(PSF)/polyethylene glycol (PEG)

Desalination 347 (2014) 52–65

Contents lists available at ScienceDirect

Desalination journal homepage: www.elsevier.com/locate/desal

Preparation, characterization and performance of ZnCl2 incorporated polysulfone (PSF)/polyethylene glycol (PEG) blend low pressure nanofiltration membranes Swapna Rekha Panda, Sirshendu De ⁎ Department of Chemical Engineering, Indian Institute of Technology Kharagpur, Kharagpur 721302, India

H I G H L I G H T S • • • • •

Polysulfone based NF membranes were cast using PEG 200 and ZnCl2 as additives. Molecular weight cut off of NF membrane was less than 180. Membranes were having negative zeta potential. About 40% NaCl rejection was obtained at pH 2 and 11 at 690 kPa. Rejections of chrysoidine R, crystal violet and congo red were above 95%.

a r t i c l e

i n f o

Article history: Received 15 January 2014 Received in revised form 1 May 2014 Accepted 21 May 2014 Available online 10 June 2014 Keywords: Polysulfone Nanofiltration Low pressure Polyethylene glycol zinc chloride

a b s t r a c t Simultaneous use of organic (PEG 200), and inorganic (ZnCl2) additives to polysulfone for casting of nanofiltration (NF) membrane was reported in this work. Dimethylformamide (DMF) was used as the solvent. The membranes were characterized in terms of surface morphology, permeability, porosity, contact angle, molecular weight cut off and zeta potential. The resultant NF membranes were tested with monovalent and divalent electrolytes sucrose, glucose and three dyes in the range of molecular weight from 263 to 697. Both NF membranes showed rejection of salts, NaCl: 40–43% and Na2SO4: 52–54%, at 690 kPa and pH 11. The corresponding permeate flux were 20 to 22 l/m2·h. Among the dyes crystal violet and congo red are rejected more than 98% and chrysoidine more than 95% over the transmembrane pressure range of 276 to 690 kPa. Electrostatic interaction was found to play a determinant role in rejection of salts, whereas, size exclusion was the major mechanism for rejection of dyes. © 2014 Elsevier B.V. All rights reserved.

1. Introduction With technological advancement, membrane technology is steadily growing in various industries due to competitive energy prices and environmental concerns. Although a number of pressure driven solvent separations had been explored till date, nanofiltration (NF) is relatively a younger technology. The use of polysulfone (PSF) as a material for the synthesis of ultrafiltration (UF) and NF membrane is widespread because of its good thermal, chemical and mechanical stability. Phase inversion method developed in the early sixties, highlights the most Abbreviations: PASB, poly[(4-aminophenyl) sulfonyl] butane diamide; mPASB, methyalated poly[(4-aminophenyl) sulfonyl] butane diamide; PEI, polyetherimide; PVP, polyvinylpyrrolidone; PES, polyether sulfone; PVDF, polyvinylidene difluoride; PI, polyimide; PA, polyamide; DMF, N,N-dimethylformamide; SPES, sulfonated polyethersulfone; PP, polypropylene; TFC, thin film composite; CS, chitosan; PPES, poly(1,4-phenylene ether ether sulfone); GO, graphene oxide. ⁎ Corresponding author. Tel.: +91 3222 283926; fax: +91 3222 255303. E-mail address: [email protected] (S. De).

http://dx.doi.org/10.1016/j.desal.2014.05.030 0011-9164/© 2014 Elsevier B.V. All rights reserved.

versatile, economically feasible and reproducible mechanisms for the fabrication of polymeric asymmetric membranes. These membranes possess a skin-layer at the top followed by a porous sub layer. Thin skin layer enables high permeability and selectivity. Limitation of PSF and polyether sulfone (PES) over other polymers is their high hydrophobicity that causes severe membrane fouling. When feed is charged to the membrane module, hydrophilic surface shows better performance against fouling. Surface hydrophilicity can be augmented by different modification routes including graft polymerization, plasma treatment, and physical pre-adsorption of hydrophilic components to the membrane surface [1–5]. Another simple and inexpensive approach for the bulk modification is the incorporation of hydrophilic additives in polymer solution itself. Additives affect the structure and state of the casting solution, altering the physical, morphological and permeation characteristics of asymmetric membranes. The use of different organic additives, like, polyethylene glycol (PEG) [6–8], polyvinylpyrrolidone (PVP) [9–11], acids [12], acetone [13], and alcohols [14,15] has been reported. Kim and Lee [16] found that smaller molecular weight of PEG

S.R. Panda, S. De / Desalination 347 (2014) 52–65

200 and 400 can be used as a pore reducing hydrophilic additive rather than a pore forming one using polyetherimide (PEI) as base polymer. Nanofiltration membranes recently have gained their importance in the field of drinking water, industrial water and waste water treatment [17,18]. Charge in NF membranes is one of the major factors for the separation of monovalent and divalent ions or organic solutes in the feed solution. Size exclusion and electrostatic interactions are two basic phenomena by which separation occurs in NF membranes [19–21]. Isloor and coworkers have done extensive work in fabrication and characterization of PSF based NF membranes using different functional additives like poly isobutylene-alt-maleic anhydride, poly[(4-aminophenyl) sulfonyl] butane diamide (PASB), and methyalated poly[(4-aminophenyl) sulfonyl] butane diamide (mPASB) [22–25]. PSF was better to prepare chlorine resistant membranes. Moreover, it was soluble in most of the organic solvents [26]. The use of inorganic salts in the polymeric solution was also reported to be a promising and effective method to prepare membranes with enhanced performance. These solutes changed the solvent properties as well as interaction between the macromolecular chains [27]. Reasonable literature is available to observe the effects of inorganic salts as additives to various casting solution of PSF, PES, polyvinylidene difluoride (PVDF) and polyimide [28–37]. These are summarized in Table 1. This literature survey highlights the following points: (i) membrane pores were reduced using zinc chloride as additive in the absence of any organic additive [28–30] and the resultant membrane was still in UF range, except in one case, where NF membrane was resulted [30]; (ii) addition of lithium compounds [31–35] and potassium perchlorate [36] led to the increase in pore size of the membranes, all being in UF range. The present work employs simultaneous use of organic (PEG 200) as well as inorganic (ZnCl2) additives (both are well known pore constrictors) [16,28–30] to PSF/DMF system in order to get NF membrane. In the first step, concentration of PEG 200 was selected to obtain highly selective dense UF membrane. In the next step, the effect of ZnCl2 was investigated in PSF-PEG/DMF system to arrive at NF membrane. The membranes were characterized in terms of permeability, molecular weight cut off (MWCO), morphological study, contact angle and surface zeta potential. Performance of NF membrane was tested in terms of rejection of monovalent and divalent salts, glucose, sucrose and various dyes. Thus, the novelty of this work is that it is the first attempt to use simultaneous inorganic (ZnCl2) and organic (PEG 200) additives on PSF/DMF system to get NF membrane. 2. Experimental 2.1. Materials PSF with an average molecular weight of 22,400 Da was procured from M/s, Solvay Chemicals, Mumbai, India and was used as the base polymer. Solvent N,N-dimethylformamide (DMF) was purchased from M/s, Merck (India) Ltd., Mumbai, India. Polyeth±ylene glycol (PEG) of average molecular weight 200 Da, 400 Da, 600 Da, 4 kDa, 20 kDa and 35 kDa were supplied by M/s, S. R. Ltd., Mumbai, India. Dextran (average molecular weight: 70 kDa) and polyethylene glycol of average molecular weight 100 and 200 kDa were procured from M/s, Sigma Chemicals and M/s, Aldrich Chemicals, USA, respectively. These neutral solutes were used to evaluate molecular weight cut-off (MWCO), of the cast membranes. The dyes chrysoidine R, MW: 262.74, and congo red, MW: 696.65, were supplied by M/s, Loba Chemie Pvt Ltd., Mumbai, India. Crystal violet, MW: 407.5, was procured from M/s, Sisco Research Laboratory, Mumbai, India. Distilled water was used as the non-solvent in the coagulation bath. All chemicals were of analytical grade and used without further purification. 2.2. Determination of coagulation value Thermodynamic stability of polymeric solution can be explained in terms of its coagulation value [28]. Coagulation value is defined as

53

weight in grams of 1:1 wt.% non solvent/solvent mixture needed to initiate a visual turbidity in 10 g of polymeric solution, which persists for 24 h at 25 °C. 2.3. Membrane preparation Flat sheet PSF membranes were prepared by phase inversion method and the steps involved in the membrane fabrication were as follows. Fixed amount of PEG was added to a premixed 18% PSF in DMF dissolved at 60 °C. The solution was magnetically stirred for at least 12 h to ensure complete dissolution of the polymer. During the whole process of stirring, the lid of the container was kept closed to prevent the loss of solvent due to evaporation. The prepared solutions were kept for at least 24 h without stirring at room temperature to remove air bubbles. In the first step of casting of membranes, non-woven polyester fabric of thickness 118 ± 22.8 μm (product number TNW006013, supplied by M/s, Hollytex Inc., New York, USA) was attached to a clean glass plate using adhesive tape. The casting solution was drawn down the fabric using a doctor's blade with an adjustable thickness fixed at 200 μm and was immediately immersed in a precipitation bath containing distilled water at room temperature to initiate the non-solvent induced phase separation. The membrane was kept in the precipitation bath for 10 min and then, it was transferred to another container having fresh distilled water for 24 h to remove the excess solvent. After that, the membrane was ready to be tested. In order to see the effect of ZnCl2 on the casting solution, another set of asymmetric membranes were cast by following all the above steps using ZnCl2, in the concentration range of 0 to 2 wt.%. Concentration of PEG of molecular weight 200 Da was varied up to 10 wt.%. Casting of the membranes was done at room temperature (25 ± 2° C). Viscosity of the polymer solution was measured using a rheometer (model: Physica MCR 301, supplied by M/S, Anton Parr, Austria). 2.4. Characterization of membranes The following characterizations were carried out for the prepared membranes. 2.4.1. Porosity (ε) and contact angle (CA) Membrane porosity plays an important role in its performance. Membrane porosity was measured by the mass loss of wet membrane after its drying. The membrane, soaked with distilled water was weighed after mopping superficial water with filter paper. Then, the wet membrane was placed in an air-circulating oven at 60 °C for 24 h and then further dried in vacuum oven before measuring the dry weight until a constant mass was obtained. From the two weights (wet sample weight, w0 and dry sample weight, w1), the porosity of the membranes was calculated using the following equation [7]. ε¼

w0 −w1 100% ρw Al

ð1Þ

where, ε is the membrane porosity, A is the membrane surface area, l is the membrane thickness and ρw is water density. In order to minimize the experimental errors, the membrane porosity of each sample was measured three times and the average values were reported. The contact angle was measured by a Goniometer (supplied by Labline instrument, Mumbai, India, manufactured by Rame-Hart instrument Co., New Jersy, USA; model number: 200-F4) using sessile drop method. The contact angle was measured at six different locations of the membrane and the average value was reported. 2.4.2. Membrane permeability Measurement of pure water permeability was carried out in a batch filtration cell [7]. Effective area of the membrane in the module was 34 cm2. First, the cell was filled with 500 ml of distilled water and

54

Table 1 Comparison and highlights of the present work with other works in literature. [28]

[29]

[30]

[31]

[32]

This study

Polymer Solvent Organic additive

PEI 15–25 wt.% NMP 73 to 85 wt.% –

PSF 20 wt.% NMP 77 to 79 wt.% –

PI 15 wt.% NMP 61–85 wt.% –

PVDF 15 wt.% DMAc, DMF NMP –

PVDF 15–22 wt.% DMAc –

Inorganic salts Casting knife thickness (μm) Gelation medium Membrane pore structure

ZnCl2 1 to 3 wt.% Not reported Water 18 °C Finger like to dense structure

ZnCl2 0 to 3 wt.% 120 Water 20 + 0.01 °C. Small and narrow pore sizes

ZnCl2 0, 6, 24 wt.% 100 Water at room temperature Small and narrow pore sizes

LiCl 0 to10 wt.% 350 Water 4 °C Finger like cavities

Average pore size Permeability Membrane MWCO Polymer solution viscosity Other characterization

NA 6 × 10−11 to 9.8 × 10−11 m/Pa·s NA 1.7 to 2.28 Pa·s SEM

NA 1.6 × 10−10 to 2 × 10−10 m/Pa·s 40 to 78 kDa NA SEM, FTIR

NA Not reported 49–60 kDa NA SEM

Salt rejection test

NA

NA

5–20 nm – NA NA SEM, ATR-FTIR, AFM, XPS, DSC With NaCl

LiCl, H2O, TiO2 nanoparticles – Water 25 °C Finger like to macrovoids type with spongy structure 14–32 nm Not reported – 0.1 to 0.25 Pa·s FESEM, EDX, contact angle

PSF 18 wt.% DMF 72–82 wt.% PEG 200 (0–2 wt.%) ZnCl2 0 to 2 wt.% 200 Water at room temperature Small, narrow, finger like pore with dense structure 1.6–14.5 nm 3 × 10−12 m/Pa·s 180 Da to 55 kDa 0.56 to 0.73 Pa·s SEM, contact angle

NA

NA

Dye rejection test

NA

NA

NA

NA

NA

Mono/oligo saccharides rejection test Type of membrane resulted Casting conditions and membrane characteristics Polymer

NA

NA

NA

NA

NA

With NaCl, Na2SO4, MgSO4, MgCl2 Chrysoidine R, crystal violet, congo red Glucose, sucrose

UF/flat sheet [33]

UF/flat sheet [34]

NF/flat sheet [35]

UF/flat sheet [36]

UF/hollow fiber [37]

UF, NF/flat sheet This study

PES 20 wt.%

PES 20 wt.%

PES 16 wt.%

PES 18 wt.%

PSF 18 wt.%

Solvent

PVDF co polymer PVDF-homo polymer DMAc 72.5 to 80 wt.%

DMF 76 to 80 wt.%

DMAc 77 to 82 wt.%

NMP

DMF 72 to 82 wt.%

Organic additive Inorganic salts Casting knife thickness (μm) Gelation medium

PVP 2.5 to 7.5 wt.% LiCl 2.5 to 7.5 wt.% 250 Water 20 ± 2 °C

– LiBr 1 to 4 wt.% – Water 25 °C

DMF 76 to 80 wt.% DMF 59 to 62 wt.% + acetone 17 to 17.9 wt.% – LiCl 1 to 4 wt.% 200 Water at room temperature

– Nano clay 1 to 5 wt.% 100 Water at room temperature

PEG 2000 to 2 wt.% ZnCl2 0 to 2 wt.% 200 Water at room temperature

Membrane pore structure

Macrovoids with finger like pores – – – – SEM, mechanical stability

Finger like pores, large macrovoids 18 to 3.5 nm – 2.1 to 36 kDa 180–800 cp MWCO, EDX, SEM

Finger like pores

Average pore size Permeability Membrane MWCO Polymer solution viscosity Other characterization

Sponge like structure, fingerlike pores – 6 × 10−10 m/Pa·s – – Contact angle

Small, narrow, finger like pore with dense structure 1.6 to 14.5 nm 9.0 to 9.2 × 10−12 m/Pa·s 180 Da to 55 kDa 0.56 to 0.73 Pa·s SEM, contact angle

Salt rejection test

NA

NA

NA

PVP 2 wt.% KClO4 1 to 7 wt.% 150 Water + isopropyl alcohol 80:20 vol % ratio Small finger like pores, nodular nano porous structure 32.8 to 127.4 nm (surface) – – 0.81–3.3 Pa·s Contact angle, AFM, milk protein rejection, SEM NA

NA

Dye rejection test

NA

NA

NA

NA

NA

Mono/oligo saccharides rejection test Type of membrane resulted

NA

NA

NA

NA

NA

With NaCl, Na2SO4, MgSO4, MgCl2 Chrysoidine R, Crystal violet and Congo red Glucose, sucrose rejection

UF/flat sheet

UF/hollow fiber

UF/flat sheet

UF/flat sheet

UF/flat sheet

UF and NF

5.2 to 7.4 nm – 55 to 95 kDa – Contact angle


Casting conditions and membrane characteristics


membranes were compacted at 690 kPa for 3 to 4 h. The permeate flux was calculated by Jw ¼ Q =A ΔT

ð2Þ

where, Q is the volumetric flow rate of permeating water, A is effective membrane area, and ΔT is the sampling time. Next, the steady state permeate flux was noted at five values of transmembrane pressure drop. A plot of Jw with transmembrane pressure drop resulted into a straight line through the origin. From the slope of this curve, membrane permeability was estimated. 2.4.3. MWCO of the membrane Solute rejection measurements were carried out in a stirred batch cell. Solutions were prepared using 10 kg/m3 using different neutral solutes, namely, PEG of different molecular weights, 400 Da, 4 kDa, 10 kDa, 20 kDa, 35 kDa, 100 kDa, 200 kDa and dextran 70 kDa. The experiments were conducted at 69 kPa transmembrane pressure (TMP) and at 2000 rpm. The permeate samples were analyzed using a digital refractometer (supplied by M/s, Cole-Parmer, Kolkota, India). Rejection values were plotted against the molecular weight of solutes in a semilogarithmic curve. Molecular weight corresponding to 90% rejection was estimated as MWCO of the membrane. After the experiment, membrane was thoroughly rinsed with distilled water and its original permeability was restored. The rejection was calculated as,

2.5. Permeation experiments of NF membranes Rejection of salt was studied using NaCl, Na2SO4, MgSO4 and MgCl2 solution. Rejection of two neutral solutes glucose, sucrose and three dyes chrysoidine R, crystal violet and congo red was also investigated. All the experiments were conducted in a cross flow filtration set-up under total recycle mode. The details of the experimental set-up are available [40]. Experiments were performed under four operating pressures, 276, 414, 552 and 690 kPa. The cross flow rate was maintained at 50 l/h. Concentration of salts, sucrose and glucose was selected as 1000 mg/l and that of all three dyes was 200 mg/l. Concentration of salt was measured via conductivity meter (model: PCSTestr 35, Eutech Instruments, Germany). Concentration of glucose and sucrose was measured using a digital refractometer. UV–vis spectrophotometer (model: Lambda 35, supplied by M/S, Perkin Elmer, Connecticut, USA) was used to measure the dye concentration. The effect of pH of salt solution was studied at a pH range 2 to 11. pH values were measured using pH meter (model: PCSTestr 35, Eutech Instruments, Germany). For each membrane and solute, experiments were carried out thrice and the average values were reported.

ð3Þ 3.1. Effects of additives

where, C0 and Cp are the concentrations of solute in feed and permeate, respectively. 2.4.4. Scanning Electron Micrograph (SEM) SEM images were analyzed using a scanning electron microscope (supplied by JEOL, Japan, model ESM-5800). First, the membrane was carefully removed from the substrate. It was then cut into small pieces, dried using a filter paper, dipped in liquid nitrogen for 1 min and then fractured cryogenically. The fractured samples were dried under vacuum. The samples were gold sputtered and then mounted on the sample pad. The observations were carried out at the top surface and on the cross-sections to evaluate the membrane thickness and morphology. 2.4.5. Skin and total membrane thickness Microstructure of the membrane controls the separation performance. Asymmetric microstructure includes a thin skin layer on the top followed by a porous substructure. Based on the morphology, skin layer thickness may be defined as the perpendicular distance from the topmost layer to the surface of the membrane where the pores begin to transit into a more interconnected network. 2.4.6. Estimation of average pore radius The average pore radius of the membranes was evaluated by using Eq. (4) [38]. −10

value at specific wavelength signifies the presence of the functional groups in the respective membrane.

3. Results and discussions

R ¼ 1−Cp =C0 100%

rm ðcmÞ ¼ 16:73 10

55

0:557

ðMWCOÞ

ð4Þ

This section is divided into three parts. First part deals with the effects of concentration of PEG-200 in polysulfone solution. The interaction of ZnCl2 with PSF matrix is presented in the second part and the synergistic effects of PEG-200 and ZnCl2 is analyzed in the third part. 3.1.1. Effects of PEG-200 Additives like PEG are well known pore former, due to their strong hydration and large excluded volume in the gelation bath. In contrast, Kim and Lee [16] found that smaller molecular weight of polyethylene glycol (PEG 200 and PEG 400) can be used as a pore reducing hydrophilic additive rather than a pore forming one. Molecular weight of the additives has a fine control over the morphology and permeability of the membranes [10,16,41]. Thus, to observe the effect of PEG on the pore size of the membrane, the effects of lower molecular weight PEG 200 in the concentration range of 2 to 10 wt.% were studied in the casting solution. As is evident from the experimental findings reported earlier [7, 8], the presence of PEG in the casting solution has two different effects. Firstly, PEG consumes some of the solvent leading to higher polymer concentration and increased viscosity of the casting solution. The polymeric solution becomes less stable due to increased viscosity, resulting to rapid demixing. Viscosity can inhibit the rate of demixing that allows the formation of a thermodynamically meta-stable state, resulting to Table 2 Composition of the casting solution for the present study. Membrane

2.4.7. Zeta potential measurement with electrolyte solution Zeta potential of the membranes was determined using an electric field enhanced cross flow filtration set-up. The set-up and the procedure of measurement were available [39]. Measurements were conducted using 1 mM NaCl in the pH range of 2 to 11. 2.4.8. Fourier Transform Infrared Spectroscopy (FTIR) analysis In order to investigate the functional changes that occurred between PSF/PEG 200 and ZnCl2, FTIR (supplied by M/s, Perkin Elmer, Connecticut, USA; model: Spectrum 100) analysis was performed. The transmittance

MPEG0 MPEG2 MPEG4 MPEG8 MPEG10 MPEZ-0 MPEGZ-0.1 MPEGZ-0.3 MPEGZ-1 MPEGZ-2

PSF

PEG 200

ZnCl2

DMF

Viscosity (cp)

wt. %

wt. %

wt. %

wt. %

Shear rate γ, 24 (s−1)

18 18 18 18 18 18 18 18 18 18

0 2 4 8 10 2 2 2 2 2

– – – – – – 0.1 0.3 1 2

82 80 78 74 72 80 79.9 79.7 79 78

– 920 – – – 920 923 1140 2590 3440

56


microvoids. Composition of various membranes we have used in our present work with codes is shown in Table 2. As observed from Fig. 1(a), PSF membranes without PEG (MPEG0) result to a spongy,

more porous membrane [7,8] and consequently, influences the membrane morphology. Secondly, being a hydrophilic additive, some of PEG leaches out [38] during phase inversion, thereby forming the

a MPEG0

MPEG2

MPEG4

MPEG8

skin

contact angle(0) Porosity (%) Permeability x 1011 (m/Pa.s)

MPEG10

b 10 8 6 4 2 80 60 40 20 100 90 80 70 60 50 40

0

2

4

6

8

10

12

Concentration PEG 200 (wt%)

Fig. 1. Effect of concentration of PEG-200 from (0 to 10 wt.%) on PSF membranes: (a) cross sectional views of SEM images (b) membrane porosity, contact angle and permeability.


57

200 concentration 2 wt.% and it was 39 kDa with 12 nm average pore diameter at this concentration of PEG. However, with further increase of PEG, MWCO increased to 43 kDa at 4 wt.% with 13 nm average pore diameter. Further addition of PEG up to 8 and 10 wt.% leads to a decrease in the porosity and permeability of the membranes. These results are supported with SEM photographs, permeability data and porosity measurements given in Figs. 1(a) and (b), respectively. Thus, with PEG 200 only PSF-DMF system results to UF membranes only [7,8,10,16].

loose morphology with few open macrovoids at the bottom layer of the membrane cross section. Thickness of the membranes measured from SEM analysis is presented in Table 3. The addition of PEG in the virgin membrane from zero to 2 wt.% (MPEG2) shows the development of smaller fingerlike pores just below the skin layer of the membrane leaving behind a relatively thick, dense, spongy mass at the bottom layer. Thus, the membrane formed at this composition is expected to show the lowest permeability and higher retention characteristics. With further addition of PEG to 4 wt.% (MPEG4), the structure is dominated by macrovoids. The finger like pore structure increases in number and size, promoting the formation of larger pores that span almost the entire membrane thickness. These data support the fact that for lower PEG concentration (2 wt.%), less amount of PEG is leached out of the membrane giving rise to small pores below the skin layer with a thicker and denser spongy bottom layer. At higher concentration (4 wt.%), leaching of PEG is more, resulting to more number of macrovoids. On further increasing the concentration to 6, 8 and 10 wt.%, viscosity of the casting solution increases. As a result, solvent-non solvent exchange rate is suppressed that results to a denser structure [14–16]. These data is in well agreement by the similar observations made by other researchers reported earlier [42,43]. Typical variation of membrane porosity, contact angle and permeability with the composition of PEG 200 is shown in Fig. 1(b). As was expected, Fig. 1(b) clearly shows that there is a decline in porosity at 2 wt.% of PEG, then it increases up to 4 wt.% and decreases thereafter up to 10 wt.%. These trends are in direct corroboration with SEM images. Highest contact angle is obtained with the membrane without any additive, i.e., 0% PEG. The membrane surface becomes more hydrophilic, with the concentration of PEG 200 in the casting solution exhibiting a decrease in contact angle. The variation of membrane permeability with PEG composition is a direct consequence of porosity variation. The permeability of PSF membrane without any additive is 3.3 × 10−11 m/Pa·s. The addition of 2 wt.% PEG 200 reduces the permeability value to 2.8 × 10−11 m/Pa·s. Further increase in PEG concentration leads to gradual increase in the membrane permeability, up to 4 wt.%. This is due to the fact that, at lower concentration of PEG, slower demixing occurs as described earlier, leading to a thick, spongy, denser bottom layer, resulting into the lowest permeability. As the concentration of PEG increases, more macrovoids appear and the membrane matrix becomes more porous, as observed from SEM images described earlier. This leads to an increase in membrane permeability up to 4 wt.%. Further increase in PEG concentration to 10 wt.%, makes the membrane more compact and pore size of the membrane is also reduced. This results to a decrease in permeability as shown in Fig. 1(b). This may be due to the combined effect of increase in viscosity and lowered thermodynamic stability by the addition of hydrophilic additive to the polymer solution [44,45]. MWCO curves for different composition of PEG 200 are shown in Table 3. The average pore size of membranes determined from MWCO data (solute separation curve) using Eq. (4), is given in Table 3. The results show that, the membranes without any additive exhibited a MWCO 55 kDa. The MWCO and the pore diameter decreased at PEG

3.1.2. Effects of ZnCl2 The effects of ZnCl2 on PSF have been extensively studied by Kim et al. [29]. They observed that ZnCl2 imparts strong association with PSF resulting to dense and homogeneous morphology of the membrane. They explained this phenomenon by an increase in viscosity with ZnCl2 concentration. They concluded that ZnCl2 associates with PSF and acts as a charged group of the polymer. This makes the PSF to behave like a polyelectrolyte that increases the viscosity of polymer solution through electroviscous effect. This observation was confirmed by FTIR studies. They also observed experimentally that critical concentration of PSF decreased with ZnCl2 concentration. Thus, at higher ZnCl2 concentration, the polymer became closely packed resulting to a denser morphology. ZnCl2 concentration up to 1 wt.% reduced the molecular weight cut off the membrane from 78 kDa to 40 kDa with N-methyl2-pyrrolidone (NMP) as the solvent [29]. It may be noted that although ZnCl2 is used as a pore constrictor, the resultant membrane is still in UF range. 3.1.3. Synergistic effect of PEG 200 and ZnCl2 As presented in the preceding section, PEG 200 at 2 wt.% and 18 wt.% PSF resulted denser membrane. Thus, effects of ZnCl2 on this composition of polymer solution were investigated. The mechanism of formation of membranes by non-solvent induced phase separation is very complicated specifically for a pentamerous system like PSF/PEG 200/DMF/ZnCl2/non-solvent. The cross sectional and top view of SEM images for PSF membrane without ZnCl2 and with ZnCl2 are shown in Fig. 2(a). As is observed from Fig. 2(a) that PSF membranes without ZnCl2, MPEGZ-0 result to a spongy, dense morphology without any macrovoids at the bottom layer of the membrane. As the additive percentage increases from zero to 0.1 wt.% (MPEGZ-0.1), the figure shows development of spongy layer with the formation of open macrovoids at the bottom part of the membrane cross section. Thus, the membrane formed at this composition is expected to show higher permeability. Further addition of ZnCl2 from 0.1 (MPEGZ-0.1) to 0.3 wt.% (MPEGZ-0.3), the morphology is dominated by wide finger like structure with bigger macrovoids at the bottom as observed from this figure. With further addition of additive at 1 wt.%, the membrane had a relatively dense morphology with porous structure, as is observed from the SEM images. The experimental finding of the thickness of the membranes is presented in Table 3. At 1 wt.% ZnCl2, the visual observation from SEM images shows a total membrane thickness and skin thickness of 160 μm and 20 μm, respectively, from Table 3. This can be compared to the membrane with 0.3 wt.% ZnCl2 having 110 and 9 μm

Table 3 Thickness, average pore size and MWCO of the cast membranes for the present study. Membrane ID

Skin thickness (μm)

Total thickness (μm)

MWCO

rm. (nm) Eq. (4)

MPEG0 MPEG2 MPEG4 MPEG8 MPEG10 MPEGZ-0 MPEGZ-0.1 MPEGZ-0.3 MPEGZ-1 MPEGZ-2

8 10 7 35 18 10 20 9 20 25

90 140 150 200 110 140 100 110 160 200

55,000 ± 1940 39,000 ± 1840 43,000 ± 1600 41,000 ± 1500 40,000 ± 1720 39,000 ± 1840 43,000 ± 2100 49,000 ± 1100 180 ± 30 –

7.3 6.0 6.3 6.2 6.1 6.0 6.4 6.8 0.4 –

± ± ± ± ± ± ± ± ± ±

0.16 0.2 0.14 0.7 0.36 0.2 0.4 0.18 0.4 0.5

± ± ± ± ± ± ± ± ± ±

1.8 2.8 3.0 4.0 2.2 2.8 2.0 2.2 3.2 4.0

± ± ± ± ± ± ± ± ±

0.15 0.11 0.12 0.12 0.10 0.11 0.13 0.14 0.01

58


for total and skin thickness. More addition of ZnCl2 to 2 wt.% resulted to the formation of a very thick (total thickness of 200 μm) and dense membrane with a skin thickness of 25 μm. The skin layer has an active role in determining the transport properties of the asymmetric membranes by solution diffusion mechanism [41]. A relatively dense skin

layer may show lower permeability with good selectivity. Thus, the membranes formed at these compositions i.e., 1 wt.% and 2 wt.% ZnCl2 are expected to show the lowest permeability and higher retention of solutes. The behavior of membrane morphology with ZnCl2 is explained as follows.

a MPEGZ-0

MPEGZ-0

MPEGZ-0.1

MPEGZ-0.1

MPEGZ-0.3

MPEGZ-0.3

MPEGZ-1

MPEGZ-1

MPEGZ-2

MPEGZ-2

Fig. 2. Effect of ZnCl2 from 0 to 2 wt.%, on PSF-PEG blend membranes. (a) Cross sectional views of SEM images: left side; right side: top views at high magnification up to 20,000× for MPEGZ-1 and MPEGZ-2 membranes (b) Coagulation value of the polymeric solution with ZnCl2 additive (c) Effect of FTIR spectra of polysulfone film with and without ZnCl2 additive.


1.6

Porosity,

15 wt% PSF

80

1.4

Contact angle Error bars: ±5%

90

70 60

1.2

Porosity (%)

Coagulation value (g)

100

90

1.0

80 50 40

70

30

Contact angle (0)

a

b

59

0.8 20 0.6

0 0.4

60

10

0.0

0.5

1.0

1.5

50 0.0

0.3

0.6

0.9

1.2

1.5

1.8

2.1

2.4

2.1

2.4

Concentration of ZnCl2 (wt %)

2.0

ZnCl2 concentration (wt%)

b

c

8 Error bars: ±5%

Permeability × 1011 ( m/Pa.s)

7

Transmittance (%)

MPEGZ-2 873.10

1150.72

MPEGZ-1

1151

873.16

6 5 4 3 2 1 0 0.0

MPEGZ-0

1151.02

1100

1000

900

800

0.6

0.9

1.2

1.5

1.8

Concentration of ZnCl2 (wt %)

873.26

1200

0.3

700

600

c

Wave number (cm-1)

100

Viscosity and coagulation value of polymeric solution play a crucial role in determining the state of the polymeric solution [28]. The change in coagulation value with ZnCl2 is plotted in Fig. 2(b). The coagulation value of the system decreases with ZnCl2 concentration leading to thermodynamically unstable system. Table 2 shows the viscosity of PSF/ DMF/PEG 200 solution solutions without and with various concentrations of ZnCl2 additive at 25 ± 0.5 °C at a shear rate of 24 s−1. Results reveal that the polymeric solution containing lower concentration of ZnCl2 has lower viscosity (0.1 and 0.3 wt.%) compared to higher ones (1 and 2 wt.%). It is apparent that the solution viscosity increases to around 2 to 3 times by the addition of ZnCl2. Explanations for such observations are available in literature [46,47]. First, additive ZnCl2 and PEG-200 intensely interact with PSF due to their polar and ionic characteristics resulting to more association of polymer chains [46]. Second, there exists solvent (DMF)–salt (ZnCl2) interaction and strong association between Zn+2 and electron donating sulfone group of polymeric network [47]. Similar findings were obtained by Kesting et al. using cellulose acetate in LiCl inorganic additive [48,49], by Bottino et al. using polyvinylidene difluoride (PVDF) in LiCl [50], by Kim et al., using PSF in ZnCl2 [29], by Ahmed et al., and Idris et al., using LiCl and LiBr in polyethersulfone (PES) [51,52]. As a result, exchange rate between solvent and non-solvent slows down leading to delayed demixing or slower phase separation, thereby, forming a dense morphology [28,

Rejection (%)

Fig. 2. (continued).

80

60

MPEGZ-1 MPEGZ-2

40

Error bars: ±5% 20 100

200

300

400

500

600

Molecular weight (Da) Fig. 3. Effect of ZnCl2 from 0 to 2 wt.%, on PSF-PEG blend membranes (a) porosity and contact angle of membranes (b) permeability of membranes (c) rejection of MPEGZ-1 and MPEGZ-2 NF membranes.

29]. By comparing the top view of the membranes prepared with ZnCl2 of different concentration, it is observed that the membrane at lower additive concentration 0.1 wt.% (MPEGZ-0.1) contains small sized pores at a magnification of 1000. With slight increase in additive concentration to 0.3 wt.% (MPEGZ-0.3), the number of pores increased but no consistent change in the size of the pores was observed at the


80 MPEGZ-1

70

MPEGZ-2 C0 : 1000 mg/l, cross flow rate: 50 l/h, ΔP: 690 kPa

60

Error bars: ±5%,

Rejection (%)

same magnification. As the additive concentration increased to 1 and 2 wt.%, the surface of the membranes became very dense. The pores of these membranes obtained were too small to be seen even at the higher magnification of 20,000. The morphological symmetry between the top surfaces of MPEGZ-1 and MPEGZ-2 could be viewed as an additional indication for the selective permeation of lower molecular weight solutes through the membrane. This observation once again supports that zinc chloride acts as a pore inhibitor rather than a pore former and this effect is augmented in the presence of PEG-200 [53]. The salt (ZnCl2) and solvent (DMF) interaction in PSF can also be explained by FTIR measurement of the polymer films [47,54]. The interaction of ZnCl2 with PEG 200 and PSF is established by taking the IR spectra of PSF membranes shown in supporting Fig. 2(c). The membranes with and without ZnCl2 were analyzed in FTIR/HATR mode. The strongest bond in PSF is generally found at 1250 cm−1 and 830 cm− 1, corresponding to sulfonic and C–H groups of polymeric chain [54]. According to Kim et al., the addition of ZnCl2 affects the peaks present in PSF chain at around 1150 cm−1 and 880 cm−1, respectively [29]. In the present study, these peaks are observed at 1151.02 and 873.26 cm−1. With the addition of ZnCl2 at 1 and 2 wt.%, the peak at 1151.02 cm−1 has been shifted to 1151 cm− 1 and 1150.72 cm− 1 (refer Fig. 2c). The corresponding shifts of peak at 873.26 cm− 1 are 873.16 cm− 1 and 873.10 cm−1. The IR spectra support the fact that PSF solution, along with PEG 200 and ZnCl2 additive shows interaction with PSF chains leading to enhancement of viscosity as observed in Table 3 [47,54]. Similar type of observation is reported by Kim et al. [29]. Typical variation of membrane porosity and contact angle with the composition of ZnCl2 is shown in Fig. 3(a). As expected, Fig. 3(a) clearly shows that there is a sharp increase in porosity up to 0.3 wt.% of ZnCl2, then it decreases sharply up to 2 wt.%. As the concentration of ZnCl2 increases, more finger like macrovoids appeared up to 0.3 wt.% as observed from SEM images described earlier and the membrane matrix becomes more porous. Contact angles of the membranes are also reported in the same figure. Up to 0.3 wt.% ZnCl2, contact angle remains almost invariant. Contact angle increases from about 35 to 43° for 1 and 2 wt.% of ZnCl2. As discussed earlier, concentration of ZnCl2 in this range results to denser membranes making them slightly hydrophobic. Variation of membrane permeability with ZnCl2 is a direct consequence of porosity variation as shown in Fig. 3(b). It is found that the membrane prepared using (MPEGZ-0.3) 0.3 wt.% ZnCl2 has the maximum porosity resulting to the highest permeability as shown in Fig. 3(b). The permeability of PSF membrane without any additive (MPEGZ-0) is found to be 2.8 × 10−11 m/Pa·s. The addition of 1 wt.% ZnCl2 (MPEGZ-1) reduces the permeability value to 9.2 × 10− 12 m/Pa·s. A further increase in ZnCl2 concentration (MPEGZ-2) leads to a gradual decrease in the

50 40 30 20 10 0

2

4

6

8

10

12

pH of NaCl solution Fig. 5. Rejection of NaCl at different pH (2 to 11) for MPEGZ-1 and MPEGZ-2 NF membrane.

membrane permeability up to 9.0 × 10−12 m/Pa·s. Permeability and solute rejection indicate the performance of any membranes. The MWCO for different composition of ZnCl2 is found out to be 39 kDa, 43 kDa and 49 kDa for 0, 0.1, and 0.3% by weight, respectively. As

a 80

Permeate flux (l/m2.h) and rejection (%)

60

- Flux, 70 60

-Rejection, Error bars: ±5%

C0 : 1000 mg/l, Operating pH: 11 Cross flow rate: 50 l/h, ΔP: 690 kPa

50 40 30 20 10 0

NaCl

Na2SO4

MgSO4

MgCl2

b

Zetapotential (mV)

-3 -6 -9 -12 -15 MPEGZ-1 MPEGZ-2 Error bars: ±5%

-18 2

4

6

8

10

12

Permeate flux (l/m2.h) and rejection (%)

80

0

- Flux, 70 60

-Rejection, Error bars: ±5%

C0 : 1000 mg/l, Operating pH: 11 Cross flow rate: 50 l/h, ΔP: 690 kPa

50 40 30 20 10 0

NaCl

Na2SO4

MgSO4

MgCl2

pH Fig. 4. Zeta potential of NF membranes as a function of pH.

Fig. 6. Permeate flux and percentage rejection of (a) MPEGZ-1 (b) MPEGZ-2 NF membranes.


corroboration with the observation of overall porosity, permeability and SEM images.

discussed earlier, 1 and 2 wt.% ZnCl2 resulted to denser membrane and hence solute rejection performances of these two membranes were separately determined with lower molecular weight solutes (glucose, PEG 200, 400 and 600) and presented in Fig. 3(c). For 2 wt.% (MPEGZ-2) membrane, 100% rejection is obtained for all the solutes. Whereas, 1 wt.% (MPEGZ-1) membrane, 90% solute rejection occurs for glucose (MW: 180). On increasing the concentration of ZnCl2 from 1 to 2 wt.%, the membrane becomes thicker and denser and MWCO decreases along with membrane permeability. These results corroborate the earlier observation [27,28]. Thus, MPEGZ-1 and MPEGZ-2 are definitely NF membranes. The data and calculations for the skin thickness and total thickness of the membranes without fabric for all the nine membranes are shown in the Table 3. It is observed from Table 3 that the thickness of the active layer and the permeability of the membranes correlate quite well with the rejection data. The skin thickness of the NF membrane MPEGZ-1 was determined to be about 20 μm of a total thickness of 160 μm. Whereas for NF membrane MPEGZ-2, the skin thickness was found to be 25 μm out of a total thickness of 200 μm. The findings are quite consistent with the fact that the former one is more permeable than the later one. The average pore radius is calculated using Eq. (4). The results for average pore radii of membranes at various concentrations of PEG 200 and zinc chloride are presented in Table 3. These values are in direct

80 70

Rejection (%)

3.2.1. Zeta potential: effects of pH The nature and variation of zeta potential with solution pH is greatly reflected by the presence of different ionisable groups present in the membrane material. For example, if the groups are weakly acidic then their dissociation takes place gradually and zeta potentials of membrane are expected to be more negative with increasing pH. If the groups are strongly acidic then their dissociation takes place very quickly and higher zeta potentials can be expected even at low pH. The membrane zeta potential in pH range 2.0 to 11.0 is shown in Fig. 4. From Fig. 4, it is clear that both zinc chloride incorporated PEG blended PSF NF membrane are negatively charged in the whole pH range. Both the membranes have high negative charge at lower and higher pH with a maximum at pH 7. These findings qualitatively identify the electrostatic contribution to rejection of salts and dyes. 3.2.2. Effect of pH on NaCl rejection Rejection of monovalent salt (NaCl) at different pH values is shown in Fig. 5. It can be observed from this figure that at low and high pH,

c 276 kPa

414 kPa

60

25

552 kPa 690 kPa Membrane: MPEGZ-1

Cross flow rate: 50 l/h, C0 : 1000 mg/l,

55

Error bars: ±5%

50

Na2SO4

Na2SO4

20

50 40 30

NaCl

45

15 NaCl

40

20

Membrane: MPEGZ-1 Operating pH: 11 Cross flow rate: 50 l/h C0 : 1000 mg/l

35

10 30 200

0

pH 6.5

pH 2

pH 11

b

400

500

600

10

5

700

Transmembrane pressure (kPa)

d 276 kPa

552 kPa 690 kPa Membrane: MPEGZ-2

Na2SO4

Cross flow rate: 50 l/h, C0 : 1000 mg/l,

55

Error bars: ±5%

50

20

50 40 30

NaCl

15

45 NaCl

40

Na2SO4 Membrane: MPEGZ-2 Operating pH: 11 Cross flow rate: 50 l/h C0 : 1000 mg/l

20 35 10 30 200

0

pH 2

pH 6.5

pH 11

300

400

500

600

700

10

Permeate flux (l/m2.h)

60

414 kPa

Rejection (%)

70

25

60

80

Rejection (%)

300


60

3.2. Characterization of the resulting NF membranes

Rejection (%)

a

61

5

Transmembrane pressure (kPa)

Fig. 7. Rejection and permeate flux profile of electrolytes: (a) Effect of NaCl pH for MPEGZ-1 membrane (b) Effect of NaCl pH for MPEGZ-2 membrane (c) Effect of TMP on rejection of NaCl and Na2SO4 MPEGZ-1 membrane (d) Effect of TMP on rejection of NaCl and Na2SO4 MPEGZ-2 membrane.


Table 4 Rejection of glucose and sucrose by MPEGZ-1 and MPEGZ-2 membranes at 690 kPa, cross flow rate 50 l/h and feed concentration 1000 mg/l. Feed solution (1000 mg/l)

Glucose Sucrose

MW (g/mol)

180 342

Membrane code MPEGZ-1

Membrane code MPEGZ-2

Flux (l/m2·h)

%R

Flux (l/m2·h)

%R

15 15.5

90 100

13 14

100 100

22

105

Rejection (%)

20

crystal violet

congo red

100

18 16

95

chrysodine R

14

congo red

12

90

chrysodine R

10 crystal violet

8

85 Cross flow rate: 50 l/h C0 : 200 mg/l

80

6 4

300

400

500

600

700

Transmembraane pressure (kPa)

b 105

20 18

congo red

100

crystal violet

16 14

chrysodine R

95

12 chrysodine R

10 90 crystal violet

85

8 6

congo red Cross flow rate: 50 l/h, C0 : 200 mg/l

80


3.2.3. Rejection performance for different inorganic electrolyte Rejection and corresponding permeate flux values of different electrolytes at 690 kPa and pH 11 are presented in Fig. 6. Results of MPEGZ-1 and MPEGZ-2 membranes are shown in Figs. 6(a) and (b), respectively. It is observed from both the figures that rejections are in the order of Na2SO4 N NaCl N MgSO4 N MgCl2. At pH 11, both the mem2 branes possess negative charge. For divalent salt Na2SO4, SO− is 4 strongly repelled by electrostatic interaction compared to Cl− in monovalent NaCl. Thus, Na2SO4 shows the maximum rejection. In case of MgCl2, rejection is dominated by Mg+2 due to its higher valency [28, 56]. Since, both the membranes are negatively charged, MgCl2 rejection is minimum. Whatever the rejection (approximately 8%) experienced by MgCl2 is due to repulsion of Cl−. For MgSO4, both Mg+2 and SO−2 4 have same but opposite valency. Thus, repulsion of SO−2 4 and attraction of Mg+2 are almost in the same order but it is stronger than repulsion of Cl− in MgCl2. The permeate flux at 690 kPa is around 19 to 22 l/m2·h for both the membranes and all the electrolytes. Fig. 7 presents the rejection profile of electrolytes with TMP and pH (pH 2, pH 7 and pH 11) for both the NF membranes, MPEGZ-1 and MPEGZ-2. It is observed from Figs. 7(a) and (b) that rejection decreases to about 40% from 43%, when TMP increases from 276 to 690 kPa at pH 2 and pH 11 for MPEGZ-1 membrane. It decreases further to 20% at pH 7. Similar type of observation is obtained for MPEGZ-2 membrane. Rejection decreases to about 42% from 47%, when TMP increases from 276 to 690 kPa at pH 2 and 11 for MPEGZ-2 membrane. It decreases further to 23% at pH 7. With increase in TMP, convective flux through the NF membrane dominates over diffusive and electrostatic repulsion resulting to permeation of more salt and reduction of rejection. At extreme pH condition (2 and 11), the membrane possesses the maximum negative charge. Thus, at these pH values electrostatic repulsion is more compared to pH 7, hence the rejection is lowest at pH 7. The effects of TMP on NaCl and Na2SO4 at pH 11 for both the membranes are shown in Figs. 7(c) and (d). As explained earlier, both the electrolytes show a reduction of rejection with TMP. This is due to dominance of convective transport of the solutes at higher TMP. However, − Na2SO4 having higher charge (SO−2 4 ) compared to Cl in NaCl, Na2SO4 is rejected more due to better electrostatic repulsion with respect to negatively charged membrane. For MPEGZ-1 membrane, rejection of Na2SO4 decreases from 55% to 52% as TMP increases from 276 to 690 kPa. However, the permeate flux increases from 16 to 23 l/m2·h due to enhanced driving force. In case of NaCl, rejection decreases from 43 to 40% and permeate flux increases from 12 to 21 l/m2·h over the same range of TMP. For MPEGZ-2 membrane, variation of rejection is from 57 to 54% for Na2SO4 and that for NaCl is from 47 to 41% as TMP

a


rejection is more. NaCl being a strong electrolyte, Cl− ions get rejected by the negatively charged membrane at higher pH. To maintain the electroneutrality, same amount of Na+ is also rejected (Donnan exclusion) leading to higher rejection of NaCl. Similarly, rejection behavior is observed at lower pH conditions. Since, membrane zeta potential is close to zero at pH 7, charge–charge interaction is also minimal leading to lower rejection of NaCl. At pH 2, 40 and 41% rejections of NaCl are obtained for MPEGZ-1 and MPEGZ-2 membranes. These values are 40% and 42% at pH 11 and 20% and 23% at pH 7. Similar V-shaped curve of rejection of monovalent salt at different pH has been reported by a commercial NF membrane by Luo and Wan et al. [55].

Rejection (%)

62

4 2

300

400

500

600

700

Transmembrane pressure (kPa) Fig. 8. Rejection and permeate flux of different dyes congo red, chrysoidine R and crystal violet, at varying TMP at their natural pH (a) MPEGZ-1 (b) MPEGZ-2.

increases from 276 to 690 kPa. Corresponding increase in permeate flux is from 15 to 20 l/m2·h for Na2SO4 and 12 to 18 l/m2·h for NaCl.

3.2.4. Rejection performance of sucrose and glucose Table 4 shows the rejection of lower molecular weight solutes like glucose and sucrose by zinc chloride incorporated NF membranes. Glucose and sucrose being the neutral solutes, transport through the NF is controlled by the pore size of the membrane and the size of the solute itself. The rejection to this kind of solute is determined by “sieving effect” not by “Donnan effect”. The membranes MPEGZ-1 and MPEGZ-2 showed a very good performance in terms of glucose rejection of 90% and 100%, respectively. Sucrose rejection by both the membrane was found to be 100%.

Table 5 Dye rejection performances by MPEGZ-1 and MPEGZ-2 membranes at their natural solution pH, at 690 kPa, cross flow rate 50 l/h and feed concentration 200 mg/l. Dye solution (feed)

MW (g/mol)

pKa

Solution pH

Chrysoidine R Crystal violet Congo red

262.74 407.98 696.65

6.0 [57] 9.4 [58] 4.1 [58]

5.1 6.0 7.5


3.2.5. Rejection and flux for different dyes NF membranes are also used for the separation of different dyes. These are congo red (CR), crystal violet (CV) and chrysoidine R (CRD). Rejection behavior and permeate flux of these dyes at different TMP is presented in Fig. 8. It is apparent that congo red and crystal violet have molecular weights above MWCO of the membranes (as reported in Table 3). Thus, due to size exclusion itself, the rejection of these dyes are 99% at 690 kPa for both the NF membranes. On the other hand, CRD has the lowest molecular weight 262.74. With CRD, the solution pH is 4.2, whereas, pKa of CRD is 6.0. Thus, dye is positively charged (protonated) at this pH. Membrane being negatively charged at pH 4, more dyes are attracted towards the membrane surface and for both the MPEGZ-1 and MPEGZ-2 membranes, rejection was low i.e., 95% at TMP of 690 kPa. Permeate (dye) flux of two membranes are also presented in the same figures. pKa values of all three dyes [57,58] are presented in Table 5. From pKa values, the dye CR and the membrane MPEGZ-1 are both negatively charged at solution pH 7.5. Thus, electrostatic interaction repels the dye particles away from the membrane, thereby reducing concentration polarization and enhancing the permeate flux. Thus, at 690 kPa TMP, permeate flux of CR was the maximum. On the other hand, CV is positively charged (as solution pH b pKa) and it is attracted towards the membrane leading to significant polarization and its flux is the lowest. At 690 kPa TMP, flux of CV is 11, whereas, that of CR was 20 l/m2·h for MPEGZ-1 membrane. In case of CRD, solution pH (4.2) is slightly less than pKa (6.0) value. Thus, it is positively charged but its charge–charge interaction is less, leading to less concentration polarization compared to CV. The permeate flux in this case is 15 l/m2·h for MPEGZ-1 membrane. In case of MPEGZ-2 membrane, similar trend is observed but the permeate flux values are slightly less compared to MPEGZ-1 membrane due to its denser morphology.

63

3.3. Comparison with commercial and lab developed nanofiltration membranes with different additives 3.3.1. Comparison with commercial nanofiltration membranes The separation performance of MPEGZ-1 and MPEGZ-2 membranes with other commercial membranes is presented Table 6. The membranes NF-1, NF-2, NF-3 and NF-5 were thin film composite membranes with polyamide skin on polyester–polysulfone support (having three layers) with a thickness of 150 to 165 μm [59]. On the other hand, MPEGZ-1 and MPEGZ-2 membranes are not thin film membranes and have comparable performance with NF-1 and NF-3 for rejection of NaCl which are 48% and 44%, respectively. NTR-7410, NTR-7450 and NTR-7250 (Nitto Denko) were thin film composite membranes made up of sulfonated polyethersulfone (SPES) and polyvinyl alcohol support [60]. NF-40 (Film Tec) was also a thin film composite (TFC) membrane made up of aromatic polyamide (PA) [60]. Among these, the performances of NTR-7250, NTR-7450 and NF-40 have comparable NaCl rejection with MPEGZ-1 and MPEGZ-2 membranes. NTR-7450 and NF-40 showed equivalent rejection of neutral solutes like sucrose and glucose with MPEGZ-1 and MPEGZ-2 membranes. On the other hand, NTR-7410 exhibited poor (15%) performance of NaCl compared to MPEGZ-1 and MPEGZ-2 membranes. NF-PES-10 (Microdyne Nadir Filtration) and OPMN-P70 (Vladipor) were also TFC membrane with PA skin on PES and polypropylene (PP) support, respectively [61]. Out of these two, NF-PES-10 showed poor and OPMN-P70 showed comparable performance in terms of NaCl rejection with respect to MPEGZ-1 and MPEGZ-2 membranes. In case of rejection of divalent salt (sodium sulfate), the developed membranes showed comparable performance with respect to NTR-7410 and NF-PES-10. For magnesium chloride, the performance of developed membrane was better than that of NTR-7410. NTR-7410 and NTR-7450 membranes

Table 6 Comparison of separation performance of MPEGZ-1 and MPEGZ-2 membranes with commercial nanofiltration membranes [59–64]. Membrane types

Commercial thin film composite NF membranes (TFC) NF-1 [59] NF–2 (NE-90) [59] NF-3 [59] NF-5 [59] NTR-7410 [60] NTR-7450 [60] NF40 [60] NTR-7250 [60] NF-PES-10 [61] OPMN-P70 [61]

Dye stuff R (%)/ MW

NaCl ΔP (kPa)/C0 (mg/l)/Cp (mg/l)/R (%)

Na2SO4 ΔP (kPa)/C0 (mg/l)/Cp (mg/l)/R (%)

MgSO4 ΔP (kPa)/C0 (mg/l)/Cp (mg/l)/R (%)

MgCl2 ΔP (kPa)/C0 (mg/l)/Cp (mg/l)/R (%)

Sucrose/ glucose R (%)

pH

– – – – 98/300 100/300 – – – –

500/2000/1040/48 500/2000/320/84 500/2000/1120/44 500/2000/260/87 1000/5000/4250/15 1000/5000/2450/51 2000/2000/1100/45 2000/2000/1000/50 –/5/4.75–4.25/5–15 –/1.5/N0.75/N50

– – – – 1000/5000/2250/55 1000/5000/400/92 1390/100/5/95 – –/10/7–4–/30–60 –

– – – – 1000/5000/4550/9 1000/5000/3400/32 2000/2000/60/97 2000/2000/40/98 – –/2/0.06/97

– – – – 1000/5000/4800/4 1000/5000/4350/13 – – – –

– – – – 5/– 36/– 98/– 98/99 – –

– – – – 1–13 1–13 2–11 2–8 0–14 2–11

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

276–690/1000 /470–430/53–57 276–690/1000 /470–430/53–57

276–690/1000/ /900/10 276–690/1000 /900/10

276–690/ 1000/920/8 276–690/ 1000//920/8

100/90

2–11

100/100

2–11

NF membrane with different additives and coating modification methods PP membrane with CS coatings [62] – 200/3500/2170/38 700/3500/2975/15 PSF blended with PSAB, mPSAB membranes [24] – 200/3500/700/75 800/3500/2975/15 PPES/CS coating and glutaldehyde concentration – 400/1000/ variation [63] 880–650/12–35 600/1000/ 900–700/10–30 800/1000/ 910–850/7–15 GO dispersed PSF membrane [64] – 400–1000/2000 /1100–840/45–58 400–1000/1000 /650–580/35–42 Present work MPEGZ-1 MPEGZ-2

95–99/ 262–696 95–99/ 262–696

276–690/1000 /600–530/40–47 276–690/1000 /600–530/40–47

64


performed equally well with MPEGZ-1 and MPEGZ-2 as far as dye separation is considered. 3.3.2. Comparison with lab developed nanofiltration membranes with additives/coating Comparison of the developed membranes with various labdeveloped membranes with additives/coating is presented in Table 6. It is observed from this table that MPEGZ-1 and MPEGZ-2 fare better than PP membrane with CS coating [62], PSF blended with PSAB and mPSAB membranes at higher TMP [24], PPES/CS coating and different concentration of glutaraldehyde as cross linking agent [63] and GO dispersed PSF membranes [64] as far as the rejection of NaCl is concerned. 4. Conclusion The major conclusions drawn from this study are summarized below: 1) Addition of 1 and 2 wt.% of ZnCl2 to 18 wt.% of PSF and 2 wt.% PEG 200 in DMF resulted to nanofiltration (NF) membranes, MPEGZ-1 and MPEGZ-2. 2) The skin thickness of NF membranes was 20–25 nm. 3) Molecular weight cut off of both the membranes was less than 180. 4) Contact angles of MPEGZ-1 and MPEGZ-2 were 62° and 58°, respectively. 5) Permeability of these membranes was about 3 × 10−12 m/Pa·s. 6) Zeta potentials of MPEGZ-1 membrane were −17 and −9 mV at pH 2 and 11, respectively. These values for MPEGZ-2 membrane were −11 mV and −10 mV. 7) NaCl rejections for MPEGZ-1 and MPEGZ-2 are about 40% and 43% for pH 2 and 11, at 690 kPa and 1000 mg/l. 8) Rejection of NaCl decreases from 47% to 44% and that for Na2SO4 is from 57% to 53% at pH 11 when TMP increases from 276 to 690 kPa for MPEGZ-2 membrane. The corresponding increase in permeate flux is from 12 to 18 l/m2·h for NaCl and 16 to 20 l/m2·h in case of Na2SO4. 9) More than 95% of Chrysoidine R and more than 98% of crystal violet and congo red were rejected for all operating TMP for both membranes. Nomenclature C0 concentration of feed (mg/l) Cp concentration of permeate (mg/l) R rejection, (%) ΔP operating pressure (kPa) MW molecular weight

Acknowledgement 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. 25-072012. 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] V. Vatanpour, S.S. Madaeni, R. Moradian, S. Zinadini, B. Astinchap, Fabrication and characterization of novel antifouling nanofiltration membrane prepared from oxidized multiwalled carbon nanotube/polyethersulfone nanocomposite, J. Membr. Sci. 375 (2011) 284–294. [2] L.E.S. Brink, S.J.G. Elbers, T. Robbertsen, P. Both, The anti-fouling action of polymers preadsorbed on ultrafiltration and microfiltration membranes, J. Membr. Sci. 76 (1993) 281–291. [3] Y. Su, W. Cheng, C. Li, Z. Jiang, Preparation of antifouling ultrafiltration membranes with poly(ethylene glycol)-graft-polyacrylonitrile copolymers, J. Membr. Sci. 329 (2009) 246–252.

[4] M. Ulbricht, K. Richau, H. Kamusewitz, Chemically and morphologically defined ultrafiltration membrane surfaces prepared by heterogeneous photo-initiated graft polymerization, Colloids Surf. A Physicochem. Eng. Asp. 138 (1998) 353–366. [5] T.D. Tran, S. Mori, M. Suzuki, Plasma modification of polyacrylonitrile ultrafiltration membrane, Thin Solid Films 515 (9) (2007) 4148–4152. [6] J.H. Kim, K.H. Lee, Effect of PEG additive on membrane formation by phase inversion, J. Membr. Sci. 138 (1998) 153–163. [7] S.R. Panda, S. De, Role of polyethylene glycol with different solvents for tailor-made polysulfone membranes, J. Polym. Res. 20 (2013) 179–195. [8] Y. Liub, G.H. Koopsa, H. Strathmanna, Characterization of morphology controlled polyethersulfone hollow fiber membranes by the addition of polyethylene glycol to the solution and bore liquid solution, J. Membr. Sci. 223 (2003) 187–199. [9] H. Matsuyama, T. Maki, M. Teramoto, K. Kobayashi, Effect of PVP additive on porous polysulfone membrane formation by immersion precipitation method, Sep. Sci. Technol. 38 (14) (2003) 3449–3458. [10] J. Marchese, M. Ponce, N.A. Ochoa, P. Pradanos, L. Palacio, A. Hernandez, Fouling behaviour of polyethersulfone UF membranes made with different PVP, J. Membr. Sci. 211 (2003) 1–11. [11] M.J. Han, S.T. Nam, Thermodynamic and rheological variation in polysulfone solution by PVP and its effect in the preparation of phase inversion membrane, J. Membr. Sci. 202 (2002) 55–61. [12] B.K. Chaturvedi, A.K. Ghosh, V. Ramachandhran, M.K. Trivedi, M.S. Hanra, B.M. Misra, Preparation, characterization and performance of polyethersulfone ultrafiltration membranes, Desalination 133 (2001) 31–40. [13] C. Barth, M.C. Gonclalves, A.T.N. Pires, J. Roeder, B.A. Wolf, Asymmetric PSF and PES membranes: effects of thermodynamic conditions during formation on their performance, J. Membr. Sci. 169 (2000) 287–299. [14] T.B. Swinyard, A.J. Barrie, Phase separation in nonsolvent/DMF/PES and nonsolvent/ DMF/PSF systems, Br. Polym. J. 20 (1988) 317–321. [15] M.J. Han, D. Bhattacharyya, Morphology and transport study of phase inversion polysulfone memrbanes, Chem. Eng. Commun. 128 (1994) 197–209. [16] I.C. Kim, K.H. Lee, Effect of poly(ethylene glycol) 200 on the formation of a polyetherimide asymmetric membrane and its performance in aqueous solvent mixture permeation, J. Membr. Sci. 230 (2004) 183–188. [17] G. Hagmeyer, R. Gimbel, Modelling the rejection of nanofiltration membranes using zeta potential measurements, Sep. Purif. Technol. 15 (1999) 19–30. [18] Y. Xu, R.E. Lebrun, Comparison of nanofiltration properties of two membranes using electrolyte and non-electrolyte solutes, Desalination 122 (1999) 95–106. [19] A. Yaroschuck, E. Staude, Charged membranes for low pressure reverse osmosis properties and applications, Desalination 86 (1992) 115–134. [20] Y. Garba, S. Taha, N. Gondrexon, G. Dorange, Ion transport modelling through nanofiltration membranes, J. Membr. Sci. 160 (1999) 187–200. [21] J.M.M. Peeters, M.H.V. Mulder, H. Strathmann, Streaming potential measurements as a characterization method for nanofiltration membranes, Colloids Surf. A 150 (1999) 247–259. [22] M. Padaki, A.M. Isloor, G. Belavadi, K.N. Prabhu, Preparation, characterization and performance study of poly(isobutylene-alt-maleic anhydride) [PIAM] and polysulfone [PSf] composite membranes before and after alkali treatment, Ind. Eng. Chem. Res. 50 (2011) 6528–6534. [23] B.M. Ganesh, A.M. Isloor, M. Padaki, Preparation and characterization of polysulfone and modified poly isobutylene-alt-maleic anhydride blend NF membrane, Desalination 287 (2012) 103–108. [24] M. Padaki, A.M. Isloor, A.F. Ismail, M.S. Abdullah, Synthesis, characterization and desalination study of novel PSAB and mPSAB blend membranes with polysulfone (PSf), Desalination 295 (2012) 35–42. [25] M. Padaki, A.M. Isloor, R. Kumar, A.F. Ismail, T. Matsuura, Synthesis, characterization and desalination study of composite NF membranes of novel poly [(4aminophenyl)sulfonylbutanediamide (PASB) and methyalated poly[(4aminophenyl)sulfonyl]butanediamide (mPASB) with polysulfone, J. Membr. Sci. 428 (2013) 489–497. [26] P.S. Zhong, N. Widjojo, T.S. Chung, M. Weber, C. Maletzko, Positively charged nanofiltration (NF) membranes via UV grafting on sulfonated polyphenylenesulfone (sPPSU) for effective removal of textile dyes from wastewater, J. Membr. Sci. 417 (2012) 52–60. [27] T.A. Tweddle, O. Kutowy, W.L. Thayer, S. Sourirajan, Polysulfone ultrafiltration membranes, Ind. Eng. Chem. Prod. Res. Dev. 22 (1983) 320. [28] J.S. Park, S.K. Kim, K. Lee, The effect of ZnCl2 on formation of asymmetric PEI membranes by phase inversion process, J. Ind. Eng. Chem. 6 (2) (2000) 93–99. [29] S.R. Kim, K.H. Lee, M.S. Jhon, The effect of ZnCl2 on the formation of polysulfone membrane, J. Membr. Sci. 119 (1996) 59–64. [30] C. Ba, C. Ba, Design of advanced reverse osmosis and nanofiltration membranes for water purification, Materials Science and Engineering, University of Illinois, Urbana-Champaign, 2010. [31] A. Bottino, G. Capannelli, S. Munari, A. Turturro, High performance ultrafiltration membranes cast from LiCl-solutiond solutions, Desalination 68 (1988) 167–177. [32] E. Yuliwati, A.F. Ismail, Effect of additives concentration on the surface properties and performance of PVDF ultrafiltration membranes for refinery produced wastewater treatment, Desalination 273 (2011) 226–234. [33] E. Fontananova, J.C. Jansen, A. Cristiano, E. Curcio, E. Drioli, Effect of additives in the casting solution on the formation of PVDF membranes, Desalination 192 (2006) 190–197. [34] A. Idris, I. Ahmed, M. Misran, Novel high performance hollow fiber ultra filtration membranes spun from LiBr solutiond solutions, Desalination 249 (2009) 541–548. [35] I. Ahmed, A. Idris, N.F.C. Pa, Novel method of synthesizing poly(ether sulfone) membranes containing two solvents and a lithium chloride additive and their performance, J. Appl. Polym. Sci. 115 (2010) 1428–1437.

S.R. Panda, S. De / Desalination 347 (2014) 52–65 [36] A. Rahimpour, S.S. Madaeni, Y. Mansourpanah, Fabrication of polyethersulfone (PES) membranes with nano-porous surface using potassium perchlorate (KClO4) as an additive in the casting solution, Desalination 258 (2010) 79–86. [37] C.D. Vecitis, J.C. Mierzwa, US Patent number WO 2013/142141 A1. [38] S. Singh, K.C. Khulbe, T. Matsuura, P. Ramamurthy, Membrane characterization by solute transport and atomic force microscopy, J. Membr. Sci. 142 (1998) 111. [39] B. Sarkar, S. De, S. DasGupta, Pulsed electric field enhanced ultrafiltration of synthetic and fruit juice, Sep. Purif. Technol. 63 (2008) 582–591. [40] S. Chhaya, G.C. Mondal, S. De Majumdar, Clarifications of stevia extract using cross flow ultrafiltration and concentration by nanofiltration, Sep. Purif. Technol. 89 (2012) 125–134. [41] A. Idris, L.K. Yet, The effect of different molecular weight PEG additives on cellulose acetate asymmetric dialysis membrane performance, J. Membr. Sci. 280 (2006) 920–927. [42] Z.G. Wang, Z.K. Xu, L.S. Wan, Modulation the morphologies and performance of polyacrylonitrile-based asymmetric membranes containing reactive groups: effect of non-solvents in the dope solution, J. Membr. Sci. 278 (2006) 447–456. [43] J. Chen, J. Li, X. Zhan, X. Han, C. Chen, Effect of PEG additives on properties and morphologies of polyetherimide membranes prepared by phase inversion, Front. Chem. Eng. Chin. 4 (3) (2010) 300–306. [44] J.H. Choi, J. Jegal, W.N. Kim, Fabrication and characterization of multi-walled carbon nanotubes/polymer blend membranes, J. Membr. Sci. 284 (2006) 406–415. [45] S.R. Panda, S. De, Effects of polymer molecular weight, concentration, and role of polyethylene glycol as additive on polyacrylonitrile homopolymer membranes, Polym. Eng. Sci. (2013), http://dx.doi.org/10.1002/pen.23792. [46] I. Ahmed, A. Idris, M. Yusof Noordin, R. Rajput, High performance ultrafiltration membranes prepared by the application of modified microwave irradiation technique, Ind. Eng. Chem. Res. 50 (2011) 2272–2283. [47] M.A. Phadke, D.A. Musale, S.S. Kulkarni, S.K. Karode, Poly(acrylonitrile) ultrafiltration membranes. I. Polymer–salt–solvent interactions, J. Polym. Sci. B Polym. Phys. 43 (2005) 2061–2073. [48] R.E. Kesting, Synthetic Polymeric Membranes, John Wiley & Sons, New York, 1985. [49] R.E. Kesting, Phase inversion membranes, ACS Symposium Series 269, American Chemical Society, Washington, DC, 1985, pp. 9131–9164. [50] A. Bottino, G. Capanelli, S. Munari, A. Turturro, High performance ultrafiltration membranes cast from LiCl solution solutions, Desalination 68 (1988) 167–177.

65

[51] I. Ahmed, A. Idris, N.F.C. Pa, Novel method of synthesizing polyethersulfone membrane containing two solvents and lithium chloride additive and its performance, J. Appl. Polym. Sci. 115 (2010) 1428–1437. [52] A. Idris, I. Ahmed, Viscosity behavior of microwave-heated and conventionally heated poly(ether sulfone)/dimethylformamide/lithium bromide polymer solutions, J. Appl. Polym. Sci. 108 (2008) 302–307. [53] R.Y.M. Huang, X. Pheng, Studies on solvent evaporation and polymer precipitation pertinent to the formation of asymmetric polyetherimide membranes, J. Appl. Polym. Sci. 57 (1995) 613. [54] Y.Q. Song, J.S. Sheng, M. Wei, X.B. Yuan, Surface modification of polysulfone membranes by low-temperature plasma-graft poly(ethylene glycol) onto polysulfone membranes, J. Appl. Polym. Sci. 78 (2000) 979–985. [55] J. Luo, Y. Wan, Effects of pH and salt on nanofiltration—a critical review, J. Membr. Sci. 438 (2013) 18–28. [56] M.N. Katarzyna, Synthesis and properties of polysulfone membranes, Desalination 71 (1989) 83. [57] M.K. Purkait, S. DasGupta, S. De, Performance of TX-100 and TX-114 for the separation of chrysoidine dye using cloud point extraction, J. Hazard. Mater. B137 (2006) 827–835. [58] R.W. Sabnis, Handbook of Acid Base Indicators, CRC press, Taylor and Francis, 2008. [59] A. Rahimpour, B. Rajaeian, A. Hosienzadeh, S.S. Madaeni, F. Ghoreishi, Treatment of oily wastewater produced by washing of gasoline reserving tanks using self-made and commercial nanofiltration membranes, Desalination 265 (2011) 190–198. [60] M. Nystrom, L. Kaipia, S. Luque, Fouling and retention of nanofiltration membranes, J. Membr. Sci. 98 (1995) 249–262. [61] M. Manttari, A. Pihlajamaki, M. Nystrom, Effect of pH on hydrophilicity and charge and their effect on the filtration efficiency of NF membranes at different pH, J. Membr. Sci. 280 (2006) 311–320. [62] M. Padaki, A.M. Isloor, J. Fernandes, K.N. Prabhu, New polypropylene supported chitosan NF-membrane for desalination application, Desalination 280 (2011) 419–423. [63] S.S. Shenvi, S.A. Rashid, A.F. Ismail, M.A. Kassim, A.M. Isloor, Preparation and characterization of poly(1,4-phenylene ether ether sulfone) PPEES/chitosan composite nanofiltration membrane, Desalination 315 (2011) 135–141. [64] B.M. Ganesh, A.M. Isloor, A.F. Ismail, Enhanced hydrophilicity and salt rejection study of graphene oxide-polysulfone mixed matrix membrane, Desalination 313 (2013) 199–207.