Polycaprolactone Blended Cellulose Acetate Thin Film Composite Membrane for Dairy Waste Treatment Using Forward Osmosis

In recent years new sustainable technology for wastewater treatment has emerged, and among them, Forward osmosis (FO) has gained importance. FO utilizes osmotic pressure difference across the semipermeable membrane as the driving force to concentrate the wastewater. Further, the surface and physical properties of the semipermeable FO membrane play a crucial role during the FO process to reduce the internal concentration polarization. In general, FO membranes are prepared using cellulose acetate (CA) polymer due to their high hydrophilic nature. However, CA membranes are mechanically unstable for the FO process. Hence, to increase the mechanical strength and exibility of CA, other polymers are blended along with it. In this present study, we have prepared a Phase-Inversion membrane using CA blended with polycaprolactone (PCL) polymers. Further, to increase the hydrophilicity of the membrane, a thin lm composite (TFC) layer of polyamide is coated using interfacial polymerization. To increase the antifouling properties of the membrane, Graphene oxide (GO) and copper oxide (CuO) nanoparticles (NPs) are incorporated inside the TFC matrix. The prepared NPs and membrane were characterized using Fourier-transform infrared spectroscopy (FTIR), Wide-angle X-ray scattering (WAXD), and contact angle. Further, the GO-CuO incorporated TFC coating has improved the hydrophilicity and antifouling properties of the membrane. It was observed that the water ux has increased up to 5 LMH, and reverse solute ux has reduced to 4 GMH. Further, the membrane was utilized to concentrate in situ prepared dairy waste. It was observed that after 60 mins of the FO process, the concentration of dairy waste has increased to 23%, with a concentration factor of 0.903. Thus, prepared TFC phase inversion membrane is potential for dairy wastewater treatment.


Introduction
The essential contribution to sustaining a healthy environment is water. Through works, it has been established that about two-thirds of the earth's surface is enclosed with water, out of which only 2.5% is fresh water and an unbiased 0.3% is accessible for human consumption. The increase in population, industrialization, and improved living standards has upsurged the demand for freshwater, thereby making people aware of its importance and vulnerability in terms of quality. This issue can be addressed by incorporating a proper wastewater treatment methodology. The traditional wastewater treatment approaches include ltration, coagulation-occulation, or biological treatment. However, among these various methods, the most acknowledged technique is Membrane technology. Membrane technology has shown signi cant progress in the last decades owing to its easy accessibility and environmental friendliness. Several membrane processes that have been tried earlier for treating wastewater are ultra ltration (UF), micro ltration (MF), nano ltration (NF), membrane distillation (MD), and reverse osmosis (RO). Amongst these, RO has received a greater appreciation for water reclamation due to its higher e ciency. This technology was anticipated to bridge the gap between economics and sustainability. (Obaid et al. 2016) Dairy processing is now one of the largest sources of industrial wastewater due to rapid industrialization.
Approximately 60-80 percent of the water consumed in the industry is contaminated. The production of several food items from raw milk resulted in the formation of a diverse run-off stream with qualitative and quanti able variations, which is restricted for further reuse. The primary characteristics of dairy wastewater are high temperature, high organic content, and a wide pH range. To avoid environmental damage, the wastewater from the dairy industries must be treated. Milk production results in the production of wastewater containing similar constituents. The type of treatment required is heavily in uenced by the composition and ow rate of the wastewater. Because of the high energy requirements and signi cant fouling of membranes, pressure-driven membrane processes cannot be used to treat this contaminated water. Also, this pressure-driven process entails a large energy requirement and is costly. A new energy-e cient technique, Forward Osmosis (FO), was introduced to encounter water stresses cautiously.
Forward Osmosis is an evolving energy-e cient membrane separation technology proposed in the 1970s.
It trails the natural phenomenon of osmosis, involving the passage of water crossways towards the membrane from the low concentration region, known as feed solution, to the high concentration region, known as draw solution. (Cath et al. 2006) This movement of water molecules across the membrane due to the osmotic pressure gradient between the two solutions. The main advantage of FO is that it does not require any external pressure, which aids in the minimal fouling of membranes in comparison to other pressure-driven processes that are already in use. The membrane used in the FO process a crucial role in water treatment. Cellulose acetate (CA) FO membrane was also frequently used until lately when Thin Film Composite (TFC) membranes took the FO platform showing an enhanced water ux performance along with better salt rejection. The currently used FO membranes are designed by forging a thin lm polymerization over the base membrane as a support layer. Optimization studies conducted on the TFC membranes reported that a base layer with a mixture of different morphologies greatly affected the membrane performance. A further insight revealed that though TFC membranes have superior selectivity, the membranes suffer serious internal concentration polarization (ICP), limiting their applications.
The synthesis of chemically stable high-performance FO membranes is still in its early phases of development. The unique advantages offered by CA make it worthwhile to investigate further into its improvement that might have been unnoticed in the past. New studies and advancements in the TFC membranes recommend producing a thin porous support layer for vindicating the effect of ICP, thereby sustaining the top dense selective layer for effective water ux. In search of a better strategy for membrane structure, a new polymer Polycaprolactone (PCL), was found useful. PCL is a biodegradable, biocompatible, and hydrophobic polymer used for various tissue engineering applications. This polymer exhibits great strength and elasticity. To address the issues concerning CA and TFC membranes, we chose to integrate PCL into the CA solution in the present study. The argument for this approach includes (1) incorporating PCL into CA enhanced its mechanical strength drastically. (2) the presence of PCL can be tailored accordingly to the hydrophilicity needs of the support layer for reducing the ICP. (3) No study has reported this combination for preparing FO membranes so far. Also, GO, and CuO NPs are incorporated to increase the antifouling properties. Hence, the potential bene t of low fouling of FO technology is utilized to concentrate dairy wastewater.

Preparation of nanoparticles
Graphene oxide is prepared using the modi ed hummers method. (Chandran et al. 2021) In general, 2 g of graphite was washed subsequently with water and ethanol to remove any contamination. Further, the graphite (Sigma) was oxidized using a mixture of H 2 SO 4 and H 3 PO 4 (9:1, Merck India) with the appropriate amount of KMnO 4 (Merck India) and stirred at 60°C for 12 h. After the oxidation reaction, the acid slurry was poured into an ice-cold distilled water with 20% H 2 O 2 (Merck India). The yellow color appeared con rmed the formation of GO NPs. Further, the yellow-colored suspension was separated and washed using distilled water and ethanol subsequently. The obtained mass after washing was dried in a vacuum oven at 60°C. Similarly, the CuO was prepared using the sole gel method. (Etefagh et al. 2013) To prepare the precursor, an appropriate amount of copper nitrate (Merck India) was added to a mixture of ethanol and water (1:1). The citric acid (Merck India) was used as the reducing agent and ethylene glycol as the stabilizer. The mixture re uxed for 4 h at 100°C, and the obtained black powder was calcinated at 600°C.

Preparation of FO membrane
The membrane was cast using the Phase-Inversion method. In general, 18 wt. % of cellulose acetate (CA-Sigma) was dissolved in an approximate volume of DMF (Merck India) and acetone (Merck India) (7:3). After 4 h of stirring, the polymer solution was poured over the glass substrate, cast using a doctors' blade, and immersed in a water bath for membrane precipitation. To remove the solvent, the precipitated membrane was washed with DI water. Similarly, CA-PCL blend membrane with 20 wt. % of PCL (Sigma) was fabricated using a similar method. To increase the hydrophilicity of the membrane, an interfacial polymerization (IP) process was used to form a polyamide selective layer on the upper surface of the CA-PCL membrane. The IP reaction was carried out using m-Phenylenediamine (MPD-Sigma) and 1,3,5-Benzenetricarbonyl trichloride (TMC-Sigma). The as-prepared CA-PCL membrane was kept on a glass plate and a 2.0 wt.% MPD aqueous solution was emptied over it. The poured solution was then reserved for 2 minutes to ensure proper saturation, followed by careful removal of excess solution. Following the drying of the aqueous solution, the top surface of the membrane was brought into contact with a 0.1 wt. percent TMC/n-hexane solution for 1 minute. The resultant CA-PCL-TFC (Thin Film Composite) membrane was washed and stored in DI water after drying at room temperature for several minutes and then in the oven for 20 minutes at 70°C. The Thin Film Nanocomposite (TFN) membrane was prepared using GO and CuO. The MPD solution with varying concentrations of GO wt.% (0, 0.005, 0.01, 0.02) was taken, followed by TMC, the solution for the preparation of TFN. The prepared TFN membranes were named CP0.005, CP0.01, CP0.02, respectively. Similarly, as prepared CuO NPs were incorporated in TFN as depicted in Scheme 1. The fabricated membrane is named CA-PCL membrane as CP, Thin-lm composite as CPT, Graphene doped TFC membrane as CPTG, and CuO NPs doped TFC membrane as CPTGC.

Characterization
Morphology of NPs was investigated using FESEM (Model: Gemini SEM 300 Brand: Carl Zeiss). Surface functionality and crystallinity were studied using FTIR (PerkinElmer Frontier MIR) and WAXD (PANalytical). Structure defects of GO were investigated using Raman spectroscopy (532 nm laser-Horiba Labram HR Evo). The blend formation of CA-PCL was investigated using FTIR (PerkinElmer Frontier MIR) and WAXD (PANalytical). The water contact angle of the membrane was measured (Kyowa, DMs-401, Japan) to gain a piece of knowledge about the hydrophilicity of the membranes. The contact angles conveyed are an average of ve measurements done at different locations on the individual membrane samples. Information regarding the mechanical stability, tensile strength, and percentage elongation of the membrane was obtained using Universal Testing Machine (UTM-Shimadzu, AG-X plus, Japan).

Performance of membrane
The performance evaluation of the membrane is made using a lab-scale FO setup. The FO arrangement involves a cross-ow cell, two pumps for the circulation of individual feed and draws solutions, respectively, and a weighing balance on both sides to measure the amount of water passed. The effective area of the FO cell used is 150 × 100mm 2 . Spacers were used along with the membrane in the cell to increase the mass ow rate. The membrane performance is evaluated in the AL-FS mode, i.e., where the selective polyamide layer of the membrane faces the feed side. The forward osmosis tests were conducted using DI water as the feed solution and NaCl as the draw solution. The hydraulic pressures of both solutions were the same (<1.0 kg/cm 2 ). Volumetric ow rates through the cell were maintained at 40 L/hr. Initial volumes of feed and draw solution were 3 ml and 1ml, respectively. After the start of the process, the changes in the weight on both sides were recorded. If the feed side, in any case, shows an increase in weight, then the process should be stopped immediately. Otherwise, the process is continued for a certain period until a constant conductivity value of the feed solution is observed. The water ux J w (Lm −2 h −1 ) through the membrane is calculated using the Eqn.  where ΔC t (g/L) and V (L) are the changes of salt concentration and feed solution volume, respectively.

Nanoparticles
The FTIR spectrum of graphite and GO is depicted in Figure 1a, and the broadening of the peak for GO in a range of 3000-3500 cm − can be used in the forward osmosis process to address the issue caused by the pristine CA membrane.
The tensile strength and percentage elongation of the prepared neat and blend membrane are depicted in Figure 2d. Three samples were tested for each individual CA, PCL, and CA-PCL membrane, and an average value was estimated and compared. From Figure 2d, it is observed that the tensile strength of neat CA is very less as compared to the PCL, demonstrating its lower resistance to deformation. The subsequent blending of PCL in the neat CA to obtain the CA-PCL membrane exhibits a drastic improvement in its tensile strength. Also, the results display that blending has increased the elastic behavior of pristine CA by 395%. As discussed early, the blending of PCL polymer will enhance the mechanical properties of CA polymer. However, the hydrophilicity of the membrane decreases due to the hydrophobic nature of PCL. To overcome this, a thin lm layer of polyamide is coated over the CA-PCL blend. Also, GO and CuO NPs were incorporated inside the thin lm to form a thin lm nanocomposite. The FTIR spectrum was employed to investigate molecular-level interaction. The FTIR spectrum, as depicted in Figure 3a, corresponds to Phase-Inversion CP, CPT, CPTG, and CPTGC membrane. It is observed from Figure 3a  Further, the peak located at 1661 cm −1 in the CPTG membrane is associated with the C=O stretching of the carbonyl group. This spectrum con rms the formation of new amide bonds between the NH 2 group of MPD and OH groups of the GO. Further, there can be observed the presence of a new peak at 700 cm −1 and a subsequent shift in the range of 1200 cm −1 which can be attributed to the presence of copper oxide. The relative hydrophilicity of both the polyamide coated and nanoparticles incorporated membrane was measured by the Contact angle shown in Figure 3b. Time-dependent measurement of water contact angle demonstrates that the nanoparticles doped membranes, CPTGC have a better wettability than the CP membrane. The amalgamation of nanoparticles has reduced the contact angle of the CPT membrane from ~67 o to ~32 o , thus increasing its hydrophilicity by 52%. This increase in the hydrophilic property of the membrane was desired to reduce the antifouling behavior and increase the water ux performance.

Membrane performance
The prepared CPT membrane was tested for different concentrations of draw solute (0.1, 0.5, and 1 M of NaCl), as shown in Figure 4a. It was observed that as the concentration of draw solute increased, the water ux through the membrane increased, indicating an increase in the driving force for the process. Similarly, an increasing trend was observed for the reverse solute ux. However, 0.5 M NaCl was selected as a draw solution due to high water ux and low reverse solute ux. Further, it was observed that when GO (0.005 wt. %) is incorporated into the thin lm matrix, the water ux changed from 4.75 LMH to 21.48 LMH at 0.5 M NaCl as the draw solution. Further, increase in the concentration of GO to 0.01wt.% lead to an average water ux of 26.635 LMH. This increase in water ux for the membrane could be attributed to an increase in the hydrophilicity of the membrane due to the presence of more hydroxyl groups in graphene oxide. After increasing the graphene oxide concentration beyond 0.01 wt.%, a reduction in water ux was observed, which might be because of the fact that a higher concentration of GO could have led to its aggregation in the polyamide layers of the membrane. The optimized GO (0.01wt.%) doped membrane was later assimilated with copper oxide (CuO) on the membrane's same active layers to further increase the hydrophilicity, antifouling, and water ux properties the membrane. The water ux and reverse solute ux performance comparison of the CP-TFC (CPT), CP-TFC-GO (CPTG), and CP-TFC-GO-CuO (CPTGC) membrane against 0.5 M NaCl as draw solute is shown in Figure 5a and b. From the gure displayed it can be concluded that after incorporating both the nanoparticles, the water ux of the membranes increased by the factor of 42. Two batches of fresh laboratory-made dairy wastewater samples were processed by the forward osmosis using a draw solute concentration of 0.5 M NaCl in the AL-FS mode. Firstly, before treating the wastewater, an antifouling study was conducted between the only GO doped and GO-CuO membranes. The better performance membrane could be used for further concentration study. As can be seen from Figure 5c, the water ux declines rapidly with time for both the membranes, but when the membrane was used again after its washing (conducted inline), it was seen that the GO-CuO doped membrane was able to recover its ux, then GO doped membrane. Also, through the literature study, it is found that as the hydrophilicity of the membrane is increased, its capacity to repel the hydrophobic protein increases, thereby increasing the fouling resistance. Therefore, this copper oxide doped membrane was used for concentrating dairy wastewater. The concentration of organic matter in both feed and draw sides were effectively characterized using the TOC (Total Organic Carbon) analyzer. During the FO process conducted for nearly 80 minutes, the feed concentration increased from ~30ppm to ~37ppm, thus, concentrating the solution up to 23% and a concentration factor of 0.903.

Conclusion
This study investigates the applicability of the novel (CA+PCL) prepared Forward Osmosis membranes towards concentrating dairy wastewater. After conducting an optimization study for an appropriate blending of two polymers, the membrane was successfully fabricated using Phase-Inversion methods.
Blending PCL into CA signi cantly improved the mechanical strength of the neat CA membranes. This blended membrane was then effectively modi ed to prepare a TFC membrane, thereby improving the membrane's hydrophilicity and stability, which resulted in the optimal desired low reverse solute ux performance of the phase inversion membrane. Further surface modi cation of this optimized membrane leading to the e cient formation of Thin Film Nanocomposite (TFN) integrating Graphene oxide and

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.