Current Issue

PM₁₀ and PM_2.5 samples were collected at the Gosan background site (33°17ʹN, 126°10ʹE) on Jeju Island, Korea, from 2013 to 2015 using PM₁₀ and PM_2.5 sequential air samplers (APM Engineering, PMS-103 and PMS-104, Korea) with a teflon filter (Pall Co., Zefluor^TM, PTFE 47 μm, 2.0 μm, USA). Flow rates for the air samplers were maintained at 16.7 L/min using a mass flow controller. The collected sample filters were sealed in petri dishes (SPL life science, PS, 52.7×12.6 mm) onsite, and then dried in a desiccator until reaching a constant weight. Dried sample filters were stored in a freezer maintained at -24°C until analysis (Kim et al., 2014). In total 315 sample filters were collected for PM₁₀ and 301 for PM_2.5. The PM₁₀ and PM_2.5 sample filters were divided into four groups based on meteorological conditions: 15 and 14 for PM₁₀ and PM_2.5 during Asian dust days, 23 and 24 during haze days, and 116 and 115 during mist days, respectively. The total number of non-event samples were 161 and 148 for PM₁₀ and PM_2.5, respectively.

Water-soluble ionic components were extracted from the sample filters by adding 0.2 mL of ethanol and 30 mL of ultrapure water (18.2 MΩ · cm) and agitating them in an ultrasonic extractor for 30 min and shaker for 1 h (200 rpm). After filtering insoluble species from the extract using a syringe filter (Whatman, PVDF syringe filter, 0.45 μm), the filtrates were used for ionic component analysis.

The major water-soluble ionic species were analyzed by ion chromatography (Modula IC, equipped with a 907 IC pump and 732 IC detector, Metrohm, Herisau, Switzerland) using the Metrohm Metrosep Cation C6-150 column with 3.0 mM nitric acid eluent for cations, and the Metrosep A-SUPP-16 column with 7.5 mM Na₂CO₃ eluent and 200 mM H₂SO₄ suppressor solutions for anions.

Standard solutions for cation (NH₄⁺, Na⁺, K⁺, Ca²⁺, and Mg²⁺) analyses were prepared using 100 μg/mL AccuStandards, and diluted to various concentrations (0.1-5.0 μg/mL) to make the standard calibration curves. The standard calibration curve calculated from the standard solutions showed good linearity; the coefficient of determination (R²) was higher than 0.9999. The standard solutions for anion (SO₄^2-, NO₃^-, and Cl^-) analyses were prepared using primary standard reagents (Sigma, 99.999% (NH₄)₂SO₄, 99.99% KNO₃, and 99.99% NaCl). The standard calibration curve was calculated using standard solutions with concentrations of 0.1-5.0 μg/mL. The standard solutions for the analysis of trace organic acid ions (HCOO^-, CH₃COO^-), F^-, and CH₃SO₃^- were made using high purity reagents (Sigma, 99.99% NaF, 99.9% CH₃COONa·3H₂O, 99% HCOONa, and 98% CH₃SO₃Na), and diluted to concentrations of 0.01-0.5 μg/mL for calculating standard calibration curves. The correlation between the standard calibration curve and standards for analyzing anions and organic acids showed good linearity; the coefficients of determination were ≥0.9999.

Based on the seven repeated analyses by ion chromatography, the instrument detection limits (IDLs) for the IC analyses were in the range of 0.3-10.7 μg/mL for the twelve analyzed ionic species, and the coefficient of variation (CV) ranged from 0.1 to 3.3%.

The mass concentrations of PM₁₀ and PM_2.5 during non-event days excluding Asian dust, haze, mist and heavy rainfall (over 3 mm), were 35.7±15.3 μg/m³ and 14.8±8.5 μg/m³, respectively. The mass concentration observed using the β-ray absorption method by Korea Meteorological Administration, was 25.8±11.2 μg/m³, which was somewhat lower than this study. The correlation coefficient between the two measurements was about r=0.75. In addition, Lee et al. (2015) also demonstrated a somewhat lower result than the PM_2.5 mass concentration, 18.6 μg/m³ measured in 2008-2012 at the Gosan area. This value was also 1.6-1.8 times lower compared to the PM₁₀ mass concentrations; 57.8, 64.5, 61.2, and 59.4 μg/m³, respectively, measured in 2009 in the major metropolitan areas of Seoul, Gyeonggi, Incheon, and Busan (Table 1).

The ionic concentrations of PM₁₀ on the non-event days were in the order; nss-SO₄^2->Cl^->NO₃^->Na⁺>NH₄⁺>K⁺>Mg²⁺>nss-Ca²⁺>CH₃COO^->HCOO^->CH₃SO₃^->F^-. The major secondary pollutants (nss-SO₄^2-, NO₃^-, and NH₄⁺) accounted for 60.0% of total ionic composition, followed by the sea salt (Na⁺, Cl^-, and Mg²⁺) at 34.2%, the organic acid (CH₃COO^- and HCOO^-) at 2.0%, and the soil components (nss-Ca²⁺) at 1.8% (Table 2 and Fig. 1).

Fig. 1. 
Relative contributions of water-soluble ionic species to PM₁₀ and PM_2.5 particle compositions.

The ionic concentrations of PM_2.5 on non-event days were in the order; nss-SO₄^2->NH₄⁺>NO₃^->Na⁺>K⁺>Cl^->nss-Ca²⁺>Mg²⁺>CH₃COO^->HCOO^->CH₃SO₃^->F^-. Of these components, nss-SO₄^2-, NH₄⁺, and NO₃^-, which are of anthropogenic origin, were as high as 4.56, 1.80, and 0.85 μg/m³, respectively. In contrast, Na⁺, Cl^-, and nss-Ca²⁺ components were relatively low in concentration. Secondary pollutants (nss-SO₄^2-, NO₃^-, and NH₄⁺) accounted for 90.4% of the total ion composition, accounting for 1.5 times more of the total than in PM₁₀. Sea salt (Na⁺, Cl^-, Mg²⁺) accounted for 5.7%, organic acid (HCOO^-, CH₃COO^-) accounted for 0.9%, and the soil component (nss-Ca²⁺) accounted for 0.8% of the total, demonstrating low relative abundances.

Comparing the distributions of ion components in PM_10-2.5 and PM_2.5, shown in Table 1, the concentration ratio of PM_10-2.5/PM_2.5 is <1 for NH₄⁺ and nss-SO₄^2-, indicating that they were more distributed in the fine particle mode. However, Na⁺, Cl^-, nss-Ca²⁺, Mg²⁺, and NO₃^- had higher abundances in the coarse particle mode, and K⁺ was uniformly distributed in coarse and fine particles.

In atmospheric aerosols of urban areas, NO₃^- concentration is usually higher than nss-SO₄^2-. This higher abundance is from NO₃^- generation related to energy use and the influence of mobile pollutants. In previous studies, the ratios of nss-SO₄^2-/NO₃^- in PM_2.5 fine particles in the urban areas of Seoul, Chuncheon, Suwon, Gwang-ju, and Chungju were 1.06, 1.62, 1.54, 1.21 and 1.03, respectively (Kang et al., 2015; Lee et al., 2009; Jung and Han, 2008). In addition, the values were 1.99, 1.48, and 1.66 in the urban areas of New York, Beijing, and Shanghai, respectively (Wang et al., 2006, 2005). In contrast, the nss-SO₄^2-/NO₃^- ratios of Baekryongdo and Deokjeokdo, which are domestic background areas, were 3.34 and 3.57, respectively, much higher than in the cities (Lee et al., 2010, 2002). The nss-SO₄^2-/NO₃^- ratio in the mountainous site of Jeju Island showed a large value of 5.3 for the PM_2.5 fine particles, indicating that the effect of anthropogenic emission due to mobile pollution sources was relatively low.

Atmospheric sulfur and nitrogen oxides are converted to sulfuric acid and nitric acid by photochemical oxidation; they are also present in aerosols in the form of sulfate or nitrate by neutralizing reactions with ammonia or a soil basic substance. Trace amounts of organic acids are also neutralized by ammonia and calcium carbonate (Seinfeld and Pandis, 1998). The acidification contribution from sulfuric acid and nitric acid was evaluated from equivalent concentrations of SO₄^2- and NO₃^-, and compared by meteorological phenomena (Table 4). As shown in the table, the sum of cationic and anioic equivalent concentrations in PM₁₀ and PM_2.5 were similar on non-event and event days, suggesting that inorganic acids, such as sulfuric and nitric acids, significantly contributed to aerosol acidification. Furthermore, the sums of the cationic and anionic equivalent concentrations were high during Asian dust and haze days compared to non-event days. However, the sums of the cationic and anionic equivalent concentrations for mist days were relatively lower than those of Asian dust and haze days.

Ammonia and calcium carbonate are the primary contributors to the neutralization of acidic substances. The degree of neutralization by these two substances can be evaluated by determining the neutralization factor (NF) through the following equations (1) and (2) (Galloway and Keene, 1989)

where [nss-SO₄^2-], [NO₃^-], [HCOO^-], [CH₃COO^-], [NH₄⁺], and [nss-Ca²⁺] are each component’s equivalent concentration. As shown in Table 5, the neutralization factors for ammonia on non-event days were 0.74 and 0.91 for PM₁₀ and PM_2.5, respectively; the degree of neutralization by ammonia was higher for fine particles. However, the neutralization factors for calcium carbonate were 0.10 and 0.03 for PM₁₀ and PM_2.5, respectively, indicating that it had a large influence on coarse particles. During Asian dust days, the neutralization factors for ammonia were 0.65 and 0.83 for PM₁₀ and PM_2.5, respectively. The neutralization factors for calcium carbonate were 0.23 and 0.03, respectively, demonstrating that the degree of neutralization by calcium carbonate was much higher in the coarse particles. However, during haze days, the neutralization factor for ammonia was 0.91 for PM_2.5, which was higher than other event days. In addition, the neutralization factor for ammonia was 0.88 during mist days.

Based on the neutralization factors, it was found that the acidic substances in fine particles were mainly neutralized by ammonia, but the neutralization in coarse particles was occurred by calcium carbonate. In particular, the degree of neutralization by calcium carbonate was higher during Asian days in PM₁₀ particles, and the degree of neutralization by ammonia during haze days were much higher in PM_2.5.

Based on the sampling days (161 days), the cluster back-trajectory analysis was performed in order to investigate the transport pathways of air masses using NOAA’s HYSPLIT4 (Hybrid Single-Particle Lagrangian Integrated Trajectory) model and GDAS meteorological data provided by the National Centers for Environmental Prediction (NCEP) (Draxler and Rolph, 2013; Kim et al., 2004). The back-trajectory analysis was created for 72 hours based on 00 UTC for the corresponding sampling date. The altitude of the starting point for this cluster back-trajectory analysis was set at 72 m above sea level for Gosan site (Fig. 4).

Fig. 4. 
Cluster back-trajectories for air masses corresponding to sampling dates at the Gosan Site.

As shown by the results of the cluster back-trajectory analysis, the pathways of air masses transported to the Gosan site were initially classified into three different categories: Cluster 1 (China continent), Cluster 2 (Korean Peninsula), and Cluster 3 (North Pacific). As shown in the figure, the frequency distribution of all transport pathways during the entire study period was 60% (97 days), 35% (56 days), and 5% (8 days) for Clusters 1, 2, and 3, respectively, with inflow pathways from China continent accounting for the largest fraction. Comparing the concentrations of major ionic components by transport pathways of air masses, NH₄⁺, nss-SO₄^2-, and NO₃^- were highest, at 2.27, 5.36, and 3.24 μg/m³, respectively, when the air mass moved from China continent (Cluster 1) (Table 6). The concentration of soil-originated nss-Ca²⁺ was also high when the air masses were moving in from mainland China. In contrast, marine components, such as Na⁺ and Cl^-, were relatively higher when air masses were moving in from the North Pacific (Cluster 3).