Fabrication and Characterization of Polyphenylsulfone/Titanium Oxide Nanocomposite Membranes for Oily Wastewater Treatment

Polyphenylsulfone (PPSU) membranes are critical for numerous applications, including water treatment, oil separation, energy production, electronic manufacturing, and biomedicine because of their low cost; regulated crystallinity; and chemical, thermal, and mechanical stability. Numerous studies have shown that altering the surface characteristics of PPSU membranes affects their stability and functionality. Nanocomposite membranes of PPSU (P0), PPSU-1%TiO 2 (P1), and PPSU-2% TiO 2 (P2) were prepared using the phase inversion method. Scanning electron microscopy and thermal analysis were performed to determine the contact angle and mechanical integrity of the proposed membranes. The results showed that the membranes contained channels of different diameters extending between 1.8 µm and 10.3 µm, which made them useful in removing oil. Thermal measurements showed that all of the PPSU membranes were stable at a temperature of not less than 240 °C, and had good mechanical properties, including tensile strength of 7.92 MPa and elongation of 0.217%. These properties enabled them to function in a harsh thermal environment. The experimental results of oil and water separation and BSA solution fouling on membrane P2 showed a 92.95% rejection rate and a flux recovery ratio of 82.56%, respectively, compared to P0 and P1.


INTRODUCTION
Oily wastewater (OW) is produced by several industrial processes, including petrochemical, chemical, metallurgical, pharmaceutical, textile, steel, leather, and food manufacturing [Ismail et al., 2019;Alsalhy et al., 2016]. Therefore, the produced OW must be appropriately treated to address environmental and human health concerns [Bahmani et al., 2021]. The hydrocarbon content of fresh water sources grows as a result of the growth of these sectors. Free-floating oil (>150 lm), unstable dispersed oil (20-150 lm), and stable emulsion (20 lm) are the three forms of oil that may be found in water [Kong and Li, 1999]. Stable emulsion droplets cannot be removed using conventional methods such as flotation, chemical coagulation, or heat treatment [Ahmad et al., 2013]. Ultrafiltration is one of the various methods being used to treat oily wastewater today [Sun et al., 2018]. Other methods are ultrasonic separation [Stack et al., 2005], adsorption [Soares et al., Fabrication and Characterization of Polyphenylsulfone/Titanium Oxide Nanocomposite Membranes for Oily Wastewater Treatment 2017], and coagulation/flocculation [Canizares et al., 2008]. However, these methods have disadvantages, such as the need for a large area and the high cost of these procedures. However, because of its unique characteristics such as ease of use and low energy consumption together with the lack of phase transition, membrane processes are considered a viable and cost-effective alternative to traditional methods for treating oilfield wastewater [Ong et al., 2014].
Membrane separation methods have become a widely accepted alternative method for separating oil from water. Microfiltration (MF), ultrafiltration (UF), nanofiltration (NF), and reverse osmosis (RO) are various forms of membrane technology. These membranes have different pore sizes, which is why they have different applications [Pendergast and Hoek, 2011]. Compared to traditional separation processes, UF has low energy cost and high oil removal efficiency without the need to use chemical additives [Alsalhy et al., 2013] .
Sulfone polymers, such as polyphenylsulfone (PPSU), have been extensively investigated for their potential use in membrane science and technology [Nayak et al., 2017]. High thermal and mechanical stability, chemical resistance, impact resistance, and hydrolytic stability are several advantages that PPSU-based membranes offer. PPSU polymers are well-suited for a wide variety of filtration processes, from ultrafiltration to reverse osmosis, because of their bulk modification capability in the polymer's skeleton and the flexibility to customize the pore size of the RO membrane and its porosity [Feng et al., 2016].
Polymeric membranes are increasingly incorporating nanoparticles (NPs) because of their specific characteristics [Salim et al., 2022], functionalization for oily wastewater treatment, and nanoscale size reactivity and large surface area   Reham et al., 2022]. Inert gas condensation, sol-gel, sputtering, spark discharge, ultrasound, coprecipitation, hydrothermal, and biological processes are applied to manufacture these NPs, which are then integrated into the polymers [Haider et al., 2019]. Additionally, inorganic nanoparticles have been used to enhance the qualities of polymeric membranes, such as antifouling, permeability, and thermal and mechanical stability by incorporating them into the composite membranes [Rahimpour et al., 2008]. TiO 2 is the most common inorganic nanoparticle because of its high hydrophilicity, photocatalytic capabilities, anti-fouling properties, and remarkable chemical and thermal durability. A thin layer of water molecules forms on the surface of TiO 2 because of its superhydrophilicity, which results in high hydration of its surface. The selfcleaning characteristic of TiO 2 is enhanced by its photocatalytic nature, which also helps to keep the surface spotless [Rabiee et al., 2014]. However, the number of nanoparticles (such as TiO 2 , SiO 2 , and carbon nanotubes) that may be put into the membrane structure before the phase inversion casting process causes considerable changes to the membrane's morphology. Coating membrane surfaces by adding inorganic nanoparticles may be regarded as an efficient alternative method to increase membrane stability, antifouling performance, and separation ef- No previous study has been conducted on the PPSU/TiO 2 ultrafiltration (UF) membrane process in oily wastewater treatment. Thus, the present study fills a research gap by incorporating TiO 2 NPs into a PPSU polymer solution at various concentrations. Characterization of PPSU/ TiO 2 was conducted using scanning electron microscopy (SEM), contact angle, tensile test, and thermogravimetric analysis (TGA). Furthermore, calculations of pure water flux (PWF), antifouling performance by BSA solution, and removal of oil for each membrane have been analyzed.

Preparation of PPSU/TiO 2 membrane
The PPSU membrane was manufactured with a slight modifi cation to the phase inversion technique previously described [Shukla et al. 2017]. Then, 17% PPSU and 1% PVP were dried into 82% NMP for 3 h at 60 °C with moderate stirring. To make the PPSU/TiO 2 nanocomposite membrane, we added 1% and 2% TiO 2 to the PPSU solution and sonicated it for 1 h using a digital sonicator (Branson Ultrasonics, USA) to obtain a transparent homogeneous solution. The evenly distributed TiO 2 NPs were then hand cast onto a clean and dry glass plate with a blade to a thickness of 100 ± 3 μm. For phase inversion, the glass plate was gently immersed in a coagulation bath containing distilled water. The membrane fl oated to the top of the water in a minute. To avoid microbial contamination, we removed the membrane, cleaned it with DI water, and kept it in an aqueous solution of 0.2% sodium. Details of the prepared membranes are shown in Table 1. The schematic of membrane fabrication is shown in Figure 1.

Scanning electron microscopy
The surface and cross-sectional morphology of membranes were studied using SEM (JEOL, Tokyo, Japan). Sputtered Pt coatings were applied to membrane samples having an area of approximately 0.5 cm 2 that were taped to support. Cross-sectional pictures were obtained by orienting the membranes perpendicular to the electron beam. A series of magnifi cations were made to capture the SEM pictures at a working distance of 6.4 mm and an accelerating voltage of 10 kV.

Contact angle measurements
Sessile drop technique was used to measure the hydrophilicity of the membranes. Goniometer (Atension, MAC 200, Biolin Scientifi c, Amsterdam, Netherlands) coupled with a digital camera and image processing software was used to measure the contact angles of the surfaces. Using a microliter syringe, we deposited water droplets with volumes of 3 1 L at fi ve points on the membrane surface. The digital camera captured a 2D picture of each water droplet's profi le on the membrane surface. Contact angles (q) were averaged from fi ve measurements.

Tensile test
A computerized universal testing machine, type D638 was used to perform the tensile test on the samples (WDW-50E). The test was performed at room temperature with a steady strain rate of approximately 5 mm per minute (Vertex 80, Bruker, UK).

Thermogravimetric analysis (TGA)
The temperature ranged from 25 °C to 600 °C for the thermogravimetric analysis of the produced membranes (Veeco, San Jose, CA, USA). A flow rate of 30 mL/min and a heating rate of 10 °C/ min were applied in a nitrogen environment. A 6.9012 mg dry sample was used. The sample mass measurement had a standard uncertainty of 1%. Calcium oxalate, delivered along with the device which was used to calibrate it.

PWF of membranes
We determined the water permeability of membranes using a cross-fl ow fi ltration cell. Before the experiment, the fl at sheet membrane was rinsed with deionized water. In the membrane module, the PPSU fl at sheet membrane was sliced into a square sheet with a surface area of 18 cm 2 , as illustrated in Figure 2. The module's intake was fi tted with a pressure gauge and attached to a feed solution tank. Deionized water was used as the feed solution to estimate the pure water fl ow, followed by measurements of the permeate water fl ux utilizing the membrane for 2 h. Permeate was taken from the membrane at regular intervals (30,45,60,75,90,105, and 120 min). PWF was calculated by the following equation (1)  ( where: (

Antifouling study of membranes
where: Jb -permeated flux. The membrane total fouling (Rt %) was calculated by the reversible and irreversible fouling and defined as follows [Tiron et al. 2017]: 1 = .
(%) = 2 − 1 × 100 The fouled membrane was washed with deionized water after filtration of the BSA solution for 2 h. The pure water flux, Jw 2 , of the cleaned membrane was measured again under the same conditions, and flux recovery ratio (FRR) was calculated as follows [Tiron et al. 2017] : 1 = .

Procedure for oily wastewater separation and rejection
Oily wastewater was collected from a local refinery in Iraq after the dissolved air flotation (DAF) process. The oil was analyzed in the sanitary laboratory at the University of Technology -Iraq using the gravimetric method. The organic phase was extracted from the aqueous phase using n-hexane as a solvent (EPA method 1664). Three 30 ml parts of n-hexane were used to extract 400 ml of oily water, and one 20 ml piece was used for the final rinse in this phase. An analytic funnel containing filter paper and 10 grams of anhydrous sodium moistened with n-hexane was used to drain the organic phase into an Erlenmeyer flask, where it was collected. The color of the extract in the flask was then examined by placing it on a stirrer plate. Gravimetric analysis was performed using technique A because the extract exhibited a mild yellowish color. The method was performed in the following stages. Magnetic stirring of the Erlenmeyer flask with silica gel (3.0 g) was conducted for 5 min. A roundbottom distillation flask had been reweighed before the extract was filtered and collected. An IKA RV 05 Basic 1-B rotary evaporator was used to remove the solvent from this flask, which was then placed under nitrogen flow for 1 h. After this length of time, the flask was weighed again. At the end of the extraction, the oil rejection R (%) of the oily wastewater was calculated as follows [Tiron et al. 2017]: where: Cp and Cf are the concentrations of all parameters in permeate and feed, respectively.

SEM analysis
The morphology of the top surface and crosssection of PPSU/TiO 2 membranes was inspected through SEM as shown in Figure 3. Figure 3a shows that the top surface of pristine PPSU had a high pore density with very small pore size. Furthermore, Figure 3b shows that the membrane fabricated from pristine PPSU had finger-like channels at the top with randomly distributed macro-voids at the bottom. Furthermore, a sponge layer can be observed as an essential layer of the cross-section because of the moderately high viscosity of the casting mixture caused by the high polymer concentration (17 wt.% PPSU).
Adding TiO 2 with 1 wt.% and 2 wt.% to the casting solution resulted in good optimal dispersion of titanium oxide in the membrane matrix, thereby modifying the morphology of the PPSU membrane. In addition, it led to an increased solubility factor between the solvent and the nonsolvent, resulting in the elimination of macrovoids, and thus the formation of a highly porous or dense structure (Figures 3d and 3f). The top surface of the fabricated membranes P1 and P2 had higher pore density with smaller pore size than the top surface of the P0 membrane (Figures 3c and 3e).

Analysis of contact angle and thickness
The contact angle is often measured to determine the types of suitable materials and ensure the optimum use of the fabricated membranes. On the basis of hydrophilicity and hydrophobicity, the membranes can be classified as those used for removal of water-based pollutants and those used for removal of organic pollutants. Figure 4 (top) shows that the prepared membranes have contact angles of 69°, 65°, and 64° for P0, P1, and P2, respectively. The contact angle after adding nanotitanium oxide became lower, which means that the surface area increased and the contact angle decreased. This condition enhanced the possibility of using the membrane to remove organic and inorganic pollutants, and thus, the membrane achieves a high probability of use in different applications [Taylor et al. 2007]. Furthermore, membrane thickness of P0, P1, and P2 was observed by SEM and summarized in Figure 4 (bottom). The thickness of all the membranes was approximately similar.

Tensile test analysis
The stress-strain curve (Fig. 5) shows that the P0 polymer and P1 is a ductile and not a brittle material, where the tensile strength value is equal to 7.318 MPa and the elongation value is 0.261% [Xu, 2019]. In the case of P2, the tensile strength value was increased and found to be 7.92 MPa, while the elongation value was 0.217%. The measurement proves that 1% addition of TiO 2 NPs did not cause a significant change in the mechanical properties of the membrane. However, a significant change was observed when the amount of TiO 2 was increased by 2%.

Thermogravimetric analysis
To ascertain the thermal stability of the polymeric membrane, we took 0.6944 grams of it and placed it on a platinum pan in the thermal analyzer. The weight loss stages were recorded at different temperatures. A temperature of 50-240 °C indicated a loss of approximately 57%, 42%, and 37% of the total weight of the PPSU, PPSU -1%TiO 2 , and PPSU -2%TiO 2 membranes, respectively. This stage was due to the beginning of polymer melting. At 240-750 °C, a loss of approximately 79%, 78%, and 68% of the total weight of the PPSU, PPSU -1%TiO 2 , and PPSU -2%TiO 2 , respectively, was observed as a result of the decomposition of chemical bonds. At 1,000°C, breakdown of the rest of the polymer and its transformation into gas resulted from combustion [Hatakeyama et al. 1999]. The results showed that all membranes were stable for use even when their composition was not changed to 240°C. Therefore, the membranes have potential to be used in applications where heat is essential. Figure 6 shows that increasing the percentage of TiO 2 increases the stability of the polymer because weight loss is least possible for the membrane containing 2% of TiO 2 .

PWF
Nanocomposite membranes P1 and P2 enhanced water permeability compared to the P0 membrane (Fig. 7), notably with increasing TiO 2 concentration from 1.0% to 2.0%. The water permeability increased to twice that of the pristine membrane P0. The average PWF for P0, P1, and P2 were 35.03, 45.75, and 54.53 L/m 2 ·hr. The increasing water flux was also due to the presence of PVP as a pore-forming additive that affected the morphology of the PPSU membrane and increased the water flux [Nayak et al. 2017].

Antifouling study
A common problem with membranes is fouling, which reduces their ability to separate and increases their energy consumption. The antifouling capabilities of the TiO 2 -modified membrane were evaluated by monitoring water flux recovery following BSA solution fouling. Fouling may be influenced by various factors, including the solutemembrane interface, polymer chemical and structural characteristics, ionic strength, and pH [Shi, 2016]. Table 2 shows the calculated FRR, Rr, and Rir fouling ratios (reversible and irreversible, respectively). Using the FRR metric, the antifouling capabilities of the modified membranes may be assessed, and P2 had an FRR of 82.56%. The FRR value improved when TiO 2 nanoparticles were added to the membrane, according to the data. Hydrophilic TiO 2 particles minimized the fouling of PPSU/TiO 2 membranes due to the presence of hydrophilic groups (hydroxyl and amine) on its surface, which led to the production of a thin water layer on the membrane surface and enabled cleaning. Water molecules and hydrophilic groups on the nanoparticles exposed to the feed solution on the membrane surface showed more intense interaction, thereby reducing membrane fouling [Cheshomi et al. 2018].
As shown in Figure 8, PPSU surface modification using TiO 2 nanoparticles was shown to be an efficient method for increasing the water flow and antifouling capabilities of the membranes, according to the findings of this study.
The resulting hydrophilic particle layer acts as a hydrophilic filtration membrane in isolating pollutants and supporting membranes from   fouling. The membranes exhibit increased permeate flux resistance depending on particle size (a larger particle size produces dynamic membranes with lower resistance) [Lu et al. 2018]. Backwash can easily remove particles in the dynamic membrane because they are not chemically bonded to one another or to the supporting membrane. As a result, fouling on the membrane is reversible. Furthermore, the membrane can be regenerated after backwashing by applying another hydrophilic particle layer. As a result, PPSU-TiO 2 -NP membranes have the advantages of easy preparation and regeneration.

Oil rejection
Increases in TiO 2 additives and PVP as a pore forming agent resulted in improved membrane oil rejection efficiency. In this study, only water molecules passed through the membrane and produced oil-free water after the oil droplets were rejected on the membrane surface. Figure 9 shows that membranes P2, P1, and P0 have 92.95%, 91.88%, and 89.90% rejection rates, respectively. Table 3 shows a comparison between the performance of membranes prepared in this study with various membranes found in the literature. The comparison was according to the removal efficiency and total pure water flux. The PPSU-TiO 2 -NP membranes have a reasonable pollutant removal efficiency and PWF compared with most membranes in previous research.

CONCLUSION
Several applications of membrane technology have been used, such as water purification and desalination, as well as oily wastewater treatment, because they are simple and fast to perform. Up to this point, most of the research has focused on the development of new membranes. Oil and water separation is a vital stage, but a study on the process and the mechanisms by which the oil droplets are rejected on the membrane surface is limited. Phase inversion was effectively performed to manufacture polymeric membranes for the treatment of oily effluent from a local refinery in Iraq. Thermal analysis, contact angle, and SEM determined the capabilities of the produced membranes. Changes in membrane morphology and pore size distribution occurred as a result of adding TiO 2 nanoparticles to the PPSU membranes. The hydrophilicity and PWF properties of PPSUs were improved when TiO 2 NP concentrations in the casting solution were increased. The P2 membrane had a higher FRR of 82.56%. Using BSA as the model foulant indicated a higher antifouling capacity. A rejection of 92.95% was observed for membrane P2, whereas 91.88% and 89.90% rejection rates were found in the oil-in-water separation experiments performed on membranes P0 and P1. The SEM images showed that the channels between the membranes provided a good water penetration route. The findings of this study are expected to pave the way for large-scale antifouling membranes based on TiO 2 NPs that can be used in a wide range of water treatment applications.