Doping of Polyethersulfone Nanofiltration Membanes Antifouling Effect Observed at Ultralow...

Dynamic Article LinksC<Journal ofMaterials Chemistry

Cite this: J. Mater. Chem., 2011, 21, 10311

www.rsc.org/materials PAPER

Dow

nloa

ded

by K

atho

lieke

Uni

vers

iteit

Leu

ven

on 1

2 M

arch

201

2Pu

blis

hed

on 2

7 M

ay 2

011

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C1J

M11

040C

View Online / Journal Homepage / Table of Contents for this issue

Doping of polyethersulfone nanofiltration membranes: antifouling effectobserved at ultralow concentrations of TiO2 nanoparticles

Arcadio Sotto,a Arman Boromand,b Stefan Balta,bc Jeonghwan Kimd and Bart Van der Bruggen*b

Received 10th March 2011, Accepted 18th April 2011

DOI: 10.1039/c1jm11040c

Doping of nanofiltration membranes with TiO2 nanoparticles was studied in the ultralow concentration

range, in the absence of photocatalysis. Blended polyethersulfone/TiO2 flat-sheet membranes were

manufactured and investigated in terms of pure water flux, permeability, fouling resistance and solute

rejection. The membranes were synthesized at four different polymer concentrations by the phase

inversion method, using 1-methyl-2-pyrrolidone (NMP) and deionized water as solvent and coagulant,

respectively. The influence of TiO2 addition was investigated in an unusually low concentration interval

(0.035–0.375 wt%). The membrane morphology was studied by determining particle size distributions

of TiO2 to explore the effect of nanoparticle aggregation. Furthermore, membranes were characterized

by hydrophilicity (contact angle), morphology (SEM), porosity, mechanical strength (bursting

pressure) and thermal analysis (TGA). Membrane fouling was studied with humic acids as model

organic foulants. Overall, a remarkable improvement in the permeability was observed with the

addition of ultralow amounts of nanoparticles to the polymer. The optimum permeability was found to

be as low as 0.085 wt%, using a constant rejection of dyes as the boundary condition. It was shown that

rejection of solutes is not negatively influenced by the increase in permeability. In addition, the

resistance against membrane fouling was found to be above 12% for the TiO2 blended membranes.

1. Introduction

Nanofiltration (NF) membranes are currently used as advanced

separation tools, aiming at the removal of pollutants and impu-

rities from aqueous solutions in the production of drinking water,

quaternary treatment ofwastewater in view of recycling, and from

non-aqueous process streams.1 NF membranes are asymmetric,

consisting of a separation layer ranging from 0.1 to 1 mm sup-

ported by a thicker layer with larger pores. Among many aspects

related to the separation itself, membrane fouling caused by the

deposition of organic pollutants on the membrane surface and/or

adsorption intomembrane pores is amajor obstacle.1,2Membrane

fouling due to adsorption of organic pollutants increases the

membranes’ hydrophobicity or could decrease pore diameters,

and thus reduce water transport across the membrane.3–5 In

addition, the formation of the fouling layer could alter other

aDepartment of Chemical and Energy Technology, Rey Juan CarlosUniversity, Tulip�an s/n, 28933 M�ostoles, Madrid, Spain. E-mail: [email protected]; Fax: +34 91 4887068; Tel: +34 91 4888153bDepartment of Chemical Engineering, K.U. Leuven, W. de Croylaan 46,B-3001 Leuven, Belgium. E-mail: [email protected];Fax: +32 16 322991; Tel: +32 16 322340cDepartment of Environmental and Material Engineering, UniversityDunarea de Jos, Galati, Romania. E-mail: [email protected]; Fax: +32 16 322991; Tel: +32 16 322340dDepartment of Environmental Engineering, INHA University,Yonghyungdong 253, Namgu, Incheon, Republic of Korea. E-mail:[email protected]; Fax: +82 32 865 1425; Tel: +82 32 860 7502

This journal is ª The Royal Society of Chemistry 2011

functional properties of the membrane surface such as membrane

charge, which affects the rejection performance.6

Attempts to develop ‘‘mixed membranes’’ in which nano-

particles are used as additives in the polymeric membranes have

received much attention recently. Fabricating polymeric

membranes with various metal oxide nanoparticles is attractive,

because metal oxides can provide specific functionalities to the

membrane, while retaining the intrinsic separation performance

of the bare membrane. Enhancing the hydrophilic nature of the

membrane is assumed to yield a better performance in terms of

permeability, antifouling properties and solute rejection.7–10 In

addition, embedding catalytic nanoparticles on the surface of

polymeric membrane allows the use of such membranes in

a catalytic membrane reactor.11,12 In aqueous systems, it is often

attempted to use nanoparticles to combine filtration on a nano-

scale with oxidation. The latter can be obtained by using e.g.,

a chemical oxidant such as ozone, but also by using the photo-

catalytic effect of well-chosen nanoparticles.13–15 The use of

nanoparticles may, however, also have a direct effect on the

membrane performance, in the absence of a (photo)catalytic

effect; this has been shown by the development of freestanding

thin-film membranes consisting of nanoparticles16 or aligned

nanotubes.17,18

Among the various types of nanoparticles that are being used

for addition to polymeric membranes are silica (SiO2), carbon

nanotubes, alumina (Al2O3), silver (Ag), zirconia (ZrO2), gold

(Au), zerovalent iron (Fe0), palladium (Pd) and titania

J. Mater. Chem., 2011, 21, 10311–10320 | 10311

http://dx.doi.org/10.1039/c1jm11040c






http://pubs.rsc.org/en/journals/journal/JM

http://pubs.rsc.org/en/journals/journal/JM?issueid=JM021028

Dow

nloa

ded

by K

atho

lieke

Uni

vers

iteit

Leu

ven

on 1

2 M

arch

201

2Pu

blis

hed

on 2

7 M

ay 2

011

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C1J

M11

040C

View Online

(TiO2).19–21 The TiO2 nanoparticles have been studied for the

enhancement of polymeric membranes because of their unique

characteristics of excellent resistance against chemicals and

photocatalytic properties.22–24 In most of these studies TiO2

nanoparticles were used as additives in high concentrations

(above 1%) to various polymers, in view of their photocatalytic

activity.1 For polymeric membranes, however, the combination

with UV irradiation may be problematic because of the risk of

membrane degradation. Several studies also suggested intrinsic

effects of TiO2 nanoparticles on the membrane performance

without UV irradiation, and a positive effect on improving

membrane permeability was reported when sufficiently high

concentrations of TiO2 are used.25–27

The addition of nanoparticles to the membrane structure is

carried out through two general ways: coating or deposition of

nanoparticles on the membrane surface28,29 and blending with the

membrane casting solution.30–32 The latter method has been used

more frequently, due to its easier operation. This method

involves the addition of TiO2 nanoparticles during membrane

synthesis. A suspension of the dissolved polymer and the nano-

particles is cast on a support layer. Mixing is critical in this

procedure: the suspension should be homogeneous, and no

aggregation should occur. In this context, TiO2/polymer thin film

composite (TFC) reverse osmosis membranes were studied as

well.22

The use of nanoparticles can have a direct influence on

membrane characteristics such as the porosity and pore size of

the membrane skin and the macrovoid morphology of the

asymmetric support.33 Lower contact angles have been observed

by adding titania nanoparticles to a polyethersulfone (PES)

membrane; Rahimpour et al.29 observed a decrease in contact

angle from 66� to 53� for TiO2-entrapped PES membranes at

a nanoparticle concentration of 6%. This, however, does not

always lead to higher fluxes due to the risk of pore plugging. The

concentration of nanoparticles is critical for this, and aggrega-

tion of nanoparticles is decisive. Aggregation of the individual

nanoparticles and the establishment of an equilibrium state

under definite thermodynamic conditions determine the size

distribution of the agglomerate of dispersed nanoparticles.34

Taurozzi et al.35 mention an increased skin layer thickness,

a higher surface porosity of the skin, suppressed macrovoid

formation, and a higher permeability of the membrane (although

a status quo and a decreased permeability have also been

observed, or a maximum permeability at intermediate nano-

particle loadings); the rejection may increase or decrease due to

the addition of TiO2 nanoparticles. The understanding of the

effect caused by the nanoparticle concentration on the membrane

performance is still not clear.

Usually, the concentration of TiO2, relative to the casting

solution, used in reported studies is above 1 wt%.25,29,30,36,37 This

is remarkable, given the effect of concentration on the perfor-

mance of the resulting membrane; it is known that high

concentrations may lead to negative effects instead of improve-

ments. Only in the work of Wu et al.31 the authors used

a concentration of 0.3 wt% on the lower end. The membrane

performance at TiO2 contents below 0.3 wt% is still not under-

stood, which is a lack in the knowledge of this research field.

In this paper a systematic study of the concentration effect of

TiO2 in the ultralow concentration range was carried out, in view

10312 | J. Mater. Chem., 2011, 21, 10311–10320

of determining unambiguously the effect of the nanoparticles on

the membrane performance in the absence of any masking effect

due to agglomeration, blocking of pores or photochemical

oxidation. An overall membrane characterization was carried

out to determine the direct effect of the nanoparticles on the

physico-chemical properties of the membranes, such as hydro-

philic character, permeability and fouling resistance. In addition,

the membrane performance during filtration of aqueous solu-

tions containing dyes was studied in view of understanding the

indirect effects of the nanoparticles on membrane permeability

and solute rejection.

2. Methods and materials

2.1. Materials

Polyethersulfone (PES, type Radel) supplied by Solvay (Bel-

gium) was employed as the base polymer. 1-Methyl-2-pyrroli-

done (NMP, 99.5%) was used as the polymer solvent. The

support layer (Viledon FO2471) used for the PES membrane

manufacturing was obtained from Freudenberg (Weinheim,

Germany).

Titanium tetraisopropoxide (TTIP) was used as a precursor of

TiO2 nanoparticles. 0.125 mol of TTIP was added dropwise to

100 mL of deionized water under vigorous stirring at room

temperature. The sol samples obtained by hydrolysis process

were irradiated in an ultrasonic cleaning bath for 1 h. To

hydrolyze the TTIP samples and obtain monodisperse TiO2

particles, the samples were subsequently aged in a closed beaker

at room temperature for 24 h. After aging, these samples were

dried at 100 �C for 8 h in air to vaporize water. Dried gel samples

obtained were calcinated further at 500 �C for 1 h. Surface

analysis and microscopic analysis using XRD (X-ray Diffrac-

tometer system, DMAX-2500, Rigaku) and TEM (CM 200,

Philips Inc) showed that the synthesized TiO2 particles were

characterized as anatase with 25 nm particle diameter.

Sigma–Aldrich humic acid (St. Louis, MO) was selected as

a model organic foulant in this study. Humic acids (HAs)

constitute a major component of natural organic matter that can

be found in wastewater streams and refers to the fraction of

humic substances obtained by chemical and biological degrada-

tion products from plant and animal residues.46

Six different dyes were used to explore the size interaction in

the interface solute–membrane pore. Organic compounds

purchased from Acros Organics (Belgium) were selected in order

to cover a large range of molecular mass. The selected dyes were

methyl-red (269.21 Da), neutral-red (288.77 Da), methylene-blue

(319.85 Da), Sudan-black (456.54 Da), Victoria-blue (506.10

Da), and Congo-red (696.67 Da).

2.2. Membrane preparation

Control PES membranes and TiO2-entrapped PES membranes

were prepared using phase inversion induced by immersion

precipitation. PES cast from four different concentrations in

N-methyl-pyrrolidone (NMP) (25, 27, 30 and 32 wt%) was used

as the polymer matrix. The TiO2-entrapped membranes were

prepared by dissolving different amounts of nanoparticles in the

corresponding volume of NMP for 3 h by mechanical stirring at

200 rpm and room temperature. The following concentrations of



Dow

nloa

ded

by K

atho

lieke

Uni

vers

iteit

Leu

ven

on 1

2 M

arch

201

2Pu

blis

hed

on 2

7 M

ay 2

011

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C1J

M11

040C

View Online

TiO2 were used: 0.035, 0.070, 0.085, 0.1, 0.125, 0.2, 0.250, 0.3 and

0.375 wt%. Subsequently, the polymer was added to the solution,

which was stirred for 24 h at 500 rpm and 40 �C. After formation

of a homogeneous solution, the films were cast with 250 mm

thickness using a filmograph (K4340 Automatic Film Appli-

cator, Elcometer) in an atmosphere with controlled relative

humidity on non-woven polyester as a support layer. Prior to the

casting, the support layer was wetted with NMP to prevent the

polymer solution of intruding in the pores of the support layer.

The prepared films were immersed in a non-solvent bath

(distilled water at 20 �C) for precipitation. The membrane was

afterwards repeatedly washed with distilled water to remove the

remaining solvent, and stored wet. For each polymer solution

composition, five identical membrane sheets were made and

tested to obtain an average value of flux and solute rejection.

2.3. Particle size distributions

AMasterSizer Laser Diffraction Particle Size Analyzer (Malvern

Instrument Ltd., Malvern, England) was used. The size distri-

bution was quantified as the relative volume of particles in size

bands presented as size distribution curves (MalvernMasterSizer

Micro Software v 2.19).

2.4. Characterization of the membrane surface

A contact angle measuring system DSA 10 Mk2 (Kr€uss, Ger-

many) was used to measure the water contact angle of the

synthesized membranes. A water droplet was placed on a dry flat

homogeneous membrane surface and the contact angle between

the water and membrane was measured until no further change

was observed. A hydrophilic membrane surface is expected to

show a low contact angle. The average contact angle for distilled

water was determined in a series of 8 measurements for each of

the different membrane surfaces.

To visualize membrane surface characteristics, scanning elec-

tron microscopy (SEM) measurements were performed. SEM

images were obtained with a Philips XL30 FEG instrument with

an accelerating voltage of 20 keV. Cross-sections were prepared

by fracturing the membranes in liquid nitrogen.

2.5. Thermal analysis

Thermogravimetric analysis scans were performed on a simulta-

neous TGA–DSC thermobalance (TGA-DCS1, Mettler-Toledo,

S.A.E.). A flow of 100 mL min�1 of nitrogen was employed as

reactive gas. The gas is fed horizontally over the cylindrical pan

that contains the catalyst sample. The heating rate was

10 �C min�1 and the samples placed in a ceramic pan (70 mL of

capacity). Thermal investigations were carried out with the main

focus on the investigation of the behavior of the individual

degradation steps of the membrane in correlation with the TiO2-

entrapment amount. Measurements were carried out for three

pieces of a flat sheet membrane for every type of membrane with

standard deviation below �0.2% of weight loss.

2.6. Porosity and mechanical properties

Porosity (Pr, %) was calculated as a function of the membrane

weight using the following equation:


Pr ð%Þ ¼�Ww �Wd

Sdr

�� 100 (1)

whereWw andWd are the weights of a membrane at equilibrium,

swelling and dry state, respectively; S the membrane area; d the

thickness and r is the density of water. The membranes were

immersed in water during 24 h prior to measurement of swelling

state. The porosity data were the average values obtained for 4

samples of each membrane.

Mechanical tests were developed in a universal testing machine

(MTS Alliance RT/5). Tensile tests were carried out at 23 �C of

temperature and 50% relative humidity at a loading velocity of

5 mm min�1. To have a good accuracy, five specimens were

evaluated for each membrane.

2.7. Permeation experiments

The prepared membranes were characterized for water flux, pure

water permeability (membrane hydraulic resistance) and dye

rejection studies using dead end filtration experiments.

Comparison of the fouling-resistant ability of the manufactured

neat and blended membranes was explored by cross-flow

experiments.

Pure water permeability and dye rejection were determined for

a wide range of membranes. Four membrane coupons of the

same membrane sheet for eight membranes of each type were

tested. Therefore, the obtained results are the average of 32

experimental values. The maximum experimental errors were less

than 5% and 8% for control and blended membranes,

respectively.

The pure water flux was determined from a compaction

experiment at a transmembrane pressure of 10 bar and a constant

temperature of 25 �C in dead-end mode with a Sterlitech HP4750

Stirred Cell. A nitrogen cylinder coupled with the pressure

regulator was connected to the top of vessel to pressurize the cell.

The active membrane area was 14.6 cm2. The thoroughly washed

membrane was cut into the desired shape and fitted in the dead

end device. The volume of the appropriate solution was 250 mL.

The initial water flux was measured 30 s after the pressurization.

Permeate was collected in a graduated cylinder for a time interval

until steady state.47

After compaction, the pure water permeability (PWP) was

determined by measuring the pure water flux (Jw) at different

transmembrane pressures (DP) from 2 to 14 bar; the slope of the

linear regression of the water flux as a function of trans-

membrane pressure was determined as the permeability. The

PWP was calculated by the following equation:

PWP ¼ Jw

DP(2)

Rejections were measured at a transmembrane pressure of 10

bar. Concentration polarization at the membrane surface is

minimized by a driven Teflon coated magnetic stirring bar on top

of the membrane. Four samples of permeate, 5 mL each, were

taken.

In order to study the effect of membrane fouling, the

membranes were tested in a cross-flow filtration set-up48 fed with

5 ppm of HAc. To compare flux decline between different

membranes, relative fluxes were defined as the relation of the

J. Mater. Chem., 2011, 21, 10311–10320 | 10313


Dow

nloa

ded

by K

atho

lieke

Uni

vers

iteit

Leu

ven

on 1

2 M

arch

201

2Pu

blis

hed

on 2

7 M

ay 2

011

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C1J

M11

040C

View Online

permeate flux to the pure water flux of the respective membrane

as follows:

RF ¼ Jv

Jw(3)

Concentration polarization is minimized by using a cross-flow

velocity of 4.5 m s�1. This feed velocity corresponds to a Rey-

nolds number of 30 000, which is situated far in the turbulent

region.

Fig. 1 TEM image of TiO2 nanoparticles prepared by a sol–gel

technique.

2.8. Fouling experiments

Cross-flow filtration experiments were carried out for humic acid

solution at a fixed concentration of 5 mg L�1 for 24 h at 25 �C and

8 bar transmembrane pressure. To determine the relative flux

corresponding to every membrane the permeation of pure water

was measured before studying the fouling resistance of

membranes in contact with HAc solution. The flux decline was

explored in terms of relative fluxes (eqn (3)). Normalization is

necessary to take the possible disparities between the different

membrane pieces into account.

2.9. Analytical methods

A Shimadzu UV-1601 double beam spectrophotometer was used

to determine the concentration of dyes. Regression factors (R2)

obtained for calibrations within the experimental concentration

range were above 0.99. The rejection R of the dissolved dyes was

calculated as follows:

R ð%Þ ¼�1� Cp

Cf

�� 100 (4)

where Cp and Cf are the permeate and feed concentrations of

dyes, respectively.

To determine the concentration of TiO2 nanoparticles in the

membrane structure, inductively coupled plasma atomic emis-

sion spectroscopy (ICP-AES, Agilent Technologies, CA, USA)

was used. The PES membranes at different TiO2 concentrations

tested were characterized at 337 nm wavelength, corresponding

to titania adsorption.

Fig. 2 TiO2 particle size distribution in organic polymeric solutions. I.

27 g of polymer dissolved in NMP doped with 0.035 wt% of TiO2; II. 27 g

of polymer dissolved in NMP doped with 0.085 wt% of TiO2; and III. 27 g

of polymer dissolved in NMP doped with 0.125 wt% of TiO2.

3. Results and discussion

3.1. Particle characterization

Transmission electron microscopy (TEM) was used to investi-

gate the particle size and the image is shown in Fig. 1.

The TiO2 particles can be seen in the form of black spots and

the sizes are different from 25 to 50 nm while some particles are

larger clusters which can be ascribed to the agglomeration of

TiO2 particles.

The incorporation of nanoparticles to the polymeric solution

could alter the inter-diffusion of water and NMP by the

hindrance effect during phase inversion, modifying the further

membrane structure. It is well-known that the nano-TiO2 parti-

cles show a tendency to aggregate due to their high specific

surface area and the hydroxyl groups on the TiO2 surface.38 The

observed size distribution of the nanoparticles at different

contents of TiO2 is shown in Fig. 2.

10314 | J. Mater. Chem., 2011, 21, 10311–10320

The particle size distribution for organic suspensions shifts to

a larger mean diameter as the TiO2 content increases, and has

a lower peak covering a wider size range. This change in the

shape of the particle size distribution for the higher concentra-

tions of nanoparticles is caused by clustering of the nano-

particles. As a result of the enhancement of the polydispersity of

the nanoparticles at higher concentrations of TiO2 it is expected

that the subsequently manufactured membranes are character-

ized by a lower uniformity in the pore size distribution as

a consequence of the differences in hindrance effects, determined

by the size diversity of nanoparticles suspended in the polymeric

solution during the phase inversion process.

3.2. Membrane characterization

Fig. 3 presents SEM images of the cross-sections of control PES

and PES/TiO2 composite membranes.

From the SEM images it is observed that the polymer

concentration has a clear effect on the membrane structure,

which can be described in terms of membrane pore size and

porosity variations. An increase in the polymer concentration in

the initial casting solution led to enhance the polymer fraction



Fig. 3 SEM pictures of cross-section area of PES and PES/TiO2

composite membranes: (a) PES 27 wt%, (b) PES 30 wt%, (c) PES 32 wt%,

(d) PES 27 wt% and 0.035 wt% TiO2, (e) PES 27 wt% and 0.085 wt%

TiO2, and (f) PES 27 wt% and 0.125 wt% TiO2.

Fig. 4 TGA curves of control and composite membranes blended with

different TiO2 contents: (a) control PES membrane, (b) 0.125 wt% TiO2

and (c) 0.350 wt% TiO2.

Dow

nloa

ded

by K

atho

lieke

Uni

vers

iteit

Leu

ven

on 1

2 M

arch

201

2Pu

blis

hed

on 2

7 M

ay 2

011

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C1J

M11

040C

View Online

volume in the phase inversion process. As a result, a denser and

thicker upper layer is formed (Fig. 3a–c).

In addition, when the polymer concentration was 27 wt%,

interconnected macrovoids between both sides of membrane

were observed with higher uniformity in the macrovoids distri-

bution. As the polymer concentration is increased, the geometry

of the macrovoids changes; the observed uniformity disappears

and in some areas, especially close to the bottom near the non-

woven fiber of the membranes, a sponge-like structure is formed.

Macrovoids are not connected with the bottom, and appear to be

blind channels. This effect is more pronounced for membranes

synthesized with 32 wt% PES. Increasing the polymer concen-

tration suppresses macrovoid formation and enhances the

tendency to form a sponge-like structure, increasing at the same

time the thickness of the skin layer.39 A comparison between the

control membrane and TiO2-entrapped membranes in Fig. 3a

and d shows that the addition of TiO2 has an apparent effect on

the membrane structure, the TiO2-entrapped membranes show

a more open structure, having more and larger macrovoids

inside.

In Fig. 3d–f a clear trend is observed for the macrovoid

dimensions, which become larger as the TiO2 concentration in

the polymeric solution increases. As a consequence of the higher

content of TiO2 added during membrane formation in the

interval from 0.035 to 0.125 wt% the finger-like macrovoids close

to the back side of membranes seem to be wider. In addition, it is

noticeable that at 0.035 wt% of TiO2 the macrovoid size distri-

bution is more uniform whereas for the higher concentrations,

closer to the concentrations used in other studies, the macrovoids

have a broad range of size and shape, i.e., the uniformity

decreases yielding a more polydisperse structure. This fact is

coherent with nanoparticle size distributions for the dispersions

shown in Fig. 2.


The results from thermal analysis using TGA are shown in

Fig. 4, where the thermal stability of control and TiO2-entrapped

membranes at different nanoparticle contents is indicated.

The TGA curves show that the decomposition temperature

(Td, defined as the temperature at 3% weight loss) for the selected

membranes was around 420 �C. This result is coherent with

reported values in the literature.40 A small increase is observed

for the residual weight at temperatures above 600 �C as a result

of TiO2 addition. Considering that the residual weight obtained

for the control membrane should be related to organic content of

PES, the difference between the percentages of the original

membrane and the membranes to which TiO2 was added should

correspond to the TiO2 fraction added. This confirms the pres-

ence of nanoparticles in the manufactured membrane

composition.

3.3. Porosity and mechanical properties

Table 1 summarizes the porosity and mechanical properties, in

which the mechanical resistance is expressed in terms of tensile

strength (MPa) and elongation at fracture (%).

As a result of the addition of TiO2 the porosity of the

membrane increases. This behavior can be related to enhanced

macrovoid formation; a higher membrane specific area favors

water sorption inside the membrane.

The bursting pressure of membranes expressed by the tensile

strength is used as a measure of the toughness of the membrane

under pressure.41 Similar results for elongation at fracture were

found for the materials studied.

As shown in Table 1, the bursting pressure of the membranes

slightly decreases and the porosity increases with the increase in

TiO2 amounts added to the membrane structure in the low

concentration range considered here, indicating that the higher

the porosity, the lower the bursting pressure of the membrane. It

is expected that the bursting pressure increases with addition of

metal oxides; however, in this case (low concentration of nano-

particles) the influence of porosity is apparently higher than the

expected effect of metal oxide addition, given the ultralow

concentration range applied here.

J. Mater. Chem., 2011, 21, 10311–10320 | 10315


Table 1 Porosities and mechanical properties’ values

TiO2

content (%) Porosity (%) Tensile strength/MPa Elongation (%)

0 32.1 � 1.4 30.6 � 0.3 20.6 � 0.80.125 39.3 � 0.6 27.6 � 0.2 20.9 � 1.20.375 39.8 � 0.7 27.0 � 0.4 21.2 � 1.2

Dow

nloa

ded

by K

atho

lieke

Uni

vers

iteit

Leu

ven

on 1

2 M

arch

201

2Pu

blis

hed

on 2

7 M

ay 2

011

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C1J

M11

040C

View Online

ICP-AES measurements to investigate TiO2 concentrations in

the membrane structure are shown in Table 2. Based on the

amount of TiO2 in the casting solution of PES in NMP for

the phase-inversion method, the concentration of TiO2 in the

membrane structure was expected to range from 0.32 to 1.28 wt

%. The ICP-AES characterization yields an experimental TiO2

concentration in the membrane structure between 0.27 and

1.15 wt%. This result indicates that the loss of TiO2 particles

during the phase-inversion method is not significant, which

confirms the stability of the membranes. Similar results were

reported by Razmjou et al.,42 but the concentration range of

TiO2 in the casting solution was between 2 and 6 wt%, which is

out of range compared to the concentrations tested in this study

(0.085–0.350 wt%). This result confirms that PES nanofiltration

membranes can be doped successfully with TiO2 nanoparticles in

the ultralow concentration range.

Fig. 5 Effect of PES content and TiO2 entrapment on hydrophilicity

(contact angle) and water permeability of membranes.

3.4. Hydrophilicity and permeability

Fig. 5a shows the measured contact angles for control PES

membranes and membranes containing 0.125 wt% of TiO2 for

different polymer concentrations, indicating that membrane

hydrophilicity increases as the polymer concentration decreases.

The data show that TiO2-entrapped membranes are more

hydrophilic than the neat PES membranes. The effect of polymer

concentration on membrane hydrophilicity from the contact

angle values should be explained in terms of the pore size and

porosity considering that the polymer used was the same in every

case. Increasing polymer concentration decreases the solvent

inter-diffusion rate and hence, the average membrane porosity

and pore size lead to a decrease in the membrane’s hydrophilicity

and water permeability. The presence of the thicker upper layer

observed from the SEM study is in agreement with the results

shown in Fig. 5. Because the water flux is inversely proportional

to the thickness of a membrane, and the sub-layer is very porous,

a lower water flux (more hydraulic resistance) is observed as the

top layer becomes thicker.

The influence of the TiO2 concentration on permeability was

investigated through pure water flux experiments as shown in

Fig. 6. In agreement with the trend shown for the contact angle

Table 2 TiO2 concentration in membrane structure by ICP-AES

TiO2 content incasting solution (wt%)

Expected TiO2

in membrane (wt%)Calculated TiO2

in membrane (wt%)

0.085 0.32 0.27 � 0.120.125 0.46 0.68 � 0.170.350 1.28 1.15 � 0.20

10316 | J. Mater. Chem., 2011, 21, 10311–10320

measurements the polymer concentration has a negative effect on

the water permeation as a consequence of the porosity decrease.

The explanation of the observed permeability increase with the

addition of TiO2 might be associated with a higher affinity to

water of nanoparticles in comparison with the hydrophobic

Fig. 6 Water affinity and permeability of 27 wt% PES/TiO2 membranes

at different TiO2 concentrations.



Fig. 7 Temporal evolution of membrane fouling performance.

Dow

nloa

ded

by K

atho

lieke

Uni

vers

iteit

Leu

ven

on 1

2 M

arch

201

2Pu

blis

hed

on 2

7 M

ay 2

011

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C1J

M11

040C

View Online

polymer. However, the presence of TiO2 not only increases water

diffusion into the growing membrane due to its higher hydro-

philic nature but also affects the interaction between polymer

and solvent molecules by the hindrance effect of nanoparticles.

As the precipitation rate decreases the two phases have more time

to separate, resulting in an increase in pore sizes. Rahimpour

et al. reported that the exchange between solvent and non-solvent

through the coagulation and skin formation favors the expansion

of finger-like pores in depth towards the other skin face resulting

in an increased pore size and porosity of TiO2-entrapped

membranes.29 It provides an improvement in the membrane

performance as water soluble polymer additives were used as

pore forming agents.43 This effect is confirmed by comparing

SEM images of control and TiO2-entrapment membranes.

As explained above, nearly all studies thus far focused on

nanoparticle concentrations above 0.3 wt%. Results are shown in

Fig. 6 for the concentration range between 0.035 and 0.375 wt%.

The hydrophilicities and hydraulic resistance of TiO2 entrapped

membranes were analyzed to examine the variations in wetting

characteristics of membranes as a function of TiO2 concentra-

tion. Considering the previous characterization for PES

membranes, 27 wt% of PES was selected.

In Fig. 6 the standard deviation of contact angle measure-

ments has also been included. The obtained low values for

standard deviations of contact angle measurements suggest

a well dispersion rate of nanoparticles throughout the membrane

surface. It can be seen that the permeability and contact angle of

membranes have a reverse proportionality, as could be expected.

The permeability increases with increasing TiO2 concentration

until 0.085 wt%, varying almost linearly. At this concentration,

water permeation through the membrane is more stable. So,

there are two well-defined behaviors associated with different

ranges of concentration. Addition of nanoparticles in the lower

concentration range, below 0.1 wt%, could disturb the interac-

tion between polymer and solvent molecules by the hindrance

effect of nanoparticles, encouraging a more easy diffusion of

solvent molecules from the polymer matrix, resulting in the

formation of a membrane with a more porous structure.44

Combining the work of Wu et al.31 and the work of Yang et al.,36

it can be concluded that the membrane porosity is altered by the

TiO2 concentration. Increase in the membrane’s porosity and

hydrophilicity as well as the more open membrane structure,

through the improvement in macrovoid formation, could be the

reasons of the permeability enhancement.

At concentrations above 0.085 wt% the permeability is still

higher than that of the control PES membranes, but the overall

trend changes, and rather decreases. However, it is important to

point out that this decrease is not relevant. Some researchers

have observed a similar behavior at higher concentrations of

TiO2.30,36 The TiO2 concentration proposed in this work is not

comparable with the literature reports, but the trend is similar,

although less pronounced. Hence, the causes could be similar.

Due to plugging of membrane pores by nanoparticles during

immersion precipitation, the permeability slightly decreases. In

addition, from the behavior of the standard deviation values

associated with contact angle measurements, it becomes clear

that the reproducibility rate decreases as a consequence of the

less uniform distribution of TiO2 on the membrane surface

(potentially determined by the polydispersity of nanoparticles in


the former polymeric solution). Aggregation of nanoparticles

due to the increase in concentration is assumed to have a negative

effect on the reproducibility. The possible formation of TiO2

aggregates affects the dispersion rate inside the membrane

structure and also on the surface, and the pore size distribution of

the formed membrane as a function of the size of the aggregates.

In some membrane areas the content of TiO2 could be really poor

or TiO2 could even be totally absent. So, the TiO2 entrapped

membranes maintain a porous structure but the increase at

higher TiO2 concentrations, above 0.085 wt%, led to a slight

decrease in permeability because of pore blockage and of

a decrease in the average pore size.

3.5. Fouling resistance

The relative fluxes of control and entrapped membranes at

different concentrations of TiO2 are shown in Fig. 7. The results

clearly show that fouling of the TiO2 entrapped membrane is

minimized. The rate and the extent of membrane fouling

decreased drastically (around 12%).

Amaximum in the resistance against fouling for the membrane

manufactured at 0.085 wt% of TiO2 was observed. This

maximum corresponds to the optimum (higher) value that was

also obtained from the permeability analysis, and also with the

lower contact angle (higher hydrophilicity). Considering that

fouling is here caused by the adsorption of pollutants on the

membrane surface, the probability of HAc adsorbed on the

membrane surface drops off with the addition of nano-TiO2 as

a consequence of membrane hydrophilicity increase. The

apparent maximum antifouling improvement by 12% is a value

similar to those obtained by other researchers, even using higher

amounts of nanoparticles.29,31Rahimpour et al.29 investigated the

fouling mitigation effect of TiO2-entrapped membranes using

casting solution containing concentrations between 2 and 6% of

TiO2. The reported values by the authors showed an increase of

around 17% for the flux of TiO2-entrapped PES in comparison

with the observed flux for neat PES membrane.

It can be concluded that it is possible to achieve an improve-

ment in antifouling membrane properties in the low concentra-

tion range proposed in this work.

J. Mater. Chem., 2011, 21, 10311–10320 | 10317


Dow

nloa

ded

by K

atho

lieke

Uni

vers

iteit

Leu

ven

on 1

2 M

arch

201

2Pu

blis

hed

on 2

7 M

ay 2

011

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C1J

M11

040C

View Online

In addition, the flux rapidly declined at the beginning of each

filtration experiment for all membranes. However, membrane

saturation, a condition of the membrane performance when the

membrane shows a stationary state on the flux, is faster reached

by the neat membranes (ca. 5 h). The entrapped membranes were

observed to have a nearly constant flux after ca. 10 h. Organic

adsorption is regarded as the first step in membrane fouling and

is strongly dependent on the physico-chemical properties of

membranes and foulants, especially the affinity of these foulants

towards the membrane material. It has been found that organic

substances have a higher tendency to adsorb and deposit on

hydrophobic surfaces and hence, adsorption occurs relatively

quickly for hydrophobic compounds resulting in a stable

normalized flux.4 It can be concluded that adsorption of humic

acids on the surface of neat membrane results in a small time

delay due to the more hydrophobic character of the polymeric

material in comparison with the improved (more hydrophilic)

TiO2 entrapped membranes. This behavior has been confirmed

previously by modeling of flux decline based on the adsorption

thermodynamics of organic compounds on the nanofiltration

membrane surface.5

Fig. 8 Observed dye rejection for manufactured membranes as a func-

tion of (a) dye molecular weight and (b) PES concentration.

Fig. 9 Comparison of pure water permeation and methylene blue

rejection for PES/TiO2 membranes with different nanoparticle contents.

3.6. Dye rejection potential

Physical sieving by pores is believed to be one of the main driving

forces in the rejection of organic components. The molar mass is

an easily available parameter to describe the molecular size;

therefore, a comparison of the rejection capacity of neat and

entrapped membranes for different commercial dyes with

increasing molar mass was determined (see Fig. 8a). Fig. 8b

shows the effect of polymer concentration on rejection of

methylene blue dissolved in aqueous solution (10 mg L�1).

The rejection trend of selected dyes by the membranes

confirms the typical tendency for the rejection of organic

compounds by nanofiltration membranes, which can be quanti-

fied by using MWCO models.45 The addition of TiO2 nano-

particles does not yield a large improvement in the rejection

potential of manufactured membranes. However, the polymer

concentration has a significant effect on the rejection perfor-

mance of membranes. The gradual increase in the rejection

potential of the membranes with higher polymer content that has

been observed is assumed to be caused by the pore size and

porosity decrease. However, the observed decrease in flux decline

as the polymer content increases is thought to be caused by the

increase in the membrane’s hydrophobicity.

The influence of TiO2 concentration on the rejection perfor-

mance of membranes was also studied, maintaining the PES

concentration at 27 wt% (Fig. 9). Permeability values obtained

for every membrane were also added in Fig. 9.

As explained above, the TiO2 nanoparticles—in addition to

their hydrophilic character—are not to be considered as a pore

formation additive during the immersion precipitation-process,

as a consequence of their hindrance effect on the solvent/water

diffusion rate. While other additives that are also used as pore

former, such as polyvinylpyrrolidone and polyethyleneglycol, are

soluble in water and hence are removed from the membrane

structure at the end of the phase inversion, the added TiO2

nanoparticles are still entrapped into the membrane. So, the

plugging of the membrane pores by nanoparticles during

10318 | J. Mater. Chem., 2011, 21, 10311–10320

immersion precipitation is another intrinsic effect that should be

considered for the understanding of TiO2 entrapped membrane

performance, also at ultralow concentrations.27

The methylene blue rejection decreases as the TiO2 content

increases up to 0.085 wt%, which is in agreement with the

permeability improvement. At higher concentrations, the rejec-

tion of dyes was improved. Along the lower concentration



Dow

nloa

ded

by K

atho

lieke

Uni

vers

iteit

Leu

ven

on 1

2 M

arch

201

2Pu

blis

hed

on 2

7 M

ay 2

011

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C1J

M11

040C

View Online

interval the observed rejection trend could be associated with an

increase in the pore size and porosity of the membrane. At

concentrations above 0.085 wt%, it is postulated that the larger

pores of the membranes are plugged by nanoparticles, causing an

increase in the rejection of organic solutes. Water molecules are

small and they pass through the membranes more easily; the

permeability is not strongly affected. As explained above, the

increase in the concentration of nanoparticles can be accompa-

nied by unfavourable phenomena such as particle aggregation,

which causes a decrease in the dispersion rate of TiO2 in the

polymeric solution, and therefore in the membrane structure.

While decreasing the specific number of (larger) nanoparticles,

which could be clusters or less aggregated nuclei, the overall

pore-forming capacity of TiO2 as additive should be lower and

therefore the pore size distribution changes, with the more

probable size closer to the average pore size of the neat

membrane. From the size interaction analysis point of view, the

dye rejection observed for TiO2-entrapped membranes at

concentrations above 0.085 wt% was similar to the rejection for

the neat membrane, so steric hindrance (pore size) of both types

of membranes is, on average, comparable. This fact has been

observed by some researchers reporting that a further increment

in the concentration of metal oxide nanoparticles (above the

optimum value) could lead to the formation of defects and stress

convergence points, which negatively affect some of the

improved membrane properties such as intrinsic membrane

resistance,29 mean pore size,31,36 mechanical stability,31

porosity,30,36 and MWCO.30

4. Conclusions

An appropriate selection of nanoparticle concentration is crucial

in order to achieve the expected improvement in TiO2 entrap-

ment membranes. The negative effect of the critical dispersion

rate of nanoparticles in the polymeric solution and hence, inside

of the membrane structure, was shown to have a determining

influence.

The particle size distributions determined for organic suspen-

sions showed that as the TiO2 content increased the average

diameter was higher, having a lower peak which covered a wider

size range.

Morphological analysis of membranes performed by SEM

showed that the TiO2-entrapped membranes have a more open

structure in comparison with control membranes, having more

and larger macrovoids inside. The presence of TiO2 was

confirmed comparing the residual weights obtained for control

and TiO2-entrapped membranes. An inverse correlation was

found between porosity and mechanical strength of tested

membranes. The effect of composition (addition of TiO2) is

lower upon the mechanical properties than the improved

porosity.

A comparison between hydrophilicity and permeability for

control and TiO2-entrapped membranes showed that the pres-

ence of nanoparticles provides a higher affinity of blended

membranes as a consequence of water affinity and pore forma-

tion properties added nanoparticles.

Thus, the unusually low range of TiO2 concentrations

proposed in this work provides a significant improvement in the

permeability and fouling resistance of blended membranes. Both


for hydrophilicity and permeability an apparent optimum value

was found for the concentration at 0.085 wt%. It should also be

noted that even at lower concentrations of nanoparticles unde-

sirable effects such as aggregation have been observed; this yields

a less improved membrane performance, in some cases similar to

the former control membranes.

Acknowledgements

This work was supported by the National Research Foundation

of Korea and FWO-Vlaanderen (Fund for Scientific Research)

through an international joint research program between

Republic of Korea and Belgium (F01-2009-000-10045-0). Arca-

dio Sotto would like to acknowledge the support provided by the

Regional Government of Madrid through project REMTA-

VARES (S2009/AMB-1588). Stefan Balta would like to

acknowledge the support provided by the European Union,

Romanian Government and Dunarea de Jos University of Gal-

ati, through the project POSDRU—6/1.5/S/15.

References

1 B. Van der Bruggen, M. M€antt€ari and M. Nystr€om, Sep. Purif.Technol., 2008, 63, 251.

2 P. Xu, J. E. Drewes, T.-U. Kim, C. Bellona and G. Amy, J. Membr.Sci., 2006, 279, 165.

3 Y. Zhao, L. Song and S. L. Ong, J. Membr. Sci., 2010, 349, 75.4 J. M. Arsuaga, M. J. L�opez-Mu~noz and A. Sotto, Desalination, 2010,250, 829.

5 L. Braeken, K. Boussu, B. Van der Bruggen and C. Vandecasteele,ChemPhysChem, 2005, 6, 1606.

6 L. D. Nghiem, P. J. Coleman and C. Espendiller, Desalination, 2010,250, 682.

7 C. Guizard, A. Bac, M. Barboiu and N. Hovnanian, Sep. Purif.Technol., 2001, 25, 167.

8 P. C. Chiang, W. T. Whang, M. H. Tsai and S. C. Wu, Thin SolidFilms, 2004, 447/448, 359.

9 A. Taniguchi and M. Cakmak, Polymer, 2004, 45, 6647.10 Z. H. Lua, G. J. Liu and S. Duncan, J. Membr. Sci., 2003, 221, 113.11 O. Y. Lu, D. M. Dotzauer, S. R. Hogg, J. Macanas, J. F. Lahitte and

M. L. Bruening, Catal. Today, 2010, 156, 100.12 J. Xu and D. Bhattacharyya, J. Phys. Chem. C, 2008, 112, 9133.13 S. H. Kim, S. Y. Kwak, B. H. Sohn and T. H. Park, J. Membr. Sci.,

2003, 211, 157.14 M. Roso, S. Sundarrajan, D. Pliszka, S. Ramakrishna and

M. Modesti, Nanotechnology, 2008, 19, 285707.15 S. Y. Kwak, S. H. Kim and S. S. Kim, Environ. Sci. Technol., 2001, 35,

2388.16 J. B. He, P. Kanjanaboos, N. L. Frazer, A. Weis, X. M. Lin and

H. M. Jaeger, Small, 2010, 13, 1449.17 B. J. Hinds, N. Chopra, T. Rantell, R. Andrews, V. Gavalas and

L. G. Bachas, Science, 2004, 303(5654), 62.18 S. A. El-Safty, A. Shahat, W. Warkocki and M. Ohnuma, Small,

2011, 7, 2.19 J. Kim and B. Van der Bruggen, Environ. Pollut., 2010, 158, 2335.20 S. Yu, X. Zuo, R. Bao, X. Xu, J. Wang and J. Xu, Polymer, 2009, 50,

553.21 L. Yana, S. L. Yu and C. B. Xiang, Polymer, 2005, 46, 7701.22 S. H. Kim, S.-Y. Kwak and B.-H. Sohn, J. Membr. Sci., 2003, 211,

157.23 W. Choi, A. Termin and M. R. Hoffmann, J. Phys. Chem., 1994, 98,

13669.24 A. D. Syafei, C. F. Lin and C. F. Wu, J. Colloid Interface Sci., 2008,

323, 112.25 T. H. Bae and T. M. Tak, J. Membr. Sci., 2005, 249, 1.26 J. B. Li, J. W. Zhu and M. S. Zheng, J. Appl.Polym. Sci., 2007, 103,

3623.27 T.-H. Bae, I.-C. Kim and T.-M. Tak, J. Membr. Sci., 2006, 275, 1.

J. Mater. Chem., 2011, 21, 10311–10320 | 10319


Dow

nloa

ded

by K

atho

lieke

Uni

vers

iteit

Leu

ven

on 1

2 M

arch

201

2Pu

blis

hed

on 2

7 M

ay 2

011

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C1J

M11

040C

View Online

28 Y. Mansourpanah, S. S. Madaeni, A. Rahimpour, A. Farhadian andA. H. Taheri, J. Membr. Sci., 2009, 330, 297.

29 A. Rahimpour, S. S. Madaeni, A. H. Taheri and Y. Mansourpanah,J. Membr. Sci., 2008, 313, 158.

30 I. Soroko and A. Livingston, J. Membr. Sci., 2009, 343, 189.31 G. Wu, S. Gan, L. Cui and Y. Xu, Appl. Surf. Sci., 2008, 254,

7080.32 X. Cao, J. Ma, X. Shi and Z. Ren, Appl. Surf. Sci., 2006, 253,

2003.33 V. V. Tarabara in Nanotechnology Applications for Clean Water, ed.

Savage, et al., William Andrew Inc., 2009, p. 59.34 B. A. Rozenberg and R. Tenne, Prog. Polym. Sci., 2008, 33, 40.35 J. S. Taurozzi, H. Arul, V. Z. Bosak, A. F. Burban, T. C. Voice,

M. L. Bruening and V. V. Tarabara, J. Membr. Sci., 2008, 325,58.

36 Y. Yang, H. Zhang, P. Wang, Q. Zheng and J. Li, J. Membr. Sci.,2007, 288, 231.

37 R. A. Damodar, S.-J. Youa andH.-H. Chou, J. Hazard.Mater., 2009,172, 1321.

38 H. Lin, C. P. Huang, W. Li, C. Ni, S. Ismat and T. Yao-Hsuan, Appl.Catal., B, 2006, 68, 1.

10320 | J. Mater. Chem., 2011, 21, 10311–10320

39 H. Susanto and M. Ulbricht in Membrane Operations. InnovativeSeparations and Transformations ed. E. Drioli and L. Giorno,Wiley-VCH, Weinheim, 2009, pp. 26–30.

40 J.-F. Li, Z.-L. Xu, H. Yang, L.-Y. Yu and M. Liu, Appl. Surf. Sci.,2009, 255, 4725.

41 N. Li, C. Xiao, S. An and X. Hu, Desalination, 2010, 250, 530.42 A. Razmjou, et al., The effects of mechanical and chemical

modification of TiO2 nanoparticles on the surface chemistry,structure and fouling performance of PES ultrafiltrationmembranes, J. Membr. Sci., 2011, DOI: 10.1016/j.memsci.2010.10.019.

43 M.-J. Han and S.-T. Nam, J. Membr. Sci., 2002, 202, 55.44 I. C. Kim, K. H. Lee and T. M. Tak, J. Membr. Sci., 2001, 183, 235.45 A. R. Costa, M. N. de Pinho and M. Elimelech, J. Membr. Sci., 2006,

281, 716–725.46 M. J. L�opez-Mu~noz, A. Sotto, J. M. Arsuaga and B. Van der

Bruggen, Sep. Purif. Technol., 2009, 66, 194–201.47 K. Boussu, C. Vandecasteele and B. Van der Bruggen, Polymer, 2006,

47, 3464–3476.48 B. Van der Bruggen, K. Everaert, D. Wilms and C. Vandecasteele, J.

Membr. Sci., 2001, 193, 239–248.



Doping of Polyethersulfone Nanofiltration Membanes Antifouling Effect Observed at Ultralow...

Documents

Transcript of Doping of Polyethersulfone Nanofiltration Membanes Antifouling Effect Observed at Ultralow...