Characterization and antifouling properties of polyethylene glycol doped PAN–CAP blend membrane

Abstract

The effects of polyethylene glycol (PEG) as an additive to a cellulose acetate phthalate–polyacrylonitrile blend membrane in the ultrafiltration range were investigated. Influences of both the molecular weight and concentration of PEG were examined. Ternary phase diagrams were generated to identify the domain of composition where thermodynamic instability occurred. Kinetic hindrance due to the presence of the additive was also estimated. The relative importance of thermodynamic and kinetic factors was investigated quantitatively to interpret the nature of the membrane morphology. The prepared membranes were characterized in terms of surface morphology by scanning electron microscopy, water permeability, pore density, molecular weight cut off, contact angle and breaking stress. The antifouling characteristics of the prepared membranes were evaluated in terms of the filtration of bovine serum albumin protein. Membranes with PEG 200 at concentrations of 2 wt% and 6 wt% showed the best antifouling properties.

---

Mr Mrinmoy Mondal is a doctoral student in the Department of Chemical Engineering at the Indian Institute of Technology, Kharagpur. He has authored 1 paper and 3 patents.

Dr Sirshendu De is a professor in the Department of Chemical Engineering at the Indian Institute of Technology, Kharagpur. He has authored more than 200 papers, 7 books, 15 patents, 10 book chapters and 2 technology transfers. He is also the recipient of various awards including the Shanti Swarup Bhatnagar Prize in the Engineering Science category in 2011.

---

1. Introduction

Due to their low energy consumption and environmental friendliness, membrane based technologies are becoming attractive operation units in various applications including water purification, pharmaceuticals, and juice processing.^1–4 Non-solvent induced phase inversion is a well-known technique for preparing asymmetric polymeric membranes.^5–12 In this process, a thin, dense skin is formed over a porous sub-structure. The pore formation is a complicated phenomenon, involving interplay of various parameters such as polymer composition, coagulant temperature, type of solvent, and nature and concentration of additives. The phase-inversion process must be carefully controlled to attain the desired morphology and performance.^13–15

Fouling is the major problem in any membrane separation. Thus, the preparation of anti-fouling membranes is an area of active research. To improve the anti-fouling characteristics of a membrane surface, various methods are employed. These methods are blending of polymers,^10,16,17 addition of inorganic salts,^18,19 surface adsorption by suitable chemicals,^20,21 chemical grafting,^22,23 UV assisted grafting,^24,25 and plasma treatment.^26,27 The common purpose of all these processes is to make the membrane surface more hydrophilic and fouling resistant.

Polymer blending and use of additives are two of the cost-effective methods. Sivakumar et al. reported the characterization and application of cellulose acetate–polyurethane and cellulose acetate–polysulfone blend membranes.^28–31 Saljoughi et al.^13,32 investigated the effects of additives, such as polyvinyl pyrrolidone (PVP) and polyethylene glycol (PEG), on a cellulose acetate blend membrane. Rahimpour et al.¹⁴ used cellulose acetate phthalate (CAP) to improve the hydrophilicity of a polyethersulfone (PES) membrane. Reports are available on polysulfone–polyurethane³³ and polyvinylidene fluoride (PVDF)–polyethersulfone blend membranes.³⁴ These membranes have been used for water treatment and industrial applications. Similarly, preparation of antifouling blood- and bio-compatible membranes for medical applications has also been attempted. Nie et al.³⁵ developed a blood and cell compatible membrane having a heparin-like surface using a blend of polyethersulfone, sulfonated polyethersulfone and carboxylic polyethersulfone. Ma et al.³⁶ reported a new multifunctional polyethersulfone membrane coated by a dopamine grafted sulfonated linear heparin-like polymer showing remarkable blood and cell compatibility. A composite layer was deposited on a polyethersulfone membrane substrate by a layer-by-layer self-assembly of graphene-based supermolecules prepared by grafting poly(styrenesulfonate) and poly(acrylamide) onto graphene oxide through free radical polymerization, and this membrane was found to be highly biocompatible and bioactive.³⁷ Similar types of membranes with antibacterial properties were prepared by Wang et al.³⁸ A high-performance antifouling and antithrombotic hemocompatible blend membrane using polyethersulfone and polyurethane was also prepared by Yin et al.³⁹ Liu et al.⁴⁰ constructed self-cross-linked polymer nanolayers for the design of versatile biointerfaces.

However, the development of blend ultrafiltration membranes having antifouling properties for applications in industrial wastewater and surface water treatment is an area of active research. Studies on the blending of polyacrylonitrile (PAN) with other polymers are scarce. Some of these blends are PES–PAN and PVDF–PAN blend membranes.^41–43 Recently, Roy and De explored PAN–CAP blend membranes in dimethylacetamide, and investigated their utility in the extraction of steviol glycosides from stevia extracts.⁴⁴ The idea was to impart more hydrophilicity to the membrane surface by blending hydrophilic CAP with hydrophobic PAN. PAN is thermally stable, chemically resistant, commercially available at relatively low cost and easily soluble in various solvents.^45–49 Thus, it may be preferred over the popular polymer, polysulfone, as a suitable material for preparing blend membranes with desirable properties such as hydrophilicity without compromising the molecular weight cut off (MWCO) of the membrane. Moreover, a PAN–CAP blend is compatible, as discussed in this work.

The use of different molecular weights and concentrations of hydrophilic polymeric additives, such as PEG and PVP, to tailor-make the surface morphology and surface properties are common.^16,50–52 In the present work, the effects of molecular weight and concentration of PEG on the PAN–CAP blend membrane have been investigated. The aim is to further enhance the hydrophilicity of the resultant membrane by incorporating hydrophilic PEG. The cast membranes are characterized in terms of permeability, MWCO, hydrophilicity and surface morphology. The anti-fouling characteristics of each membrane have been examined using bovine serum albumin (BSA) solution.

2. Experimental

2.1 Materials

CAP was purchased from M/s G.M. Chemie Pvt. Ltd, Mumbai, India. PAN (homopolymer average molecular weight of 150 kDa) was procured from M/s Technorbital Advanced Materials Pvt. Ltd., Kanpur, India. N,N-Dimethylformamide (DMF) was purchased from M/s Merck (India) Ltd., Mumbai, India. PEG reagents with average molecular weight of 200 Da, 400 Da, 1.5 kDa, 4 kDa, 6 kDa, 10 kDa, 20 kDa and 35 kDa were supplied by M/s S. R. Ltd., Mumbai, India and dextran (average molecular weight: 70 kDa) was procured from M/s Sigma Chemicals, USA. These neutral solutes were used to evaluate the MWCO of the cast membranes. PEG was also used as additives during casting of the membranes. BSA was procured from M/s S. R. Ltd., Mumbai, India (molecular weight: 67 000 Da). Distilled water was used as the non-solvent in the coagulation bath. The membranes were cast on a non-woven polyester fabric with a thickness of 118 ± 22.8 μm (product number TNW006013) supplied by M/s Hollytex Inc., New York, USA.

2.2 Selection of composition of PAN–CAP blend membrane

The total polymer concentration was fixed at 19 wt% in DMF. Four compositions of PAN and CAP, i.e., 19 wt% PAN; 15 wt% PAN and 4 wt% CAP; 4 wt% PAN and 15 wt% CAP; and 19 wt% CAP, were studied. The blend composition, 4 wt% PAN and 15 wt% CAP, was found to have the maximum hydrophilicity along with the desirable MWCO (the results are elaborated in Section 3.1). To further enhance the hydrophilicity of the blend membrane without compromising the MWCO, effects of molecular weight and concentration of PEG additive on this blend composition were investigated in this study. The molecular weight of PEG was varied as 200, 400, 1500, 4000 and 6000 Da at 1 wt%. The concentration of PEG with a molecular weight of 200 Da was varied as 1, 2, 4, 6, 8 and 10 wt% to observe the effect of concentration.

2.3 Membrane preparation

Flat sheet blend membranes were prepared by phase inversion method and hand casting. For PAN–CAP–PEG membrane, a fixed amount of PEG with different molecular weights and concentrations were added to a premixed solution of 15 wt% CAP and 4 wt% PAN in DMF and dissolved at 60 °C. The solution was prepared with constant mechanical stirring for 6 hours at 60 °C to mix it completely, and it was then cast on a non-woven polyester fabric (attached to a glass plate) with the help of a stainless steel casting knife, set at a fixed gap of 150 μm and a manual drawdown speed of 30 mm s⁻¹. The casting solution with the glass plate was immediately put into a gelation bath (to minimize the evaporative effect) containing distilled water for phase inversion at room temperature. The membrane was kept in the bath for 24 hours to complete the phase inversion. The abovementioned procedure was followed for casting the PAN–CAP membranes with the compositions presented in the earlier section.

2.4 Ternary phase diagram

The ternary phase diagram was constructed from the cloud point data obtained by the usual titration method.⁵³ Phase diagrams were generated to observe the effect of PEG molecular weight (at a fixed concentration) and concentration (at a fixed molecular weight). For the former case, three compositions of casting solution were considered, i.e., no PEG, PEG-1500 Da and PEG-6000 Da, both at 1 wt%. To observe the effect of PEG concentration, two concentrations of PEG-200, i.e., 6 and 10 wt%, were selected. The polymer solution was stirred for 6 h. Distilled water was added dropwise to the polymer solution with stirring using a 100 μL micro-pipette to perform the titration. At the first sign of turbidity, the addition of distilled water was stopped, and the cloudy solution was stirred for an additional 30 min to see whether the turbid solution becomes clear. If the suspension remained turbid, the composition was recorded as the cloud point. The composition at the cloud point was determined from the amount of water, DMF and polymer present in the solution.

2.4.1 Thermodynamic and kinetic instabilities of casting solution. Miscibility gap (MG) and degree of shift in binodal curve (DSBC) determine the thermodynamic properties of the casting solution.⁵³ The thermodynamic instability of a polymeric solution with an additive is represented by an enhancement parameter, T. Large value of T indicates thermodynamic instability. The detailed calculation involved for the estimation of T for the various compositions of casting solution is presented in Appendix A.

In addition to thermodynamic instability, solvent–nonsolvent kinetics during phase inversion is an important factor that controls the final pore size distribution in the membrane. It is quantified by the kinetic hindrance parameter, K. A decrease in K indicates a denser membrane matrix. K can be quantified by measuring the concentration of solvent (DMF) in the gelation bath as a function of time. The detailed calculations involved in the estimation of K are presented in Appendix A.

2.5 Characterization of the membranes

2.5.1 Membrane permeability. The prepared membranes were compacted in an unstirred batch cell at 690 kPa for 3 hours. The batch cell (effective filtration area 33 cm²) was filled with 500 ml of distilled water, and then steady state permeate flux was measured at five different trans-membrane pressure drops. The permeate flux was calculated by

J_w = Q/A₀ΔT₀ (1)

where J_w is pure water flux; Q is volume of permeate collected in ΔT₀ and A₀ is effective membrane surface area. A plot of J_w against trans-membrane pressure drop resulted in a straight line passing through the origin, the slope of which gave the membrane permeability.

2.5.2 Molecular weight cut-off (MWCO) of the membrane. The MWCO of each of the membranes was calculated by measuring the rejection of neutral solutes of different molecular weights (4, 6, 10, 20, 35, 70, 100 and 200 kDa). A solution of 10 kg m⁻³ was prepared in distilled water and fed to the batch cell with continuous stirring. The experiments were conducted at 70 kPa transmembrane pressure drop at 2000 rpm. The permeate was analyzed and the percentage rejection (%R) was measured as follows:

R = (1 − C_P/C_F) × 100% (2)

where C_P is the concentration of permeate and C_F is the concentration of feed. Rejection values were plotted against the molecular weights of the solutes in a semi-logarithmic curve. The molecular weight corresponding to 90% rejection was estimated as MWCO.

2.5.3 Contact angle. Contact angle of the membrane was measured by a Goniometer, supplied by M/s Rame Hart, New Jersey, USA, using the sessile drop method.⁵⁴ The contact angle at six different locations were measured and the average value was reported.

2.5.4 Pore density. Pore density was calculated using the classical Hagen–Poiseuille equation. It was assumed that the membrane had straight cylindrical pores. Thus, the volumetric flow rate is given byr_avg was calculated using the following equation:⁵⁵

r_avg (cm) = 16.73 × 10⁻¹⁰ (MWCO)^0.557 (4)

where ΔP is transmembrane pressure drop; r_avg is average pore radius; μ is water viscosity; and L is the thickness of the membrane (i.e., total thickness–thickness of the fabric). If n_c = number of pores/membrane surface area, the total flux (J_w) is

J_w = Q × n_c (5)

Using eqn (3) and (5), the following expression of pore density is obtained:

2.5.5 Pore volume distribution and average pore radius by BET analysis. The pore volume distribution of the prepared membrane samples was measured by Brunauer–Emmett–Teller (BET) analysis. The BET instrument was supplied by Quantachrome instruments, Florida, USA (model: AUTOSORB-1).

2.5.6 Surface morphology. The surface morphology of the membranes was studied with a scanning electron microscope (SEM) (model: ESM-5800, JEOL, Japan).

2.5.7 Atomic force microscopy (AFM). An atomic force microscope (model: 5500 AFM, Agilent Technologies, USA) was used to measure the surface roughness of the membranes under tapping mode.

2.5.8 Mechanical strength. The mechanical strength of the membranes, in terms of breaking stress, was studied by a universal electronic strength measuring instrument procured from M/s Tinius Olsen Ltd., Redhill, England of model H50KS. All the measurements were carried out at room temperature and at a strain rate of 20 mm min⁻¹. The average of three values of each sample was reported.

2.5.9 Differential scanning calorimetry. The thermal analyses of pure and blend membranes were carried out in a differential scanning calorimeter (M/s TA instrument Ltd., New castle, Delaware USA; model DSC Q₂₀). The experiment was carried out in a two-step heating–cooling cycle in nitrogen atmosphere. The analysis was carried out utilizing 5 mg of the membrane sample in an aluminium pan and heating it from 0 to 300 °C at a heating rate of 10 °C min⁻¹.

2.5.10 Antifouling experiment using BSA. BSA solution of 500 mg l⁻¹ at pH 7.0 and transmembrane pressure drop of 138 kPa was used to conduct antifouling experiments.

2.5.11 Flux recovery ratio (FRR) and flux decline ratio (FDR) of membrane. The membrane gets fouled due to the deposition of solutes after each experiment. The antifouling characteristics of the membrane were quantified with the help of two parameters, i.e., flux decline ratio (FDR) and flux recovery ratio (FRR). FRR is related to the irreversible membrane fouling and was calculated by measuring the pure water flux of all the membranes before and after BSA experiments. FDR quantifies the flux decline during a particular experiment with protein solution. FRR is defined aswhere J_w2 is the water flux of the membrane after the experiment and J_w1 is the water flux of the membrane before the experiment. FDR is defined aswhere, J₁ is the initial flux of the membrane with BSA and J_s is the flux at the end of one hour.

2.6 Analysis

The rejection of neutral solutes for MWCO analysis was measured by a digital refractometer (supplied by M/s, Cole-Parmer, Kolkata, India). The concentration of BSA before and after the experiments was measured with a UV-visible spectrophotometer (supplied by M/s Perkin-Elmer, Connecticut, USA, model: Lambda 35) at a wavelength of 280 nm.

3. Results and discussions

3.1 Selection of composition of PAN–CAP blend membrane

The variation of contact angle and MWCO of PAN–CAP is presented in Table 1. It is observed that the hydrophilicity of the membrane decreases as the CAP concentration increases in the blend, thus reducing the contact angle. For example, PAN 19 wt% has a contact angle of 86°, which decreases to 72° for 15 wt% CAP blended in PAN. The MWCO of the membrane increases with CAP concentration. For example, MWCO increases from 12 kDa to 85 kDa with increase in CAP concentration from 0 to 19 wt%. Membrane with CAP 19 wt% shows the highest hydrophilicity (contact angle of 65°), but the MWCO (85 kDa) is higher than that of PAN (4 wt%)–CAP (15 wt%) blend membrane (38 kDa). This particular PAN–CAP blend results in the highest hydrophilicity and desired MWCO among the blends. Therefore, the membrane of this composition was selected to study the effects of PEG additives to further enhance the hydrophilicity.

Table 1 Contact angle and MWCO of the membrane for various compositions of PAN and CAP

Synthesized membrane	Contact angle (°)	MWCO (Da)
PAN 19 wt%	86 ± 2	12 000 ± 1100
PAN 15 wt%, CAP 4 wt%	78 ± 3	21 000 ± 800
PAN 4 wt%, CAP 15 wt%	72 ± 2	38 000 ± 700
CAP 19 wt%	65 ± 3	85 000 ± 900

---

3.2 Ternary phase diagram, thermodynamic instability and kinetic hindrance

The phase diagram of water/PAN–CAP/DMF/PEG is shown in Fig. 1. On the introduction of PEG into the casting solution, the curve moves closer to the polymer/solvent axis (Fig. 1(a)). Therefore, less water is needed for the precipitation of the polymer compared to water/DMF/PAN–CAP. This indicates that the additive lowers the thermodynamic stability of the casting solution. The addition of a higher molecular weight PEG (6 kDa) shifts the binodal curve further towards the left, indicating lesser thermodynamic stability and lower water requirement for the precipitation of the polymer. Similarly, the effects of the concentration of PEG are clear from Fig. 1(b). Thermodynamic instability of the solution increases with the concentration of PEG-200.

Fig. 1 Ternary phase diagram of Water/PAN–CAP–PEG/DMF system with the variation of (a) PEG molecular weight and (b) concentration of PEG-200.

Thermodynamic stability of the casting solution is more quantitatively interpreted from the thermodynamic enhancement parameter, T, as elaborated in the Appendix. The values of T for various compositions of casting solution are presented in Table 2. At a fixed concentration (1 wt%), T increases with the molecular weight of PEG. Its value is 13 for PEG-1500 and 56.3 for PEG-6000, indicating the enhanced thermodynamic instability of the casting solution, resulting in the possibility of the formation of a porous membrane. The effects of the concentration of PEG are also apparent from the T values presented in Table 2.

Table 2 Thermodynamic properties of casting solution

Membrane	MG	DSBC (%)	δDMF/additive	χDMF/additive	T
PEG-200 0 wt%	2.7	0	—	—	—
PEG-200 6 wt%	2.4	11.0	3.55	0.39	4.4
PEG-200 10 wt%	2.1	22.0	3.55	0.39	8.7
PEG-1500 1 wt%	2.4	11.0	6.11	1.16	13.0
PEG-6000 1 wt%	1.7	37.0	6.99	1.52	56.3

---

T increases from 0 (for 0 wt% PEG-200) to 8.7 (for 10 wt% PEG-200). However, membrane morphology is not only dependent on thermodynamic considerations, but also on the kinetics of solvent-nonsolvent demixing. The kinetics of solvent–nonsolvent demixing is quantified by kinetic parameter K, as explained in Appendix A.

Values of K for the various compositions of casting solutions are presented in Table 3. The enhancement of ‘K’ indicates a quick demixing, leading to the formation of a porous membrane, and lower value of K results in delayed demixing, and thus a denser membrane. As observed from Table 3, K increases from 5.9 to 8.0 with the addition of 6 wt% PEG-200 in the casting solution. Actually, the relative magnitude of T and K for the different casting solutions dictates the interplay of thermodynamic and kinetic effects. The addition of 6 wt% PEG-200 to the casting solution results in increase in both T and K, indicating that both effects work in tandem, leading to the formation of a porous membrane. On the other hand, in the case of 10 wt% PEG-200, kinetic parameter K decreases to 6.2 × 10⁻⁹ and thermodynamic parameter T increases to 8.7 compared to 6 wt% PEG-200 (4.4). As discussed by Sadrzadeh and Bhattacharjee,⁵³ an alteration in the kinetic parameter has a more prominent effect than thermodynamic instability. Therefore, for 10 wt% PEG-200, kinetic hindrance is more dominant than thermodynamic instability, and membrane with 10 wt% PEG-200 is expected to have a denser morphology. In the case of PEG with higher molecular weights (1500 and 6000 Da), thermodynamic parameter, T, increases significantly (13 and 56.3, respectively, compared to 0 for no PEG) compared to marginal increases in the kinetic parameter (7.9 × 10⁻⁹ and 9.6 × 10⁻⁹, respectively, from 5.9 × 10⁻⁹). Moreover, both the effects are synergistic in nature. Thus, for higher molecular weight of PEG, the membrane structure is expected to be more porous.

Table 3 Kinetics properties of casting solution

Synthesized membrane	m (min^1/2)	t0 (min)	Viscosity (μ) (Pa s)		Correlation coefficient	K (× 10^−9)
PEG-200 0 wt%	0.046	0.19	19.5	1.39	0.998	5.9
PEG-200 6 wt%	0.055	0.28	28.3	1.67	0.972	8.0
PEG-200 10 wt%	0.050	0.34	30.9	1.55	0.989	6.2
PEG-1500 1 wt%	0.060	0.42	31.1	1.85	0.990	7.9
PEG-6000 1 wt%	0.063	0.51	32.5	1.93	0.983	9.6

---

3.3 SEM

Effects of molecular weight of PEG (at 1 wt%) on the blend membrane were studied using 200, 400, 1500, 4000 and 6000 Da PEG. The SEM images of the cross section of the resultant membranes are presented in Fig. 2(a). As observed from Fig. 2(a), there is a thin and dense skin followed by a porous substructure. This is a common feature of phase inversion membranes. A general observation from this figure is that teardrop-like pores start appearing under the skin layer with increase in the molecular weight of PEG in the blend. The size and length of the pores increase with PEG molecular weight. The skin layer is clearly visible in Fig. 2(a)(i)–(iii). From Fig. 2(a)(i) and (iii), skin layer thickness can be estimated to be less than 3 μm. The width of macropores is 20 to 40 μm for 200 Da PEG; 20 to 40 μm for 400 Da PEG; 20 to 45 μm for 1500 Da PEG; and 30 to 50 μm for 4000 and 6000 Da PEG. The lengths of the macropores increase with the molecular weight of PEG. For 6000 Da PEG, the macropores cover almost the full cross section of the membranes. PEG is a well-known pore former.⁵² Being hydrophilic, it leaches out during phase inversion to the anti-solvent (water), leaving behind macropores. With increase in molecular weight of PEG, the larger sized PEG molecules create bigger pores, and consequently the membrane becomes more porous. These observations are in corroboration with the thermodynamic and kinetic parameters discussed in Section 3.2, indicating that porous membranes are formed as the molecular weight of PEG increases in the blend.

Fig. 2 (a) Cross section of membrane with various molecular weights of PEG (1 wt%) doped in blend: (i) 200 Da, (ii) 400 Da, (iii) 1500 Da, (iv) 4000 Da, and (v) 6000 Da. (b) Cross section of membrane with various weight % of PEG-200 doped in blend: (i) 0, (ii) 1, (iii) 2, (iv) 4, (v) 6, (vi) 8, and (vii) 10.

Because PEG-200 resulted in a less porous structure or denser membrane,⁵⁶ it was selected to study the effects of additive concentration. The concentration of PEG-200 was varied from 1 to 10 wt%. The SEM images of the cross section are presented in Fig. 2(b). As observed from Fig. 2(b), pore size increases with PEG concentration up to 6 wt%, and then the morphology becomes denser till 10 wt% (as discussed earlier). For the PAN–CAP blend membrane without PEG, the size of macropores is less than 25 μm (Fig. 2(b)(i)). For 1 wt% PEG-200, teardrop-like macropores are visible in the cross section with a pore width in the range of 10 to 40 μm. The size of pores increases when 2 wt% PEG-200 is added, and the pore width is in the range of 25–40 μm. By increasing PEG concentration to 4 and 6 wt%, wider (10 to 50 μm) and longer macropores are formed. For 8 and 10 wt% of PEG, the morphology of the membrane matrix changes from the porous to a denser structure (as discussed earlier).

It is observed in Fig. 2(b)(vii) that no regular contour of macropores is visible and a dense structure is formed. This phenomenon can be explained on the basis of the dominance of phase inversion kinetics as discussed earlier. The size and number of pores formed during phase inversion is determined by the interdiffusion of water (non-solvent) and solvent from the membrane matrix into the gelation bath. It is a proven fact that PEG imparts hydrophilicity to a membrane matrix, and in the gelation bath, the hydrophilicity induces water to move freely into the matrix and the solvent to move out of it, leading to bigger sized pores. This observation is in corroboration with the effect of variation of thermodynamic and kinetic parameters with the composition of casting solution, as discussed in an earlier section. Beyond 8 wt% PEG, the kinetic hindrance parameter becomes dominant, resulting in a delayed demixing, thus leading to the formation of a denser membrane matrix.

3.4 Permeability and pore density

The variation of membrane permeability and pore density as a function of molecular weight of PEG is shown in Fig. 3(a).

Fig. 3 (a) Pore density and permeability with variation of the molecular weight of PEG. (b) Pore density and permeability with the variation of concentration of PEG-200.

As observed from the SEM images presented in Fig. 2(a), the membrane becomes more porous with increasing molecular weight of PEG, leading to an increase in the membrane permeability. The permeability of the membrane increases from 6 × 10⁻¹¹ m Pa⁻¹ s⁻¹ to 26 × 10⁻¹¹ m Pa⁻¹ s⁻¹ as the molecular weight of PEG increases from 200 to 6000 Da. However, a reverse trend is observed in the case of pore density. Pore density decreases from 2 × 10¹⁶ m⁻² to 7 × 10¹⁵ m⁻² as the molecular weight of PEG increases from 200 to 6000 Da. As observed from SEM images, for cross-section displayed in Fig. 2(a), the pore size increases, but the number of pores decreases with PEG molecular weight, thus leading to a decrease in pore density. Therefore, although pore density decreases, membrane permeability increases. The formation of bigger and longer macropores, as discussed earlier, is responsible for this behaviour.

Fig. 3(b) demonstrates the variation of membrane permeability and pore density with concentration of PEG-200.

As observed from SEM images (Fig. 2(b)), the membrane pore size increases up to 6 wt% and then decreases. The same trend is reflected in membrane permeability. Permeability increases from 5.5 × 10⁻¹¹ to 9 × 10⁻¹¹ m Pa⁻¹ s⁻¹ as PEG concentration increases from 0 to 6 wt%, and it decreases to 2.2 × 10⁻¹¹ m Pa⁻¹ s⁻¹ for 10 wt% PEG. However, pore density decreases up to 8 wt% and increases at 10 wt%. Because the membrane becomes denser at 10 wt% compared to that at 8 wt%, the number of pores per square meter of membrane increases at 10 wt%. However, increase in pore density and decrease in pore size balance occurs in such a way that the permeability of 8 wt% and 10 wt% PEG-200 remains almost invariant. Results of the variation of permeability with membrane composition confirm the observations presented in Sections 3.2 and 3.3.

3.5 Molecular weight cut-off (MWCO) and contact angle

Variations of MWCO and contact angle for various membranes are shown in Fig. 4. The effects of the molecular weight of PEG are presented in Fig. 4(a), and those of their concentration are shown in Fig. 4(b). The observations from these figures are in direct alignment with the effect of PEG molecular weight, as shown in the SEM image (Fig. 2(a)). The MWCO of the membranes increases from 38 to 106 kDa as the molecular weight of PEG increases from 200 to 6000 Da. Because hydrophilicity increases with the molecular weight of PEG, the contact angle decreases. The corresponding change is from 67° to 47°.

Fig. 4 (a) MWCO and contact angle with the variation of the molecular weight of PEG. (b) MWCO and contact Angle with variation of concentration of PEG-200.

The effects of PEG concentration are evident in Fig. 4(b). As discussed earlier, the membrane becomes more porous up to 6 wt% of PEG, thus increasing MWCO from 38 to 56 kDa. Membranes become denser beyond this concentration, and hence the MWCO decreases to 39 kDa. The trend of contact angle is in line with the variation of membrane permeability. As the permeability increases up to 6 wt% PEG, the membrane becomes more hydrophilic and its contact angle decreases from 72° to 55°. Beyond 6 wt%, the permeability decreases, inducing more hydrophobicity on the membrane surface. Contact angle increases to 65° when PEG concentration increases to 10 wt%.

3.6 Pore size distribution

As described earlier, the cumulative pore volume (cm³ g⁻¹) distributions of the membranes are measured using a BET surface area analyzer. Four typical pore size distributions corresponding to 0 wt%, 6 wt%, 10 wt% of PEG-200 and 1% PEG-6000 are presented in Fig. 5.

Fig. 5 Variation of pore volume distribution with PEG-200 concentration and PEG molecular weight.

It is clear from Fig. 5 that the cumulative pore volume for 6 wt% of PEG-200 is more than that of no PEG (0 wt%). This indicates that the number of pores and permeability of membrane with 6 wt% PEG are more than that with no PEG. This measurement is in corroboration with kinetic hindrance and thermodynamic parameters in Section 3.2, SEM images in Fig. 2(b) and permeability values (Fig. 3(b)) of these two membranes. On the other hand, with 10 wt% of PEG-200, the cumulative pore volume is less than that of 6 wt%. This clearly indicates that the membrane with 10 wt% PEG is denser than that with 6 wt%. Thus, this observation again confirms the trends shown by SEM images (Fig. 2(b)) and permeability values (Fig. 3(b)).

As presented in Table 4, the average pore size of the membrane with 6 wt% PEG is 89 Å, which is higher than that of 10 wt% PEG, which is 80 Å. The membrane with no PEG has an average pore size of about 79 Å. This shows that the average pore size of 10 wt% PEG-200 and that of no PEG is almost the same. However, the membrane with no PEG has a higher permeability (5.5 × 10⁻¹¹ m Pa⁻¹ s⁻¹) compared to 10 wt% PEG (2.2 × 10⁻¹¹ m Pa⁻¹ s⁻¹). This is due to the fact that the pore density of the membrane without PEG is considerably higher (1.7 × 10¹⁶ m⁻²) compared to that of 10 wt% PEG membrane (6.4 × 10¹⁵ m⁻²). Interestingly, the MWCO values of these two membrane are considerably close, i.e., about 38 kDa. This trend is in accordance with the cumulative pore size distribution of these two membrane measured by BET, as shown in Fig. 5, and also the average pore size presented in Table 4. Thus, the MWCO and permeability of the membrane are not always one-to-one. The cumulative pore volume for the membrane with 1 wt% PEG-6000 is also shown in Fig. 6. It shows that as observed in the SEM images, this membrane is more porous and its average pore size is 105 Å, which is in agreement with the permeability (26 × 10⁻¹¹ m Pa⁻¹ s⁻¹) and MWCO of this membrane (105 kDa).

Table 4 Comparison of pore radius from eqn (4) and BET analysis

Synthesized membrane	Pore radius from MWCO correlation (Å)	Average pore radius from BET surface area analysis (Å)
PEG-200 0 wt%	59 ± 1.5	79 ± 4
PEG-200 1 wt%	60 ± 3	—
PEG-200 2 wt%	61 ± 4	—
PEG-200 4 wt%	72 ± 2.5	—
PEG-200 6 wt%	74 ± 2.6	89 ± 3
PEG-200 8 wt%	68 ± 1.7	—
PEG-200 10 wt%	60 ± 4	80 ± 4
PEG-400 1 wt%	68 ± 3.5	—
PEG-1500 1 wt%	73 ± 2.3	—
PEG-4000 1 wt%	79 ± 2.8	—
PEG-6000 1 wt%	105 ± 1.8	92 ± 3

---

Fig. 6 Atomic force microscopy (AFM) images of membranes with (a) PEG-200 0 wt%, (b) PEG-200 6 wt%, (c) PEG-200 10 wt%, and (d) PEG-6000 1 wt%.

3.7 Atomic force microscopy (AFM)

The three-dimensional AFM images of the typical four membranes corresponding to 0, 6 and 10 wt% of PEG-200 and 1 wt% PEG-6000 are presented in Fig. 6. It is observed in this figure (Fig. 6a and b) that a bumpy structure appears on the top surface as the concentration of PEG increases. PEG, a well-known hydrophilic substance, accumulates more on the top surface during phase inversion. With increase in concentration on PEG, more PEG molecules compete for their place in the upper layer of the membrane. Thus, average surface roughness increases from 70 to 133 nm as the concentration of PEG increases from 0 to 6 wt%. This trend also confirms the observation made by Sadeghi et al.⁵⁷ that surface roughness increases for more porous membranes. As observed in Fig. 3(b), the permeability of the 6 wt% PEG membrane (9 × 10⁻¹¹ m Pa⁻¹ s⁻¹) is more than that without PEG (5.5 × 10⁻¹¹ m Pa⁻¹ s⁻¹). However, for 10 wt% PEG (Fig. 6(c)), the average roughness decreases to 94 nm. Two opposing phenomena take place in this case. First, more PEG molecules would come to the surface because the concentration of PEG is high, which would enhance surface roughness. Second, at 10 wt% PEG, water influx to the membrane matrix is highly impeded (as discussed earlier) and a dense morphology sets in. Denser membranes have lower surface roughness.⁵⁸ The second effect becomes dominant at 10 wt% PEG concentration, and the surface roughness becomes less (94 nm). For PEG 6000, bigger PEG molecules accumulate near the surface, making it quite rough with an average roughness of 188 nm.

3.8 Tensile strength

The variation of the breaking stress with the molecular weight of PEG and concentration is presented in Fig. 7. Because the membrane becomes more porous with the increase in the molecular weight of PEG, the breaking stress is also reduced. For example, breaking stress decreases from 18 to 16 MPa as the molecular weight of PEG increases from 200 to 6000 Da. Similarly, the membrane becomes more porous up to 6 wt% of PEG, thus the breaking stress decreases up to that concentration. Beyond 6 wt%, the membrane becomes dense, improving the breaking stress.

Fig. 7 Tensile strength with the variation of (a) molecular weight of PEG and (b) concentration of PEG-200.

3.9 Differential scanning calorimetry

The DSC thermogram of the PAN–CAP blend membrane and those with different PEG additives are presented in Fig. 8. It is observed from this figure that the PAN–CAP blend membrane and membranes with PEG show a single glass transition temperature (T_g), indicating that the polymers are compatible. The possible interaction between the nitrile group of PAN and hydroxyl group of CAP is schematically presented in Fig. 9. The T_g value is the highest at 89 °C for the PAN–CAP blend membrane, and it is the lowest for 1 wt% PEG-6000 (82 °C). It is known that lower T_g of membrane indicates a more porous structure,^59,60 and this observation is in corroboration with the observations made in the earlier sections.

Fig. 8 Differential scanning calorimetry (DSC) curves for (a) PEG-200 0 wt%, (b) PEG-200 10 wt%, and (c) PEG-6000 1 wt% membrane.

Fig. 9 Possible interaction of PAN with CAP.

3.10 Permeate flux decline for BSA solution

Profiles of permeate flux for BSA solution (0.5 kg m⁻³) for the various membranes are shown in Fig. 10. It is observed from Fig. 10(a) that flux values at a particular time of filtration are higher for 200, 400, 4000 and 6000 Da PEG, in that order. This is in corroboration with the permeability values of these membranes. For a given membrane, the permeate flux declines over the filtration period due to concentration polarization.⁶¹ The flux decline trends for the different concentrations of PEG are shown in Fig. 10(b). The flux profiles are again in the same order as the permeability values of these membranes. At a fixed time, the permeate flux increases from 0 to 6 wt% PEG, and it decreases for 8 and 10 wt%.

Fig. 10 Permeate flux variation of BSA solution with different (a) molecular weights of PEG and (b) concentrations of PEG-200.

3.11 Rejection and permeate flux of BSA solution

The rejection of BSA by the various membranes and the permeate flux at the end of 1 hour are shown in Fig. 11.

Fig. 11 Permeate flux after 1 h and BSA rejection with variation of (a) molecular weight of PEG and (b) concentration of PEG-200.

The flux values follow the expected trend, as discussed earlier. BSA rejection can be explained according to the variation of MWCO of these membranes. Because the MWCO of the membranes increases with the molecular weight of PEG (Fig. 4(a)), the rejection of BSA also decreases in the same order. Thus, BSA rejection decreases from 95% to 12% as the molecular weight of PEG increases from 200 to 6000 due to increase in the pore size of the membrane (refer Fig. 10(a)). Similarly, the variation of BSA rejection with PEG concentration is in direct alignment with the MWCO values of these membranes. As the MWCO of the membrane increases up to 6 wt% PEG, the rejection of BSA decreases from 92% to 84%. Then, the MWCO of the membrane decreases up to 10 wt%, leading to the corresponding increase in rejection to 95% at 10 wt%. A brief comparison study of BSA rejection by various tailor-made membranes is given in Table 5.

Table 5 Comparison of the BSA rejection

Author	Polymer	Additive (PEG)	Solvent	% BSA rejection
Chakrabarty et al.^62	PSF 12 wt%	400 Da, 6000 Da and 20000 Da 0 to 5 wt%	NMP and DMAc	11% to 56.4%
Ma et al.^63	PSF 18 wt%	400 Da, 800 Da 1500 Da, 4000 Da, 10000 Da, 20000 Da 0 to 10 wt%	DMAc	40% to 90%
Vijayalakshmi et al.^64	CA/PC	600 Da 0 to 10 wt%	NMP	70% to 95%
Amirilargani et al.^65	PES/PAN	400 Da, 600 Da, 1500 Da, 6000 Da	DMF	74.8% to 93.8%
Present study	CAP/PAN	200 Da, 400 Da, 1500 Da, 6000 Da 1 to 10 wt%	DMF	14% to 94.72%

---

It is observed from this table that the lower concentration of polysulfone (PSF) leads to lower rejection (maximum 56%) of BSA.⁶² However, the higher concentration of PSF (18 wt%) and lower molecular weight of PEG lead to a higher rejection of BSA of up to 90%.⁶³ CA based membranes are reported to show BSA rejection from 70% to 95% corresponding to various concentrations of PEG as additive.⁶⁴ Amirilargani et al.⁶⁵ reported BSA rejection between 75% and 94% corresponding to the various molecular weights of PEG in the PES/PAN blend membrane. In the present case, the PAN/CAP blend membrane exhibits a wide range of BSA rejection (14% to 95%) for various molecular weights of PEG (200 to 6000 Da) in the concentration range of 1 to 10 wt%. Thus, the present membrane is comparable with the reported ones with respect to BSA rejection.

3.12 FRR and FDR

The antifouling capacities of various membranes are quantified in terms of FRR and FDR. Thus, high FRR, low FDR and moderately high retention of BSA indicate antifouling properties for a membrane. From Fig. 12(a), it is observed that the membrane with PEG-200 shows an FRR of 60% and an FDR of 78%. The membrane with 6000 Da PEG shows a maximum of 80% FRR and about 65% FDR. On the other hand, PEG-200 results in 91% rejection of BSA, whereas PEG-6000 shows only 14% rejection. Thus, considering the rejection of BSA as a factor, the membrane with PEG-200 is desirable.

Fig. 12 (a) Flux recovery ratio (FRR) and flux decline ratio (FDR) of membrane with the variation of molecular weight of PEG. (b) Flux recovery ratio (FRR) and flux decline ratio (FDR) of membrane with variation of the concentration of PEG-200.

Effects of PEG concentration on antifouling performance are shown in Fig. 12(b). It is observed that for all the membranes, the rejection of BSA is above 85%. For 8 and 10 wt% PEG, rejection is about 95%. However, the FRR values are considerably poor in both the cases (about 50%). On the other hand, the performances of 2 and 6 wt% PEG membranes are considerably close; the FRR values are 75% and 78%, respectively, and corresponding FDR values are 77% and 75%. 2 wt% PEG membrane rejects 91% BSA, whereas 6 wt% membrane rejects 84% BSA. Therefore, as far as antifouling properties are concerned, both these membranes perform equally well.

It is interesting to note that the hydrophilic and antifouling properties of the developed membranes are similar to those reported in the literature. A PAN based membrane was reported to have a contact angle of 60° for 8 wt% of PEG-400 with a MWCO of 74 kDa,⁴⁵ and for PSF based membrane, it was 67° for 10 wt% of PEG-400 with a MWCO of 55 kDa.⁵² In the present work, 6 wt% PEG-200 in a PAN–CAP blend further reduces the contact angle to 55° with a MWCO of 56 kDa. Therefore, the addition of PEG to a PAN–CAP blend results in the formation of a more hydrophilic membrane with a reduced MWCO compared to a PAN–PEG membrane and an equivalent MWCO with respect to a PSF–PEG one. A PEG doped PAN–CAP blend membrane has more antifouling characteristics for 500 mg l⁻¹ BSA solution compared to a PAN–PEG membrane. The FDR value of a PAN–PEG membrane is more than 82%, and that of a PAN–CAP–PEG membrane is less than 72%, whereas the rejection of BSA in both the membranes is comparable. Therefore, the use of PEG additive in a PAN–CAP blend results in a more hydrophilic and antifouling membrane compared to those of PSF–PEG and PAN–PEG blends.

4. Conclusion

The effects of PEG as an additive on the characteristics and performance of a PAN–CAP blend membrane have been investigated in detail. The following conclusions can be drawn from this study:

(i) Increase in the molecular weight of PEG results in a more porous membrane. The permeability of the membrane increases from 6 × 10⁻¹¹ m Pa⁻¹ s⁻¹ to 26 × 10⁻¹¹ m Pa⁻¹ s⁻¹ as the molecular weight of PEG increases from 200 to 6000 Da. On the other hand, the pore density decreases with PEG molecular weight. Thus, the increase in permeability results from the increase in pore size of the membrane, and not by the increase in pore density.

(ii) Permeability of the membrane increases up to 6 wt% PEG-200 (9 × 10⁻¹¹ m Pa⁻¹ s⁻¹), and then it decreases to 2.5 × 10⁻¹¹ m Pa⁻¹ s⁻¹.

(iii) The MWCO of the membrane increases from 38 kDa to 105 kDa as molecular weight of PEG increases from 200 to 6000 Da. The corresponding contact angle decreases from 66° to 47°, making the membrane more hydrophilic.

(iv) MWCO reaches a maximum of 56 kDa at 6 wt% PEG-200, and then it is reduced to 38 kDa at 10 wt%. The contact angle decreases to 54° at 6 wt% PEG and increases to 65° at 10 wt%.

(v) The membranes with 2 wt% and 6 wt% PEG-200 showed equal antifouling properties. These membranes resulted in FRR values of 75% and 78%, FDR values of 77% and 75% and BSA rejection of 91% and 84%, respectively.

Appendix A

A.1 Estimation of thermodynamic parameter

The distance between the polymer–solvent axis and the part of the binodal curve parallel to this axis is MG. The degree of shift in binodal curve (DSBC) indicates the thermodynamic change of the casting solution due to the addition of additive. DSBC is estimated as follows:⁶⁶

Thus, the overall thermodynamic enhancement parameter, T, of the polymer due to the additive is quantified as⁶⁷

T = DSBC × χ_{solvent/additive} (A2)

where χ_{solvent/additive} is the interaction parameter between the additive and the solvent. For no additive, T is zero, and the large value of T indicates enhanced thermodynamic instability leading to the quick demixing and formation of a porous membrane. χ_{solvent/additive} in the above equation is the Flory–Huggins solvent polymer interaction parameter and is calculated as⁶⁷where V₁, R and δ are the molar volume of the solvent, ideal gas constant and solubility parameters, respectively. The molar volume of the solvent, DMF, was V₁ = 77.09 cm³ mol⁻¹.⁶⁸ The group contribution method is used to calculate the solubility parameter of additives of different molecular weights.^69–72 The Hansen solubility parameters of the additives and DMF are presented in Table 6. Literature data was used for the calculation of the solubility parameters of DMF.⁶⁸

Table 6 Hansen solubility parameters of additives and DMF

Material	n (MWp/MWm)	ρ (g cm^−3)	V (cm^3 mol^−1)	δd (MPa)^1/2	δp (MPa)^1/2	δh (MPa)^1/2	δt (MPa)^1/2
PEG-0.2 kDa	4	1.128	177	15.65	5.33	13.45	21.31
PEG-1.5 kDa	34	1.10	1364	16.10	1.74	9.45	18.75
PEG-6 kDa	136	1.07	5607	15.56	0.8367	8.74	17.44
DMF^55	—	—	—	17.4	13.7	11.30	24.86

---

A.1.1 Calculation of Hansen solubility parameter. Interactions between dispersion forces (δ_d), polar interactions (δ_p) and hydrogen bonding (δ_h) of the structural groups are considered to estimate the Hansen solubility parameter (δ_t).

δ_t² = δ_d² + δ_p² + δ_h² (A4)

where δ_d, δ_h and δ_p are calculated using the group contribution method.^69–72

A.2 Kinetic hindrance effect due to additives

This effect can be quantified by the diffusion of the solvent in the coagulation bath and the concentration profile of the solvent and is represented by the following equation:^53,73,74where γ, D_m, C, A, Ĉ_ns, C_t, C_∞, V and t₀ are the cast film compaction factors, diffusion coefficient of solvent, volume fraction, area of the cast film, volume fraction of pure nonsolvent (Ĉ_ns = 1), concentration of solvent in coagulation bath at time t, concentration of solvent in coagulation bath at infinite time, fixed volume of nonsolvent into which the solvent diffuses and time lag for sensing organics in the coagulation bath, respectively. Therefore, a plot between C_t/C_∞ versus t^0.5 results in a straight line with slope (m)

From the abovementioned equation, knowing the values V = 1000 cm³, A = 232 cm², Ĉ_ns = 1 and γ = 1, the ratio is estimated. The dimensionless kinetic hindrance parameter is expressed as⁴²

where M, A and μ are the mass, surface area and viscosity of the cast film, respectively.

The kinetic data on the polymer precipitation were obtained from the solvent leaching rate.^53,73,74 The polymer solution was cast on the non-woven fabric (attached to a glass plate) at room temperature, and it was then immediately placed into a deionized water bath. The samples were taken from the gelation bath using a micro syringe. The solvent concentration in the coagulation bath was measured with time in terms of refractive index using a digital refractometer (M/s, Cole-Parmer, Kolkata, India).

A.2.1 Estimation of kinetics parameter. The solvent concentration (C_t/C_∞) in the coagulation bath for different polymer compositions is shown in Fig. 13. It is observed from this figure that the rate of solvent–nonsolvent demixing increases in the order of 10 wt% PEG-200, 6 wt% PEG-200, PEG-1500 (1 wt%) and PEG-6000 Da (1 wt%). The additive increases the viscosity of the casting solution, promoting kinetic hindrance.

Fig. 13 Solvent concentration profile in the gelation bath. Inset: variation of solvent concentration with t^0.5 in gelation bath.

Nomenclature

A0	Membrane surface area, m^2
CP	Concentration of permeate, kg m^−3
CF	Concentration of feed, kg m^−3
Ct	Concentration of solvent in coagulation bath at time t, kg m^−3
C∞	Concentration of solvent in coagulation bath at infinite time, kg m^−3
Ĉns	Volume fraction of pure nonsolvent
Dm	Diffusion coefficient of solvent, m^2 s^−1
DSBC	Degree shift in binodal curve, eqn (A1)
FRR	Flux recovery ratio, eqn (7)
FDR	Flux decline ratio, eqn (8)
Jw	Pure water flux, m^3 m^−2 s^−1
Jw1	Water flux of the membrane before experiment, m^3 m^−2 s^−1
Jw2	Water flux of the membrane after experiment, m^3 m^−2 s^−1
J1	Initial flux of the membrane with BSA, m^3 m^−2 s^−1
Js	Flux at the end of one hour, m^3 m^−2 s^−1
K	Kinetic hindrance parameter, eqn (A7)
M	Mass of the cast film, kg
MG	Miscibility gap, eqn (A1)
L	Membrane thickness, m
nc	Pore density, number of pores per m^2
Q	Volumetric flow rate, m^3 s^−1
R	Rejection, %
ravg	Average pore size, m
T	Thermodynamic enhancement parameter, eqn (A2)
t0	Time lag for sensing organics in the coagulation bath, s
V	Fixed volume of nonsolvent into which the solvent diffuses, m^3
V1	Molar volume of the solvent, m^3 mol^−1

Greek symbols

Δ	Solubility parameters
ΔP	Transmembrane pressure drop, kPa
ΔT0	Sampling time, s
Γ	Cast film compaction factors
X	Interaction parameter between additive and the solvent, eqn (A2)
μ	Viscosity of water, Pa s

Acknowledgements

This work is partially supported by a grant from the Board of Research in Nuclear Sciences, Department of Atomic Energy, Government of India, Mumbai, under the scheme no. 2012/21/03-BRNS, Dt. 25-07-2012. Any opinions, findings and conclusions expressed in this paper are those of the authors and do not necessarily reflect the views of BRNS.

References

1. S. S. Madaeni, Water Res., 1999, 33, 301–308.
2. B. Nicolaisen, Desalination, 2002, 153, 355–360.
3. J. Radjenovic, M. Petrovic, F. Venturac and D. Barcelo, Water Res., 2008, 42, 3601–3610.
4. Y. He, Z. Ji and S. Li, Sep. Purif. Technol., 2007, 57, 366–373.
5. M. Mulder, Basic Principles of Membrane Technology, Kluwer Academic Publishers, Dordrecht, 1997.
6. M. J. Han and S. T. Nam, J. Membr. Sci., 2002, 202, 55–61.
7. B. Jung, J. K. Yoon, B. Kima and H. W. Rheeb, J. Membr. Sci., 2004, 243, 45–57.
8. Y. Liu, G. H. Koops and H. Strathmann, J. Membr. Sci., 2003, 223, 187–199.
9. W. Y. Chuang, T. H. Young, W. Y. Chiu and C. Y. Lin, Polymer, 2000, 41, 5633–5641.
10. J. H. Kim and K. H. Lee, J. Membr. Sci., 1998, 138, 153–163.
11. M. Amirilargani, E. Saljoughi and T. Mohammadi, Desalination, 2009, 249, 837–842.
12. M. Amirilargani, E. Saljoughi and T. Mohammadi, J. Appl. Polym. Sci., 2010, 115, 504–513.
13. E. Saljoughi and T. Mohammadi, Desalination, 2009, 249, 850–854.
14. A. Rahimpour and S. S. Madaeni, J. Membr. Sci., 2007, 305, 299–312.
15. E. Saljoughi, M. Sadrzadeh and T. Mohammadi, J. Membr. Sci., 2009, 326, 627–634.
16. B. Jung, J. K. Yoon, B. Kima and H. W. Rheeb, J. Membr. Sci., 2004, 243, 45–57.
17. Y. Liu, Y. Su, Y. Li, X. Zhao and Z. Jiang, RSC Adv., 2015, 5, 21349–21359.
18. M. A. Phadke, S. S. Kulkarni, S. K. Karode and D. A. Musale, J. Polym. Sci., Part B: Polym. Phys., 2005, 43, 2074–2085.
19. S. Yang and Z. Liu, J. Membr. Sci., 2003, 222, 87–98.
20. L. E. S. Brink, S. J. G. Elbers, T. Robbertsen and P. Both, J. Membr. Sci., 1993, 76, 281–291.
21. A. V. R. Reddy, D. J. Mohan, A. Bhattacharya, V. J. Shah and P. K. Ghosh, J. Membr. Sci., 2003, 214, 211–221.
22. Y. Su, W. Cheng, C. Li and Z. Jiang, J. Membr. Sci., 2009, 329, 246–252.
23. X. Yang, B. Zhang, Z. Liu, B. Deng, M. Yu, L. Li, H. Jiang and J. Li, J. Mater. Chem., 2011, 21, 11908–11915.
24. M. Ulbricht and A. Oechel, Eur. Polym. J., 1996, 32, 1045–1054.
25. J. Frahn, G. Malsch and F. H. Schwarz, Macromol. Symp., 2001, 164, 269–276.
26. T. D. Tran, S. Mori and M. Suzuki, Thin Solid Films, 2007, 515, 4148–4152.
27. M. Ulbricht and G. Belfort, J. Membr. Sci., 1996, 111, 193–215.
28. M. Sivakumar, R. Malaisamy, C. J. Sajitha, D. Mohan, V. Mohan and R. Rangarajan, J. Membr. Sci., 2000, 169, 215–228.
29. M. Sivakumar, D. Mohan and R. Rangarajan, Polym. Int., 1998, 47, 311–316.
30. M. Sivakumar, R. Malaisamy, C. J. Sajitha, D. Mohan, V. Mohan and R. Rangarajan, Eur. Polym. J., 1999, 35, 1647–1651.
31. M. Sivakumar, D. Mohan and R. Rangarajan, J. Membr. Sci., 2006, 268, 208–219.
32. E. Saljoughi, M. Amirilargani and T. Mohammadi, Desalination, 2010, 262, 72–78.
33. G. Arthanareeswaran, S. Velu and L. Muruganandam, J. Polym. Eng., 2011, 31, 125–131.
34. L. Wu, J. Sun and Q. Wang, J. Membr. Sci., 2006, 285, 290–298.
35. S. Nie, M. Tang, C. Cheng, Z. Yin, L. Wang, S. Suna and C. Zhao, Biomater. Sci., 2014, 2, 98–109.
36. L. Ma, H. Qin, C. Cheng, Y. Xia, C. He, C. Nie, L. Wanga and C. Zhao, J. Mater. Chem. B, 2014, 2, 363–375.
37. H. Zhou, C. Cheng, H. Qin, L. Ma, C. He, S. Nie, X. Zhang, Q. Fua and C. Zhao, Polym. Chem., 2014, 5, 3563–3575.
38. L. Wang, B. Su, C. Cheng, L. Ma, S. Li, S. Nie and C. Zhao, J. Mater. Chem. B, 2015, 3, 1391–1404.
39. Z. Yin, C. Cheng, H. Qin, C. Nie, C. He and C. Zhao, J. Biomed. Mater. Res., Part B, 2015, 103, 97–105.
40. X. Liu, J. Deng, L. Ma, C. Cheng, C. Nie, C. He and C. Zhao, Langmuir, 2014, 30, 14905–14915.
41. A. V. R. Reddy and H. R. Patel, Desalination, 2008, 221, 318–323.
42. M. C. Yang and T. Y. Liu, J. Membr. Sci., 2003, 226, 119–130.
43. Y. Xiuli, C. Hongbin, W. Xiu and Y. Yongxin, J. Membr. Sci., 1998, 146, 179–184.
44. A. Roy and S. De, J. Food Eng., 2014, 126, 7–16.
45. S. R. Panda and S. De, Polym. Eng. Sci., 2014, 54, 2375–2391.
46. N. Schamagl and H. Buschatz, Desalination, 2001, 139, 191–198.
47. B. Jung, J. K. Yoon, B. Kim and H. W. Rhee, J. Membr. Sci., 2004, 243, 45–57.
48. J. W. Lim, J. M. Lee, S. M. Yun, B. J. Park and Y. S. Lee, J. Ind. Eng. Chem., 2009, 15, 876–882.
49. E. Shekarian, E. Saljoughi and A. Naderi, J. Polym. Res., 2013, 20, 1–9.
50. J. Wang, Z. Yue, J. S. Ince and J. Economy, J. Membr. Sci., 2006, 286, 333–341.
51. X. Chen, Y. Su, F. Shen and Y. Wan, J. Membr. Sci., 2011, 384, 44–51.
52. S. R. Panda and S. De, J. Polym. Res., 2013, 20, 179–195.
53. M. Sadrzadeh and S. Bhattacharjee, J. Membr. Sci., 2014, 441, 31–44.
54. J. F. Li, Z. L. Xu, H. Yang, L. Y. Yu and M. Liu, Appl. Surf. Sci., 2009, 255, 4725–4732.
55. S. Singh, K. C. Khulbe, T. Matsuura and P. Ramamurthy, J. Membr. Sci., 1998, 142, 111–127.
56. I. C. Kim and K. H. Lee, J. Membr. Sci., 2004, 230, 183–188.
57. I. Sadeghi, A. Aroujalian, A. Raisi, B. Dabir and M. Fathizadeh, J. Membr. Sci., 2013, 430, 24–36.
58. R. R. Olmos, E. R. Hernandez, L. Z. F. Lopez, S. W. Lin and H. E. Gomez, J. Chil. Chem. Soc., 2008, 53, 1705–1708.
59. M. Ali, M. Zafar, T. Jamil and M. T. Z. Butt, Desalination, 2011, 270, 98–104.
60. S. R. Panda and S. De, RSC Adv., 2015, 5, 23599–23612.
61. M. C. Porter, Handbook of Industrial Membrane Technology, Crest Publishing House, New Delhi, India, 2005.
62. B. Chakrabarty, A. K. Ghoshal and M. K. Purkait, J. Membr. Sci., 2008, 309, 209–221.
63. Y. Ma, F. Shi, J. Ma, M. Wu, J. Zhang and C. Gao, Desalination, 2011, 272, 51–58.
64. A. Vijayalakshmi, D. L. Arockiasamy, A. Nagendran and D. Mohan, Sep. Purif. Technol., 2008, 62, 32–38.
65. M. Amirilargani, A. Sabetghadam and T. Mohammadi, Polym. Adv. Technol., 2012, 23, 398–407.
66. J. H. Kim, B. R. Min, H. C. Park, J. Won and Y. S. Kang, J. Appl. Polym. Sci., 2001, 81, 3481–3488.
67. L. Shuguang, J. Chengzhang and Z. Yuanqi, Desalination, 1987, 62, 79–88.
68. A. F. M. Barton, CRC Handbook of Solubility Parameters and Other Cohesion Parameters, CRC Press, Boca Raton, FL, 2nd edn, 1991.
69. D. W. Van Krevelen and K. T. Nijenhuis, Properties of Polymers, Elsevier, 4th edn, 2009.
70. J. F. Li, Z. L. Xu, H. Yang, C. P. Feng and J. H. Shi, J. Appl. Polym. Sci., 2007, 107, 4100–4108.
71. C. Özdemir and A. Güner, Eur. Polym. J., 2007, 43, 3068–3093.
72. M. Maniruzzaman, J. S. Boateng, M. Bonnefille, A. Aranyos, J. C. Mitchell and D. Douroumis, Eur. J. Pharm. Biopharm., 2012, 80, 433–442.
73. A. V. Patsis and E. H. Henriques, J. Polym. Sci., Part B: Polym. Phys., 1990, 28, 2681–2689.
74. A. C. Sun, W. Kosar, Y. Zhang and X. Feng, Desalination, 2013, 309, 156–164.

---