Introduction

Humic acid (HA) with carbonyl, quinone, carboxyl, and hydroxyl functional groups in aromatic and organic ring frameworks, as one of the main components of organic substances soluble in natural waters, plays an influential role in improving soil fertility. However, its presence in the environment due to the formation of HA-metal complex ions leads to the displacement of the metal during diffusion in the soil (Nazri et al. 2021; Teow et al. 2017). Also, the conventional treatment of water contaminated with HA causes chlorine to interact with HA, producing a series of carcinogenic substances for humans. Therefore, it is necessary to remove HA before the chlorination process in wastewater treatment plants. Among several purification methods, including coagulation, electrocoagulation, flotation, oxidation, and membrane separation, membrane separation acts as an innovative technology to separate water from pollutants such as water/oil mixture, microorganisms, proteins, and humic substances (Algamdi et al. 2019; Kumar et al. 2016), and advantages such as compact design, easy operation, environmental friendliness, low energy consumption, no need for chemicals, selectivity, and operation at room temperature have increased this process in recent decades (He et al. 2020; Li et al. 2020; Tajik et al. 2024). However, fouling and, consequently, the high operating cost of the membrane is a subject that deserves discussion, and much research has been done on it. In fact, when the membrane surface is contaminated, not only does it reduce the flux, but it also increases the operation and maintenance costs (Kusworo et al. 2020; Tajik et al. 2023). Improving the characteristics of the membrane surface with a hydrophilic modifier (including polymer materials and inorganic nanoparticles) can significantly increase the antifouling performance of the membrane. Hydrophilic polymers such as polyvinyl alcohol, polyethylene glycol (PEG) derivatives, and zwitterionic polymers have been used to increase membrane flux and reduce membrane fouling (Luo et al. 2020).

Polyvinyl fluoride (PVDF) membrane is known as the most widely used material among all existing polymeric membranes due to its outstanding properties. Nevertheless, PVDF polymer has a profoundly high fouling due to its inherent hydrophobic property. Therefore, various techniques have been reported to make the membrane surface hydrophilic, including mixing hydrophilic polymers or embedding inorganic nanoparticles, chemical modification, and stabilization of Polymers with hydrophilic moieties on the surface of membranes by coating (Makhetha and Moutloali 2018; Mat Nawi, et al. 2020; Yong et al. 2019).

Tannic acid (TA) is a low-cost, safe, and environmentally friendly polyphenol extracted from plants, such as tea leaves, oak wood, nettle, and Chinese gall. HA has anti-carcinogenic, antimutagenic, and antioxidant properties with the ability to form macromolecules and metal ions and is widely used as a coagulant or polymeric coagulant for water and cationic dyes purification (Oulad et al. 2020; Xu et al. 2018). PEG is also a polyether with two functional groups, hydroxyl, and ether, which consists of repeated units of ethylene glycol [–(CH2CH2O)n] and is known as macrogol as well (D’souza and Shegokar 2016).

Much research has been done on PVDF polymeric membranes integrated with nanoparticles such as TA, graphene oxide (GO), and titanium dioxide (TiO2) to remove various pollutants; some examples are mentioned below. In Zhao et al. (2019) studies, a PVDF membrane was coated with graphene oxide nanoparticles, which caused water permeability to be reduced from 1845 to 375 L/m2.h, hydrophilicity, and negative surface area to be increased from − 23.4 to − 46.8. (The water contact angle was reduced from 73° to 56°.) In the research of Li et al. (2019), PVDF membrane coating was done by TA nanoparticles and polyvinyl pyrrolidone additive, which increased pure water flux (> 16,000 L/m2.h), antifouling properties (FRR > 99.5% and emulsion rejection > 99%) and hydrophilicity (54°) in the modified membrane. In the studies of Teow et al. (2020), the PVDF X500 TiO2 membrane was presented for the removal of HA due to its smooth surface and reactive surface layer, good antifouling properties (FRR = 78.24%), and initial water flux of 58.81 L/m2.h. In the research of Jiang et al. (2021), with a single-step coating of HA/sodium periodate solution on the PVDF hydrophobic membrane, increased permeability flux (> 2400 LMH/bar) and oil removal > 98% for the hydrophilic and underwater super oleophobic modified PVDF membrane (water contact angle reduction from 121 to 32) was observed. Sun et al. (2021) coated dopamine on PVDF/TiO2 membrane and observed high removal efficiency (99%) and permeate flux (1389 L/m2.h) for oily wastewater. According to the research of Karimi et al. (2021), the synthesis of PVDF membrane and its modification by Cu2S 0.2wt% to remove reactive blue dye was obtained, which had a high water flux of 248.25 L/m2.h and FRR = 92.4%. Yan et al. (2022) coated the hydrogel on the PVDF membrane surface, which reduced the oil pollutant fouling on the membrane surface. Also, pure water flux increased by 214%, permeability increased by 127%, and oil recovery was observed by 99.6% (Yan et al. 2022). Recently, Ren et al. (2022), by surface modification of PVDF membrane with propyl sulfated chitosan, witnessed high hydrophilicity (39°) and antifouling properties (FRR = 90.4%) after two cycles. In the studies of Abdulazeez et al. (2023), a PVDF membrane was incorporated with iron in order to remove oil. It indeed increased the pure water permeance from 59.9 L/m2.h in the pristine membrane to 300.5 L/m2.h, and oil rejection was more than 99% (Abdulazeez et al. 2023).

In this research, the surface modification of the PVDF pristine membrane was done by coating TA nanoparticles and PEG additives to remove HA. TA nanoparticles and PEG additives are both environmentally friendly and nontoxic, able to form hydrogen bonds. In this vein, the –OH present in TA nanoparticles communicate with the –O– present in PEG additives and form a hydrogen bond, which, by coating the membrane surface, improves the antifouling properties of the membrane, hydrophilicity, and permeability flux of pure water. Investigated parameters in the test include operating pressure (0.5, 1, 1.5, 1.8, and 2 bar), pollutant concentration (30, 50, 80, and 100 ppm), PEG: TA ratio (0:0, 1:0, 0:1, 4:1, and 1:4), and pH (5, 7, and 9) were optimized. Also, the separation performance and removal efficiency, membrane properties, morphology, roughness, hydrophilicity, and membrane antifouling properties were investigated in detail.

Experimental

Materials

HA was purchased from Sigma-Aldrich to prepare synthetic wastewater. PVDF polymer and dimethylformamide (DMF) solvent for membrane synthesis and TA nanoparticles, PEG 4000, DMAc solvent, and ethanol for membrane surface modification were obtained from Merck. HCl and NaOH, obtained from Merck, were used for pH adjustment. Deionized water was used in all stages of the research.

Synthesis of PVDF polymeric membrane

This part was presented in SI, Sect. "Introduction". Part 1.

PVDF UF membrane modification

This part was presented in SI, Sect. "Introduction". Part 2.

Membrane characterization

Fourier transform infrared spectroscopy (FTIR-ATR) (PerkinElmer Frontier model) was performed to identify the functional groups on the surface of the membranes. Membrane morphology and roughness were respectively identified by scanning electron microscope TESCAN model manufactured by Mira 3-XMU company and atomic force microscope manufactured by NT-MDT company and model TS-150. EDS spectroscopic analysis was used by scanning electron microscope TESCAN model manufactured by Mira 3-XMU company to observe the elements on the surface of the membrane. The water contact angle test (manufactured by CAG-20 company, Jikan model) was performed to determine the hydrophilicity of the membrane surface. As a matter of fact, for its measurement, the membrane was placed on a smooth surface. Water drops with a volume of approximately ten microliters was released on the membrane’s surface, and photos were taken after reaching the equilibrium state. For each sample, the average contact angle at different points was considered. A dynamic light scattering device manufactured by Horiba Jobin Jyovin company, the SZ-100z model, was used to measure the particle size and zeta potential of HA particles. The porosity and average diameter of membrane pores were obtained to identify the effect of additives and nanoparticles on membrane morphology and permeability.

To obtain the porosity of the membranes, the wet and dry weight of a specific area of each membrane after 24 h of staying in deionized water and oven at 60 °C, respectively, were measured and obtained according to Eq. (1) (Ni et al. 2021). Due to the small number of nanoparticles, the insignificant effect of its density was ignored in the calculation of the membrane density, and only the density of water was considered.

$$\varepsilon =\frac{{m}_{1}-{m}_{2}}{Al{\rho }_{\text{w}}}$$
(1)

where \({m}_{1}\) and \({m}_{2}\) are, respectively, the wet and dry mass of the membrane (gr), A the effective surface of the membrane (cm2), l the thickness of the membrane (cm), \({\rho }_{\text{w}}\) the water density (gr/cm3), and ε is the porosity of the membrane.

According to Eq. (2), Gorot Leford-Ferry equation, the average pore diameter was estimated (Sakarkar et al. 2020).

$$ D_{m} = \sqrt {\frac{{\left( {2/9 - 1/75\varepsilon } \right) \times 32\eta lQ}}{\varepsilon AP}} $$
(2)

where η is the water viscosity (8.9*10−4 Pa.s), Q is the water flow rate (m3/s), P is the applied pressure (Pa), and Dm is the average diameter of the pore.

Preparation of synthetic wastewater and determination of effective parameters

First, 1 gr of HA pollutant (C9H9NO6) was dried at 105 °C for 2 h to prepare synthetic wastewater. Then, it was dissolved in 100 ml of 0.1 M NaOH, made up to volume with deionized water, and placed on the stirrer for 4 h. Then, it was placed in an ultrasonic bath for better homogenization, and finally, pH = 7 was reached with 4 M HCl. From the 1000 ppm stock solution, the desired concentrations of synthetic HA solution were prepared by dissolving the stock solution in specific amounts of deionized water (Kumar et al. 2016).

In order to find the optimal mode, the performance of HA separation from wastewater by the pristine and modified membranes was checked using the OFAT method with three repetitions. In this regard, at first, according to the permeate flux and pollutant removal percentage, the optimal initial pressure and concentration were respectively obtained for the M0 membrane. Based on that, these parameters were measured for all membranes. Finally, the optimum pH for the final selected membrane was investigated. Table 1 shows the parameters examined in this research and their ranges. The range of parameters was selected based on the references of Singh and Purkait (2016); Shi et al. (2018).

Table 1 Investigated parameters
Full size table

Filtration test

The filtration test was carried out in the dead-end laboratory pilot, according to Fig. S1, with an effective separation area of 11.34 cm2 and an effective volume of 250 ccs. The required operating pressure was applied by a nitrogen capsule.

The mass of pure water and HA was measured to obtain the membrane flux. The filtration test was continued until the flux reached at least 50% of the initial flux. The total duration of the test was 135 min, in which pure water was passed in the first 30 min and the last 15 min. The output mass of HA was measured every 3 min in the first 15 min and then every 15 min until it reached a stable value in about 90 min. To ensure the accuracy of the data, the information on each membrane was obtained after three tests, and the flux of the tests performed was calculated according to Eq. (3) (Mohd Yatim and Boon Seng 2019).

$$J=\frac{M}{t.A.\rho }$$
(3)

where M is the mass of wastewater (g), t is the duration of the solution penetration (h), A is the active area of the membrane (m2), ρ is the density of water (g/L), and J is the flux of wastewater (L/h.m2).

The pollutant removal rate in the samples was calculated using Eq. (4) (Xu et al. 2020).

$$R\%=\left(1-\frac{{C}_{p}}{{C}_{f}}\right)\times 100$$
(4)

where Cp and Cf are the HA concentration of the permeate and feed solutions, respectively, and R is the percentage of pollutant removal. (The absorbance of the solution was read by spectrophotometer at λmax = 254 nm.)

Fouling and resistances measurements of membranes

The pure water flux was measured before and after the HA flux measurement to determine the antifouling ability of the membranes. FRR(%) is membrane recovery ratio, Rr (%) is reversible resistance, Rir (%) is irreversible resistance, and Rt (%) is the total resistance of the membrane, which are obtained from relations (5), (6), (7), and (8) were respectively calculated. RFR(%) is also the sum of reversible and irreversible resistances (Khoo et al. 2022; Fan et al. 2021).

$$\text{FRR}(\%)=[\frac{{J}_{\text{w}2}}{{J}_{\text{w}1}}]\times 100$$
(5)
$${R}_{\text{r}}(\%)=[\frac{{J}_{\text{w}2}-{J}_{p}}{{J}_{\text{w}1}}]\times 100$$
(6)
$${R}_{\text{ir}}(\%)=[\frac{{J}_{\text{w}1}-{J}_{\text{w}2}}{{J}_{\text{w}1}}]\times 100$$
(7)
$$ R_{{\text{t}}} \left( \% \right) = \left[ {\frac{{J_{{{\text{w}}1}} - J_{p} }}{{J_{{{\text{w}}1}} }}} \right] \times 100 = R_{{\text{r}}} + R_{{{\text{ir}}}} = {\text{RFR}}\left( \% \right) $$
(8)

In Eqs. (5–8), Jw1 is the pure water flux in the first min of the test before the effluent filtration test, Jw2 is the pure water flux in the final min of the test after the effluent filtration test, and Jp is the effluent flux.

Results and discussion

Membrane separation performance

This section examines the membrane that is most optimized in terms of pressure, pollutant concentration, PEG:TA ratios, and pH. Additionally, the investigation focused on water contact angle, membrane resistance abilities, and membrane porosity to further ensure the optimal membrane.

Operating pressure optimization

At first, the pure water flux was measured by passing the pure water before and after the wastewater passed through the membrane in order to investigate the operating pressure for the optimal M0 membrane, the fouling, and membrane resistances (Fig. 1). As can be seen in Fig. 1 a, b, the amount of flux decreased slightly at pressures of 0.5 and 1 bar with the passage of time; however, with increasing pressure at pressures of 1.5, 1.8, and 2 times, the flux reduction has been significantly reduced in the first 15 min of the experiment. In fact, in the first min, the sharp and sudden decrease in the observed flux was due to membrane fouling and pollutant absorption on the clean surface of the membrane. Interestingly, among the particles that passed through the membrane pores, some of which had diameters almost equal to the membrane pores, were absorbed by the membrane wall, which resulted in a remarkable decrease in the flux in the first min. As it is evident, the increase in pressure is associated with the increase in flux. However, due to the higher pressure and also the rate of permeability, there is no room for pollutant absorption on the surface of the membrane, and a dense cake layer is formed on the surface of the membrane. As a result, the chance of membrane rejection and, subsequently, the removal percentage has decreased with increasing pressure. Alpatova et al. (2015) also mentioned similar results.

Fig. 1
figure 1

a Flux-time, b HA Removal (%), c General flux, d Jw1, Jw2, and Jp Flux changes passing through the M0 membrane at different pressures during 90 min ((Nazri et al. 2021) = 50 ppm, pH = 7)

Full size image

Figure 1 c, d, respectively, shows the total flux passing through the membrane in terms of time and the values of Jw1, Jw2, and Jp passing through the M0 membrane at different pressures. According to these two forms, as expected, the pure water flux also increased with the increase of the applied pressure. This increase in flux was low at low pressures and more tangible at pressures higher than 1.5 bar, as in the research of Singh and Purkait (2016), similar results were presented. According to the results, the pressure of 1.5 bar, in which the increase in flux (117.38 L/m2.h) and adequate rejection (77.63%) was observed, was chosen as the optimal pressure to continue the research.

Pollutant concentration optimization

After choosing the optimal operating pressure, the pollutant concentration was optimized for the M0 membrane (Fig. 2). As can be seen in Fig. 2 a, b, in all concentrations, both the amount of flux and the percentage of removal have decreased over time, and also, the sudden drop of flux has happened in the first 15 min of pollutant passage. According to what was observed, increasing the concentration of HA means forming a thicker cake layer on the membrane, which reduces the flux. On the other hand, in high pollutant concentrations, the formation of the HA cake layer on the membrane surface was faster, which reduced the membrane absorption ability. It, in turn, led to the removal percentage decrease, which was also mentioned in the research of Singh and Purkait (2016). According to the investigation of the membrane flux and the pollutant removal percentage at different concentrations, the permeate flux and membrane rejection decreased slightly at higher pollutant concentrations.

Fig. 2
figure 2

a Flux-time, b HA Removal (%), c General flux, d Jw1, Jw2, and Jp Flux changes passing through the M0 membrane at different concentrations of HA during 90 min (P = 1.5 bar, pH = 7)

Full size image

Figure 2 c, d, respectively, shows the total flux passing through the membrane in terms of time and the values of Jw1, Jw2, and Jp passing through the M0 membrane at different pollutant concentrations. As can be seen that with the increase in pollutant concentration (from 30 to 80 ppm), the total flux (from 153.56 to 58.74 L/m2.h), and the final net water flux (from 227.72 to 127.74 L/m2.h) has decreased due to the increased membrane surface fouling; although, the amount of initial pure water flow before passing the pollutant with different concentrations was the same due to the similarity of the conditions. According to the results, to continue the research, the concentration of 80 ppm with the permeate flux of 89.39 L/m2.h and HA removal of 75.59% was chosen as the optimal concentration.

PEG: TA ratio optimization

In order to select the optimal membrane, flux, removal efficiency, water contact angle, membrane properties, porosity and pore size, and antifouling properties were investigated for all modified membranes at 1.5 bar pressure and 80 ppm pollutant concentration. As seen in Fig. 3 a, with the increase in PEG additives percentage, the membrane’s permeability increased due to the hydrophilicity of the membrane surface by PEG hydroxyl bonds, although this trend has not had a significant impact on rejection. On the other hand, with an increase in the percentage of TA nanoparticles, membrane rejection has also increased. However, it did not significantly improve the membrane permeability, which was probably due to the decrease in the number and size of the membrane pores. In other words, increasing the concentration of nanoparticles has decreased the porosity of the membrane and provided a dense layer on the TA coated membrane’s surface, leading to higher pollutant removal. Despite this phenomenon, a decrease in the permeate flux of the mentioned membrane has occurred (Li et al. 2019).

Fig. 3
figure 3

a Separation performance, b Jw1, Jw2, and Jp Flux changes passing through all kinds of membranes during 90 min (P = 1.5 bar, (Nazri et al. 2021) = 80 ppm, pH = 7)

Full size image

All the membranes had relatively high rejection and approximately good permeability. As observed, the M1 membrane, covered only with PEG additives, had the highest permeability (131.21 L/m2.h), and the M2 membrane, which was covered only with TA nanoparticles, had the highest rejection (88.38%). Therefore, the optimal membrane is the membrane that has the least fouling and the highest removal percentage. Among the membranes M3, M4, and M5, which included nanoparticles and additives, the M3 membrane, although it had a high flux (127.62 L/m2.h), had a lower rejection (81.49%) than the two. Among M4 and M5 membranes, the M4 membrane had higher permeability (72.43 L/m2.h) and rejection (86.62%).

Figure 3 b shows Jw1, Jw2, and Jp values for modified membranes. According to this Figure, the membranes modified by nanoparticles and additives have higher pure water fluxes than the pristine M0 membrane, due to the increased hydrophilicity of the membranes modified by nanoparticles and additives. As can be seen, the M1 membrane had the highest pure water flux (1303.65 L/m2.h). The M1 membrane was modified by PEG additives, which contained a hydroxyl bond. The membrane’s high hydrophilic properties increased in pure water flux in this vein. The M5 membrane, which included PEG additives and TA nanoparticles, had a high hydrophilic effect (735.24 L/m2.h) compared to other membranes. The relatively low flux of the pristine membrane was due to its superhydrophobic property; these results were also reported in the research of Liu et al. (2021). Furthermore, the flux drop in the last 15 min (313.69 L/m2.h) compared to the first 30 min of pure water flux (463.75 L/m2.h) indicates membrane fouling; indeed, a decrease in the flux drop in the M3 membrane indicated a lower level of fouling in this membrane in comparison to other membranes.

Water contact angle

The water contact angle was measured to determine the hydrophilicity of the membrane surface. According to Fig. 4, the contact angle of the water droplet with the membrane surface decreased by modifying the membrane surface, which can be attributed to the presence of hydrophilic hydroxyl (OH) and carboxyl (COOH) groups present in TA nanoparticles and PEG additives. The hydrophilicity of the membrane also meant a decrease in the membrane surface fouling and an increase in flux (for example, M3 membrane), indicating the presence of nanoparticles on the membrane surface. On the other hand, the nanoparticles that were equal to or smaller than the membrane pores caused the walls of these pores to close, which led to a decrease in the membrane flux (for example, M4 membrane). Therefore, by coating the membranes with the nanoparticles and creating a surface layer, although the membrane surface became hydrophilic, no continuous trend was observed in the membrane flux. Khezraqa et al. (2022) also reported similar results.

Fig. 4
figure 4

Contact angle of the mentioned membranes

Full size image
Antifouling characterization

The permeability fluxes of HA and pure water were measured to check the antifouling characteristics of the membranes. According to Fig. 5 a, FRR and Rr for all modified membranes were higher than the original pristine membrane, and Rir was the opposite. Figure 5 b also shows the total membrane resistances. The M2 membrane coated with TA nanoparticles had high hydrophilicity and low water contact angle, which reduced the ability to absorb sediment on the membrane surface; in a way that the presence of water on the surface of the membrane acts like a layer that did not allow the pollutant to enter the membrane pores. Therefore, it had the highest FRR (72.39%). According to Fig. 5 a, b, the increase in Rr by modifying the membranes was due to the reduction of surface hydrophobicity caused by the hydrogen bonding of TA nanoparticles and PEG additives on the membrane polymer surface. In modified membranes, M1 (57.92%) and M2 (56.86%) coated with PEG additives and TA nanoparticles, respectively, had higher Rr than other membranes. Also, M4 (47.93%) and M5 (53.81%) had more suitable reversible resistances than the M0 membrane (28.47%). In Kumar et al.’s (2016) studies, an increase in %FRR and Rr in the polysulfone membrane with GO-TiO2 nanoparticles in regard to the separation of HA has been reported. In general, the reduction of Rir and the increase of Rr in the coated membranes compared to the pristine membrane indicates the reusability of the membranes by hydraulic washing (Kallem et al. 2021).

Fig. 5
figure 5

a Membrane resistances, b total membrane resistance during 90 min (P = 1.5 bar, (Nazri et al. 2021) = 80 ppm, pH = 7)

Full size image
Porosity

In order to better compare the membranes and choose the optimal membrane, porosity was also investigated. Membrane porosity is measured according to the number and size of pores in the surface layer of the membrane. An increase in the number and size of holes in the surface layer of the membrane means an increase in porosity and, as a result, an increase in the permeate flux. Coating membranes with nanoparticles and additives improve their hydrophilicity, which also positively improves porosity. According to Table 2, the modified membranes M2 (49.38%), M4 (46.62%) and M5 (37.03%) had lower porosity than the M0 pristine membrane (56.98%). Membrane M1 (73.46%), covered merely with additives, had the highest porosity, which was the justification for the high flux of this membrane. M3 membrane (65.74%), which had an equal ratio of nanoparticles and additives, also had high porosity. Similar results have been presented in the studies of Kumar et al. (2016); Krishnan et al. (2022). Further, by coating nanoparticles on the membrane surface, the average diameter of the pores increases, which is due to the improvement of hydrophilic properties of membranes by coating nanoparticles. As can be seen in Table 2, M5 (48.48 nm) and M1 (44.59 nm) had the highest average diameter, followed by M2 (40.92 nm), M4 (34.24 nm), and M3 (29.89 nm). Needless to say, they all had higher average diameters than M0 (21.59 nm) (Sakarkar et al. 2020).

Table 2 Characterization, permeate flux, removal, and resistance of the considered membranes
Full size table

According to the experiments conducted in this research, a significant increase in pure water permeability flux (446.03 L/m2.h), membrane rejection (86.62%), and recovery ratio (64.17%) the M4 membrane compared to pure water permeability flux (265.64 L/m2.h), membrane rejection (75.59%) and recovery ratio (62.12%) of the M0 membrane. Furthermore, an increase in reversible resistance (47.93%) and a decrease in irreversible resistance (35.82%) were observed in the M4 membrane compared to reversible (28.47%), and irreversible resistance (37.87%) of the primary membrane have been observed. On the other hand, with the coating of TA nanoparticles on the membrane surface, the membrane pores were closed with a coating of nanoparticles, which reduced the porosity in the M4 membrane (46.62%) and also the effluent flux (72.43 L/m2.h) compared to the porosity (56.98%) and the effluent flux (89.39 L/m2.h) of M0 membrane, and the average particle diameter in M4 membrane (34.24 nm) after M5 and M1 membranes, is more favorable than M0 membrane (21.59 nm). With these in mind, the M4 membrane was finally chosen as the optimal membrane for further research.

pH optimization

Finally, the effect of pH on the flux and HA removal efficiency of the selected membrane M4 was investigated ( Fig. 6a–c). According to Fig. 6 a, b, with the increase in pH from 5–9, the flux decreased from 167.96 to 27.08 L/m2.h and the increase in membrane rejection from 80.79 to 96.26% was observed. It was justified according to dynamic light diffraction analysis (Fig. 6 c). Dynamic light diffraction analysis was performed to measure HA zeta potential and its particle size to determine the chemical properties and surface charge created on the membrane surface at pH 5, 7, and 9. In fact, with the increase in pH from 5 to 9, the decrease in the HA zeta potential (at pH 5, 7, and 9, − 14.56 mV, − 21.93 mV, and − 23.86 mV, respectively) and the increase in HA particle size (at pH 5, 7, and 9, respectively, 126.3 nm, 142.1 nm, and 160.3 nm) were observed, in a way that the more acidic the condition, the steeper the slope, and with the increase in pH, the slope decreased to the point where an almost flat line was observed in the alkaline condition. With the increase in the size of the HA particles and the HA zeta potential becoming more negative, it can be concluded that at low pH, the solution has little stability and settles in a viscous and bulky form in the solution. In fact, the functional groups of carboxyl (–COOH) and hydroxyl of HA are protonated, which causes a decrease in density, a decrease in electrostatic repulsion, shrinking of the polymer, and a decrease in the size of the HA particles, which leads to an increase in the average pore size and ultimately an increase in permeability. At high pH, the opposite of this situation occurs; that is, the carboxyl groups are separated into carboxyl ions (–COO−), which causes electrostatic repulsion, increase in density, enlargement of polymer, increase in HA particle size, and as a result, decrease in HA average pore size and ultimately decrease in permeability. In other words, pH has an influential role in HA particle size and membrane surface charge according to the type of membrane and pollutant. In the studies of Mondal et al. (2019), the water flux has decreased in pH 3 to 12 for the glucose pollutant.

Fig. 6
figure 6

a Flux-time, b Flux & HA removal (%)-pH c) HA particle size and zeta potential-pH M4 membrane during 60 min (P = 1.5 bar, (Nazri et al. 2021) = 80 ppm)

Full size image

Figure 7 a, b, respectively, shows the total permeate flux through the membrane in terms of time and the values of Jw1, Jw2, and Jp permeating the optimal membrane M4 at different pHs. According to these two Figures and according to what was said in the previous part, it can be seen that with the increase in pH, the flux through the membrane, including the initial and the final pure water flux, has decreased. According to the results, pH = 5 had the highest total flux (167.96 L/m2.h), the initial pure water flux (512.70 L/m2.h), and the lowest amount of HA removal (80.79%) and vice versa pH = 9 had the lowest total flux (27.08 L/m2.h) and initial pure water flux (373.24 L/m2.h), and the highest amount of HA removal (96.26%). Thus, pH = 7, which had a favorable total flux (72.43 L/m2.h), initial pure water flux (446.03 L/m2.h), and HA removal rate (86.62%), was selected as pH Optimum.

Fig. 7
figure 7

a General flux, b Jw1, Jw2, and Jp Flux changes passing through the M4 membrane at Different pH of HA during 60 min (P = 1.5 bar, (Nazri et al. 2021) = 80 ppm)

Full size image

Membrane characterization

The membranes underwent characterization through analysis. In this regard, ATR-FTIR, FESEM, AFM, and EDS analyses were conducted to comprehensively examine the surface structure and chemical bonding, surface morphology, topography, and elemental composition of membrane surfaces.

ATR-FTIR analysis

ATR-FTIR spectrum was used to examine the structure and chemical bonding of the surface of all membranes. Figure 8 shows the spectra of all membranes. Compared to the prime pristine membrane, the M4 membrane had an absorption peak at 1729.81 cm−1 due to the stretching vibration of C=O related to the hydrophilic carboxyl group (COOH) present in TA, indicates the proper coverage of TA nanoparticles on the membrane surface. These numbers were observed in the other four modified membranes with a slight change caused by the effect of coated nanoparticles and additives on the membrane surface with different percentages. Based on all the spectra of the modified membranes, it can be concluded that TA nanoparticles and PEG additives are uniformly dispersed on the PVDF membrane, which has also indicated similar results in the studies of Shi et al. (2018) and Turk et al. (2021).

Fig. 8
figure 8

ATR-FTIR spectra of all five membranes

Full size image

FESEM and AFM analysis

Surface morphology and topography were investigated by FESEM and AFM analysis of the pristine membrane (M0) and the selected modified membrane (M4), respectively. As can be seen in the cross-sectional images in Fig. 9 a, the M0 membrane (PEG: TA = 0:0) has a smooth and uniform surface, indicating the polymer membrane’s uniformity and homogeneity. The number of pores in this membrane is less but bigger. The cross-sectional images of the M0 membrane show high membrane density. In the modified M4 membrane (PEG: TA = 4:1), the number of pores has increased, but their size has become smaller, which ensures the coating of nanoparticles and additives on the surface of the membrane. Due to the hydrogen bond created between TA nanoparticles and PEG additives in the initial casting solution, masses have been made, which made the distribution of pores on the surface of the coated membrane not homogeneous. According to the FESEM images shown in Fig. 9 b, in the cross-sectional images of the membrane, two porous and dense spongy parts can be seen, the porous part has asymmetric finger-shaped pores, and the spongy part determines the permeability rate (Jiang et al. 2021; Yang et al. 2020). The M4 membrane has more regular pores and spongy parts, more porosity, and larger finger pores compared to the M0 membrane. In fact, with the increase in the percentage of nanoparticles and the hydrophilicity of the polymer matrix, the number of coarse pores decreased, and the finger-shaped holes became more regular in the M4 membrane. The formation of more pores in the finger-shaped part and fewer in the spongy part of the membrane indicates an increase in hydrophilicity and, consequently, membrane flux. Similar results have been presented in the studies of Li et al. (2019).

Fig. 9
figure 9

Images of the M0 and M4 membranes a FESEM top surface (500 nm), b FESEM cross Sect. (20 µm), and c AFM (10 µm \(\times \) 10 µm)

Full size image

Surface roughness is effective on membrane permeability, fouling, and, ultimately, rejection. Therefore, AFM analysis was performed for the M0 membrane, which was the pristine membrane, and the M4 membrane, which was more suitable in terms of membrane properties, hydrophilicity, and surface morphology than other modified membranes. As can be seen in Fig. 9 c, with the coating of the membrane surface in the M4 membrane, Ra increased from 30.98 to 37.10, which can be attributed to the porosity improvement. The highest points of the membrane surface are lighter than depressions and pores. The higher the coefficient of the membrane surface area means the creation of more bumps and depressions on the surface and, indeed, a higher surface roughness. Root mean square roughness can have very different results depending on the occurrence and distribution of surface roughness. The higher the Rq, the more uniform the surface roughness, which in the modified M4 membrane had a value of 48.83, which indicated the proper distribution of TA nanoparticles on the membrane surface and was also higher than the M0 membrane (39.24). In fact, it can be concluded that the membrane polymer solution creates a denser structure and a softer surface. By increasing the amount of nanocomposite and simultaneously with its migration to the membrane surface, the anionic interactions related to the polymer strands cause wrinkles in the modified membrane; accordingly, the roughness of the membrane surface increases (Zhang et al. 2021).

EDS analysis

According to Fig. 10 a, the elemental composition of M0 and M4 membrane surfaces was determined by EDS. As can be seen, by covering the surface of the membrane, the amount of carbon elements (from 51.27 to 75.14), oxygen (from 1.05 to 16.11), and nitrogen (from 0.81 to 1.05) in the modified M4 membrane compared to the pristine M0 membrane has increased, and the amount of fluorine (from 46.86 to 7.7) has decreased, which is due to the hydroxyl and ether functional groups present in TA nanoparticles and PEG additives. Similar results have been reported in the studies of Fahrina et al. (2022). These results are also observed in the EDS mapping images, according to Fig. 10 b, the nanoparticles and additives are uniformly and homogeneously dispersed on the M4 membrane surface. In fact, the fluorine element in the M0 membrane was high due to the PVDF polymer in the pristine membrane. Additionally, the increase in the carbon element in the modified membrane (M4) compared to the M0 membrane was due to the coating of the membrane with TA nanoparticles and PEG additives.

Fig. 10
figure 10

a EDS spectra b EDX mappings of the membrane surface of (1) M0, and (2) M4 membranes and total mappings of (3) M0 and (4) M4 membranes

Full size image

Conclusion

In order to achieve optimal removal efficiency, the PVDF pristine membrane was surface modified with TA nanoparticles and PEG additives in this investigation. Compared to the unmodified pristine membrane, an improvement in pollutant absorption was observed in the membrane modified with TA nanoparticles. In fact, the interaction between the –OH in TA nanoparticles and the –O– in PEG additives resulted in the formation of hydrogen bonds, enhancing the membrane’s antifouling properties, hydrophilicity, pure water permeate flux, and the pollutant removal efficiency. Still, an increase in the percentage of nanoparticles did not mean an increase in pollutant absorption. Further, the modified membranes did not always exhibit an increase in permeability due to membrane fouling caused by the adsorption of HA particles on the membrane surface. The results showed that by increasing the pressure, the weight percentage of PEG additives, and decreasing the weight percentage of TA nanoparticles, permeability increased and rejection decreased. Also, with the increase of the pollutant concentration, permeability and rejection both decreased. Furthermore, at alkaline pHs, a decrease in permeability and an increase in rejection were observed. According to the analyses performed for different membrane types, the M4 membrane provided the best performance. Reducing the water contact angle with the coating of nanoparticles and additive on the membrane surface due to the increase of hydrophilic hydroxyl (OH) and carboxyl (COOH) groups, as well as the increase and regularization of pores in the finger-shaped part- so that the coarse pores decreased, but the number of pores increased, causing an increase in pure water flux, which can be related to the coating of these pores by TA nanoparticles. According to the results of this research, the membrane coating method, by affecting the morphology and surface properties of the membrane, effectively changes the flux and removal efficiency of the membrane, as well as the antifouling properties of the membrane.