Removal of Scale-Forming Ions and Oil Traces from Oil Field Produced Water Using Graphene Oxide/Polyethersulfone and TiO2 Nanoribbons/Polyethersulfone Nanofiltration Membranes

Treatment of produced water in oil fields has become a tough challenge for oil producers. Nanofiltration, a promising method for water treatment, has been proposed as a solution. The phase inversion technique was used for the synthesis of nanofiltration membranes of polyethersulfone embedded with graphene oxide nanoparticles and polyethersulfone embedded with titanium nanoribbons. As a realistic situation, water samples taken from the oil field were filtered using synthetic membranes at an operating pressure of 0.3 MPa. Physiochemical properties such as water flux, membrane morphology, flux recovery ratio, pore size and hydrophilicity were investigated. Additionally, filtration efficiency for removal of constituent ions, oil traces in water removal, and fouling tendency were evaluated. The constituent ions of produced water act as the scaling agent which threatens the blocking of the reservoir bores of the disposal wells. Adding graphene oxide (GO) and titanium nanoribbons (TNR) to polyethersulfone (PES) enhanced filtration efficiency, water flux, and anti-fouling properties while also boosting hydrophilicity and porosity. The PES-0.7GO membrane has the best filtering performance, followed by the PES-0.7TNR and pure-PES membranes, with chloride salt rejection rates of 81%, 78%, and 35%; oil rejection rates of 88%, 85%, and 71%; and water fluxes of 85, 82, and 42.5 kg/m2 h, respectively. Because of its higher hydrophilicity and physicochemical qualities, the PES-0.7GO membrane outperformed the PES-0.7TNR membrane. Nanofiltration membranes embedded with nanomaterial described in this work revealed encouraging long-term performance for oil-in-water trace separation and scaling agent removal.


Introduction
Water scarcity and sustainable water use have long been a global problem, thus industrial wastewater treatment has become a top priority. Produced water, which is a frequent expression in the oil and gas industry for wastewater generated as a byproduct in the oil and gas industry, has been a challenge for many years because there is a legal obligation to dispose of it without harming the environment. The problem is posed by a large amount of produced water, as American oil fields typically produce ten barrels of water for every barrel of oil produced, although the global ratio is more commonly 3:1 [1]. Oil fields can dispose of produced water in a variety of ways, including discharging it in the sea in the case of offshore sites after treatment to comply with environmental regulations that are tightening up, while traditional techniques are insufficient to treat produced water to PVP and TiO 2 to synthesize the UF membrane and employed it for filtration of synthetic oily feed with94.7% oil rejection, 88% salt rejection and 35 kg m −2 h −1 water flux [17]. Yuliwati et al. had an oil rejection of 98% and water flux of 82.5 kg m −2 h −1 using a PVDF-TiO 2 ultrafiltration membrane [18].
Despite some of the mentioned research utilizing GO or TiO 2 in water treatment [10][11][12][13][14][15], however, there is no detailed report focused on the use of NF membranes in the filtration of a realistic produced water sample from an oil field, with a focus on both scale-forming ions and oil content, as both must be treated before being discharged into the environment or disposed into injection wells (to maintain permeability without reservoir bores blockage). Alammar et al. [16] and Mishra et al. [17] employed alternative membrane construction parameters and synthetic oily water rather than genuine oil field samples. Additionally, the salt rejection was investigated in terms of TDS (total dissolved salts), but there was no detailed study of the most significant scale-forming ions (chloride, calcium, magnesium, and barium), which pose a high danger of clogging reservoir pores in water injection wells. Therefore, water samples from oil fields must be used to reflect the true composition of produced water. The produced water includes a variety of compounds such as corrosion inhibitors, biocide, scale inhibitors, oxygen scavengers, H 2 S scavengers, and other chemicals used during drilling and production operations, dissolved and dispersed oil compounds (hydrocarbons), corrosion products, and sulfate reduced bacteria (SRB). These chemicals have the potential to diminish or fail the filtering process and can wreak havoc on filtering efficiency and fouling results. Therefore, this work aims to introduce a category of low-cost nanofillers by incorporating titanium nanoribbons (TNR) and graphene oxide (GO) into the PES membrane to solve for the first time the problem of the contaminated water produced during the oil production industry and to get rid of the oil traces, calcium, chloride, magnesium, and barium ions. This technique can reduce the blocking of the oil reservoir pores and prevent the decrease of reservoir permeability. Therefore, in this study, TNRs and GO nanomaterials are fabricated by hydrothermal and modified Hummer methods and incorporated into PES membranes to increase filtration efficiency and overcome low water flux and membrane fouling. The results of filtration of produced water samples collected from oil fields containing calcium, chloride, magnesium and bromide ions with oil traces are comparatively studied using pure PES, PES-GO, and PES-TNR in terms of filtration efficiency, water flux, and fouling tendency. In other words, the filtration of realistic water samples from oil fields has been investigated in this study to prove the usefulness of this method to be directed to manufacturing and play a role in protecting the environment and solving the produced water problem in the oil production industry. Moreover, the advantages of the manufactured membranes become obvious by comparing salt and oil rejection results with a commercial membrane.

Preparation of GO Nanostructure
Graphene oxide nanoparticles were fabricated by a modified Hummer's method by using 1 g of graphite powder added to a mixture of 120 mL 96%wt H 2 SO 4 and 14 mL 85%wt H 3 PO 4 in an ice bath, 6 g of KMnO 4 powder added slowly over 1 h on a magnetic stirrer, then the solution was left for 1 day at 50 • C in a water bath. The combination of acid and potassium permanganate will result in the intercalation of acid in the graphite layer and oxidation into graphene oxide flakes. Following that, 800 mL deionized water Polymers 2022, 14, 2572 4 of 20 and 10 mL H 2 O 2 were added drop by drop alternately in an ice bath to reduce the residual of the KMnO 4 , resulting in a lot of air bubbles and a reddish yellow color, and the solution was left to settle for 24 h before being filtered. Finally, the precipitate is washed with deionized water till the solution becomes neutral (pH = 7) and then dried at 60 • C to obtain GO powder [19].

Synthesis of TNRs
Synthesis of titanium nanoribbons was carried out using a hydrothermal technique by adding 4 g of titanium dioxide to 500 mL sodium hydroxide with a concentration of 10 M, then, using a magnetic stirrer, the mixture was stirred for half an hour untiled homogenous solution appeared. Then, the mixture was poured into a 1 L Teflon-lined stainless-steel autoclave. The autoclave was placed in the oven for 24 h at 170 • C and after cooling the resultant powder was rinsed with 0.1 M hydrochloric acid and distilled water. The obtained white powders were left to dry at 80 • C for 24 h before calcination at 450 • C for 2 h [20].

Synthesis of PES-GO and PES-TNR Membranes
Membranes were synthesized by the phase inversion method [21] by adding nanomaterial powder TNR in the case of PES-TNR membranes and GO in the case of PES-GO membranes to 50 mL DMF gradually. The quantity of nanostructured powder was added according to the reported weight% in Table 1. The powder was added during the stirring for~3 h followed by the addition of DMF to reach a 100 mL mixture. The mixtures were kept under stirring for over 24 h until they formed a homogenous emulsion. Then 17.5 g of PES powder was added gradually to the emulsion over an ultrasonic vibrator at 45 • C. Then, stirring was continued for 12 h. After that, the emulsion was cast on a clean glass plate using a thin-film casting device with 120 µm thickness. Finally, the cast membranes were immersed in distilled water at 45 • C to be hardened then removed and dried in the air [22]. Membrane hydrophilicity before and after adding nanomaterials was determined by measuring the contact angle using the sessile method by contact angle analyzer (KRÜSS, DSA100, Hamburg, Germany) for the synthesized membranes.
The porosity of the synthesized membranes, which is pores volume divided by total membrane volume, was determined by immersing the membrane in distilled water. Then, excess water was removed using filter paper and the weight (W 1 ) of the membrane was measured. Then, the membrane was left to dry for 4 h at 85 • C and the weight of the dry membrane (W 2 ) was measured [22]. The membrane porosity can be estimated by Equation (1): where ε m is porosity (%), W 1 and W 2 are the weight of the wet and dry membranes(in grams), ρ w is the density of pure water (0.999 g cm −3 ) and ρ p is the density of PES (1.37 g cm −3 ). The mean pore radius was then determined using the filtration velocity method according to the Guerout-Elford-Ferry Equation (2): where η is the water viscosity (8.9 × 10 −4 Pas), Q is the volume of the permeate pure water per unit time (m −3 ·s), l is the membrane thickness (m) and ∆P is the operation pressure 0.3 MPa.

Morphologies and Structures
Field emission-scanning electron microscopy (FE-SEM, Model quanta 250 field emission gun, FEI, Shinagawa, Japan) and a transmission electron microscope (TEM JEOL, Model 2010, Tokyo, Japan) were used for the characterization of membrane morphologies. The TNR's and GO's crystallographic structures were investigated using X-ray diffraction (Philips X'Pert Pro MRD, Malvern Panalytical Ltd., Malvern, UK) by an EMPYREAN diffractometer with Cu Kα radiation operated at 30 mA and 40 kV in the 2θ range 5-80 • .
2.6. Membrane Performance 2.6.1. Filtration Process The filtration procedure was carried out using a pressurized system as shown in Figure S1 (Supplementary data) that included a pressurizing pump that sucked the produced water sample and pumped it to the membrane cell at a discharge pressure of 0.3 MPa. Many bench-scale experiments of filtration membranes with variable applied pressure have employed a pressurizing pump [8,[23][24][25]. For dye and salt separation, Fang et al. employed a peristaltic pump with a low-pressure feed of 0.2 MPa for nanofiltration employing the PES-iron ion complex [26]. For water treatment, Hudaib et al. employed a ceramic TiO 2 nanofiltration membrane with a feed pressure of 0.2 MPa [27]. For wastewater treatment, Amin et al. utilized 0.27 MPa on PES/GO nanoribbons and PES/GO nanosheets [28]. Filtration was performed three times using a circular flat sheet specimen with a surface area of 28.2 cm 2 and a feed pressure of 0.3 MPa. The filtration system has a three-opening flat sheet membrane module for feeding, concentrating, and permeating [29]. Using a high-pressure pump, the feed was continually delivered to the membrane module from a closed feeding tank (50 L). The permeated water was then collected from the membrane modules downstream. The findings of manufactured membranes were compared to those of a commercial nanofiltration membrane, NTR-7450. Nitto-Denko of Switzerland provided the NTR-7450 membrane. The top layer of the NTR-7450 membrane is sulfonated polyethersulfone (SPES), which is covered with an ultrathin polymeric composition layer, according to the manufacturer. The top barrier's semi-permeability is determined by its properties (solvent flux and solute rejection). SPES contains aromatic rings, oxygen, and sulfonic groups [30].

Water Flux
The water flux of the synthesized pure-PES, PES-TNR and PES-GO membranes was investigated by using deionized water as feed water and observing the quantity of permeate every 5 min, where there was a decline in permeate quantity for the first hour, then the permeability rate increased and remained constant. The quantity of water was measured when the rate became constant [31]. Water flux (J w ) was determined using the following equation. where J w is pure water flux (kg/m 2 h), M is mass of permeate (kg), A is the effective area of membrane (m 2 ) and ∆t is the filtration time (h).

Filtration Efficiency
Filtration was performed on a sample of produced water taken from an oil field in the West Bakr Company in Egypt's western desert, with chemical concentrations as listed in Table 2. Filtration efficiencies of produced water from its constituent ions were investigated by comparing the concentration of chloride, calcium, magnesium, and barium before and after filtration. After one hour of commencing the filtering process, permeate water samples were taken when the water flux became steady. Ion concentrations were determined according to ASTM D-4327 for chloride anions and ASTM D-6919 for calcium, magnesium, and barium cations, respectively, using ion chromatography. The Dionex ICS 1100 was used, which was outfitted with high-capacity columns (AS9 and CS12) for anions and cations [32][33][34]. The efficiency of produced water filtration from oil in water content was investigated by measuring oil in water content using the Eracheck analyzer device according to the ASTM D-7678 standard method [35]. Analyses were carried out by adding 5 cyclohexane to 45 mL of tested sample and aggressively shaking the mixture, then allowing it to rest until two phases of water at the bottom and the hydrocarbon-Cyclohexane mixture at the top were separated. The top layer was then withdrawn with a syringe and combined with 0.1 g sodium sulfate as a drying agent to remove any water traces. Finally, the extracted hydrocarbon portion was submitted to an infrared device with wave number 1370-1380 cm −1 and multiplied results with the concentration factor. The concentration factor is determined by dividing the sample volume by the solvent volume (concentration Factor = 9). Salt and oil rejection (R%) is determined using the following equation: where Co is feed ion concentration or oil content (as indicated in Table 2), and C is ion concentration or oil content in the permeate.

Fouling Tendency
Fouling tendencies of produced water were investigated by a dynamic method using produced water with the same composition and oil in water concentration as a fouling agent. The first filtration using synthesized membranes using deionized water was performed for one hour at a discharge pressure of 0.3 MPa and water flux was recorded (J w1 ), then produced water was used to perform filtration at the same parameters of time and pressure. The membranes were then washed with 1 M NaOH at 0.3 MPa and water for 30 min. The cleaning chemical NaOH was chosen because of its quick cleaning outcomes and overall cleaning effectiveness [36]. Filtration used pure water while recording related water flux (J W2 ), the flux recovery ratio (F RR ) [37] was calculated using the following equation: To evaluate the alkaline resistance of the membranes, PES-0.7 GO and PES-0.7 TNR membranes were immersed in 1 M NaOH for three days, and water flux and salt rejection were compared to previous results.
Double successive filtrations were carried out to see if it could result in increased filtration efficiency. In a comparative study, the salt rejection and oil rejection for single filtration and double-successive filtration were assessed.

Hydrophilicity of Synthesized Membranes
The sessile drop method [38] is applied to determine the contact angles of the synthesized membranes by measuring the angle between a deionized water droplet and a dry membrane surface immediately after dropping the droplet with a 'I'-shaped needle on the membrane surface. The process is repeated five times and the average of the readings is reported [39]. Figure 1A shows that increasing the nanomaterial percentage in the membrane results in lower contact angle readings, indicating that increasing the nanomaterial percentage increases hydrophilicity [40,41]. Pure-PES membranes have the largest contact angle and are the least hydrophilic. The hydrophilic nature of the TNR [42] and GO nanomaterial is primarily responsible for this considerable shift. The hydrophilicity of GO results from the presence of C=O, C-O-C, and C-OH functional groups [43]. During the NIPS (Non-solvent induced phase separation) process, nanomaterial spontaneously moves to the polymer/water interface to reduce the interface energy. After the membranes have been fully formed, the majority of the nanomaterial particles migrate to the membrane surface and stabilize, enhancing hydrophilicity. Figure 1B shows that increasing the ratio of nanomaterial increases the porosity of the membranes. The microstructure of generated membranes is influenced by the thermodynamic and kinetic components of the phase inversion process. Nanomaterial was used to increase porosity. The thermodynamic effect of adding nanomaterial resulted in uniform dispersion and the hydrophilic character of the added nanomaterial increased hydrophilicity and accelerated non-solvent adsorption, increasing the polymer film's thermodynamic instability. During the phase inversion process, quick demixing resulted in the production of membranes with a larger porosity. During non-solvent induced phase separation (NIPS), the high affinity of the solvent to cover embedded nanomaterial compared to polymer chains results in the creation of membranes with a larger porosity [44]. The membrane pore size is affected by the phase transition velocity of water-solvent during phase inversion. The mean pore size is estimated using Equation (2) and presented in Figure 1B for all studied membranes. The mean pore size of PES-GO and PES-TNR membranes increased as the amount of GO and TNR added increased, as shown in Figure 2B, which is due to the hydrophilicity properties of GO and TNR, which is compatible with the contact angle and porosity results. This is also consistent with the findings of Nurul et al. [45], who found that additive loadings increased the hydrophilicity and surface porosity of membranes. The PES-GO and PES-TNR membranes developed in this work had a reduced contact angle, higher porosity, and larger mean pore size, which resulted in superior hydrophilic characteristics.

Membrane Morphology
Figure 2A-C shows a cross-sectional and top view SEM image of the pure-PES membrane matrix with micro pits of near-identical sizes spread throughout. These SEM images illustrate the spongy-like structure of the pure-PES membrane. The SEM image in Figure  2C shows an asymmetric membrane with top and bottom porous surfaces. The cross-sectional SEM images in Figure 2D,F for PES-0.7GO indicate a denser and less porous top layer of the membrane, although the formation of macro voids was present in the sublayer. Macro void lengths ranged from 8 to 14 μm. Figure 2E depicts the top morphology of the PES-GO membrane. The number of pores increased, and their diameters decreased when GO was introduced [45]. Figure 2G,I shows cross-sectional SEM images for the PES-0.7TNR membrane with macro voids ranging from 12 to 18 μm. Figure 2H shows the top layer of the PES-TNR membrane, whereas some of these pores exhibit coalescence. The insertion of TNRs results in the growth of lateral, larger pores at the top layer and macro voids in the sub-layer [25]. The overall thickness ranges from 80 to 100 μm for both PES-GO and PES-TNR membranes.   Figure 2C shows an asymmetric membrane with top and bottom porous surfaces. The cross-sectional SEM images in Figure 2D,F for PES-0.7GO indicate a denser and less porous top layer of the membrane, although the formation of macro voids was present in the sub-layer. Macro void lengths ranged from 8 to 14 µm. Figure 2E depicts the top morphology of the PES-GO membrane. The number of pores increased, and their diameters decreased when GO was introduced [45]. Figure 2G,I shows cross-sectional SEM images for the PES-0.7TNR membrane with macro voids ranging from 12 to 18 µm. Figure 2H shows the top layer of the PES-TNR membrane, whereas some of these pores exhibit coalescence. The insertion of TNRs results in the growth of lateral, larger pores at the top layer and macro voids in the sub-layer [25]. The overall thickness ranges from 80 to 100 µm for both PES-GO and PES-TNR membranes.

Membrane Morphology
The hydrophilic character of the added GO and TNR nanomaterial, which promotes solvent and non-solvent mass transfer rates, is credited with the creation of nanopores and macro voids [46][47][48]. According to SEM micrographs, graphene oxide nanoparticles were efficiently dispersed in the polymer matrix and did not show significant aggregation on the membrane surface, which may be attributed to the graphene oxide carbon-based structure [46]. Anna Rabajczyk et al. [47,48] demonstrated that dispersing TNR or GO in DMF (solvent) while stirring prevents nanoparticle agglomeration.
Transmission electron microscopy (TEM) has long been recognized as a helpful method for describing the morphology of GO and TNR. TEM images, Figure 3A,B, of the calcinated TNR illustrate the influence of the calcination temperature on the surface of the produced nanoribbons. The conversion of hydrogen titanate to TiO 2 -B NRs resulted in the formation of pores with diameters ranging from 2 to 10 nm [49]. These pores are formed because the interlayered OH groups escape the structure [50,51]. Figure 3C illustrates the graphene oxide plates as flat surfaces. The effective synthesizing of the GO nanostructure is confirmed by the TEM image in Figure 3C. By oxidizing graphene extensively, the graphitic stack's flakes were chemically exfoliated into mono-and multilayer sheets [52,53]. The hydrophilic character of the added GO and TNR nanomaterial, which promotes solvent and non-solvent mass transfer rates, is credited with the creation of nanopores and macro voids [46][47][48]. According to SEM micrographs, graphene oxide nanoparticles were efficiently dispersed in the polymer matrix and did not show significant aggregation on the membrane surface, which may be attributed to the graphene oxide carbon-based structure [46]. Anna Rabajczyk et al. [47,48] demonstrated that dispersing TNR or GO in DMF (solvent) while stirring prevents nanoparticle agglomeration.
Transmission electron microscopy (TEM) has long been recognized as a helpful method for describing the morphology of GO and TNR. TEM images, Figure 3A,B, of the calcinated TNR illustrate the influence of the calcination temperature on the surface of the produced nanoribbons. The conversion of hydrogen titanate to TiO2-B NRs resulted in the formation of pores with diameters ranging from 2 to 10 nm [49]. These pores are formed because the interlayered OH groups escape the structure [50,51]. Figure 3C illustrates the graphene oxide plates as flat surfaces. The effective synthesizing of the GO nanostructure is confirmed by the TEM image in Figure 3C. By oxidizing graphene extensively, the graphitic stack's flakes were chemically exfoliated into mono-and multilayer sheets [52,53].   [55]. However, for TiO2, the (−403) peak is the most intense, while for NaTiO2, the (−211) peak is the most intense. Figure 4B shows the tetragonal GO crystalline phase in GO nanoparticles as seen on the XRD chart. The presence of a tetragonal GO phase along the (001), (220), (221), and (322) orientations is indicated by the presence of a GO peak at 2theta of 10.5°, 21°, 26.7°, and 42.3° (card No. 96-590-0025) [56], the (10.5°) peak is the most intense. This could be attributed to the fluctuation of equilibrium surface concentration in the direction of minimizing surface energy or setting the segregation energy near zero.   [55]. However, for TiO 2 , the (−403) peak is the most intense, while for NaTiO 2 , the (−211) peak is the most intense.  Using values of W (full width at half maximum) for XRD peaks, the average crystallite size of TiO2, NaTiO2, and GO is estimated using Scherrer's equation:  Figure 4B shows the tetragonal GO crystalline phase in GO nanoparticles as seen on the XRD chart. The presence of a tetragonal GO phase along the (001), (220), (221), and (322) orientations is indicated by the presence of a GO peak at 2theta of 10.5 • , 21 • , 26.7 • , and 42.3 • (card No. 96-590-0025) [56], the (10.5 • ) peak is the most intense. This could be attributed to the fluctuation of equilibrium surface concentration in the direction of minimizing surface energy or setting the segregation energy near zero.

Structural Properties
Using values of W (full width at half maximum) for XRD peaks, the average crystallite size of TiO 2 , NaTiO 2 , and GO is estimated using Scherrer's equation: where Ds, θ, and λ are the crystallite size, Bragg's angle, and X-ray wavelength, respectively [57,58]. Texture coefficients (TCs) and dislocation densities [59] are also calculated.

Evaluation of Filtration Efficiency
For this investigation, ion concentration at feed and permeate water were measured. The permeated water was subsequently tested by ion chromatography in the same way, yielding the findings shown in Table 4. Figure 5 shows the salt rejection of the various membranes with close values, with the maximum efficiency forPES-0.7GO, followed by PES-0.7TNR, and the lowest efficiency for pure PES. PES with a higher nanomaterial ratio has a higher salt rejection, which can be interpreted as higher hydrophilicity. This finding is in line with that of Zylla et al., who discovered that decreasing nanofiltration membrane contact angles enhanced salt rejection [61]. Sulfonic and ether groups were commonly found in the PES layer on top of the NF membrane. In addition, the skin layer has mean radius pores of 6.5-11 nm. As a result, a complicated process including electrostatic contact and steric hindrance was responsible for the separation of solutes by NF membranes. When comparing findings to commercial NTR-7450 membrane salt rejection, it seems that the NTR-7450 membrane has somewhat higher salt rejection than pure PES and slightly lower salt rejection than PES-TNR and PES-GO. The hydrophobic character of the commercial membrane can be linked to the reduction in salt rejection at NTR-7450 [54]. The commercial membrane is covered with an ultrathin polymeric layer, which results in the membrane being hydrophobic despite the presence of sulfonic acid groups. The results show that membranes may be used at high concentrations of generated water. GO-toluidine/PES membranes yielded encouraging findings for Zhongzhen et al. [62]. Increased hydrophilicity causes the diffusion rate of water molecules to rise, whereas salt ions are trapped in pores.   Table 5 shows the removal of oil from water utilizing synthesized membranes. The concentration of oil in water at feed water was 120 ppm. The oil rejection increased from 71% to 88% by increasing the ratio of GO content from 0% to 0.7% and increased to 85%  Table 5 shows the removal of oil from water utilizing synthesized membranes. The concentration of oil in water at feed water was 120 ppm. The oil rejection increased from 71% to 88% by increasing the ratio of GO content from 0% to 0.7% and increased to 85% by increasing the TNR to 0.7%, i.e., the highest oil in water rejection (88%) was reported for PES-0.7 GO versus 85% for PES-0.7 TNR and71%for the pure PES, which is considered the lowest reported value for our prepared membranes. This behavior is reversely correlated to the values of the contact angle. This suggests that the difference in filtering effectiveness is due to membrane hydrophilicity, which produces repulsion between hydrophilic membrane surfaces and hydrophobic oil droplets, allowing more water to enter the pores [63]. As observed in SEM images in Figure 2, the integration of 0.7% GO or 0.7%TNR in the PES membrane also increases pore density and porosity, which can contribute to the increase in oil rejection. Furthermore, as demonstrated in Table 5, our optimized membranes have a higher oil rejection than the commercial NTR-4750 membrane, which has 40% oil rejection. This is due to the commercial membrane's hydrophobic nature.

Water Flux and Flux Recovery Ratio
Water flux of the synthesized membrane was evaluated by observing the permeate volume per unit time at a discharge pressure of 0.3 MPa. Figure 6 demonstrates that the PES-0.7GO membrane had a higher water flux than the PES-0.7TNR membrane, while the pure-PES membrane had the lowest water flux. This is owing to the PES-0.7GO membrane's higher hydrophilicity than the PES-0.7TNR membrane, while pure PES has the lowest hydrophilicity [40,64]. The ability of water molecules to permeate across the selective layer of the membrane is increased by increasing the hydrophilicity. It might also be attributed to the inclusion of nanostructures, which results in the creation of macro voids and an increase in porosity and mean pore radius. The increase in porosity and mean pore radius after introducing nanomaterial is consistent with the results of water flux [65]. The findings of employing GO as a nanofiller material for nanofiltration membranes are consistent with the reported data by Hao et al. [25] and Amin et al. [61]. TiO 2 nanomaterial, which is utilized in nanofiltration, also improves the water flux [66][67][68].
The fouling tendency of produced water was explored utilizing a flux recovery ratio (FRR) based on pure water flux before (Jw 1 ) and after (Jw 2 ) using synthesized membranes in produced water filtration. The higher FRR value indicates a higher antifouling tendency to get lower fouling. When comparing the antifouling tendency of pure-PES, PES-TNR, and PES-GO membranes, PES-0.7GO outperforms PES-0.7TNR, while pure PES has the lowest antifouling tendency. The hydrophilic interactions between the membrane surface and feed solution constituents increase electrostatic forces, resulting in improved antifouling. One of the most common methods for developing antifouling membrane surfaces is to increase the surface hydrophilicity, for example, by adding hydrophilic monomers or nanoparticles [66].
to get lower fouling. When comparing the antifouling tendency of pure-PES, PES-TNR, and PES-GO membranes, PES-0.7GO outperforms PES-0.7TNR, while pure PES has the lowest antifouling tendency. The hydrophilic interactions between the membrane surface and feed solution constituents increase electrostatic forces, resulting in improved antifouling. One of the most common methods for developing antifouling membrane surfaces is to increase the surface hydrophilicity, for example, by adding hydrophilic monomers or nanoparticles [66].

Alkaline Resistance of Membranes
Only a few commercially available NF membranes, notably PES-based membranes, can tolerate high caustic conditions. The salt rejection and water flux stabilities in an alkaline environment were used to estimate the membranes' alkaline resistance [69,70]. Both PES-0.7-GO and PES-0.7TNR demonstrated consistent salt rejection and water flux after three days of immersion in 1 M NaOH solution, as illustrated in Figure 7A,B. Both PES-0.7GO and PES-0.7TNR may be used as alkali-resistant NF membranes with increased filtering performance, according to this discovery. Due to the presence of aromatic groups, alternating sulphonyl (SO 2 ) groups, and ether (-O-) linkages, PES has great thermal stability, a high glass transition temperature, and strong structural and chemical stability [71,72]. These findings are in line with those of Mehwish et al., who discovered that TiO 2 NP can improve the chemical stability of PES membranes [66]. Yeh et al. have shown the remarkable stability of GO membranes in water [67]. The successful use of 1 M NaOH as a cleaning agent with a high FRR demonstrates membrane resilience to cleaning. The chemical stability of PES, which is attributable to the sulfonyl and ether groups, contributes to cleaning resistance. ity, a high glass transition temperature, and strong structural and chemical stability [71,72]. These findings are in line with those of Mehwish et al., who discovered that TiO2 NP can improve the chemical stability of PES membranes [66]. Yeh et al. have shown the remarkable stability of GO membranes in water [67]. The successful use of 1 M NaOH as a cleaning agent with a high FRR demonstrates membrane resilience to cleaning. The chemical stability of PES, which is attributable to the sulfonyl and ether groups, contributes to cleaning resistance.

Evaluation of Successive Filtration
The antifouling capabilities of PES-GO and PES-NR membranes, as well as the possibility of enhancing salt and oil rejection, were investigated further using a dynamic cyclic filtering experiment for generated water samples. Figure 8A demonstrates that in the second filtering cycle, the total rejection of ions and oil increased. Figure 8B shows how the fouling impact of produced water lowers water flux from 78 to 70 kg·m −2 h −1 for PES-0.7GO and 74 to 69 kg·m −2 ·h −1 for PES-0.7TNR during the second filtering cycle. The high flux recovery ratios of PES-0.7GO (89.7%) and PES-0.7TNR (93.2%) are demonstrated by the minor decrease in water flux. The limited decrease in water flux, salt rejection, and oil rejection refers to a very limited leaching effect for the used nanomaterials. The findings are in line with those of Li et al., who discovered that nanomaterial leaching from PES-TiO 2 membranes was low for the first five hours, then stopped, whereas leaching was only strong at PES membranes substantially laden with nanomaterial [73]. Additionally, Cheng et al. observed that employing a GO-PES membrane grafted with polyampholyte hydrogel results in very low graphene oxide leaching, with less than 2% graphene oxide leached from the membrane [74]. This finding encourages businesses to use membranes to treat generated water in oil fields in order to comply with environmental regulations. are in line with those of Li et al., who discovered that nanomaterial leaching from PES-TiO2 membranes was low for the first five hours, then stopped, whereas leaching was only strong at PES membranes substantially laden with nanomaterial [73]. Additionally, Cheng et al. observed that employing a GO-PES membrane grafted with polyampholyte hydrogel results in very low graphene oxide leaching, with less than 2% graphene oxide leached from the membrane [74]. This finding encourages businesses to use membranes to treat generated water in oil fields in order to comply with environmental regulations.

Conclusions
Nanofiltration membranes made of PES, PES-GO, and PES-TNR were created using the phase inversion approach. Membranes were used to filter generated water samples taken from an oil field in order to create a realistic atmosphere. The addition of nanomaterials increased membrane performance in several ways. Membranes' physiochemical characteristics were studied. The addition of TNR or GO to the PES matrix causes macro voids to form on the bottom surface of the membrane in both PES-TNR and PES-GO. When the weight of the embedded nanostructure is increased, the contact angle decreases from 81 • for pure PES to 40 • and 42 • for PES-0.7GOand PES-0.7 TNR, respectively. Porosity rose from 17% for pure PES to 44% for PES-0.7TNR and 42% for PES-0.7GO, respectively, while water flux increased from 42.5 kg/m 2 h to 85 and 82 kg/m 2 h. By increasing the weight percent of integrated nanomaterial, chloride, calcium, magnesium, and barium concentrations were reduced while water flux was increased. The PES-0.7GO membrane has the highest filtering efficiency, followed by the PES-0.7TNR and pure-PES membrane, with chloride salt rejection of 81%, 78%, and 35%, respectively. Furthermore, increased hydrophilicity is the main cause of increased oil rejection using the membranes of PES-0.7TNR (88%) and PES-0.7GO (85%) over the pure-PES membrane (71%) and commercial NTR-7450 (40%) due to repulsion forces between hydrophobic oil traces and the hydrophilic nature of nanostructure-incorporated membrane surfaces. Fouling tendency decreased with the addition of nanomaterial, with the lowest fouling tendency at the PES-0.7GO membrane, followed by the PES-0.7TNR membrane, and the greatest fouling tendency at the pure-PES membrane, which may be ascribed to fouling mitigation through enhanced hydrophilicity. Cleaning membranes with NaOH was tried before measuring flux recovery ratio, and the FRR was found to be quite high. Membrane resistance to alkali after three immersions demonstrated the membrane's great resistance to alkaline media. Overall, the findings support the use of membranes in the oilfield to treat generated water.