Development of Novel Fouling-Resistant Hollow Fibre Nanocomposite Membrane Augmented with Iron Oxide Nanoparticles for Efficient Rejection of Bisphenol A from Water: Fouling, Permeability, and Mechanism Studies

Recently, frequent discharge of water-ladened emerging organic pollutants such as Bisphenol A has elicited serious concern due to its pernicious effects on public health and the ecological environment. In this study, hematite nanoparticles (Fe2O3) were synthesized via the sol–gel auto-combustion technique and incorporated as a nanofiller to fabricate a PVDF-PEG nanocomposite hollow fibre membrane with enhanced antifouling properties. A series of membranes comprising various loadings (1.0–2.0 wt%) of Fe2O3 NPs were fabricated through the phase inversion technique and thoroughly analyzed. The developed Fe2O3-membrane fibres were thoroughly characterized. The performance of the membrane fibres was investigated through permeation flux, BPA rejection, as well as antifouling characteristics. Based on the results obtained, the resultant nanocomposite membrane fibres exhibited superior performance in comparison with the pristine fibre. Also, the nanocomposite membrane with 1.5 wt%-Fe2O3 NPs exhibited remarkable performance with − 43.7 mV, 56.3º, 191.85 L/m2 h, 86.7%, and 12% of negatively charged zeta potential, least contact angle, water permeation flux, BPA rejection, and minimum weight loss, respectively. Besides, the 1.5 wt-Fe2O3 NPs nanocomposite membrane demonstrated superior antifouling performance after the third filtration, accomplishing a higher percent of FRR (77.35%) along with RFR of 21.29%, respectively. Hence, based on the performance of the fabricated hollow-fibre membranes loaded with Fe2O3 NPs, efficient antifouling membranes was achieved which can be suitably applied in the purification of industrial wastewater.


Introduction
Water pollution is one of the contemporary global issues resulting from a scarcity of pristine water that affects millions of people around the world, and this problem is becoming more acute as a result of population increase, rapid industrial advancement, and climate change [1]. In the last decade, pollution of the water resources and the environment has had an adverse impact on the ecosystems and the lives of animals. With the expeditious advancement in industrialization and the growing economy, the purification of drinkable water is becoming increasingly indispensable in our society since water quality is deteriorating daily due to the superfluous exploitation of harmful chemicals, including emerging contaminants [2].
Particularly, endocrine-disrupting compounds (EDCs) are among the most extensively released emerging contaminants into the environment via human-induced activities. Moreover, most of these compounds have drawn urgent attention due to their detrimental effects on the environment, potential noxiousness to humans, ecological safety, and bioaccumulation [3]. Among these compounds, Bisphenol A (BPA) (2,2-bis (4-hydroxyphenyl) propane), is the most essential industrial chemical, mainly used as a monomer to manufacture unsaturated polyester-styrene resins, epoxy resins, flame retardants, and polycarbonate. Similarly, it is extensively utilized in food can coatings, paints, adhesives, and lotions, and it is being generated in superfluous quantities globally, with an annual output surpassing 3.8 million tons [4]. However, effluent from wastewater treatment facilities, landfilled plastic waste leachate, BPA production units, paint industries, and plastic industry effluent, as well as e-waste, remains the main source of BPA pollution [5]. Similarly, BPA can trigger multidimensional harmful effects, including immune toxic, hepatotoxic, carcinogenic, mutagenic, and breast cancer in living organisms [6]. Undoubtedly, reports have indicated that persistent exposure to BPA could induce severe human health diseases, particularly infertility, diabetes, liver damage, and heart disorder [7].
Recently, the removal of emerging contaminants, particularly BPA, from water is rapidly drawing the interest of the scientific community [8]. The separation of EDCs using membrane technology is one of the promising economical techniques and is suitable for efficient purification due to its facile synthesis, ease of operation, zero chemical supplements or phase changes, comparatively low energy requirement, and superior separation performance [9]. Recently, the UF membrane in particular has had greater potential in the water treatment process and is regarded as an efficient barrier for suspended substances, colloids, microorganisms, and particles, and has been increasingly used for drinking water supply owing to lower energy requirements, low cost, efficient performance, and a reduced operating pressure which indicates an advantage of UF over other hydraulic pressure driven processes [10]. Also, the UF technique is a suitable option for the pretreatment process for drinking 1 3 water purification. Thus, a FO membrane that allows for a sufficient rejection of trace organic contaminants while maintaining a high water flux would be desirable.
It is pertinent to mention that hollow fibre membranes are one of the preferentially applied configurations in industrial separation as compared to flat sheet membranes, as a result of their excellent properties including elevated efficiency per unit volume, superior surface area to volume ratio of a module, enhanced filtration, and rejection performance, mechanically self-sufficiency, and their superior operational flexibility compared to flat sheet membranes [11].
Poly(vinylidene fluoride) (PVDF) is the most preferred polymer and is widely utilized to manufacture UF membranes among the various polymers employed in membrane fabrication by virtue of its remarkable chemical resistance, thermal stability, mechanical properties, and practicability [12]. However, its persistent hydrophobicity, which could trigger severe fouling, has remained a significant drawback, undermining the applications of PVDF membranes, particularly in wastewater and drinking water purification [13].
Undesirably, fouling remains to be a major drawback and limiting factor in membrane purification technology most especially when water is the separation medium as well as in industrial membrane applications [14]. Typically, the polymer membranes are easily fouled by the adhesion of natural organic matter (NOM) due to their inherent hydrophobicity [15]. The accretion on the membrane surface as well as inside the pores of natural organic matter and inorganic solutes creates a further impediment that advances to membrane fouling [16]. Membrane fouling has detrimental impacts on the membrane including flux and membrane lifetime decline, heightened cleaning frequency, and energy demand [17,18]. Therefore, numerous efforts are made to alleviate membrane fouling including expansion of the hydrophilicity and antifouling characteristics of the membranes via blending modification [19], surface grafting [20], and surface coating via incorporation of nano additives [21] has been regarded as an efficient technique to enhance membrane resistance to fouling [22].
Recently, nano-composite membranes have been receiving enormous attention owing to their superior characteristics and better separation performance as well as excellent permeate flux. Nano-composite membranes are fabricated by incorporating nanoparticles into the membrane dope matrix [23]. Nanoparticles exhibited a unique large surface-to-volume ratio and micropore structure together with surface functional groups (OH) [24]. These properties make it compatible with the formation of a stable matrix structure and strong linkage with most of the polymers used for membrane preparation [25]. Various nanoparticles are widely utilized to modify polymeric membranes to enhance their hydrophilicity [26].
However, previous reports have indicated that most of the widely applied nanoparticles are photo-catalytically propelled to efficiently address the fouling problem [27]. This suggests that the presence of ultraviolet radiation is sine qua non to precede the antifouling performance. Furthermore, the issue persists in releasing superoxide and antifouling radicals, which could seriously endanger the stability of the composite membrane matrix structure.
Meanwhile, iron oxide NPs have drawn significant attention from the scientific community owing to their wide scientific applications [28]. In particular, hematite (Fe 2 O 3 ) NPs are highly stable (strong affinity to form a stable linkage with the polymer/co-polymer), low-toxic, inexpensive, amply available, and eco-friendly as compared to other iron oxides [29]. Interestingly, Fe 2 O 3 NPs are highly hydrophilic with minimal fouling tendency (strong mitigation of the fouling tendencies of the polymeric membrane), outstanding compatibility, and the ability to produce a hydroxyl functional group through the deprotonation reaction as compared to other metal-based NPs [30]. Thus, the incorporation of Fe 2 O 3 NPs into the (PVDF-PEG) membrane polymeric matrix can enhance its surface negativity [37]. Moreover, the blending of Fe 2 O 3 NPs into the polymer membrane matrix was expected to augment its fouling decline propensity [31], thereby making the membrane matrix hydrophilic.
Given these, the present study focuses on modifying the structure of an in-house fabricated PVDF-PEG ultrafiltration membrane at various loadings of Fe 2 O 3 NPs to augment its hydrophilic properties to enhance permeability flux, increase BPA rejection, curtail fouling, which can result in minimal energy consumption. Thus, it is envisaged that Fe 2 O 3 NPs may play a more considerable role in elevating the permeability and antifouling capability of membranes. More explicitly, Fe 2 O 3 NPs have not been explored as a nanofiller to fabricate a PVDF-PEG/Fe 2 O 3 composite membrane for the removal of BPA from drinking water. Stemming from this gap, we investigated the influence of varying Fe 2 O 3 NPs concentrations (1.0, 1.25, 1.5, and 2.0 wt.%) on the membrane properties as well as separation performance. The developed membrane fibres were analyzed based on hydrophilicity, thermal strength, morphological and elemental composition, surface charges, porosity, pore size, BPA rejection, as well as antifouling properties. PVDF was used as the principal bulk polymer and was utilized after oven drying at 80℃ for 48 h, and DMF represented the primary solvent to dissolve PVDF. PEG (MW: 6000 g/mol., CAS NO: 25322-68-3) was procured from R&M Chemicals, Essex, U.K. (co-polymer) and was utilized to enhance the pore formation. Anhydrous citric acid (C 6 64-17-5) were obtained from R&M Chemicals, Ltd., Malaysia) and utilized as doping and post-treatment solvents. Bisphenol A (99% purity) reagent, KCl, HCl, and NaOH (AR) analytically graded obtained from Sigma-Aldrich (West Chester, PA, USA). Distilled water was utilized to prepare standard solutions throughout the experiment. Hematite (Fe 2 O 3 ) NPs with particle size (MW: 159.69 g/mol., BET ˂50 nm) was prepared in-house.

Synthesis of Hematite (Fe 2 O 3 ) NPs
The sol-gel auto-combustion method was employed during the synthesis of hematite (Fe 2 O 3 ) NPs with slight modification [32]. The precursor materials are citric acid (C 6 H 8 O 7 ·H 2 O) and iron (III) nitrate hydrate Fe (NO 3 ) 3 ·9H 2 O, and metal salt, while ethanol and distilled water were used as a dissolving medium. Firstly, the precursors were measured utilizing a weighing balance. The calculated amount of iron (III) nitrate was dissolved in 50 ml. of distilled water to acquire a definite solution and then stirred magnetically for 20 min, and the experiment was subsequently succeeded by adding a suitable amount of ethanol to the solution to improve the reduction potential of the process. Afterward, the desired quantity of ammonia solution (NH 4 OH) was dropwise added to the solution with continuous agitation at room temperature to regulate the pH (7)(8). Next, metal salt in a stoichiometric ratio with citric acid (2:1) was gently introduced to the solution and heated under 70-80 ℃ for 2 h. The resultant homogenized solution was steadily agitated using a magnetic stirrer and heated at 110 ℃ to turn the sol into a red gel, and the temperature was thereafter raised to 250 ℃ for combustion to occur. The resultant fluffy powder was later sintered at 550 ℃ for 5 h to eliminate the organic components and obtain a dried single hexagonal Fe 2 O 3 NPs.

Membrane Dope Formulation
Prior to dope formulation, the PVDF pellets and Fe 2 O 3 NPs were oven-dried at 110 ℃ for 24 h to exterminate the moisture level and preserved using a desiccator to cool down to ambient temperature. Afterward, a calculated quantity of Fe 2 O 3 NPs was dispersed and continuously sonicated in DMF using a digital ultrasonic water bath (Branson, 3800) at 60 ℃ for 1 h to prevent the accretion of NPs. After that, LiCl 2 ·H 2 O was gradually added into the dispersion and steadily stirred at 250 rev per min under a temperature of 80 ℃ for 24 h and enabled to degas. Primarily, this step is to attain excellent mixing of the solution and achieve satisfactory dispersal of NPs in the solution and get rid of air bubbles. Then the mixture was further allowed to cool at room temperature, and the dried PVDF pellets as well as PEG were dispersed into the solution and sonicated at 380 rpm under a temperature of 100 ℃ for 24 h using a hot plate with a stirrer (Monotaro; C-MAG HS9, Malaysia) to obtain a homogenized dope solution. The resultant solution was further sonicated for 30 min. and then placed separately for a few hours to expunge the air bubbles. The composition of the dope formulation and the varying quantities of the Fe 2 O 3 NPs in the solutions are presented in Table 1.

Spinning Process for the Fabrication of Nano-composite PVDF/PEG-Fe 2 O 3 Hollow Fibre Membrane
A dry-jet wet spinning procedure was used in the fabrication of hollow fibre polymeric membranes. Essentially, similar spinning conditions were employed throughout all the fabricated membrane fibres. The dope solutions were swirled via a dry-jet wet-spinning procedure utilizing a spherical spinneret. The outer and inner diameters of the circular spinneret were 1.15 and 0.55 mm, correspondingly. Distilled water and tap water were applied as the bore liquid and external coagulants in the fabrication of the membrane. Furthermore, the extrusion rate, pick speed control, external coagulant temperature, air gap, collecting drum speed, room humidity, and ambient temperature remains steady at 2.5 ml/min, 7 rpm, 25 ℃, 10 cm, 7 rpm, 72.7%, and 29.5 ± 1 ℃ correspondingly. The swirled fibres were immersed in a continual flow water bath for 24 h to remove solvent residues. Water was utilized as a bore fluid and a non-solvent bath. Following the formation of the fibres, they were preserved in a water bath overnight to allow solvent exchange and finally air-dried.

Analysis of Membrane Morphology
The surface, as well as the cross-sectional morphologies of the fabricated nanocomposite membrane fibres, were examined using (SEM: S-3400N) scanning-electron-microscope. Briefly, membrane fibres were fractured into liquid nitrogen for 5 min to provide a sharp breaking of the fibre to display the cross-section configuration. Next, the immersed fibres were coated with a thin gold layer, while the voltage acceleration was kept steady at 20 kV during the image capturing. In addition, the elemental composition, imaging, and dispersion of Fe 2 O 3 NPs within the nanocomposite membrane fibres were investigated utilizing SEM-Thermo Scientific (Hitachi & S-3400N) scanning electron microscope/energydispersive X-ray. The fractured membrane fibres were positioned on the sticky tape attached to the metal plate and studied at 20 kV.

Analysis of Porosity and Hydrophilicity
The porosity of the membrane was computed by determining the weight of the membrane samples in wet and dry conditions using the gravimetric method [33]. Ten (10) pieces for each membrane sample of 20 cm of the same length were airtightly enclosed at each end using adhesive resin and were cut and immersed in water for 48 h at room temperature. Next, dry tissue paper was used to remove the superficial water drops from the surface of the membrane samples and then measured as a wet membrane (M w ). Wet membrane samples were dried for 24 h under 80 ℃ before being remeasured as the dried membrane (M d ). Consequently, the porosity (ℇ) for each sample was computed using Eq. (1): where ℇ represents the porosity in %, M d denotes the weight of the dry membrane (g), M w is the weight of the wet membrane (g), ρ PVDF denotes the density of PVDF (1.78 g/cm 3 ), while ρ H2O depicts the density of water (0.998 g/cm 3 ). The average pore radius of the membrane samples was calculated using the Guerout-Elford-Ferry equation using the water flux and porosity data (2) [8], as follows: where Q (m 3 /s) denotes water flux, A depicts the effective area (m 2 ) of the membrane, η represents viscosity (8.9 × 10 −4 Pa s), r m denotes membrane pore radius (rm, nm) and ΔP is the transmembrane pressure (0.3 MPa).
The hydrophilicity of the fabricated fibres was determined using a goniometer (GmbH OCA 15pro, Data-Physics) on the basis of static contact angle. RO water was utilized as the analytical liquid, such that the droplets were obtained and noted with the fitted camera within 20 s to avert evaporation. The hydrophilicity was qualitatively assessed by calculating the contact angle from the shape of the drop. The calculated contact angles were computed using SCA20 software. Each measurement was repeated ten times on different membrane spots under room temperature, and the mean value was calculated as the contact angle (ϴ) to mitigate error.

Analysis of Surface Charge
The membrane surface charge was investigated using (Zetasizer Nano Plus 3, Zeta/nanoparticle analyzer, Malvin, USA) Zeta potential analyzer equipped with an adjustable and holding cell at room temperature. Besides, a 1.0 mM KCl was utilized as a base electrolyte at a varying pH (2-10) applying 0.1 M NaOH or HCl. The prepared samples were positioned on the sample griper using glue tape. Afterward, the analyzer was pre-adjusted to 400 mbar peak pressure to secure laminar flow [34]. The surface charge was computed using the Helmholtz Smoluchowski procedure.

Analysis of Membrane Thermal Stability (TGA)
The thermal resistance (stability) and weight decline (degradation process) of neat and nanocomposite membranes with respect to varying temperatures were conducted using a thermogravimetric analyzer (Mettler Toledo, TGA-DSC HT 3). The analysis was conducted in a regulated nitrogen atmosphere. The sample mass (5-10 mg) was filled in an alumina pan, pre-heated to 100 ℃ for 15 min to eliminate moisture, and subsequently heated with a 10 °C/min heating rate with corresponding temperatures ranging from 25 to 600 °C.

Analysis of Stability of Fe 2 O 3 Nanoparticles on the Membrane
Leaching of the nanoparticles in the application of nanocomposite membranes is an important index to be considered. To measure the amount of Fe ions leaching out from the composite membrane and investigate Fe 2 O 3 NPs stability, for this test, a composite membrane sample (M4) for the evaluation of Fe leaching with an area of 0.01615 m 2 was immersed in a container. Then, 40 mL of deionized water was added, and the container was sealed and placed on an orbital shaker and stirred at 180 rpm for 48 h at room temperature. The membranes were removed, and the presence of Fe ions in the resultant solution was analyzed Finally, the resultant water samples were collected and analyzed for the presence of Fe ions by ICP-MS (PerkinElmer, NexION350D) with argon as carrier gas. The mean value based on triplicate measurements was used for analysis.

Membrane Performance Evaluation
A filtration test was performed using a laboratory scale dead-end system that was designed for this study at room temperature during the experiment to assess the permeate flux and separation performance of the pristine and nanocomposite membrane fibres at minimal working pressure. The system comprised of a feed tank, control valve, UF membrane compartment, pressure gauge, membrane fibres, peristaltic pump, and permeate tank. The feed solution is released from the feed tank through gravity to the membrane compartment via the control valve. The control valve, which was connected to the membrane compartment controls the flow while the pressure is being monitored through the pressure gauge. A peristaltic pump (Model No.:YT25H), purchased from Guangdong, China, provides the required suction pressure during the filtration process. The schematic representation of the experimental configuration is presented in (Fig. 1). Each of the modules contains 20 pieces of the membrane of a similar equal length of 0.3 m. To evaluate the permeate flux, the hollow fibre membrane module was initially compacted and conditioned with distilled water for 30 min at 0.4 MPa under room temperature to attain stable-state conditions (steady flux) during filtration experiments, while the subsequent filtration process was carried out at reduced pressure (0.3 MPa). Following compaction, the water flux was recorded at different transmembrane pressures (TMP, ranging from 0.3 to 0.7 MPa), and no less than three readings were recorded to achieve an average value. The influence of the applied TMP was investigated using each of the fabricated membranes at room temperature. The permeate volume was determined using a volumetric cylinder at 2 min intervals under TMP of 0.3 MPa. The permeate flux, J w (L/m 2 h), was computed using Eq. (3): where V denotes the volume of water permeated (L), effective area (A e , m 2 ) of the membrane, as well as filtration time Δt (t, h), were measured [35].
Prefiltered (0.45 µm) tap water was obtained from the Sewage Treatment Plant, located at Putrajaya 1, Regional Sewage treatment in Putrajaya, Malaysia.

Bisphenol A Rejection (R%) Study
The rejection tests of the swirled fibres were studied with respect to BPA removal from the feed solution. Membrane fibres were subjected to filtration of feed solution with the initial BPA concentration of 38.97 µg/L using the same setup. The membrane fibres were initially immersed in the feed water for 90 min to create adhesion of BPA on the fibres. This step aids in saturating the porous structure of the membrane fibres [36]. The apparent concentrations of BPA in the feed solution and permeate samples were analyzed using the ultra-high-performance chromatography (UHPLC) procedure. BPA rejection, R, is defined, as a percentage and was computed using Eq. (4): where R depicts BPA rejection (%), C p denotes the concentration of BPA in the permeate (µg/L), whereas C f describes the concentration of BPA in the feed solution (µg/L).

Adsorption of BPA Pollutant Filtration Protocol
The investigation of the adsorption of BPA using pristine and modified membranes was performed by filtration using the same setup operated under a TMP of 0.3 MPa at room temperature. The feed solution has an initial concentration of 38.97 µg/L and the filtration was conducted for 24 h to guarantee that the adsorption was saturated. The adsorption quantity during filtration (Q f , μg/cm 2 ) of each sample was computed using Eq. (5): where C p and V p are the concentration (μg/L) and volume (L) of BPA in the permeate solution, respectively; C f and C' f are the initial and final concentration of BPA (μg/L) in the feed solution, respectively; and V f and V' f are the initial and final volume (L) of BPA in the feed solution, respectively.

Analytical Methods
The BPA concentrations in the solutions and samples were quantified using (Thermo Scientific Dionex, Ultimate 3000, USA) ultra-high-performance chromatography (UHPLC). In the case of water samples after adsorption, the biochar was separated from the solution by a magnetic field for 25 min. The separation condition of BPA was conducted on the (Hypersil Gold aQ 100 × 2.1 mm, 1.9 µm) column fitted to UHPLC for BPA analysis under 30 °C. The injection volume of 20 μL was introduced into the column and chromatographically separated using a mobile phase containing 95% water (Optima@LCMS, Fisher) and 5% acetonitrile (Optima@LCMS, Fisher) at a 0.4 mL/min flow rate, at a 230 nm detector wavelength. The BPA concentration was computed from the calibration curve and a good linear correlation with R 2 = 0.9938 was observed in the concentration range of 10 µg/L to 50 µg/L.

Fouling and Reusability Studies
The membranes were subjected to water filtration for a period of 200 min and then cleaned with distilled water to get rid of residual particles (foulants) from the membrane surface. Afterward, cleaned membranes were then reused to investigate the reusability potential in relation to permeability flux with respect to time.
The membranes were subjected to feed solution filtration for a period of 200 min. and rinsed with distilled water to get rid of residual particles (foulant) attached to the surface of the membrane. The washed membranes were subsequently reapplied again to study the reusability potential with respect to permeability flux with respect to time. Given this, membrane fibres with various concentrations of nano-Fe 2 O 3 (a) 0, (b) 1.0, (c) 1.25, (d) 1.5, and (e) 2.0 wt% were exposed to three filtration cycles with 200 min. for each cycle to examine the anti-fouling characteristics of the membranes. The membranes were rinsed with distilled water for 30 min. after each filtration cycle.

Membrane Fouling Analysis
To assess the reusability and antifouling characteristics of the membrane fibres and mimic practical conditions, the membranes were exposed to three continuous filtration cycles for 9 h using a known concentration (38.97 µg/L) as the feed solution. The antifouling performance of membrane fibres were assessed following 180 min of uninterrupted filtration for each completed cycle. Then the antifouling efficiency and reusability of the fibres were examined using the (%FRR) flux recovery ratio as well as (%RFR) relative flux recovery indices using Eqs. (6) and (7), respectively. The membrane fibres were conducted under three cycles of filtration. Each cycle was conducted for 96 min and was then reapplied for a subsequent cycle of filtration following a 30-min physical washing utilizing distilled water. The membrane module, after the DI water cleansing, was also backwashed with DI water for 30 min under the same operating pressure as for the filtration process and the filtration run was also repeated for a second cycle.
The water flux (J L ) was determined using the corresponding filtrate volume (mL) attained at an interval of 24 min during the tests. Upon the completion of each filtration cycle, the permeate volume, permeation flux as well as the solute rejected were calculated. Next, the applied membranes were cleansed with distilled water for 10 min and backwashed with DI water for 30 min. The washed membranes were reused for the 2 nd filtration cycle for a further 180 min, using a similar approach as earlier indicated. The filtration cycle was carried out three times to assess the antifouling performance of the membranes.
where J w depicts water flux, J f represents feed flux, and J w2 denotes the second pure water flux after cleaning, (L/m 2 .h). Table 2 presented the physicochemical characteristics of the feed solution which were determined following standard procedure (APHA, 2005) alongside the drinking water quality standards.

Morphological Analysis of Hollow Fibre Membranes
fibres. It can be noticed that the pristine fibre (a) possessed three typical layers having thin layers equally at an outer and inner segment of the membrane (Fig. 2). The centre layer represents sandwich-type morphology comprising little finger-type pores at each side near the ultra-thin layers. Though, the modified nanocomposite membranes exhibited distinct scenarios. The finger-shaped pores grow longer and the amount of the micro-pores structure considerably expanded owing to the increasing Fe 2 O 3 NPs loading (Fig. 2b-d) [37]. Besides, some asymmetric structures characterized by sublayer and skin layers composed of spongy-like cavities, finger-like projections, closed ends, and macro-void structures were also observed near the interior ultra-thin layer at an elevated concentration of Fe 2 O 3 NPs. This phenomenon is consistent with previously reported studies [38,39]. By increasing the hydrophilic nanofillers Fe 2 O 3 NPs within the membrane matrix, the sublayer is efficiently modified. Observably, there was progressive expansion in the macro void structures, finger-like projections, as well as spongy sections due to the influence of the Fe 2 O 3 additive (Fig. 2b-d). In the nanocomposite membranes, a handful of horizontal channels were observed, and these channels can enhance the hydrophilicity and possibly increase membrane water permeability [40]. Herein, added PEG polymer serves as a hydrophilic pore-building agent in all membranes and influences the membrane pore morphology. However at 2.0 wt%-Fe 2 O 3 NPs loading, the pores form was greatly repressed and this perhaps is a result of the non-uniform dispersal of NPs [41]. The inhomogeneous dispersion of the Fe 2 O 3 NPs at elevated loading caused the development of aggregated particles in the matrix structure [42]. Hence, the resultants membrane is composed of a dense form with repressed pore sizes as revealed in Fig. 2e. From Fig. 2b and c, uniform dispersion of Fe 2 O 3 NPs was noticed, as well as better interactions and affinities at the loadings between the matrix formulation. Though mild agglomeration was noticed in Fig. 2c.  Despite the compatibility, 1.5 wt%-Fe 2 O 3 modified membranes exhibit a free accretion of NPs within the matrix, which could advance the hydrophilicity and conceivably augment the membrane permeate flux as illustrated in Fig. 3d. However, at higher Fe 2 O 3 NPs loading (2.0 wt%), non-uniform dispersion and a clump of particles appeared (Fig. 3e). Studies have indicated that the existence of clumped particles within the dope matrix raises the viscosity, as well as causes inconsistent nucleation together with irregular crystallization action in the process of phase separation [43]. Figure 4 illustrates the results of the porosity for the pristine and modified hollow fibre nanocomposite membranes at varying Fe 2 O 3 NPs loadings. It can be observed from Fig. 4, that the pristine PVDF/PEG membrane has the least porosity with 64.51%, compared with the nanocomposite membranes. The porosity of the Fe 2 O 3 -modified membranes increased from 69.12 to 84.35% as a result of the Fe 2 O 3 NPs loadings. The upsurge in porosity of composite membranes emerged, which is attributable to enhanced membranes hydrophilicity and it possesses a strong correlation with the quantity of Fe 2 O 3 NPs supplement trapped within the membrane matrix. Notably, the membrane loaded with 1.5 wt% Fe 2 O 3 NPs possessed superior porosity with 84.35%. The elevated porosity indicated ample channels and pores which are beneficial for water permeation. Studies have reported that pore channels of hydrophilic membranes can be facilely wetted by water molecules [44].

Analysis of Membrane Porosity
However, at 2.0 wt% Fe 2 O 3 NPs loading, the porosity of the modified nanocomposite membrane declined to 77.94%, which was caused by a rise in the viscosity of the dope that hindered the interaction rate between the solvent and water with a consequential decrease in pore volume within the membrane matrix [45] and clogging of the membrane pores, consequently resulting in the decline of the Fe 2 O 3 NPs modified nanocomposite membrane porosity. Moreso, the decline in porosity at higher concentrations of Fe 2 O 3 NPs might be attributable to the high viscosity influence and agglomeration of the dope [46]. The porosity results agree with prior studies [47]. Similarly, the pore size expansion via modification, as observed in the morphological images (Fig. 2), will augment the water permeability of the membrane.
Hence, the superfluous loading of Fe 2 O 3 NPs in the dope mixture could undermine the structure of the pores and reduce the porosity and water permeability of the membrane. Further, the mean pore size of the nanocomposite membrane fibres has expanded as compared to the pristine PVDF/PEG membrane as displayed in Fig. 4. This could be associated with the hydrophilic behavior of the Fe 2 O 3 NPs which has a significant role in enhancing the rate of interaction between the solvent-free and the solvent throughout the phase inversion process, enabling the formation of numerous free voids within the membrane [48]. However, at higher Fe 2 O 3 NP loadings, the mean pore size of the modified membrane (i.e., M5) has declined from 65.38 ± 2.2 nm in M4 to ∼62.55 ± 2.0 nm in M5. This perhaps was associated with the increased viscosity of the dope matrix, thereby, hampering the solvent-free and solvent interaction throughout the phase inversion process [49].

Analysis of Membrane Hydrophilicity
The hydrophilic properties play a significant role in the filtration performance of a membrane. Principally, the membrane's hydrophilicity could be determined via the water contact angle. In addition, the lower the contact angle, the higher the membrane hydrophilicity [50]. The results of the water contact angle for the pristine PVDF/PEG and modified membranes with varying Fe 2 O 3 loadings (0-2.0 wt%) is revealed in Fig. 5. As observed from Fig. 5, the pristine membrane has the highest value of water contact angle of 87.3º. Noticeably the water contact angle continues to decline as Fe 2 O 3 loadings increases. The water contact angle of modified membranes declines from 75.6º to 56.3°. Among the nanocomposite membranes, the membrane modified with 1.5wt% Fe 2 O 3 NPs recorded a relatively least water contact angle of 56.3°. Hence, this more hydrophilic property is based on the NPs surface hydroxyl groups (FeO-H) which interplay with water via robust hydrogen bonding as well as chemisorption [51].
Additionally, the magnitude of the contact angle is a result of surface tension that takes place between the membrane and water, which is dropped during the process of sessile drop. Moreso, the reduced interfacial energy due to the exchange of hydrophilic Fe 2 O 3 NPs composite solution to the water/membrane interface throughout the phase inversion operation could also cause a decline in the water contact angle [52]. The increase in weight percent of Fe 2 O 3 NPs loadings allows some composite to surge towards the membrane surface to draw more water than pristine PVDF/ PEG membrane. Hence, PVDF/PEG-Fe 2 O 3 composite membranes offer comparatively superior hydrophilicity, which is favourable for the augmentation of water permeability. However, there is a surge in the water contact angle when the Fe 2 O 3 loading is raised to 2.0 wt%, indicating a decline in the water contact angle. This phenomenon might be due to the agglomeration of Fe 2 O 3 NPs at higher loading. The experimental results obtained in the current study are consistent with the literature [53]. Generally, a lesser contact angle results in superior hydrophilicity as well as enhances the water flux and resistance to fouling [54]. In this scenario, a corresponding increment in the surface hydrophilicity of modified membranes can be ascribed to the strong compatibility of Fe 2 O 3 NPs with water molecules.

Analysis of Surface Charges of Hollow Fibre Membrane
The surface charges of pristine and nanocomposite membranes were investigated to demonstrate the migration and accretion of Fe 2 O 3 NPs at the membrane surface. The study of the membrane surface zeta potential is beneficial to investigate the interactivity between the charged compound and the membrane in the solution stream [55]. Zeta potential describes the nature of charges of the membrane surface and this plays an essential role in contaminants removal rate, and permeation flux together with the anti-fouling strength of the membrane [56]. Thus, it is important to determine the surface charge of the fabricated pristine and nanocomposite hollow fibre membranes within a definite pH range. The results of the surface charges for the fibre membranes with respect to the pH (2-10) range are depicted in Fig. 6. The isoelectric point of the unmodified membrane was observed at 3.22 pH, which is consistent with the literature [57].
The incorporation of Fe 2 O 3 NPs into the matrix dope noticeably transformed the surface negativity of the fibre membranes as a result of the development of oxidation effects as well as acidic oxides [58]. Based on Fig. 6, the surface of the membranes turns out to be more negatively charged with the increase in Fe 2 O 3 NPs loadings through the deprotonation process. In the course of the spinning operation, the NPs reacting with the membrane surface were successfully hydrolyzed to develop some OH − groups subject to the availability of water. Also, the presence of Fe 2 O 3 NPs stimulates the deprotonation action on the surface of the membrane. Consequently, the surface develops more negative charges. Notably, the membrane loaded with 1.5 wt% Fe 2 O 3 NPs possessed the maximum negative value of − 43.7 mV zeta potential, while the unmodified membrane had the least value of − 25.61 mV zeta potential at pH 10 (Fig. 6). The membrane loaded with 1.5 wt% Fe 2 O 3 NPs content recorded superior zeta potential as compared with the membrane with 2.0 wt% Fe 2 O 3 NPs (Fig. 6). Comparatively, the surface zeta potential of modified membranes progressively expands owing to increased NPs loadings from (1.0-2.0 wt%). This augmentation provides an enormous amount of OH − group as well as a broad modification within the membrane structure. In addition, this phenomenon will enhance the filtration capacity as well as advance salt permeation with respect to hydrophobic pollutant rejection. The presence of Fe 2 O 3 NPs with hydroxyl as well as carboxyl groups in the dope formulation promotes deprotonation on the surface of the membrane, thus, making it more negatively charged [59]. Based on the experimental zeta potential result, it can be inferred that to obtain in-depth surface negativity, the concentration of Fe 2 O 3 NPs should be retained between 1.0 and 1.5 wt%.
The augmented negative zeta potential of the modified membranes can expand the electrostatic repulsion between the membrane surface and the negatively charged BPA, thus influencing its rejection performance [60]. In addition, the zeta potential analysis elucidates the possible electrostatic interaction, which depends on the feed solution chemistry.

Analysis of Thermal Stability of Hollow Fibre Membranes
One of the most essential characteristics of membranes is their thermal resistance. A thermal gravimetric study was employed to examine this property. Thermo-gravimetric stability analysis (TGA) of the fabricated pristine and composite membranes was studied according to the weight losses at an increasing temperature (25-600 ℃). Figure 7 shows the TGA curve of the pristine and modified membranes. It can be noticed that there were two major distinct degradation stages (weight-loss) for the pristine and modified membranes modified with Fe 2 O 3 NPs as revealed in the TG curve.
From the TGA curve (Fig. 7) it can be observed that a higher amount of weight loss was observed in the PVDF-PEG pristine membrane, and membranes modified with 1.0 wt%-Fe 2 O 3 , 1.25 wt%-Fe 2 O 3 NPs loadings at a temperature range of 451.6-490.6 ℃ recording a weight loss of 63%, as compared with the 1.5 wt%-Fe 2 O 3 (386-462 ℃) and 2.0 wt%-Fe 2 O 3 (479 ℃) with a weight loss of 12% and 48% respectively. Speedy material loss was observed in the pristine (black line) and a few composite membranes (M2 and M3) between 451 and 490 ℃, due to decomposition. It was also observed from Fig. 7, that the incorporation of Fe 2 O 3 favoured thermal stability, as evidenced by the lower weight loss compared to the pristine PVDF/PEG membrane. The decline in the thermal stability of the pristine and composite membranes modified with 1.0 wt%-Fe 2 O 3 , 1.25 wt%-Fe 2 O 3 NPs loadings (M2 and M3) may be due to lower interfacial interaction between PVDF/PEG and Fe 2 O 3 NPs at an elevated temperature. Above 496 ℃ decomposition rate during the temperature elevation became steady for pristine membrane, membrane modified with 1.0 wt% and 1.25wt%-Fe 2 O 3 NPs membrane samples. The pristine membrane exhibited the highest material loss (over 75% weight reduction) as compared with Fe 2 O 3 loaded composite membranes while the test temperature attained 600 ℃ (Fig. 7). Noticeably, the membrane modified with 1.5 wt%-Fe 2 O 3 NPs recorded the least weight loss of 53.5%, even at the highest degradation temperature of 600 ℃.
The pristine and modified composite membranes at 1.0 wt%-Fe 2 O 3 , and 1.25 wt%-Fe 2 O 3 ) loadings, had the first weight loss region (451.6-490.6 ℃) with 63% weight loss, which can be characterized by the evaporation of the residual solvent and adsorbed water. Similarly, significant decompositions (T d ) occurred with the pristine membrane (black color) and modified membranes (1.0 wt%-Fe 2 O 3 , and 1.25 wt%-Fe 2 O 3 ), and this was owing to poor bonds between the matrix structure which readily dissociate with the temperature rise. In this region, the pristine and modified membranes with (1.0 wt%-Fe 2 O 3 , and 1.25 wt%-Fe 2 O 3 ) possess superior %weight loss than the membrane modified with 1.5 wt% as well as 2.0 wt%-Fe 2 O 3 . All the modified membranes have better and more stable decomposition temperatures (T d ) than the pristine membrane (black curve). Initially, the main decomposition (T d1 ) was noticed at a temperature between 451.6 an 490.6 ℃, and the second decomposition (T d2 ) occurred between 495 and 590 ℃ as shown in Fig. 7. Conversely, at higher decomposition temperatures, the modified membranes (M4 and M5) only demonstrate a nearly steady T d between 496 to 546 ℃ and 395 to 590 ℃ respectively. Hence, the significant increase in the thermal stability was a result of the Fe 2 O 3 NPs which principally enhance the bonds between the matrix structure [61]. At 600 °C, the membrane was modified with 1.0 wt%, Fe 2 O 3 NPs, and 1.25 wt% Fe 2 O 3 NPs loading, and the pristine membrane had 73.43, 70.68, and 75.6% residual weight, respectively. Though, the modified membranes with Fe 2 O 3 NPs loading of 1.5 and 2.0 wt% had lesser residual weight with 53.54% and 56.83% respectively.
Among the modified composite membranes, the membrane modified with 1.5 wt% Fe 2 O 3 NPs yielded the least residual weight loss of 53.54% even at an elevated temperature of 600 ℃. The decline in residual weight might be characterized by the existence of inconsistent crosslinks which yielded susceptible and weak bonds within the structure [62]. Hence, this signifies that uniform dispersion of Fe 2 O 3 NPs is suitable for extending the temperature of decomposition. The obtained TGA result is consistent with existing literature [63].

Analysis of Fe 2 O 3 NPs Stability and Fe Leaching
The leaching of NPs as filtrate from membrane matrix into permeate streams during the filtration process. Hence, this anomaly could compromise membrane functionality and performance and possibly induce pollution as well as environmental risks. The weak adhesion and compatibility between the NPs and polymer chains is one of the underlying challenges sabotaging the viable application of nanocomposite membrane and the leaching out of the NPs is an indispensable index. The ICP results revealed that the concentration of Fe ion leaching from the Fe 2 O 3 composite membrane was about 0.37 ppm during a 48 h process, confirming the better stability of the Fe 2 O 3 composite membrane attributable to strong interactions between Fe 2 O 3 NPs and the PVDF/PEG membrane matrix, which could be very beneficial for industrial separation as well as environmental applications. Figure 8 illustrates the permeability results of the fabricated hollow fibre membranes investigated under varying (TMP) operating pressures (0.3, 0.5, and 0.7 MPa). Noticeably, the permeate flux of all the hollow fibre membranes increased due to a rise in filtration pressure, which could be explained by the higher driving force presented by increased pressures (Fig. 8). Also, the water flux rises with an upsurge in operating pressure based on Darcy's equation [64].

Permeability Assessment
Though the permeate flux of modified nano-composite hollow fibre membranes varied with the loadings of Fe 2 O 3 NPs, compared to the pristine PVDF/PEG membrane. Noticeably, Fe 2 O 3 -modified membranes exhibited higher permeate fluxes with a rise in working pressures. This was perhaps owing to the combination of the enhanced hydrophilicity, and higher porosity of The obtained flux pattern signifies that the addition of Fe 2 O 3 NPs possesses a positive effect on the permeate flux, and this could be by virtue of its capacity to reinforce the hydrophilicity through the deprotonation process in water media to generate an OH − functional group on the membrane surface [65]. Conversely, this phenomenon appears to be different at higher Fe 2 O 3 NPs loading of 2.0 wt% with  (Fig. 8). This may be due to the dense structure and inhibited pores, which revealed the major effect of the excessive agglomerated particles existing in the dope solution [66].
Noticeably, the permeate flux largely relies on the quantity as well as the magnitude of pores available on the surface of the membrane [67]. Given that at an elevated Fe 2 O 3 NPs loading, agglomeration of Fe 2 O 3 NPs resulted in lesser pores, reduced porosity, and decline permeate flux in the composite membrane. Besides, excessive Fe 2 O 3 NPs could stimulate pore clogging, which unfavourably affects the pure water flux of the membrane [68]. Similarly, a rise in operating pressure (TMP) likewise has the propensity to negatively stimulate the compaction of the membrane which could result in a decline in porosity [69]. However, the presence of Fe 2 O 3 NPs within the membrane matrix minimizes the compaction effect on the flux of the permeate. Figure 9 shows the BPA removal efficiency for the pristine and modified membranes under a TMP of 0.3 MPa. Observably, the membrane with 1.5 wt% Fe 2 O 3 NPs loading exhibited superior BPA removal with 86.7% and 5.02 µg/L concentration, while the pristine membrane had 53.6% BPA removal and 18.21 µg/L concentration. It can be observed that the influence of the Fe 2 O 3 NPs blend on the BPA removal are more obvious. The result obtained in this study is consistent with literature studies on the removal of BPA using nanocomposite membranes for water purification [70]. A drop in BPA removal (80.5%) was noticed when the Fe 2 O 3 NPs content was increased to 2.0 wt.% in the matrix dope, in comparison to the membrane loaded with 1.5 wt%-Fe 2 O 3

BPA Removal Performance
NPs. This observable trend was because of the structural and morphological changes of the membrane surface caused by the non-uniform dispersion of the Fe 2 O 3 NPs which stimulated a decline in the efficient interacting surface area necessary to reject the BPA [71].
The accretion of the NPs restricted the hydrophilic influence while the main portion of the matrix was free of the supplement (Fe 2 O 3 NPs). Hence, there exists a partial and erratic removal efficiency of the composite membrane with increased or surplus NPs concentration. Considering this, it could be inferred that attaining a homogenized dispersion of the supplement at a specific concentration is essential to obtaining excellent removal efficiency. The increase and incorporation of Fe 2 O 3 NPs within the matrix dope present the surface more negatively, and this could be because of the deprotonation action [70]. Figure 10 represents the BPA removal mechanism attributed to the surface negativity of the membrane. Generally, microcontaminants easily adsorb to the membrane surface. The BPA removal rate was a result of adsorption onto the membrane. Since adsorption is the principal mechanism for the removal of microcontaminants using an ultrafiltration membrane [72]. Principally, EDCs possess electron-sharing functional groups namely phenolic hydroxyl, carboxylic acid, and a primary amine, while Fe 2 O 3 NPs also exhibit charged end groups [73]. On this basis, there is a potential for a charge transfer complex that enhances the adsorption of BPA by the Fe 2 O 3 NPs-modified PVDF/PEG membrane.

BPA Removal Mechanism
Also, the higher BPA removal is attributed to its high hydrophobicity which led to more beneficial and easier retention on nanocomposite membranes. The reason may not be devoid of the higher Log K ow (3.46) which is greater than 2.0, indicating more affinity to sorption and aggregation. This suggests that the adsorption effect controls the removal of BPA. The principal interaction forces are a result of solvation and hydrophobic effects [73]. The existence of aromatic rings and functional groups in PVDF and BPA molecules  allows such attractive effects (interactions). Considering this, the removal of BPA enhances filtration performance as well as improves fouling mitigation. In recap, the removal of BPA by the UF membrane fibres was principally controlled by the adsorption mechanism between the hydroxyl bonding groups of BPA and the membrane surface containing Fe 2 O 3 NPs. The removal of BPA by the UF membrane fibres via the adsorption mechanism was achieved since the functional groups (aromatic and phenolic) existing in Fe 2 O 3 NPs could selectively adsorb the BPA molecules.
Generally speaking, the BPA molecules adsorbed by the UF membrane could gradually deposit on the membrane surface to form a ''filter cake,'' which could increase the molecular penetration resistance.

BPA Adsorption During Filtration
The pristine PVDF/PEG (M1) and membrane modified with 1.5 wt% Fe 2 O 3 (M4) were further utilized in filtration and the abilities in removing BPA were investigated. As shown in Fig. 11a, the initial percent of BPA removal by M1 was 64.7%, indicating that BPA was adsorbed onto the membrane during the filtration process [74]. In addition, the percent BPA removal by M1 swiftly declined to zero within 60 min, implying that the pristine PVDF/PEG membrane cannot efficiently and effectively remove BPA. Conversely, the removal of BPA by the composite membrane blended with Fe 2 O 3 NPs as a hydrophilic supplement was distinctly superior to that by pristine PVDF/PEG membrane. The percent BPA removal was reduced after 120 min and reached zero at about 210 min, whereby the membrane was saturated with BPA. Figure 11b, shows the adsorption quantity of BPA for the pristine and modified membranes. The result showed that the quantity of BPA adsorbed by M4 was about 22.70 μg/cm 2 , which is 7 times greater than that of M1 (3.85 μg/cm 2 ).

Membrane Antifouling Assessment and Reusability Study
Accumulation of foulants on the pore walls and surface of a polymeric membrane produces fouling. Since various studies have reported that polymeric membrane is susceptible to acute fouling because of the robust adhesive interaction that occurs between the interfaces [75,76]. This phenomenon indicates that subduing the hydrophobic nature via the blending of Fe 2 O 3 nanofillers within the membrane dope formulation and structure can reduce foulants accretion, thus improving the antifouling characteristics. On this note, pristine and modified hollow fibre membranes with varying loadings of Fe 2 O 3 NPs (1.0-2.0 wt%) were exposed to three cycles of the filtration process, The obtained experimental results are demonstrated in Fig. 12. Afterward, the membrane fibres were subjected to washing for 20 min with distilled water following every filtration cycle. Thereafter, from 250 to 700 min, water flux measurements were conducted again. Figure 12 shows the results of fluxes in three cycles of filtration against filtration time. In general, an observable reduction trend in flux was noticed for all the membrane fibres (including pristine as well as modified membranes) with varying loadings of Fe 2 O 3 NPs (0-2.0 wt% with corresponding filtration time as a result of the accumulation of foulants on the surface of membranes. Noticeably, the pristine membrane possessed a flux of 80 L/m 2 ·h, whereas the membrane modified with (1.5 wt%) attained a permeate flux of 173 L/m 2 ·h following 200-min permeation time. The permeate flux further drops to 65 and 140 L/m 2 ·h after 700 min, respectively. Principally, polymeric membranes are more vulnerable to fouling because of the robust adhesive bonding existing between the available interfaces, [77]. The blending and integration of hydrophilic Fe 2 O 3 NPs into the membrane dope formulation could selectively adsorb the BPA molecules owing to the presence of functional groups (aromatic and phenolic) existing in Fe 2 O 3 NPs.
It has been reported that strong antifouling property indicates that flux recovery ratio (FRR) should be as high as possible while relative flux reduction (RFR) should be as low as possible [78][79][80]. Figure 13 illustrates the experimental results of the FRR and RFR for the pristine and modified membrane fibres. The obtained RFR for pristine membrane (M1) and Fe 2 O 3 -modified hollow fibre membranes (M2, M3, M4, and M5) were 32.51, 26.49, 23.27, 17.67, and 20.32%, respectively (Fig. 13). Based on the result of the first cycle, the RFR of the pristine PVDF/PEG (M1) membrane was higher at 32.51%, signifying that the PVDF/PEG membrane suffered more permeability flux decline because of fouling, while the membrane modified with (1.5 wt%-Fe 2 O 3 NPs) recorded the least RFR with 17.67%, which indicates that the composite membranes have superior cleaning efficiency. Inorganic hydrophilic NPs play an essential role in curtailing fouling.
Similarly, the FRR of nanocomposite hollow fibre membranes (M2: 84.92%, M3: 87.38%, M4: 91.57%, and M5: 89.68, respectively) were higher compared to the pristine PVDF/PEG membrane (79.85%). The membrane modified with 1.5 wt%-Fe 2 O 3 NPs (M4) recorded the highest FRR (91.57%) while the pristine PVDF/PEG membrane (M1) had the least FRR value (79.85%). The lower value of FRR for pristine membrane (M1) is because a hydrophobic waterrepellent surface typically lacks ionic, polar groups as well as hydrogen bonding sites, hence indicating a weak attraction between the water molecules and the surface of the membrane binding together [81].
Evidently, upon the completion of the third filtration cycle, the composite membrane with 1.5 wt%-Fe 2 O 3 NPs still attained an elevated %FRR (77.35%) as well as an RFR of 21.29%. Generally, the %FRR, as well as %RFR, declined with respect to the filtration cycles. Noticeably, the first cycle of filtration had the highest %FRR compared with the second as well as third filtration cycles due to fouling. The remarkable antifouling performance of the composite membranes was a result of the successful blending of Fe 2 O 3 NPs within the membrane dope matrix. The incorporation of NPs to the membrane matrix dope could enhance the hydrophilicity of the membrane as well as develop an enormous thin water layer on the surface of the membrane, which curtailed the hydrophobic surface sorption of foulant particles [22]. Among all the modified membranes, the superior FRR was obtained for M4 can be attributed to its augmented hydrophilicity and decreased surface roughness upon the incorporation of Fe 2 O 3 NPs, since a lesser surface roughness indicates lower adsorption [82].
Also, this suggests that uniform dispersion of Fe 2 O 3 NPs within the membrane matrix exhibits enormous expanded interacting surface sites, thereby enhancing antifouling in addition to flux recoverability [83]. Besides, studies have established that the membranes with more hydrophilicity exhibit a lesser propensity to be fouled [84].
As clearly noticed in Fig. 5, Fe 2 O 3 NPs-modified membranes were more hydrophilic in comparison with the pristine PVDF/PEG membrane. The hydrophilic surface of composite membranes attracts a considerable amount of water molecules and develops a compress hydration layer, which restrains the adsorption of foulant droplets on the membrane surface [85]. Membrane hydrophilicity as well as surface roughness are influential variables that control the fouling characteristics of the membrane [86]. Membranes with more hydrophilic surfaces are less susceptible to fouling, owing to the reduced hydrophobic interaction between the surface of the membrane and the foulants [78].
As earlier elucidated in previous sections, the membrane hydrophilicity was enhanced following modification with Fe 2 O 3 NPs. The enhanced antifouling properties of PVDF/ PEG-Fe 2 O 3 NPs composite membranes show the robust influence of enhanced hydrophilicity and increased surface roughness.
This work also demonstrates that the incorporation of inorganic hydrophilic NPs can expand membrane antifouling properties, which is in harmony with the previous studies [87]. Findings from this study agreed with previous reports [88] where various nanoparticles were reported to have favourably enhanced the RFR and FRR of polymeric membranes. Overall, it can be deduced that there exists a noticeable augmentation in the anti-fouling behaviour of the nanocomposite hollow fibre membranes blended with varying loadings of Fe 2 O 3 NPs.

Performance Appraisal with Literature
The findings of the current study was succinctly assessed with previous reports that used metal oxide nanoparticles to transform polymeric membranes. Although the available report on the application of nanocomposite membrane modified with Fe 2 O 3 NPs for removal of BPA from water is very limited, this demonstrates the novelty of this study. Considering this, Fe 2 O 3 NPs were synthesized and utilized to enhance the hydrophilicity and surface negativity of PVDF-PEG membrane for BPA removal using permeate flux, BPA removal, and antifouling as the principal performance indicators. At a concentration of 1.5 wt% Fe 2 O 3 NPs, an outstanding water permeability flux as well as the BPA rejection of 191.96 L/m 2 h, and 86.7% were attained, respectively. Following the 3 cycles of filtration in 700 min, the flux recovery ratio stays relatively consistent at 77.35%.
For instance, Lu et al., (89) investigated the influence of various metal oxides NPs including (CuO, TiO 2 , Fe 2 O 3 , CeO 2 , and MnO 2 ) on ceramic ultrafiltration membranes. For the studied membranes, Fe 2 O 3 NPs exhibited the most hydrophilic with the least fouling tendency, and highest COD rejection (95.4%). The hydrophilicity properties of the studied metal oxides rise with contact angles (26.6º, 29.8º, 52.8º, 63.6º, 71.7º) corresponding with Fe 2 O 3 , TiO 2 , CuO, CeO 2, and MnO 2 respectively. Al-Hobaib et al., (90) incorporated maghemite nanofillers in a mixed matrix membrane (MMM) to study the influence of varying loading of maghemite blended into a polyamide membrane. It was observed that superior hydrophilicity was achieved with the reduction of water contact angle from (74º-29º), increased water flux from 26 to 44 L/m 2 h at 0.3 wt%, and excellent salt rejection of 98%. Karimi & Homayoonfal [91] uses iron oxide (Fe 3 O 4 ) and Zirconium oxide as a composite supplement to enhance the hydrophilicity, porosity water flux, and fouling behaviour of polyacrylonitrile membrane fibres. Their findings indicate that the blending of the composite NPs concurrently, considerably improves the hydrophilicity with a reduction in water contact angle from 38º to 21º, porosity (27-33%) fouling reduction from 22% to zero, 23% greater flux, and cephalexin rejection of 68%. Barati et al. [51] reported a considerable enhancement of surface hydrophilicity, organic rejection, and fouling behaviour when a commercial ceramic membrane was impregnated with in situ iron oxide nanoparticles. Conclusively, this investigation revealed the significance of Fe 2 O 3 NPs in enhancing the surface charges and hydrophilicity properties of polymeric membrane. Thus, it can be inferred from the study that improved performance of the polymeric membrane was achieved.

Conclusions
In recap, the PVDF-PEG membrane fibres embedded with Fe 2 O 3 NPs were fabricated via phase inversion technique and characterized using advanced analytical techniques including SEM, EDX, porosity, contact angle, TGA, zeta potential as well as their potential ability to remove BPA from water. The novel membranes were tested for the treatment of tap water containing BPA. The experimental results revealed that the Fe 2 O 3 nanofillers significantly influence the morphology, surface charges, hydrophilicity, permeability, BPA removal, and antifouling properties of the membrane fibres. The conducted research shows that the permeate flux has increased from 57.39 L/m 2 h in the pristine PVDF/PEG membrane to 191.85 L/m 2 h in the M4 (Fe 2 O 3 -1.5 wt%) composite membrane. It also has to be mentioned that the fabricated PVDF-PEG/Fe 2 O 3 membranes were more negatively charged, thermally stable, and hydrophilic when compared to the pristine PDVF/PEG membrane. Based on the experimental findings of the current study, the Fe 2 O 3 nanofillers significantly influence the morphology, surface charges, hydrophilicity, permeability, BPA removal, and antifouling character of the membranes. It is confirmed that PVDF-PEG/Fe 2 O 3 membranes play an essential role in the BPA rejection and antifouling properties.