Evaluation of Sampling Device
To evaluate the device recovery, polymers under size 20 µm were used. The synthetic polymer species used were PP, PS, PET, PVC, and HDPE. Recovery was measured by the difference in weight of the STS filter in the device before and after filtration. The recovery rates of synthetic polymers used in the study are presented in Table 1. The average recovery rate was 94.083%. The recovery rate based on the weight difference may not be accurate because of factors such as moisture, temperature, and vibration. Therefore, thermal extraction desorption-gas chromatography–mass spectrometry (TED-GC–MS) was conducted for accurate weight measurement. Thermal analysis showed a recovery rate of 87.015% (Table 1). This result has an error range of around 10% compared to the weight analysis method, and it could be confirmed that the device used in this study showed a relatively high recovery rate.
Table 1
Recovery efficiency of sampling device using manufactured polymers of size < 20 µm assessed by weighing the filter and by thermal extraction desorption-gas chromatography-mass spectrometry (TED-GC-MS)
Species of microplastics (a) | Recovery (%) |
PP | 97.287 |
PS | 94.190 |
PET | 97.341 |
PVC | 86.539 |
HDPE | 95.060 |
Average recovery of sampling device | 94.083 |
Recovery using TED-GC–MS (PS) | 84.269 |
(a) PP, polypropylene; PS, polystyrene; PET, polyethylene terephthalate; PVC, polyvinyl chloride; HDPE, high-density polyethylene
Table 2 shows the results of blank sample analysis by µ-Raman spectroscopy. The blank samples were prepared before filtering the environmental samples using 4 L HPLC-grade water. The results of blank samples showed the presence of 0.25–4.25 particles/L. This result indicates concentration of microplastics which caused device and cross-contamination in filtration process. Through blank sample analysis, it is possible to minimize the error of analysis results. Manta trawl and plankton net have been used for research on microplastics in aqueous systems. Although these devices are useful for separating particles of a specific size range, they have components made of plastics. Plastic fragments generated from these devices during sampling can cause cross-contamination, which affects the results of microplastic analysis. The blank sample analysis manta trawl and plankton net showed 1.85–31 microplastic particles (Bikker et al., 2020; Mu et al., 2019a; Mu et al., 2019b). Through comparing blank sample results, it was confirmed that the device manufactured in this study had relatively little cross-contamination than other devices. Further, these net-based sampling devices are suitable for studying microplastics of sizes over 100 µm in aqueous media surfaces but are not suitable for those under 20 µm, which are targeted in this study. Therefore, manufacturing a new sampling device with minimal cross-contamination while selectively separating 20 µm plastic was necessary. The results of recovery rate and the blank analysis showed that the manufactured device was highly efficient and had low cross-contamination.
Table 2
Microplastic concentrations in blank samples (HPLC-grade water) after filtering through manufactured sampling device
Species of microplastics | Concentration (particles/L) |
P-Lake | N-River |
PP | 0.25 | 4.25 |
PS | 0.00 | 0.50 |
PET | 0.50 | 0.00 |
PVC | 0.25 | 0.00 |
PE | 1.00 | 3.00 |
Size Of The Microplastics
The microplastics were divided by their sizes as follows: <5 µm, 5–20 µm, 20–50 µm, and > 50 µm. The concentrations of microplastics over 50 µm size was 7.70 particles/L and 8.37 particles/L in P-Lake and N-River, respectively (Fig. 3). The proportion of over 50 µm size to total results was 3.53% and 2.62%, respectively, in P-Lake and N-River. The ratio is the smallest, and it can be seen as the effect of screening using 50 µm pore size sieves in the sample collection process. The concentrations of microplastics of size 20–50 µm were 43.14 particles/L and 73.81 particles/L in P-Lake and N-River, respectively. The concentrations of microplastics of size 5–20 µm, which were targeted in this study, were 133.61 particles/L and 202.07 particles/L in P-Lake and N-River, respectively. The concentrations of microplastics of size < 5 µm were 33.33 particles/L and 33.24 particles/L in P-Lake and N-River, respectively. The proportion of microplastics of size 5–20 µm was 61.3% and 63.3% of the total microplastics particles in P-Lake and N-River, respectively. The proportion of 5–20 µm size microplastics was highest among the total microplastics. The efficiency of multiple stage size classification of the manufactured device was high for sampling in aqueous media, particularly fresh water, for isolation of microplastics. However, the proportion other than the target size accounts for more than 30%. Microplastics have various shapes such as circular, fiber, and fragments. In particular, in the case of the fiber shape, the size of the cross-sectional area and the overall size may be different. As a result, some losses may occur in the filtration process. Therefore, it is assumed that some fiber-type microplastics not be screened or size-classified.
Concentration Of Total Microplastics
Figure 4 shows the concentration of total particles and microplastics. The concentrations of both total particles and microplastics were higher in N-River than those in P-Lake. There may be a correlation between the total particle concentration and the total microplastics concentration. Therefore, in this study, a factor Rmp was used to indicate the proportion of microplastic particles among total particles.
$${R}_{mp} = \frac{particles of microplastics}{Total particles}$$
The Rmp in P-Lake and N-River are 1.6×10− 2 and 1.1×10− 2, respectively. In other words, P-Lake has a relatively higher microplastic concentration compared to the total particle than N-River. Several studies have shown that hydrological characteristics such as river type, flow rate, water depth, algae, and discharge amount affect microplastics (Mani et al. 2020; Yan et al.2021; Zhang et al. 2020). In addition, morphological characteristics such as dams and dam reservoirs are also considered to have an effect (Wu et al. 2020; Corcoran et al. 2019). As shown in Table 3, the flow rate of the N-River was 62.945 m3/s, which was higher than that of the P-Lake (0.102 m3/s). In addition, the precipitation from November 1 to the sampling date was slightly larger in N-River. As a result, the flow rate and precipitation of the N-River were relatively greater than that of the P-Lake, which affected the concentration of particulate substances, that is, the SS concentration (Table 3). Furthermore, it can be seen that it affects the concentration of microplastics. However, the total depth of P-Lake is 21 m, which is deeper than the N-River (8 m). In addition, the P-Lake area is a dam which artificially build for water supply. Construction structures such as dams and reservoirs affect the behavior and movement of microplastics (Sarkar et al. 2021; Liro et al. 2020).
In addition, artificial structures act as a factor that improves the sedimentation condition of microplastics (Di et al. 2018; Kumar et al. 2021; Watkins et al. 2019). The P-Lake area which is artificially constructed and has deeper water depth has higher microplastic sedimentation conditions than the N-River area. As a result, the Rmp value of the P-Lake having a higher precipitation condition might be high. Therefore, the shape, structure, flow rate, precipitation, and depth of the river could affect not only the precipitation state of microplastics but also the concentration. In particular, through the Rmp factor presented in this study, it might be a factor that can indirectly indicate the microplastic precipitation state, flow rate, and the structure at the sampling point.
Table 3
Temperature, suspended solid (SS), flow rate, and precipitation recorded in sampling site
Site | Total depth (m) | Temperature of water (°C) | SS(a) (mg/L) | Flow rate (a) (m3/s) | Precipitation |
Duration (b) (days) | Amount (mm) |
P-Lake | 21.0 | 12.2 | 3.2 | 0.102 | 23 | 28.7 |
N-River | 9.8 | 13.46 | 6.56 | 62.945 | 25 | 39.7 |
(a) The result of SS concentration and flow rate was obtained from the Water Environment Information System of the Ministry of Environment of South Korea. (b) The precipitation measurement duration for sampling sites is from November 1 to the sampling date (P-Lake = November 23, N-River = November 25).
Concentration Of Microplastics By Layer And Species
Figure 5 shows the concentration of microplastics at three layers of depth at the sampling points. The concentrations in the upper layers of P-Lake and N-River were 23.95 particles/L and 182.07 particles/L, respectively. The concentration in the upper layer of N-River was higher than that of P-Lake. In the case of the middle layer, the proportion was similar, with P-Lake at 25% and N-River at 22%. However, the microplastic concentration in the lower layer was 138.38 particles/L at P-Lake, which was higher than that in the N-Lake (65.27 particles/L). In P-Lake, the microplastic concentration in upper layer was lower than that in the lower layer. In N-River, the microplastic concentration in upper layer was higher than that in the lower layer. Similar to the relation between the total particles and SS concentration, the environmental factors could possibly affect the distribution of microplastics by layers. The flow rate in the two sampling sites was the biggest difference (Table 3). The cross-sectional distance of the sampling point is 567 m for P-Lake and 207 m for N-River. The cross-sectional area according to the depth is 11,907 m2 for P-Lake and 2,028 m2 for N-River. Therefore, the velocity is 8.57 × 10− 6 m/s for P-Lake and 31.03 m/s for N-River. It means that the velocity of the P-Lake is very small compared to the N-River, and the number of Reynolds is small. Reynolds number is also a factor that distinguishes laminar flow and turbulence of fluids. In relatively small P-Lakes, laminar flow can be governed, and in N-River, turbulence can be dominated (Kumar et al. 2021; Waldschläger et al. 2019). But the concentration of microplastics will not be affected by only the Reynolds number. This is because very complex factors such as microplastic density, shape, hydrological properties, aggregation, aging, and biological system (Jódar-Reyes et al. 2006; Liu et al. 2020; Miao et al. 2021; Song et al. 2019). However, in the case of P-Lake, where laminar flow prevails, the concentration of lower layer may be relatively high due to the sedimentation of microplastics. Conversely, in the case of N-River, where turbulence is dominant, the concentration of upper layer may be higher than that of lower layer due to inter-layer mixing or convection during the process of microplastics settling.
Concentration Of Microplastics By Species
The distribution of microplastics based on their type and depending on the depths is shown in Fig. 6. Grey color indicates PE, green color is PP, blue is PS, and red is PET. PE and PP were found in the highest proportions in all the layers and sites. The composition of PE was 6–72% and PP was 18–75%. The abundant concentrations of these two plastic species could be because of the quantity of their production. The highest produced plastic groups are PE (36%), PP (21%), and PVC (12%), which are followed by PET, polyurethane, and PS (< 10% each) (Geyer et al. 2017). PE and PP with the highest production ratio are mainly used as materials for packaging.
PET was found only in the lower layers. The concentration of PET in the lake and river was 9.7 particles/L (7%) and 3.0 particles/L (5%), respectively. The density of PE, PP, and PS is approximately 0.90–1.10 g/cm3, while that of PET is about 1.38–1.41 g/cm3. Because the density of PET is relatively higher than that of other plastics, it can be assumed to sink lower and was measured only in the lower layer. Similarly, High-density PE (HDPE) sinking in deeper seawaters has been reported (Dai et al. 2018). The proportion of PS decreased, from 35–4% in the lake and 19–5% in the river, as the depth increased. Aged PS had increased surface roughness and decreased average pore diameter (5.1 ± 0.2 nm) than pristine PS (39.3 ± 0.5 nm). Further, the specific surface area increased from 2.0 ± 0.1 m2/g to 7.9 ± 0.2 m2/g (Fu et al. 2021; Sun et al. 2020). The porosity could influence the buoyancy of PS. Because PS gains buoyancy with aging, its proportion can be assumed to decrease with increase in the depth. PS is mainly used as thermal packaging material by foaming and buoyancy material in fishing. Therefore, because of the buoyancy characteristics of PS, the distribution of PS decreased with increasing depth.