One of the most important environmental concerns is the contamination of soil, underground, and surface water via wastewater containing heavy metal ions. These metal ions are usually discharged into the environment by the untreated wastewater of various chemical industry activities. Unlike organic contaminants, heavy metals cannot be degraded or destroyed. Hence the presence of these contaminants in the environment and their accumulation in living organisms can cause destructive effects on the ecology, health of animals, and human beings. Manganese element as one of the important heavy metals has been extensively used in a variety of industries such as ceramics, dry battery cells and electrical, coils, paint, painting materials, glass, match, catalysts manufacturers, galvanized sheets, and especially in the steel production . In addition to the discharge of the untreated wastewater of these industries into the environment, anthropic activities and the burning of oil and coal are other main source of water pollution by manganese . On the other hand, 56Mn as one of the manganese radioisotopes is formed during normal operation of a nuclear power plant’s reactors as a result of the neutron activation of 55Mn in the primary circuits of water reactors. The corrosion of structural material of nuclear power plants such as stainless steel equipment and concrete is the main source of the presence of this element (55Mn) in the reactors cooling water. Also, 54Mn is another radioisotops of manganese which is a versatile radiotracer for use in various applications . Therefore, the presence of these radioisotopes in the environment with the half-lives of t1/2=312 d (54Mn), t1/2=2.57h (56Mn), and gamma emitters can create risks for biological systems. The recommended safe level for manganese concentration in drinking water by European and developed countries such as the United Kingdom, United States, Japan, and Canada is 50 μg.L-1 . The presence of an excess amount of this ion in the water and its entrance into the body of a human being by ingestion can cause diseases such as Parkinson’s disease, coronary heart disease, birth defects, impotence in men, insomnia, and alterations in bone formation . Therefore, the removal of even trace amounts of this element and its radioisotopes from wastewater has been of particular interest to scientists around the world. Various techniques such as chemical precipitation, adsorption, biological and contact catalytic oxidation, coagulation/flocculation, membrane filtration, ion exchange, etc. [6-8] have been applied for the removal and separation of manganese ions from wastewater. Among these methods, the nanofiltration process as an intermediate process between reverse osmosis and ultrafiltration is an attractive and effective method for the treatment of effluents. At the same time, this technique benefits from relatively high permeation flux, high rejection efficiency, unique removal capability for divalent ions, low energy consumption, and easy operation .
The main core of membrane technology is the membrane material . Polymeric membranes due to their great capabilities in the field of wastewater treatment and the production of different types of asymmetric nanofiltration membranes are most often used in these processes . Some of the most important polymeric materials that are commonly used in the fabrication of membranes are aromatic polyamides (PA), cellulose acetate (CA), polyvinylidene fluoride (PVDF), polyethersulfone (PES), polysulfone (PSU), and polypropylene (PP) . Among them, polyethersulfone has attracted the attention of many researchers  in the removal and separation processes. That is because of its specific properties such as film formation property, good compatibility with hydrophilic additives, thermal and mechanical stability, acceptable chemical resistance to acids and oxidizing substances like fluorine and hydrogen peroxide along with tolerance to a wide range of temperatures and pH, low sensitivity to UV, easy processing, and a broader range of pore sizes . However, the poor inherent antifouling property of this polymeric membrane has limited its application as an efficient membrane in the nanofiltration process . Hence, to overcome this respective shortcoming and enhancement of wettability and performance of PES, a large number of mixed matrix membranes (MMMs) have been extensively synthesized in the last few years . This type of membrane takes advantage of the merits of both polymers and filler materials. In this regard, one of the most effective approaches is the embedding of the hydrophilic nanoparticles in the polymeric matrix . The introduction of nanoparticles into the polymeric membranes not only can improve their antifouling properties but also modify the hydrophilicity, strength, stiffness properties, and water permeability of these polymers . Some of the nanoparticles that have been used in the fabrication of MMMs based on polyethersulfone are; graphene oxide, functionalized multiwalled carbon nanotubes, activated carbon, zinc oxide, iron oxide, tin oxide, aluminum oxide, titanium oxide, cerium oxide, copper oxide, and so on [16-18]. The super hydrophilic character of Fe3O4 nanoparticles makes them promising toward large quantities of OH functional group attachments. Therefore, the incorporation of Fe3O4 nanoparticles can appreciably improve the hydrophilicity of the PES membrane, thereby enhancing the water permeability and resistance of antifouling while treating manganese ions (19). Also, these nanoparticles can adsorb metal ions and increase their rejection from aqueous solution in the membrane process. Nevertheless, a review of the kinds of literature indicates that no studies have been yet done on the removal of manganese by MMMs containing PES and Fe3O4 nanoparticles. Therefore, this study aims to fabricate MMMs by introducing Fe3O4 nanoparticles into PES and investigating their ability of them in the reject of manganese ions from aqueous solution.
Manganese chloride, N-methyl pyrrolidone (NMP), polyvinylpyrrolidone (PVP), Fe3O4, and other chemicals and reagents were obtained from E. Merck or Fluka companies. Polyethersulfone with a molecular weight of 58000 was provided from BASF SE, Germany.
Synthesis of Fe3O4 nanoparticles
The synthesis of Fe3O4 nanoparticles in the solution phase was carried out as follows . 2.22g of Iron(III) chloride and 0.866 g of Iron(II) chloride were dissolved in 40 ml of demineralized water. While the temperature was gradually increased up to 70 °C under nitrogen atmosphere and refluxing conditions, the solution was stirred at 1000 rpm. After 20 min, a specific value of ammonia solution (0.5 M) was added and then the mixture was stirred for 30 min. Afterward, 2 mL of citric acid solution (2.6 mM) was added to the mixture. The temperature of this mixture was gradually raised to 90 °C while the mixture was stirred and refluxed. Finally, after cooling the mixture, the black Fe3O4 nanoparticles were filtered, washed five times with demineralized water, and dried in an oven at 50 °C for 48 hours.
The flat sheet membranes including neat PES and PES/Fe3O4 were fabricated by wet phase inversion and conventional casting methods. To do this, at first, the determined amount of neat polyethersulfone was mixed with N-methyl pyrrolidone as a solvent. After an additional 0.25 wt. % polyvinylpyrrolidone, this mixture was sonicated for 30 min and stirred for 24 h to distribute and solve entirely these particles in NMP. Also, to prepare the MMMs, the various amounts of prepared Fe3O4 were added to NMP and homogenized by ultrasonic bath. After that, PVP and PES were added and dispersed. The air bubbles into these polymeric solutions were omitted using an ultrasonic bath for 30 min and allowed to settle overnight. To prepare flat sheet membranes, these mixtures were cast on a glass plate with a casting knife with a thickness of 100µm and immediately they were immersed in a demineralized water bath. The prepared polymeric films were kept in demineralized water for 24 hours. Finally, these membranes were dried at 25 °C for 24h .
Membrane porosity and mean pore size
The overall membrane porosity (ε) can be calculated by Eq. (1) as follows:
where Wd, Ww, Vm, and ρf are the weight of the dried membrane (g), the weight of the wet membrane (g), membrane volume (cm3), and water density (g.cm-3 ), respectively. The experiments were carried out three times for each sample to minimize the experimental errors.
The Guerout–Elford–Ferry equation (Eq. (2)) was also applied to estimate the mean pore sizes of prepared membranes by using pure water flux and porosity .
where η, Q, ∆P, L, A, and ε are the water viscosity (8.9 × 10−4 Pas), permeated water flow rate (m3.s-1), operating pressure (1 Mpas), membrane thickness (80 × 10−6 m), membrane area
(33 × 10−4 m2) and the porosity, respectively.
Field emission scanning electron microscopy (FESEM) model S-4160, Hitachi (Japan) was used to characterize the morphology of the synthesized nanoparticles and flat sheet membranes. Transmission electron microscopy (TEM) micrographs were obtained with a transmission electron microscope (CM30, Phillips) using the acceleration voltage of 200 kV. The sessile drop method using a contact angle measurement instrument (OCA15EC, Data Physics, Germany) was utilized to measure the static contact angle. Optical microscope images were immediately captured for subsequent contact angle measurement. Inductively coupled plasma-optical emission spectroscopy (ICP-OES) technique was used to the amount of manganese ions. To evaluate the performance of the fabricated neat PES and MMMs for manganese ions rejection, a cross-flow stainless steel nanofiltration (NF) setup was selected (Fig. 1). As shown in Fig. 1, the NF setup is outfitted with a membrane cell with an effective area of 33 cm2. The operating pressure of both sides of the cell was monitored by two pressure gauges (Wika, Korea). In addition, a high-pressure dosing pump (Jesco, Germany) was used to transfer the feed solution into the membrane cell. Other main constituents used on the NF setup were two flow meters on permeate and retentate streams, one pressure control valve, and one pressure safety valve.
RESULTS AND DISCUSSION
FESEM and TEM images were utilized to study the structure and morphology of the fabricated flat sheet membranes and nanoparticles. Fig. 1 shows a TEM image of Fe3O4 nanoparticles which proves the nanometric structure of these prepared nanoparticles. Also, as can be seen in Fig. 2, iron oxide particles are relatively spherical and their size is less than 30 nm.
Cross-section and top surface SEM images of mixed matrix membranes and the neat PES have been presented in Fig. 3. As can be seen in Fig. 3a, the relatively thin and dense active skin layer has formed the top surface of all fabricated membranes which plays the key role in the rejection of ions in the solution. Also, the typical porous and similar asymmetric structures with this active skin layer supported by the fully developed finger-like cavities and macroporous sublayer can be seen in fabricated flat sheet membranes. This observed structure is typical for PES membranes manufactured using N-methyl pyrrolidone as a solvent and DMW as a coagulation batch . Also, these images show that with addition of Fe3O4 doesn’t change the asymmetric structure of these membranes. As can be seen in Fig. 3, by increasing these nanoparticles in the casting solution, in addition to the relative increase in thicker active layer, the finger-like structures of these fabricated MMMs became longer, straight, and wider compared to the nest PES (Fig. 3b). With addition of nanoparticles, the casting solution viscosity is increased and it could result in the delayed exchange of NMP and DMW as a solvent and non-solvent, respectively . This phenomenon led to the formation of a less porous membrane, a relatively smaller number of finger-like macro voids with denser and thicker skin layers .
The membrane surface hydrophilicity property is one of the important parameters that can influence permeate flux and antifouling ability of the membranes. As proved in previous studies, higher hydrophilicity properties of the membranes will lead to higher permeation flux and a lower tendency toward biofouling, too . Therefore, the water contact angle measurement as a common method was used to characterize the hydrophobicity of the prepared membrane surfaces. Table 1 summarizes the obtained results of the water contact angle at three random locations for the fabricated membranes to minimize the experimental error. As can be seen in Table 1, the amount of contact angles was reduced after the incorporation of Fe3O4 into the neat PES membrane structure. A lower contact angle indicates that the fabricated mixed matrix membrane should have a more hydrophilic surface and better wettability to water which results in higher water flux and antifouling properties.
Table 2 presents the porosity and mean pore size of prepared neat PES and Mixed matrix membrane. As can be seen, the porosity and mean pore size for the fabricated mixed matrix membranes are higher and larger, respectively, compared to the neat PES membrane. This can be assigned to increase of phase inversion rate mainly by utilizing hydrophilic Fe3O4 nanoparticles in the casting solution. In addition, the possible interactions between nanoparticles and polymer binder can reduce the interactions between polymer-polymer chains which leads to the production of large micro-voids . The decrease of porosity and pore size of nanocomposite membrane at a high additive ratio may be attributed to the increase of casting solution viscosity which reduces the mass exchange rate .
Permeate flux and manganese ion rejection experiments
Permeation tests including permeate flux and the amount of manganese ion rejection were performed to evaluate the efficiency of the prepared polyethersulfone membranes and its mixed matrix membranes. For this purpose, the operational conditions were firstly optimized for the rejection of manganese ions using the flat sheet membrane of the neat PES, and after that, the rejection behavior of the mixed matrix membranes was investigated and compared in the obtained optimum conditions.
Fig. 4 shows the manganese ion rejection percentage and the amount of permeate flux as a function of PES concentration. The used membranes were fabricated by wet phase inversion and conventional casting method using the preparation of four solutions containing 15, 18, 20, and 22 wt.% PES in NMP as a solvent. As shown in Fig. 4, with increasing PES concentration in casting solution, the rejection percentage of manganese ion is increased while permeate flux is decreased. This can be explained by increasing the casting solution viscosity as a result of the rise in the concentration of polymer. Consequently, the rate of the solvent transfer into the non-solvent and the rate of the non-solvent into the polymer will be decreased which leads to a lower porosity and a lower permeability of the membrane . At the same time, decreasing the size of the membrane pore results in I) the rise in sieving property of the prepared flat sheet membrane and consequently, the rise in the manganese ion rejection and II) the decrease in the value of solvent passing through the membrane and thus, the deduction of the amount of permeate flux or permeability . To test and evaluate the membrane’s performance, an optimal concentration of polymer should be selected, therefore an appropriate permeate flux should be combined with a good rejection of manganese ions. The results show that the rejection of manganese ions increases with the rise of polymer concentration. Accordingly, the highest possible concentration of polymer would be advantageous for this rejection, but it is disadvantageous for the value of permeation flux. The optimal concentration was fixed as 20 wt.% polyethersulfone. In the following of this research work, all the membranes made with this concentration will be studied.
Another important factor that can influence the membrane surface charge and the type of metal species is the solution pH. Therefore, this factor can affect the rejection properties and the amount of permeate flux . For this reason, the pH of the solution is usually the key factor for controlling the rejection process . The relationship of permeate flux and manganese rejection percentage with the solution pH is illustrated in Fig. 5. Based on the solubility product of Mg(OH)2 (pKsp=11), the predominant species of manganese ion at pH values above 7 is Mg(OH)2. Therefore, these experiments were performed up to 6.5. As shown in Fig. 5, both rejection percentage and permeate flux increase with a rise in the amount of the solution pH. It is well known that the negative surface charge of the PES membrane originates from the functional groups of PES (O=S=O) . A higher solution pH value results in an increment in the negative surface charge of the PES membrane which leads to stronger electrostatic charge repulsion between the anion in the solution (chloride ion) and membrane surface. This increases the rejection of chloride ions and consequently, the rejection of manganese ions increases due to the Donnan effect . On the other hand, at low pH where the hydronium ion concentration is high, a progressive neutralization of negative sites of the PES membrane surface takes place and the negative surface charge of the membrane reduces. Therefore, chloride ions can easily pass through the membrane and following that, to maintain the charge balance of the solution, manganese ion crosses the membrane . Also, the increase of membrane permeability at higher pH may be due to the expansion of the cross-linked-membrane polymer network and the highest net driving force due to the lowest osmotic pressure at the membrane surface as a result of the reduction of concentration polarization effect .
Information on the effect of pressure on the rejection of manganese ions and permeate flux was obtained by variation of the operating pressure in the range of 7 to 15. The results are shown in Fig. 6. As can be seen, the increase in the operating pressure results in an increase in permeate flux. This increase could be due to the enhancement of driven force (pressure) and adsorption of water on the membrane surface which leads to the increment of solvent passage through the membrane . Moreover, as shown in Fig. 6, manganese ion rejection declines with the rise in the operating pressure. This phenomenon could be explained by the fact that the increasing pressure will cause more manganese ion transports and aggregates on the surface of the PES membrane through convective transport. Therefore, the concentration polarization will increase and result in a decrease in manganese ion rejection by decreasing the charge effect .
Fig. 7 presents the curve illustrating the efficiency of manganese ion rejection and permeate flux by the neat PES membranes studied in terms of the initial concentration of manganese ions in the solution. The operating pressure was held constant at 10 bar. As can be seen from this figure, an increase in manganese ion concentration leads to the decline of manganese ion rejection and permeate flux. The decline of permeate flux and the rejection of manganese ion with the increasing of the feed solution concentration can be explained as follows:
- I) As the concentration increases, the concentration polarization effect becomes more obvious and weakens the electrostatic forces between the membrane and the ions in the solution which finally results in the passing of manganese ions through the membrane and decreases its rejection [29, 32].
- II) the increase of manganese ion concentration in the feed solution enhances osmotic pressure which leads to the decline of permeate flux and thus the rejection decrease due to the decrease of solvent passing through the membrane .
To evaluate the performance stability of the flat sheet membrane of PES, long-term tests were performed. Fig. 8 shows permeate flux and manganese ion rejection behavior versus time.
As shown in Fig. 8, with elapsing the experimental time (after 45 min), the amount of permeate flux declined. This behavior can be due to the effect of various factors such as the formation of concentration polarization and pores clogging as a result of ion aggregation on the surface of the membrane  that hinder the solvent from passing through the membrane and increase manganese ion rejection, too.
The influence of Fe3O4 content on manganese ion rejection percentage and permeate flux was studied. The obtained results have been shown in Fig. 9 and indicate that the rejection percentage of manganese ion was enhanced with a rise in the amount of Fe3O4 in the PES membrane up to 0.1 wt.%. The reason for that is the homogeneous distribution of Fe3O4 nanoparticles throughout the PES membrane which can form a dense top layer and improve the rejection percentage of manganese ions . Moreover, by adding and well dispersion of these nanoparticles in PES membrane (up to 0.1 %wt.) and consequently increment of available active sites and higher surface area of nanoparticles, manganese ion rejection percentage increases because of the inherent properties of adsorption of nanoparticles . At higher nanoparticle concentrations (>0.1 %Wt.), the possibility of nanoparticles agglomeration reduces the water flux. In addition, pores filling/blockage would decrease the size of membrane channels which decline the water flux . The higher nanoparticle percentage in MMMs decreased salt rejection slightly. Decrease of manganese ion rejection for MMMs with 0.2 %wt. nanoparticles is assigned to its high porosity and larger pore size which make possible the salt percolation through the membrane.
In this test, the best efficiency was found in MMMs containing PES/Fe3O4 (0.1 wt.%). Despite the increment of the hydrophilicity properties of MMMs containing PES and Fe3O4 nanoparticles (Table 1), permeate flux is associated with a decrease and this can be due to the increase in the viscosity of the casting solution by adding Fe3O4 concentration. As mentioned before, this decreases the rate of NMP as a solvent into the demineralized water (non-solvent) and the rate of the non-solvent into the polymer which leads to a smaller porosity and a lower value of permeate flux .
Table 3 shows the performance of the prepared MMMs (PES/Fe3O4) along with the neat polyethersulfone membrane. As can be seen, the rejection percentage of the prepared MMMs is higher than the neat PES membrane which indicates an increase in the performance of these membranes as a result of increasing these nanoparticles. Also, this table shows that the permeate flux amount of MMMs has decreased as compared with the neat PES membrane. Despite the increase in hydrophilicity properties of the membranes as a result of the increase of these nanoparticles, the amount of permeate flux in MMMs is reduced. This can be described concerning the mutual effect of these nanoparticles on membrane porosity increment against reduction of mean pore size which decreases the membrane flux (Table 2).
To compare a summary of some research reports investigated for manganese removal by membrane processes has been presented in Table 4.
In this study, wet phase inversion and conventional casting methods were used to prepare the nanofiltration membranes of the neat PES and MMMs. The hydrophilicity properties of the neat PES and the rejection performance of manganese ions were improved using Fe3O4 nanomaterials. The best performance of the neat PES was achieved at 20 wt.% PES, pressure 10 bar, lower concentration of manganese ions, and pH=6.5. The investigation of MMMs performance prepared by the addition of nanomaterials in the structure of PES indicates that under similar operational conditions, the rejection percentage of manganese ions has significantly been improved, while their permeability of them was declined. The best performance of the fabricated membranes was found to belong to MMMs containing PES membranes (20%) and Fe3O4 (0.1 wt.%). Therefore, we can conclude that this prepared membrane can used for the removal of manganese and its radioisotopes (54Mn, 56Mn) from aqueous solution and nuclear wastewater, too.
CONFLICT OF INTEREST
The authors hereby declare that there is no conflict of interest.