Membrane Fouling Mechanisms in Combined Microfiltration-Coagulation of Algal Rich Water Applying Ceramic Membranes

In this paper, we investigated the membrane fouling mechanism according to the coagulant dosage in algal rich water using a ceramic membrane. The algae that were used in this experiment were Microcystis sp. of cyanobacteria, and the fouling mechanism was analyzed through irrigation and filtration resistance through a constant flow operation. The experimental results showed that the filtration resistance decreased as the coagulant dosage increased, but the irreversibility at above optimum coagulant dosage increased. Additionally, as the coagulant dosage increased, the resistance value due to cake and adsorption contamination decreased, and membrane fouling by adsorption was dominant in comparison with cake fouling and adsorption fouling. The specific cake resistance was decreased as the coagulant dosage increased. The characteristics of the cake layer according to the coagulant dosage were found to loosely form the cake layer by increasing micro-size algae as the coagulant dosage increased. The results of this experiment confirmed the membrane fouling mechanism according to coagulant dosage when the ceramic membrane filtered algal rich water.


Introduction
It is expected that global climate changes due to global warming will bring environmental changes that are different from the past. In particular, there is concern that a large outbreak of Cyanobacteria will be prolonged by an increase in water temperature during summer in water environments, such as rivers or lakes. In addition, the occurrence of algal blooms causes serious problems in conventional water treatment processes, including coagulation, flotation, filtration, and disinfection and then eventually causes undrinkable water. Recently, low pressure membrane filtration, such as ultra filtration (UF) and micro filtration (MF), has attracted increasing attention from algal rich water treatment areas because of the higher production of treated water and the lower operating costs. According to preliminary studies [1][2][3], a membrane process was able to completely remove algal cells, and a hybrid process combined with activated carbon was able to remove extracellular organic matter (EOM). However, membrane fouling remains an important issue until membrane technologies are widely applied. It has been reported that algal cells, including similar substances, such as polysaccharides, proteins, lipids, and humus mainly cause membrane fouling of algae-rich water [4]. The forms of membrane fouling include adsorption or gelation between algal particles and the membrane surface and formation of pore blockage or cake layers [5].
Numerous control strategies were proposed for mitigating the fouling resistances of membrane: feed pretreatment (coagulation, adsorption), change of feed characteristics (temperature, pH, dissolved oxygen), membrane modification (grafting, coating), optimizing operational parameters (cross-flow),

Experimental Set-Up
All of the experiments were conducted using laboratory-scale membrane set-up, which mainly included a raw water tank, a constant level water tank, a peristaltic pump, a pressure transducer, and a data acquisition system. Table 2 shows the specifications of the membranes used in this experiment. The immersed flat type MF membrane module (Cembrane, Denmark) was made of silicon carbide (SIC), with an effective membrane area of 0.0652 m 2 and a nominal pore size of 0.1 µm. The reactor (effective volume of 16.4 L) was fed with raw solution through the constant level tank and the effluent was directly drawn from the membrane module by the peristaltic pump (EMS-2000S, Korea). A pressure transducer (PTP708 Tuopo Electric, Korea), which was connected to a laptop computer, was used to continuously monitor the transmembrane pressure (TMP).

Operating Conditions
All of the ceramic membrane fluxes were maintained at 60 LMH, and the reactor temperature was adjusted to 20 ± 0.5 • C using a water bath. Air scrubbing was able to be operated stably at 1 LMH through preliminary experiments. The experiments were performed by applying it. Prior to the use, each membrane was washed and then flushed using ultrapure water under the same conditions. The coagulation conditions were adjusted to pH 7.5 ± 0.3 with 0.01 N H 2 SO 4 and 0.01 N NaOH.

Measurement of Resistance
Membrane fouling due to algal deposition was studied by measuring the following equations at a constant permeate flux (60 LMH) and water temperature (20 ± 0.5 • C). In the filtration experiments, the specific membrane resistance was first measured using ultrapure water.
where R c the resistance of algal cake (1/m) ∆P the transmembrane pressure (Pa) J the filtration flux (m 3 /m 2 /s) R m the resistance of membrane (1/m) R c can be made from the perspective of specific cake resistance, as follows: Furthermore, α is directly affected by the cake pressure gradient ∆P, and the functions of α and ∆P take the following forms [13][14][15][16][17]: α = α 0 (∆P) n α 0 Empirical constant n cake compressibility factor n has values between 0 and 1. It means non-compressibility as it approaches 0 and compressibility as it approaches 1 [18].
The reversibility and irreversibility according to the coagulant injection rate were analyzed while using the following methods. Membrane fouling was calculated by the following equations in defining the filtration resistance that was obtained by filtering ultrapure water through the initial membrane, the final filtration resistance of the fouled membrane by filtering using the target raw water to be treated, and the filtration resistance that was obtained by filtering using ultrapure water after physical cleaning as R 0 , R 1 , and R 2 , respectively: RF means reversible fouling, IF means irreversible fouling, and the total membrane (TF) becomes 1.

Measure of EPS
The extraction of EPS (Extracellular polymeric substances) was performed by the thermal extraction method in this study [19]. Protein analysis was measured by Bradford Assay method and polysaccharide was measured by the phenol sulfate method. All of the above analyses were conducted in duplicate, and their average values were reported. Figure 1 shows the changes in the ceramic membrane TMP (Transmembrane pressure) according to the coagulant dosage rate from 0 to 250 mg/L. It is observed that the rate of increase in TMP markedly decreases with increasing PaCl dosage. It has been reported algal cells, including similar substances, such as polysaccharides, proteins, lipids, and humus, and extracellular organic matter (EOM) cause the membrane fouling by algae [20]. Moreover, it is known that membrane fouling decreases with increasing the coagulant injection into these substances, because they are coagulated into larger aggregates [21]. The increase in TMP decreased rapidly with increasing the coagulant injection rate also in the results of these experiments. It is considered that algae particulate matter lowered the fouling resistance by coagulation. Figure 2a shows the reversible (R r ) and irreversible (R ir ) filtration resistance according to the coagulant injection rate. As shown in the figure, R r and R ir decreased with increasing the coagulant injection rate. The reversible fouling resistance (∆R r , −9 × 10 9 X) showed a higher rate of decrease than the irreversible fouling resistance (∆R ir , −4 × 10 8 X), according to the coagulant injection rate. In addition, Figure 2b shows the results of the irreversible analysis. Although the irreversibility decreases to 200 mg/L, it shows a tendency to increase from 250 mg/L. Wu et al. reported that membrane fouling was reduced within the reasonable coagulant dosage rate [22]. However, they reported that the excessive coagulant dosage increased the Zeta potential of raw water and then caused electrostatic repulsion to accelerate fouling by forming dense fouling layers. A similar tendency was found in this paper, and PaCl that was above the proper dosage (200 mg/L) dose increased irreversible fouling.

Flux Curves and Reversibility Analyses
Membranes 2019, 9, x FOR PEER REVIEW 4 of 11

Measure of EPS
The extraction of EPS (Extracellular polymeric substances) was performed by the thermal extraction method in this study [19]. Protein analysis was measured by Bradford Assay method and polysaccharide was measured by the phenol sulfate method. All of the above analyses were conducted in duplicate, and their average values were reported. Figure 1 shows the changes in the ceramic membrane TMP (Transmembrane pressure) according to the coagulant dosage rate from 0 to 250 mg/L. It is observed that the rate of increase in TMP markedly decreases with increasing PaCl dosage. It has been reported algal cells, including similar substances, such as polysaccharides, proteins, lipids, and humus, and extracellular organic matter (EOM) cause the membrane fouling by algae [20]. Moreover, it is known that membrane fouling decreases with increasing the coagulant injection into these substances, because they are coagulated into larger aggregates [21]. The increase in TMP decreased rapidly with increasing the coagulant injection rate also in the results of these experiments. It is considered that algae particulate matter lowered the fouling resistance by coagulation. Figure 2a shows the reversible (Rr) and irreversible (Rir) filtration resistance according to the coagulant injection rate. As shown in the figure, Rr and Rir decreased with increasing the coagulant injection rate. The reversible fouling resistance (∆Rr, −9 × 10 9 X) showed a higher rate of decrease than the irreversible fouling resistance (∆Rir, −4 × 10 8 X), according to the coagulant injection rate. In addition, Figure 2b shows the results of the irreversible analysis. Although the irreversibility decreases to 200 mg/L, it shows a tendency to increase from 250 mg/L. Wu et al. reported that membrane fouling was reduced within the reasonable coagulant dosage rate [22]. However, they reported that the excessive coagulant dosage increased the Zeta potential of raw water and then caused electrostatic repulsion to accelerate fouling by forming dense fouling layers. A similar tendency was found in this paper, and PaCl that was above the proper dosage (200 mg/L) dose increased irreversible fouling.   Figure 2 shows Ra and Rc to analyze the membrane fouling mechanism with coagulant dosage in algal rich water. As shown in Figure 2a, the adsorption resistant (Ra) showed a higher rate than cake resistance (Rc) regarding filtration using membranes in algal rich water. The rate for Ra showed a tendency to gradually increase with an increasing coagulation injection rate. Figure 2b shows each injection rate condition in filtration resistance logarithmic values. log (Ra) and log (Rc) tended to decrease with increasing coagulant injection rate and log (∆Rc) (−0.0031X) showed a higher decreasing trend than log (Ra) (−0.0037X) (Figure 3b).  Figure 2 shows R a and R c to analyze the membrane fouling mechanism with coagulant dosage in algal rich water. As shown in Figure 2a, the adsorption resistant (R a ) showed a higher rate than cake resistance (R c ) regarding filtration using membranes in algal rich water. The rate for R a showed a tendency to gradually increase with an increasing coagulation injection rate. Figure 2b shows each injection rate condition in filtration resistance logarithmic values. log (R a ) and log (R c ) tended to decrease with increasing coagulant injection rate and log (∆R c ) (−0.0031X) showed a higher decreasing trend than log (R a ) (−0.0037X) (Figure 3b).   As shown in Figure 4, ∆log (Ra) changes from 0.443 to 0.385 according to the coagulant injection rate from 0 mg/L to 250 mg/L and showed a tendency to decrease with an increasing injection rate. Although ∆log (Ra) showed the overall decrease, it increased the above proper coagulant dosage (200 mg/L). It was confirmed that the above proper coagulant dosage in algal rich water was more closely related to the fouling by cake than to the fouling resistance by adsorption.  As shown in Figure 4, ∆log (R a ) changes from 0.443 to 0.385 according to the coagulant injection rate from 0 mg/L to 250 mg/L and showed a tendency to decrease with an increasing injection rate. Although ∆log (R a ) showed the overall decrease, it increased the above proper coagulant dosage (200 mg/L). It was confirmed that the above proper coagulant dosage in algal rich water was more closely related to the fouling by cake than to the fouling resistance by adsorption.

Specific Cake Resistance and Compressibility
As shown in Figure 5, the specific cake resistance showed a tendency to decrease from 1.62 × 10 13 to 2.23 × 10 12 with an increasing PaCl dosage from 0 to 250 mg/L. According to Lee et al. [23], the floc structure had an effect on the specific cake resistance, and specific cake resistance increased with a decreasing floc structure, because it formed a more compact cake layer on the membrane surface. Moreover, Tabatabai et al. [24] reported that coagulation substantially reduced the fouling potential and compressibility of the AOM cake/gel layer by creating highly porous cakes. The membrane fouling resistance also decreased with an increasing PaCl dosage rate in this experiment, and it is considered that the specific cake resistance decreases because this floc forms a loose porous cake layer

Specific Cake Resistance and Compressibility
As shown in Figure 5, the specific cake resistance showed a tendency to decrease from 1.62 × 10 13 to 2.23 × 10 12 with an increasing PaCl dosage from 0 to 250 mg/L. According to Lee et al. [23], the floc structure had an effect on the specific cake resistance, and specific cake resistance increased with a decreasing floc structure, because it formed a more compact cake layer on the membrane surface. Moreover, Tabatabai et al. [24] reported that coagulation substantially reduced the fouling potential and compressibility of the AOM cake/gel layer by creating highly porous cakes. The membrane fouling resistance also decreased with an increasing PaCl dosage rate in this experiment, and it is considered that the specific cake resistance decreases because this floc forms a loose porous cake layer on the membrane surface. Figure 6 is the analysis results of polysaccharide in feed and permeate water. As the PaCl dosage rate increased, the polysaccharide concentration in the permeate water decreased. These results suggest that the small particles form large agglomerates by dosing coagulant, and it is thus removed by membrane. It was also confirmed that these agglomerates formed high porosity cake. on the membrane surface. Figure 6 is the analysis results of polysaccharide in feed and permeate water. As the PaCl dosage rate increased, the polysaccharide concentration in the permeate water decreased. These results suggest that the small particles form large agglomerates by dosing coagulant, and it is thus removed by membrane. It was also confirmed that these agglomerates formed high porosity cake.   Figure 7 shows the results of the compaction coefficient of the cake layer through correlations with specific cake resistance, according to TMP. It can be said that the compressible cake index (n) is the value that is determined by the raw water characteristics because membranes with the same materials were used in the experiments to determine the compaction of the coefficient and the raw water conditions were changed. As shown in Figure 7, the initial specific cake resistance (α0) decreased continually from 4.30 × 10 10 to 7.75 × 10 8 , according to PaCl dosage rate from 0 to 250 mg/L. on the membrane surface. Figure 6 is the analysis results of polysaccharide in feed and permeate water. As the PaCl dosage rate increased, the polysaccharide concentration in the permeate water decreased. These results suggest that the small particles form large agglomerates by dosing coagulant, and it is thus removed by membrane. It was also confirmed that these agglomerates formed high porosity cake.   Figure 7 shows the results of the compaction coefficient of the cake layer through correlations with specific cake resistance, according to TMP. It can be said that the compressible cake index (n) is the value that is determined by the raw water characteristics because membranes with the same materials were used in the experiments to determine the compaction of the coefficient and the raw water conditions were changed. As shown in Figure 7, the initial specific cake resistance (α0) decreased continually from 4.30 × 10 10 to 7.75 × 10 8 , according to PaCl dosage rate from 0 to 250 mg/L.  Figure 7 shows the results of the compaction coefficient of the cake layer through correlations with specific cake resistance, according to TMP. It can be said that the compressible cake index (n) is the value that is determined by the raw water characteristics because membranes with the same materials were used in the experiments to determine the compaction of the coefficient and the raw water conditions were changed. As shown in Figure 7, the initial specific cake resistance (α 0 ) decreased Membranes 2019, 9, 33 9 of 11 continually from 4.30 × 10 10 to 7.75 × 10 8 , according to PaCl dosage rate from 0 to 250 mg/L. On the other hand, the compressible cake index n decreased from 0.50 to 0.29, according to PaCl dosage rate from 0 to 200 mg/L, but it showed a higher value (0.36) at the dosage of 250 mg/L than that at the dosage of 200 mg/L. Table 3 summarizes the specific cake resistance and the compressible cake index. Specific cake resistance showed linear increase with increasing PaCl dosage, but compressible cake index tended to decrease within the proper dosage rate. However, it was confirmed that the dosing PaCl that above the proper dosage increased compressible cake index.
On the other hand, the compressible cake index n decreased from 0.50 to 0.29, according to PaCl dosage rate from 0 to 200 mg/L, but it showed a higher value (0.36) at the dosage of 250 mg/L than that at the dosage of 200 mg/L. Table 3 summarizes the specific cake resistance and the compressible cake index. Specific cake resistance showed linear increase with increasing PaCl dosage, but compressible cake index tended to decrease within the proper dosage rate. However, it was confirmed that the dosing PaCl that above the proper dosage increased compressible cake index.
On the basis of these results, the increase in coagulant dosage within the proper dosage rate in membrane filtration according to coagulant dosage in algal rich water caused a decrease in the initial cake resistance and less impaction of the cake layer that was attached to the membrane layer. The initial cake resistance decreased, but the compressible cake index increased in injecting the coagulants above the proper injection amount. As mentioned in the introduction, it is considered that a decrease in the electrostatic repulsion between the membrane surface and algal particles formed a dense cake layer on the membrane surface. On the basis of these results, the increase in coagulant dosage within the proper dosage rate in membrane filtration according to coagulant dosage in algal rich water caused a decrease in the initial cake resistance and less impaction of the cake layer that was attached to the membrane layer. The initial cake resistance decreased, but the compressible cake index increased in injecting the coagulants above the proper injection amount. As mentioned in the introduction, it is considered that a decrease in the electrostatic repulsion between the membrane surface and algal particles formed a dense cake layer on the membrane surface.

Conclusions
This paper examined membrane fouling according to coagulant dosage using raw water that was sampled from the lake where algae occurred. This study was performed to understand the cake properties of algae particles that were attached to the membrane surface under various coagulation conditions. The following conclusions were drawn: 1.
The rate of increase in TMP decreased with increasing PaCl dosage when filtering membranes in algal rich water. It was confirmed that PaCl dosage rate and TMP were important to each other.

2.
The reversible and irreversible fouling resistance decreased with an increasing PaCl dosage rate. The irreversible rate increased above the optimal PaCl dosage (200 mg/L as PaCl).

3.
Fouling resistance showed a tendency to decrease with an increasing PaCl dosage rate. As a form of membrane fouling, the adsorption resistant accounted for a higher proportion than cake resistance. In particular, cake resistance showed a higher decreasing trend than adsorption resistance. It is considered that an increase in the floc size according to coagulant played a causative role.

4.
The specific cake resistance and compressible index were analyzed to examine the cake layer properties according to the PaCl dosage rate. As a result, the cake resistance decreased with an increasing PaCl dosage, but the compressible index showed a tendency to increase above the proper coagulant dosage. It is considered that the calculation of the proper coagulant dosage is an important factor controlling membrane fouling in membrane process of algal rich water.