Comparisons of Spatial and Temporal Variations in PM_2.5-Bound Trace Elements in Urban and Rural Areas of South Korea, and Associated Potential Health Risks

Abstract

PM_2.5-bound trace elements were chosen for health risk assessment because they have been linked to an increased risk of respiratory and cardiovascular illness. Since the Korean national air quality standard for ambient particulate matter is based on PM_2.5 mass concentration, there have only been a few measurements of PM_2.5 particles together with trace elements that can be utilized to evaluate their effects on air quality and human health. Thus, this study describes the trace elements bound to PM_2.5 in Seoul (urban area) and Seosan (rural area) using online nondestructive energy-dispersive X-ray fluorescence analysis from December 2020 to January 2021. At both the Seoul and Seosan sites, S, K, Si, Ca, and Fe constituted most of the PM_2.5-bound trace elements (~95%); major components such as S, K, and soil (estimatedcalculatedcalculated based on oxides of Si, Fe, Ca, and Ti) were presumably from anthropogenic and crustal sources, as well as favorable meteorological conditions. During winter, synoptic meteorology favored the transport of particles from severely contaminated regions, such as the East Asian outflow and local emissions. The total dry deposition flux for crustal elements was 894.5 ± 320.8 µg m⁻² d⁻¹ in Seoul and 1088.8 ± 302.4 µg m⁻² d⁻¹ in Seosan. Moreover, potential health risks from the trace elements were estimated. Cancer risk values for carcinogenic trace elements (Cr, As, Ni, and Pb) were within the tolerable limit (1 × 10⁻⁶), suggesting that adults and children were not at risk of cancer throughout the study period in Seoul and Seosan. Furthermore, a potential risk assessment of human exposure to remaining carcinogens (Cr, As, Ni, and Pb) and non-carcinogens (Cu, Fe, Zn, V, Mn, and Se) indicated that these trace elements posed no health risks. Nevertheless, trace element monitoring, risk assessment, and mitigation must be strengthened throughout the study area to confirm that trace-element-related health effects remain harmless. Researchers and policymakers can use the database from this study on spatial and temporal variation to establish actions and plans in the future.

1. Introduction

Fine particulate matter (PM_2.5) plays a crucial role in determining visibility and affects climate change and human health [1,2,3,4,5]. Furthermore, the chemical components of PM_2.5 may be associated with a number of adverse health effects (e.g., respiratory diseases, cardiovascular disease, pregnancy/birth outcomes) [6,7,8].

In recent decades, PM_2.5 pollution has become a severe environmental concern in urban and rural areas in Asia [9,10,11,12,13,14,15,16]. Field measurements have shown that PM_2.5 is a complex mixture of organic materials, inorganic salts, and trace elements [17,18,19,20]. Si, Fe, Ca, K, Ti, Cr, Mn, Zn, Ni, Cu, Pb, As, V, and Ba are significant contributors to PM_2.5-bound trace elements [21,22,23,24]. Their sources include regional aerosols, local road traffic, coal combustion, soil, and metal industries [4,10,25]. Due to their long residence time in the atmosphere, trace elements bound to PM_2.5 can be transported downwind [26]. Moreover, PM_2.5-bound trace metals may be deposited in terrestrial and aquatic ecosystems via the processes of dry deposition—which is a consistent and reliable atmospheric cleansing process—or wet deposition [27,28,29].

The health effects of trace elements must be addressed, because their potential carcinogenic and non-carcinogenic effects on human health have raised concerns [30,31,32]. As stated by the International Agency for Research in Cancer, trace metals such as Cd, Ni, Ar, Pb, and Cr can cause cancer in humans and animals [33,34,35,36]. When the CR is 10⁻⁴, there is a serious risk of cancer. When the CR is between 10⁻⁶ and 10⁻⁴, the risk is acceptable. When CR is 10⁻⁶, there is less risk or the risk may be ignored [33,34]. Additionally, according to several epidemiological studies, PM_2.5 exposure has been associated with substantial health risks [37,38]. The complex characteristics of ambient PM_2.5 necessitate knowledge of its sources, sinks, and temporal and spatial variations. Trace elements can be utilized as markers for source apportionment studies of these particles, owing to their distinctive source characteristics.

Few studies on PM_2.5 and trace elements have compared urban and rural areas in South Korea [39,40,41]. To fill these gaps in PM_2.5 research, we conducted simultaneous PM_2.5 monitoring studies at representative sites in the urban (Seoul) and rural (Seosan) regions of South Korea during the winter season. Seoul is affected by local anthropogenic emissions and regional outflows that cause high PM_2.5 concentrations [42]. Seosan was chosen based on recent studies that found high PM_2.5 concentrations in the area. However, the causes of this are yet to be known [42,43]. In this study, we evaluated the temporal and spatial variability of PM_2.5 and trace elements, along with their possible health risks to humans throughout the winter of 2020–2021 at rural and urban sites. The findings concerning trace elements, their sources, and their estimated potential risks to humans in Seoul (urban) and Seosan (rural) will have important implications for controlling air pollution and improving wellbeing. This research strengthens the existing research by estimating potential health risks in the context of comparing two different locations in South Korea and similar locations worldwide.

2. Materials and Methods

2.1. Study Site

Simultaneous measurements were conducted in Seoul (urban area) and Seosan (rural area) from 15 December 2020 to 15 January 2021. Seoul and Seosan are approximately ~99 km apart. Seoul is a city with a population of approximately 10 million and an area of 605 km² (Figure S1) [44]. The measurements in Seoul were performed at an air-quality measurement site run by the National Institute of Environmental Research (NIER) in the metropolitan area of Bulgwang-dong, Eunpyeong-gu, Seoul (37.61° N, 126.93° E). The site is surrounded by many highly congested roads and residential complexes [44,45,46].

Seosan is a rural area in Chungcheong-do with a population of approximately 0.18 million and an area of 742 km² (Figure S1) [43,46]. This area is considered an “agro-industrial area”, which combines farms, power plants, and markets. The measurements were conducted at an air-quality monitoring station in the Chungcheong region operated by the NIER (36°78′ N, 126°49′ E).

2.2. Measurements

The concentrations of PM_2.5 and PM_2.5-bound trace elements were simultaneously quantified hourly at the two sites from 15 December 2020 to 15 January 2021. The β-ray absorption technique was used to measure the mass concentrations of PM_2.5 (BAM1020, Met One Instruments Inc., Grants Pass, OR, USA). Si, K, S, Ca, Ti, Fe, V, Mn, Cr, Ni, Cu, As, Zn, Se, Ba, Br, and Pb were quantified using an ambient online nondestructive energy-dispersive X-ray fluorescence monitor (Xact625i, Ambient Metal Monitor, COOPER ENVIRONMENTAL, Tigard, OR, USA) [47,48]. The method detection limit (MDL) varied from 1.17 to 17.8 ng m⁻³ for Si, S, Cl, and K (Table S1), [47,49,50] and from 0.1 to 0.39 ng m⁻³ for Ba, Ca, Fe, Mn, Pb, V, Cr, Ni, and Br. For Se, Cu, Zn, and As, the MDL ranged from 0.063 to 0.08 ng m⁻³ [49,50]. A high percentage of samples were below the detection limit for Ba, Ni, and V in Seoul, and for V, Si, and Cr in Seosan (

Table 1. Summary statistics of trace element concentrations (ng m⁻³) in PM_2.5 quantified in Seoul and Seosan, South Korea, from 15 December 2020 to 15 January 2021.

	Seoul			Seosan
Element	Mean ± SD a	% BDL b	Median	Mean ± SD	%BDL	Median
Si	306 ± 527	30	101	460 ± 761	47	147
S	1444 ± 908	0.9	1168	1265 ± 673	2.3	1153
K	250 ± 131	0.9	219	294 ± 203	2.2	239
Ca	81 ± 113	0.9	47	101 ± 160	2.2	53.0
Ti	10.6 ± 10	0.9	7.8	12 ± 15	2.2	6.6
V	0.7 ± 0.5	33	0.5	0.7 ± 0.7	56	0.4
Cr	1.2 ± 1.0	7.7	0.9	1.3 ± 1.4	17	0.8
Mn	13 ± 8.2	0.9	11	16 ± 22	2.6	10
Fe	183 ± 135	1.0	142	210 ± 201	2.2	153
Ni	0.8 ± 0.7	58	0.7	0.7 ± 0.8	12.6	0.4
Cu	4.4 ± 4.0	0.9	2.9	4.1 ± 4.0	3.4	3.1
Zn	49 ± 85	0.9	37	51 ± 41	2.6	39
As	4.7 ± 3.8	0.9	3.8	3.9 ± 3.3	3.3	3.1
Se	1.0 ± 1.0	7.4	0.5	1.4 ± 1.4	7.9	1.0
Br	7.3 ± 6.4	0.9	4.8	8.8 ± 6.2	2.3	7.5
Ba	2.2 ± 1.8	61	1.7	-	−100	-
Pb	21 ± 11	0.9	19	21.0 ± 11.8	4.3	19.7
Temperature (°C)	−3.5 ± 6.3	-	-	−2.64 ± 5.63	-	-
Relative humidity (%)	56.9 ± 15.7	-	-	70.0 ± 17.1	-	-
Wind speed (m/s)	2.38 ± 1.09	-	-	1.79 ± 1.67	-	-
Wind direction (degree)	205 ± 110.1	-	-	146 ± 138	-	-
Rainfall (mm)	0.76 ± 0.93	-	-	0.33 ± 0.25	-	-

^a Standard deviation; ^b below detection limit.

).

2.2.1. Dry Deposition Calculations

Dry deposition flux (DDF) is calculated by multiplying element concentrations by dry deposition velocities [24,51]. DDFs (µg m⁻² d⁻¹) of trace elements were estimated for the Seoul and Seosan sites. Herein, the dry deposition rates of specific species were expected to be 2 cm s⁻¹ for dust-originated elements, 1 cm s⁻¹ for anthropogenic elements, and 0.5 cm s⁻¹ for remaining trace elements; these were selected based on previous research on these species of PM_2.5 [24,52,53,54]. In prior works, similar dry deposition rates were estimated using the DDF of atmospheric trace elements [24,55].

2.2.2. Exposure Estimations

Non-carcinogenic and carcinogenic health risk assessments for trace elements (Figure S4) were conducted using the methods of the Integrated Risk Information System and the United States Environmental Protection Agency database [56,57,58,59]. PM_2.5-bound trace elements pose potential health risks to children and adults through dermal contact, inhalation, and ingestion; therefore, the dermal absorbed dose (DAD), chemical daily intake (CDI), and exposure concentration (EC) were assessed [59] (Equations (1)–(3), respectively). The exposure assessment formulae are as follows:

$${\mathrm{DAD}}_{\mathrm{dermal}}=\frac{\mathrm{C}\times \mathrm{SA}\times \mathrm{AF}\times \mathrm{ABS}}{\mathrm{BW}}\times \frac{\mathrm{EF}\times \mathrm{ED}}{\mathrm{AT}}\times \mathrm{CF},$$

(1)

$${\mathrm{CDI}}_{\mathrm{ingestion}}=\frac{\mathrm{C}\times \mathrm{IngR}}{\mathrm{BW}}\times \frac{\mathrm{EF}\times \mathrm{ED}}{\mathrm{AT}}\times \mathrm{CF},$$

(2)

$${\mathrm{EC}}_{\mathrm{inhalation}}=\mathrm{C}\times \frac{\mathrm{InhR}\times \mathrm{EF}\times \mathrm{ED}}{\mathrm{BW}\times \mathrm{PEF}\times \mathrm{ATn}},$$

(3)

where C is the concentration (µg m⁻³). The exposure duration (ED) is 6 y for children and 24 y for adults. IngR is the ingestion rate (mg day⁻¹), and EF is the exposure frequency (180 d y⁻¹ for residents). The ingestion rate is 200 and 100 mg d⁻¹ for children and adults, respectively. Body weight (BW) is 70 and 15 kg for adults and children, respectively; CF (conversion factor) is 10⁻⁶ kg mg⁻¹; ET (exposure time) is 24 h d⁻¹; AT (the time average for non-carcinogenic elements) is equal to ED × 365 days, where for carcinogens it is equal to 70 years × 365 days [24,59,60] (Table S2). The average time (ATn) for non-carcinogens is ATn = 365 d × ED × 24 h d⁻¹, while for carcinogenic elements is ATn = 70 y × 24 h × 365 d y⁻¹. AF (the skin adherence factor) is 0.2 mg cm⁻¹ h⁻¹; SA denotes the exposed skin area (3300 and 2800 cm² for adults and children, respectively). According to the US EPA, the dermal absorption factor (or ABS) is 0.03, 0.1, and 0.01 for As, Pb, and the remaining elements, respectively (Table S2) [24,59,61].

2.2.3. Non-Carcinogenic Risk Characterizations

To calculate non-carcinogenic risks, we used the hazard index (HI = ΣHazard Quotient (HQ). The non-CR caused by PM_2.5-bound trace elements (Cr, Mn, Ni, Cu, V, Zn, As, and Pb) was determined utilizing the following equations (Equations (4)–(7)):

$${\mathrm{HQ}}_{\mathrm{ingestion}}=\frac{\mathrm{CDI}}{\mathrm{RfDo}} ,$$

(4)

$${\mathrm{HQ}}_{\mathrm{dermal}}=\frac{\mathrm{DAD}}{\mathrm{RfDo}\ast \mathrm{GIABS}} ,$$

(5)

$${\mathrm{HQ}}_{\mathrm{inhalation}}=\frac{\mathrm{EC}}{RfCi\ast 1000 \mu \mathrm{g} {\mathrm{mg}}^{-1}} ,$$

(6)

HI = HQ_ingestion + HQ_dermal + HQ_inhalation,

(7)

where the oral reference dosage is in mg kg⁻¹ d⁻¹, while the gastrointestinal absorption factor and the inhalation reference concentration are in µg m⁻³, and they are denoted by RfDo, GIABS, and RfCi, respectively (Table S2). The probable non-cancer consequences were measured using the HI, estimated as the sum of the HQ values [57,60,62]. When the HI values of non-carcinogenic trace elements were below one, they did not pose a risk to adults or children [56]. The potential for non-carcinogenic risks increases if the HI increases above one [59].

2.2.4. Cancer Risk Characterizations

CR is the likelihood of developing cancer due to lifelong contact with carcinogens [10]. Intended for regulatory purposes, the acceptable or tolerable hazard of emerging cancer throughout of a human’s life ranges from 1 × 10⁻⁴ to 1 × 10⁻⁶ [63]. The CR posed by trace elements (Cr, Ni, As, and Pb) through various paths—including inhalation, ingestion, and dermal contact—was estimated and used to calculate the total cancer hazard (CR_risk) as follows (Equations (8)–(11)) [56,59]:

$${\mathrm{CR}}_{\mathrm{ingestion}}=\frac{\mathrm{CDI}}{\mathrm{SF}0} ,$$

(8)

$${\mathrm{CR}}_{\mathrm{dermal}}=\frac{\mathrm{DAD}}{\mathrm{SF}0\ast \mathrm{GIABS}} ,$$

(9)

$${\mathrm{CR}}_{\mathrm{inhalation}}=\mathrm{IUR}\ast \mathrm{EC} ,$$

(10)

$${\mathrm{CR}}_{\mathrm{risk}}={\mathrm{CR}}_{\mathrm{ingestion}}+{\mathrm{CR}}_{\mathrm{dermal}}+{\mathrm{CR}}_{\mathrm{inhalation}},$$

(11)

where the inhalation unit risk is in µg m⁻³, the inhalation reference concentration is in µg m⁻³, and the oral slope factor is in mg kg⁻¹ d⁻¹, denoted by IUR, RfCi, and SFo, respectively. A CR of greater than 10⁻⁴ indicates a severe risk of cancer [64]. In contrast, a CR of less than 10⁻⁶ indicates either a low risk or an acceptable risk [64]. Previous research has determined that the concentration proportion of Cr(VI) and Cr(III) is 1:6. As a result, the Cr(VI) concentration used to calculate the CR was adjusted to 1/7 of the total Cr concentration observed [24].

2.2.5. Principal Component Analysis (PCA)

Principal component analysis (PCA) was used in the Origin application (OriginPro, Version 2022, OriginLab Corporation, Northampton, MA, USA) to identify potential sources of PM_2.5-bound trace elements in Seoul and Seosan. PCA was used to identify site-specific sources using winter data by site [65,66]. PCA was performed using a varimax rotation of the data matrix, and the principal components (PCs) or factors (Fs) were extracted in accordance with a set of quality criteria: the number of eigenvalues greater than 1 (Kaiser criterion), with more than 70–85% of cumulative variance explained by the corresponding PCs. The loadings of obvious variables were near one, while those of non-obvious variables were near zero. Variables with a loading of 0.5 were considered to be factor indicators.

2.2.6. Enrichment Factor

The enrichment factors (EFs) were calculated in order to determine the degree of contamination of each element in PM_2.5, and to determine whether it primarily came from anthropogenic or natural sources [44,67]. The definition of EF is the ratio of a particular element’s concentration (X_i) to Fe (X_reference, a reference element for crustal dust), normalized to the same ratio in the upper continental crust and calculated as follows using Equation (12) [68]:

$$\mathrm{Enrichment} \mathrm{Factors} =\frac{{\left(\frac{X_i}{X_{\mathrm{reference}}}\right)}_{Aerosol}}{{\left(\frac{X_i}{X_{\mathrm{reference}}}\right)}_{Crust}}$$

(12)

where ()aerosol and ()crust refer to the PM_2.5 sample and continental crust, respectively. The reference element used in this study was iron (Fe), which is also supported by previous research in Korea [44].

2.2.7. Potential Source Contribution Function (PSCF)

The PSCF was calculated using PM_2.5-bound trace element concentrations, and air-mass back-trajectories were calculated using NOAA’s HYSPLIT4 (Hybrid Single-Particle Lagrangian Trajectory) model with Global Data Assimilation System (GDAS) grid meteorological data to identify the likely locations of the regional sources for long-range transboundary aerosols [69,70]. This calculates the probability that a source exists at latitude i and longitude j. The PSCF for the ijth grid cell, PSCF_ij, was determined for PM_2.5 values as follows:

$$PSC{F}_{ij}=\frac{m_{iij}}{n_{ij}}$$

(13)

where n_ij is the total number of trajectory endpoints that fall into the ijth cell, and m_ij is the number of endpoints in the same cell corresponding to trajectories associated with concentration values at each receptor site above a predefined criteria value [70]. As a result, cells with a high PSCF value are expected to produce high concentration values at receptor locations and, thus, can be reasonably believed to be potential source areas [69]. The details of the PSCF input parameters are given elsewhere [70].

2.2.8. Tukey’s Test and Paired t-Test

The daily mean concentrations of PM_2.5 and PM_2.5-bound trace elements between Seoul and Seosan were compared using one-way analysis of variance (ANOVA) with Tukey’s test and paired t-tests [71,72]. These tests were carried out to determine whether there were any significant differences in the mean PM_2.5 and PM_2.5-bound trace element concentrations between the Seoul and Seosan sites. The paired-sample t-test’s statistical significance level was set at 5%.

3. Results and Discussion

3.1. Overview of Trace Elements and PM_2.5 in Urban and Rural Areas

During the present study, in Seoul, S had the highest median concentration of all trace elements in PM_2.5 (1446.9 ng m⁻³), followed by K (251.1 ng m⁻³) and Si (237.5 ng m⁻³); in Seosan, S also had the highest median concentration (1265.2 ng m⁻³), followed by Si (306.1 ng m⁻³) and K (294.5 ng m⁻³) (Table 1). For all elements, the mean concentrations were higher than the median concentrations, indicating the impact of episodic high-concentration events (Table 1). A similar variance in mean and median concentrations has been reported in Bhopal, India, possibly as a result of the impact of episodic high-concentration events [24]. Furthermore, PM_2.5 concentrations were higher than 35 µg m⁻³ during episodic events, owing to high concentrations of S, K, and crustal elements at both study sites (Figure S2). In Seoul, the mean concentrations for Ni and S varied between 0.5 ng m⁻³ and 1446.9 ng m⁻³, while those for V and S in Seosan varied between 0.51 ng m⁻³ and 1265.2 ng m⁻³. S and crustal elements (Si, Ca, and Fe) were the major components detected in both Seoul and Seosan, which is consistent with the findings of previous studies conducted at these locations during winter [19,39]. The mean S concentration was significantly higher in Seoul than in Seosan, and the opposite trend was true for the concentrations of crustal elements (Si, Fe, Ca, and Ti). The paired t-test revealed that the differences between Seoul and Seosan based on S, K, Ca, and Fe were statistically significant (Tukey’s test and paired t-test, p < 0.05; Table S4). In Seoul, high S concentrations were primarily emitted from burning coal [73]. The higher crustal element concentrations in Seosan compared with those in Seoul may have been due to the characteristics and amounts of resuspended dust. K concentrations were higher in Seosan than in Seoul, which could be attributed to wood burning for cooking and agricultural waste burning to prepare fields for the next harvest.

Anthropogenic elements, such as Zn, Pb, Mn, Cu, and As, were prevalent at both sites. The concentrations of Pb, Cu, V, Ni, Se, and Zn at both sites were similar. Zn and Cu are tracers of brake and tire wear [74]. The primary sources of atmospheric Pb in previous studies have been identified as vehicular and ship emissions, heavy oil combustion, transboundary coal combustion emissions, waste incineration, and recirculation of historic leaded gasoline [74]. Moreover, Fe, Pb, and Zn emissions in China in 2013 were determined to have originated from five main emission sectors: industrial, residential sources, transportation, power generation, and windblown dust [75].

Trace elemental ratios are reliable tracers of pollutants emitted to the atmosphere by coal combustion, oil combustion, waste burning, and industrial emissions [74,76,77]. As tracers for coal and oil combustion, respectively, the V/Pb and V/Ni ratios can be used [78]. In general, coals have a high V content, and a V/Pb ratio greater than 1 suggests coal combustion emissions, while a lower V/Pb ratio implies vehicle exhaust emissions. The waste and emissions from non-coal-burning industries have a V/Pb ratio of <<1. The ratios of V/Pb were found to be 0.03 ± 0.03 in Seoul and 0.03 ± 0.04 in Seosan. The V/Pb ratio was <<1, indicating that these elements may be emitted as a result of industrial activity or the burning of waste. The impurities V and Ni have been widely used as tracers of petroleum combustion, since they are always present in heavy oil [79]. The V/Ni ratio in heavy oil combustion is between 3 and 4, the ratio in ship emissions is between 2.5 and 5, and increased Ni emissions from oil burning result in lower ratios [74,80,81,82]. The V/Ni ratio was found to be 1.33 ± 2.08 in Seoul and 1.04 ± 1.17 in Seosan, suggesting that these elements may be emitted during oil combustion. The V/Ni ratios of above 0.7 suggested that these elements were emitted as a result of oil burning [82].

During the entire monitoring period, the daily mean concentration of PM_2.5 in Seosan (31.2 ± 13.3 µgm⁻³) was higher than that in Seoul (23.0 ± 12.4 µgm⁻³) [42] (Text S1). During the study period, the daily mean values of temperature (T), relative humidity (RH), and wind speed (WS) were −3.5 ± 6.3 °C, 57 ± 16%, and 2.4 ± 1.1 m s⁻¹ in Seoul, respectively, and −2.6 ± 5.6, 70 ± 17.1%, and 1.8 ± 1.7 m s⁻¹ in Seosan, respectively. In comparison to the Seoul site, Seosan had a higher relative humidity and a lower wind speed. The higher PM_2.5 concentrations in Seoul compared to Seosan were mostly attributable to the lower temperature, lower wind speed, and higher relative humidity during the measurement period [42]. According to previous studies, high PM_2.5 concentrations result from its accumulation under high RH, low WS, and low T [42,83,84,85].

3.2. Comparison of Variation in the Trace Elements of PM_2.5 in Urban and Rural Areas

The diurnal variation in PM_2.5-bound trace elements is important because it suggests information about the short-term air quality, sources, and formation characteristics of these components on an hourly basis [26]. Figure 1 displays the diurnal variations in PM_2.5 and its trace elements in Seoul (dashed black line) throughout the entire measurement period. The main peak in hourly PM_2.5 concentration in Seoul occurred between 9:00 and 14:00 (Figure 1). S displayed a high concentration between 5:00 and 13:00, and a second peak between 15:00 and 20:00. In Seoul, no apparent diurnal variations in K concentrations were observed. The Si, Fe, Ca, and Ti concentrations peaked between 7:00 and 14:00, which may be associated with crustal sources. Because of potential similarities in their sources, Cu, Pb, Zn, and Mn also showed concentration peaks between 7:00 and 14:00. The Br concentration peaked around 10:00 and 19:00 at the Seoul site.

Figure 1 shows the diurnal variation in PM_2.5 and its elements in Seosan (red solid line). A bimodal distribution of hourly PM_2.5 concentrations was observed, with a major peak between 20:00 and 22:00, and a comparable peak between 8:00 and 13:00 (Figure 1). S showed a high concentration peak between 8:00 and 14:00, followed by modest peaks at 16:00–18:00 and 20:00–23:00. K demonstrated an intense concentration peak between 8:00 and 14:00. Si, Ca, Fe, and Ti all showed high peaks between 9:00 and 14:00, which could be attributed to the similar sources of these elements. Zn, Cu, Pb, and Mn all showed major concentration peaks between similar hours of 9:00 and 14:00, which may also be indicative of similar origins. The Br concentration peaked around 10:00 and 22:00 at the Seosan site.

In general, almost all measured elements’ concentrations increased in the morning and decreased in the afternoon in both the rural and urban areas (Figure 1). The morning peak could be attributed to vehicle emissions during traffic rush hours [86]. The peak concentrations of Cu, Zn, Mn, and Pb might be related to tire and brake-lining wear [87,88,89]. Source apportionment studies of Korea have suggested that Br and Pb may be emitted from vehicular and/or burning (e.g., biomass, waste material, coal, oil) activities [90,91]. However, the nighttime peak may have been due to decreases in T, WS, and planetary boundary layer height (PBLH), along with an increase in relative humidity (RH). At night, the lower PBLH tends to limit the dispersion of PM [92]. The PBLH was typically high during the day, which is suitable for pollution dispersion; thus, PM_2.5 and most elements displayed lower concentrations during the afternoon [92,93]. Furthermore, S had higher hourly concentrations in Seoul than in Seosan; however, the remaining trace elements had higher concentrations in Seosan than in Seoul. A more intense morning peak in K concentration was observed in Seosan than in Seoul, possibly due to agricultural burning activities in the rural area [43,94]. At both sites, Cr, As, Ni, V, and Se displayed clear peaks throughout the day, possibly owing to industrial emissions [95].

3.3. Origins of Trace Elements in PM_2.5

Principal component analysis (PCA) was used to identify the sources of PM_2.5-bound trace elements at the Seoul and Seosan sites (

Table 2. Varimax-rotated principal factor (F) loadings for PM_2.5-bound trace elements at the Seoul and Seosan sites during the winter period.

	Seoul			Seosan
	F1	F2	F3	F1	F2	F3	F4
Si	0.01	0.98	−0.09	0.98	0.03	0.12	0.02
S	0.87	0.17	0.20	−0.11	0.88	0.25	0.07
K	0.55	0.68	0.34	0.58	0.64	0.31	0.10
Ca	0.03	0.98	−0.09	0.94	0.06	0.12	−0.09
Ti	0.09	0.98	−0.03	0.97	0.04	0.16	0.03
V	0.31	0.91	−0.08	0.83	0.29	0.02	0.23
Cr	0.88	0.36	0.17	0.46	0.52	0.44	−0.20
Mn	0.72	0.56	0.22	0.14	0.20	0.86	0.04
Fe	0.41	0.90	0.00	0.89	0.23	0.37	−0.03
Ni	0.81	0.31	0.16	0.50	0.81	0.05	−0.02
Cu	0.85	0.28	0.03	0.30	0.17	0.77	−0.05
Zn	0.62	−0.01	0.01	0.11	0.29	0.90	0.08
As	0.02	−0.13	0.95	0.05	−0.18	0.00	0.93
Se	0.87	0.27	0.05	0.13	0.88	0.21	−0.17
Br	0.95	0.14	−0.07	0.12	0.87	0.31	−0.01
Ba	0.70	−0.25	0.14	No data	No data	No data	No data
Pb	0.27	−0.01	0.93	0.04	0.33	0.61	0.59
Variance	9.05	3.76	1.69	7.94	3.02	1.78	1.16
Percentage of variance (%)	53.2	22.1	9.95	49.6	18.8	11.1	7.26
Cumulative (%)	53.2	75.3	85.3	49.6	68.5	79.6	86.8

F = factor.

, Text S2). PCA is widely used in studies to identify the sources of various aerosol components [10,44,96]. The factor loadings (F1 to F4) were determined using a varimax rotation for each element [97]. The number of factors was determined by an eigenvalue of one or more [97]. In this study, the three factors identified in Seoul and four in Seosan are described as follows:

Three factors contributed 85.3% of the total explained variation in the Seoul data. The first factor was anthropogenic activity (e.g., coal burning, industrial, biomass burning, and vehicular activities), with high S, Mn, Ni, Cu, Se, Br, Zn, and Ba loadings. The second factor was related to soil dust resuspension, with a high loading of crustal elements (Si, Ca, Ti, and Fe). The third factor revealed the presence of Pb and As sources, which included metal smelting, crustal sources, coal combustion, and burning of municipal waste [98,99,100].

Four factors contributed approximately 86.8% of the total variance in the Seosan data (Table 2). The first factor was the resuspension of soil dust, which had high loadings of crustal elements (Si, Ca, Ti, and Fe) (Table 2). The second factor had high loadings of S, K, Ni, Se, and Br, which were determined to be associated with coal and biomass burning. The third factor, characterized by high Cr, Mn, Cu, and Zn loadings, was related to traffic and industry. The fourth factor in Seosan was the high loadings of As and Pb. According to the analysis, K was mainly associated with crustal sources in Seoul and combustion sources in Seosan. As previously reported, V and Ni are well-known tracers for oil combustion sources [101]. However, in the PCA results, Ni and V were separate factors. Thus, source apportionment tools, such as positive matrix factorization (PMF), may be required to gain a better understanding of the sources of Ni and V in PM_2.5 sources in Seoul and Seosan.

Emissions from anthropogenic sources must be drastically reduced to the greatest extent practicable in comparison to emissions from natural sources. More important information for risk management can be obtained around the sampling site from a thorough assessment of the risk levels of metals from anthropogenic sources than from that of metals from natural sources (Figure 2). If the calculated values of EF for a certain element are close to unity, it means that crustal dust is the predominant source [67,102]. On the other hand, EFs greater than 10 suggest a considerable contribution from anthropogenic sources [67,74,103]. At both the Seoul and Seosan sites, S, Cu, Zn, As, Se, Br, and Pb were identified as anthropogenic elements with EF values greater than 10 (Figure 2). In a previous study conducted in Seoul, Si, Ti, and Fe were suggested as having crustal origins, while S, Zn, and Pb were suggested as anthropogenic elements using EF calculations [39].

To explore the possible source locations of PM_2.5 and its elements, the potential source contribution function (PSCF) was used (Figure 3a,b and Text S2). Text S2 in the Supplementary Materials provides information about backward air-mass trajectories and their use for PSCF. The results indicate that Southern China (i.e., Shandong, Anhui, Jiangsu, Henan, and Hebei provinces) was the likely source of PM_2.5 arriving at Seoul (Figure 3a), whereas that arriving at Seosan mainly originated from Jiangsu, Henan, and Anhui. Shandong, Anhui, Jiangsu, Henan, and Hebei were probable sulfur source locations for both sites. Industry in China has become more concentrated and agglomerated along the coast in the east, north, and south; Shandong, Anhui, Jiangsu, Henan, and Hebei are notable examples [104,105]. For K, the potential source locations for Seoul included Shandong, Jiangsu, and Hebei, whereas those for Seosan were the Jiangsu, Anhui, and Shanghai metropolitan areas. According to a previous study, the eastern provinces of China (including Shandong, Jiangsu, Hebei, Anhui, and Shanghai) are well known for their high PM_2.5 concentrations and biomass burning throughout the winter [106,107]. The fires were visible during the study period in the eastern regions of China (Figure S3). For soil, potential source regions for Seoul included Shandong, Jiangsu, and Hebei, whereas those for Seosan included Inner Mongolia and the coasts of Jiangsu, Anhui, and Zhejiang provinces.

3.4. Estimated dry Deposition Fluxes of Elements in PM_2.5 in Urban and Rural Areas

Elements—especially metals that are toxic and deposited on soil and plants—can harm human health and the environment due to their entry and accumulation in food chains [53,108]. The atmospheric deposition of pollutants, which is known as dry deposition, can be measured as DDF and is a vital method for removing contaminants from the air [53,55,109].

Table 3. Average dry deposition flux (μg m⁻² d⁻¹) for PM_2.5-bound elements in Seoul and Seosan during the winter period.

Element	Dry Deposition Velocity (cm s−1)	Mean ± SD (Seoul)	Mean ± SD (Seosan)
Si	2.0	411.6 ± 689.9	530.5 ± 801.8
S	1.0	1253.8 ± 654.0	1096.4 ± 451.6
K	2.0	435.2 ± 169.5	510.4 ± 194.8
Ca	2.0	144.3 ± 172.6	175.2 ± 219.8
Ti	2.0	18.6 ± 16.1	20.0 ± 18.7
V	1.0	0.5 ± 0.3	0.4 ± 0.4
Cr	0.5	1.0 ± 0.6	1.0 ± 0.6
Mn	1.0	11.4 ± 5.2	13.7 ± 10.4
Fe	2.0	319.9 ± 208.3	363.2 ± 242.3
Ni	0.5	0.5 ± 0.4	0.6 ± 0.4
Cu	0.5	1.9 ± 1.3	1.8 ± 0.9
Zn	0.5	21.3 ± 11.9	21.9 ± 9.6
As	0.5	2.0 ± 0.8	1.7 ± 0.7
Se	0.5	0.4 ± 0.4	0.6 ± 0.4
Br	0.5	3.2 ± 1.8	3.8 ± 1.5
Ba	0.5	0.7 ± 0.4	0.0 ± 0.0
Pb	0.5	9.1 ± 2.7	9.1 ± 2.5

summarizes the average DDF for each element of PM_2.5 at the urban and rural sites. The average DDF values were within similar ranges at the two sites: from 0.5 ± 0.4 (Ni and V) to 1253.8 ± 654.0 μg m⁻² d⁻¹ (S) in Seoul, and from 0.3 ± 0.4 (V) to 1096.4 ± 451.6 μg m⁻² d⁻¹ (S) in Seosan (Table 3). The DDF values of S and primarily crustal elements (Si, Ca, and Fe) were higher than those of primarily anthropogenic elements (including, but not limited to, Mn, Zn, Cu, and Pb) for all samples at these study sites (Table 3).

The DDF values of crustal elements (such as Si, Ca, Fe, and Ti) were higher than those of anthropogenic elements (such as Zn, Pb, Mn, Cu, and As) at both locations, because crustal elements were generally found in bare soils and contributed to fine PM mass via resuspension during the study period [24,110,111]. According to prior research, even though heavy trace metals form a small fraction of particle mass and have lower DDFs than crustal elements, they are hazardous to both adults and children [12,112]. The standard deviations were comparable to or slightly greater than the mean DDF values, indicating significant daily variability in DDF (Table 3).

The similar dry deposition velocities at the Seoul and Seosan sites demonstrated that the DDF was controlled primarily by element concentrations, and the DDFs of pollutants were affected by their concentrations. Thus, the measured fluxes and concentrations should theoretically exhibit strong correlations [113,114]. In theory, particles with minimal changes in their DDF should have a strong association between their concentration and flux, because changes in concentration mainly affect flux [51,110]. Additionally, the RH was higher in Seosan (mean 70%) than in Seoul (57%), which might have played a role in the higher DDF for major elements in Seosan than for those in Seoul (Table 3). Furthermore, high RH results in increased particle size via hygroscopic growth, which can significantly increase the particle deposition rate [51].

3.5. Health Risk Assessments

3.5.1. Exposure Assessments for Trace Elements

The values for CDI, DAD, and EC correspond to ingestion, dermal contact, and inhalation, respectively, as shown in Figure S4. During the study period, adults and children in Seoul and Seosan were mainly exposed to trace elements through ingestion. The potential exposure of adults and children at both sites to S, K, and Si was higher via all three routes than for the remaining elements. The S and K emissions in Seoul were associated with many different sources, including coal burning, industrial activities, biomass burning, and vehicular activity (Table 2, PCA). In contrast, in Seosan, S and K were mainly associated with coal combustion and biomass burning (Table 2, PCA). At both locations, exposure to Si was primarily due to crustal sources.

3.5.2. Non-Carcinogenic Health Risks

The non-carcinogenic potential health risks due to elements (Pb, Cr, Cu, V, Mn, Ni, Zn, and As) through various exposure pathways (i.e., dermal contact, ingestion, and inhalation) at the Seoul and Seosan sites were estimated via the HQ (

Table 4. Hazard quotient (HQ) and hazard index (HI) of components in PM_2.5 via ingestion (HQing), dermal (HQder), and inhalation (HQinh) for adults and children in rural and urban areas in South Korea during the winter.

Element	Adult			Children
	HQing	HQder	HQinh	HQing	HQder	HQinh
Seoul
Cr	2.8 × 10−07	3.7 × 10−05	1.3 × 10−13	2.6 × 10−06	1.5 × 10−04	3.1 × 10−13
Mn	2.0 × 10−07	8.3 × 10−06	3.0 × 10−12	1.9 × 10−06	3.3 × 10−05	6.9 × 10−12
Ni	1.7 × 10−08	1.1 × 10−07	7.7 × 10−17	1.6 × 10−07	4.2 × 10−07	1.8 × 10−16
Cu	7.8 × 10−08	1.7 × 10−08	3.6 × 10−16	7.3 × 10−07	6.8 × 10−08	8.4 × 10−16
V	6.1 × 10−08	1.6 × 10−05	2.8 × 10−16	5.7 × 10−07	6.2 × 10−05	6.6 × 10−16
Zn	1.2 × 10−07	3.8 × 10−08	5.3 × 10−16	1.1 × 10−06	1.5 × 10−07	1.2 × 10−15
As	1.1 × 10−05	5.3 × 10−06	5.0 × 10−14	1.0 × 10−04	2.1 × 10−05	1.2 × 10−13
Pb	4.2 × 10−06	1.9 × 10−05	1.9 × 10−14	3.9 × 10−05	7.4 × 10−05	4.5 × 10−14
HQ	1.6 × 10−05	8.5 × 10−05	3.2 × 10−12	1.5 × 10−04	3.3 × 10−04	7.4 × 10−12
HI = ΣHQ	1.0 × 10−04			4.8 × 10−04
Seosan
Cr	2.8 × 10−07	3.8 × 10−05	1.4 × 10−13	2.7 × 10−06	1.5 × 10−04	3.2 × 10−13
Mn	2.4 × 10−07	1.0 × 10−05	3.6 × 10−12	2.3 × 10−06	4.0 × 10−05	8.3 × 10−12
Ni	2.5 × 10−08	1.5 × 10−07	1.1 × 10−16	2.3 × 10−07	6.0 × 10−07	2.6 × 10−16
Cu	7.2 × 10−08	1.6 × 10−08	3.3 × 10−16	6.8 × 10−07	6.3 × 10−08	7.7 × 10−16
V	5.1 × 10−08	1.3 × 10−05	2.4 × 10−16	4.8 × 10−07	5.2 × 10−05	5.5 × 10−16
Zn	1.2 × 10−07	3.9 × 10−08	5.5 × 10−16	1.1 × 10−06	1.6 × 10−07	1.3 × 10−15
As	9.0 × 10−06	4.4 × 10−06	4.1 × 10−14	8.4 × 10−05	1.7 × 10−05	9.6 × 10−14
Pb	4.2 × 10−06	1.9 × 10−05	1.9 × 10−14	3.9 × 10−05	7.4 × 10−05	4.5 × 10−14
HQ	1.4 × 10−05	8.4 × 10−05	3.8 × 10−12	1.3 × 10−04	3.3 × 10−04	8.8 × 10−12
HI = ΣHQ	9.8 × 10−05			4.6 × 10−04

). In Seoul, the HQ values for adults were 1.6 × 10⁻⁵, 8.5 × 10⁻⁵, and 3.2 × 10⁻¹², while those for children were 1.5 × 10⁻⁴, 3.3 × 10⁻⁴, and 7.4 × 10⁻¹², through ingestion, dermal, and inhalation contact, respectively. In Seosan, the HQ values for adults were 1.4 × 10⁻⁵, 8.4 × 10⁻⁵, and 3.8 × 10⁻¹², while those for children were 1.3 × 10⁻⁴, 3.3 × 10⁻⁴, and 8.8 × 10⁻¹², through ingestion, dermal contact, and inhalation, respectively. At both sites, the HQ values associated with As, Cr, Pb, Ni Cu, Zn, and Mn were primarily attributed to dermal contact in adults (8.46 × 10⁻⁵) and children (3.35 × 10⁻⁴). Moreover, at both sites, the dermal exposure to Cr had the highest estimated HQ values for both adults (3.67 × 10⁻⁵) and children (1.45 × 10⁻⁴) (Table 4). The estimated HQ values were higher for children than for adults at both locations for all exposure routes (Table 4).

The estimated HI was higher in Seoul than in Seosan for adults and children (Table 4). In Seoul and Seosan, the HI was higher for children (4.8 × 10⁻⁴ and 4.6 × 10⁻⁴, respectively) than for adults (1.0 × 10⁻⁴ and 9.8 × 10⁻⁵, respectively). One-way ANOVA with Tukey’s test and paired t-tests was used to assess the statistical differences in HI between Seoul and Seosan (Tables S4 and S5). The HIs of the elements at the Seoul and Seosan sites did not reveal any significant differences (p > 0.05). Furthermore, the exposures to PM_2.5-bound trace elements were not significantly different between Seoul and Seosan (p > 0.05). The HI was within the permissible limit (HI < 1) at both sites, indicating no potential non-carcinogenic health risks to humans. According to a previous study conducted in the Chinese province of Shandong, the health risks from As and Zn primarily occurred via dermal contact and ingestion [60]. Adults were more likely to be harmed by dermal contact, whereas children were more likely to be harmed by ingestion [115].

3.5.3. Carcinogenic Risk

In Seoul and Seosan, the potential CRs of As, Ni, Cr, and Pb were estimated for inhalation, ingestion, and dermal contact (Figure 4). In Seoul, for both adults and children, the CR values were higher for ingestion (3.2 × 10⁻⁹ and 7.5 × 10⁻⁹, respectively) than for dermal contact (1.5 × 10⁻⁹ and 1.4 × 10⁻⁹, respectively) and inhalation (4.2 × 10⁻¹⁸ and 2.4 × 10⁻¹⁸, respectively). At Seosan, for both adults and children, the CR values were also higher for ingestion (3.0 × 10⁻⁹ and 6.9 × 10⁻⁹, respectively) than for dermal contact (1.47 × 10⁻⁹ and 1.46 × 10⁻⁹, respectively) and inhalation (5.0 × 10⁻¹⁸ and 2.9 × 10⁻¹⁸, respectively). Furthermore, throughout the research period, the CR values for Cr, Ni, As, and Pb via all exposure routes were higher in Seoul than in Seosan. In Seoul, the highest CR value was for As ingestion by children (CR, 3.9 × 10⁻⁹), whereas the highest CR value in Seosan was for Pb ingestion by children (CR, 3.32 × 10⁻⁹). The CR values for Cr, As, Ni, and Pb in PM_2.5 at both sites were less than 1 × 10⁻⁶ for adults and children for all exposure routes; therefore, the risk of developing cancer from these toxic elements would be considered acceptable under most regulatory programs (Figure 4). According to a study conducted in rural and urban areas of northern Zhejiang Province, China, PM_2.5-bound elements posed significant non-CRs and CRs to the general public [10].

4. Conclusions

In this study, the trace constituents of PM_2.5 in Seoul and Seosan, South Korea, were measured during winter. The most abundant elements at the study sites were S and crustal elements (Si, Ca, Fe, and Ti). Additionally, the morning element concentrations were higher than the afternoon and evening element concentrations, which could be attributed to high traffic volumes and favorable weather conditions (lower T, lower WS, and higher RH). PM_2.5 concentrations were higher than 35 µg m⁻³ during episodic events, primarily due to high concentrations of S, K, and crustal elements at both study sites. Most elements exhibited spatial variation in their average concentration. For example, S was higher in Seoul, whereas crustal elements were higher in Seosan. The DDFs of the crustal elements (Si, Fe, and Ca) were greater than the DDFs of the primarily anthropogenic elements (Cu, Ni, Pb, Zn, Cr) in both Seoul and Seosan. During the winter, children were exposed to more elements by ingestion, dermal contact, and inhalation than adults in Seoul and Seosan. The health risk assessment for element exposure suggested no significant carcinogenic or non-carcinogenic effects in children and adults for either of the study sites. Compared with other elements, As and Pb were found to have higher CR values for adults and children at both study sites. In Seoul, higher CR levels were mostly linked to anthropogenic sources (traffic emissions, coal combustion, and biomass burning), whereas in Seosan they were linked to coal and biomass burning. We also suggest that future scientists and policymakers use these trace element concentration and health risk estimation results with caution, because they are specific to Seoul and Seosan.