Application of enhanced coagulation in drinking water treatment from the degraded catchment to reduce NOM: a major DBP precursor Open Access

Ensuring access to clean drinking water is a cornerstone of public health and crucial for achieving the UN's sustainable development goal six by 2030. Surface water serves as a primary source of potable water for many communities worldwide (Scanlon et al. 2023), which is processed through drinking water treatment plants (DWTPs). These treatment plants employ various methods, including adsorption, membrane filtration, ion exchange, advanced oxidation processes, and biological processes (Sillanpää et al. 2018; Elma et al. 2022).

However, in several regions, especially in developing nations, traditional treatment systems rely on simpler methods such as coagulation, sedimentation, sand filtration, and chlorination due to their cost-effectiveness (Pandit & Kumar 2019; Treacy 2019). The efficiency of these processes, particularly coagulation, is pivotal for subsequent treatment stages and for reducing health hazards during disinfection (Ghernaout 2020b; Wang et al. 2021; Knap-Bałdyga & Żubrowska-Sudoł 2023).

The efficacy of coagulation, however, faces challenges, notably in dealing with changes in natural organic matter (NOM) concentration and composition. These challenges are exacerbated by factors such as climate change, seasonal variations, and human activities (Qiu et al. 2019; Hadadi et al. 2022; Anderson et al. 2023). For example, in Ethiopia, interventions like extensive agriculture in highland regions have significantly altered surface water quality, leading to increased soil erosion and NOM levels (Sishu et al. 2021; Derseh et al. 2022; Sishu et al. 2024). Especially, since the 1960s, the land uses have been mainly changed to agricultural and increased the share of cultivated land from 39% to the current 77% (Zeleke 2000; Zeleke & Hurni 2001; Tamene et al. 2006). The cultivated land at sloped triggering soil erosion resulted in degradation. The studies showed that on average, soil loss is estimated to be 25 ̶37 Mg ha⁻¹ yr⁻¹ within a rainy period of 3–4 months (Lemann et al. 2016; Zimale et al. 2016; Haregeweyn et al. 2017).

Understanding the composition and characteristics of NOM is essential for devising effective treatment strategies. NOM compounds vary in polarity, acidity, and molecular mass, influencing their behavior during treatment (Loganathan et al. 2020; Ghernaout 2020a). NOM can be broadly classified into two main groups based on their polarity: hydrophilic and hydrophobic components. The hydrophilic fractions of NOM primarily comprise nitrogenous compounds, aliphatic carbon, and low molecular weight, while hydrophobic NOM mostly consists of humic substances and high molecular weight (Sharp et al. 2006). The shift in NOM composition in surface water is anticipated to be a direct result of soil exposure to solar radiation and the degradation of soil humic substances (Oni et al. 2012; Slavik et al. 2023). Hydrophobic components are most prevalent for the formation of chlorinated disinfection byproducts (CDBPs), mainly trihalomethanes (THMs), but hydrophilic components present unique challenges in removal and contribute to the formation of halogenated DBPs during and after chlorination (Ma et al. 2013).

Several types of coagulants, spanning aluminum-based, ferric-based, polymeric, and bio-coagulants, have been utilized to tackle the removal of NOM. Conventional water treatment, employing the traditional coagulant aluminum sulfate, typically achieves the removal of 60–70% of hydrophobic fractions and 30–40% of hydrophilic fractions (Elma et al. 2022). Nonetheless, conventional methods often struggle to maintain consistent removal rates, particularly in areas like the Ethiopian highlands, which are characterized by extreme seasonal fluctuations in NOM levels. Hybrid coagulation, which amalgamates metallic coagulants with polymers such as polydiallyldimethylammonium chloride (polyDADMAC), has demonstrated advantages over the conventional use of aluminum sulfate (Sillanpää et al. 2018; Ghernaout 2020a). However, the inclusion of polyDADMAC during water treatment may contribute to the formation of N-nitroso dimethylamine, a nitrogenous DBP (An et al. 2019).

Enhanced coagulation emerges as a promising solution, optimizing coagulant dosage and pH control to achieve higher NOM removal rates and mitigate the formation of chlorinated DBPs (USEPA 1999; Ghernaout 2020a; Wang et al. 2021). Despite its proven effectiveness, traditional coagulation methods persist in many Ethiopian water treatment processes.

This research endeavors to characterize NOM in surface water sources within degraded catchments in the Ethiopian highlands and enhance NOM removal through enhanced coagulation processes. The Angereb reservoir, supplying water to Gonder, Ethiopia, serves as a case study, drawing from a catchment impacted by soil degradation and agricultural activities. By improving understanding and implementing enhanced coagulation, this study aims to inform and potentially revise DWTP-operating conditions to ensure a safer drinking water supply.

The landscape of the watershed is characterized by a series of hills rising to 2,870 meters above sea level, interspersed with ridges and valleys. The region experiences a moderate mean annual temperature ranging from 17 to 25 °C, accompanied by a variable annual rainfall of 2,077 mm (Mengesha et al. 2013). Land usage data retrieved from ESA WorldCover 10m 2020 v100 (Zanaga et al. 2022) reveal that agriculture dominates the land cover, occupying a substantial 80% of the watershed. Additionally, plantation forests, grasslands, and shrublands collectively cover 18% of the area, with the reservoir itself contributing to less than 1% of the total catchment.

Between March and August 2023, we gathered raw water samples at the DWTP inlet and treated water samples at the outlet. To capture seasonal changes, we took samples during both wet and dry periods. Employing a grap method, we collected these samples using 1 L amber glass bottles. Immediately after collection, we dispatched them to the Abay Basin Environmental Laboratory, where they were stored in a refrigerator at 4 °C, shielded from light to mitigate chemical reactions and biological processes until analysis.

Turbidity, color, water temperature, conductivity, dissolved total solids, pH, and residual chlorine were assessed onsite, while other parameters underwent laboratory analysis. Alkalinity, pH, turbidity, true color, UV@254, and total organic carbon (TOC) were analyzed in the laboratory as per standard procedures. Water pH was directly measured using a pH meter from a Vante 900p multi-parameter kit in the field. The color of the water was measured by using a spectrophotometer and compared against the platinum cobalt standard. Turbidity, total suspended solids (TSS), alkalinity, total hardness, aluminum, iron, ammonia, and sulfate levels were determined with a Wagtech 8000M spectrophotometer (Plaintext, UK), following the methods outlined by Rice et al. (2012). UV@254 was measured at 254 nm using 1 cm quartz cells and a UV/VIS carry 60 (Agilent, USA). The UV@254 absorbance was utilized to compute specific ultraviolet absorbance (SUVA) as per Equation (1). TOC in both untreated and Jar test-treated waters was quantified using the persulfate wet digestion method with a DR 2400 spectrophotometer (Hach). This method aided in determining the NOM removed through various coagulation techniques.

Enhanced coagulation experiments followed the methodology outlined in USEPA (1999), juxtaposed with the standard coagulation process. Both aluminum sulfate and ferric chloride coagulants were employed. The efficacy of enhanced coagulation was evaluated based on its capacity to mitigate THM precursors and UV@254, rather than the conventional focus on turbidity. Consequently, the dosage of coagulants in enhanced coagulation tests was contingent upon the removal targets for TOC, UV@254, turbidity, color, and SUVA, as stipulated in the USEPA (1999) guidelines. Dosages spanned from 20 to 140 mg L⁻¹.

Table 1 provides a summary of the untreated water results for TOC, UV@254, SUVA, and various physicochemical parameters, showing seasonal fluctuations in their values.

Typically falling within the alkaline range, pH levels ranged from 7.3 to 7.8, with higher values observed during dry periods. Importantly, these pH levels were within the safe drinking range as recommended by USEPA (2022) and WHO (2017). The pH of untreated raw water plays a pivotal role in the dosage of coagulating chemicals (Naceradska et al. 2019; Zhao et al. 2022).

During wet periods, recorded turbidity and TSS values were notably higher by up to 47 times compared to dry periods. This significant seasonal variation has implications for the coagulation process and dosing strategies. Similar observations have been made in studies of tropical highland lakes receiving water drainage from degraded catchments (Womber et al. 2021).

During the wet season, TOC exhibited a 30% increase compared to the dry season. Moreover, the UV absorbance at 254 nm of raw water during the wet period was notably elevated. SUVA values remained below 2 L/mg ⁻¹ m⁻¹, indicating lower aromaticity and suggesting a higher hydrophilic nature (Liu et al. 2020). Nitrate nitrogen (N-NO₃) levels were higher during the wet season but remained below the WHO drinking water limit of 10 mg L⁻¹ N-NO₃ in both seasons (WHO 2017). Ammonia nitrogen (N-NH₄), a potential freshwater pollutant, was undetectable in the dry season, likely due to the high alkalinity of the water, which inhibits the presence of the ionic (NH₄⁺) form of ammonia that typically occurs in acidic to neutral water (Reddy et al. 1984). This finding aligns with previous observations in Lake Tana (Sishu et al. 2024). Conversely, iron (Fe) concentration was slightly higher in the dry season, indicating the influence of redox processes. During dry periods, the reduction of ferrous substances from bed sediment likely contributes to increased concentrations, surpassing the WHO guideline limit (WHO 2017).

Table 2 provides results of physicochemical water quality indicators obtained after treatment at the Angereb DWTP, encompassing parameters such as TOC, UV@254, SUVA removal efficiency, and others. Notably, the removal efficiency was influenced by the seasonal variability of these parameters.

During dry periods, the pH of untreated water decreased from 7.8 to 7.1 post-treatment, whereas in wet seasons, the decline was more pronounced from 7.3 to 6.4. The treatment process involves prechlorination followed by the addition of Al₂(SO₄)₃·18H₂O and polyDADMAC as chemical coagulants. This resulted in a reduction in Fe concentration, aligning it well below the WHO guideline limit (WHO 2017).

The treatment process demonstrated significant efficacy in reducing the turbidity of the raw water shown in Table 1 from 25 to 1.5 NTU, as seen in Table 2, which was a 94% reduction in dry seasons and 99% in wet seasons. Similarly, TSS were substantially removed at comparable rates to turbidity, indicating a primary focus on turbidity removal in the traditional treatment process (Sun et al. 2019).

However, the process exhibited lower efficiency in terms of reducing CDBP precursors. The TOC removal rate was only 28%, and UV absorbance @ 254 removal ranged from 14 to 31%. These removal rates fall short of the conventional water treatment process, which typically achieves 30–40% hydrophilic TOC removal (Elma et al. 2022). Therefore, further efforts in TOC removal are imperative to mitigate the risk of DBP formation.

The pH stands out as a pivotal factor affecting the efficacy of the coagulation process. In the refined coagulation approach, pH optimization can be achieved through two methods: (1) adjusting with an acid or base before fine-tuning the coagulant dosage or (2) directly applying coagulant within a dosage range spanning from several milligrams per liter to hundreds of milligrams per liter (Wang et al. 2021). In this particular investigation, the latter approach was adopted.

Upon the addition of 140 mg L⁻¹ of ferric chloride, the pH decreased from 7.3 to 5.1 in the wet season, while the same dose of alum caused a drop from 7.3 to 5.4. Furthermore, in the dry period, the pH of the water declined from 7.8 to 6.2 after ferric chloride addition and to 6.5 after alum addition (see Figure 3(a)). Notably, during the dry period, the raw water exhibited higher alkalinity compared to the wet season, rendering it more resilient to pH alterations.

The optimal pH range for achieving maximum TOC removal with alum was determined to be between 6.0 and 6.7. Typically, aluminum-based coagulants exhibit their best coagulation performance when the pH is near the point of minimal coagulant solubility (Saxena et al. 2018). At lower pH levels, hydrolysis of alum salt generates positively charged coagulant species (Al(OH)₂⁺), while higher pH levels prompt the formation of large hydroxide precipitates, facilitating internal particle bridging and sweep flocculation (Yu et al. 2015). For ferric chloride, the highest TOC removal efficiency was observed at pH levels below 5.2.

Jar test experiments were utilized to optimize the dosage of alum and ferric chloride coagulants. Results indicated that increasing coagulant dosages enhanced TOC removal, as illustrated in Figure 3(c). During the wet season, the maximum TOC removal achieved for alum was 42% at a dosage of 80 mg L⁻¹, while during the dry season, it reached 34%. Conversely, ferric chloride exhibited a maximum TOC removal of 50% at a dosage of 100 mg L⁻¹ in the wet season and 40% at 120 mg L⁻¹ in the dry season. Ferric chloride demonstrated efficient coagulation at low pH levels (3–5) (Sun et al. 2019).

Comparative analysis revealed that ferric chloride could coagulate a larger percentage of TOC compared to alum. This suggests that a portion of organic matter is more amenable to coagulation by ferric hydroxide floc rather than aluminum hydroxide floc. Ferric hydroxide flocs possess a greater number of active adsorption sites in contrast to aluminum hydroxide flocs (Jarvis et al. 2006). This disparity likely accounts for the higher TOC removal observed with ferric chloride utilization (Duan & Gregory 2003).

During the dry season, the UV₂₅₄ absorbance removal in water reached approximately 20%, almost reaching full removal at a coagulant dose of 60 mg L⁻¹ for both alum and ferric chloride. Conversely, in the wet season, utilizing the same dose, the UV₂₅₄ removal percentage surged to around 50%. These findings align with prior research, indicating that chemical coagulation effectively eliminates aromatic compounds from TOC at moderate coagulant doses (Beauchamp et al. 2020).

Comparing seasons, the wet season demonstrates a higher UV₂₅₄ removal rate. This could be attributed to sediment transport, which carries humic substances, thereby increasing TOC concentrations during heavy rainfall. The enhanced coagulation improves the removal of TOC and UV₂₅₄ which are used as surrogates for the organic DBP precursors, resulting in improved removal of precursors for THMs and other DBPs. Figure 3(b) illustrates this relationship, depicting TOC as a function of chlorine required.

The production of halogenated DBPs resulting from the chlorination of drinking water is directly proportional to the aromatic carbon content of the organic constituents in the water (Singer 1999). Increasing the NOM level increases the level of DBP precursors, which is a major cause of DBP formation (Singer 1999; Barrett et al. 2000; Hrudey 2008; Mazhar et al. 2020).

The higher NOM level also resulted in high chlorine consumption as seen in Figure 3(b). The chlorine demand and TOC levels were significantly correlated both in the dry and wet seasons under alum and ferric-utilized coagulation processes. The observed r values were near 0.97, as seen in Table 3. The higher NOM level associated with high chlorine demand might lead to the formation of chlorinated DBPs (Xie 2003; Uyak & Toroz 2007).

Under both dry and wet season conditions, optimal turbidity and color removal were achieved with a coagulant dose of 20 mg L⁻¹ for both alum and ferric chloride, as detailed in the accompanying Table 4. While additional coagulant dosage resulted in the detection of aluminum and iron ions in treated water, levels remained within WHO guideline limits (WHO 2017).

In this study, a first attempt was made to apply the enhanced coagulation process to potentially reduce NOM as a major DBP precursor in drinking surface water drained from a degraded catchment in the Ethiopian highlands. Our findings show that the humic content of the soil is predominantly characterized by hydrophilic NOM, as shown by the value below SUVA 2 L/mg ⁻¹ m⁻¹. Therefore, we evaluated the application of enhanced coagulation for hydrophilic NOM removal by measuring parameters including TOC and UVA@254. Both TOC and UVA@254 were removed up to 50% with an enhanced coagulation process using ferric chloride. In contrast, the existing conventional treatment process that involves prechlorination followed by aluminum sulfate coagulation removed less than half of the enhancement. Nonetheless, further research is necessary to evaluate the economic feasibility of transitioning from the current coagulation process to enhanced coagulation.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.