Measurements of size distribution, hygroscopicity, and volatility of submicrometer sea spray particles produced by the bubble busting of artificial and natural seawater were conducted to determine their mixing state and volume fractions of hygroscopic and nonhygroscopic species or volatile and nonvolatile species. The particles sprayed from artificial seawater having insoluble silica particles were found to be an external mixture of two groups of particles having hygroscopic growth factors (HGFs) of 1.33 (an internal mixture of nonhygroscopic silica particles and hygroscopic salt species) and 1.68 (a similar mixture having more salt species) when the mass ratio of insoluble particles to dissolved salts was higher than 2. For sea spray particles from natural seawater, the external mixing was not significantly observed because of a high concentration of dissolved salts. The HGFs of sea spray particles (80–140 nm) from natural seawater were in the range of 1.70–1.76, which were lower than from pure artificial seawater (1.87), and the HGFs had no change before and after membrane filtration of seawater, suggesting that the sea spray particles from natural seawater contained a significant amount of nonhygroscopic dissolved organic matter in addition to hygroscopic salt species. The volume fraction of the nonhygroscopic species ranged from 20% to 29%, and the highest value was observed for seawater samples from the site where strong biological activity occurred, suggesting that biological materials played an important role in the formation of nonhygroscopic organic matter. Volatility measurements also identified the existence of volatile organic species in single particles from natural seawater, with the volume fraction of volatile species evaporated at 100°C being 4%–5%.
The ocean covers more than 70% of the earth’s surface and is the most important contributor for atmospheric aerosols (Andreae 1995). The global annual production of marine aerosols from the ocean is reported to be about 1300–3300 Tg yr−1 (Andreae 1995), significantly affecting the earth’s radiation balance (Slingo 1990; Murphy et al. 1998; Haywood et al. 1999; Solomon et al. 2007), cloud formation by serving as cloud condensation nuclei (CCN) (Dinger et al. 1970; Orellana et al. 2011; Fuentes et al. 2011; Moore et al. 2011), and biological cycle because of interactions between the air and seawater (Marks et al. 1996; O’Dowd et al. 2004; Deguillaume et al. 2008; Hultin et al. 2011). Typically, seawater contains dissolved salts or ions (sodium, chloride, sulfate, magnesium, calcium, etc.), dissolved gases, dissolved organics (either released directly to seawater from primary production or final product of degraded detritus of biological materials in seawater) (Verdugo and Santschi 2010), and insoluble species (bacteria, virus, clay minerals, dust, biological debris, etc.), which subsequently contribute to the formation of primary and secondary marine aerosols.
The bubble-burst process from the sea surface is known as an important mechanism in the generation of primary marine aerosols. In this process, film and jet droplets are typically produced by the bubble bursting of seawater (O’Dowd et al. 1997; Spiel 1998; Bowyer 2001; Sellegri et al. 2006; Keene et al. 2007; Fuentes et al. 2010), which includes dissolved and insoluble species (Wells and Goldberg 1992; Mårtensson et al. 2003). Thus, sea spray particles produced via the bubble-bursting process under high wind speeds (>4 m s−1) mainly consist of inorganic salts having varying fractions of organic matter. It has been shown that although inorganic salts are a major constituent of the sea spray particles, organic matter components can also be an important component—especially in the submicrometer size range (O’Dowd et al. 2004; Facchini et al. 2008; Rinaldi et al. 2010). Ovadnevaite et al. (2011) reported sea spray particles enriched by organic matter with mass concentration up to 3–4 μg m−3 in CCN active sizes, suggesting that primary organic aerosols significantly impact the CCN budget. It was reported that an increase in the phytoplankton population in seawater can enhance the production of biological marine aerosols (O’Dowd et al. 2004; Ceburnis et al. 2008; Facchini et al. 2008). During phytoplankton blooms, water insoluble organic matter (WIOM) was found to be a dominant chemical constituent of primary marine aerosols (size range: 0.125–0.5 μm), reaching a mass concentration of 45%–73% at Mace Head, Ireland (O’Dowd et al. 2004; Ceburnis et al. 2008; Facchini et al. 2008). In addition, it was reported that biogenic volatile organic compounds (BVOCs) (e.g., dimethylsulphide, isoprene, monoterpenes, and halocarbons) can be emitted from seawater, contributing to the formation of secondary marine aerosols via the gas-to-particle conversion process (Charlson et al. 1987; Carpenter et al. 2000; Yassaa et al. 2008; Gantt et al. 2009).
The size distribution, mixing state, and compositional structure of submicrometer sea spray particles were reported to be dependent on the salinity, temperature, surfactant concentration, and biological activity in the seawater (Monahan et al. 1986; O’Dowd et al. 1997; Mårtensson et al. 2003; Tyree et al. 2007; Sellegri et al. 2006, 2008; Keene et al. 2007; Modini et al. 2010; Fuentes et al. 2010; Hultin et al. 2010). Previous laboratory and filed measurements suggest that marine organic aerosols are externally and/or internally mixed with sea salt species, although some contradictory results exist (Middlebrook et al. 1998; Gantt and Meskhidze 2013). It was reported that the organic fraction and mixing state were related to the particle size, chemical, and biological composition of seawater, and wind speed (Gantt and Meskhidze 2013). In addition to physical and chemical properties of sea spray particles, supplementary measurements of meteorological parameters and chemical and biological properties of seawater are needed to better understand the mixing state of marine primary organic aerosols (POAs) (Gantt and Meskhidze 2013). The mixing state in size-resolved submicrometer sea spray particles is difficult to be characterized because of the complex mixing state of soluble species (e.g., dissolved salt and organic species) and insoluble materials, both of which contribute to the formation of primary marine aerosols (Fuentes et al. 2010; Modini et al. 2010). The role of insoluble species in sea spray particles is not well understood.
In this study, filtration of seawater using microfiltration (MF) and ultrafiltration (UF) membranes (Park et al. 2009b; Park et al. 2009, 2011) was first conducted to obtain seawater samples that have only dissolved species (after UF membrane filtration) or a mixture of dissolved species and insoluble submicrometer particles (<0.45 μm) (after MF membrane filtration and before UF membrane filtration). This separation enables us to better understand their contribution to the formation of submicrometer sea spray particles. Here, a laboratory-scale bubble-bursting system was constructed in order to produce sea spray particles. Nonhygroscopic silica particles in the submicrometer size range (i.e., they can be separated from hygroscopic salt species in hygroscopic measurement) were added into artificial seawater to examine the role of insoluble species in sea spray particles. Natural seawater samples from coastal areas (West Sea, East Sea, and South Sea) in Korea were used to produce sea spray particles by using this bubble-bursting system. Number size distributions of sea spray particles with varying bubble flow rates and salinity were determined using a scanning mobility particle sizer (SMPS) (20–600 nm) and particle size distribution (PSD) analyzer (500 nm–10 μm). The mixing state, hygroscopicity, and volatility of the size-selected sea spray particles were subsequently determined using the hygroscopicity and volatility tandem differential mobility analyzer (HVTDMA) technique, which is able to provide useful insights into the mixing state and fraction of volatile organic species of size-selected submicrometer particles (Rader and McMurry 1986; Park et al. 2009a, 2010). In addition, a transmission electron microscopy (TEM)–energy dispersive spectroscopy (EDS) analysis of the sea spray particles was conducted to examine their morphology and elemental composition.
2. Experimental method
Figure 1 presents a schematic of the bubble-bursting system to produce sea spray particles and aerosol measurement systems. The volume of the bubble chamber is 2.0 L, and it was filled with a 1.0-L seawater sample. The bubble chamber was made of aluminum, and the inside of the chamber was coated with Teflon. A sintered glass filter with pore sizes of 20–40 μm (Millipore, United States) was placed on the end of tube, located at ~3.5 cm below the water surface. Clean compressed air was introduced through the glass filter (i.e., bubble flow) to generate bubbles. Simultaneously, another clean air source was supplied to the upper part of the chamber (i.e., airflow) to generate whitecaps on the seawater sample surface. Both the bubble and airflow rates were controlled using flow controllers (Dwyer Instruments, United Kingdom). The temperature inside the chamber and the airflow rate were fixed at 25°C and 5.0 L min−1, respectively, while the bubble flow rate was varied. Bubbles generated by this system were then dried using a series of two diffusion dryers (each dryer has silica gels in the outer tube in the length of 53.5 cm) (TSI 3062, TSI, United States), and the size distribution, hygroscopicity, volatility, morphology, and elemental composition of the dried particles were measured using the aerosol instruments described below.
The size distribution of the dried particles (20 nm–10 μm) was continuously measured using an SMPS [differential mobility analyzer (DMA) (TSI 3081, TSI, United States)] and condensation particle counter (CPC) (TSI 3022A, TSI, United States) (20–600 nm), and PSD analyzer (TSI 3063, TSI, United States) (500 nm–10 μm). In the SMPS, the aerosol and sheath flow rates were 0.3 and 3 L min−1, respectively; for the PSD, the aerosol flow rate was 1 L min−1. Since the SMPS and PSD used different size measures (DeCarlo et al. 2004), the aerodynamic-equivalent diameters measured with the PSD were converted to mobility-equivalent diameters measured with the SMPS, assuming particle density information (particle density varied to have a consistent size distribution in the overlapping size range from SMPS and PSD). The hygroscopicity and volatility of particles were measured using an HVTDMA system (Rader and McMurry 1986; Park et al. 2009a, 2010). In brief, the HVTDMA system consists of two regular DMAs (TSI 3081, TSI, United States), a humidification system, a heated tube, and an ultrafine condensation particle counter (UCPC) (TSI 3010, TSI, United States). The first DMA selects particles of a certain size and these particles are sent into a humidifier or a heated tube, and subsequently routed to the second DMA and CPC to measure particle size change under the elevated relative humidity (~85% RH) or increased temperature (~100°C). The hygroscopic growth factor (HGF) was determined as the ratio of particle mobility diameter at the increased RH to that under the dry condition, whereas the shrinkage factor (SF) was determined as the ratio of particle mobility diameter at increased temperature to that at room temperature (~25°C). Next, the Zdanovskii–Stokes–Robinson (ZSR) model was used to estimate the volume fractions of components of a mixture based on their measured HGFs (Stokes and Robinson 1966). Finally, the dried sea spray particles were collected onto a TEM copper grid (Electron Microscopy Science, United States) using a nanometer aerosol sampler (TSI 3089, TSI, United States) for morphological and elemental analysis using TEM (JEM-2100, JEOL, Japan) and EDS (INCAxsight, Oxford Instruments, United States), respectively.
Artificial seawater (Sigma Aldrich, United States), silica particles (SINOS-3M, ABC Nanotech, South Korea) in deionized (DI) water, and silica particles in artificial seawater were tested to examine the mixing state of the submicrometer sea spay particles. The artificial seawater contained chloride (55%), sodium (31%), sulfate (8%), magnesium (4%), potassium (1%), and other elements (2%) (Mårtensson et al. 2003; Park et al. 2011; Fuentes et al. 2011). A total dissolved solid (TDS) meter (Oakton Instruments, Singapore) was used to determine the mass concentration (ppm) of dissolved species in the artificial seawater. Silica particles were used to represent insoluble submicrometer particles in seawater; they exist in the submicrometer size range with a mean size of 121 nm (polydispersed) and a density of 1.32 g cm−3. The number concentration of silica particles in water (particles per milliliter) was calculated from their known density, size, and mass concentration in water (ppm) under well-dispersed conditions. All solution samples were prefiltered using a MF membrane (mixed cellulose ester membrane, 0.45-μm pore size) (Advantec, Japan) to remove particles larger than 0.45 μm, and were homogeneously mixed using a magnetic stirrer.
Natural seawater samples were obtained from coastal areas near Pohang (latitude: 36°05′05″, longitude: 129°38′26″) (East Sea of Korea), Taean (latitude: 36°74′22″, longitude: 126°07′43″) (West Sea of Korea), and Yeosu (latitude: 34°57′36″, longitude: 127°79′88″) (South Sea of Korea) in July 2010; the selected sampling sites had nearby few anthropogenic sources. Seawater sampled at the Taean site is known to have the strongest biological activity (e.g., higher concentrations of biological materials such as chlorophyll-a, bacteria, and virus) compared to the Yeosu and Pohang sites (Yoo 2008). After sampling, each sample was placed in a portable icebox and stored in a refrigerator (at 4°C) before analysis, in order to prevent contamination during transport. Since natural seawater contains a complex mixture of dissolved species and insoluble particles, MF and UF membrane filtration systems (Park et al. 2009b; Park et al. 2009, 2011) were used to obtain seawater samples that only have dissolved species (after UF membrane filtration) and a mixture of dissolved species and insoluble particles (after MF membrane filtration and before UF membrane filtration).
A dead-end filtration unit was used for MF membrane filtration (pore size: ~0.45 μm) to remove insoluble particles larger than 0.45 μm from the seawater. The MF membrane–filtered seawater was further filtered using a polyethersulfone (PES) UF membrane (pore size: ~5 nm) (GE Osmonics, France) via a laboratory-scale cross-flow filtration unit (Park et al. 2009b). The MF membrane–filtered solution contained both insoluble particles (<0.45 μm) and dissolved species, whereas the UF membrane–filtered solution had only dissolved species because all insoluble particles were removed by the UF membrane. To sample the particles remaining in the UF membrane, the UF membrane was first rinsed with DI water for 15 min to remove dissolved species that could be crystallized on the UF membrane surface due to the concentration polarization. Next, the UF membrane was placed in DI water and sonicated for 3 min to extract insoluble submicrometer particles from the membrane. The solution was then aerosolized using the bubble-bursting system to produce sea spray particles. In addition, the number concentration of insoluble submicrometer particles (20–450 nm) (particles per milliliter) in natural seawater was determined using the membrane filtration–differential mobility analyzer (MF-DMA) technique (Park et al. 2009b; Park et al. 2009, 2011). A constant output atomizer (TSI 3076, TSI, United States) was used to measure the number of insoluble particles in seawater by the MF-DMA system.
To determine the biological materials in natural seawater, the mass concentration of chlorophyll-a (>0.7 μm) and number concentration of bacteria (200–450 nm) were measured. Chlorophyll-a was measured using a UV spectrophotometer (Optizen 2120UV, Mecasys, South Korea), and bacteria was measured using an epifluorescence microscopy method after staining using a 4′-6′ diamino-2-phenylindole (DAPI) solution (Sigma-Aldrich, Germany).
3. Results and discussion
Size distributions of dried particles bubbled from artificial seawater having a TDS of 32 000 ppm are shown in Fig. 2a. Typically, a bimodal size distribution was observed for the sea spray particles. The first mode (~118 nm) appeared due to film droplet formation (small droplets: 0.25–5 μm), while the second mode (~638 nm) was produced by jet droplet formation (large droplets: 3–50 μm) (Andreas 1998; Massel 2007). Next, effects of the bubble flow rate and salinity on the size distribution of the sea spray particles were also examined. With an increase in the bubble flow rate from 2 to 4 L min−1, the number concentration of particles increased by ~1.6 times (Fig. 2a). The higher bubble flow rate caused the number of bubbles to increase, leading to a higher number of particles. The salinity also played an important role in the number of the sea spray particles. In Fig. 2b, as the salinity was increased from 2000 to 32 000 ppm, the number of particles increased by 15 times. During bubble bursting, with a subsequent drying process, a higher number of crystallized salt particles is produced; a similar result was also found in previous studies (Mårtensson et al. 2003; Keene et al. 2007; Tyree et al. 2007).
Since the sea spray particles produced via bubble bursting from natural seawater can contain insoluble matter, insoluble silica particles were added to the artificial seawater to examine the effect of insoluble species on the size distribution of sea pray particles—especially in the submicrometer size range. Silica particles (mode diameter: ~121 nm) having different concentrations (0.01%–0.1% weight) were added into the artificial seawater (32 000-ppm TDS). The solution was filtered using an MF membrane to remove particles larger than 0.45 μm before bubble bursting. As shown in Fig. 3, with the addition of silica particles in the artificial seawater, the number concentration of sea spray particles increased. When the number concentration of silica particles in artificial seawater (particles per milliliter) was increased by 2 times, the resulting number concentration of sea spray particles in air (particles per cubic centimeter) increased by 1.2 times. The number concentration of silica particles in artificial seawater was calculated from a known density, size, and mass concentration of silica particles in the artificial seawater, and the number concentration of sea spray particles (20–600 nm) in air was measured using the SMPS. Figure 3 also showed that a higher bubble flow rate led to an increase in the number of sea spray particles. Our data suggest that although dissolved salt species are a major component of submicrometer sea spray particles, the existence of insoluble submicrometer particles in seawater can lead to an increase in their number concentration. Note that the tested concentration of TDS (~32 000 ppm) in the artificial seawater used here was in the same order of magnitude as typically exists in natural seawater (Bigg et al. 2004; Park et al. 2011).
The mixing states of the size-selected sea spray particles from DI water and artificial seawater having insoluble silica particles were then examined by the HVTDMA system. The HGFs of 100-nm particles produced from DI water with silica particles (2.0% weight), artificial seawater (500 and 5000 ppm of TDS) with silica particles (2.0% weight), and artificial seawater with no particles (5000 ppm of TDS) were measured using the HVTDMA system. In Fig. 4, the HGF of 100-nm particles produced from the DI water with silica particles is ~1.00, whereas that of particles from pure artificial seawater is ~1.87 (i.e., insoluble silica particles are nonhygroscopic, whereas the TDS in artificial seawater are highly hygroscopic). When particles were produced from artificial seawater (5000 ppm of TDS) having silica particles (2.0% weight), it was clearly observed that the particles were an external mixture of two groups of particles having HGFs of 1.33 and 1.68, respectively. Particles having an HGF of 1.33 represent an internal mixture of nonhygroscopic silica particles and hygroscopic salt species [volume fractions of silica and salts are 76% and 24%, respectively, based on the ZSR model (HGF = 1.0 for nonhygroscopic species and HGF = 1.87 for hygroscopic sea salt species; Stokes and Robinson 1966), whereas particles having an HGF of 1.68 are a similar mixture having more salt species (volume fractions of silica and salts are 34% and 66%, respectively, based on the ZSR model). When the TDS in the artificial seawater with silica particles (2% weight) was decreased to 500 ppm, particles having such a high GF disappeared, while when the TDS in the artificial seawater was increased to 20 000 ppm, particles having a low GF became negligible relative to particles having a high GF. The external mixing of silica particles was observed when the TDS in the artificial seawater with silica particles (2.0% weight) was smaller than 10 000 ppm and higher than 500 ppm. Data suggest that particles having a low HGF can be considered significant in sea spray particles when the mass ratio of insoluble particles to dissolved salts is much higher than 2. The number of silica particles in artificial seawater (2% weight) is calculated to be 1013 particles mL−1 by using the mean size and density of silica particles. Note that the TDS of natural seawater is around 32 000 ppm. Thus, the number of insoluble particles should be higher than ~1014 particles mL−1 in the seawater to observe significant external mixing of sea spray particles based on our calculation. Our measured concentration of insoluble particles smaller than 0.45 μm was ~1011 particles mL−1, which was much smaller than the value. Ovadnevaite et al. (2011) found the external mixing of marine aerosols (low GF and high GF) in the ambient atmosphere at coastal size in summer. This might occur because their number of insoluble particles in seawater was much higher than ours. Also, they observed no external mixing (i.e., only high GF) in other days, which is similar to our observation. However, a direct comparison would not be feasible because of the lack of information on the mass ratio of insoluble particles to dissolved salts in the seawater. In addition, the ambient measurements of marine aerosols might be affected by other sources and various metrological conditions, and it would not be easy to separate primary and secondary marine aerosols.
HVTDMA measurements were conducted on sea spray particles produced from natural seawaters sampled from the coastal areas around the West Sea, East Sea, and South Sea of South Korea. The natural seawater samples before being sprayed were filtered using an MF membrane (0.45 μm) to remove large particles. The general properties of natural seawater are summarized in Table 1. The pH, conductivity, and salinity of seawater sampled from Pohang, Taean, and Yeosu were similar (within 5%). However, the ultraviolet absorbance values at 254 nm (UVA254) (i.e., indicator for the existence of aromatic organic compounds) (Weishaar et al. 2003) for seawater from Taean and Yeosu were higher than from Pohang. This difference suggests that the amount of organics in the seawater from the Taean and Yeosu sites was relatively higher compared to the seawater from the Pohang site. We found that the mass concentrations of chlorophyll-a (size range: >0.7 μm) in the seawater sampled from Taean and Yeosu were also ~3 times higher than was found in the Pohang samples. The number concentrations of insoluble submicrometer particles (20–450 nm) and bacteria (200–450 nm) were the highest value at the Taean site, suggesting that the biological activity was the strongest at the Taean site.
Figure 5 shows that the HGFs of sea spray particles (~100 nm) produced from natural seawater were 1.76, 1.76, and 1.70 at Pohang, Yeosu, and Taean, respectively, all of which are lower than obtained from pure artificial seawater (1.87). These values occurred because the natural seawater included nonhygroscopic organic matter, leading to a lower HGF than pure sea salt particles. By using the ZSR model (HGF = 1.0 for nonhygroscopic species and HGF = 1.87 for hygroscopic sea salt species) (Stokes and Robinson 1966), the volume fractions of nonhygroscopic species in the single particles was calculated to be in the range of 20%–29%. Ming and Russell (2001) reported that the presence of 30% organic species led to reduce the HGF by 15% based on their modeling data. Several HGF measurements also reported 10%–30% organic fraction in the submicrometer sea spray particles (Swietlicki et al. 2000; Zhou et al. 2001). Nonhygroscopic organic matter can be soluble or insoluble in natural seawater, which will be clarified later. In any case, there were few less-hygroscopic particles (i.e., no external mixing) observed. The external mixing of sea spray particles was observed only when the mass ratio of insoluble particles to dissolved salts was higher than 2 when tested using the artificial seawater in the previous section. One possibility for the existence of few particles having low GF in sea spray particles is that the smaller amount of insoluble particles exists relative to the TDS in the given size range in current seawater samples.
When the natural seawater was further filtered by the UF membrane (i.e., to remove most insoluble particles), the HGFs of the 100-nm sea spray particles after UF filtration were 1.71–1.76, quite similar to those before UF filtration (see Fig. 5). As such, our data suggest that nonhygroscopic insoluble particles were not included in the 100-nm sea spray particles in a sufficient amount to affect the observed HGF before and after UF filtration. The lower HGF of particles from natural seawater than from artificial seawater having only dissolved salt species suggests that a significant amount of dissolved nonhygroscopic organic matters were included in the 100-nm particles.
After the UF filtration of natural seawater, the remaining particles in the UF membrane surface were collected, and their HGF was measured after aerosolization (see section 2). The mode diameter of these particles was around 41 nm. In Fig. 5, the HGF of the 40-nm particles was 1.52–1.56, much lower than for particles produced from seawater after UF membrane filtration, suggesting that nonhygroscopic insoluble particles exist, though not in a sufficient amount to affect the HGF of the 100-nm particles before and after UF filtration, which is consistent with the observation discussed earlier. Some dissolved salt species remaining on the surface the UF membrane could be crystallized during the collection procedure (Ning and Troyer 2007), thereby contributing to the observed HGF of these particles.
The HGFs of other sizes of sea spray particles (80, 100, 120, and 140 nm) produced from natural seawater from the Taean and Yeosu sites were also measured using the HVTDMA system. The volume fractions of the nonhygroscopic species in the single particles as a function of size were determined using the ZSR method (Fig. 6). The fractions ranged from 22% to 25% with no significant dependence on size (80–140 nm). However, a difference was observed between the two sites. Volume fractions of the nonhygroscopic species in sea spray particles from the Taean site were much higher than from Yeosu for all sizes. The difference is caused by the higher amount of nonhygroscopic dissolved organic species at the Taean samples. This finding is consistent with the stronger biological activity in the Taean samples.
Volatility measurements of the size-selected particles (100 nm) produced from natural seawaters (at a heater temperature of 100°C) were conducted to obtain the fraction of volatile species in the sea spray particles, calculated using the SF. In Fig. 7, the volume fractions of volatile species are 4%–5%, much lower than the fraction of nonhygroscopic species estimated above. Note that a certain amount of dissolved organic species that can evaporate at 100°C are included in the volatile fraction of the sea spray particles (the organic species that can evaporate at >100°C were not included in this fraction). It was previously reported that the organic mass fraction was typically larger for submicrometer particles (Barker and Zeitlin 1972; Hoffman and Duce 1976; O’Dowd et al. 2004; Facchini et al. 2008). And Modini et al. (2010) reported that submicrometer (71–77 nm) sea spray particles contained an organic volume fraction of ~8%, based on VTDMA measurements at a peak temperature of 583°C.
Dried sea spray submicrometer particles from natural seawater were also collected on a TEM grid in order to investigate their morphology and elemental composition. Figure 8 shows several mixtures having elements possibly originated from organics (C and O), salts (Na, Cl, Mg, S, and Cl), iron colloids (Fe), and clay minerals (Al, Si, and Ca). Quantification of such types of particles was not conducted here because of the limited number of particles examined in the TEM–EDS. Also, since volatile organic species is very sensitive to electron beam in the TEM, their accurate measurement would be difficult. Subsequent TEM–EDS data support the fact that organic matter is included in submicrometer sea spray particles.
Number size distributions of sea spray particles under varying bubble flow rates and salinity were determined using SMPS (20–600 nm) and a PSD analyzer (500 nm–10 μm), and the mixing state, hygroscopicity, and volatility of size-selected sea spray particles were determined using the HVTDMA technique. Pretreatment of seawater samples by MF and UF membranes before being sprayed led to seawater samples that had only dissolved species or a mixture of dissolved species and insoluble submicrometer particles, enabling us to better understand their contribution to the formation of sea spray particles. Data showed that increasing the bubble flow rate and salinity of seawater also caused the number of sea spray particles to increase. Then, with the addition of insoluble silica particles to the artificial seawater, the number of sea spray particles further increased, and an external mixture of two groups of particles (less and more hygroscopic particles) was observed when the mass ratio of insoluble particles (~100 nm) to dissolved salts was higher than 2. Such external mixing was not observed for sea spray particles from natural seawater because of the high concentration of dissolved salt species (>30 000 ppm TDS). Another interesting finding was that the HGF of sea spray particles from natural seawater (80, 100, 120, and 140 nm) was lower (1.70–1.76) than from a pure artificial seawater solution (1.87), suggesting that sea spray particles from natural seawater include an internal mixture of nonhygroscopic dissolved organic matter and hygroscopic salts species. The volume fractions of the nonhygroscopic species in the single particles were calculated to be in the range of 20%–29%. The volume fraction of the nonhygroscopic species was not found to be dependent on the size (80–140 nm), but varied with location. The volume fraction of the nonhygroscopic species was the highest at the Taean site—where the strongest biological activity occurred—suggesting that biological organic matter was the main contributor of nonhygroscopic species in submicrometer-sea-spray particles. The volatility data also showed that volatile species in the single particles at 100°C exist, at volume fractions of 4%–5%. Finally, the existence of organic species as an internal mixture with salt species was supported by subsequent TEM–EDS data.
The research described in this paper was supported by the National Leading Research Laboratory program funded by the National Research Foundation of Korea (NRF) (Grant 2011-0015548) and partially supported by the National Research Foundation of Korea Grant NRF-C1ABA001-2011-0021066 funded by the South Korean government.