To improve the performance of the membrane process in the treatment of oily wastewater, the combined effects of pretreatment, membrane modification, and optimization of operating parameters on the microfiltration membrane system were investigated. First, coagulation and adsorption were used as pretreatment steps. Polyaluminium chloride and ferric chloride were employed as coagulants, and granular activated carbon was used as an adsorbent. In the optimal coagulation condition (1 g/L polyaluminium chloride, pH 7.5), chemical oxygen demand (COD) was reduced by 96%, while in the optimal adsorption condition, in which large amounts of activated carbon were utilized, 48% of COD was eliminated. A membrane of polyethersulfone containing SiO2-g-polymethacrylic acid (PMAA) nanoparticles was then prepared by the non-solvent-induced phase separation method. To reduce fouling and increase the flux of the membrane, the SiO2 nanoparticles were first activated with amine groups and then PMAA was grafted onto the surface of the particles. Subsequently, the operating parameters were studied to optimize the performance of the polyethersulfone (PES)/SiO2-g-PMAA membrane using the response surface methodology (RSM) method. The results indicated that the flux of the modified membrane for pretreated wastewater was 72.2% higher than that of the PES membrane and non-pretreated wastewater at an optimum pressure of 2 bar and a flow rate of 3.5 L/min.

  • A combination of conventional wastewater treatment and membrane-based process was studied.

  • A polyethersulfone membrane modified with SiO2-g-PMAA nanoparticles was prepared.

  • The effects of coagulation and adsorption as the pretreatment were investigated.

  • The response surface methodology model was used to determine the best-operating conditions (cross flow velocity (CFV), transmembrane pressure (TMP), and feed concentration).

  • The combined process showed high MF performance and chemical oxygen demand removal.

Graphical Abstract

Graphical Abstract
Graphical Abstract

Considering the increasing amount of oily wastewater generated by a variety of process industries, it is imperative to treat the wastewater before discharging it into the sewer system. Due to the growing demand for freshwater, water scarcity, and stringent pollution control regulations, numerous methods have been proposed, and one of the high-tech and significant solutions is membrane-based separation, particularly the cost-efficient microfiltration and ultrafiltration for treating stable emulsified oil from wastewater (Mo & Huang 2003; Chakrabarty et al. 2008; Lee et al. 2009; Salahi et al. 2010; Kehrein et al. 2021). Despite the remarkable advantages of membrane filtration, the decrease in permeate flux is still a major problem that hinders the expanding implementation of the membrane process on large scales (Guo et al. 2004; Murić et al. 2014; Shang et al. 2015). To address this significant drawback, causing restriction in extensive utilization, experiments have been conducted on membrane modification (Sun et al. 2013; Chang et al. 2015; Ghandashtani et al. 2015), feed pretreatment (Abdessemed & Nezzal 2005; Lerch et al. 2005; Barbot et al. 2008; Abbasi et al. 2012; Hashlamon et al. 2017; Mikhak et al. 2019), and optimization of operating conditions (Lobo et al. 2006; Rezaei et al. 2011; Abadikhah et al. 2014; Fouladitajar et al. 2016). In various studies, common methods, such as coagulation and adsorption, were adopted as suitable pretreatments in combination with membrane filtration because of their low cost and ease of operation (Fabris et al. 2007; Wu et al. 2009; Ang et al. 2016; Yu et al. 2021). Coagulation plays an important role in charge neutralization, agglomeration of oil droplets, formation of flocs, settling, and finally removal of flocs (Zouboulis & Avranas 2000; Ahmad et al. 2006). Coagulants such as ferric and aluminium salts and polyaluminium chloride are widely used in wastewater treatment (Ahmad et al. 2006; Dong et al. 2015; Ren et al. 2021). In addition, the type and dosage of coagulants, pH, and mixing time should be determined under optimal conditions to enhance the efficiency of the procedure (Barbot et al. 2008; Lee et al. 2009; Martín et al. 2011). In the adsorption process, activated carbon is a conventional adsorbent used to remove a variety of organic compounds, including oil from wastewater owing to its large surface area, large pore volume, and availability (Mohammadi & Esmaeelifar 2005; Yang et al. 2011; Sarfaraz et al. 2012). It is evident that the adsorption/coagulation step has been successfully utilized to improve filtration performance by reducing the concentration of low molecular weight organic pollutants through adsorption or floc structure modification through coagulation. Moreover, amending the hydrophilicity of membranes can affect membrane fouling by forming more hydroxyl groups on the membrane surface and increasing resistance to oil adhesion. So, different procedures are applied for this intention such as blending hydrophilic polymers or inorganic materials with a hydrophobic polymer; among them, using inorganic nanoparticles in polymeric casting solution has grabbed more attention because it reinforces the membrane permeability and reduces fouling (Wu et al. 2008; Yan et al. 2009; Vatanpour et al. 2012). In order to improve the dispersion properties of nanoparticles and abate the aggregation phenomenon on the membrane's surface and in the membrane's bulk, which lead to blocking the pores, nanoparticle surface modification by the grafting method was commonly used (Gao et al. 2011; Zhu et al. 2014; Yin & Zhou 2015). Furthermore, changing the operating conditions, such as the cross-flow velocity, transmembrane pressure, and oil concentration in the feed, could affect the membrane operation enhancement (Hua et al. 2007; Abadi et al. 2011; Badrnezhad & Beni 2013). Although much research has been done in this regard, there are no efforts based on the simultaneous combination of all these three mentioned approaches, which seems to be necessary to investigate.

The aim of the present study is to investigate the performance of combination processes of pretreatment, membrane modification, and the change of operating parameters to improve membrane performance. In this way, firstly, the effect of pretreatment was investigated by using the coagulation and adsorption process according to the chemical oxygen demand (COD) removal ability in the same synthesis wastewater and under different conditions. Secondly, the polyethersulfone membrane modified with SiO2-g-polymethacrylic acid (PMAA) nanoparticles was provided by the phase inversion method. The interfacial properties of the SiO2 nanoparticles were ameliorated by grafting hydrophilic PMAA onto the surface of the nanoparticles (NPs). Finally, the best-operating conditions such as cross flow velocity (CFV), transmembrane pressure (TMP), and feed concentration were considered to optimize the flux and COD removal of the modified membrane in a microfiltration (MF) procedure using the response surface methodology (RSM) model to evaluate the filtration capability of the membrane in oily wastewater.

Feed preparation

The synthesized oil-in-water (O/W) emulsion consisted of sunflower oil purchased from Behshahr Ind. Co., and the dispersed oil phase was prepared by mixing oil and surfactant in distilled water at a mixing speed of 12,000 rpm for 30 min. The surfactant used was Tween 80, Merck, at a concentration of 100 ppm. It should also be mentioned that the dynamic laser scattering (DLS) method (Nano ZS (red badge) ZEN3600, Malvern, UK) was utilized to determine the particle size distribution (PSD) of the feed solutions. This parameter was measured two times, first when the emulsion was prepared and then after 4 h which was more than the experiments’ required time. The provided curves in Figure 1(a) and 1(b) clearly show that the mean oil droplet diameter was about 0.1 μm and no significant diversification occurred in a certain period of time. Therefore, it can be concluded that the oil-in-water emulsion was stable during the experiments.
Figure 1

The mean oil droplet diameter (a) after emulsion preparation (b) after 4 h.

Figure 1

The mean oil droplet diameter (a) after emulsion preparation (b) after 4 h.

Close modal

Coagulation experiments

The standard jar test was carried out in the laboratory experiment and as a batch test using the Jar Tester (Velp Scientifica FP4). Procedures included 1 min of rapid mixing (150 rpm), followed by 30 min of slow mixing (40 rpm) and 1 h of settling. The supernatant was taken from the top zone of each beaker for the determination of the COD test. The conventional potassium dichromate oxidation procedure according to the standard method EPA 410.4 and an atomic adsorption spectrophotometer (Jasco, V-550) were used to analyze the COD values. Polyaluminium chloride (PAC) and ferric chloride (FeCl3), provided by RSCO (Iran) with a purity of 99%, were used as coagulants at dosages of 0.05, 0.1, 0.5, 1, 2, and 4 g/L. In addition, diluted 0.1 M HCl and NaOH, purchased from MERCK Company (Germany), were used to adjust the pH of the solution to the desired values using a pH meter (Metro ohm model 744).

Adsorption experiments

Commercially available granular activated carbon (GAC), obtained from Norit Netherlands BV, was utilized for the experiments and characterized before using it in our laboratory (Fouladi Tajar et al. 2009). The general properties of GAC are listed in Table 1. In the batch studies, 5 g of GAC was added to a flask containing an O/W emulsion at the desired concentration. The flask was shaken continuously at 25 °C and 150 rpm in a shaker (N-BIOTEK 205v). It should be noted that the shaking provided good contact between the high surface area of GAC and dissolved oil particles. At the end of each step, the supernatant liquid was filtered and the COD was determined. Equilibrium time was defined as the contact time required for the reduction of COD in solution to reach equilibrium. The amount of COD adsorbed by the solid (q) was calculated from the expression:
formula
(1)
where C0 is the initial concentration, Ce is the residual COD, V is the volume solution (L), and m is the mass of the adsorbent. During the tests, the contact time was set to 3 h. The effects of changing pH and adsorbent concentrations were also studied to determine the effects of these parameters on COD reduction.
Table 1

General characteristics of GAC

ValueUnit
Iodine number 975 – 
Methylene blue adsorption 20 g/100 g 
Total surface area (BET) 1,100 m2/g 
Apparent density 470 kg/m3 
Particle size>8 mesh (2.36 mm) Max. 15 Mass% 
Particle size<30 mesh (0.6 mm) Max. 5 Mass% 
ValueUnit
Iodine number 975 – 
Methylene blue adsorption 20 g/100 g 
Total surface area (BET) 1,100 m2/g 
Apparent density 470 kg/m3 
Particle size>8 mesh (2.36 mm) Max. 15 Mass% 
Particle size<30 mesh (0.6 mm) Max. 5 Mass% 

Materials for membrane preparation

Polyethersulfone (PES, E6020p) was obtained from BASF (Germany); N-methyl-2-pyrrolidone (NMP) and the SiO2 nanoparticles (average particle size of 10 nm, the specific surface area of more than 600 m2/g, and with a hydrophilic group) provided by MERCK Company (Germany), and polyvinylpyrrolidone (PVP) of k25 grade (MW ∼ 25,000 g/gmol) was purchased from Rahavard Tamin Pharmaceutical Company (Iran). Distilled water was utilized as a non-solvent in the coagulation bath.

Nanoparticle modification

To modify the surface of nanoparticles, firstly, the activation step is required. For this purpose, 0.6 g of completely dried nanoparticles were cast into 50 mL of toluene and the UP200S ultrasonic processor (Hielscher, Germany) was used to ensure the better dispersion of nanoparticles for 20 min. Then, an amount of 5.5 mL of the activator triaminopropylmethyl methoxylan (APTMS, Merck, Germany) was added to the solution. In order not to evaporate the solvent, the return condenser was applied and the temperature of the solution was adjusted with the oil bath at 90 °C, and the reaction took 24 h to complete. After completion of the reaction, the solution should be centrifuged and the obtained product was washed using toluene and water. Finally, an oven was used for 24 h at 60 °C to dry the activated SiO2 nanoparticles. Hereafter, for the formation of a polymeric methacrylic acid chain on the surface of the nanoparticles, 0.5 g of activated nanoparticles were dissolved in 45 mL of water and again ultrasonicated for better distribution. A three-neck flask was used in this reaction, and a mercury thermometer was utilized to adjust the temperature and return the condenser to prevent the evaporation of the solvent. The reaction was carried out under a nitrogen atmosphere and parafilm was used to completely seal the system. Then, 2.8 mL of methacrylic acid monomer was added to the solution, and after 30 minutes and ensuring a full evacuation of air, ammonium peroxylate (0.4 wt.% monomer) as an initiator was added. It should be noted that during the reaction, the solution was completely mixed with a magnetic stirrer at 70 °C. The polymerization was completed after 7 h, and the product was separated using a centrifuge (Hettich Universal 320, Germany) at 8,000 rpm for 7 min. Finally, the product was washed several times with ethanol and then dried at 60 °C (Khodadousti 2019).

Membrane preparation, filtration module, and process apparatus

Based on previous works (Ghandashtani et al. 2015; Tashvigh et al. 2015; Armand et al. 2017; Boshrouyeh et al. 2018; Khodadousti 2019; Ghorabi et al. 2021), the PES–SiO2 membrane was prepared under the desired and specified preparation conditions as shown in Table 2. The non-solvent-induced phase separation (NIPS) was used to prepare the membrane. First, a vacuum oven was used at 110 °C for 12 h to dry PES before preparing a casting solution. Then, NPs were added to the solvent, and to improve their scattering, they should be sonicated for 20 min before the determined amounts of PES and PVP according to Table 2 were added to the solution. Next, the solution was stirred at 1,000 rpm for 12 h, and in the end, the casting solution was allowed to settle for half a day to eliminate possible bubbles. Subsequently, the provided solution was poured onto a clean glass plate and cast with a 180-μm casting knife. The cast film should be put in a 25 °C distilled water coagulation bath for phase-inversion initiation. After the completion of the phase-inversion and removing the solvent from the cast film, the membrane was soaked in distilled water for 24 h. Further details about the preparation of the module and the experimental procedure have been described in previous research papers (Rezaei et al. 2011; Fouladitajar et al. 2013; Valizadeh et al. 2015).

Table 2

Membrane preparation conditions

ParametersValue
Nanoparticle concentration (% w/w of polymer) 1.5 
Nanoparticle size (nm) 10 
Coagulation water bath temperature (°C) 25 
PES concentration (wt.%) 14 
PVP concentration (wt.%) 
Membrane thickness (μm) 180 
ParametersValue
Nanoparticle concentration (% w/w of polymer) 1.5 
Nanoparticle size (nm) 10 
Coagulation water bath temperature (°C) 25 
PES concentration (wt.%) 14 
PVP concentration (wt.%) 
Membrane thickness (μm) 180 

The performance of the prepared membrane was measured by permeate flux and oil rejection that were calculated by using Equations (2) and (3), respectively:
formula
(2)
formula
(3)

Here, J refers to the constant flux, Mp is the permeate flow mass, A is the active membrane area, t is the predetermined time of filtration, and for oil rejection, Cp and Cf are the oil concentration in permeate and feed streams (mg/L), respectively.

The fabricated membrane was applied in a cross-flow pattern process under different operating conditions. The three-level and three-factor central composite design (CCD) was utilized for statistical analysis of membrane efficiency according to various operating parameters, e.g., TMP, feed flow rate, and COD of feed, in which levels are shown in Table 3 in coded and actual values.

Table 3

Code and level of factors used for CCD

VariablesCoded variablesLevels in coded and actual values
α− 101+ α
Feed flow rate (L/min) 1.35 1.5 2.5 3.5 3.65 
Pressure (bar) 0.93 1.5 1. 2. 07 
Concentration (ppm) 200 600 3,300 6,000 6,400 
VariablesCoded variablesLevels in coded and actual values
α− 101+ α
Feed flow rate (L/min) 1.35 1.5 2.5 3.5 3.65 
Pressure (bar) 0.93 1.5 1. 2. 07 
Concentration (ppm) 200 600 3,300 6,000 6,400 

Pretreatment

Coagulation

For the coagulation step as a pretreatment, first, the effects of pH and coagulant dosage were investigated for two types of coagulants (ferric chloride and polyaluminium chloride) according to the COD removal. This was due to the fact that appropriate pH can neutralize the negative charge of colloidal particles and form a link between them to form flocs and consequently accelerate the settling process (Valizadeh et al. 2015). Also, to overcome the poor performance of coagulation, caused by overdosing, and to minimize the dosing cost, it was essential to determine the optimal dosage. The results are shown in Figures 2,34, and it should be mentioned that the initial COD of synthetic effluent was measured and was equal to 6,000 ppm. As shown in Figures 2 and 3, the highest removal efficiency was 97.95% at pH 6 for 1 g/L of FeCl3, and 98.5% for 1 g/L of polyaluminium chloride at pH 7.5, which can be attributed to the precipitation of aluminium hydroxide in water at pH 7, while its solubility increased in acidic and alkaline zones. From Figure 4, it also can be concluded that the good efficiency of COD elimination was achieved by increasing the coagulant concentration up to 1 g/L at certain optimal pH values, which were 96.85% for FeCl3 and 97.61% for PAC. However, higher concentrations of coagulants resulted in a slight decrease in efficiency, which was due to the particle charge inversion, restabilization of oil particles, and the prevention of floc formation. Therefore, comparing the effectiveness of two coagulants represented the slightly better performance of PAC in comparison with FeCl3, and this result was also consistent with the finding of Fouladitajar et al. (2013).
Figure 2

Effects of pH on COD and the percentage of COD removal in oil-in-water emulsion after using FeCl3.

Figure 2

Effects of pH on COD and the percentage of COD removal in oil-in-water emulsion after using FeCl3.

Close modal
Figure 3

Effects of pH on COD and the percentage of COD removal in oil-in-water emulsion after using PAC.

Figure 3

Effects of pH on COD and the percentage of COD removal in oil-in-water emulsion after using PAC.

Close modal
Figure 4

Effects of coagulant dosage on COD and COD removal in oil-in-water emulsion.

Figure 4

Effects of coagulant dosage on COD and COD removal in oil-in-water emulsion.

Close modal

Adsorption

For the adsorption step, the equilibrium time, effects of pH, and adsorbent dosage were studied. The equilibrium time was determined at 25 °C and different contact times (from 0 to 240 min) with an adsorbent dosage of 5 g/L without any pH adjustment. The results showed that the COD removal efficiency after 30 min was 13.12% (171.7 mg/g) and remained constant, indicating that the granular activated carbon was saturated in a short time and could not achieve high removal efficiency. In addition to this, Figure 5 shows that, at a concentration of 5 g/L adsorbent, a decrease in pH value (4) increases the removal efficiency up to 19.38%. Thus, the pH of the solution can affect the surface charge of the adsorbent, the degree of ionization, and the dissociation of the functional groups on the adsorbent surface (Zhou et al. 2011). Moreover, as shown in Figure 6, the optimum removal rate is achieved at high amounts of GAC as the surface area of the particles increases. For example, at a concentration of 20 g/L of GAC, the COD can only be removed up to 48%, which is not economically feasible due to the recovery problem at this amount of adsorbent. Therefore, it can be deduced that coagulation had better performance in comparison with adsorption as a pretreatment step.
Figure 5

Effect of pH on COD removal in adsorption.

Figure 5

Effect of pH on COD removal in adsorption.

Close modal
Figure 6

Effect of adsorbant dosage on COD removal at 25 °C, pH 4, and C0 = 6,000 ppm.

Figure 6

Effect of adsorbant dosage on COD removal at 25 °C, pH 4, and C0 = 6,000 ppm.

Close modal

Membrane filtration

Characterization of SiO2-g-PMAA

To ensure the formation of polymeric chains on SiO2 nanoparticles, the Fourier transform infrared spectroscopy (FTIR) spectra of SiO2 and SiO2-g-PMAA were measured using a BRUKER (Tensor 27) FTIR spectrophotometer (Table 4).

Table 4

FTIR spectra of SiO2 and SiO2-g-PMMA

Wave number (cm−1)
Functional groups
SiO2SiO2-g-PMMA
3,422 3,259 O–H stretching of the hydroxyl group 
 1,716 C = O stretching of the –COOH group 
 957 O–H out-of-plane bending of –COOH 
Wave number (cm−1)
Functional groups
SiO2SiO2-g-PMMA
3,422 3,259 O–H stretching of the hydroxyl group 
 1,716 C = O stretching of the –COOH group 
 957 O–H out-of-plane bending of –COOH 

As shown in Table 4, the bare SiO2 sample illustrates strong absorption at 3,422 cm−1 due to the different hydroxyl groups on the nanoparticles, while this value is lower for SiO2-g-PMAA due to the reaction between activating agents and nanoparticles. According to SiO2-g-PMAA, the stretching vibrations of the carbonyl group are detected at 1,716 cm−1 (Zhou et al. 2011). In addition to this, the absorption peaks at 3,259 and 957 cm−1 are determined as stretching vibrations of the O–H band and out-of-plane stretching vibrations of hydroxyl in carboxyl groups, respectively. So, due to these discrepancies in the absorption spectra, it can be concluded that the polymer was grafted onto the nanoparticles’ surface and the SiO2-g-PMAA was formed (Gao et al. 2011; Khodadousti 2019).

Modified PES–SiO2 membrane surface morphology

Figure 7 illustrates the impact of nanoparticle modification on the surface structure of the membrane. As can be seen, the addition of the nanoparticles to the PES membrane in the polymer solution increases the surface pores and contributes to a more uniform distribution. By adding modified nanoparticles to the membrane structure, due to the greater compatibility that is produced in the polymer solution, the surface pores are more regular and as a result, a membrane with more hydrophilicity and desirable properties is obtained (Khodadousti 2019).
Figure 7

Field emission scanning electron microscopy (FESEM) image of membrane surface. (a) Bare polyethersulfone (PES) and (b) PES/SiO2-g-polymethacrylic acid (PMAA) (1.5 wt.%).

Figure 7

Field emission scanning electron microscopy (FESEM) image of membrane surface. (a) Bare polyethersulfone (PES) and (b) PES/SiO2-g-polymethacrylic acid (PMAA) (1.5 wt.%).

Close modal

RSM model for permeate flux

RSM and CCD were exerted to study the effect of the considered and influent factors on the response variables, as well as to analyze the interaction between them and to optimize the microfiltration of the oil-in-water emulsion process. The effects of the pretreated and unpretreated wastewater COD (CFeed), TMP, and liquid flow rate (QL) on permeate flux and COD removal were studied at three levels: high (+1), low (−1), and center points (0) as well as star point (±α) set at the outer value of ±1.15. The designed experiments including the coded values of the factors and responses are presented in Table 5.

Table 5

CCD and responses for MF of oily wastewater

RunInput variables
Responses
QL A (L/min)TMP B (bar)CFeed C (ppm)Flux (L/m2·h)Rejection (%)
−1 −1 15.14 99.6 
−1 108.21 95.1 
α 86.57 95.8 
α 21.97 98.8 
+α 46.06 98.2 
−1 −1 54.32 96.3 
40.28 98.94 
36.21 99.02 
39.67 98.9 
10 −1 24.95 99.2 
11 +α 30.29 99.3 
12 32.81 99.04 
13 α 21.64 99.2 
14 −1 25.24 99 
15 −1 −1 50.5 96.1 
16 +α 57.71 98.3 
17 42.44 98.9 
18 35.36 99 
19 −1 −1 −1 28.07 97.5 
20 40.39 98.4 
RunInput variables
Responses
QL A (L/min)TMP B (bar)CFeed C (ppm)Flux (L/m2·h)Rejection (%)
−1 −1 15.14 99.6 
−1 108.21 95.1 
α 86.57 95.8 
α 21.97 98.8 
+α 46.06 98.2 
−1 −1 54.32 96.3 
40.28 98.94 
36.21 99.02 
39.67 98.9 
10 −1 24.95 99.2 
11 +α 30.29 99.3 
12 32.81 99.04 
13 α 21.64 99.2 
14 −1 25.24 99 
15 −1 −1 50.5 96.1 
16 +α 57.71 98.3 
17 42.44 98.9 
18 35.36 99 
19 −1 −1 −1 28.07 97.5 
20 40.39 98.4 

The system behavior was clarified by a reduced quadratic regression model in terms of the actual factors to fit the experimental data for permeate flux as follows:
formula
Subjected to:
formula

It is worth mentioning that the CFeed was determined on the basis of both non-pretreated and pretreated wastewater using 0.05, 0.5, and 1 g/L PAC as a coagulant, which resulted in the feed COD reduction to 3,300, 600, and 200 ppm, respectively.

According to the obtained results, the permeate flux versus filtration time for each run is shown in Figure 8. It can be clearly seen that in all experiments, two areas emerge: a sharp decrease at the beginning of filtration due to the particle deposit formation on the membrane surface as well as internal fouling and then a slow decrease until a steady-state permeate flux was reached.
Figure 8

Time dependency of permeate flux for all experiment runs.

Figure 8

Time dependency of permeate flux for all experiment runs.

Close modal

To test the significance of the regression coefficients, a statistical student test was used, and to verify the adequacy of the developed model, an analysis of variances (ANOVA) with omitted non-significant terms (p-value greater than 0.05) from the predictive model was performed, which is summarized in Table 6. For statistical validation, the F-value should be as high as possible, while the p-value must be as low as possible, and based on this tip, the model F-value of 32.89 and p-value of less than 10−4 showed that the model was significant. In addition, there was only a 0.01% chance of error because of noise. Moreover, the correlation coefficient R2 value for permeate flux is 0.96, indicating an acceptable fit of the quadratic models to the experimental data. The value of the adjusted R2 is 0.9246, which could confirm the accuracy of the quadratic models. The value of lack of fit (<0.00001) and the adequate precision ratio that was discovered to be 23.44 is desirable, and all these statistical estimators revealed the validated developed model to predict and simulate the MF process in statistical concepts. The order of significant factors in this model is as follows: C: Feed concentration first-order main effect > A: feed flow rate first-order main effect > B: TMP first-order main effect > C2 quadratic main effect of Feed concentration > BC interaction of TMP and Feed concentration > AC interaction of feed flow rate and Feed concentration > B2 quadratic main effect of TMP > AB interaction of feed flow rate and TMP.

Table 6

ANOVA tables and statistical parameters for the reduced regression quadratic RS model (response: permeate flux)

SourceSum of squaredfMean squareF-valuep-value
Model 9,074.63 1,134.33 32.89 <0.0001 
A – Feed flow rate 1,929.52 1,929.52 55.94 <0.0001 
B – Pressure 1,608.93 1,608.93 46.64 <0.0001 
C – Concentration 3,761.31 3,761.31 109.04 <0.0001 
AB 169.46 169.46 4.91 0.0487 
AC 322.07 322.07 9.34 0.0109 
BC 434.83 434.83 12.61 0.0045 
B2 270.70 270.7 7.85 0.0172 
C2 819.56 819.56 23.76 0.0005 
Residual 379.43 11 34.49   
Lack of fit 313.42 52.24 3.96 0.0764 
Pure error 66.01 13.2   
Total 9,454.05 19    
R2 0.9603     
R2-adjusted 0.9246     
Coefficient of variance (CV) (%) 14.10     
Adequate precision 23.44     
SourceSum of squaredfMean squareF-valuep-value
Model 9,074.63 1,134.33 32.89 <0.0001 
A – Feed flow rate 1,929.52 1,929.52 55.94 <0.0001 
B – Pressure 1,608.93 1,608.93 46.64 <0.0001 
C – Concentration 3,761.31 3,761.31 109.04 <0.0001 
AB 169.46 169.46 4.91 0.0487 
AC 322.07 322.07 9.34 0.0109 
BC 434.83 434.83 12.61 0.0045 
B2 270.70 270.7 7.85 0.0172 
C2 819.56 819.56 23.76 0.0005 
Residual 379.43 11 34.49   
Lack of fit 313.42 52.24 3.96 0.0764 
Pure error 66.01 13.2   
Total 9,454.05 19    
R2 0.9603     
R2-adjusted 0.9246     
Coefficient of variance (CV) (%) 14.10     
Adequate precision 23.44     

To evaluate the adequacy and reliability of the statistical model, plots of the normal probability of the residuals and the residuals versus predicted responses for permeate flux were taken into the consideration. From this, it can be concluded that the model was suitable for predicting the permeate flux of oil-in-water emulsions. The mentioned plots are provided in the Supplementary Materials (S1).

The three-dimensional plots in Figures 9 and 10 show both the main and interaction effects of the various variables on the permeate flux of the fabricated membrane. The figures show that an increase in TMP and feed flow rate enhances the permeate flux, whereas an increase in feed concentration leads to a decrease in permeate flux. This means that the built-up concentration polarization and the formation of a cake layer were drastic at higher oil concentrations. Moreover, the principal effect of feed concentration was two times larger than that of feed flow rate and TMP. Figure 9 also illustrates the interaction between feed concentration and TMP. As can be seen, the effect of TMP was remarkable at lower feed concentrations, while the effect of feed concentration was notable at higher TMPs. For example, at feed COD of 600 mg/L, increasing the TMP from 1 to 2 (bar) resulted in a 48% increment in response, whereas only a 30% increment was noted at a concentration of 6,000 (mg/L). This means that at high CFeed, and TMP, more oil droplets can pass through the membrane pores, eventuate in radius reduction, and consequently, in pore constriction and/or blocking.
Figure 9

Permeate flux decline as a function of pressure and COD concentration.

Figure 9

Permeate flux decline as a function of pressure and COD concentration.

Close modal
Figure 10

Permeate flux decline as a function of feed flow rate: (a) feed concentration and (b) pressure.

Figure 10

Permeate flux decline as a function of feed flow rate: (a) feed concentration and (b) pressure.

Close modal

In addition, some other interactions between QLCFeed and QL–TMP were observed (Figure 10(a) and 10(b)). As presented, the permeate flux was improved by increasing the feed flow rate and its impact was more drastic at both lower pressure and feed concentration, because of the lower compaction of oil droplets at the membrane surface. Thus, the main reason for this improvement is the disruption of the gel layer due to the generation of shear stress and more turbulence in the boundary layer of the membrane surface. The results are in agreement with the studies of Chakrabarty et al. (2008) and Rezaei et al. (2011). Therefore, to achieve the higher permeate flux, the MF should be performed at high TMP (2 bar), QL (3.5 L/min), and low CFeed (600 ppm), which can be achieved by pretreatment.

RS model for rejection

The same procedure and simulation were applied to the second response variable, rejection (%).

Based on the ANOVA (Table 7), it can be deduced that the developed RS model is significant from the statistical concept. The reduced quadratic regression model in terms of the actual factors was developed as follows:
formula
Table 7

ANOVA tables and statistical parameters for the reduced regression quadratic RS model (response: oil rejection)

SourceSum of squaredfMean squareF-valuep-value
Model 33.60 4.80 251.28 <0.0001 
A – Feed flow rate 2.48 2.48 129.77 <0.0001 
B – Pressure 1.39 1.39 72.97 <0.0001 
C – Concentration 21.67 21.67 1,134.56 <0.0001 
AC 0.19 0.19 9.90 0.0084 
BC 0.18 0.18 9.58 0.0093 
B2 0.43 0.43 22.34 0.0005 
C2 4.85 4.85 253.97 <0.0001 
Residual 0.23 12 0.019   
Lack of fit 0.18 0.025 2.31 0.1865 
Pure error 0.054 0.011   
Total 33.83 19    
R2 0.9932     
R2-adjusted 0.9893     
CV (%) 0.14     
Adequate precision 51.988     
SourceSum of squaredfMean squareF-valuep-value
Model 33.60 4.80 251.28 <0.0001 
A – Feed flow rate 2.48 2.48 129.77 <0.0001 
B – Pressure 1.39 1.39 72.97 <0.0001 
C – Concentration 21.67 21.67 1,134.56 <0.0001 
AC 0.19 0.19 9.90 0.0084 
BC 0.18 0.18 9.58 0.0093 
B2 0.43 0.43 22.34 0.0005 
C2 4.85 4.85 253.97 <0.0001 
Residual 0.23 12 0.019   
Lack of fit 0.18 0.025 2.31 0.1865 
Pure error 0.054 0.011   
Total 33.83 19    
R2 0.9932     
R2-adjusted 0.9893     
CV (%) 0.14     
Adequate precision 51.988     
Subjected to:
formula

It should be noted that only significant terms were retained and non-significant factors (p > 0.05) were eliminated from the model. The model's F-value of 259.08 and p-value of less than 10−4, as well as R2 and R2-adjusted values of 0.9957 and 0.9919, show that the model is adequate. This model also confirms that COD removal (%) is the main function of feed concentration compared to TMP and QL. The high correlation between the predicted versus actual values of oil removal was illustrated in the Supplementary Material (S2).

To investigate the main and interaction effects of different variables on the 3D surface of rejection based on the developed RS model Figures 11 and 12 are given. These plots illustrate that increasing the TMP and QL had a regressive effect on the oil rejection, while increasing the feed concentration led to an improvement in rejection since increasing the oil concentration formed a thicker concentration polarization layer on the membrane and play a significant role as a hindrance to maintaining the oil droplets. However, higher TMPs can make the deformable oil droplets pass through the membrane pores, and also the increase of QL can effectively reduce the thickness of the formed layer. As can be seen in Figure 11, the effects of TMPs are more prominent at lower feed concentrations, while the effects of feed concentration were more significant at higher TMPs. For instance, the enhancement of TMP from 1 to 2 bar at a low feed concentration (600 ppm) resulted in a 1.2% reduction in the rejection response, while at a high feed concentration (6,000 ppm) a 0.4% decrease was obtained, and a greater increase in the response (3.2%) was also observed by increasing the feed concentration at TMP 2 bar. This behavior confirms the effectiveness of forming a thicker concentration polarization layer on the membrane.
Figure 11

Rejection surface plots as a function of pressure and feed concentration.

Figure 11

Rejection surface plots as a function of pressure and feed concentration.

Close modal
Figure 12

Rejection surface plots as a function of feed flow rate and feed concentration.

Figure 12

Rejection surface plots as a function of feed flow rate and feed concentration.

Close modal

Figure 12 shows the considerable effects of QL at lower feed concentration and also feed concentration at lower QL due to the influence of the reducing effect of the deposition layer, which improves membrane rejection efficiency at higher QLs.

Thus, the formation of a concentration polarization layer on the membrane not only affects the flux reduction but also has a slight impact on the rejection.

Overall, the best performance of the modified MF procedure in the case of flux increment was obtained by increasing the TMP and feed flow rate to 2 bar and 3.5 L/min, respectively, and by employing 0.5 g/L coagulant to lower the COD of the feed to 600 ppm. Although the results acquired from the COD removal showed a different performance, they could be easily neglected compared to the great importance of flux increment (Fouladitajar & Ashtiani 2015).

Comparison of PES and PES/SiO2-g-PMAA membrane performance in separating pretreated and initial synthetic oily wastewater

To investigate the effect of membrane modification on flux change, an experiment was conducted to compare the performance of the PES and PES/SiO2-g-PMAA membrane for both pretreated and non-pretreated wastewater under optimal operating conditions (Table 8).

Table 8

PES and PES/SiO2-g-PMAA membrane performance in treating oily wastewater

MembraneFeed concentrationPressure (bar)Feed flow rate (L/min)Permeate flux (L/m2·h)Rejection (%)
PES Pretreated (600 ppm) 3.5 72.14 94.7 
PES Initial (6,000 ppm) 3.5 30.08 97.2 
PES/SiO2-g-PMAA Pretreated (600 ppm) 3.5 108.21 95.1 
PES/SiO2-g-PMAA Initial (6,000 ppm) 3.5 40.39 98.4 
MembraneFeed concentrationPressure (bar)Feed flow rate (L/min)Permeate flux (L/m2·h)Rejection (%)
PES Pretreated (600 ppm) 3.5 72.14 94.7 
PES Initial (6,000 ppm) 3.5 30.08 97.2 
PES/SiO2-g-PMAA Pretreated (600 ppm) 3.5 108.21 95.1 
PES/SiO2-g-PMAA Initial (6,000 ppm) 3.5 40.39 98.4 

According to Table 8, it can be stated that the membrane modification contributed to a 33.3 and 25.5% increase in the permeate flux of the pretreated oily wastewater and the initial one, respectively. The chief reasons for this improvement were the equal dispersion of nanoparticles in the polymer matrix and the reduction of their agglomeration, which led to a better interaction of nanoparticles with PES and consequently to an improved membrane performance (Khodadousti 2019). The table also shows the prominent impact of pretreatment on the permeate flux amelioration. This means that both membrane modification and pretreatment play a major role in improving the performance of the membrane process under suitable operating conditions.

To focus on improving the filtration efficiency of the microfiltration process, the combined effects of feed pretreatment, membrane hydrophobicity modification, and the optimization of operating parameters were investigated. For this purpose, two types of coagulants, polyaluminium chloride, and ferric chloride at different concentrations and pH values, as well as activated carbon as an adsorbent for the pretreatment process, were used. Since polyestersulphon has lower hydrophilicity and needs to be modified before use, SiO2 nanoparticles were employed. To prevent the nanoparticles from accumulating in the membrane, a modification step was performed using the radical polymerization method to bind polyaminoacrylic acid on the surface of the nanoparticles. The effect of varying the operating parameters on the membrane performance was also investigated using the RSM method. The main conclusions can be summarized as follows:

  • • The polyaluminium chloride showed better performance as a pretreatment method than ferric chloride and activated carbon at an optimal dosage of 1 g/L and pH 7.5, and resulted in the 96% removal of COD. Therefore, this approach was selected as the best pretreatment method.

  • • Changing the operating parameters during the microfiltration process showed a significant effect on improving the flux by increasing the TMP and the feed flow rate up to 2 bar and 3.5 L/min, respectively. While this increment in pressure and the feed flow rate had a regressive effect on membrane COD removal, an approximate 3% reduction in the quality of the permeate was completely negligible in comparison with the great flux improvement.

  • • An increase of 72.2% in the flux of the membrane and pretreated wastewater compared to the initial wastewater without pretreatment and the pure PES membrane illustrates the great positive impact of this combined procedure on better membrane performance.

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

The authors declare there is no conflict.

Abadi
S. R. H.
,
Sebzari
M. R.
,
Hemati
M.
,
Rekabdar
F.
&
Mohammadi
T.
2011
Ceramic membrane performance in microfiltration of oily wastewater
.
Desalination
265
,
222
228
.
doi:10.1016/j.desal.2010.07.055
.
Abbasi
M.
,
Reza Sebzari
M.
&
Mohammadi
T.
2012
Effect of metallic coagulent agents on oily wastewater treatment performance using mullite ceramic MF membranes
.
Sep. Sci. Technol.
47
,
2290
2298
.
doi:10.1016/S1004-9541(13)60617-5
.
Abdessemed
D.
&
Nezzal
G.
2005
Tertiary treatment of a secondary effluent by the coupling of coagulation-adsorption-ultrafiltration for reuse
.
Desalination
175
,
135
141
.
doi:10.1016/j.desal.2004.09.026
.
Ahmad
A. L.
,
Sumathi
S.
&
Hameed
B. H.
2006
Coagulation of residue oil and suspended solid in palm oil mill effluent by chitosan, alum and PAC
.
Chem. Eng. J.
118
,
99
105
.
doi:10.1016/j.cej.2006.02.001
.
Ang
W. L.
,
Mohammad
A. W.
,
Benamor
A.
&
Hilal
N.
2016
Chitosan as natural coagulant in hybrid coagulation-nano filtration membrane process for water treatment
. Journal of Environmental Chemical Engineering 4 (4), 4857–4862.
doi:10.1016/j.jece.2016.03.029
.
Armand
S. B.
,
Fouladitajar
A.
,
Ashtiani
F. Z.
&
Karimi
M.
2017
The effect of polymer concentration on electrospun PVDF membranes for desalination by direct contact membrane distillation
. European Water 58, 21–26.
Badrnezhad
R.
&
Beni
A. H.
2013
Ultrafiltration membrane process for produced water treatment: experimental and modeling
.
J. Water Reuse Desalin.
3
,
249
259
.
doi:10.2166/wrd.2013.092
.
Barbot
E.
,
Moustier
S.
,
Bottero
J. Y.
&
Moulin
P.
2008
Coagulation and ultrafiltration: understanding of the key parameters of the hybrid process
.
J. Membr. Sci.
325
,
520
527
.
doi:10.1016/j.memsci.2008.07.054
.
Boshrouyeh
M.
,
Tohid
G.
,
Farzin
T.
&
Ashtiani
Z.
2018
Experimental investigation and mathematical modeling of nano-composite membrane fabrication process: focus on the role of solvent type
. Asia Pacific Journal of Chemical Engineering 13 (6), e2260.
doi:10.1002/apj.2260
.
Chakrabarty
B.
,
Ghoshal
A. K.
&
Purkait
M. K.
2008
Ultrafiltration of stable oil-in-water emulsion by polysulfone membrane
.
J. Membr. Sci.
325
,
427
437
.
doi:10.1016/j.memsci.2008.08.007
.
Chang
Q.
,
Wang
X.
,
Wang
Y.
,
Zhang
X.
,
Cerneaux
S.
&
Zhou
J.
2015
Effect of hydrophilic modification with nano-titania and operation modes on the oil–water separation performance of microfiltration membrane
.
Desalin. Water Treat.
1
8
.
doi:10.1080/19443994.2014.1002010
.
Fabris
R.
,
Lee
E. K.
,
Chow
C. W. K.
,
Chen
V.
&
Drikas
M.
2007
Pre-treatments to reduce fouling of low pressure micro-filtration (MF) membranes
.
J. Membr. Sci.
289
,
231
240
.
doi:10.1016/j.memsci.2006.12.003
.
Fouladi Tajar
A.
,
Kaghazchi
T.
&
Soleimani
M.
2009
Adsorption of cadmium from aqueous solutions on sulfurized activated carbon prepared from nut shells
.
J. Hazard. Mater.
165
,
1159
1164
.
doi:10.1016/J.JHAZMAT.2008.10.131
.
Fouladitajar
A.
,
Ashtiani
F. Z.
,
Okhovat
A.
&
Dabir
B.
2013
Membrane fouling in micro filtration of oil-in-water emulsions: a comparison between constant pressure blocking laws and genetic programming (GP) model
.
Desalination
329
,
41
49
.
doi:10.1016/j.desal.2013.09.003
.
Fouladitajar
A.
,
Zokaee Ashtiani
F.
,
Valizadeh
B.
,
Armand
S. B.
&
Izadi
R.
2016
Application of gas/liquid two-phase flow in cross-flow microfiltration of oil-in-water emulsion: permeate flux and fouling mechanism analysis
.
Desalin. Water Treat.
57
,
4476
4486
.
doi:10.1080/19443994.2014.992048
.
Gao
B.
,
Li
D.
&
Lei
Q.
2011
Preparation of high PMMA grafted particle SiO2 using surface initiated free radical polymerization
.
J. Polym. Res.
18
,
1519
1526
.
doi:10.1007/s10965-010-9557-3
.
Ghandashtani
M. B.
,
Zokaee Ashtiani
F.
,
Karimi
M.
&
Fouladitajar
A.
2015
A novel approach to fabricate high performance nano-SiO2 embedded PES membranes for microfiltration of oil-in-water emulsion
.
Appl. Surf. Sci.
349
,
393
402
.
doi:10.1016/j.apsusc.2015.05.037
.
Guo
W. S.
,
Vigneswaran
S.
,
Ngo
H. H.
&
Chapman
H.
2004
Experimental investigation of adsorption-flocculation-microfiltration hybrid system in wastewater reuse
.
J. Membr. Sci.
242
,
27
35
.
doi:10.1016/j.memsci.2003.06.006
.
Hashlamon
A.
,
Mohammad
A. W.
&
Ahmad
A.
2017
The effect of wastewater pretreatment on nano filtration membrane performance Ali Hashlamon
.
J. Water Reuse Desalin.
45
52
.
doi:10.2166/wrd.2016.083
.
Hua
F. L.
,
Tsang
Y. F.
,
Wang
Y. J.
,
Chan
S. Y.
,
Chua
H.
&
Sin
S. N.
2007
Performance study of ceramic microfiltration membrane for oily wastewater treatment
.
Chem. Eng. J.
128
,
169
175
.
doi:10.1016/j.cej.2006.10.017
.
Kehrein
P.
,
Jafari
M.
,
Slagt
M.
,
Cornelissen
E.
,
Osseweijer
P.
&
Posada
J.
2021
A techno-economic analysis of membrane-based advanced treatment processes for the reuse of municipal wastewater
.
11
,
705
725
.
doi:10.2166/wrd.2021.016
.
Khodadousti
S.
2019
Preparation and characterization of novel PES-(SiO2-g-PMAA) membranes with antifouling and hydrophilic properties for separation of oil-in-water emulsions
. Polymers for Advanced Technologies 30 (9),
1
12
.
doi:10.1002/pat.4651
.
Lee
B. B.
,
Choo
K. H.
,
Chang
D.
&
Choi
S. J.
2009
Optimizing the coagulant dose to control membrane fouling in combined coagulation/ultrafiltration systems for textile wastewater reclamation
.
Chem. Eng. J.
155
,
101
107
.
doi:10.1016/j.cej.2009.07.014
.
Lerch
A.
,
Panglisch
S.
,
Buchta
P.
,
Tomita
Y.
,
Yonekawa
H.
,
Hattori
K.
&
Gimbel
R.
2005
Direct river water treatment using coagulation/ceramic membrane microfiltration
.
Desalination.
179
,
41
50
.
doi:10.1016/j.desal.2004.11.054
.
Lobo
A.
,
Cambiella
Á.
,
Benito
J. M.
,
Pazos
C.
&
Coca
J.
2006
Ultrafiltration of oil-in-water emulsions with ceramic membranes: influence of pH and crossflow velocity
.
J. Membr. Sci.
278
,
328
334
.
doi:10.1016/j.memsci.2005.11.016
.
Martín
M. A.
,
González
I.
,
Berrios
M.
,
Siles
J. A.
&
Martín
A.
2011
Optimization of coagulation-flocculation process for wastewater derived from sauce manufacturing using factorial design of experiments
.
Chem. Eng. J.
172
,
771
782
.
doi:10.1016/j.cej.2011.06.060
.
Mikhak
Y.
,
Mohammad
M.
,
Torabi
A.
&
Fouladitajar
A.
2019
Refinery and Petrochemical Wastewater Treatment
.
Elsevier Inc
.
doi:10.1016/B978-0-12-816170-8.00003-X
.
Mohammadi
T.
&
Esmaeelifar
A.
2005
Wastewater treatment of a vegetable oil factory by a hybrid ultrafiltration-activated carbon process
.
J. Membr. Sci.
254
,
129
137
.
doi:10.1016/j.memsci.2004.12.037
.
Murić
A.
,
Petrinić
I.
&
Christensen
M. L.
2014
Comparison of ceramic and polymeric ultrafiltration membranes for treating wastewater from metalworking industry
.
Chem. Eng. J.
255
,
403
410
.
doi:10.1016/j.cej.2014.06.009
.
Ren
L.
,
Liu
C.
&
Meng
T.
2021
Effects of micro-flocculation pretreatment on the ultra-filtration membrane fouling caused by different dissolved organic matters in treated wastewater
.
11
,
597
609
.
doi:10.2166/wrd.2021.051
.
Rezaei
H.
,
Ashtiani
F. Z.
&
Fouladitajar
A.
2011
Effects of operating parameters on fouling mechanism and membrane flux in cross-flow microfiltration of whey
.
Desalination
274
,
262
271
.
doi:10.1016/j.desal.2011.02.015
.
Salahi
A.
,
Gheshlaghi
A.
,
Mohammadi
T.
&
Madaeni
S. S.
2010
Experimental performance evaluation of polymeric membranes for treatment of an industrial oily wastewater
.
Desalination
262
,
235
242
.
doi:10.1016/j.desal.2010.06.021
.
Sarfaraz
M. V.
,
Ahmadpour
E.
,
Salahi
A.
,
Rekabdar
F.
&
Mirza
B.
2012
Experimental investigation and modeling hybrid nano-porous membrane process for industrial oily wastewater treatment
.
Chem. Eng. Res. Des.
90
,
1642
1651
.
doi:10.1016/j.cherd.2012.02.009
.
Shang
X.
,
Kim
H. C.
,
Huang
J. H.
&
Dempsey
B. A.
2015
Coagulation strategies to decrease fouling and increase critical flux and contaminant removal in microfiltration of laundry wastewater
.
Sep. Purif. Technol.
147
,
44
50
.
doi:10.1016/j.seppur.2015.04.005
.
Sun
W.
,
Liu
J.
,
Chu
H.
&
Dong
B.
2013
Pretreatment and membrane hydrophilic modification to reduce membrane fouling
.
Membranes (Basel)
3
,
226
241
.
doi:10.3390/membranes3030226
.
Tashvigh
A. A.
,
Ashtiani
F. Z.
&
Fouladitajar
A.
2015
Genetic programming for modeling and optimization of gas sparging assisted microfiltration of oil-in-water emulsion
.
3994
.
doi:10.1080/19443994.2015.1096830
.
Valizadeh
B.
,
Zokaee
F.
,
Fouladitajar
A.
,
Dabir
B.
,
Seyed
S.
,
Baraghani
M.
,
Borhan
S.
,
Salari
B.
&
Kouchakiniya
N.
2015
Scale-up economic assessment and experimental analysis of MF–RO integrated membrane systems in oily wastewater treatment plants for reuse application
.
Desalination
374
,
31
37
.
doi:10.1016/j.desal.2015.07.017
.
Vatanpour
V.
,
Madaeni
S. S.
,
Khataee
A. R.
,
Salehi
E.
,
Zinadini
S.
&
Monfared
H. A.
2012
TiO2 embedded mixed matrix PES nanocomposite membranes: influence of different sizes and types of nanoparticles on antifouling and performance
.
Desalination
292
,
19
29
.
doi:10.1016/j.desal.2012.02.006
.
Wu
G.
,
Gan
S.
,
Cui
L.
&
Xu
Y.
2008
Preparation and characterization of PES/TiO2 composite membranes
.
Appl. Surf. Sci.
254
,
7080
7086
.
doi:10.1016/j.apsusc.2008.05.221
.
Wu
B.
,
An
Y.
,
Li
Y.
&
Wong
F. S.
2009
Effect of adsorption/coagulation on membrane fouling in microfiltration process post-treating anaerobic digestion effluent
.
Desalination
242
,
183
192
.
doi:10.1016/j.desal.2008.04.005
.
Yan
L.
,
Hong
S.
,
Li
M. L.
&
Li
Y. S.
2009
Application of the Al2O3-PVDF nanocomposite tubular ultrafiltration (UF) membrane for oily wastewater treatment and its antifouling research
.
Sep. Purif. Technol.
66
,
347
352
.
doi:10.1016/j.seppur.2008.12.015
.
Yang
Y.
,
Chen
R.
&
Xing
W.
2011
Integration of ceramic membrane microfiltration with powdered activated carbon for advanced treatment of oil-in-water emulsion
.
Sep. Purif. Technol.
76
,
373
377
.
doi:10.1016/j.seppur.2010.11.008
.
Yu
T.
,
Xu
C.
,
Chen
F.
,
Yin
H.
,
Sun
H.
,
Cheng
L.
&
Bi
X.
2021
Microcoagulation improved the performance of the UF–RO system treating the effluent from a coastal municipal wastewater treatment plant: a pilot-scale study
.
J. Water Reuse Desalin.
11
(
2
),
177
188
.
doi:10.2166/wrd.2021.099
.
Zhou
G.
,
Fung
K. K.
,
Wong
L. W.
,
Chen
Y.
,
Renneberg
R.
&
Yang
S.
2011
Immobilization of glucose oxidase on rod-like and vesicle-like mesoporous silica for enhancing current responses of glucose biosensors
.
Talanta
84
,
659
665
.
doi:10.1016/j.talanta.2011.01.058
.
Zhu
L. P. J.
,
Zhu
L. P. J.
,
Jiang
J. H.
,
Yi
Z.
,
Zhao
Y. F.
,
Zhu
B. K.
&
Xu
Y. Y.
2014
Hydrophilic and anti-fouling polyethersulfone ultrafiltration membranes with poly(2-hydroxyethyl methacrylate) grafted silica nanoparticles as additive
.
J. Membr. Sci.
451
,
157
168
.
doi:10.1016/j.memsci.2013.09.053
.
Zouboulis
A. I.
&
Avranas
A.
2000
Treatment of oil-in-water emulsions by coagulation and dissolved-air flotation
.
Colloids Surf. A Physicochem. Eng. Asp.
172
,
153
161
.
doi:10.1016/S0927-7757(00)00561-6
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).

Supplementary data