Identification of microplastics and their removal efficiency in drinking water treatment plants in tropical areas: a case study of the Dago Pakar drinking water treatment plant, Bandung, Indonesia Open Access

This study investigates microplastic (MP) concentrations, characteristics, and removal efficiency in different treatment units of drinking water treatment plants (DWTPs) at different seasons and sampling times. The result shows that MP levels were higher during the rainy season compared to the dry season. Based on statistical analysis, there are significant differences and correlations between the number of MPs in the water and the seasonal variations of sampling times. Furthermore, even though the number of MPs was slightly higher in the morning compared to the other times, there were no significant differences or correlations between the MPs and the sampling times. The MP size was dominated by the size of 300–1,000 μm (88 and 96% for rainy and dry seasons, respectively) while fibre and transparent are the dominant shape and colours of MPs. The dominant polymer types found were polyester and polyethylene terephthalate. The average removal efficiencies of MPs by Dago Pakar DWTP in the rainy and dry seasons were 58.9 and 39.3%, respectively. Based on these results, it can be seen that the Dago Pakar DWTP has not significantly removed the MPs contained in the raw water source thus a post-treatment unit is needed.

Microplastics (MPs) are plastic particles less than 5 mm in size and larger than 300 μm (Andrady 2011). MPs are widespread in the environment and have been found in varying concentrations in marine water, wastewater, fresh water, food, air, and bottled and tap water (Novotna et al. 2019). MPs can be divided into categories. The first is primary MPs which include any plastic fragments that are already 5 mm or less in size before entering the environment. Secondary MPs arise from the degradation or breakdown of larger plastic products through natural weathering processes after entering the environment such as river water. These include microfibre from clothes, microbeads and plastic polymers (Biao et al. 2024). MPs can have harmful effects on organisms and human health through biomagnification as they are high-level actors in the food chain (Andrady 2011).

Currently, a few studies have been conducted to investigate the MP abundance and removal in DWTPs. Pivokonsky et al. (2018) reported that MPs were present in three different DWTPs supplied from various kinds of water bodies. The abundance of MPs in these raw water and treated water were about 1,473 ± 34 MP/L to 3,605 ± 497 MP/L and 443 ± 10 MP/L to 628 ± 28 MP/L, respectively. With relatively conventional technologies, i.e. coagulation, flocculation, sand and or GAC filtration, and ozonation the removal efficiency of DWTP is about 70–82%. Mintenig et al. (2019) found that even in the low numbers (from 0 to 7 particles per m³), MP is detected in drinking water from groundwater sources which is probably caused by abrasives of plastic materials used during drinking water purification and transport. Wang et al. (2020) studied the characteristics of MPs contained in raw water and effluent resulting from each treatment unit in an advanced drinking water treatment plant (DWTP). It was reported that the numbers of MPs in the raw water and treated water is about 6,614 ± 1,132 and 930 ± 71 MP/L (86% of removal efficiency). The study itself aims to develop the knowledge on removal behaviour of MPs in DWTPs and it is of significance in controlling MPs' pollution in DWTPs. Another study by Pivokonsky et al. (2020) reported that MPs were present in all water samples taken from both the raw water and treatment units DWTP. MPs were found in all water samples and their average abundance ranged from 1,473 ± 34 to 3,605 ± 497 particles/L in raw water and from 338 ± 76 to 628 ± 28 particles/L in treated water, depending on the WTP. Zhang et al. (2020) and Skaf et al. (2020), evaluate the micron- and nano-sized MP removal efficiency during drinking water treatment. Other research related to the abundance and removal of MPs in the water resource and DWTP also has been reported by several researchers (Sarkar et al. 2021; Xue et al. 2021; Sharifi & Attar 2022).

Seasonal changes can significantly affect the concentration of MPs in an environment. Studies often demonstrate that MPs are more abundant during dry seasons compared to wet seasons. This discrepancy is primarily due to variations in water flow and sediment dynamics, which can resuspend and transport MP particles. As a result, the number of MPs can vary throughout the year, influenced by factors such as temperature, rainfall, and river flow, all of which play a key role in their distribution.

MP has also become an important environmental and health issue in Indonesia, as indicated by the increasing number of MP-related publications in this country (Isfarin et al. 2024). MP distribution in surface water such as in a river (Alam et al. 2019; Sembiring et al. 2020) and reservoirs (Ramadan & Sembiring 2020) in Indonesia have been reported by previous researchers. Sembiring et al. (2021a, 2021b) evaluated the performance of rapid sand filter single media and cloth filter to remove MPs presented in water. Radityaningrum et al. (2021, 2023) investigated the presence of MPs in the Surabaya River which is used as raw water for water supply and their removal efficiency on treatment units of selected DWTP. Therefore, many water sources and unit treatments of DWTP may contain a considerable amount of MPs, suggesting a critical need to investigate their fate and transport during drinking water treatment.

The Cikapundung River runs through the city of Bandung, located in West Java, Indonesia. It originates from Lembang in the northern part of the city and flows southward, where it meets the Citarum River. Despite being one of the main sources of water supply for Bandung, the river is classified as heavily polluted. The primary source of this pollution is domestic waste, including human waste and detergents, which has increased as more people move to live along the riverbanks. Additional pollutants come from industrial activities, agriculture, and farming. Currently, there has been a proliferation of residential properties along the riverbank within the city boundary. There are some scattered settlements of farmers within the catchment area of this river, especially in the farmlands, as well as some villas. These settlements are likely the primary sources of plastic waste disposed of in the river, aside from the waste brought into the water by wind or runoff. This plastic waste may degrade into small particles, becoming a source of MPs that can contaminate the Cikapundung River. The presence and occurrence of MPs in the Cikapundung River have not been investigated so far and it becomes important since the water is used as the source of raw water for Dago Pakar DWTP. Dago Pakar DWTP is situated in the northern urban area of Bandung City, Indonesia. This plant uses raw water from the Cikapundung River and employs pre-sedimentation, aselator (coagulation–flocculation–sedimentation), filtration, and disinfection units in its water treatment process. The coagulation–flocculation process is conducted hydraulically using waterfalls and baffles. Liquid coagulants, specifically poly-aluminium chloride (PAC), are applied at a constant dosage of 10–15 parts per million (ppm) throughout the year. The media used in the RSF consist of anthracite and silica sand. Chlorine gas is utilized as the disinfectant and is injected into the water after the RSF, resulting in a residual chlorine level between 0.8 and 1 milligram per litre (mg/L).

This study aims to investigate the concentrations, characteristics, and removal efficiency of MPs in different treatment units of Dago Pakar DWTP, which is supplied with fresh water from the Cikapundung River. To identify the concentrations of MPs in raw water and water undergoing treatment, we analysed their abundance in both rainy and dry seasons at each stage of the DWTPs. The characterization of MP was conducted by measuring the size, shape, colour, and type of the polymers. Moreover, removal efficiency was calculated to determine which treatment unit had the greatest impact on MP removal and to evaluate DWTP performance as well. Statistical analysis was used to determine if there were significant differences and correlations between MP abundance and the sampling times. The benefit of this study is to find out whether the configuration of drinking water treatment processes can significantly reduce MPs.

Water samples were collected from the intake and several units of the DTWP, situated in the northern urban area of Bandung City, Indonesia. This plant utilizes raw water from the Cikapundung River and employs pre-sedimentation, aselator (coagulation–flocculation–sedimentation), filtration, and disinfection units in its water treatment process.

The size of MPs in this study is defined as between 300 and 5,000 μm (NOAA 2016). To separate MPs from the water sample, 500 mL of the sample was filtered using Whatman GF/C paper (Glass microfibre filter 1.2 μm) following the method outlined in a previous study (Barrows et al. 2018). The glass bottle container was then washed with deionized water to ensure no MPs were left behind. Before use, the washed water was tested with deionized water to ensure it was free of MPs.

After filtering, the Whatman filter paper was transferred to a petri dish and then dried in an oven at 60 °C for 24 h. The extracted MPs were observed using a binocular microscope (Olympus CX-21) with a magnification of 10–40 times to aid in their identification, counting, and measuring their size, shape, and colour. The procedures for MP identification followed the technical guidelines and provisions from a previous study (Crawford & Quinn 2016).

The suspected MPs were analyzed using attenuated total reflectance – Fourier transform infrared spectrophotometry (ATR – FTIR) with an ALPHA II – Platinum FTIR Spectrometer equipped with Platinum Diamond-ATR by Bruker. The per cent transmission was recorded in the range of 500–4,000 cm⁻¹ with a resolution of 1 cm⁻¹ and 20 scan numbers (Sarkar et al. 2021). The MPs were identified by comparing their characteristic bands in the FTIR spectrum with the reference spectrum of the plastic polymer (Wang et al. 2017).

The efficiency of MP removal was determined by comparing the number of MPs removed in the DWTP unit. The percentage efficiency was calculated by subtracting the outlet MPs from the inlet MPs, then dividing by the inlet and expressing the result as a percentage. The percentage removal efficiency can be calculated using the equation: %R = [(C_in – C_out)/C_in] × 100%, where C_in is the number of MPs in the inlet and C_out is the number of MPs in the outlet.

The study presented the abundance, size, shape, and colour of MPs as the number of MPs per unit volume of collected water (L). Data analysis was conducted using Microsoft Excel. The analysis of MP abundance and characteristics focused on variations in season (rainy and dry seasons) and sampling time (morning, noon, and afternoon). The Mann–Whitney U test and the Kruskall–Wallis test were used to determine if there were significant differences in MP concentration based on different sampling times and seasons. Additionally, the Spearman correlation test was performed to identify any correlation between MP abundance and sampling time and season.

In Figure 2(a), during dry season 1, it was observed that the number of MPs was slightly higher at the effluent of the sedimentation unit compared to the number at the coagulation inlet. The water velocity at the coagulation slots is high enough for MPs to breakdown due to hydraulic shear. This high turbulence at the coagulation slots could be a contributing factor. A similar result was also observed at the filtration outlet. There are several possible causes for these observations, including the accumulation of MPs in units, incomplete removal of particles during the draining period, and potential MP contamination from the outdoor environment since the water treatment plant is located outdoors. It's also possible that the fine sizes of MPs are being broken up, leading to an increase in the number of MPs due to the shearing force of water flow in porous media (Horton et al. 2017). A similar result was also observed by Sarkar et al. (2021), who argued that the increase in MPs after filtration was due to screening effects.

A statistical analysis was conducted to determine the significance and correlation between the abundance of MPs in the water and seasonal variation. The results of the Mann–Whitney U test show significant differences in MP abundance between the rainy and dry seasons, with a p-value of 0.000 (p-value (0.000) ≤ p-value (0.05)). Additionally, the Spearman Test analysis indicated a significant correlation between the abundance of MPs in the water and the seasonal variations of sampling times, with a p-value of 0.000 (p-value (0.000) ≤ p-value (0.05)).

In Figure 2(b), the abundance of MPs in the sampling sites was measured at different times: morning, noon, and afternoon. The number of MPs was slightly higher in the morning compared to the other times, possibly due to peak domestic activities such as washing clothes. However, there were no significant differences or correlations between the numbers of MPs and the sampling times. This was confirmed by statistical analysis (Kruskall–Wallis value: χ² = 0.098, df = 2, p-value = 0.952, and Spearman p-value = 0.770), indicating that the time of day did not have a significant impact on the abundance of MPs in the water samples.

In this study, the characteristics of MPs were identified based on their size, shape, colour, and type of polymers. The analysis aims to validate that the observed particles are indeed MP particles, using visual criteria as demonstrated in a previous study (Horton et al. 2017).

In Figure 3(b), it is evident that during the dry season, the distribution of MP sizes follows a similar pattern to that of the rainy season, with the 300–1,000 μm size range being the most prevalent. However, unlike the rainy season, the percentage number of this size range decreases as the level of treatment in the DWTP increases. The average percentage number of MP particles at different stages of treatment in the DWTP are as follows: intake (96%), inlet of coagulation–flocculation (95%), outlet of sedimentation (90%), and outlet filtration (85%). Additionally, a small fraction of MPs ranging from 1,001–2,000 and 2,001–5,000 μm were also observed in the DWTP (1–8% and 2–6%, respectively). Notably, during the dry season, MP particles in the 1,001–2,000 μm and 2,001–5,000 μm sizes were found to be less abundant compared to the rainy season, possibly due to increased sun exposure causing fragmentation into smaller sizes. Cole et al. (2011) and Xue et al. (2021) have reported that the abundance of MPs in freshwater varies in size and number due to dynamic aquatic conditions and diverse human activities. Furthermore, the size of MPs found in freshwater is influenced by various factors, including their source and formation process (Cole et al. 2011). MPs originating from primary sources tend to be larger than those from secondary sources. MPs from secondary sources undergo fragmentation into smaller sizes due to aquatic conditions such as waves, currents, and sunlight.

During the rainy season, the percentage of fibre-type MP particles at each sampling location, including the inlet of intake, inlet of aselator (coagulation–flocculation–sedimentation), outlet of sedimentation, and outlet of filtration units, ranged from 95 to 99%. In the dry season, the percentage of fibre-type particles fluctuated between 91 and 100%. Textiles are identified as a significant source of fibre-shaped MPs. Synthetic fibres from textiles can detach during production, use, and disposal, making textiles a major contributor of MPs in the environment. The process of washing textiles, in particular, releases a significant amount of synthetic fibres into the environment. This occurs when the fibres detach, fragment, and are carried away by the wastewater into rivers (Sarkar et al. 2021).

MPs may escape from the formed floc instead of being removed by settling. MPs float on the surface of the water because of their low density. Meanwhile, the slightly increased amounts of MPs in the filtration unit may be caused by a backwash process that does not completely remove micro-sized particles. The high velocity of backwash water during backwash can break up the fibre-type shaped MPs into smaller sizes, therefore increasing the number of MP particles. Some of the samples might be taken during the ripening period of filter running, apart from screening effects that breakdown the MPs into smaller sizes. The ripening period is the process of conditioning granular media shortly after backwash, which can take some minutes to almost an hour. During the ripening period, the filter waste, including MPs trapped in the media porous, comes out through the filtration outlet and causes the water not to meet the standard quality. Similar results were also reported by Sarkar et al. (2021), who stated that the higher number of MPs in the filtration outlet was due to screening effects.

The percentage of fragment-type shaped MPs varied from 1 to 7% in both rainy and dry seasons. Meanwhile, film-type shaped MPs were only found in the coagulation inlet at 2% and in small amounts (1%) in the sedimentation outlet during the rainy and dry seasons, respectively. These film-type particles can be completely removed by coagulation–flocculation–sedimentation units. In the rainy season (Figure 4(a)), the fragment-type shape was not completely removed from the DWTP, likely due to the coagulation–flocculation process being ineffective in entrapping MFs in the floc formation. In the dry season (Figure 4(b)), the complete removal of fragment-type shaped MPs was probably due to their wider specific surface area or electrostatic properties, making them fit better with the coagulant being used at the time compared to fibre-type shaped MPs, thus making it easier to form a floc and precipitate. When the floc is not well-formed, the removal efficiency of the unit is reduced. The formation of floc due to the coagulation–flocculation process depends on several factors. The important factors are the electrostatic characteristics of colloids or particles, pH, type of coagulant, and design factor (Crittenden et al. 2012).

Based on the data, it can be concluded that in both the rainy and dry seasons, the dominant shape of MPs found is fibre-shaped. A study by Radityaningrum et al. (2021) also reported that fibre-shaped MPs were predominantly found in both raw water (94.1%) and treated water (81.5%).

The results of this study were compared with research conducted by Radityaningrum et al. (2021). It was stated that the most common MP colours found in raw water were black (52.7%) and blue (25.7%). In treated water, blue (33.8%) and black (27.7%) were the dominant colours. The slight variations in results may be attributed to different sources of pollutants and environmental conditions of water sources. Particles that are colourless or transparent are believed to come from polypropylene polymers, while white particles are derived from polyethylene, and opaque (non-translucent) particles come from LDPE (Hidalgo-Ruz et al. 2012). Colour can also be a useful indicator for determining the source of MPs (Castro et al. 2016) since it reflects the original materials from which the MPs are derived.

Several samples with different colours, shapes, and sampling sites were also tested and compared with the reference spectrum of the polymer, as shown in Figures 6(b)–6(f). Based on the spectrum comparison, despite a lot of noise, the peak similarity indicates that the tested samples closely match the polyester spectrum.

In this test, not all types of MPs and not all known types of polymers were tested due to limitations of the standard spectrum. Out of the several MPs that were tested, 212 out of 235 particles (90.2%) were confirmed to be plastic polymers. However, the remaining 23 particles couldn't be identified as plastic polymers due to the limitations of the standard spectrum. Therefore, the analysis was conducted using MP concentration data obtained through quantification based on categories developed by Horton et al. (2017). The summary of sampling locations, MP types, and colours with polymer types is presented in Table 1.

The MP removal efficiency at each unit process was determined to evaluate the performance of DWTP Dago Pakar (Table 2). During the rainy season, the aselator unit exhibited the highest MP (MPs) removal efficiency at 31.9%, surpassing other unit processes such as the pre-sedimentation unit at 18.9% and the filtration unit at 7.8%. Consequently, DWTP Dago Pakar achieved a total MP removal efficiency of 58.9%. On the other hand, during the dry season, the filtration unit demonstrated the highest MP removal efficiency at 31.9%, outperforming the pre-sedimentation and aselator unit. In the dry season, the total MP removal efficiency was 39.3%. The differences between rainy and dry seasons were likely due to the dominant size and abundance of the MPs. During the dry season, the abundance of MPs was lower compared to the rainy season, hence a lower percentage of removal is considered normal under these conditions (Kankanige & Babel 2021).

The MP removal efficiency of DWTP Dago Pakar was found to be lower than what was reported in previous studies (Wang et al. 2020; Kankanige & Babel 2021; Radityaningrum et al. 2021). Wang et al. (2020) reported that the removal efficiency of MPs at conventional water treatment plants ranged from 58.9 to 70.5%. Similarly, Kankanige & Babel (2021) conducted a study at a DWTP in Thailand and found that the removal efficiency during the dry and rainy seasons was 67.7 and 57.2%, respectively. Moreover, Radityaningrum et al. (2021) stated that total MP removal efficiencies in DWTP Surabaya I and II, which had similar characteristics to DWTP Dago Pakar, were 54 and 76%, respectively.

The DWTP Dago Pakar has lower efficiency compared to other WTPs, possibly due to differences in the size distribution of MPs and polymer type, as well as the coagulant added. In the case of DWTP Dago Pakar, the smaller size of MPs dominates, and the optimal coagulant dosage is not carefully adjusted for varying conditions of raw water quality. Additionally, the low efficiency in removing MPs at DWTP Dago Pakar may be attributed to suboptimal operation and maintenance processes. Existing conditions indicate that there is no fine/bar screen in the intake section, allowing fine particles/waste to be carried to the next unit. Likely, MP particles are also transported to the next unit. Furthermore, after evaluating the design, it was found that the actual detention time of the pre-sedimentation unit in DWTP Dago Pakar (5 min) is shorter than the design criteria (0.5–3 h). With such a low detention time, it is likely that the particle deposition process, including MP particles, will not be optimally operated. In the coagulation unit, the coagulant is only added at one point using a plastic tube, resulting in uneven distribution of the coagulant. As a result, the aluminium sulphate coagulant, which should be able to optimally remove MPs, is not achieving the desired effect. Additionally, the evaluation result of the detention time in the flocculation unit was shorter, which was 12 min, than the design criteria. This condition makes irregular velocity gradient changes, causing turbulence and floc will break. If the formed floc breaks, the MP particles will be difficult to deposit and tend to float because the MP has a low density. Meanwhile, in the sedimentation and filtration unit, the automatic opening and closing of the sludge drain valve and the automatic filter washing are not working properly. The cleaning process of these two units is not optimally operated and can cause small particles to remain, such as MPs. The detention time in the flocculation unit was found to be shorter than the design criteria, at 12 min. This resulted in irregular velocity gradient changes, causing turbulence and potential breakage of the floc. If the floc breaks, it will be difficult to deposit MP particles, which tend to float due to their low density. Additionally, the automatic opening and closing of the sludge drain valve and the automatic filter washing in the sedimentation and filtration unit are not functioning properly. As a result, the cleaning process of these units is not operating optimally, potentially leading to the retention of small particles, including MPs.

In this study, the concentrations, characteristics, and removal efficiency of MPs in different treatment units of Dago Pakar Drinking Water Treatment were investigated at different seasons and sampling times. The result shows that MP levels were higher during the rainy compared to the dry season (rainy season: 20.6, 16.7, 11.5, and 11.7 MPs/L, and dry season: 6.7, 6.2, 6.4, and 4.4 MPs/L for intake, influent aselator, effluent sedimentation, and effluent filtration, respectively). Based on statistical analysis there are significant differences and correlations between the abundance of MPs in the water and the seasonal variations of sampling times, with a p-value of 0.000 (p-value (0.000) ≤ p-value (0.05)). Moreover, even though the number of MPs was slightly higher in the morning compared to the other times (noon and afternoon), there were no significant differences and correlations between the numbers of MPs and the sampling times (Kruskall–Wallis value: χ² = 0.098, df = 2, p-value = 0.952, and Spearman p-value = 0.770). The MP size was dominated by the size of 300–1,000 μm (88 and 96% for rainy and dry seasons, respectively) while fibre and transparent are the dominant shape and colours of MPs. The dominant polymer types found were polyester and PET. During the rainy season, the aselator unit exhibited the highest MP (MPs) removal efficiency at 31.9%, while during the dry season, the filtration unit demonstrated the highest MPs removal efficiency at 31.9%, outperforming the aselator unit. The average removal efficiencies of MPs by Dago Pakar DWTP in the rainy and dry seasons were 58.9 and 39.3%, respectively. Based on these results, it can be seen that the Dago Pakar DWTP has not significantly removed the MPs contained in the raw water source. The addition of an ultrafiltration membrane as a post-treatment can be an alternative to removing remaining MPs besides protecting the quality of water resource from pollution. Moreover, further research should focus on identifying water physicochemical parameters that correlate with MP abundance in DWTPs to provide useful information on the environmental impacts caused.

This research was supported by Riset Peningkatan Kapasitas of ITB.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.