Fabrication and Characterization of PAA‑g‑PES Polymers and Resultant Membranes: Combating Fouling through the Introduction of Hydrophilic PAA Brushes

Abstract

The adoption of polymeric membranes for reclamation of natural organic pollutant-impacted water has been of recent interest. However, membrane fouling has been an obstacle in the application of membranes over a longer period. Fouling, which is mainly caused by organic materials, is a universal challenge in membrane technology for water treatment. The hydrophobic nature of poly­(ether sulfone) (PES) membranes has a high affinity toward organic compounds. To elevate this challenge, in this study, hydrophilic brushes of poly­(acrylic acid) (PAA) were grafted onto the PES backbone using free radical graft polymerization. The successful grafting of PAA brushes on PES was confirmed through characterization using Fourier transform infrared spectrophotometry, ¹H NMR, ¹³C NMR, and thermogravimetric analysis, with a grafting yield ranging from 2.11 to 6.7%. The thermal stability, functionalities, morphology, wettability, and surface roughness of the PAA-g-PES membranes were also investigated as these characteristics play an important role in the overall performance of the membrane system. The improved hydrophilicity was seen through the observed decrease in the membranes’ water contact angle (77° for the pristine PES membrane to 46° for the highest modified PAA-g-PES membrane) in response to the increased PAA content. The antifouling propensity of the fabricated PAA-g-PES membranes was tested through the seven-cycle fouling–backwashing processes using bovine serum albumin (BSA) model protein as well as the river water sourced from Centurion, South Africa, as the feeds. It was observed that the flux recovery percentages after the seventh cycle (6.5 h) were 20.7 and 25.5% for the pristine PES membranes and 55.3 and 59.8% for the PAA-g-PES membrane with the highest grafting yield for the BSA-laden water and the river water, respectively. The improved recovered flux suggests that the PAA-g-PES properties enhanced the fouling resistance of the modified membranes, giving the membranes a better chance of applicability in a real-life filtration system.

1. Introduction

Increasing river water contamination from illegal discharge and dumping contributes to a decrease in water quality, potentially resulting in negative health effects in many communities around the world. This is caused by the release of polluted or incompletely treated water into river water bodies by various industries such as pharmaceuticals, textiles, metals, agriculture, etc. − Membrane technology is among the technologies that offer great promise in producing clean water of acceptable quality. This technology is very selective, flexible, dependable, and efficient for eliminating toxins from wastewater. Its flexibility is primarily influenced by the membrane’s composition and pore size. − The major drawback of the membrane technology is fouling, which basically happens when the pollutants are deposited on the membrane surface or inside its pores. Numerous approaches have been explored to prevent membrane fouling, and those include traditional methods (pretreatment of the feed solution, backflushing, cleaning with harsh chemicals, and air sparging) and antifouling strategies. , The antifouling strategies that have been explored comprise zwitterion coating, enhancing surface hydrophilicity, photocatalytic cleaning, combining the fouling-resistance and fouling-release mechanism, electrically enhanced antifouling, hydrophilic dynamic membranes, and magnetic pickering emulsions for fouling-free separation. Additionally, integrating membrane systems with other technologies is a practical consideration in solving the membrane fouling phenomenon, which is a bottleneck in its widespread adoption for various types of polluted water and wastewater. ,

The fabrication of superhydrophilic, smooth, and neutral membrane surfaces is a strategy utilized to efficiently reduce the interaction between the membrane surface and the foulants, thereby reducing fouling. Hydrophilic materials are introduced to membrane surfaces by surface grafting through the anchoring of polymers via a chemical covalent reaction; this process is reported to be achieved through several surface activation processes. The challenge in surface grafting is limited to flat-sheet membranes, whereby direct activation through external sources, if practical, results in variable and unpredictable grafting efficiencies and control. Therefore, the modification of polymers via grafting before membrane fabrication is viewed as a better way of gaining control of the grafting process, leading to predictable membrane properties and outcomes. Several methods for introducing hydrophilic side chains onto the PES polymer have been reported. These mostly require, first, the functionalization of PES with a reactive group that is subsequently reacted with a hydrophilic polymer through the “click” chemistry of similar approaches. These additional steps have an impact on the complexity of the modification process. Direct modification using the radical addition of monomers with controlled growth of the side chains is more attractive. Such direct methods are exemplified by direct radical polymerization using either benzoyl peroxide (BPO) or persulfate radical methods or living radical polymerization using catalysts such as atom transfer reduced polymerization (ATPR). , One of the most frequently used hydrophilic side chains is poly­(acrylic acid) (PAA). PAA is a synthetic polymer of acrylic acid consisting of carboxylic groups on every second carbon atom of its primary main chain. This cheap, easily processable, nontoxic polymer is characterized by ionizable hydrophilic properties, high acid/base and water absorption capabilities, and a high negative charge density associated with the carboxylate group in the PAA chain, making it a good candidate for membrane applications. −

Herein, the chemical grafting of AA onto PES using radical initiation is revisited with a view to optimizing the polymer grafting and subsequently studying the effects of low grafting yields on the performance of the resultant membranes. , It was envisaged that the grafting of PES with PAA would impart its stated positive attributes, resulting in the long-term stability of the resultant membranes in practice. This report details the investigation of the effect of different grafting contents of PAA on PES on pure water flux, protein rejection, and the antifouling profile relative to those of the pristine PES membrane. Specifically, the fouling and fouling recovery processes were followed using surface FTIR and SEM analyses of fouled and backwashed membranes using BSA as a probe foulant and river water.

2. Results and Discussion

2.1. PAA-g-PES Polymer Characterization

The successful grafting of PAA on the PES polymer was determined using a combination of FTIR, ¹H NMR, ¹³C NMR, and TGA as indicated earlier, and the results are discussed below.

2.1.1. FTIR of the Grafted Polymer, PAA-g-PES

The grafting of PAA on the PES polymer was studied using ATR-FTIR spectroscopy, as illustrated in Figure . Successful grafting was determined through the analysis of characteristic bands attributed to specific functional groups on the unmodified PES, AA, and target product PAA-g-PES. For instance, the unmodified PES polymer displayed characteristic bands in line with prior reported literature, i.e., the sulfonyl group at 1144 and 1102 cm^–1, the 1,4-disubstituted aromatic benzene rings at 1484 and 1577 cm^–1, and bridging ether at 1230 cm^–1. In addition, the single band at 1100 cm^–1 is now accompanied by extra bands indicative of the 1,3,4-trisubstituted phenyl ring, confirming the grafting as desired (Figure c). The free PAA spectrum exhibited a broad band at around 2757–3418 cm^–1 associated with acidic OH and alkane CH groups (Figure a). An intense band around 1696 cm^–1 was assigned to the carboxylic stretching of PAA, accompanied by another intense peak due to C–O stretching at 1236 cm^–1. The C–H bending of the PAA was also observed at 1434 cm^–1. , Similarly, these established PAA bands are used for confirmation of the grafting process. For instance, subsequent to grafting PAA onto the PES, carboxylic stretching was observed at around 1722 cm^–1. Furthermore, C–H bending, which increased with increasing AA monomer concentrations, was found around 1044 and 995 cm^–1 in PAA-g-PES spectra. Similarly, the carboxylic band at 1700 cm^–1 shifted slightly to lower frequencies to 1696 cm^–1 (Figure b). The band at 1400 cm^–1 also shifted to higher values, while that at around 1044 cm^–1 shifted in the opposite direction (Figure c). All of the changes observed on modified PES are a confirmation of the successful grafting of PAA onto the PES backbone.

1.

(a) FTIR spectra of pristine PES, PAA, and PAA-g-PES at different monomer concentrations. (b,c) Expanded sections indicating emergent bands with an increasing AA monomer used.

2.1.2. ¹H NMR and ¹³C NMR of PES and PAA-g-PES Polymers

The grafting of PAA with acrylic monomers under a radical initiation process was also interrogated using both ¹H NMR and ¹³C NMR analyses. ¹H NMR was used as a complementary technique to FTIR in that the evolution of the phenyl rings from the 1,4-disubstituted structure to the 1,3,4-trisubstituted structure could clearly be identified (Figure a). The chemical structures and ¹H NMR of pristine PES and PAA-g-PES polymers are illustrated in Figure a’ and b’, respectively. For the pristine PES, two sets of doublets were observed for the two chemical environments, c and f, around 7.28 and 7.80 ppm chemical shifts, as previously reported. On the other hand, the successful grafting of PAA onto the PES polymer backbone should result in the evolution of some phenyl rings from the 1,4-disubstituted to the 1,3,4-trisubstituted forms, leading to the emergence of new peaks in the aromatic region accompanied by bands in the aliphatic region due to the PAA side chain (Figure b’), here shown for the highest grafting polymer (PAA₁@PES₁). From the spectrum, the new peaks observed were the doublets and triplets emanating from the trisubstituted rings, as indicated in the aromatic region and the two types of triplets in the aliphatic region of the NMR spectrum from the attached PAA brushes (Figure b’). These observations clearly confirm the successful grafting of PAA side chains onto the main PES backbone. The relative peak areas for the two types of phenyl rings indicate a low grafting yield, in line with the gravimetric analysis done earlier. The observed peak shifts align with what was observed in the literature. Figure b presents ¹³C NMR of the pristine PES and PAA-g-PES with the highest concentration. ¹³C analysis further supported the successful grafting of the PAA brushed onto the PES polymer backbone. The peaks for the benzene carbons at 119.92, 130, 136.61, and 159.37 ppm were assigned for carbons a, b, c, and d, respectively (Figure b). Post grafting, there was a slight peak shift for the PES ring carbons, which is an indication of the attachment of the PAA onto the polymer backbone. Also, the carboxyl carbon peak (e) of the acrylic acid brushes was observed at around 162.35 ppm.

2.

(a) ¹H NMR spectra of (a’) pristine PES and (b’) PAA-g-PES with the highest concentration. (b)¹³C NMR of (a’) pristine PES and (b’) PAA-g-PES with the highest concentration.

2.1.3. TGA of PES and PAA-g-PES Polymers and Membranes

TGA is a quantitative technique for evaluating the weight loss in samples resulting from an increase in the temperature. The samples were heated at a rate of 10 °C min^–1 from 30 to 800 °C (Figure ). The thermal stability experiments of polymers (PES and PAA-g-PES) (Figure a) were primarily conducted to establish the onset of decomposition or loss of the grafted side chains, which will also be indicative of the grafting yields. Four weight loss stages were observed upon exposing the polymer increasingly to higher temperatures. The first decomposition occurred around 100 °C, and it was mainly attributed to the elimination of moisture from the polymer samples. The second weight loss occurring between 140 and 225 °C was from the decomposition of the PAA side chains. The third and fourth observed decompositions were from 580 to 600 °C, and that around 700 °C was due to the decomposition of PES polymer chains, as reported in a prior article. It must be noted that the onset of the thermal degradation of PAA alone is between 230 and 300 °C with complete decomposition around 500 °C. This clearly indicated that the self-polymerized PAA byproduct was successfully removed from the final product used for membrane fabrication. The results demonstrated the thermal stabilities of PAA-g-PES that are appropriate for water purification and related applications. The increasing mass loss observed with the content of the AA monomer added can be used to estimate the amounts of grafts created on the PES polymer backbone (Table ). The trends here are in agreement with those observed using gravimetric analysis and deduced using eq . The thermal stability experiments of the membranes (A0–A5) (Figure b) are primarily conducted to measure the residual solvent content and stability of the fabricated membranes. The membranes went through multiple decomposition stages, which included first the membrane degradation through the loss of the PAA chain with concomitant weight loss between 132 and 234 °C, followed by that of the main polymer chain decomposition, consistent with the previous studies. ,

3.

Thermal stabilities of (a) pristine PES and PAA-g-PES polymers at different concentrations and (b) fabricated membranes (A0–A5) exhibiting mass loss attributed to the loss of PAA side chains (A) and the degradation of the main PES chain (B).

1. Gravimetric Weight (%) and the Estimated Weight Loss Obtained from the Thermograms of the Nanocomposites.

membrane ID	nanocomposite ratio (PAA:PES)	gravimetric yield (%)	TGA weight loss (%)
A0	0:1
A1	0.3:1	2.11	2.02
A2	0.5:1	2.72	2.66
A3	0.67:1	3.51	3.36
A4	0.83:1	4.43	4.22
A5	1:1	6.76	6.64

2.2. PAA-g-PES Membrane Characterization

2.2.1. FTIR Spectra of Membranes

Figure presents ATR-FTIR spectra of the fabricated membranes from (a) 4000–2000 cm^–1 and (b) 2000–650 cm^–1 wavenumbers. The pristine PES membrane exhibited typical adsorption peaks at 1150 and 1201 cm^–1, which correspond to the symmetric and asymmetric stretching of the sulfonyl groups. The aromatic ether adsorption peak was noted at 1240 cm^–1. Additionally, the aromatic benzene ring stretching adsorption peaks were noted between 1280 and 1580 cm^–1. In addition to transformation of the single to three bands, this indicated the transformation of the 1,4-disubstituted to the 1,3,4-trisubstituted phenyl rings between 1000 and 1200 cm^–1. This is in line with the work reported by Moradi et al. These characteristic peaks were also noted in the PAA-g-PES membranes. The carboxylic stretching of PAA in PAA-g-PES membranes was observed around 1762 cm^–1, and this is in correspondence with work reported by Zhu et al. where they modified the PES through corona-induced graft polymerization. Two emerging peaks were noted at 1451 and 995 cm^–1, while an increase in peak intensity was noted around 1033 cm^–1. Furthermore, an increase in the intensity of peaks at 2853 and 2923 cm^–1 was observed as the PAA grafting yield increased. This was due to the alkane stretching of alkane groups found in the PAA brushes.

4.

FTIR spectra of pristine PES and PAA-g-PES membranes at different PAA concentrations: (a) 4000–2000 cm^–1 and (b) 2000–650 cm^–1.

2.2.2. AFM

The effect of PAA on the PES surface topology was investigated using AFM. AFM images of A0–A5 membranes are shown in Figure with their parameters (R _a and R _q). The AFM images exhibit large valleys and peaks as well as bright regions that are associated with variations in height and areas of greater elevation across the membrane surface. The pristine PES membrane (A0) exhibited a higher surface roughness of 72.4 nm, and the roughness was improved upon the introduction of hydrophilic PAA brushes (A5: 5.9%). This is associated with the hydrophilic nature of the PAA, which quickens the demixing processes of the polymer dope solution, resulting in smoother membrane surfaces. The results obtained for the modified membranes showed that the surface roughness of the membranes decreased significantly from 10.2 to 5.9 nm for the A1 to A5 membranes. It has been observed that membranes with a higher PAA monomer have smoother surfaces. Accordingly, the relative hydrophilic contribution of PAA facilitates the demixing process, leading to smoother surfaces. This is excellent as the membranes with decreased surface roughness reduce the contact between the membrane surface and the foulants, thereby enhancing their antifouling properties.

5.

AFM images of A0–A5 membranes.

2.2.3. WCA

The wettability or hydrophilicity of the fabricated membranes was determined by measuring the water contact angle with the membrane surface-active side. The results of the WCA are listed in Figure . The pristine PES membrane exhibited a WCA of 77°, which reflects the hydrophobic nature of the PES membrane. The membranes (A1–A5) modified with PAA chains exhibited decreased WCA ranging from 67 to 46°; this shows the effect of incorporating water-loving functional groups of PAA onto the PES. The modified membrane’s WCA was affected by the grafting yield. As the monomer concentration of PAA increased, the hydrophilicity enhanced significantly, and this is in line with the literature. ,

6.

Water contact angles of A0–A5 membranes exhibiting a decreasing trend with an increasing PAA content.

2.2.4. Membrane Morphology

It has been established that membrane morphology is mostly influenced by factors, such as the viscosity of the polymer casting solution and polymer concentration. Thus, solution viscosity can influence the kinetics of membrane formation, surface, and cross section of the membrane; this depends on the exchange rate between the solvent and nonsolvent during the demixing process. The fillers used in polymers and the casting temperature can also influence the solution viscosity. In this study, the as-fabricated membranes exhibited a general thin, selective skin layer supported by a porous substructure, as shown in Figure . The wider finger-like pores in the substructures of the modified membranes (A1, 0.024; A5, 0.063 μm) are attributed to the fast solvent and nonsolvent exchange during the membrane fabrication process. /The modified membranes exhibited well-defined and improved finger-like substructures compared to the pristine PES membrane. The same observations were noted for the top and bottom surfaces of the pristine PES membrane; the modified membranes exhibited bigger pore sizes. Upon increasing the grafting yield, the appearance of pores toward the membrane bottom appeared to be relatively larger compared to those seen on the pristine membrane. This was attributed to the quickened demixing in response to the presence of an increased concentration of the PAA hydrophilic functional groups in the casting solutions. These observations are in line with the literature.

7.

SEM micrographs showing the top surface morphology, bottom surface morphology, and cross section of A0–A5 membranes.

2.3. Membrane Performance

2.3.1. Water Uptake and Porosity

The porosity and water uptake (Figure ) of the membranes were conducted to determine the structural effect of the incorporated poly­(acrylic acid). The pore size and thickness of the membrane, which play a role in the porosity and water uptake capabilities, were also determined using ImageJ software. The average thickness of the membranes increased from 114 μm for the pristine PES membrane to 132 μm for the modified membranes (Figure ). The pristine membrane (A0) had smaller pore sizes of 0.024 μm. On the other hand, the pore sizes of the modified membranes increased with an increasing PAA content, ranging from 0.033 μm (A1) to 0.063 μm (A5). It is speculated that the open internal pore structure is an indication of a faster demixing process as a result of higher polymer hydrophilicity. This showed that the PAA side chains grafted onto the PES polymer facilitated increased pore sizes and membrane porosity, leading to adequate water channels, as seen in the SEM micrographs (Figure ). The relatively increasing porous membrane structures with the PAA content resulted in increased water flux, and this is in line with the reported literature. , The calculated membrane porosity increased from 131 to 147%, while water uptake increased from 257 to 302% (Figure ). The observed enhancements of the measured parameters were attributed to the increased hydrophilic character of the grafted PES polymer as a result of the increasing content of carboxylic acid functional groups, in agreement with the literature report.

8.

Membrane thickness, water uptake, porosity, and average pore sizes of fabricated membranes (A0–A5) indicating their increasing trends with the PAA content.

2.3.2. Pure Water Flux and BSA Retention Rate

The effect of the PAA content grafted onto the PES polymer backbone on pure water flux and the BSA retention rate is clearly illustrated in Figure . The PWF of the pristine PES membrane (A0) was 679 L m^–2 h^–1. The PAA-g-PES membranes (A1–A5) showed an increase in PWF as it increased from 740 to 1092 L m^–2 h^–1. The increase in PWF was expected, as the modified membranes showed high surface hydrophilicity (WCA) (Figure ), overall porosity, and wider water channels, as seen in the cross-sectional micrograph (Figure ). The rejection and antifouling capabilities of the fabricated membranes were studied using the well-characterized and frequently used protein probe, BSA, as a model pollutant and foulant. The conformation of BSA is known to be pH-responsive, with the normal form persisting between pH 5 and 8, with an ellipsoid shape of 40–140 Å, and the other forms existing at lower pH (<4) ranges all with an isoelectric point of 4.8. − The BSA retention rate was measured by using UV–vis spectrophotometry using eq . The calculated absorption rate showed an increase in BSA retention from 52% observed for the pristine membrane (A0) to 94% realized for the PAA-g-PES membranes (A3–A5), with A1 and A2 exhibiting slightly lower values of ca. 60 and 70%, respectively. The extent of rejection might be related partly to the mutual negative charges possessed by both BSA and PAA grafts at the neutral pH of the deionized solutions used. The hydrophilic surface character also led to increased attraction of water molecules, resulting in a strong hydration layer that is known to alleviate the adsorption of BSA on the membrane surface. , The results indicated the excellent effect of incorporating PAA grafts on the PES backbone.

9.

Pure water flux and BSA rejection of the fabricated membranes (A0–A5) showing their increasing trends with an increasing PAA content.

2.3.3. Antifouling Performance Studies

The antifouling properties of the fabricated membranes (A0–A5) were studied by measuring pure water flux recovery (J _w2) after fouling the membranes with BSA solution (1000 ppm) at neutral pH and river water. The flux versus time of pure water flux filtration (J _w1), BSA filtration/river water (J _p), and pure water flux (J _w2) after backwashing the membrane with DI water for 10 min are shown in Figure S1a,b. The pure water fluxes of the modified membranes were all higher than those of a pristine PES membrane under the regimes studied. This observation was attributed to the presence of the PAA side chain property and its higher water absorption capability. Thus, as the content of PAA increased on the PES, higher fluxes were obtained. Furthermore, as earlier stated, increased porosity and wider water channels compared to the pristine PES membrane also facilitated higher PWF. During the BSA-laden water and river water filtration, i.e., the fouling process of the membranes (J _p), a decrease in flux was observed for all the membranes, as expected from the formation of a physically deposited BSA and pollutants from the river water layer on the membrane surface that restricted free water passage. The BSA fouling process was significant compared to river water, as shown in Figure S1a,b. The layers were removed through backwashing with water to different degrees depending on the membrane resistance toward BSA or river water pollutant interactions and the grafting yield. As anticipated, the pristine PES membrane showed the lowest filtration of the protein and river water pollutants as well as flux recovery. The relatively higher hydrophobicity and neutral surface of the pristine PES membrane and its rougher surface were thought to be the reasons for its poor performance. The modified membranes exhibited remarkable PWF recovery, especially membranes with a higher PAA content, attributed to the formation of the hydration layer and density of negative surface charges that resulted in relatively higher reversible fouling behavior of the modified membranes toward BSA and river water pollutants.

2.3.4. Relative Fouling Propensity and Recycling Potential of the Membranes

The relative flux recovery ratio (FRR), total fouling ratio (R _t), reversible fouling ratio (R _r), and irreversible fouling ratio (R _ir) of the pristine PES and PAA-g-PES-modified membranes are presented in Figure a,b. These fouling parameters of the membranes were derived from the data of the first deionized water flux (J _w1) versus filtration of BSA solution (J _p) at neutral pH or river water and pure water flux after backwashing the membrane with deionized water (J _w2) cycle. The pristine PES membrane showed the lowest antifouling properties, with an FRR of 37% for BSA filtration. The FRR for the PAA-g-PES membranes ranged from 65 (A1) to 92% (A5). Additionally, the total fouling ratio (R _t) of the pristine membrane was very high (78%), steadily decreasing as the PAA content increased reaching the lowest value of 39% for the membrane with the highest PAA grafting. It was also evident that the rougher membrane (A0) was difficult to backwash with pure water as the initial flux was far from being recovered, resulting in an irreversible fouling ratio of 63%, while that of the membrane containing with the highest PAA was only 8%. The same observations were noted for river water filtration; the pristine PES membrane exhibited the highest total fouling ratio (60.3%) and FRR (49.2%), and this was expected since the pristine PES membrane has a rougher surface. However, after modification of the membrane surfaces, excellent antifouling resistance was noted. The membrane with the highest PAA concentration exhibited a 97.3% FRR and an R _t of 33.7%. The river water fouled the membranes less than BSA-laden water. The observed low antifouling property of the pristine PES membrane was attributed to its rougher surface (Figure ) and hydrophobic nature (Figure ). The improvement shown by the PAA-grafted membranes was attributed to the influence of the PAA polymer chain content; that is, increasing the PAA content led to a higher FRR. The decrease in fouling propensity of the PAA-g-PES membranes indicated that these foulants deposited on the membrane surfaces can be easily backwashed off and the water flux closer to the initial PWF regained.

10.

Fouling resistance parameters of the fabricated membranes (A0–A5) indicating the positive effect of an increasing grafted PAA content for (a) BSA and (b) river water.

Seven fouling–backwashing cycles (Figure ) for BSA and river water configurations were used for the determination of the reusability of the fabricated membranes over time. All of the membranes showed a decrease in flux recovery percentages due to the accumulation of pollutants on the membrane surfaces and pores, with the pristine PES (A0) membranes being the worst. The fabricated membranes showed better resistance toward river water compared to the BSA solution owing to the high concentrations (1000 ppm) of BSA pollutants. The membrane with the highest AA concentration (A5) exhibited excellent fouling resistance and reusability. The PAA-grafted membranes depicted a prolonged lifespan with an excellent filtration ability over seven cycles. Membranes with enhanced hydrophilicity and reduced surface roughness are excellent at preventing the adhesion of pollutants on the surface, thereby prolonging the membrane lifespan. These observations are in line with the prior literature. ,

11.

Antifouling properties of membranes at seven cycles of (a) BSA and (b) river water.

2.3.5. Surface Characterization of the Fouled Membranes

Membrane fouling is a major challenge in membrane separation processes. Fouling happens when dispersed or suspended solids are deposited onto the membrane surface, forming a cake layer attached at the pore opening and within the pore, clogging the pore, resulting in relatively poor membrane performance and shortening their life span. Figures a and b show the FTIR spectra of the membranes after 30 min of BSA filtration and after backwashing with deionized water, respectively. For BSA filtration (Figure a), the fouled membranes presented new peaks at around 3287, 1654, and 1575 cm^–1, which are attributed to the presence of deposited BSA on the membrane surface. The vibrational mode around 3287 cm^–1 was assigned to N–H and O–H functional groups. The band at 1654 cm^–1 was assigned to the stretching vibration of CO in the amide I groups in the BSA protein, while the band at 1575 cm^–1 was observed as a response to the strong primary amine scissoring. , The intensity of both bands at 1654 and 1575 cm^–1 decreased as the amount of grafted PAA increased, finally disappearing for membranes with a high PAA content. This suggests less attachment/fouling of the BSA protein on the membrane surface in membranes with a higher PAA content. Further bands that showed increasing intensity were found around 2905–2990 cm^–1, which are typical of the underlying PAA-g-PES membrane as less BSA was bound to the surface. After the fouled membranes were backwashed with deionized water (Figure b), all of the intense peaks associated with BSA (at 3287, 1654, and 1575 cm^–1) seen on the fouled membranes disappeared. This was an indication that backwashing with deionized water was effective and that reversible fouling was higher for the fabricated membranes without adjusting pH or adding a harsher backwashing liquid. Higher fouling reversibility can be attributed to the presence of the grafted PAA brushes on the PES chain, which are believed to form a tighter hydration layer on the membrane surface that ensures less or delayed attachment of BSA protein to the membrane surface. In addition, the increased peak intensity of the stretching vibration of CO in the amide I groups in the BSA protein noted at 1654 cm^–1 on the fouled membranes decreased significantly after backwashing, indicating that it was effectively removed. The river water filtration (Figure S2a,b), on the other hand, did not alter the functional groups of the membranes; this was due to less attachment of pollutants on the membrane surface.

12.

FTIR spectra of (a) membranes fouled with BSA solution at (i) 4000–650 cm^–1 and (ii) 2000–650 cm^–1 and (b) membranes after backwashing with deionized water, (i) 4000–650 cm^–1 and (ii) 2000–650 cm^–1.

The SEM images of the BSA-fouled membranes are presented in Figure S3a,b. These micrographs were dominated by a cake layer and specks with no previously porous surface, as observed in Figure of the clean membranes. The fouled PES (A0) membrane presented a surface dominated by the highest number of foreign particles and no traces of a porous surface reminiscent of the original membrane surface, probably due to a denser cake layer. The membranes containing grafted PAA presented surfaces with decreasing foreign matter as well as the reappearance of a porous sublayer for membranes containing a higher PAA content (A4 and A5). These observations indicated that the thickness of the cake layer remaining on the surface was decreasing. In fact, fouled A5, and to some extent A4, revealed an increasing porous appearance, indicating the presence of little or no cake layer formation. This was in agreement with the surface FTIR results in Figure a, which indicate decreasing BSA-associated bands and increasing bands of the underlying membranes. The SEM micrographs after backwashing with deionized water are shown in Figure S3b. The pristine PES (A0), as well as the A1 and A2 membranes, exhibited fewer foreign particles and no visible pores on the surface compared to the fouled membranes, indicating less efficient removal of BSA and hence presenting relatively higher irreversible fouling properties than membranes with higher PAA contents (A3, A4, and A5). In fact, no foreign particles were observed in the fouled membranes with a higher PAA content, while the porous features were also becoming apparent. Figure S3b further indicates that the cake layer was effectively removed with backwashing with deionized water for membranes (A3, A4, and A5) with higher PAA contents. These observations from SEM analysis of fouled membranes (A0, A1, and A2) are therefore in agreement with those found from surface FTIR studies (Figure b), where the BSA-based functional group bands disappeared. The cake layer that formed was easily reversed with an increasing PAA content, showing the importance of incorporating the hydrophilic chains of PES to increase the antifouling properties of the membranes.

3. Conclusions

Direct grafting of the AA monomer onto the PES polymer backbone via free radical graft polymerization was successfully achieved and confirmed using ¹H NMR, ¹³C NMR, and FTIR analysis. Furthermore, gravimetric methods were also used to compliment the spectrochemical techniques to confirm the increasing grafting as designed. The membranes fabricated from the resultant PAA-g-PES polymer through the phase inversion method also showed the presence of PAA on the membrane surfaces using FTIR. The WCA and AFM results confirm that the water-loving carboxylic acid functional group of PAA enhanced the surface hydrophilicity and led to relatively smooth surfaces. Additionally, the morphological aspects of the membrane improved as the modified membranes exhibited increasing surface pore sizes, increasing membrane porosity, and wider subsurface water channels. These properties facilitated higher water fluxes, BSA fouling resistance, and BSA rejection, leading to a higher-quality effluent. This hydrophilic character of the PAA-g-PES polymer is thought to have led to fast solvent and nonsolvent exchange during the membrane fabrication process, which also resulted in hydrophilic membrane surfaces that tend to attract water molecules. The last attribute resulted in higher fluxes and fouling resistance for the fabricated membranes reported herein. The fouling layer and its removal using deionized water were confirmed using both SEM and FTIR, which indicated its successful BSA and river water removal indicative of the high fouling resistance of the membranes toward BSA promising an extended membrane lifespan. The membranes are showing potential treatment for nonpotable standards for reuse purposes such as landscape irrigation, toilet flushing, and construction.

4. Materials and Methods

4.1. Materials

All chemicals and reagents were of analytical grade and used without further purification. Benzoyl peroxide (BPO), acrylic acid (AA), bovine serum albumin (BSA), N-methyl-2-pyrrolidone (NMP, 99.7%), and PES powder (M _w = 58,000–61,000 g mol^–1) were sourced from Sigma-Aldrich/Merck. Freshly deionized (DI) water was used in the grafting experiments.

4.2. Grafting of AA onto the PES Polymer Backbone

The free radical graft polymerization method was used to graft AA monomers onto the PES polymer backbone (Scheme ). The PES powder (12 g, average 0.207 mmol, average of 389 repeat units) was dispersed in a three-neck round-bottom flask containing 200 mL of DI water. The mixture was heated in an oil bath, and nitrogen gas was used to degas the mixture before adding BPO. Different amounts (4 g, 56 mmol to 12 g, 166.5 mmol) of AA were slowly added to the PES mixture. The reaction was left to run for 3 h under nitrogen at 70 °C. The resultant PAA-g-PES powder was filtered and washed with DI water to remove the unreacted AA and self-polymerized poly­(acrylic acid) oligomers and polymers. The beige powder was oven-dried at 50 °C for 12 h. The gravimetric grafting yield (GY) was calculated using eq :

$$\mathrm{GY}(\%)=\frac{W_{\mathrm{f}}-W_{\mathrm{i}}}{W_{\mathrm{i}}}\times 100\%$$ (1)

where W _f is the grafted PES powder weight after grafting and W _i is the initial PES weight.

1. Grafting of AA onto the PES Polymer and the Formation of PAA-g-PES.

4.3. PAA-g-PES Membrane Fabrication

Membranes were fabricated using the nonsolvent-induced phase separation (NIPS) method. The casting mixtures containing PES or various PAA-g-PES polymers of different grafting densities dissolved in an NMP solvent over 12 h were prepared (Table S1). The polymer solutions were stored in an airtight, evacuated container for 24 h to allow the dissipation of trapped gases. The flat-sheet membranes were prepared by casting the polymer solution on a glass plate using a casting knife (set at 200 μm). The glass plate was then immersed in a coagulation bath of DI water at RT (25 ± 0.2 °C) after standing for about 30 s in air. The formed membrane was detached from the glass plate, rinsed several times with DI water to remove any remaining solvent, and stored in DI water at 4 °C until it was used.

4.4. Characterization

4.4.1. Characterization of PAA-g-PES Polymers

Fourier transform infrared spectroscopy (FTIR) was utilized to confirm the functional groups (FGs) present in PAA-g-PES polymers and membranes. The FTIR spectra of the materials were obtained with a PerkinElmer Spectrum 100 FTIR (Bruker, Karlsruhe, Germany) spectrometer in the 650–4000 cm^–1 range. Thermogravimetric analysis (TGA) was utilized to determine the thermal stability of the nanofillers and membranes. TGA was conducted using a PerkinElmer STA6000 (Waltham, MA, USA) thermogravimetric analyzer. Approximately 25 mg of a sample was heated (20–800 °C, 10 °C min^–1) under a N₂ atmosphere (20 mL min^–1), and then, the weight loss was recorded. The compositions of the pristine and modified PES powder were studied by ¹H NMR and ¹³C NMR spectra recorded on a Bruker 400 MHz Avance NMR spectrometer. The samples were dissolved in a DMSO-d ₆ solvent to obtain a homogeneous solution.

4.4.2. Characterization of the Surface of the PAA-g-PES Membrane

The surface area and cross section of the membranes were investigated by SEM (TESCAN VEGA 3, Czech Republic) set at a 20 kV acceleration voltage. Prior to the cross-sectional analysis, the membranes were fractured in liquid nitrogen to get a clean cut. For both the surface and cross-sectional analyses, the membranes were coated with gold before analysis. The surface roughness of the membranes was measured using atomic force microscopy (AFM, Nanoscale IV, Veeco, USA) with the spring constant of 0.12 N m^–1 through the contact mode in dry air. The membranes were dried at RT for 2 days before analysis. The water contact angle was used to determine the surface hydrophilicity of the membranes. The tests were carried out at RT on the surface energy evaluation system DataPhysics optical contact angle (OCA) 15 EC (G10, KRUSS, Hamburg, Germany) equipped with video capture. Each dry membrane sample was fixed on a glass plate with double-sided tape. Thereafter, a maximum of 10 water droplets were carefully dosed on the membrane sample, and the contact angle between the membrane and water droplet was measured after 30 s. The average of five measurements was selected as the final contact angle value. The porosity and water uptake were established by weighing the mass of dry and wet membranes according to established protocols. , The membranes were dried overnight and cut using a membrane support ring, and then, their mass (dry weight, W _d) was taken. For wet weight (W _w), the cut membranes were immersed in DI water for 24 h, the excess water was removed from the surface, and thereafter, the membranes were weighed. The water uptake and porosity percentages were determined by eqs and , respectively:

$$\mathrm{water}\mathrm{uptake}=\frac{W_{\mathrm{w}}-W_{\mathrm{d}}}{W_{\mathrm{d}}}\times 100$$ (2)

$$\epsilon =\frac{W_{\mathrm{w}}-W_{\mathrm{d}}}{{\rho }_{\mathrm{w}}\times A\times \delta }\times 100$$ (3)

where ρ_w is the density of DI water (0.998 g mL^–1 at 20 °C), δ is the thickness of the membrane measured on the cross section using ImageJ software, and A is the effective area of the membrane.

4.5. Membrane Filtration Performance Test

The flux performance of the fabricated membranes was assessed through a pure water flux (PWF) experiment studied in a dead-end cell filtration cell with DI water. Before the filtration process, the membrane was compacted for 15 min at 300 kPa to obtain a constant flux. Subsequently, the filtration of pure water was done at 200 kPa for 30 min, and the volume was measured. The PWF was calculated according to Darcy's law (eq ), expressed as follows:

$$J_{\mathrm{flux}}=\frac{Q}{t\times A}$$ (4)

where Q is the volume of the permeate (L), t is the permeation time (h), and A is the effective area of the membrane (0.00126 m²).

The rejection ability of the membranes was evaluated by filtration of 1 g L^–1 BSA-laden water under an operation pressure of 150 kPa, and the permeate was collected in a beaker for 30 min. The BSA rejection was determined by measuring the BSA absorbance of the permeate using a UV–vis spectrophotometer. The rejection of BSA solutes was calculated using eq :

$$\mathrm{removal}\mathrm{efficiency}=\frac{C_{\mathrm{o}}-C}{C_{\mathrm{o}}}\times 100$$ (5)

where C _o is the influent concentration (mg L^–1) of a given pollutant and C is the corresponding effluent concentration (mg L^–1).

The membrane antifouling properties were determined using the given BSA and river water. First, J _w1 was investigated by passing pure water through the membranes for 30 min. Then, J _p was determined by passing BSA or river water through the membranes for 30 min. The membrane was then backwashed with DI water for approximately 10 min at 200 kPa followed by passing pure water for another 30 min to determine the J _w2. In the backwashing process, the accumulation of foulants on the top surface of the membranes was removed by a controlled reversal flow of pure water to maintain the performance and prolong the lifespan of the membranes. The membrane antifouling properties were assessed by determining the flux recovery ratio (FRR), total fouling (R _t), reversible fouling ratio (R _r), and irreversible fouling ratio (R _ir), using eqs , , , and , respectively, ,

$$\mathrm{FRR}=\frac{J_{\mathrm{w}2}}{J_{\mathrm{w}1}}\times 100$$ (6)

$$R_{\mathrm{t}}=\left(1-\frac{J_{\mathrm{p}}}{J_{\mathrm{w}1}}\right)\times 100$$ (7)

$$R_{\mathrm{r}}=\left(\frac{J_{\mathrm{w}2}-J_{\mathrm{p}}}{J_{\mathrm{w}1}}\right)\times 100$$ (8)

$$R_{\mathrm{ir}}=\left(\frac{J_{\mathrm{w}1}-J_{\mathrm{w}2}}{J_{\mathrm{w}1}}\right)\times 100=R_{\mathrm{t}}-R_{\mathrm{r}}$$ (9)

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.4c05443.

• Antifouling performance, SEM of fouled and cleaned membranes, FTIR of river water fouled membranes, and casting solution composition (PDF)

Funeka Matebese: conceptualization, methodologies, visualization, formal analysis, investigation, and writing of the original draft. Mabore J. Raseala: formal analysis, investigation, and review and editing. Meladi L. Motloutsi: formal analysis, investigation, and review and editing. Richard M. Moutloali: conceptualization, methodologies, visualization, validation, resources, review and editing, and supervision.

The authors declare no competing financial interest.