**1. Introduction**

The supply of high-quality freshwater is a crucial problem in rural areas. In many cases, water resources are of inferior quality (i.e., peat water), making it inconsumable without implementing advanced treatments. Peat water is one of the water sources that are still untapped. It is characteristically acidic (pH 5.9) and high in natural organic matter (NOM), identified using three standard parameters of the non-specific indicator: dissolved organic carbon (DOC, 36.40 mg/L), UV absorbance 254 nm (0.955 1/cm), and organic substances (113.76 mg KMnO4/L). NOM in peat water may exert odors, aromatization, biological instability, and corrosion of water distribution networks [1]. Conventional water treatments for removal NOM have been widely applied by standalone processes such

**Citation:** Elma, M.; Pratiwi, A.E.; Rahma, A.; Rampun, E.L.A.; Mahmud, M.; Abdi, C.; Rosadi, R.; Yanto, D.H.Y.; Bilad, M.R. Combination of Coagulation, Adsorption, and Ultrafiltration Processes for Organic Matter Removal from Peat Water. *Sustainability* **2022**, *14*, 370. https:// doi.org/10.3390/su14010370

Academic Editors: Mohd Rafatullah and Masoom Raza Siddiqui

Received: 3 December 2021 Accepted: 27 December 2021 Published: 30 December 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

as coagulation–flocculation and sedimentation [2], activated carbon adsorption [3], and filtration [4]. However, they do not provide optimal treatment for removing NOM.

Previous studies have reported types of water and wastewater treatments that contain high NOM using standalone coagulation, with about 60–70% removal of hydrophobic fraction of NOM and 30–40% of hydrophilic fractions [5–8]. Another work also reported performance of adsorption for NOM removals of up to 98% that were obtained by powdered activated carbon (PAC) with an optimum dosage 500 mg/L [9], which can remove organic materials with a molecular weight (MW) ranging from 0.5–1 to 1–3 kDa. Nevertheless, it could not remove NOM with an MW of <0.5 kDa [10,11]. In addition, the adsorbent may be saturated due to the complete occupation of the adsorption site, while reactivation of the adsorbent results in a complex operation, which may lead performance to decrease [12].

Membrane technology is an advanced treatment process for treating NOM in water, such as wetland or peat water [13–17]. Several studies were reported successful treatment of wetland saline water by pervaporation using silica-based membranes [18–21], wetland saline water by pure silica membrane and organosilica-based membranes [22–28]. Another study showed ultrafiltration (UF) membrane for removal fraction of NOM from peat water [29]. The UF technology is more applicable and better for reducing NOM in water compared with pervaporation. The pervaporation setup is more complex than UF. However, despite NOM's effective removal by the membrane, it is also able to decline membrane performance through membrane fouling [30,31]. Fouling is a major factor that may decrease membrane flux during the separation process, especially the ultrafiltration.

Membrane fouling in peat water treatment is mainly caused by NOM through both the hydrophilic and the hydrophobic fractions [32]. The most common methods to reduce membrane fouling are by altering the physical and chemical properties of the membrane materials by adding additives in the fabrication stage [33–36] or by applying pretreatment of the feed in the operational stages [32,37]. In this study, both coagulation and adsorption were investigated for the first time as a pretreatment of UF for the treatment of real peat water.

This paper reports a preliminary study on NOM removal from peat water by using both aluminum sulfate-based coagulation and PAC-based adsorption as a pretreatment of UF. The polysulfone (PSf) UF membrane was first prepared and characterized. Before being used for the pretreated peat water filtration. The NOM composition in the peat water samples was then characterized. Finally, actual peat water was treated using a series of treatments, namely aluminum sulfate-based coagulation, PAC-based adsorption, and filtration using the developed Psf-UF membrane.

### **2. Materials and Methods**

## *2.1. Peat Water Characterization*

The peat water sample was taken from Banjar Regency, South Kalimantan, Indonesia. Preliminary characterization of peat water included measuring pH using a pH meter (Hanna Hi2211), KMnO4—oxidizable organic substances using the permanganate titrimetric method, and aromatic organic matter absorbance of UV254, and DOC by a total organic carbon analyzer (Shimadzu TOC-L). The permanganate titration method was conducted according to Standard (SNI 06-6989.22-2004). The UV<sup>254</sup> parameter was measured by a UV visible (UV-1600 Spectrophotometer). On the other hand, DOC was analyzed by high-temperature catalytic oxidation with non-dispersive infrared (NIDR) detection. As a pretreatment, the samples were filtered using Whatman 0.45 µm before being tested by TOC analyzer. Meanwhile, specific UV absorbance (SUVA254, L/mg.m) was used to represent TOC normalized aromatic moieties (UV254). Meanwhile, specific UV absorbance (SUVA254, L/mg.m) was used to represent TOC normalized aromatic moieties (UV254) by dividing of the UV<sup>254</sup> with the DOC value.

### *2.2. Membrane Preparation and Characterization*

The dope solution for Psf UF membrane preparation was made using 18 wt.% of Psf (Merck) as the polymer, 64 wt.% dimethylacetamide (DMAc, Merck) as the solvent, and polyethylene glycol (PEG, Merck) PEG 600 as the additive (18 wt.%). According to earlier reports, the membranes were prepared using the phase inversion method [35]. The polymer, solvent, and additive were mixed and stirred until homogeneous. Then, the solution was left idle overnight to release the entrapped bubbles. Subsequently, the dope solution was cast on a glass plate at a wet casting thickness of 165 µm using a casting applicator. The phase inversion was then continued by immersing the cast film into a coagulation bath containing nonsolvent solution comprising of DMAc 35 wt.% and KCl 0.5 wt.% in water.

The hydraulic resistance of the prepared membrane was characterized by measuring the clean water permeability, and a scanning electron microscopy (SEM) was used to determine the membrane PSf morphology and membrane thickness. The pore size of the membranes was determined using image-J software from the surface of the SEM image [38]. The permeability test was conducted by flowing the distilled water on a dead-end system filtration device. The permeate volume was then measured every 5 min intervals for 60 min operation time under different pressures of 1, 1.5, 2, 2.5, and 3 bar.

### *2.3. Coagulation, Adsorption, and Ultrafiltration*

The coagulation tests were done by varying doses of aluminum sulfate (one of the most common coagulants) in a range of 125–250 mg/L using the Jar-test method at adjusting pH 6 (regulated by drop-wise adding 0.1 M NaOH (Merck)) with a working volume of 1.2 L. During the jar test, the coagulant mixture in peat water was stirred at 100 rpm for 1 min, followed by slow stirring at 40 rpm for 20 min and sedimentation for 20 min, according to a protocol reported earlier [39]. The range of the coagulant dosage was defined based on a previous study [31].

The PAC adsorption tests were done using the Jar-test under varying 20–200 mg/L doses. It was carried out for the pretreated peat water through coagulation/flocculation. The feed and PAC (particle size of 100 mesh; surface area of 800 m2/g, Merck) were mixed with a rotary shaker at 180 rpm for 3 h.

After coagulation and adsorption, the treated supernatant underwent a UF—200 mL of the supernatant was filtered using the developed UF PSf membrane by using a standard dead-end filtration cell (Figure 1) according to a protocol detailed elsewhere [30]. The filtrations were done at variable pressures of 1, 1.5, 2, 2.5, 3 bar for 60 min at room temperature of feed (25 ◦C), stirred at 50 rpm. A gas compressor generated the pressure, and the permeate volume was collected every 5 min. All of the coagulatin, adsorption, and filtration experiments were done in triplicate. *Sustainability* **2022**, *13*, x FOR PEER REVIEW 4 of 12

> that the peat water had a neutral pH, similar to previous reports where the pH value on surface water ranged from 5.0–8.1 [31]. The high NOM content was indicated by the DOC values [39,40], the absorbance value of UV254, which are high compared with the results obtained by Kang and Choo [41] and Jeong et al. [42] of UV254 < 0.1 cm−1 for surface water. However, compared with the results obtained elsewhere Herwati, Mahmud, and Abdi [30], Mahmud, Abdi, and Mu'min [31], Saputra [37], Aisyahwalsiah [39] showed the UV254 absorbance of the peat water samples was relatively low. Similar to the UV254 absorbance value, the SUVA254 values of the peat water sample deviated from others (Zularisam et al. [43]). Based on their reports, the SUVA254 characteristic of peat water contained a high hydrophobic fraction. The SUVA254 values of the peat water sample in this study ranged at 2–3 L/mg.m suggesting the mixture of hydrophobic and hydrophilic substances NOM characteristics, with a large range of MWs. Similar results were found in previous research on surface water that reported SUVA254 of < 2 L/mg.m (low hydrophobic character) [41,42].

1 pH 6.3 6.3 6.3 6.3 6.3 0

3 UV254 absorbance 1/cm 0.968 1.005 0.977 0.955 0.976 0.02

5 SUVA254 L/mg.m 2.659 2.761 2.684 2.624 2.682 0.006

The surface and cross-section SEM images of the prepared membrane are shown in Figure 2. The UF PSf membrane had a tight pore arrangement and sponge-like cross-section morphology without macrovoids (large cavities), which was similar to the membrane structure reported earlier [35]. Based on image-J surface SEM image processing (Figure 3),

mg KMnO4/L

carbon) mg/L 36.40 - - 36.40 36.4 -

**Week** 

**Average STDEV I II III IV** 

120.08 126.4 120.08 113.76 120.08 5.16

**Figure 1.** Illustration of the ultrafiltration dead-end experimental setup. **Figure 1.** Illustration of the ultrafiltration dead-end experimental setup.

**Table 1.** Characteristics of the peat water sample.

**No Parameter Units** 

2 DOC (dissolved organic

KMnO4 organic substances

*3.2. Characterisation of UF-PSf Membranes* 

4

**3. Results** 

*3.1. Peat Water Characteristics* 

### **3. Results**

### *3.1. Peat Water Characteristics*

The characterizations of peat water were carried out for four periods to monitor the changes of NOM content of the peat water samples, as summarized in Table 1. It shows that the peat water had a neutral pH, similar to previous reports where the pH value on surface water ranged from 5.0–8.1 [31]. The high NOM content was indicated by the DOC values [39,40], the absorbance value of UV254, which are high compared with the results obtained by Kang and Choo [41] and Jeong et al. [42] of UV<sup>254</sup> < 0.1 cm−<sup>1</sup> for surface water. However, compared with the results obtained elsewhere Herwati, Mahmud, and Abdi [30], Mahmud, Abdi, and Mu'min [31], Saputra [37], Aisyahwalsiah [39] showed the UV<sup>254</sup> absorbance of the peat water samples was relatively low. Similar to the UV<sup>254</sup> absorbance value, the SUVA<sup>254</sup> values of the peat water sample deviated from others (Zularisam et al. [43]). Based on their reports, the SUVA<sup>254</sup> characteristic of peat water contained a high hydrophobic fraction. The SUVA<sup>254</sup> values of the peat water sample in this study ranged at 2–3 L/mg.m suggesting the mixture of hydrophobic and hydrophilic substances NOM characteristics, with a large range of MWs. Similar results were found in previous research on surface water that reported SUVA<sup>254</sup> of < 2 L/mg.m (low hydrophobic character) [41,42].

**Table 1.** Characteristics of the peat water sample.


### *3.2. Characterisation of UF-PSf Membranes*

The surface and cross-section SEM images of the prepared membrane are shown in Figure 2. The UF PSf membrane had a tight pore arrangement and sponge-like cross-section morphology without macrovoids (large cavities), which was similar to the membrane structure reported earlier [35]. Based on image-J surface SEM image processing (Figure 3), the surface pore size of the membrane was 0.061 µm, falling under UF range of 0.001– 0.1 µm [44].

The polysulfone membrane used in this study had a pore size of less than 0.1 µm (Figure 3). However, from the SEM image it is not possible to determine exactly what the pore size is. However, the pore size distribution can determine by utilized Image-J software following the previous research used digital SEM image data. The result of processing SEM images by Image-J can be seen in Figure 3B,C. The results show the average pore diameter of the membrane is 0.061 µm. Based on literature, the polysulfone membrane used in this work can be categorized as well as ultrafiltration membrane.

µm [44].

µm [44].

*Sustainability* **2022**, *13*, x FOR PEER REVIEW 5 of 12

**Figure 2.** Ultrafiltration polysulfone membrane surface microstructure with (**A**) magnification of 1000× and (**B**) magnification 2500×, and cross-section microstructure with (**C**) magnification of 1000× and (**D**) magnification 2500×. **Figure 2.** Ultrafiltration polysulfone membrane surface microstructure with (**A**) magnification of 1000× and (**B**) magnification 2500×, and cross-section microstructure with (**C**) magnification of 1000× and (**D**) magnification 2500×. **Figure 2.** Ultrafiltration polysulfone membrane surface microstructure with (**A**) magnification of 1000× and (**B**) magnification 2500×, and cross-section microstructure with (**C**) magnification of 1000× and (**D**) magnification 2500×.

the surface pore size of the membrane was 0.061 µm, falling under UF range of 0.001–0.1

**Figure 3.** Membrane surface SEM image (2500×) processed with Image-J for pore size determination (**A**) before editing, (**B**) after threshold and (**C**) outline image result. **Figure 3.** Membrane surface SEM image (2500×) processed with Image-J for pore size determination (**A**) before editing, (**B**) after threshold and (**C**) outline image result. **Figure 3.** Membrane surface SEM image (2500×) processed with Image-J for pore size determination (**A**) before editing, (**B**) after threshold and (**C**) outline image result.

#### The polysulfone membrane used in this study had a pore size of less than 0.1 µm The polysulfone membrane used in this study had a pore size of less than 0.1 µm *3.3. Coagulation-Flocculation*

*3.3. Coagulation-Flocculation* 

*3.3. Coagulation-Flocculation* 

the addition of Al2(SO4)3 coagulant [37].

(Figure 3). However, from the SEM image it is not possible to determine exactly what the pore size is. However, the pore size distribution can determine by utilized Image-J software following the previous research used digital SEM image data. The result of processing SEM images by Image-J can be seen in Figure 3B,C. The results show the average pore diameter of the membrane is 0.061 µm. Based on literature, the polysulfone mem-(Figure 3). However, from the SEM image it is not possible to determine exactly what the pore size is. However, the pore size distribution can determine by utilized Image-J software following the previous research used digital SEM image data. The result of processing SEM images by Image-J can be seen in Figure 3B,C. The results show the average Figure 4 shows that for aluminum sulfate dosings of 125–250 mg/L, the removal efficiency of NOM increased from 125 mg/L to 175 mg/L and then a slight decrease until 250 mg/L. Beyond that value, the NOM removal efficiency decreased. The loading restabilization can explain the pattern on the NOM removals as a function of the dosing rate on the addition of Al2(SO4)<sup>3</sup> coagulant [37].

brane used in this work can be categorized as well as ultrafiltration membrane.

Figure 4 shows that for aluminum sulfate dosings of 125–250 mg/L, the removal efficiency of NOM increased from 125 mg/L to 175 mg/L and then a slight decrease until 250

zation can explain the pattern on the NOM removals as a function of the dosing rate on

brane used in this work can be categorized as well as ultrafiltration membrane.

pore diameter of the membrane is 0.061 µm. Based on literature, the polysulfone mem-

Figure 4 shows that for aluminum sulfate dosings of 125–250 mg/L, the removal efficiency of NOM increased from 125 mg/L to 175 mg/L and then a slight decrease until 250

the addition of Al2(SO4)3 coagulant [37].

**Figure 4.** NOM removal rate represented by oxidation with KmnO4 and UV254 absorbances as a function of doses of alum in the coagulation–flocculation test. **Figure 4.** NOM removal rate represented by oxidation with KmnO<sup>4</sup> and UV<sup>254</sup> absorbances as a function of doses of alum in the coagulation–flocculation test.

The oxidation of organic substances by KMnO4 peaked at the optimum dose of 175 mg/L. Meanwhile, a slight decrease in UV254 removals was observed. In the process of coagulation–flocculation, the dominant fraction of removed NOM is the one that hydrophobic or with large MWs as detailed elsewhere [31,45]. In addition, according to Suslova et al. [46] that KmnO4 can oxidize various types of organic components irrespective of the The oxidation of organic substances by KMnO<sup>4</sup> peaked at the optimum dose of 175 mg/L. Meanwhile, a slight decrease in UV<sup>254</sup> removals was observed. In the process of coagulation–flocculation, the dominant fraction of removed NOM is the one that hydrophobic or with large MWs as detailed elsewhere [31,45]. In addition, according to Suslova et al. [46] that KmnO<sup>4</sup> can oxidize various types of organic components irrespective of the MWs.

MWs. The NOM removal in the coagulation-flocculation process was achieved optimum at a dose of 175 mg/L corresponding to organic substances KmnO4 and UV254 absorbance of 77.78% and 75.24%, respectively. The removal of NOM obtained in this study was higher than the previous studies [47,48]. After adding the coagulant, the coagulation rate decreased, as well as the pH value from 6 to 3.65. It was caused by the reaction of aluminum The NOM removal in the coagulation-flocculation process was achieved optimum at a dose of 175 mg/L corresponding to organic substances KmnO<sup>4</sup> and UV<sup>254</sup> absorbance of 77.78% and 75.24%, respectively. The removal of NOM obtained in this study was higher than the previous studies [47,48]. After adding the coagulant, the coagulation rate decreased, as well as the pH value from 6 to 3.65. It was caused by the reaction of aluminum sulfate with water that produces H<sup>+</sup> ions. The acidification of water lowered the coagulation/flocculation efficiencies.

### sulfate with water that produces H+ ions. The acidification of water lowered the coagula-*3.4. Coagulation-Adsorption*

NOM bonding [50].

tion/flocculation efficiencies. *3.4. Coagulation-Adsorption*  The PAC adsorption was carried out after the coagulation/flocculation of the raw peat water sample. Figure 5 shows the rate of NOM removal in the PAC adsorption process. The NOM removal rate was higher than the standalone coagulation-flocculation or adsorption processes reported earlier [32,49]. Increased NOM removal at low pH during the PAC adsorption can be attributed due to the low pH of the solution due to the preced-The PAC adsorption was carried out after the coagulation/flocculation of the raw peat water sample. Figure 5 shows the rate of NOM removal in the PAC adsorption process. The NOM removal rate was higher than the standalone coagulation-flocculation or adsorption processes reported earlier [32,49]. Increased NOM removal at low pH during the PAC adsorption can be attributed due to the low pH of the solution due to the preceding coagulation/flocculation stage [39]. The pH has a significant effect on activated carbon adsorption and the removal efficiency is higher in acidic than in neutral and alkaline conditions. The presence of H<sup>+</sup> ions in solution leads to competition between H<sup>+</sup> ions and NOM bonding [50].

ing coagulation/flocculation stage [39]. The pH has a significant effect on activated carbon adsorption and the removal efficiency is higher in acidic than in neutral and alkaline conditions. The presence of H+ ions in solution leads to competition between H+ ions and

**Figure 5.** The NOM removal rate represented by organic substances of KMnO4 and UV254 absorbances as function of PAC dosages. **Figure 5.** The NOM removal rate represented by organic substances of KMnO<sup>4</sup> and UV<sup>254</sup> absorbances as function of PAC dosages.

The removal rate of KmnO4—oxidizable organic substances in the coagulation–adsorption process is higher than the UV254 (representing aromatic moieties). The results were the opposite of the coagulation/flocculation process, which could be attributed to the NOM content with large MW removed in the coagulation–flocculation process. As such, only the low MW NOM remained in the PAC adsorption. In previous research, the adsorption process could remove NOM hydrophilic fractions with small MWs [9–11]. The removal rate of KmnO4 organic substances was higher than the UV254 absorbance. These The removal rate of KmnO4—oxidizable organic substances in the coagulation–adsorption process is higher than the UV<sup>254</sup> (representing aromatic moieties). The results were the opposite of the coagulation/flocculation process, which could be attributed to the NOM content with large MW removed in the coagulation–flocculation process. As such, only the low MW NOM remained in the PAC adsorption. In previous research, the adsorption process could remove NOM hydrophilic fractions with small MWs [9–11]. The removal rate of KmnO<sup>4</sup> organic substances was higher than the UV<sup>254</sup> absorbance. These results indicated that KmnO<sup>4</sup> organic substances easily oxidized NOM due to their small MWs as stated elsewhere [51].

results indicated that KmnO4 organic substances easily oxidized NOM due to their small MWs as stated elsewhere [51]. The best PAC dose was 120 mg/L of PAC, judging from the highest removal of KmnO4 and UV254 parameters of >90%. In addition, the efficiency of NOM removal slightly increased up to 120 mg/L. Due to the NOM removal of 120 to 200 mg/L being relatively similar, the 120 mg/L of PAC was chosen because it does not need too much PAC. The removal efficiencies of KMnO4 organic substances and UV254 under the optimum dosing were 95.83 and 91.83%, respectively, as shown in Figure 5. The results obtained in this study were in line with others, e.g., Lee et al. [32] and Aisyahwalsiah [39] The best PAC dose was 120 mg/L of PAC, judging from the highest removal of KmnO4 and UV<sup>254</sup> parameters of >90%. In addition, the efficiency of NOM removal slightly increased up to 120 mg/L. Due to the NOM removal of 120 to 200 mg/L being relatively similar, the 120 mg/L of PAC was chosen because it does not need too much PAC. The removal efficiencies of KMnO<sup>4</sup> organic substances and UV<sup>254</sup> under the optimum dosing were 95.83 and 91.83%, respectively, as shown in Figure 5. The results obtained in this study were in line with others, e.g., Lee et al. [32] and Aisyahwalsiah [39] using PAC as the adsorbant. Nonetheless, higher NOM removals were obtained in this study. After adding PAC, the pH increased as reported by others [49] due to the soluble ash, which is rinsed out of the media during use, and the effluent pH will eventually approach neutral.

### using PAC as the adsorbant. Nonetheless, higher NOM removals were obtained in this *3.5. Coagulation-Adsorption-Membrane Experiments*

study. After adding PAC, the pH increased as reported by others [49] due to the soluble ash, which is rinsed out of the media during use, and the effluent pH will eventually approach neutral. *3.5. Coagulation-Adsorption-Membrane Experiments*  The permeability of pure water (aquadest), pretreatment feed, and non-pretreated The permeability of pure water (aquadest), pretreatment feed, and non-pretreated feed are shown in Figure 6. Based on the results, pure water permeability was obtained of 38–180 L/h.m<sup>2</sup> by the prepared UF polysulfone membrane. This result exhibits the higher water flux of pure water permeability by low transmembrane pressure compared to commercial polysulfone membrane (Merck) of 150–350 L/h.m<sup>2</sup> (6–20 bar), which was reported by Adams et al. [52].

feed are shown in Figure 6. Based on the results, pure water permeability was obtained of 38–180 L/h.m2 by the prepared UF polysulfone membrane. This result exhibits the higher water flux of pure water permeability by low transmembrane pressure compared to commercial polysulfone membrane (Merck) of 150–350 L/h.m2 (6–20 bar), which was reported

by Adams et al. [52].

**Figure 6.** The performance of ultrafiltration of pretreated peat water at different pressures. Aquadest denotes distilled water and represents the permeability of clean water. **Figure 6.** The performance of ultrafiltration of pretreated peat water at different pressures. Aquadest denotes distilled water and represents the permeability of clean water.

Figure 6 shows that the NOM removal efficiency decreased by increasing transmembrane pressure. The results obtained are by previous research, in which the magnitude of NOM rejection is inversely proportional to the applied pressure [53,54]. The deformation of the membrane most probably causes it due to high pressure, which causes membrane compaction that constricts the pore size and the thicker foulant layer that became the sec-Figure 6 shows that the NOM removal efficiency decreased by increasing transmembrane pressure. The results obtained are by previous research, in which the magnitude of NOM rejection is inversely proportional to the applied pressure [53,54]. The deformation of the membrane most probably causes it due to high pressure, which causes membrane compaction that constricts the pore size and the thicker foulant layer that became the secondary filter on top of the PSf membrane.

ondary filter on top of the PSf membrane. The NOM removal efficiency reflected from the KMnO4 organic substances was higher than the UV254 absorbance obtained in the adsorption process with PAC. The rejection of NOM by membrane was determined by the adequate pore size [53] and an addi-The NOM removal efficiency reflected from the KMnO<sup>4</sup> organic substances was higher than the UV<sup>254</sup> absorbance obtained in the adsorption process with PAC. The rejection of NOM by membrane was determined by the adequate pore size [53] and an additional dynamic layer formed on the membrane surface.

tional dynamic layer formed on the membrane surface. In addition to the NOM removal rate, water flux value was also an indicator of the optimum pressure. Figure 6 shows that the water flux value was directly proportional to the pressure. The smallest water flux value was obtained at 1 bar of 13.3 L/h.m2, and the highest water flux was at 3 bar of 92.5 L/h.m2. The permeability of each pressure to percent removal of NOM for KMnO4 organic substances and UV254 parameters were determined to determine the optimum pressure. The highest water flux was obtained with the removal rate of KMnO4 and UV254 of 94.79 and 94.66%, respectively. The UV254 rejection of the polysulfone membrane in this work is extremely high over commercial PSf membrane In addition to the NOM removal rate, water flux value was also an indicator of the optimum pressure. Figure 6 shows that the water flux value was directly proportional to the pressure. The smallest water flux value was obtained at 1 bar of 13.3 L/h.m2, and the highest water flux was at 3 bar of 92.5 L/h.m<sup>2</sup> . The permeability of each pressure to percent removal of NOM for KMnO<sup>4</sup> organic substances and UV<sup>254</sup> parameters were determined to determine the optimum pressure. The highest water flux was obtained with the removal rate of KMnO<sup>4</sup> and UV<sup>254</sup> of 94.79 and 94.66%, respectively. The UV<sup>254</sup> rejection of the polysulfone membrane in this work is extremely high over commercial PSf membrane that was only able to remove about 41% of NOM at 6 bar [52].

that was only able to remove about 41% of NOM at 6 bar [52]. The water permeability of treated peat water was smaller than the clean water permeability (Figure 6) due to membrane fouling. However, it was higher by almost two-fold than peat water permeability without pretreatment, which was also similar reported in earlier studies [31,41,54]. It was shown that the pretreatment contributed substantially to The water permeability of treated peat water was smaller than the clean water permeability (Figure 6) due to membrane fouling. However, it was higher by almost two-fold than peat water permeability without pretreatment, which was also similar reported in earlier studies [31,41,54]. It was shown that the pretreatment contributed substantially to reducing the membrane fouling [53].

reducing the membrane fouling [53]. The permeability decrease in the pretreated peat water filtration can be attributed to the fouling by the residual NOM that escaped from the pretreatment. However, previous works by Kang and Choo [41] and Zhang*,* et al. [55] ascribed the small water permeability to the use of PAC. The bonding of NOM with PAC particles caused the PAC-NOM particles to become an additional foulant that blocks the membrane pores or forms a cake layer on the membrane surface. In this study, the PAC was separated. Hence the foulant was The permeability decrease in the pretreated peat water filtration can be attributed to the fouling by the residual NOM that escaped from the pretreatment. However, previous works by Kang and Choo [41] and Zhang, et al. [55] ascribed the small water permeability to the use of PAC. The bonding of NOM with PAC particles caused the PAC-NOM particles to become an additional foulant that blocks the membrane pores or forms a cake layer on the membrane surface. In this study, the PAC was separated. Hence the foulant was originated from residual NOM in the feed.

originated from residual NOM in the feed. Overall findings suggested that the application of coagulation–adsorption pretreatment of UF is promising to reduce the fouling potential on the feed as indicated by in-

creasing the water permeability value and the removal rate of NOM represented by DOC, KMnO<sup>4</sup> organic substance, and absorbance UV254. KMnO4 organic substance, and absorbance UV254. In addition, the results obtained were also reinforced with SEM UF-PSF membrane

In addition, the results obtained were also reinforced with SEM UF-PSF membrane image after treatment. The SEM images in Figure 7 also show the thickness of the UF-PSf membrane (determined from the cross-section SEM image) after being compressed at 3 bars was 85.4 µm. The pore structure of the membrane after passing the feed water was approximately the same as the pristine membrane. It could be seen on the surface of the membrane there is only a thin layer which is thought to be a cake layer. The additional cake layer helped enhance the rejection of NOM and the final quality of the permeate. It is worth noting that a significant difference in thickness was seen from data in Figures 2C and 7B. The high variability was originated from the cutting process. image after treatment. The SEM images in Figure 7 also show the thickness of the UF-PSf membrane (determined from the cross-section SEM image) after being compressed at 3 bars was 85.4 µm. The pore structure of the membrane after passing the feed water was approximately the same as the pristine membrane. It could be seen on the surface of the membrane there is only a thin layer which is thought to be a cake layer. The additional cake layer helped enhance the rejection of NOM and the final quality of the permeate. It is worth noting that a significant difference in thickness was seen from data in Figures 2C and 7B. The high variability was originated from the cutting process.

creasing the water permeability value and the removal rate of NOM represented by DOC,

**Figure 7.** SEM image of used ultrafiltration polysulfone membrane after ultrafiltration process: (**A**) surface section and (**B**) cross-sectional. **Figure 7.** SEM image of used ultrafiltration polysulfone membrane after ultrafiltration process: (**A**) surface section and (**B**) cross-sectional.

### **4. Conclusions**

**4. Conclusions** 

of the manuscript.

*Sustainability* **2022**, *13*, x FOR PEER REVIEW 9 of 12

This study demonstrated the advantages of combining the coagulation–adsorption process and membrane filtration to treat fouling-prone actual peat water. The coagulationadsorption showed a positive effect as a pretreatment for the ultrafiltration. The pretreated feed showed a lower membrane fouling propensity. The optimum coagulation/flocculation and adsorption condition was at Al2(SO4)3 dosing of 175 mg/L and PAC dosing of 120 mg/L, respectively. Higher filtration pressure enhanced the peat water permeability. The optimum pressure on the hybrid process was 3 bar with a permeability value of 92.5 L/m2.h and an organic removal rate of 95%. The findings highlight the importance of the hybrid system for treating challenging feeds that otherwise proven diffi-This study demonstrated the advantages of combining the coagulation–adsorption process and membrane filtration to treat fouling-prone actual peat water. The coagulationadsorption showed a positive effect as a pretreatment for the ultrafiltration. The pretreated feed showed a lower membrane fouling propensity. The optimum coagulation/flocculation and adsorption condition was at Al2(SO4)<sup>3</sup> dosing of 175 mg/L and PAC dosing of 120 mg/L, respectively. Higher filtration pressure enhanced the peat water permeability. The optimum pressure on the hybrid process was 3 bar with a permeability value of 92.5 L/m<sup>2</sup> .h and an organic removal rate of 95%. The findings highlight the importance of the hybrid system for treating challenging feeds that otherwise proven difficult when applying a standalone system. Moreover, long-term studies are still required to accurately gauge the performance of the hybrid system for treatment of peat water.

cult when applying a standalone system. Moreover, long-term studies are still required to accurately gauge the performance of the hybrid system for treatment of peat water. **Author Contributions:** Conceptualization, M.E. and M.M.; methodology, A.E.P.; software, A.E.P.; validation, A.E.P., M.E. and M.M.; formal analysis, A.R.; investigation, A.E.P.; resources, E.L.A.R.; data curation, A.E.P.; writing—original draft preparation, A.E.P.; writing—review and editing, **Author Contributions:** Conceptualization, M.E. and M.M.; methodology, A.E.P.; software, A.E.P.; validation, A.E.P., M.E. and M.M.; formal analysis, A.R.; investigation, A.E.P.; resources, E.L.A.R.; data curation, A.E.P.; writing—original draft preparation, A.E.P.; writing—review and editing, M.E.; visualization, A.E.P.; supervision, M.M., C.A., M.E. and M.R.B.; project administration, A.E.P., R.R., D.H.Y.Y.; funding acquisition, M.E. All authors have read and agreed to the published version of the manuscript.

M.E.; visualization, A.E.P.; supervision, M.M., C.A., M.E. and M.R.B.; project administration, A.E.P., **Funding:** This research received no external funding.

R.R., D.H.Y.Y.; funding acquisition, M.E. All authors have read and agreed to the published version **Funding:** This research received no external funding. **Acknowledgments:** The authors thank the Engineering Faculty and Materials and Membranes Re-**Acknowledgments:** The authors thank the Engineering Faculty and Materials and Membranes Research Group (M2ReG), Lambung Mangkurat University for the facilities. Muthia thanks the Applied Research of Universities Grant 2021–2023, Basic Research Grant 2021–2022, and World Class Research Grant 2021–2023 Directorate General of Higher Education, Ministry of Education, Culture, Research, and Technology, Republic of Indonesia.

search Group (M2ReG), Lambung Mangkurat University for the facilities. Muthia thanks the Applied Research of Universities Grant 2021–2023, Basic Research Grant 2021–2022, and World Class **Conflicts of Interest:** The authors declare no conflict of interest.
