Oily Wastewater Treatment Using Polyamide Thin Film Composite Membrane Technology

In this study, polyamide (PA) thin film composite (TFC) reverse osmosis (RO) membrane filtration was used in edible oil wastewater emulsion treatment. The PA-TFC membrane was characterized using mechanical, thermal, chemical, and physical tests. Surface morphology and cross-sections of TFCs were characterized using SEM. The effects of edible oil concentrations, average droplets size, and contact angle on separation efficiency and flux were studied in detail. Purification performance was enhanced using activated carbon as a pre-treatment unit. The performance of the RO unit was assessed by chemical oxygen demand (COD) removal and permeate flux. Oil concentration in wastewater varied between 3000 mg/L and 6000 mg/L. Oily wastewater showed a higher contact angle (62.9°) than de-ionized water (33°). Experimental results showed that the presence of activated carbon increases the permeation COD removal from 94% to 99%. The RO membrane filtration coupled with an activated carbon unit of oily wastewater is a convenient hybrid technique for removal of high-concentration edible oil wastewater emulsion up to 99%. Using activated carbon as an adsorption pre-treatment unit improved the permeate flux from 34 L/m2hr to 75 L/m2hr.


Introduction
The increased need for edible oil all over the world has resulted in the development of many edible oil plants, leading to the disposal of huge amounts of wastewater. Palm, olive, soybean, cottonseed, and sunflower are considered the main sources for extracting edible oil [1]. Persistent market research (PMR) has reported that the global oil market value is predicted to increase to 130.3 billion US$ by 2024, recording a compound annual growth rate of 5.1%. In 2018, global consumption of vegetable oils was wastewater [37]. Azmi et al. studied a UF and RO hybrid membrane system on palm oil/water emulsion. Removal of BOD and turbidity up to 98.9% and 99%, respectively, were obtained [38]. In another study, the same research team used sand filtration and coagulation/flocculation before the membrane unit for oily wastewater treatment [39]. A remarkable reduction in COD (95%) and BOD (95%) was recorded. In this study, an activated carbon adsorption unit was used for pretreatment. The present investigation aimed to identify the optimum conditions allowed for the permeate stream suitable for recycling or to be discharged into water sewage with no risk of toxicity to the ecosystem. Coupling of the adsorption unit to RO was investigated in a hybrid mode. The treatment processes were compared based on their efficiency in COD and turbidity removals. It was concluded that this coupling of treatment units was able to purify edible oil wastewater emulsion and treated water to meet standards.

Materials and Methods
Polyoxyethylene sorbatin monolaurate "Tween 20" (C 58 H 11 O 26 , MWt = 1227.72 g/mol) was used as a surfactant. Oily wastewater was synthesized by mixing 5 mLof soybean oil and 1 L of distilled water in the presence 1 mL of Tween 20. The oily wastewater emulsion was mixed mechanically for 3.5 h at a mixing rate of 300 rpm. The stability of the oily wastewater emulsion was observed after 24 h. The mixture showed a uniform and stable white emulsion. The PA-TFC RO membrane in a spiral wound structure was investigated. The Membrane model (TW 30-1812-75) was installed-length: 0.26; diameter: 0.05 m; maximum service temperature and pressure: 45 • C and 1.034 MPa, respectively; applied pressure: 0.86 MPa; and allowable range of pH: 2 to 10. The pores' size ranged from 0.67 nm to 0.78 nm [40]. The maximum flow rate of was 0.0076 m 3 /min. The granular activated carbon (GAC) filter was installed as an adsorption pretreatment unit. The column dimensions were 63.5 mm and (height) 254 mm. The GAC filter was fixed to reduce total suspended solids and chemicals resulting in undesired tastes and odors and volatile organic chemicals (VOCs). The GAC is capable of delivering up to 0.0038 m 3 per minute of treated water. A high-pressure pump (PKM60) with dimensions 0.0254 m × 0.0254 m, Q max "0.04 m 3 /min", and "2850 rpm" was installed to pump the water-oil emulsion across the membrane surface. An analysis of oil wastewater and treated water was conducted in terms of turbidity, oil and grease, and COD.

Chemical Oxygen Demand (COD)
Chemical oxygen demand for all oily wastewater samples (feed, permeate, and rejection) was measured by COD photometer (LaMotte, Chestertown, MD, USA). Sample (0.2 mL) was added into digested solution (High range: 200-15,000 mg/L) consisting of a strong oxidizing agent (potassium dichromate), mercuric sulfate, and sulfuric acid. The water mixture was then heated for 120 min in a COD heater at 150 • C. By the end of oxidation, the COD was measured using a spectrophotometer.

Turbidity
The samples of feed, permeate, and rejection were analyzed for total suspended solids in terms of turbidity using a turbidity meter (Lovibond meter, Lovibond company, Sarasota, FL, USA) expressed as nephelometric turbidity units (NTUs).

Thermogravimetric Analysis (TGA)
Thermal stability of polyester support, polysulfone, and polyamide layers were characterized using TGA (TGA Q500 V20.10 Build 36, TA Company, Tokyo, Japan). Thermogravimetric analysis is considered a proper method to quantitatively examine the thermal stability of polymeric membrane materials. The thin film composite RO membrane was freeze-fractured and gold coated using an S150A SPUTTER COATER SEM (Edwards High Vacuum = 7.5 mbar). Surface morphology and cross-sections of TFC membranes were investigated using scanning electron microscopy (SEM, Quanta 250 FEG, FEI Company, Tokyo, Japan).

Mechanical Testing
The strength of the RO membrane was measured using a (Tinius Olsen H25KS, Shanghai, China) tensile instrument by providing load coaxial style.

Average oil Particle Size
The size distribution of oil droplets in the oily wastewater feed was determined using a laser particle size analyzer (Nicomp Nano size, 380 ZLS, Santabarbara, CA, USA).

Dynamic Wetting
Each sample was cut to a size of 10 mm × 50 mm with sharp scissors. The TFC-PA specimen was immersed into oily wastewater and water samples. The weight of the absorbed water was measured. The dynamic water absorption was plotted versus time.

Contact Angle
To study the effect of edible oil on the hydrophobicity of the PA-THF RO membrane, the surface contact angle of the membrane was measured based on the sessile drop method via a contact angle measuring instrument (OCA 15EC contact angle model, GmbH, Filderstadt, Germany). The reported results were the average contact angle of deionized water droplets and oily wastewater at five different locations on each sample.

Oil and Grease
The oil concentration was determined using the hexane extraction gravimetric method. One hundred and twenty-five milliliters of oily wastewater samples were added to ethanol (20 mL) and sulfuric acid (20 mL). Mixtures were well shaken in a separation funnel. After fifteen minutes, two layers were separated, and the supernatant layer containing the oil (organic layer) was collected. The weight (wt) of the beaker before and after separation was recorded. Then, the oil concentration was calculated as given in Equation (1). (1)

Experimental Setup and Operation
The tank, made from acrylic for mixing the water-oil emulsion, is shown in Figure 1. The tank's dimensions were 0.4 m in diameter and 0.7 m in height with a total volume of 0.05 m 3 and a rotational speed of 300 rpm for the stainless-steel mixer.  Figure 2a shows the schematic diagram of the PA-TFC RO membrane filtration unit and the facilities. The membrane was washed using de-ionized water for fifteen minutes before the treatment process to ensure that the RO had reached maximum force. The synthetic wastewater/oil emulsion was fed in mechanically stirred tanks, 0.04 m 3 , in the presence of surfactant (Tween 20). The oily wastewater was pumped to the RO unit under a constant pressure of 8.5 bar. The oil concentration in wastewater varied from 3 g/L to 6 g/L. The treated water was collected from the RO unit, determined by collected permeate water in vessel. The permeate flow rate was measured regularly to calculate flux. The permeate, rejected, and feed water contents were tested for COD, oil and grease, and turbidity. The rejected wastewater stream was analyzed to determine oil content in the oily wastewater. To control the membrane fouling, a pre-treatment adsorption unit of activated carbon was installed. In the feed pretreatment unit, adsorption GAC was applied to remove suspended solids and organic constituents. The pretreatment process plays a crucial role for treatment process of high concentration edible oil wastewater in oil industries that would block the membrane and reduce its lifetime. A process flow diagram (PFD) of the pre-adsorption unit and membrane filtration process are shown in Figure 2b. After oily wastewater treatment using the hybrid system (i.e., GAC and RO membrane), the treated water sample was tested for possible irrigation or discharge into the sewage network. The oil contents were determined in the rejected stream for recycling highly concentrated oily wastewater streams in the production of soap as a step towards a zero discharge system [41]. Turbidity and COD were measured in each process to study the efficiency of oil removal.  Figure 2a shows the schematic diagram of the PA-TFC RO membrane filtration unit and the facilities. The membrane was washed using de-ionized water for fifteen minutes before the treatment process to ensure that the RO had reached maximum force. The synthetic wastewater/oil emulsion was fed in mechanically stirred tanks, 0.04 m 3 , in the presence of surfactant (Tween 20). The oily wastewater was pumped to the RO unit under a constant pressure of 8.5 bar. The oil concentration in wastewater varied from 3 g/L to 6 g/L. The treated water was collected from the RO unit, determined by collected permeate water in vessel. The permeate flow rate was measured regularly to calculate flux. The permeate, rejected, and feed water contents were tested for COD, oil and grease, and turbidity. The rejected wastewater stream was analyzed to determine oil content in the oily wastewater. To control the membrane fouling, a pre-treatment adsorption unit of activated carbon was installed. In the feed pretreatment unit, adsorption GAC was applied to remove suspended solids and organic constituents. The pretreatment process plays a crucial role for treatment process of high concentration edible oil wastewater in oil industries that would block the membrane and reduce its lifetime. A process flow diagram (PFD) of the pre-adsorption unit and membrane filtration process are shown in Figure 2b. After oily wastewater treatment using the hybrid system (i.e., GAC and RO membrane), the treated water sample was tested for possible irrigation or discharge into the sewage network. The oil contents were determined in the rejected stream for recycling highly concentrated oily wastewater streams in the production of soap as a step towards a zero discharge system [41]. Turbidity and COD were measured in each process to study the efficiency of oil removal.

Percentage oil rejection and flux
In the RO membrane unit, the separation efficiency was determined in terms of rejection (%) of COD, turbidity or oil and grease as given in Equation (2) [5]: The CP is the concentration of oil in permeate stream; CF is the oil concentration in the feed. The water flux (Jω) is the volume of permeate (V) that is collected per unit membrane surface area (a) per unit time (t). Water flux is determined as given in Equation (3):

Percentage Oil Rejection and Flux
In the RO membrane unit, the separation efficiency was determined in terms of rejection (%) of COD, turbidity or oil and grease as given in Equation (2) [5]: The CP is the concentration of oil in permeate stream; CF is the oil concentration in the feed. The water flux (Jω) is the volume of permeate (V) that is collected per unit membrane surface area (a) per unit time (t). Water flux is determined as given in Equation (3): 7 of 17

Surface Morphology
Polyamide thin film composite contains three layers of thin film PA, polysulfone, and a supported layer of polyester. The SEM images of the TFC membrane were investigated before and after oily wastewater treatment to analyze microscopic structure of the membrane. Cross-sectional images of the TFC were taken after the cutting with freeze-fracturing. Figure 3a shows cross-sectional images which show the RO polyamide membrane consisting of three layers: a thin selective layer on the upper surface (0.22 µm), a microporous interlayer (about 17.21 µm), and polyester acting as structural support (92.61 µm). The SEM micrographs of pristine membrane are shown in Figure 3b. Membrane surface morphology after wastewater treatment is shown in Figure 3c. It was found that oil droplets absorbed on the surface of the TFC membrane. In the absence of the GAC unit, most of the membrane pores were blocked, resulting in a decrease in membrane flux. Zhao et al. (2015) reported similar results on polyvinylchloride ultrafiltration membranes blocking in oily wastewater treatment [42]. Polyamide thin film composite contains three layers of thin film PA, polysulfone, and a supported layer of polyester. The SEM images of the TFC membrane were investigated before and after oily wastewater treatment to analyze microscopic structure of the membrane. Cross-sectional images of the TFC were taken after the cutting with freeze-fracturing. Figure 3a shows cross-sectional images which show the RO polyamide membrane consisting of three layers: a thin selective layer on the upper surface (0.22 µm), a microporous interlayer (about 17.21 µm), and polyester acting as structural support (92.61 µm). The SEM micrographs of pristine membrane are shown in Figure 3b. Membrane surface morphology after wastewater treatment is shown in Figure 3c. It was found that oil droplets absorbed on the surface of the TFC membrane. In the absence of the GAC unit, most of the membrane pores were blocked, resulting in a decrease in membrane flux. Zhao et al. (2015) reported similar results on polyvinylchloride ultrafiltration membranes blocking in oily wastewater treatment [42].

Contact Angle
Hydrophilicity is one of the most crucial parameters that influences the permeability and antifouling process of the membrane surface. A low contact angle refers to the high hydrophilicity of

Contact Angle
Hydrophilicity is one of the most crucial parameters that influences the permeability and antifouling process of the membrane surface. A low contact angle refers to the high hydrophilicity of the surface. Contact angles of the TFC membrane were measured using advancing contact angle methods. As shown in Figure 4, the contact angle of water on the surface was 33 • , while the average contact angle of oily wastewater on the membrane surface was 62.9 • . These results suggest that the membrane in water was more hydrophilic than oil.
Membranes 2020, 10, x FOR PEER REVIEW 8 of 17 8 the surface. Contact angles of the TFC membrane were measured using advancing contact angle methods. As shown in Figure 4, the contact angle of water on the surface was 33°, while the average contact angle of oily wastewater on the membrane surface was 62.9°. These results suggest that the membrane in water was more hydrophilic than oil.  Figure 5 shows the results of the stress-strain behavior of the TFC membrane. The tensile strength was found to be 52 MPa, and the Young's modulus was 1000 MPa. The mechanical testing results are listed in Table 1.    Figure 5 shows the results of the stress-strain behavior of the TFC membrane. The tensile strength was found to be 52 MPa, and the Young's modulus was 1000 MPa. The mechanical testing results are listed in Table 1. the surface. Contact angles of the TFC membrane were measured using advancing contact angle methods. As shown in Figure 4, the contact angle of water on the surface was 33°, while the average contact angle of oily wastewater on the membrane surface was 62.9°. These results suggest that the membrane in water was more hydrophilic than oil.  Figure 5 shows the results of the stress-strain behavior of the TFC membrane. The tensile strength was found to be 52 MPa, and the Young's modulus was 1000 MPa. The mechanical testing results are listed in Table 1.    Thermogravimetric analysis (TGA) of the reverse osmosis membrane was performed under nitrogen atmosphere as shown in Figure 6. Four weight-loss stages can be observed in the TGA curve. Weight loss at an average temperature 100 • C can be attributed to water loss. The second stage at an average temperature 304-551 • C refers to the thermal degradation range of the supported polyester layer. The third stage at an average temperature 450-550 • C can be attributed to polyamide loss. The fourth stage shows the weight loss of the polysulfone layer, starting from 590 • C. Thermogravimetric analysis (TGA) of the reverse osmosis membrane was performed under nitrogen atmosphere as shown in Figure 6. Four weight-loss stages can be observed in the TGA curve. Weight loss at an average temperature 100°C can be attributed to water loss. The second stage at an average temperature 304-551°C refers to the thermal degradation range of the supported polyester layer. The third stage at an average temperature 450-550°C can be attributed to polyamide loss. The fourth stage shows the weight loss of the polysulfone layer, starting from 590°C.

Fourier Transform-Infrared Spectroscopy FTIR
The FTIR was used to analyze the chemical composition of the top surface of membrane (PA), and PSF and polyester-supported layer. The infrared spectra of TFC membrane showed the main characteristic peaks of PA layer and polysulfone-supporting layer as shown in Figure 7 and listed in Table 2.

Fourier Transform-Infrared Spectroscopy FTIR
The FTIR was used to analyze the chemical composition of the top surface of membrane (PA), and PSF and polyester-supported layer. The infrared spectra of TFC membrane showed the main characteristic peaks of PA layer and polysulfone-supporting layer as shown in Figure 7 and listed in Table 2.

Effect of Oil Particle Size on Separation Efficiency
Oil particles that are captured on the PA-TFC RO membrane surface affect separation efficiency across membrane filtration. The PA-TFC RO membrane's capacity may vary depending on the particle size variation in the oily wastewater as shown in Figure 8. The distribution of oil droplets' size was measured for six samples of emulsion to study the effect of oil particle size on the separation efficiency. It was found that the separation performance improves as the oil particles' sizes increase. This can be attributed to the coalescence of oil particle droplets where these are retained on thin film membrane. Thus, rejection efficiency increased from 94% to 99% with the increase of oil concentration as well as the average droplet size. Similar trends were observed by Chakrabarty et al. [43] and Mittal et al. [44].

Effect of Oil Particle Size on Separation Efficiency
Oil particles that are captured on the PA-TFC RO membrane surface affect separation efficiency across membrane filtration. The PA-TFC RO membrane's capacity may vary depending on the particle size variation in the oily wastewater as shown in Figure 8. The distribution of oil droplets' size was measured for six samples of emulsion to study the effect of oil particle size on the separation efficiency. It was found that the separation performance improves as the oil particles' sizes increase. This can be attributed to the coalescence of oil particle droplets where these are retained on thin film membrane. Thus, rejection efficiency increased from 94% to 99% with the increase of oil concentration as well as the average droplet size. Similar trends were observed by Chakrabarty et al. [43] and Mittal et al. [44].

Dynamic Wetting
All membranes demonstrated water absorption curves that rose sharply within 20 s. Further immersion occurred after 20 s up to 90 s, inducing a plateau in all curves. Early in the immersion period, instantaneous water absorption was observed for all membranes. After 20 s all curves plateaued, showing that all membranes could only absorb a certain number of water molecules. This observation confirms the hydrophilic behavior of the TFC membrane. The absorption curve of the TFC membrane can be seen in Figure 9. Oil contents considerably decreased the water absorption properties of the RO membrane. This can be attributed to the higher contact angles.

Dynamic Wetting
All membranes demonstrated water absorption curves that rose sharply within 20 s. Further immersion occurred after 20 s up to 90 s, inducing a plateau in all curves. Early in the immersion  period, instantaneous water absorption was observed for all membranes. After 20 s all curves plateaued, showing that all membranes could only absorb a certain number of water molecules. This observation confirms the hydrophilic behavior of the TFC membrane. The absorption curve of the TFC membrane can be seen in Figure 9. Oil contents considerably decreased the water absorption properties of the RO membrane. This can be attributed to the higher contact angles.

Effect of Time on Permeate Flux
The fluxes of permeate oily wastewater were taken every 50 min, starting from 5 min. Figure 10 shows that the initial flux in case of RO and RO and adsorption pretreatment hybrid system were 34 L/m 2 ·h and 75 L/m 2 ·h, respectively. The initial flux was the maximum value at constant pressure 8.5 bar. This phenomenon can be attributed to the lowest value of mass transfer resistance at the beginning. At the membrane surface, permeate flow rate reduces with time and concentration polarization increases. A decrease in flux was recorded as a result of increased mass transfer resistance, so fouling is often considered in the form of decreasing permeate flux over time [45][46][47][48][49]. Membrane blocking leads to a decrease in membrane performance because of suspended solids accumulating on the outer layer. This could be observed in the RO unit without an adsorption unit. Differences in fouling behavior between the oily wastewater treatment using RO and the RO hybrid system were expected as a result of the differences in the rate of fouling.

Effect of Time on Permeate Flux
The fluxes of permeate oily wastewater were taken every 50 min, starting from 5 min. Figure 10 shows that the initial flux in case of RO and RO and adsorption pretreatment hybrid system were 34 L/m 2 ·h and 75 L/m 2 ·h, respectively. The initial flux was the maximum value at constant pressure 8.5 bar. This phenomenon can be attributed to the lowest value of mass transfer resistance at the beginning. At the membrane surface, permeate flow rate reduces with time and concentration polarization increases. A decrease in flux was recorded as a result of increased mass transfer resistance, so fouling is often considered in the form of decreasing permeate flux over time [45][46][47][48][49]. Membrane blocking leads to a decrease in membrane performance because of suspended solids accumulating on the outer layer. This could be observed in the RO unit without an adsorption unit. Differences in fouling behavior between the oily wastewater treatment using RO and the RO hybrid system were expected as a result of the differences in the rate of fouling.

Effect of Oil Concentration on Separation Efficiency
The effect of the oil concentration on the separation performance was determined at a constant pressure of 0.85 MPa, as listed in Table 1, which describes the maximum concentration of suspended solid in terms of turbidity and COD removal. The results are presented in Figures 11 and 12, showing that the oil removal increases with an increase in feed concentration. Separation efficiency improves with the increase in oil concentration which can be attributed to the formation of larger oil droplets.

Effect of Oil Concentration on Separation Efficiency
The effect of the oil concentration on the separation performance was determined at a constant pressure of 0.85 MPa, as listed in Table 3, which describes the maximum concentration of suspended solid in terms of turbidity and COD removal. The results are presented in Figures 11 and 12, showing that the oil removal increases with an increase in feed concentration. Separation efficiency improves with the increase in oil concentration which can be attributed to the formation of larger oil droplets. Furthermore, oil droplets separated by the membrane form a gel layer on the membrane surface [43]. Thus, removal can be improved. However, the flux decreases in the absence of an adsorption pretreatment unit. The average removal of oil in terms of COD reached 97.4%. The performance of membrane showed high COD removal. The COD of the permeate stream was within the allowable levels (1100 ppm) for the discharge of wastewater into sewage networks.  Figure 11. Effect of feed concentration on permeate COD. Figure 11. Effect of feed concentration on permeate COD.
14 Figure 11. Effect of feed concentration on permeate COD.

Effect of Adsorption Pretreatment unit on Membrane Filtration Performance
Permeate flux decreased as a result of membrane fouling and concentration polarization. Using activated carbon pretreatment unit led to a decrease in suspended solids' load on membrane filtration. Oil removal improved to 99.9% in the presence of a pre-treatment unit as listed in Table 2. Oil concentration in terms of COD in the permeate stream was less than 100 mg/L which meets the national discharge standards of wastewater.

Effect of Adsorption Pretreatment unit on Membrane Filtration Performance
Permeate flux decreased as a result of membrane fouling and concentration polarization. Using activated carbon pretreatment unit led to a decrease in suspended solids' load on membrane filtration. Oil removal improved to 99.9% in the presence of a pre-treatment unit as listed in Table 4. Oil concentration in terms of COD in the permeate stream was less than 100 mg/L which meets the national discharge standards of wastewater.

Conclusions
The reverse osmosis thin film composite membrane showed good results in the treatment of oily wastewater produced from the edible oil industry. Removal of turbidity, COD, and oil removal at fixed flux was achieved using RO and an adsorption unit hybrid system. An activated carbon unit as a pretreatment unit showed better efficiency and higher permeate flux. Membrane filtration using a reverse osmosis membrane was conducted to remove edible oil and suspended solids from oily wastewater. An RO membrane was characterized using FTIR, SEM, contact angle, mechanical analysis, and TGA. The PA-TFC membrane showed a promising separation process for treatment of wastewater effluent from edible oil. Significant removal of COD and turbidity can be accomplished using RO. Higher flux and better membrane performance can be achieved using an RO membrane filtration coupled with an activated carbon unit. A concentration of oil up to 6000 mg/L can be effectively improved to 99.9%.