Synthesis and characterization of novel sulfanilic acid–polyvinyl chloride–polysulfone blend membranes for metal ion rejection

Abstract

Near-complete removal of heavy metals, namely Cd(II), Cr(VI) and Pb(II), has been attempted by a membrane purification process using a blend of modified polyvinyl chloride (PVC) and polysulfone (PSf), prepared by the diffusion induced phase separation (DIPS) method. The prepared novel material was characterized by NMR, ATR-IR spectroscopy and DSC. The sulphonyl groups incorporated into PVC enhance the hydrophilicity and are substantiated by water uptake, contact angle (CA) and flux studies. The obtained properties of the blend membrane like increased surface roughness and porosity are observed from AFM and SEM analysis. An enhanced rejection of ∼95% which is about 1.15, 1.41 and 1.37 times better than the commercially available NF 270 membrane was observed, for Cd(II), Cr(VI) and Pb(II) respectively. The work was further extended to study the antifouling property and the interference of other existing metal ions on the performance. An improved antifouling property with 98.5% rejection for bovine serum albumin (BSA) and a 75.6% flux recovery ratio (FRR) was achieved. The study gains significance in exploring the incorporation of sulphonyl groups in to polymers, to enhance membrane performance.

---

1. Introduction

The concept of water salvaging has gained a tremendous boost in recent times due to the rise of contaminants being discharged into water resources.¹ Among the various pollutants, heavy metals are abundantly present in polluted waters and are highly toxic. Human activities such as electroplating, wood processing, leather tanning, and anodizing produce wastewater containing heavy metals causing immense intimidation to mankind.² Heavy metals such as lead (Pb), cadmium (Cd) and chromium (Cr) can cause various disorders in humans such as nephrotoxicity, neurotoxicity, effects on the cardiovascular system, and effects on cellular anti-toxicity defences, damage to oxidative DNA repair systems, and kidney and liver problems.^3–5 Moreover, they are non-biodegradable and tend to accumulate in the tissues of living organisms⁶ when ingested even at very low concentrations. Hence, in the present era an effective removal method for such heavy metals is necessary.⁷ There are several methods such as chemical precipitation, ion-exchange, membrane filtration, the membrane filtration technique and adsorption⁸ for heavy metal removal.

The membrane filtration technique has shown great promise towards water purification.⁹ The feasible operating conditions, energy savings and high efficiency have created new interest in heavy metal removal with membrane technology.¹⁰ Membrane technology is a servomechanism technique in separation and purification research. Several efforts from researchers are continuing for the betterment of membrane separation in respect to better productivity with minimal selectivity, ease of operational conditions and the ability to be scaled up at low costs,¹¹ including the use of other membrane processes such as reverse osmosis (RO) and nanofiltration (NF).^12,13 However, UF membranes are not suitable for metal ion rejection due to their larger pore size compared to the size of metal ions¹⁴ in comparison to nanofiltration membranes which have shown good results towards the removal of heavy metals.¹⁵ But this setback can be resolved by modifying the polymer chemically by incorporating functional groups which in turn induce a charge on the membrane surface.¹⁶ The charge on the membrane can considerably mitigate fouling and can append to improve the efficiency and life span of the membrane.¹⁷

Polymers show great promise for the production of membranes due to their good flexibility, toughness and separation capability.¹⁸ PVC is one of the most versatile polymers with good flexibility, chemical resistance, and film forming property, as well as being economical and easy to modify.^19,20 PVC has been extensively used in industrial applications, namely as a packing material, cable insulation, and in pipes. However, its use as a membrane for water filtration is limited due to its hydrophobicity and membrane fouling property.²¹ The polymer properties can be changed by chemical modification. In this regard, chemical modification of PVC promises to render the necessary properties like membrane charge and hydrophilicity. The surface charge and hydrophilicity are very necessary tools for improvement in the selectivity of ions and productivity in membrane separation. There are reports already available in the literature on PVC modification, where Tooma modified PVC with ethyl acrylate monomer for vacuum membrane distillation (VMD) application.²² A methylimidazolium group-modified PVC doped with phosphoric acid membrane was used for proton exchange by Che.²³ Conversely there are a few reports on PVC blends and composite membranes for salt rejection by Cui using blend PVC/P(DMA-co-MMA) precursors,²⁴ a PVC–silica mixed-matrix membrane as a membrane bioreactor by Bilad²⁵ and CO₂ gas separation using a PVC/Pebax composite membrane by Ahmadpour.²⁶ But none have studied the removal of metal ions. Hence, the aim of this study is to prepare a novel polymeric membrane by a novel approach to obtain a synergistic effect of optimum antifouling, selectivity and productivity.

PVC was modified using sulfanilic acid (SA) by a simple chemical reaction. The present paper reports a novel functional group as a substituent, a novel chemical modification method for PVC modification, and its use in a membrane for water purification. To the best of our literature knowledge, such work has not been reported yet. Modified PVC is further blended with PSf (with different compositions) to regain its lost strength after modification via the DIPS method. The chemical modification of PVC was confirmed by NMR, ATR-IR and DSC. A blend membrane has properties like enhanced surface roughness and porosity, which were observed from AFM and SEM analysis. The sulphonyl group present in the modified PVC enhances the hydrophilicity and it was confirmed by water uptake, contact angle measurements and flux studies. The prepared blend membranes were used for rejection of metal ions, as the blends were charged and charge can play a very important role in the rejection of heavy metals in the ultrafiltration range. To validate the applicability of the blend membranes to real samples, the rejection in the presence of interfering ions is also studied and the results were also compared with the commercially available membrane NF 270 under the same operating conditions. The antifouling properties of the prepared membrane were also examined and the effects of modification on the membrane performance were discussed.

2. Experimental

2.1. Materials and methods

Udel PSf with a molecular weight of 35 000 Da, PVC with a high molecular weight and 1-sulfanilic acid (4-(H₂N)C₆H₄SO₃H, molecular weight: 173.19) were obtained from Sigma-Aldrich Co, India. Lead nitrate, cadmium nitrate and potassium dichromate were used as the source for Pb(II), Cd(II) and Cr(VI) respectively. Chemicals like dimethyl sulfoxide (DMSO), triethylamine and methyl-2-pyrrolidone (NMP) were of Merck brand. The commercial membrane NF 270 was purchased from Sterlitech, USA.

2.2. Synthesis of PVC–SA

Modification of PVC is being studied by Herrero et al.²⁷ using different functional groups. Similarly, Huang et al.²⁸ reported aminoethanol substitution on PVC for Heck reactions. In our approach, the reaction was carried out using a two-neck round bottom flask mounted on a magnetic stirrer which is equipped with a hot plate. 1 g of SA was dissolved in 50 mL DMSO, together with 10 mL of triethylamine to abstract the proton from the SO₃H group of SA. Subsequently, 1 g of PVC was added to the same solution and the reaction mixture was stirred for 24 hours at 60–65 °C at a constant rate of heating in nitrogen atmosphere. The reaction scheme is proposed in Fig. 1. The reaction was quenched by precipitating out in a mixture of cold water/methanol (1 : 2) after 24 hours. The precipitate was then washed with distilled water several times to remove the excess solvent and dried overnight in an oven.

Fig. 1 Schematic representation of the reaction scheme for PVC modification.

The % of modification (% M) and % modifying efficiency (% ME) of the graft PVC polymer were calculated using the following equation as described elsewhere:²⁹

2.3. Preparation of the blend membrane

The membrane was prepared by the DIPS method. The DIPS process is based on the phenomenon of precipitation of the polymer in a controlled manner, when it comes in contact with water to form a solid porous film. 4 g of PVC was dissolved in 16 g (15.54 mL) of NMP and stirred for 24 hours at 60 °C to obtain a viscous solution. The obtained viscous solution was then casted on a glass plate using a glass rod. The thickness of the membrane was maintained using double-sided tape. The glass plate was dipped in a coagulation bath containing distilled water at room temperature.³⁰ The membrane was peeled off by the phase inversion method and the obtained membrane was washed and stored in distilled water for 24 hours to remove any excess NMP and also to gain mechanical strength. Similarly, a neat PSf membrane was prepared by the same method by adding PSf instead of PVC. The same procedure was extended to prepare blend membranes; in this case both the polymers were added in different concentrations. A series of blends was prepared with different compositions as depicted in Table 1.

Table 1 Composition of concentrations of the polymers for membrane preparation

Sr. No.	Membrane	Composition (%)
		PSf	PVC/SA–PVC	NMP
1	PSf	20	—	80
2	PVC	—	20	80
3	3% PVC–SA/PSf	19.4	0.6	80
4	5% PVC–SA/PSf	19.0	1.0	80
5	8% PVC–SA/PSf	18.4	1.6	80
6	10% PVC–SA/PSf	18.0	2.0	80

---

2.4. Characterization of the membrane

2.4.1. ¹H NMR and ATR-IR studies. ATR-IR spectra of the membranes were recorded using a Bruker ECO-ATR spectrophotometer in the range 600 to 4000 cm⁻¹. The polymer powder was properly dried under vacuum for 48 hours before recording the spectra.

¹H NMR was recorded on a Perkin-Elmer EM 300 MHz spectrometer using tetramethylsilane (TMS) as the internal standard. The samples were dissolved in dimethyl sulfoxide (DMSO) at 60 °C before analysis, as the polymer was partially soluble at room temperature.

2.4.2. DSC analysis. Differential scanning calorimetry (DSC) was recorded on a Perkin Elmer Pyris 1 instrument at a heating rate of 3 °C min⁻¹. Scans were carried out from 50 to 900 °C to determine the change in the glass transition temperature after modification of the polymer.

2.4.3. Ion-exchange capacity. Measure of the ion-exchange capacity (IEC) provides information on the density of the functional groups present in the membrane matrix. Functional groups provide charge to the membrane, which is important for the conductivity and transport properties of the membrane.³¹ In this case the SO₃H groups present in the membranes have the capacity to exchange ions. The prepared blend membranes were immersed in 1 M HCl for 24 hours in order to completely ionize the functional groups to H⁺, if any unionized groups are present. The membranes were then rinsed with distilled water and further dipped in 1 M NaCl solution for the next 24 hours to bring about the exchange of cations to form membrane-Na⁺. The IEC was then calculated by back-titrating the obtained HCl solution against 0.01 M NaOH solution using phenolphthalein as an indicator. The IEC is given as:where A is the concentration of NaOH solution used, B is the volume of NaOH solution consumed and m_dry is the dry weight of the deprotonated membrane.³²

2.4.4. Atomic force microscopy (AFM). AFM analysis proves to be a very important tool to study and analyze the topography of the membranes. The analysis was undertaken using APE Research model F 80 AFM with SPM data analysis software. Lead cantilever probes were used for sample scanning. The cantilever probes were oscillated on a membrane sample size of 1 cm × 1 cm for an effective area of 5 μm. The scanner stopped moving the sample toward the probe when the amplitude decreased by 10% owing to the tip–sample interactions. Once engaged, the oscillation amplitude was automatically adjusted by the servo system for optimal imaging quality. At least three replicates were performed for each membrane sample and the surface roughness was reported.

2.4.5. SEM analysis. SEM images were recorded using a Shimadzu scanning electron microscope. The samples were first dipped and cracked in liquid nitrogen in order to get a clear image. The surface of the membrane was then gold sputtered to capture the images.

2.4.6. Water uptake and contact angle. The water uptake capacity was determined using distilled water in order to know the affinity of the membranes towards water. The membranes were scrupulously washed with distilled water and then dried for 24 hours in a vacuum desiccator. The dried membranes of area 1 cm² were cut and immersed in distilled water for 24 hours. The swollen membranes were taken out, wiped of excess water using tissue paper and weighed swiftly. The percent of water uptake was calculated using the equation:where W_w and W_d are the weights of the wet and dried membranes respectively.³³

The surface hydrophilicity of the membrane was measured by standard goniometer Ramehart-200 F1 as per the sessile droplet method. To minimize the errors, the contact angle was observed at five different positions of the membrane and the average is reported.

2.4.7. Water flux. Pure water flux experiments were carried out at room temperature using a self-constructed dead end cell filtration unit (ESI S.1†), which was fixed with a nitrogen cylinder in order to provide the required pressure. The modified blend membrane and neat PVC membrane were cut into a round shape and fixed into the static cell with a 5 cm² area. The pure water permeability of the membrane was calculated directly by measuring the permeate in terms of liters per meter square per hour (L m⁻² h⁻¹). The flux study was also carried out for the NF 270 membrane in a similar manner; the membrane was dipped in distilled water for 24 hours prior to use.

The slope for the plot of water flux and the applied pressure gives the hydraulic permeability coefficient (L_p), with the intercept fixed at zero and calculated from the equation:³⁴

L_p (m s⁻¹ Pa⁻¹) = slope × 2.77 × 10⁻¹⁰ (5)

2.4.8. Rejection of different metal ions. The rejection of the different metal ions Pb(II), Cd(II) and Cr(VI) was investigated using a blend of PVC–SA/PSf membranes of different compositions and a NF 270 membrane in a dead end filtration unit, the same equipment as was used for the water flux study. The filtration cell was placed on a magnetic stirrer to agitate the feed solution and avoid fouling. The feed solution was prepared as mentioned in the literature, where a low concentration chromium solution was used based on the UF application process which is known to be useful only for dilute solutions.³⁵ A 10 ppm concentration was maintained constantly for all the individual metal ion solutions, with 30 ppm as the concentration for the mixture of metal ions. The pressure was varied from 100 to 500 kPa and after every 10 mL of the permeate sample collection, the concentration of the metal ions was analyzed using a 55 AA Atomic Absorption Spectrophotometer Agilent Technologies in an air–acetylene flame. The accuracy of the AAS measurements was evaluated using the feed solutions as standards. Less than 5% deviations were recorded between the AAS readings and the actual feed concentrations. The percentage rejection was calculated using the following equation:where C_p and C_f are the concentrations of the permeate and feed respectively.³⁶

To investigate the competitive behavior of different metal ions towards rejection, the experiments were studied in the presence of the mixture of metal ions.

2.4.9. Antifouling study. A standard protein solution of bovine serum albumin (BSA) was used to determine the antifouling properties of the blend membranes. Initially the pure water flux (J_w1) was measured at room temperature with a fixed pressure of 500 kPa. In the proceeding step, the filtration unit tank was filled with a 200 ppm solution of BSA and the flux was observed every 10 minutes for a continuous period of 1 hour. After the filtration, the membrane was washed with warm water for 30 minutes, followed by measurement of the pure water flux (J_w2). The fouling resistance of the membrane was measured in terms of the flux recovery ratio (FRR) using the formula:³⁷

The % of BSA rejection was calculated as per the equation given in Section 2.4.8. The concentration of BSA was calculated using a UV-visible spectrophotometer at λ_max = 280 nm.

In order to get a complete idea of fouling, the total fouling ratio (R_t), reversible fouling ratio (R_r) and irreversible fouling ratio (R_ir) were calculated by the following equations:³⁸

where j_p is the permeation flux at a filtration time of 120 minutes when the feed was the BSA solution.

Furthermore, the best performing membrane (3% PVC–SA/PSf) was subjected to several cycles of filtration/rinsing as per the same procedure as that explained above, to find out the membrane life span and to validate the resistance towards fouling.

3. Results and discussion

Modification of PVC is vital to enhance the hydrophilicity and this can be done by incorporating hydrophilic groups into the polymer chain. This property improves the membrane performance in terms of flux and rejection, but may lead to severe loss of mechanical strength of the membrane which is not appropriate for its use. Hence, the blending of modified PVC with a much stronger polymer such as PSf regains its lost strength. The blend membrane will have the optimum properties like improved mechanical strength, surface charge and hydrophilicity for heavy metal ion rejection.

3.1. Characterization

3.1.1. ATR-IR. The modification of PVC with SA was substantiated by IR spectra. The spectrum showed the occurrence of some extra stretching peaks in the modified PVC–SA compound due to the presence of functional groups like N–H and –SO₃H. A broad new peak obtained between 3000 and 3600 cm⁻¹ is due to the stretching frequency of the O–H bond of the –SO₃H group and N–H stretching respectively. Furthermore, N–H bending was observed at 1393 cm⁻¹, and C–N stretching was observed at 1086 cm⁻¹.³⁹ A sharp peak observed at 1629 cm⁻¹ resembles the C=C stretching for the aromatic groups. All these peaks were absent in the PVC spectrum, and the peaks arising at 1240 and 1084 cm⁻¹ may be attributed to C–C and C–H deformation and stretching respectively as shown in the spectra (ESI S.2†). C–H stretching was observed at 2934 and 2930 cm⁻¹ in both the cases of PVC–SA and PVC respectively. Hence the incorporation of SA in the PVC chain is confirmed.

3.1.2. ¹H NMR. In the ¹H NMR spectra (ESI S.3†) it can be clearly observed that the peak at δ 6.4 and δ 7.2 ppm represents the aromatic proton peaks giving a doublet. A singlet observed at δ 5.1 ppm is due to the NH proton. The –CH and –CH₂ protons show a multiplet at δ 1.75 and δ 2.35 ppm due to the presence of a long alkane chain present in its vicinity. It has merged with the intense peaks of DMSO at δ 3.3 and δ 2.5 ppm⁴⁰ giving broadened peaks. The expected peak for O–H appears to have been merged with the peak at δ 2 ppm hence is not visible. The spectrum is in accordance with the peaks of the expected product.

The extent of modification as calculated from the equation given in the experimental section is shown in Table 2.

Table 2 Details of PVC modification

Weight ratio of PVC : SA	% modification (% M)	% modifying efficiency (% ME)
1 : 1	31.2	31.2

---

3.1.3. Differential scanning calorimetry. The thermal behavior of the PVC polymer and the modified PVC was studied from the DSC plots. The DSC plots (ESI S.6†) show their glass transition temperatures (T_g), where T_g1 and T_g2 represent the first glass transition temperature and the second glass transition temperature respectively. The PVC polymer showed T_g1 at 85.27 °C and T_g2 at 276.79 °C and this is in accordance with the T_g temperature reported for neat PVC.⁴¹ However PVC–SA showed T_g1 at 58.58 °C and T_g2 at 233.04 °C. The T_g values of PVC–SA are less compared to those of neat PVC. This is because during the chemical modification of the polymer, the polymer chain is shortened which shows a remarkable impact on the physical properties of the polymer. This indicates the better heat tolerance of PVC over PVC–SA. The different transition temperatures T_g1 and T_g2 are observable because of the change in their physical state into three different phases during the course of heat treatment. The state below the T_g1 point represents the glassy state of the polymers and the area between T_g1 and T_g2 represents the rubbery state. Once the temperature crosses the T_g2 temperature, the polymers get converted into a liquid state.

3.1.4. Ion-exchange capacity. Table 3 represents the ion-exchange capacity of the blend membranes of different compositions of PVC–SA in the PSf matrix. From the table it can be clearly seen that as the composition of PVC–SA increases from 3 to 10%, the ion-exchange capacity of the blend membrane also increases. This substantiates the presence of –SO₃H and –NH functional groups in the polymer matrix of the membrane.

Table 3 Ion-exchange capacity of the blend membranes^a

PSf : PVC–SA membrane	Ion-exchange capacity (mequiv. g^−1), IEC = AB/mdry
a An increase in the ion-exchange capacity of the membranes implies that the density of the functional groups is greater in the membranes containing a greater concentration of PVC–SA.
90 : 10	0.1673 ± 0.006
92 : 8	0.1348 ± 0.022
95 : 5	0.1146 ± 0.020
97 : 3	0.09421 ± 0.009

---

3.1.5. Water uptake capacity. The SO₃H group enhances the hydrophilicity of the prepared blend membranes.³⁶ From Fig. 2 it can be noticed that neat PSf shows the lowest water uptake capacity followed by the PVC membrane and this is because of the highly hydrophobic nature of the PSf and PVC membrane, which resists the water to get absorbed. This is also supported by the contact angle measurement values obtained as 76.7 and 78.8 for the PSf and PVC membranes respectively.

Fig. 2 Water uptake capacity and contact angles of the blend membranes.

The PVC membrane shows a slightly higher water uptake due to the macroporous nature of the membrane as can be seen from the SEM image (Fig. 4a). The water uptake capacity of the blend membrane is much higher and increases proportionately with the increase in concentration of functional groups to enhance the hydrophilicity of the membranes. The contact angle decreases from 73.6 to 64.8 while the water uptake increases from 22.5% to 82%, proportionately for the corresponding increase in the PVC–SA concentration in the blends. The sulfanilic group being hydrophilic facilitates absorption of water and hence an increase in functional groups proportionately enhances the affinity towards water. This is because the presence of hydrophilic groups makes it easy for the water to get absorbed in the membrane and the greater the number of functional groups present, the greater the affinity towards water will be and the greater the absorption will be.

3.1.6. AFM analysis. To investigate the surface topology of the membranes, AFM analysis was carried out and the hydrophilicity was correlated to the surface roughness of the membranes. Fig. 3 gives the 3D surface images of all the membranes with their corresponding R_a values.

Fig. 3 3D images of AFM: (a) neat PVC, (b) 3% PVC–SA/PSf, (c) 5% PVC–SA/PSf, (d) 8% PVC–SA/PSf, and (e) 10% PVC–SA/PSf membranes.

An increase in the content of PVC–SA increases the number of valleys (differentiated by light and dark areas) leading to an increased surface roughness from 6.9 nm to 48.4 nm. The trend observed is due to the increase in hydrophilic groups which enhances the rate of the phase inversion process, due to its mobility towards water surroundings. This observed phenomenon is in good agreement with the results obtained by Nazri et al.²⁹ and Liu et al.⁴²

3.1.7. SEM analysis. SEM analysis is another very essential tool to study the cross section morphology of the prepared membranes. It gives vital information about the microporous structure of the membranes, which is the prime source responsible for the selectivity and productivity of the membranes. Fig. 4 reveals the SEM images of the neat PSF, PVC and PVC–SA/PSf blend membranes.

Fig. 4 Cross section SEM images of (a) PSf, (b) PVC, (c) 3% PVC–SA/PSf, (d) 5% PVC–SA/PSf, (e) 8% PVC–SA/PSf, and (f) 10% PVC–SA/PSf membranes. The scale represents 100 μm in all the images.

As it can be seen, the neat PVC membrane has long and more finger-like structures without much differentiation in the layers, whereas the neat PSf membrane depicts three distinct layers. The upper dense layer (skin layer), middle layer (finger-like projections) and bottom layer (sponge-like with big pores) are responsible for the selectivity, productivity and strength.⁴³ In the PSf membrane, the bottom layer is very thick as compared to the other membranes. The addition of PVC–SA to PSf separates the three different structural layers with a substantial density of the skin layer being visible in comparison to neat PSf. An increase in the concentration of PVC–SA diminishes the appearance of sub-layers. The skin layer becomes less dense and they appear as a merge of the middle and bottom layers in 5% PVC–SA/PSf. The selective layer has completely merged with the middle layer to give one single layer with a broad finger-like structure in the 8% PVC–SA/PSf membrane. Whereas the big hollow finger-like projections (marked in red for better understanding) are randomly observed in the 10% PVC–SA/PSf membrane running from the top to the bottom of the membrane, where all three of the layers have merged to give one single layer. During the phase inversion process, the non-solvent penetrates at a faster rate and the solvent (NMP)–non solvent (water) interchange occurs very quickly to give big pores.^44,45 This observation is in good accordance with the roughness values obtained in AFM analysis. Similarly as the composition of PVC–SA is lowered the interchange of the solvent and non-solvent takes place at a slow rate giving it the dense pore network.

3.2. Performance study

3.2.1. Pure water flux. In order to find the rate of permeation of water, a flux study was carried out at different pressures from 100 to 500 kPa, and it is depicted in Fig. 5 that an increase in pressure increases the flux in all the cases due to the stress on the membrane pores at a higher pressure. An increase in the content of PVC–SA increases the flux favorably due to enhanced hydrophilicity. Nevertheless, the flux of the neat PVC is near to the flux of the 5% PVC–SA/PSf membrane and is much more than that of the 3% PVC–SA/PSf membrane, even though PVC is highly hydrophobic in nature. The macroporous nature of the PVC membrane with big long pores (as explained by the SEM image) gives it a reasonably good flux rate. PSf, being an ultrafiltration membrane and highly hydrophobic with a very low flux or zero flux,⁴⁶ adversely affects the flux of the blend. However the PSf membrane showed a thick bottom layer in SEM analysis. It may be also one of the reasons for the low flux rate. But beyond the addition of 5% PVC–SA to PSf, the flux enhances considerably due to the increase in hydrophilicity and bigger pore structure in the membrane, which lets the water pass through the membrane with ease.

Fig. 5 Pure water flux of NF 270, neat PVC and the blend membranes.

Whereas, if we compare the flux of the commercially available membrane NF 270 it is considerably high compared to that of the PVC–SA/PSf blend membranes which may be attributed to its hydrophilicity.

The flux graph is a model fit, and a trend line was drawn to study the hydraulic permeability. Table 4 depicts the L_p value of the membranes. From the table it is clear that L_p of the membrane decreases with a decrease in the concentration of functional groups present in the blend membrane. However, L_p of the neat PVC is near to that of the 5% PVC–SA/PSf membrane due to its high flux from its macro-size pores. Also L_p of the NF 270 membrane is the highest due to its pretty high flux compared to the other membranes.

Table 4 The hydraulic permeability coefficient (L_p) of the membranes

PSf : PVC–SA membrane	Hydraulic permeability coefficient (Lp) (m s^−1 Pa^−1)
Neat PVC	2.70 × 10^−12
90 : 10	12.44 × 10^−12
92 : 8	6.89 × 10^−12
95 : 5	3.12 × 10^−12
97 : 3	0.26 × 10^−12
NF 270	32.02 × 10^−12

---

3.2.2. Rejection of metal ions. To check the workability of the prepared membranes, they were applied in rejection experiments for the removal of toxic heavy metals like Pb(II), Cd(II) and Cr(VI). 10 ppm of Pb(NO₃)₂, Cd(NO₃)₂ and K₂Cr₂O₇ was batch filtered at various pressures (in neutral conditions). The rejection of all the metal ions showed interestingly a 90–95% rejection of all the metal ions at 100 kPa pressure (ESI S.7†). The behavior of the membranes for metal retention can be explained as follows: in neutral pH the functional group SO₃H will be in its ionized form as SO₃⁻ giving it a negative charge. The metal ions exist as M⁺ (M = Cd and Pb) in the free ionic state, leading to an attraction of the metal ions towards the membrane functional groups, either forming metal–ion complexation or ionic interactions (⁻O⋯Mⁿ⁺⋯O⁻).¹⁷ Whereas, Cr(VI) exists in two ionic forms, such as Cr₂O₇²⁻ and CrO₄²⁻ species at neutral pH. Hence, instead of attraction, repulsion takes place between the negatively charged Cr species and SO₃⁻ groups which leads to its rejection.⁴⁷ However in the case of Cd(II) and Pb(II), both the ions are positively charged leading to attraction. So, the interaction (repulsion and attraction) between the membrane surface and ions leads to the rejection of metal ions.

However, the 10% PVC–SA/PSf membrane showed the lowest rejection compared to all the other membranes even though it had the maximum density of functional groups in its vicinity. This can be because of the larger pore size (Fig. 4f) and flux of the membrane (Fig. 5). This indicates that even though the rejection depends on the charge, porosity plays a very important role. If the pores are large and more numerous, than the ions can easily pass through the pores of the membranes due to the applied pressure. Hence, there is no chance for charge interaction. Instead the 3% PVC–SA/PSf membrane results in maximum rejection. As explained in the SEM section, the hydrophilic component PVC–SA beyond a particular limit causes the loss of the selective skin layer. The same trend is observed in the 5% PVC–SA/PSf and 8% PVC–SA/PSf membranes. It also was confirmed by AFM images, where the surface roughness increased with the increasing concentration of PVC–SA. This is because of the hydrophilic functional group moving towards water molecules of the coagulation bath, which causes the greater membrane surface roughness. Furthermore the rejection performance was decreased with respect to the pressure increases from 100 to 500 kPa. The metal ions are forced out from the membrane, decreasing the rejection at high pressures. This suggests the importance of the morphology of the membranes in metal ion retention.

However, the metal ion rejection was almost similar because of the nearly similar size of 0.43 nm, 0.426 nm and 1.34 nm for Pb(II), Cd(II) and Cr(VI) metal ions respectively in their hydrated states.^17,48 The trend of rejection follows the order of the 10% < 8% < 5% < 3% PVC–SA/PSf membrane. In the 10% PVC–SA/PSf membrane, the rejection was much less due to the greater charge and porosity. Hence, it is necessary to attain an optimum size and charge of the membrane to obtain good results. The obtained result is consistent with the reported value of 99.44% of lead removal (at pH 5.7) by Gherasim⁴⁹ and 98% retention of iron(III) ions by Bernat et al.⁵⁰ (at pH 2).

It is observed in the figure (ESI S.7†) that the rejection pattern for all three of the metal ions by the NF 270 membrane shows a maximum of 78% rejection for Cd(II) ions followed by Pb(II) (76%) and Cr(VI) (60%). Also there is variability in the rejection pattern over the increasing pressure range which is due to the fouling of the membrane over time. However the selectivity of the commercial membrane towards metal ions is not acceptable. Hence it can be concluded that the prepared membranes gave a better rejection efficiency compared to the NF 270 membrane, however it lacked the flux of the NF 270 membrane. The lack of flux in the blend membrane is overcome by its enhanced rejection, thus making it compatible with the existing membranes.

3.2.3. Antifouling study. Fouling is one of the major drawbacks of the membrane as it decreases the life span and flux of the membrane. The adsorption of proteins on the membrane surface causes membrane fouling. Generally the adsorption takes place because of hydrophobic and electrostatic interaction between proteins and the membrane. In order to validate the antifouling capacity of the prepared membranes the rejection of a BSA protein solution was investigated at room temperature at 500 kPa pressure.^51,52 The water flux of the membranes was studied before and after the BSA rejection (ESI S.8†). The figure indicates drastic decreases in the flux during BSA filtration because of adsorption or deposition of protein on the membrane surface.

The flux recovery ratio (FRR) of the prepared membrane is shown in Fig. 6. 59.2% of the FRR and 74% of total fouling (Fig. 7) was observed in the PVC membrane. The modification of PVC increases the FRR up to 79.65% and decreases the total fouling up to 40%. Increasing the composition of PVC–SA in the membrane tends to increase the FRR and lower the fouling. This is because of the increasing hydrophilicity and membrane surface charge.

Fig. 6 Flux recovery ratio and BSA rejection of the membranes.

Fig. 7 Fouling parameters of the membranes.

In studying complete fouling, the reversible and irreversible fouling parameters are very important. In reversible fouling, the foulant can be removed by washing because these are loosely bonded on the membrane surface. Whereas, in irreversible fouling, the foulant cannot be removed because these are stuck inside the membrane pores. The total fouling, reversible and irreversible fouling ratios were calculated and are presented in Fig. 7. Reversible fouling decreases from the 3% to 10% PVC–SA/PSf membrane. However, irreversible fouling increases from the 3% to 10% PVC–SA/PSf membrane. It is the effect of the hydrophilic membrane surface and pore size. BSA is hydrophobic in nature, which loosely adheres on the hydrophilic membrane surface.⁵³ Hence, increasing the hydrophilicity of the membrane decreases the reversible fouling. The foulant can be easily removed from such membrane surfaces by washing. On the other hand, the pore size of the membrane goes on increasing with respect to the concentration of PVC–SA, which allows the BSA particles to go inside the membrane matrix and block the pores, which causes more irreversible fouling.

However during the fouling experiment, rejection of BSA was also observed. Fig. 6 presents the percent of BSA rejection from the membranes. The result obtained is similar to that of metal ion rejection with the highest of 98.5% of BSA being rejected by the 3% PVC–SA/PSf membrane. However, the rejection of BSA is greater in comparison to that of the metal ions due to the larger size of proteins, unlike the small sized metal ions.

3.2.4. Recyclability test. In order to estimate the reproducibility of the membrane performance, two cycles of water and BSA flux were performed in a similar fashion as mentioned above for the 3% PVC–SA/PSf membrane as it showed the highest rejection capability. It is observed that there was a slight decline of the water flux in the second cycle after the BSA rejection; however the FRR was not very much affected. It decreased from 69% to 62%, indicating a high efficiency of the membranes towards foulants. But the cycle was not able to extend further as the flux had increased drastically; this is due the rupturing of the pores during the cleaning process with warm water. Hence, it was clear that the incorporation of functional groups not only increased the hydrophilicity but also the efficiency of the blend membranes (Fig. 8).

Fig. 8 Reproducible characteristics of the 3% PVC–SA/PSf membrane during 2 cycles of BSA filtration.

3.2.5. Effect of interfering ions on rejection. The study of interfering ions on rejection is very essential to assess the efficiency of the membrane to reject a particular metal in the presence of other competing metal ions. Thus rejection was investigated with a mixture of all the ions, namely Pb(II), Cd(II) and Cr(VI), studied individually earlier. The efficient 3% PVC–SA/PSf membrane was selected as it showed the highest rejection for individual metal ions. All the remaining parameters were kept constant.

The effect of counter ions on the rejection of a specific ion can be observed (ESI S.9†). The mixture concentration in the feed is the sum of 10 ppm of each metal ion, which is overall 30 ppm concentration of the feed. One more intention of this study is to know the performance of the membranes in a highly concentrated feed sample. The results reveal that there is a considerable decrease in the metal ion rejection. The rejection of Cr(VI) is about 68%, and for Cd(II) and Pb(II) it is about 62%. It can be pointed out that the amount of Cr(VI) rejected is more compared to the other two ions present. This is accounted for by the difference in the rejection mechanism of Cr(VI) ions (repulsion), whereas attraction for Pb(II) and Cd(II) ions causes the difference. The rejection was decreased in mixed ion rejection up to 10% in Cr(VI), 18% in Cd(II) and 13% of Pb(II) as compared to single ion rejection. In a higher concentration of the feed solution, accumulation of ions near the surface causes saturation of functional groups and allows some of the ions to pass through the membrane. Also the mobility of the ions decreases with an increase in concentration leading to reduced overall rejection, which is observed by Jyothi et al.⁴⁶ However, a characteristic feature of charged membranes is less rejection at a high feed concentration and vice versa.⁵⁴ An unexpectedly high rejection was observed at high pressure. This is because a high concentration of ions can lead to the deposition of metal ions on the membrane surface, which causes membrane fouling with time. Meanwhile, a decrease in the water flux rate (ESI S.9†) as compared to pure water flux, is evidences for the membrane fouling. A similar observation was found for the NF 270 membrane, where the rejection of metal ions varies with pressure unlike that of the blend membrane, where the rejection remains constant throughout the experiment. Besides, the NF membrane shows a greater affinity towards Pb(II) ion rejection over the Cr(VI) ions distinct from the blend membrane. Also the flux of the membrane has reduced to half of its initial pure water flux, which is due to the large amount of fouling that takes place on the membrane surface and also leads to the constant amount of rejection over the complete set of experiment.

The present study demonstrates an improved membrane for the removal of heavy metal ions as compared to the existing NF 270 membrane but with a lack of flux; however the study opens a lot of scope for improvement in the area of productivity in the near future.

4. Conclusions

A blend of modified PVC with the PSf membrane matrix showed a high impact on heavy metal removal. The modification was carried out by a newly developed method using sulfanilic acid. ¹H NMR, ATR-IR spectroscopy and DSC confirmed the modification of PVC with sulfanilic acid. The blend membrane showed enhanced hydrophilicity and surface charge, which was confirmed by water uptake, water flux and IEC respectively. AFM and SEM analysis substantiated an improved surface roughness and porosity, due to the modified PVC. A good rejection of 90–95% was observed for the 3% PVC–SA/PSf membrane for Cd, Cr and Pb metal ions. The blend membranes show enhanced flux and antifouling property compared to the neat PVC and PSf membranes. The antifouling study demonstrated a good FRR value of 75.6% for the 10% PVC–SA/PSf membrane. The study gains significance in synthesizing an optimized and functionalized PVC–SA blend for good metal ion rejection and with a good antifouling property.

Acknowledgements

The authors sincerely acknowledge the Indian Ministry of Drinking Water and Sanitation (no. 11017/33/2012-WQ) for financial support.

References

1. A. Vijayalakshmi, D. L. Arockiasamy, A. Nagendran and D. Mohan, Sep. Purif. Technol., 2008, 62, 32–38.
2. G. Arthanareeswaran, P. Thanikaivelan, J. A. Raguime, M. Raajenthiren and D. Mohan, Sep. Purif. Technol., 2007, 55, 8–15.
3. M. Hua, S. Zhang, B. Pan, W. Zhang, L. Lv and Q. Zhang, J. Hazard. Mater., 2012, 211–212, 317–331.
4. M. H. Ajmal Khan and S. Parvez, J. Trace Elem. Med. Biol., 2015, 31, 53–60.
5. F. Tan, X. Liu, X. Quan, J. Chen, X. Li and H. Zhao, Anal. Methods, 2011, 3, 343–347.
6. C. H. Huang, T. H. Hsieh and W. Y. Chiu, Carbohydr. Polym., 2015, 116, 249–254.
7. Y. Mansourpanah, K. Alizadeh, S. S. Madaeni, A. Rahimpour and H. Soltani Afarani, Desalination, 2011, 271, 169–177.
8. F. Ferrero, C. Tonetti and M. Periolatto, Carbohydr. Polym., 2014, 110, 367–373.
9. A. Jayalakshmi, S. Rajesh, S. Senthilkumar and D. Mohan, Sep. Purif. Technol., 2012, 90, 120–132.
10. P. Daraei, S. S. Madaeni, N. Ghaemi, E. Salehi, M. A. Khadivi, R. Moradian and B. Astinchap, J. Membr. Sci., 2012, 415–416, 250–259.
11. R. Kumar, A. M. Isloor and A. F. Ismail, Desalination, 2014, 350, 102–108.
12. C. Hegde, A. M. Isloor, M. Padaki, P. Wanichapichart and Y. Liangdeng, Desalination, 2011, 265, 153–158.
13. Y. Zhang, J. Tian, H. Liang, J. Nan, Z. Chen and G. Li, J. Environ. Sci., 2011, 23, 529–536.
14. S. Liu, Z.-S. Li, C.-Y. Wang and A.-Y. Jiao, Sep. Purif. Technol., 2013, 116, 131–136.
15. P. Cañizares, A. Pérez, R. Camarillo and R. Mazarro, J. Membr. Sci., 2008, 320, 520–527.
16. Z. Yu, Z. Sui and C. Tai-Shung, Environ. Sci. Technol., 2015, 49, 10235–10242.
17. C. Mbareck, Q. T. Nguyen, O. T. Alaoui and D. Barillier, J. Hazard. Mater., 2009, 171, 93–101.
18. Y. Yang and P. Wang, Polymer, 2006, 47, 2683–2688.
19. H. Rabiee, M. H. D. A. Farahani and V. Vatanpour, J. Membr. Sci., 2014, 472, 185–193.
20. A. Gholami, A. R. Moghadassi, S. M. Hosseini, S. Shabani and F. Gholami, J. Ind. Eng. Chem., 2014, 20, 1517–1522.
21. S. Mei, C. Xiao, X. Hu and W. Shu, Desalination, 2011, 280, 378–383.
22. M. A. Tooma, T. S. Najim, Q. F. Alsalhy, T. Marino, A. Criscuoli, L. Giorno and A. Figoli, Desalination, 2015, 373, 58–70.
23. Q. Che, Z. Zhu, N. Chen and X. Zhai, Mater. Des., 2015, 87, 1047–1055.
24. Y. Cui, Z. K. Yao, K. Zheng, S. Y. Du, B. K. Zhu, L. P. Zhu and C. H. Du, J. Membr. Sci., 2015, 492, 187–196.
25. M. R. Bilad, L. Marbelia, C. Laine and I. F. J. Vankelecom, J. Membr. Sci., 2015, 493, 19–27.
26. E. Ahmadpour, A. A. Shamsabadi, R. M. Behbahani, M. Aghajani and A. Kargari, J. Nat. Gas Sci. Eng., 2014, 21, 518–523.
27. M. Herrero, P. Tiemblo, J. R. Labarta, C. Mijangos and H. Reinecke, Polymer, 2002, 43, 2631–2636.
28. X.-J. Huang, F. Dong, L. Chen and Y.-Q. Li, Monatshefte für Chemie - Chemical Monthly, 2008, 139, 1447–1451.
29. N. A. M. Nazri, W. J. Lau, M. Padaki and A. F. Ismail, J. Polym. Res., 2014, 21, 594.
30. M. S. Jyothi, M. Padaki, R. Geetha Balakrishna and R. K. Pai, J. Mater. Res., 2014, 29, 1537–1545.
31. C. Klaysom, B. P. Ladewig, G. Q. M. Lu and L. Wang, J. Membr. Sci., 2011, 368, 48–53.
32. Z. Wang, H. Ni, M. Zhang, C. Zhao and H. Na, Desalination, 2009, 242, 236–244.
33. V. Nayak, M. S. Jyothi, R. G. Balakrishna, M. Padaki and A. F. Ismail, ChemistryOpen, 2015, 4, 278–287.
34. M. Padaki, A. M. Isloor and P. Wanichapichart, Desalination, 2011, 279, 409–414.
35. M. K. Aroua, F. M. Zuki and N. M. Sulaiman, J. Hazard. Mater., 2007, 147, 752–758.
36. M. Padaki, A. M. Isloor, P. Wanichapichart and A. F. Ismail, Desalination, 2012, 298, 42–48.
37. S. Zinadini, A. A. Zinatizadeh, M. Rahimi, V. Vatanpour and H. Zangeneh, J. Membr. Sci., 2014, 453, 292–301.
38. M. Padaki, D. Emadzadeh, T. Masturra and A. F. Ismail, Desalination, 2015, 362, 141–150.
39. L. P. Donald, G. M. Lampman, G. S. Kriz and J. R. Vyvyan, Introduction to Spectroscopy, Brooks/Cole Cengage Learning, Boston, 2008.
40. M. Padaki, A. M. Isloor, R. Kumar, A. Fauzi Ismail and T. Matsuura, J. Membr. Sci., 2013, 428, 489–497.
41. K. Aouachria and N. Belhaneche-Bensemra, Polym. Test., 2006, 25, 1101–1108.
42. B. Liu, C. Chen, T. Li, J. Crittenden and Y. Chen, J. Membr. Sci., 2013, 445, 66–75.
43. M. Padaki, A. M. Isloor, A. F. Ismail and M. S. Abdullah, Desalination, 2012, 295, 35–42.
44. V. Vatanpour, S. S. Madaeni, R. Moradian, S. Zinadini and B. Astinchap, J. Membr. Sci., 2011, 375, 284–294.
45. S. Zinadini, A. A. Zinatizadeh, M. Rahimi, V. Vatanpour, H. Zangeneh and M. Beygzadeh, Desalination, 2014, 349, 145–154.
46. M. S. Jyothi, V. Nayak, M. Padaki, R. Geetha Balakrishna and A. F. Ismail, Desalination, 2014, 354, 189–199.
47. K. Y. Wang and T.-S. Chung, J. Membr. Sci., 2006, 281, 307–315.
48. S. Habibi, A. Nematollahzadeh and S. A. Mousavi, Chem. Eng. J., 2015, 267, 306–316.
49. C.-V. Gherasim and P. Mikulášek, Desalination, 2014, 343, 67–74.
50. X. Bernat, A. Pihlajamäki, A. Fortuny, C. Bengoa, F. Stüber, A. Fabregat, M. Nyström and J. Font, J. Membr. Sci., 2009, 334, 129–137.
51. R. Kumar, A. M. Isloor, A. F. Ismail and T. Matsuura, J. Membr. Sci., 2013, 440, 140–147.
52. R. Kumar, A. M. Isloor, A. F. Ismail, S. A. Rashidc and T. Matsuura, RSC Adv., 2013, 3, 7855–7861.
53. R. S. Hebbar, A. M. Isloor, K. Anandab and A. F. Ismail, J. Mater. Chem. A, 2016, 4, 764–774.
54. L. M. Ortega, R. Lebrun, J.-F. Blais and R. Hausler, Desalination, 2008, 227, 204–216.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra02590k

---