approximately 0.1 µm are 0.35 and 0.25, respectively, and the SSAs of dust (soil) with a diameter of approximately 0.3 µm are 0.85 and 0.95, respectively. Based on these calculations, we classified light-absorbing/scattering particles as follows.

  1. [40] Sulfate and sea salt are light-scattering particles.

  2. [41] OC is a weakly light-absorbing particle in near-UV light and a light-scattering particle in visible light.

  3. [42] BC (uncoated) is a light-absorbing particle in both near-UV and visible light.

  4. [43] Dust (soil) is a semi–light-absorbing particle in near-UV and visible light.

[44] Since soil in this study has a diameter of approximately 0.5 µm, whose SSA is estimated to be between 0.75 and 0.85 in the near-UV region and 0.9 in the visible region, it is classified as a semi–light-absorbing particle in near-UV and visible light.

[45] Humic-like substances (HULIS) is an important component from soil, and its optical character has been investigated by Hoffer et al. [2006]. Absorption of solar radiation by HULIS is not negligible. Since HULIS is a kind of high–molecular weight organic component, we consider HULIS as a type of OM.

[46] We considered the optical properties of each type of particle. Since the relative abundance of Type A was 55% in Fukue and 60% in Tokyo, we found that non–light-absorbing (scattering) particles are dominant in both remote (Fukue) and urban (Tokyo) atmospheres in Japan. Type B particles contain a soil particle with a diameter of approximately 0.5 µm, which is a semi–light-absorbing particle in the near-UV and visible regions. Since the relative abundances of Type B are 20% in Fukue and 35% in Tokyo, this cannot be neglected in the calculation of radiative forcing in either of these locations. Since the relative abundance of Type C is 25% in Fukue, its influence on the radiative forcing calculation is expected to be large there.

[47] Aerosol particles are known to act as cloud condensation nuclei (CCN). This depends on their size and chemical composition, which change hygroscopic properties and the efficiency of cloud formation [Mochida et al., 2010]. Particles observed in this study were associated with sulfate and OM in the atmosphere. Since both sulfate and OM, which is aged in Fukue, are hydrophilic, they can act as CCN in the East China Sea region, to which the aerosol particles are transported. This may influence the precipitation pattern on both global and regional scales [Mukai and Nakajima, 2009].

[48] It is useful to compare our findings with the mixing state of the fine aerosol particles used in the simulation. For example, Goto et al. [2011] assumed in their simulation that 50% of BC was mixed with biogenic secondary organic aerosol and that the rest was externally mixed. Sulfate, soil, and sea salt were also considered to have been externally mixed. Our observations suggest that BC and soil are internally mixed with sulfate and OM in the atmosphere. Soot observed in Shanghai was found with various trace elements such as K with biogenic origins; Si and Ca from vehicle engines; Fe, Mn, and Zn from coal combustion; and S from the atmosphere [Fu et al., 2012]. These results suggest that most of the BC emitted on the Chinese continent was internally mixed with sulfate and OM. Kajino et al. [2012] specified four types of aerosols, namely, Aitken mode, accumulation mode (ACM), BC aggregates (AGR), and coarse mode (COR). ACM consisted of OM, sulfate, nitrate, chloride, and water; AGR additionally contained BC. Thus, ACM were light-scattering particles, and AGR were light-absorbing particles. Soil was present only in COR. The Kajino model is also different from our observations. Soil particles are mixed with sulfate and OM in fine particles (ACM), and the relative abundances of Type B are 20% and 35% in Fukue and Tokyo, respectively, indicating that Type B particles should be included in ACM in the East Asian region. The data in Table 2 show that more than half of the fine particles consisted of sulfate and OM and were mixed with neither BC nor soil. In Fukue, more than half the fine particles were light scattering, 20% were semi–light absorbing, and 25% were light absorbing. These numbers can be used as constraints for the simulation. In addition, these fractions were different in the Tokyo Metropolitan area, suggesting that the fractions of light-absorbing/scattering particles can vary greatly by area. The particle model used for the aerosol simulation may also require revision.

[49] Structural information is important for the calculation of the optical properties of aerosol particles. Since we observed that BC is located at the edge and/or on the surface of sulfate in Type C particles owing to phase separation and segregation, the position of BC is probably off-center in the aerosol particles in the atmosphere. This is in agreement with several previous studies [Adachi et al., 2010; Fu et al., 2012; Cappa et al., 2012; Sedlacek et al., 2012]. On the other hand, soil particles are found in the host materials for Type B particles, suggesting that Type B particles may enhance light absorption as a result of the lens effect of the OM, which affects the optical properties of aerosol particles. Again, we require more information about the mixing state, chemical composition, and structural information of individual aerosol particles.

5 Conclusions

[50] We have measured the chemical composition and structures of fine aerosol particles collected at Fukue Island, to the west of Kyushu Island, Japan, using Q-AMS and TOF-SIMS. Comparison of the Q-AMS measurement results with those of previous studies made at the same location was similar, suggesting that the particle composition we observed is typical for this site. TOF-SIMS was capable of analyzing the structure and chemical composition of aerosol particles with diameters of less than 1 µm. Based on the results of TOF-SIMS, we classified the aerosol particles into three types: Type A, sulfate covered with OM; Type B, soil-containing particles associated with sulfate and OM; and Type C, BC and sulfate aggregates associated with OM (see Figure 14). Sulfate was found inside Type A, and OM was found outside; this is thought to be due to phase separation of AS and organics. As for Type B, soil was found inside the sulfate, and OM was found on the outer surface. The size of the soil was about 500 nm, indicating that the origin of the soil was not desert sand but roads, ground, fields, etc. In Type C, BC was found at the edge and/or the surface of the sulfate and formed BC-sulfate aggregates. During the observation period, the relative abundances of Type A, B, and C particles were 55%, 20%, and 25%, respectively, while in Tokyo, they were 60%, 35%, and 5%. More than half of the fine particles were light scattering, which may require the revision of the DRF calculation in simulations. Particles observed in this study were associated with sulfate and OM in the atmosphere, and they can act as CCN under certain conditions. This may influence the precipitation pattern on both global and regional scales. The relative abundance of Type C, i.e., BC-containing particles, is quite different, suggesting that Type C particles could be an indication of transboundary long-range air pollution, in this case from mainland China. We believe that the TOF-SIMS analysis with Q-AMS is useful and provides additional information to that obtained using TEM and SP2.


[51] We would like to thank the reviewers whose comments contributed significantly to the revision of our manuscript. We greatly appreciate the support we have received from GERF (B-1006, A-1101), MEXT-Shingakujutsu (ASEPH), and JST-SIRCP (elucidation of the impacts of the spatial distribution and changes in chemical composition of absorptive aerosols (EC) and dispersive aerosols (OC, metallic elements, and ionic elements) on radiation). We thank S. Tagata and Y. Miura (Kogakuin University) for their help with the sampling and TOF-SIMS analysis.