Ph

SSa

b

c

a

ARR2AA

KEEMBHSF

1

tnEfrgomfiotmn

iT

0d

Journal of Membrane Science 390– 391 (2012) 235– 242

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science

jo u rn al hom epa ge: www.elsev ier .com/ locate /memsci

reparation and characterization of surface modified electrospun membranes forigher filtration flux

atinderpal Kaura,b,∗, Dipak Ranaa, Takeshi Matsuuraa, Subramanian Sundarrajanb,eeram Ramakrishnab,c

Industrial Membrane Research Institute, Department of Chemical and Biological Engineering, University of Ottawa, 161 Louis Pasteur St., Ottawa, ON, K1N 6N5, CanadaNational University of Singapore, Faculty of Engineering, HEM Laboratory, 2 Engineering Drive 3, Singapore 117576, SingaporeKing Saud University, Riyadh 11451, Saudi Arabia

r t i c l e i n f o

rticle history:eceived 29 June 2011eceived in revised form3 November 2011ccepted 25 November 2011vailable online 6 December 2011

eywords:lectrospinning

a b s t r a c t

Popular polymers such as polysulfone and poly(vinylidene fluoride) (PVDF) when electrospun into amembrane, have a much higher contact angle and hence are more hydrophobic when compared to thevirgin polymeric material. In direct liquid penetration it is more beneficial if the membrane is hydrophilicso that the flux is not compromised and has less tendency to foul. Hence, this research is focused on gener-ating a highly hydrophilic electrospun membrane (EM) based on PVDF material by blending this polymerwith several different types of surface modifying macromolecules (SMMs) prepared from urethane pre-polymer with poly(ethylene glycol)s (PEGs) of various average molecular weights (400, 600, and 1000 Da)and poly(propylene glycol) of average molecular weights 3500 and 425 Da. One of the SMMs, with PEG

◦

lectrospun membranesicrofiltration

lendingydrophilicurface modifying macromoleculesiber

1000 Da, had a significant impact on the hydrophilic nature (0 SCA with water) of the EM as compared tothe blend casted membrane. This could possibly be due to the orientation of the SMMs hydrophilic groupsadopted during electrospinning on the surface, whereas they are either encapsulated or submerged inother SMMs. The water flux at a given pressure of blended EM was higher than the non-blended elec-trospun PVDF membrane. This study highlights the potential benefits of this new hydrophilic polymericmaterial in the membrane field, which can achieve high-flux rates at low pressure.

© 2011 Elsevier B.V. All rights reserved.

. Introduction

Polymer nanofibers have found a broad range of applica-ions in the fields of catalysis, electronics, protective clothing,anocomposites, filtration, bioengineering and biotechnology [1].lectrospinning is one of the techniques that has been widely usedor the fabrication of nanofibers from various synthetic and natu-al polymers. Electrospinning is far from a new technique since itoes all the way back to 1902 [2]. It is a simple technique basedn the principal of subjecting a high voltage (5–30 kV) on a poly-er solution/melt which drives the solution to be stretched into

bers. The fibers tangle with one another before they are collectedn a grounded plate or drum. The fibers accumulate over time on

he collector and eventually a membrane is formed. The entangle-

ents of the fibers in the membrane give rise to an interconnectedon-woven structure (making it highly porous) which can be

∗ Corresponding author at: National University of Singapore, Faculty of Engineer-ng, HEM Laboratory, 2 Engineering Drive 3, Singapore 117576, Singapore.el.: +65 6516 4272.

E-mail address: satinpal@gmail.com (S. Kaur).

376-7388/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.memsci.2011.11.045

useful for applications such as filtration, tissue scaffolds andimplant coating films. In addition, the fiber size can be optimizedto sub-micron to nano-meter range, thus giving rise to a largesurface area to volume ratio. In the last decade there has beenintense research in exploring the benefits of electrospun media inseparation technology with new and exciting applications beingdiscovered and realized [3].

When popular polymeric membrane materials such as polysul-fone and poly(vinylidene fluoride) (PVDF) are electrospun, theyexhibit high contact angles of ∼130–140◦ which are higher thanconventional membranes [4–6]. This hydrophobic nature is highlyundesirable for the pressure driven membrane processes in thewater treatment applications. For such applications, membraneswith hydrophilic surfaces are preferred, as a hydrophilic surface isless susceptible to membrane fouling. Also, they tend to increasethe flux and rejection. Fouling is a phenomenon where solutesor particles deposit onto a membrane surface or into membranepores and hence membrane performance is degraded. It is a major

obstacle to the widespread use of membrane technology since itis a major cause for flux decline [7]. Chemicals such as amines,polyamines, polyelectrolytes, and many more have been used toinhibit the fouling. Various solvents such as acids and alcohols have

236 S. Kaur et al. / Journal of Membrane Science 390– 391 (2012) 235– 242

Table 1Preparation composition of the SMMs.

SMM MDI, g, in 50 mL DMAc PPG, g, in 100 mL DMAc PEG, g, in 50 mL DMAc

aoatAw

ohpmpwwomoh

Scasteaflem

2

2

S(M>bS6pCalS

2

tw

SMM-1000 (MDI–PPG3500–PEG-1000) 7.5 (0.03 mol)

SMM-600 (MDI–PPG425–PEG-600) 7.5 (0.03 mol)

SMM-400 (MDI–PPG425–PEG-400) 7.5 (0.03 mol)

lso been used to modify the membrane surfaces. Various method-logies (such as blending, radiation or chemical grafting, coating,nd chemical vapor deposition) have been adopted in the litera-ure to make these electrospun membranes more hydrophilic [8].mong them, blending is one of the easiest and most convenientays [9,10].

Therefore, the present study is focused on the fabricationf hydrophilic EM by blending technique. Hence, blending of aydrophilic surface modifying macromolecule (SMM) to the PVDFolymeric solution before electrospinning was carried out. Surfaceodifying macromolecules (SMMs) based on polyurethane were

repared from the synthesis of bis(p-phenyl isocyanate) (MDI)ith poly(ethylene glycol)s (PEGs) of number average moleculareights (400, 600, and 1000 Da) and poly(propylene glycol)s (PPGs)

f number average molecular weights (425 and 3500 Da). Differentolecular weights of PPG were selected to investigate the influence

f the spacer length in polyurethane on the surface properties andydrophilicity.

In comparison, membranes were also prepared by blending ofMM and PVDF solutions using the phase inversion technique. Theomparison allows us to gain an insight into the influence of the twodopted techniques (electrospinning and phase inversion) on theurface properties. During phase inversion, the SMMs are supposedo migrate to the membrane surface [7], which will have three ben-fits: (1) an asymmetric structure of the membrane is achieved; (2)

more hydrophilic surface is achieved; (3) less fouling and higherux are achieved. Hitherto, no reports are available regarding theffects of SMMs on the surface properties of electrospun nanofiberembranes.

. Experimental

.1. Materials

Acetone (Chromasolv grade for HPLC, >99.9% purity,igma–Aldrich Company, St. Louis, MO), N,N-dimethylacetamideDMAc, anhydrous, 99.8% purity, Aldrich Chemical Company, Inc.,

ilwaukee, WI), tetrahydrofuran (THF, Chromasolv grade for HPLC,99.9%, Sigma–Aldrich Company, St. Louis, MO), 4,4′-methyleneis(phenyl isocyanate) (MDI, 98% purity, Sigma–Aldrich, Inc.,t. Louis, MO, USA), poly(ethylene glycol) (PEG, typical Mn 400,00, and 1000 Da, Sigma Chemical Company, St. Louis, MO),oly(propylene glycol) (PPG, typical Mn 425, and 3500 Da, Sigmahemical Company, St. Louis, MO) were purchased and useds received. Poly(vinylidene fluoride) (PVDF, average molecu-ar weight 4.41 × 105) was purchased from Arkema Singapore,ingapore.

.2. Preparation of surface modifying macromolecules

The SMMs were synthesized by a two-step solution polymeriza-ion method. To eliminate the effects of moisture, all glass-waresere dried overnight at 120 ◦C. The first polymerization step was

Fig. 1. Chemical stru

70 (0.02 mol) 20 (0.02 mol)8.5 (0.02 mol) 12 (0.02 mol)8.5 (0.02 mol) 8 (0.02 mol)

conducted with a predetermined composition to form urethanepre-polymer. To a solution of vacuum distilled methylene bis(p-phenyl isocyanate) (MDI, 0.03 mol) in 50 mL of degassed DMAcwas added 0.02 mol of degassed PPG (Mn, either 425 or 3500 Da)in 100 mL of degassed DMAc. The mixture was stirred for 3 h at48–50 ◦C. To this solution, 0.02 mol of PEG (either Mn, 400, 600or 1000 Da) dissolved in 50 mL of degassed DMAc was furtheradded drop-wise and the solution was stirred for 24 h at 48–50 ◦C.The solution was then added drop-wise into a 4 L beaker filledwith distilled water in 24 h under vigorous stirring to precipi-tate the SMM. Depending on the molecular weight of PEG, theSMMs so prepared were called, respectively, SMM-400, SMM-600 and SMM-1000. It should be noted that PPG of Mn 425 Dawas used to synthesize SMM-400 and SMM-600 while PPG ofMn 3500 was used to synthesize SMM-1000 (see Table 1). TheSMM-1000 was gel like, while SMM-400 and SMM-600 were elas-tomeric. All SMMs were cut into smaller pieces and dried inan air circulation oven at 50 ◦C until the weight became con-stant. The molar ratio of monomers used in the SMMs synthesiswas constant at MDI:PPG:PEG = 3:2:2. Table 1 summarizes thenumber of moles and weights of the reactants employed tosynthesize the various SMMs. All SMMs have a common namewhich is poly(4,4′-diphenylenemethylene propylene-urethane)-co-poly(4,4′-diphenylenemethylene ethylene-urethane) with bothends capped by PEG. The chemical structure of the SMM is reflectedin Fig. 1.

2.3. SMM characterization

The glass transition and melting temperature of the variousSMM additives were characterized by using a differential scanningcalorimeter (DSC equipped with a universal analysis 2000 programDSC Q1000, TA Instruments, New Castle, DE). The SMM sample wasannealed at 260 ◦C for 10 min and then quenched to −50 ◦C, andscanned at a heating rate of 10 ◦C/min. The molecular weight, num-ber average molecular weight (Mn) and weight average molecularweight (Mw), of the synthesized SMMs were measured by gel per-meation chromatography (GPC) using a Waters model 410 (Milford,MA) equipped with Waters 410 refractive index detector. Threeultra-styragel columns (103, 104, and 106 A) were used at roomtemperature with tetrahydrofuran (THF) as the mobile phase. TheSMM molecular weight was calculated using the universal cali-bration curve provided with the Millennium 32 software for dataacquisition.

2.4. Preparation of electrospun membranes (EMs)

PVDF solution of 20% (w/v) concentration was prepared in amixture of DMAc and acetone at a ratio of 2:3. A syringe pump

(74900 series, Cole-Parmer Instrument Company, Vernon Hills, IL)was utilized to supply the polymer solution at a constant flowrate of 4 mL/h during electrospinning. A voltage of 15 kV wasapplied by a transformer (DW-P503-1C, Beijing Shining Technical &

cture of SMM.

S. Kaur et al. / Journal of Membrane Science 390– 391 (2012) 235– 242 237

Table 2Characteristics of different SMMs.

Polymer m n p q Mn (kDa) Mw (kDa) Tg (◦C) Tg (◦C)Onset Midpoint

8

3

2

CBTTtfboSTPbwt

2i

(aetttdrwPSra

2

bHTodaev

Qmwb1eCo

tvi

SMM-400 9.59 7.02 4.79 8.6SMM-600 10.07 7.02 5.03 13.2SMM-1000 3.96 60.03 1.98 22.3

ommercial Centre, Xisanqu, Tiantongyuan, Changping District,eijing, PR China) to draw nanofibers from the prepared solution.he fibers were collected on a grounded 100 cm2 aluminum plate.he relative humidity was controlled between 15 and 18% and theemperature at 15 ◦C. After the electrospun membranes (EMs) wereormed, they were heated at 60 ◦C for 1 h. Subsequently, the mem-ranes were heated up to 157 ◦C to improve the structural integrityf the membrane. The fiber diameters were determined from theEM image using the ImageJ software (http://rsb.info.nih.gov/ij/).he SMM blended EMs were prepared by adding SMM (8 wt% ofVDF) to the 20% (w/v) PVDF solution. The control EM without SMMlending will be hereafter referred to as EM-PVDF. The EMs blendedith SMM-400, SMM-600 and SMM-1000 will be hereafter referred

o as EM-400, EM-600 and EM-1000, respectively.

.5. Preparation of asymmetric membranes (AMs) by the phasenversion technique

The asymmetric membranes (AMs) were prepared from 20%w/v) PVDF solution dissolved in a mixture of DMAc and acetonet a ratio of 2:3. A thin strip of solution was poured almost at thedge of a clean glass plate and immediately spread by a blade acrosshe glass plate. The glass plate together with the cast polymer solu-ion film was then placed in a cold water bath. After several minuteshe membrane was removed from the cold water bath and stored ine-ionized water. To prepare the casting dope for the SMM incorpo-ated membrane, a polymer solution containing 15% (w/v) of PVDFas first prepared and then SMM (8 wt% of PVDF) was added to the

VDF solution. The control asymmetric PVDF membrane withoutMM blending is hereafter referred to as AM-PVDF. The asymmet-ic membranes blended with SMM-400, SMM-600 and SMM-1000re referred to as AM-400, AM-600 and AM-1000.

.6. Membrane characterization

Elemental analysis of the surface of the EMs was performedy X-ray photoelectron spectroscopy (XPS) using a Kratos AxisIS Mono-Al X-ray photoelectron spectrometer (Manchester, UK).he X-ray source was operated at 15 kV, 10 mA, 150 W, the take-ff angle was 90◦ (vertical to sample surface) and the detectionepth was not more than 10 nm. EMs were also characterized by

differential scanning calorimeter (DSC, TA instrument SDT Q600quipped with TA instrument’s universal analysis 2000 softwareersion 3.9a).

Field emission scanning electron microscopy (FE-SEM, FEI-UANTA 200F, The Netherlands) was used to observe the surfaceorphology of EMs and AMs. The membranes were sputteredith a thin layer of gold before being placed in the SEM cham-

er. Coated samples were examined at an accelerating voltage of5 kV. Scanning electron microscopy–energy dispersive spectrom-ter (SEM–EDX, model Tescan Vega-II XMU VPSEM, Tescan USA Inc.,ranberry Twp., PA) was used to provide the atomic% at the surfacef the EMs and AMs.

Static contact angle (SCA) measurements were performed onhe EMs and AMs using an Advanced Surface Technologies, Inc.,ideo contact angle (VCA) Optima Surface Analysis System, Biller-ca, MA. A water drop of 0.5 �L was dispersed on the membrane

10.9 36.1 19.74 29.0615.7 47.9 8.97 18.2819.7 38.4 22.85 29.59

surface and the SCA determined using the system software. Theasymmetric membranes were dried at 50 ◦C overnight before SCAwas measured.

The pore size distribution, bubble point and mean flow poreof EMs were determined using a capillary flow porometer (PorousMaterials Inc., Ithaca, NY) which was able to detect pore size from0.013 to 500 �m. The membranes were completely wetted withwetting liquid GalwickTM (Porous Materials Inc., Ithaca, NY) andpressure was applied on one side.

Circular EMs of 25 mm in diameter with an effective area of4.1 cm2 were stamped out and subsequently used for flux stud-ies. All tests were conducted on an Amicon stirred cell model 8010,which was able to withstand a maximum operating pressure of75 psig, with a feed capacity of 10 mL. The permeation cell wasconnected to an 800 mL water bath. The nitrogen gas was usedto supply pressure to the feed water. The pressure was slowlyincreased from 0 to 20 psig and the corresponding water flux wasmeasured by weighing water collected during a predeterminedperiod. The pressure of gas was detected using a digital gauge (Meri-gauge 3900, Meriam Instruments, Cleveland, OH). Constant stirringwas applied during the collection of pure water.

3. Results and discussion

3.1. Surface modifying macromolecules (SMMs)

The structure of the SMMs in terms of the number of repeatingunits (m, n, p, and q (see Fig. 1)) was obtained as follows. The valuesof the n and q were calculated from the average molecular weightof PPG and PEG, respectively. The values of m and p were calculated(assuming that all of the added MDI and PPG were consumed) usingthe number average molecular weight Mn of the SMM obtainedfrom GPC experiments. The results are listed in Table 2 togetherwith number average (Mn) and weight average (Mw) molecularweight.

The glass transition temperatures (Tg) at the onset and the mid-point of the thermograph were determined by DSC. The results arealso depicted in Table 2. It has been observed that as the molecularweight of PPG increases (from 425 to 3500 Da, or n = 7.02–60.03 inTable 2), the Mn of SMM also increases.

3.2. Influence of different SMMs on fiber size

Table 3 summarizes the fiber diameters of the different EMs (thelast column). The fiber diameter of EM-1000 was 0.15 �m largerthan EM-PVDF. Interestingly, fiber diameters of the EM-400 andEM-600 were much smaller than EM-PVDF and EM-1000, i.e. thefiber size of EM-1000 was larger by ∼3.7 times and ∼4.3 timesthan EM-400 and EM-600, respectively. Fig. 2 shows the surfacearchitecture of the different EMs observed by SEM. There is a directrelationship between the fiber size and the size of the largest poremeasured by the bubble point method (Table 3, the first column).For example, when the fiber diameter increased by 3.7 times from

EM-400 to EM-1000, the bubble point increased by ∼1.5 times. Sim-ilarly, as the fiber diameter increased by ∼4.3 times from EM-600 toEM-1000, the largest pore size increased by ∼2.0 times. Other poresizes are also included in Table 3 to show the degree of pore size

238 S. Kaur et al. / Journal of Membrane Science 390– 391 (2012) 235– 242

Table 3Pore size distributions and fiber diameters of the different EMs.

EMs Largest pore (bubble point) diameter (�m) Mean flow pore diameter (�m) Smallest pore (�m) Fiber size (�m)

EM-PVDF 4.77 2.09 1.47 1.00 ± 0.52EM-400 3.14 0.96 0.47 0.31 ± 0.10EM-600 2.32 0.68 0.44 0.27 ± 0.09EM-1000 4.74 2.33 1.02 1.15 ± 0.55

DF, (b

dda

3

fdEtoS[tE

According to Table 4, the F content that represents the hydrophobicPVDF decreases from EM-PVDF to SMM blended EMs (EM-400, EM-600 and EM-1000) due to the absence of F in SMM. However, the Fcontent of EM-1000 is not necessarily the lowest. Thus, the atomic

Table 4Static contact angle (SCA) and surface atomic composition by XPS of the variousblended EMs.

Membrane ID SCA (◦) XPS results, atomic conc. (wt%)

F (1s) O (1s) N (1s) C (1s)

Fig. 2. Surface architecture of (a) EM-PV

istribution. The above observation indicates that, when the fiberiameter decreases, the number of fibers per unit area increasesnd the larger pores are split into smaller pores.

.3. Influence of SMM on hydrophilicity

The three SMMs synthesized had different effects on the sur-ace hydrophilicity of the EMs. As the static contact angle (SCA)ata summarized in Table 4 show the contact angles of EM-400 andM-600 (∼140◦ for both) and EM-PVDF (∼131◦) are much higherhan a heat pressed PVDF film (∼87◦). Significantly higher SCAsf EMs are often recorded for various polymers and the increase in

CAs is attributed to its inherent roughness and trapped air pockets11–13]. On the contrary, EM-1000 had a SCA of 0◦. To understandhe observed remarkable differences in SCAs, particularly betweenM-1000 and other EMs, XPS analysis was conducted. As the results

) EM-400, (c) EM-600, and (d) EM-1000.

of the XPS analysis, the atomic concentrations of fluorine (F), oxy-gen (O), nitrogen (N) and carbon (C) are also listed in Table 4.

EM-PVDF 131.54 ± 4.47 51.11 0.56 0 48.34EM-400 139.79 ± 4.70 44.02 3.04 0.74 52.20EM-600 140.00 ± 3.10 40.19 4.55 0.98 54.28EM-1000 0 42.10 4.84 0.06 53.00

S. Kaur et al. / Journal of Membrane Sc

Table 5EDX results for the various EMs and theoretical atomic compositions (in the bracket)when SMMs are uniformly distributed in EMs.

Membrane ID EDX results, atomic conc. (wt%)

F O N C

EM-PVDF – – – –(61.3) (0) (0) (38.7)

EM-400 50 4 6 42(56.40) (2.26) (0.35) (40.99)

EM-600 54 3 3 40(56.40) (2.36) (0.31) (40.93)

ch

atsbs1ttiaaSgatgto

pAtmcemwbcEi

EM-1000 2 30 4 62(56.40) (2.56) (0.09) (41.00)

ompositions obtained by XPS cannot explain the extremely highydrophilic nature of EM-1000.

In the case of PVDF, it has only three elements (H, C and F)nd its N composition was obviously zero as was expected. Onhe other hand, SMMs have additional elemental groups presentuch as N and O. Hence, it would be expected that all the threelended membranes would have a high N and O peak in the XPSpectrum. However, the N content was almost negligible for EM-000 when compared to the other two blended membranes, buthe O content for the EM-1000 membrane was marginally higherhan the other two blended membranes. The content of O increasesn the order SMM-400 < SMM-600 < SMM-1000. This was expecteds can be explained as follows. The higher molecular weight PEGnd PPG were used to synthesize SMM-1000 when compared toMM-400 and SMM-600 and hence the former has a higher oxy-en content. But, the N content for EM-400 and EM-600 was 0.74nd 0.98 respectively, which indicates that the N–H group for thesewo membranes was majorly at the surface and hence the amideroup (–NH–C O) may have contributed to a higher SCA comparedo EM-PVDF. The SCA and XPS analyses suggest that the orientationf the three SMMs within the PVDF blended fiber was different.

SEM–EDX analysis summarized in Table 5, on the other hand,rovided results remarkably different from the XPS analysis.tomic compositions were also calculated based on the assump-

ion that the SMM was uniformly distributed in the EMs using theolecular structures of the SMMs listed in Table 3 and the SMMs’

ontent in the EMs. The results are also listed in Table 5 in the brack-ts. It is to be noted that the content of F, the marker for PVDF,easured by EDX is significantly lower than the calculated F value,hile the contents for O and N, the markers for the SMM, measured

y EDX are significantly higher than the calculated values, indi-

ating the surface migration of the SMM in the EMs. In particular,M-1000 showed exceptionally low F content and high O content,ndicating a high degree of surface coverage by SMM-1000. The N

Fig. 3. Schematic illustration of the SM

ience 390– 391 (2012) 235– 242 239

content was not as high as expected from the remarkable increaseof O, about which discussions will be made later more in detail.

The degree of the surface coverage by SMM can be calculatedusing the atomic composition at the surface. The evaluation of thesurface coverage by three components of SMM blended EMs; i.e.PVDF repeat unit, polyurethane repeat units (those including PEGand PPG soft segments combined) and PEG end-capping groups,was done by using the F and C content of EM-1000. The resultswere

PVDF repeat unit: 58.1%Polyurethane repeat unit: 3.6%End-capping group: 38.3%,

indicating that a substantial part of the membrane surface is cov-ered by PEG end-capping group.

Classifying the type of SMM configuration at the EM surfaceinto four modes depicted in Fig. 3, the above results show thatthe configuration of SMM-1000 at the EM surface belongs to type4 (exposed). On the other hand, SMM-400 and SMM-600 did notmigrate to the surface as much as SMM-1000 and many SMMmolecules belong to type 1 (embedded).

EDX further allows us to obtain atomic compositions at differentparts of the EM. Thus, EDX analysis of EM-1000 was conducted on53 spots as shown in Fig. 4. It was revealed that the results aregrouped into two categories; one with high N contents of average10 wt% and the other low N contents of nearly equal to 0 wt%. SinceN belongs only to the middle polyurethane section of SMM, the highN content suggests that the surface configuration of SMM-1000 istype 3 (submerged). Thus, SMM-1000 was either in the 3rd or 4thconfiguration, enhancing, in either case, the surface hydrophilicityas compared to PVDF.

3.4. Influence of SMM-1000 on filtration flux

The pure water flux of EM-PVDF and EM-1000 is given in Fig. 5.From the figure, the water flux of EM-1000 is 20% higher than EM-PVDF. Since both EMs’ structures are similar (see Fig. 2 and Table 3)the observed increase in water flux seems due to the increase inhydrophilicity from EM-PVDF to EM-1000. It is interesting to notethat EM-PVDF looked opaque while EM-1000 appeared transparentafter the permeation test.

3.5. Influence of SMM on thermal behavior

The EMs were further characterized by their thermal behavior.Table 6 shows the enthalpy of fusion (�Hf) and the melting point(Mp) of the EMs (DSC thermograms are given in the supporting

M configuration on a single fiber.

240 S. Kaur et al. / Journal of Membrane Science 390– 391 (2012) 235– 242

Fig. 4. Elemental analysis on se

isdePlbtotEticlcPTi

TT

Fig. 5. Water permeation flux of non-blended EM-PVDF and EM-1000.

nformation). The Mp of the SMM blended membranes changedlightly, either upwards or downwards, from the EM-PVDF. Theeviation of Mp from the semi-crystalline PVDF membrane can bexplained using polymer-diluent crystallization concept. The pureVDF is associated mostly with � crystalline phase. When PVDFamellae are blended, at least partially, by amorphous SMM, thelended system becomes associated more with � crystalline struc-ure due to the presence of SMM. A similar observation, appearancef double melting peaks, is noticed in our previous research [10]. Onhe other hand, in the case of �Hf, decrease in �Hf was noticed fromM-PVDF to EM-400. The blending of EM-400 decreases the crys-alline nature of PVDF and thereby increases amorphous structuren the blend. It has been reported in the literature that increasing theoncentration of PEG-b-poly(methyl methacrylate) (high molecu-ar weight) in PVDF blend resulted in an increase in the amorphous

ontent [14]. As the length of the PEG increased from PEG-400 toEG-600 (EM-400 to EM-600) the enthalpy of fusion increased.his could be due to the increase in crystalline nature with anncrease in PEG length thereby allowing a higher amount of packing

able 6hermal properties of electrospun membranes (EMs).

EMs �Hf (J/g) Melting point (Mp) (◦C)

EM-PVDF 47.66 163.9EM-400 26.5 167.5EM-600 51.8 162.9EM-1000 73.3 164.3

veral fibers of EM-1000.

during crystal growth and hence more energy is required to meltthe polymer chains. This indicates that length of PEG and/or PPGchains greatly influences the crystallization of the PVDF chainsegments. It is already reported in literature that PEG-200 is amor-phous in nature, whereas PEG-400 and PEG-600 are crystalline innature [15]. It is to be noted here that in the case of EM-1000 addi-tional exothermic crystallization peak at higher temperature wasobserved. In general, the endothermic melting peak is observedat higher temperature than the exothermic crystallization peak. Itis suggested that the formation of the strong hydrogen bondingbetween fluorine (PVDF) and hydrogen (SMM) in the liquid crys-talline structure takes place. Notably, the additional peak has beenobserved for the particular SMM (SMM-1000) due to the highermolecular weight of PPG and PEG in comparison to the other SMMs(SMM-400, and SMM-600).

3.6. Influence of SMM in the asymmetric membranes prepared bythe phase inversion method

The SCAs of asymmetric membranes (AMs) are summarized inTable 7. The SCA of AM-PVDF is very close to the SCA of PVDF film(∼87◦). Small changes are noticed from AM-PVDF to AM-400 andAM-600 but they are within the error range. On the other hand, asignificant change is noticed from AM-PVDF to AM-1000 (by 36◦).

Interestingly, comparison of Tables 4 and 7 reveals large differ-ences between the SCAs of EMs and AMs. First, SCAs of EM-PVDF,EM-400 and EM-600 are much larger than those of AM-PVDF, AM-400 and AM-600, which is ascribed to the presence of a largequantity of air in EMs. From SEM images of EMs (Fig. 2) and AMs(Fig. 6) it is obvious that the porosity of EMs is larger than AMs.Therefore, EMs contain a larger amount of air, which increasesthe SCA. Conversely, the SCA of EM-1000 (nearly 0◦) is muchsmaller than AM-1000 (53.9◦). This can be attributed to a muchgreater coverage of the EM surface, in exposed configuration (seeFigs. 3 and 4), by SMM-1000 than the AM surface. This was con-firmed by SEM–EDX analysis of AMs, the results of which aresummarized in Table 7.

Comparing among AMs, the F content (marker of PVDF) of AM-1000 was the lowest while its O content (marker of SMM) was the

highest. It indicates the highest degree of SMM-1000 migration tothe surface and hence the surface becomes the most hydrophilic.This coincides with the lowest SCA of AM-1000 among all AMs.Comparing Table 5 (atomic compositions of EMs) with Table 7

S. Kaur et al. / Journal of Membrane Science 390– 391 (2012) 235– 242 241

Table 7SCA and EDX results of AMs (the theoretical values of the different elements is the same as those in Table 5).

AMs Surface contact angle (SCA,◦) EDX results, atomic conc. (wt%)

F O N C

AM-PVDF 90.10 ± 9.31 – – –AM-400 102.15 ± 4.56 48.15 ± 4.28 3.60 ± 0.66 4.22 ± 1.46 44.04 ± 4.08AM-600 85.89 ± 2.86 50.14 ± 4.31 4.49 ± 2.25 7.25 ± 1.97 38.13 ± 3.60AM-1000 53.90 ± 9.70 46.66 ± 7.69 7.02 ± 1.93 4.87 ± 2.50 41.45 ± 5.05

SMM

(aSdpfT

hoeewmhv

gp

Fig. 6. SEM of 20% (v/w) phase inverted membranes: (a) without

atomic composition of AMs), those of EM-400 and EM-600 arelmost the same as AM-400 and AM-600. Therefore, much higherCAs of EM-400 and EM-600 than those of AM counter parts areue not to higher hydrophobicity of the material, but to the higherorosity of EMs. Comparing EM-1000 and AM-1000, F content of theormer is much lower and O content much higher than the latter.his explains the much lower SCA of EM-1000 than AM-1000.

A question arises as to why the surface of EM-1000 could beighly covered by SMM-1000, while the degree of SMM coveragen the AM-1000 surface was not as high as EM-1000. This may beither due to the much smaller dimension of EM-1000 (fiber diam-ter was 1.15 �m) as compared to AM-1000 (membrane thicknessas ∼115 �m) or due to disruption of the highly packed macro-olecules in the surface layer, which prevents protrusion of highly

ydrophilic SMM end-capping groups to the surface, by the high

oltage applied during electrospinning process.

Fig. 6 shows the SEM images of AMs. The figure reveals theradual increase of pore size from AM-PVDF to AM-1000. This wasrobably caused by the increase in the rate of water influx in the

, (b) with SMM-400, (c) with SMM-600, and (d) with SMM-1000.

solvent/non-solvent exchange process as the surface hydrophilicityincreases by the enhanced SMM migration from SMM-400 to SMM-1000. During the electrospinning process, the SMMs and PVDFchains are phase separated and there is a greater possibility of theSMMs being at the surface, while during the phase inversion pro-cess the SMMs slowly migrate to the surface and the amount ofSMMs on the surface may be smaller.

4. Conclusions

SMM-1000 played a significant role in increasing thehydrophilicity of the blended electrospun membrane, whencompared to SMM-400 and SMM-600 of slightly increasedhydrophobicity. This was due to the orientation the SMMs adoptedduring electrospinning in which hydrophobic part of SMM-400

and SMM-600 was majorly at the surface when compared to SMM-1000. When the SCA of AM-1000 with EM-1000 was compared,its contact angle did not reach 0◦ but reduced to ∼54◦ instead.This shows that besides surface modification agents, the nature

2 ane Sc

oiEid

A

wropntTN

A

t

R

[

[

[

[

[

lithium ion battery, J. Membr. Sci. 334 (2009) 117–122.

42 S. Kaur et al. / Journal of Membr

f membrane formation also plays an important role in influenc-ng the hydrophilicity of the membrane. The pure water flux ofM-1000 was 20% higher than that of non-blended EM-PVDF. Thisndicates that the hydrophilicity of a highly porous membraneoes contribute significantly to the water flux.

cknowledgements

The author (S. Kaur) is thankful for the fellowship for researchork at University of Ottawa. This work is supported by the Envi-

onment and Water Industry (EWI) Development Council (Govt.f Singapore) through the funded project “Development of lowressure, high flux UF and NF membranes based on electrospunanofibers for water treatment” and NUS Nanoscience and Nano-echnology Initiative (NUSNNI), National University of Singapore.he authors gratefully acknowledge the financial support from theatural Sciences and Engineering Research Council of Canada.

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at doi:10.1016/j.memsci.2011.11.045.

eferences

[1] S. Ramakrishna, K. Fujihara, W.E. Teo, T.C. Lim, Z. Ma, An Introduction to Elec-

trospinning and Nanofibers, World Scientific Pub., Singapore, 2005.

[2] J.F. Cooley, Apparatus for electrically dispersing fluids, US Patent 692,631(1902).

[3] K. Yoon, B.S. Hsiao, B. Chu, Functional nanofibers for environmental applica-tions, J. Mater. Chem. 18 (2008) 5326–5334.

[

ience 390– 391 (2012) 235– 242

[4] R. Gopal, S. Kaur, C.Y. Feng, C. Chan, S. Ramakrishna, S. Tabe, T. Matsuura,Electrospun nanofibrous polysulfone membranes as pre-filters: particulateremoval, J. Membr. Sci. 289 (2007) 210–219.

[5] S. Kaur, Z. Ma, R. Gopal, G. Singh, S. Ramakrishna, T. Matsuura,Plasma-induced graft copolymerization of poly(methacrylic acid) on elec-trospun poly(vinylidene fluoride) nanofiber membrane, Langmuir 23 (2007)13085–13092.

[6] C. Feng, K.C. Khulbe, T. Matsuura, R. Gopal, S. Kaur, S. Ramakrishna, M. Khayet,Production of drinking water from saline water by air-gap membrane distil-lation using polyvinylidene fluoride nanofiber membrane, J. Membr. Sci. 311(2008) 1–6.

[7] D. Rana, T. Matsuura, Surface modifications for antifouling membranes, Chem.Rev. 110 (2010) 2448–2471.

[8] R. Gopal, Z. Ma, S. Kaur, S. Ramakrishna, Surface modification and application offunctionalized polymer nanofibers, in: G.A. Manssori, T.F. George, G. Assoufid,L. Zhang (Eds.), Molecular Building Blocks for Nanotechnology: From Diamon-doids to Nanoscale Materials and Applications, Springer-Verlag, New York, NY,2006, pp. 72–91.

[9] D.E. Suk, G. Chowdhury, T. Matsuura, R.M. Narbaitz, P. Santerre, G. Pleizier,Y. Deslandes, Study on the kinetics of surface migration of surface modi-fying macromolecules in membrane preparation, Macromolecules 35 (2002)3017–3021.

10] H.T. Dang, C. Amelot, D. Rana, R.M. Narbaitz, T. Matsuura, Performance of anewly developed hydrophilic additive blended with different ultrafiltrationbase polymers, J. Appl. Polym. Sci. 116 (2010) 2205–2215.

11] R.N. Wenzel, Resistance of solid surfaces to wetting by water, Ind. Eng. Chem.2 (1936) 988–994.

12] A.B.D. Cassie, S. Baxter, Wettability of porous surfaces, Trans. Faraday Soc. 40(1944) 546–551.

13] A. Singh, L. Steely, H.R. Allock, Poly[bis(2,2,2-trifluoroethoxy)phosphazene]superhydrophobic nanofibers, Langmuir 21 (2005) 11604–11607.

14] Q. Xiao, X. Wang, W. Li, Z. Li, T. Zhang, H. Zhang, Macroporous polymer elec-trolytes based on PVDF/PEO-b-PMMA block copolymer blends for rechargeable

15] M.K. Park, H.S. Kim, J.H. An, J. Kim, New oligomeric ether plasticizers for solidpolymer electrolytes: synthesis and electrical properties of oligomeric PEO hav-ing bis(five-membered cyclic carbonate)s at chain ends, J. Ind. Eng. Chem. 11(2005) 222–227.