Self-assembled silica nanocrystal-based anti-biofouling nanofilter membranes

Article abstract of DOI:10.1039/c2ra21135a

Herein, we report synthesis of organosiloxane 3-(2-((3-aminopropyl) diethoxysilyl)ethylthio)-5-(4-((3-aminopropyl)diethoxy silyl)phenyl)-4H-1,2,4- triazol-4-amine (TS) by Barbiar-Grignard reaction. Hybrid nanofiltration (NF) membranes were prepared from T

Full text of DOI:10.1039/c2ra21135a

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Self-assembled silica nanocrystal-based anti-biofouling
nanofilter membranes3
Cite this: RSC Advances, 2013, 3, 458
Ajay K. Singh, Ravi P. Pandey, Amaranadh Jasti and Vinod K. Shahi*
Herein, we report synthesis of organosiloxane 3-(2-((3-aminopropyl)diethoxysilyl)ethylthio)-5-(4-((3-
aminopropyl)diethoxy silyl)phenyl)-4H-1,2,4-triazol-4-amine (TS) by Barbiar–Grignard reaction. Hybrid
nanofiltration (NF) membranes were prepared from TS and poly(vinyl alcohol) (PVA) via sol–gel process
followed by cross-linking and grafting of phosphonic acid groups. Physicochemical properties of these
membranes revealed their stable (thermal, mechanical, and chemical) and chlorine tolerant nature. HRTEM
analysis reveals homogenous silica distribution in the membrane matrix. The cross-link density of the
membrane and the preparation conditions were studied in terms of NF performance. An Escherichia coli
bacterium was used to study antibacterial activity and anti-biofouling properties of the hybrid membrane.
The short-term bacterial stability test showed that membrane TS-60a has good anti-fouling properties.
Moreover TS-60a membrane showed excellent stability and anti-biofouling capability in long-time
operation.
Received 7th June 2012,
Accepted 29th October 2012
DOI: 10.1039/c2ra21135a
www.rsc.org/advances
Examination of fouled NF membranes exhibited more than
50% (w/w) of dry fouling layer, which was of biological
Introduction
origin.18 To avoid biofouling, generally anti-adhesion and
antimicrobial strategies for NF/RO membranes have been
adopted.19 In antimicrobial strategies generally silver nano-
particles have been used, which does not comply with the
regulations of the United States Environmental Protection
Agency (USEPA) for water quality (silver content should be less
than 3.4 ppb). Leaching of silver nanoparticles may pose a
regulatory risk for applications in water treatment.19 Charged
or hydrophilic modified NF membranes with a smooth surface
are less sensitive to biofouling.19 Anti-biofouling NF mem-
branes are challenging because of their architecture (pore size,
surface roughness and charged nature), hydraulic resistance,
separation performance, mechanical stability, chemical stabi-
lity, cost and flexibility of assembly.
Population growth, industrialization, and contamination
(personal care products, agricultural pesticides, heavy metals,
micro-pollutants, drugs, chemicals, and microbes) are serious
problems for providing safe and clean drinking water.1–4
Membrane technologies for water desalination/purification
have a wide range of advantages and attract wide attention for
developing new types of stable and sustainable membranes.4
Nanofiltration (NF) membranes show many advantages such
as low energy consumption, and high permeation flux
performance over reverse osmosis (RO).5,6 The surface charge
of NF membranes plays an important role due to Donnan
exclusion and is responsible for high flux, salt rejection and
fouling resistivity.7,8
Generally, NF membranes have been derived from poly-
amide,9 sulphonated polysulfone,10 and sulphonated polyphe-
nylene oxide.11 At present, NF membranes are prepared by
different techniques such as phase inversion,12 chemical
cross-linking,13 photo-radiation plasma,14 low-temperature
plasma,15 and layer-by-layer deposition technique.16
Polyamide membranes are chemically unstable in the pre-
sence of sanitizing agents, especially chlorine.17 Furthermore,
the membrane is itself susceptible to fouling (organic fouling,
inorganic fouling, colloidal fouling, and biofouling).
Attempts were made to develop organic/inorganic hybrid
NF membranes due to their rigidity, stability (cross-linking),
and ease of tailoring the pore architecture and processabil-
ity.20 Thus, there is great opportunity for development of
organic/inorganic hybrid membranes using polysiloxanes and
poly(vinyl alcohol) (PVA). Polysiloxanes are used for biomedi-
cal applications,21–25 while PVA forms chlorine-tolerant stable
ether-type linkages due to acetalisation with formaldehyde.26
We reported 3-((4-(5-(2-((3-aminopropyl)diethoxysilyl)ethylthio)-
1,3,4-oxadiazol-2-yl)phenyl)diethoxysilyl)propan-1-amine (APDSMO)-
based quaternary ammonium-grafted hybrid NF membranes
with antimicrobial activity.27 But, the 1,2,4-triazole ring also
showed antibacterial and antifungal activity.28 Thus, the hybrid
NF membrane with 1,2,4-triazole ring is also expected to exhibit
antimicrobial activity, good performance and stabilities.
Electro-Membrane Processes Division, Central Salt and Marine Chemicals Research
Institute, Council of Scientific & Industrial Research (CSIR), G.B. Marg, Bhavnagar-
364002 (Gujarat), India. E-mail: vkshahi@csmcri.org; vinodshahi1@yahoo.com;
Fax: +91 278 2567562/6970; Tel: +91-9925125760
Electronic supplementary information (ESI) available. See DOI: 10.1039/
3
c2ra21135a
458 | RSC Adv., 2013, 3, 458–467
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
Herein, we report phosphonic acid-grafted hybrid NF
membranes, containing 4-amino-1,2,4-triazole. Membrane
preparation was achieved in aqueous media by acid-catalyzed
sol–gel followed by cross-linking. Grafting of the phosphonic
acid group, tailoring of pore texture and incorporation of
antimicrobial active compounds, were responsible for anti-
biofouling and high membrane performance.
Experimental section
Materials
Aminopropyltriethoxysilane (APTEOS) (99%) and polyethylene
glycol (PEG; different molecular weights) were obtained from
Sigma-Aldrich chemicals. Poly(vinyl alcohol) (Mw: 125 000),
formaldehyde (37% in water), phosphorous acid, hydrochloric
acid, sulphuric acid, p-chlorobenzoic acid, hydrazine hydrate,
carbon disulphide, sodium hydroxide, sodium chloride, sucrose,
magnesium chloride, sodium hypochlorite, N-methylpyrrolidone,
dimethylformamide, hexane, iodine crystals, magnesium turn-
ings, tetrahydrofuran (THF), streptomycin, fluconazole, acetone
and methanol were obtained from SD. Fine Chemicals, India.
Solvents were used after proper distillation and double-distilled
water was used for experiments.
Synthesis of organosiloxane (TS)
Synthesis of 3-(2-chloroethylthio)-5-(4-chlorophenyl)-4H-1,2,4-
triazole-4-amine (TC) was achieved through multistep synth-
esis (Scheme 1; details of TC synthesis are included in section
S1, ESI ). Synthesis of organosiloxane (TS) has been reported
3
earlier.20,27,29 In a general preparation method, to a three-
necked round-bottom flask (250 ml, equipped with a magnetic
stirrer, nitrogen inlet, drying tube, condenser, and addition
funnel), a mixture of Mg turnings (1.5 g) and APTEOS (44.4 ml;
0.2 mol) in THF (30 ml) and iodine crystals were added under
constant stirring conditions. After 30 min, 20 mL solution of 3-
(2-chloroethylthio)-5-(4-chlorophenyl)-4H-1,2,4-triazol-4-amine
(TC) (2 mmol) in THF was added dropwise over 30 min, then
further refluxed at 10 h. After 10 h the clear solution turned to
a greenish yellow liquid. The reaction mixture was cooled at
room temperature and THF was removed by vacuum rotary
evaporator. Viscous liquid was washed with 30 ml hexane and
dried under vacuum.
Yield: 80% (yellow coloured transparent semisolid). IR (KBr)
cm21: 3395 (–NH), 2981 (–Ar–H), 1590 (CLN triazole), 1528
(NH), 1317 (N–N), 1013 (Si–O–C), 760 (Si–Ph), 681(C–Si); 1H
NMR (D2O): d 7.67 (Ph–Si), 7.28 (Ph–triazole), 3.47, 1.02 (–O–
CH2–CH3), 3.43 (–OCH2CH3), 2.63, 1.47 (–SCH2CH2), 2.57,
(–CH2–CH2–Si–) 0.99(–CH2Si–Ph), 0.43,0.45 (–CH2Si); 13C NMR
(DMSO-d6): d 174.68, 173.55, 136.3, 130.54, 128.18, 17.66,
57.23, 42.49, 30.25, 23.07, 16.91, 10.27. Anal. calcd for
[C24H46N6O4SSi2] (570.91): C, 50.45; H, 8.15; N, 14.76; S, 5.64,
Si, 9.87%. Found: C, 50.5; H, 8.13; N, 14.71; S, 5.61; Si, 9.84%.
Scheme 1 Schematic reaction steps involved in preparation of organosiloxane
(TS).
Obtained solutions were transformed into a gel by adding 1 M
HCl and pH was maintained according to Table 1. The
resultant white viscous gels were cast on a PVC sheet and the
desired thickness was maintained with a doctor blade. After
partial drying (10 min), thin films were gelated (precipitated)
in hexane at 10 uC for 20 min. The membrane was further
dried at room temperature for 24 h (slow evaporation of
solvent) followed by in a vacuum oven at 80 uC for 24 h. The
obtained membrane was cross-linked by formal reaction under
varied cross-linking conditions (Table 1). Cross-linked mem-
branes were subjected to phosphorylation using 1 : 1 (w/w)
formaldehyde and phosphorus acid, for 3 h at 70 uC. Different
prepared membranes were designated as TS-X(y), where X is
the weight percentage of TS, and y denotes different cross-
linking conditions, as described in Table 1.
Preparation of hybrid NF membrane
Instrumental characterization of the membranes
PVA was dissolved in distilled water (10 wt%) under stirring
conditions for 24 h. Further, TS (desired amount) was added to
PVA solution under constant stirring (300 rpm) for 6 h.
FTIR spectra of dried membrane samples were obtained
Spectrum GX series 49 387 spectrometer in the range of 4000–
450 cm21. The IR spectrum for a synthesized intermediate was
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 458–467 | 459

View Article Online
Paper
RSC Advances
Table 1 Cross-linking densities and preparation conditions for different membranes
Membrane preparation conditions
Membrane
Formaldehyde concentration (wt%)
Time (h)
pH
T/uC
Cross-linking density (mol g21
)
TS-30
TS-40
TS-50
2.5
2.5
2.5
2.5
2.0
1.5
1.0
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
3
3
3
3
3
3
3
4
2
1
3
3
3
3
3
3
3
2
2
2
2
2
2
2
2
2
2
3
4
5
6
2
2
2
60
60
60
60
60
60
60
60
60
60
60
60
60
60
70
50
40
0.009
0.0302
0.061
TS-60a
TS-60b
TS-60c
TS-60d
TS-60e
TS-60f
TS-60g
TS-60h
TS-60i
TS-60j
TS-60k
TS-60l
TS-60m
TS-60n
0.0808
0.0696
0.0454
0.0062
0.0542
0.0122
0.0818
0.0622
0.0316
0.0102
0.0002
0.082
0.011
0.0068
Error limits: Formaldehyde concentration (¡0.1 wt%), time (¡0.05 h), pH (¡0.05), temperature (¡2 uC), and cross-linking density (¡0.0002).
obtained by the KBr pellet method. Wide-angle X-ray diffracto-
grams of the nanocomposite membranes were recorded using
Philips Xpert X-ray diffractometer with Cu-Ka (1.54056)
radiation. 1H and 13C were used to characterize the synthe-
sized material recorded by an NMR spectrometer (Bruker 500
MHz) in a D2O and d6-DMSO solvents.
The thermal degradation processes and stabilities of the
membranes were investigated using a thermo gravimetric
analyzer (Mettler Toledo TGA/SDTA851 with Star software)
under a nitrogen atmosphere with a heating rate of 10 uC
min21 from 30 to 450 uC. Differential scanning calorimetry
(DSC) measurements were carried out in a temperature range
of 30–450 uC with a heating rate of 5 uC min21. The dynamic
mechanical stabilities of the composite membranes were
evaluated by using a Mettler Toledo dynamic mechanical
analyzer 861 instruments with Star software under nitrogen
with a heating rate of 10 uC min21 from 30 to 320 uC to verify
the effect of the silica content in the form of a polymeric
membrane.
The surface morphology of thoroughly dried membranes
was studied by a JEOL 1200EX transmission electron micro-
scope (TEM). The JEOL 1200EX transmission electron micro-
scope with tungsten electron source was operated at an
accelerating voltage of up to 120 kV. Atomic force microscopy
(AFM) images of dried membranes were recorded using an
NTEGRA AURA (NTMDT). Semi-contact mode SPM NSG 01 tip
was used to determine the surface roughness. The tip used
had a radius of curvature of approximately 10 nm and the
natural frequency for the cantilever was 300 kHz. In this
measurement mode, the cantilever where the tip is located
oscillates with its natural frequency and the sample topogra-
phy is obtained from the subsequent changes in the oscillation
amplitude. Differences in viscoelastic properties can be
detected from the changes in the oscillation phase. For
scanning electron microscopy (SEM), gold sputter coatings
were carried out on desired membrane samples at pressures
ranging between 1 and 0.1 Pa. The sample was loaded in the
machine, which was operated at 1022 to 1023 Pa with EHT
15.00 kV with 300 V collector bias using Leo microscope to
record SEMs. The optical densities of microbial solutions were
evaluated by using a VARIAN 50 bio UV-Vis spectrophotometer
instruments. The Total Organic Count (TOC) content in the
feed and permeate solutions was obtained by employing a
customized acid digestion method, followed by analysis with a
liquid TOC elementar.
Membrane permeation measurements
Water flux across hybrid membranes was measured in a two-
compartment cell separated by circular membrane disk (4.0
cm2) placed at the bottom of the cell with top active layer
towards the feed solution. Feed compartment was well stirred
by a mechanical stirrer (300 rpm). Electrolyte concentration in
the feed and permeate was determined by conductivity
measurements (Toshniwal TCM 15).30,31 For the neutral
organic molecule solutions, a customized total organic carbon
(TOC) digestion method was used to quantitatively measure
the concentration of neutral organic solutes in the permeate
and in the retentate. Calibration plots for all three types of
analyses were run with standard feed solutions prior to the
studies to ensure accuracy of the measurements. Water flux
and solute rejection were estimated by following equations:
permeate (l)
membrane area (m2)|time (h)
Flux (l m{2
h
{1)~
(1)
ꢀ
ꢁ
permeate concentration
Rejection~ 1{
(2)
feed concentration
The chlorine-tolerant nature of the membranes was assessed
in terms of flux and rejection performance, after membrane
treatment in sodium hypochlorite (pH 5–6) aqueous solution
(5000 ppm) solution for 24 h.
460 | RSC Adv., 2013, 3, 458–467
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
A carbon peak for –OCH2– (57.23 ppm) confirmed the presence
of a –OCH2CH3 group in TS.
Antibacterial activity
Antibacterial properties of hybrid (TS-X) NF membranes were
tested against Escherichia coli bacteria by using reported
methods.20,27 Standard testing protocol for water-insoluble
antimicrobials was used.32,33 Antibacterial tests were per-
formed in triplicate to insure reproducibility.
Guide peaks in FT-IR spectra of TC (Fig. S3, ESI ) confirmed
3
reaction between 5-(4-chlorophenyl)-1,2,4-triazole-4-amine-2-
thiol and 1,2-dichloroethane in the presence of weak base.
TC showed two medium intensity bands at ca. 745 and 661
cm21 (n(Ph–Cl) and n(C–Cl)). After Barbiar–Grignard reaction
these two bands disappeared and new bands appeared at ca.
776 and 691 cm21 (n(Si–Ph) and n(Si–C) vibration band).
Presence of triazole ring in TC and TS was confirmed by the
strong intensity band at ca. 1080 and 1095 cm21 (C–N–C
stretching vibration) (Fig. 1). TS showed medium intensity
bands at 3433 cm21 and 1552 cm21 (n(N–H), while the band at
ca. 1596 cm21 (n(CLN) was assigned to the triazole ring.33 In
the case of TS-60a membrane, a broad band at ca. 1400–1445
cm21 and a strong band at ca. 1649 cm21 confirmed
phosphorylation of the –NH group. The absorption band at
around 1260 cm21 was a characteristic band of the –Si–O–Si–
Results and discussion
Membrane structure
The monomer precursor (TS) was synthesized by Barbier–
Grignard reaction using TC and APTEOS (Scheme 1). H NMR
spectra for TS showed two peaks at 7.26 and 7.67 ppm due to
1
phenyl ring protons (Fig. S1, ESI ). The presence of triazole
3
and phenyl rings were confirmed by 13C NMR signals at d value
174.6, 173.5 and 136.3, 130.5, 128.1, 117.6 ppm (Fig. S2, ESI ).
3
Fig. 1 FT-ATR spectra of TS-60a and TS-30 membrane.
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 458–467 | 461

View Article Online
Paper
RSC Advances
asymmetrical stretching vibration. This confirmed molecular
level hybridisation between organic and inorganic segments.34
The band at ca. 1021 cm21 (–C–O–C– stretching vibration for
acyclic ether ring of PVA) showed formal cross-linking with
Membrane fouling is controlled by interaction between the
membrane surface and solute such as protein, carbohydrates
and bacteria. Thus surface roughness of the membrane is
important in determining the susceptibility to membrane
fouling. Previously, a number of studies suggested that surface
roughness may be the most important factor to avoid
membrane fouling.19 Membrane surface images were obtained
in semi-contact mode of AFM study and the pictures are
systematized in Fig. 4. For the surface roughness ana-
lysis, measurement of a small portion of the membrane
(1.4 61.4 mm) were selected and NOVA P9 INTEGRA software
was used for analysis. Obtained data concluded that mem-
brane roughness decreased with increasing organosiloxane
(TS) content. TS-60a membrane has the lowest surface rough-
ness 55 nm.
formaldehyde. These results confirmed formation of
a
phosphonic acid-functionalized cross-linked hybrid mem-
brane. On the basis of spectral and elemental analysis,
structures for TS and hybrid NF membranes were proposed
(Scheme 1 and Scheme S1 (ESI)).
Surface morphologies of hybrid NFM
TEM images (TS-60a membrane as representative case) (Fig. 2)
revealed homogeneous silica distribution in the membrane
matrix. HRTEM (Fig. 2a&b) analysis confirmed 2–5 nm particle
size of silica and inter-planar distance was below 0.24 nm
confirming the crystalline silica in the membrane matrix.35
WXRD analysis showed an increase in the amorphous region
with TS content (reduced PVA content), which is a favourable
condition for formation of Si–O–C bonds (availability of more
Thermal, mechanical, and chemical stability
Thermal stabilities of pristine and hybrid membranes were
analysed by thermogravimetric analysis (TGA) (Fig. S6, ESI ).
3
silica) (Fig. S4, ESI ). The absence of evident peaks in TS
showed its amorphous nature.
Scanning electron microscopy results are systematized in
The NF membranes showed two-steps of weight loss. The first
step (50–170 uC) occurred due to dehydration, while the
second step (280–450 uC) occurred due to degradation of the
organic segment. After 450 uC, no significant weight loss was
observed. TS-60a membrane showed higher weight loss (60%)
in 300–400 uC, indicating that more phosphoric acid is
contained in this membrane. At 500 uC, the remaining weight
indicates that membrane TS-60a was thermally stable in
comparison to other synthesized membranes. DSC thermo-
3
Fig. 3 and Fig. S5, (ESI ). Surface analysis of membrane TS-60a
3
& TS-30 showed the upper surface is dense (Fig. 3a&c), while
the bottom surface in the membrane has a porous nature
(Fig. 3b&d). The bottom surface of the TS-60a membrane looks
like a honeycomb structure (Fig. 3b). Membrane cross-section
SEM images showed that the membranes are arranged with a
layer morphology and the upper surface in dense in nature
while lower surface is porous in nature (Fig. 3c&f). The surface
morphology of the synthesised organosiloxane looks like a
grams (Fig. S7, ESI ) revealed dependence of T on organosi-
3
g
loxane (TS) content. The increase in glass transition
temperature (Tg) (85–114 uC) from TS-30 to TS-60a membrane
indicates that cross-linking and ionic interactions between
respectively organic and inorganic segments increased. The Tg
value of pristine PVA membrane was 78 uC.36
Mechanical stability is another problem for NF membranes.
We determined Young’s modulus with DMA analysis. With the
help of the Young’s modulus, membranes’ cross-linking
densities were determined by following equation.37
crushed stone (Fig. S5a, ESI ) and the elemental mapping data
3
support their elemental analysis (Fig. S5b, ESI ). The success-
3
ful grafting of the membrane surface by phosphonic acid was
confirmed by the phosphorous mapping (green dot) in TS-60a
and TS-30 membrane (Fig. S5c&d, ESI ). The synthesized
3
organosiloxane (TS) elemental analysis is further supported by
the EDX analysis (Fig. S5e, ESI ).
3
Fig. 2 TEM images of NF membrane (TS-60a): (a) low resolution and (b) high resolution (linear inter-planar lines were separated by 0.24 nm).
462 | RSC Adv., 2013, 3, 458–467
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
Fig. 3 SEM images of (a) TS-60a, (b) TS-60a (bottom portion), (c) cross-section of TS-60a, (d) TS-30, (e) TS-30 (bottom portion), (f) cross-section of TS-30 membrane.
resistivity of the hybrid membranes was evaluated by their
weight loss under different environments for 30 days (Fig.
E’
3dQRT
r~
(3)
S10b, ESI ). About 2.9–5.0% weight loss in 30 days under
3
different environments (NMP, DMSO, DMF and THF) con-
firmed solvent resistivity. The chemical stability test indicates
that the triazole moiety was stable in the membrane matrix.
Oraganosiloxane is chemically bonded with plasticizer (PVA)
and cross-linked with formaldehyde; thus there is no chance of
its leaching out from the membrane matrix.
where, the Young’s modulus, E9, was determined from the
measurements with a dynamic mechanical analyser under 10
Hz frequency at varying temperature (30–320 uC) (Fig. S8, ESI ),
3
d is the membrane density, Q is the front factor (=1), R is
universal gas constant, and T is absolute temperature.
Cross-linking densities for different membranes are
depicted in Table 1. Cross-linking data revealed that with
increasing silica content (TS) in the membrane matrix,
Young’s modulus increased due to (Si–O–C) bonds formed in
organic PVA and organosiloxane (TS). Membrane preparation
conditions (temperature, pH, concentration of cross-linking
agent (formaldehyde), and cross-linking time) also play an
important role in mechanical stability. Thus it is necessary to
optimize the TS content in the membrane matrix, cross-
linking density and Young’s modulus to achieve mechanically
stable hybrid membranes.
Inter-chain spacing in the membrane matrix
Hybrid NF membranes were scanned in reflection mode at 2h
angle (Fig. S4, ESI ) and maximum peak intensity and inter-
3
planar distance values are given in Table S1 (ESI). The inter-
planar distance decreased with silica content in the membrane
phase, which may be due to the enhanced membrane
compactness. Diffraction patterns of hybrid NF membranes
confirmed reduction in the crystalline region with silica
content. Thus, incorporation of silica in the membrane phase
enhanced the amorphous region. Inter-planar distance of less
than 0.35 nm (value to ensure moderate amount of bonds Si–
O–C) also indicated condensation reaction (removal of a water
molecule) and hybridization of organic and inorganic seg-
ments by covalent bond (Si–O–C).35
The chlorine-tolerant nature of the hybrid membrane was
assessed in terms of weight loss after treatment of the
membrane in 5% NaClO(aq) solution for different time
intervals (0–72 h) (Fig. S9, ESI ). Weight loss for the hybrid
3
membrane increased with silica content. Moreover, no
appreciable weight loss was observed after 25 h treatment.
The acid stability of the membrane was also investigated by
weight loss of the membrane in equilibrium with H2SO4 (1–8
Membrane filtration performance
Membrane flux for different solvents (water, methanol,
ethanol, THF, DMF, DMSO, acetic acid and hexane; selected
due to varied dielectric constant) varied in a similar fashion to
their dielectric constant: water . methanol . ethanol .
M) for 30 days (Fig. S10a, ESI ). The membrane showed about
3
0–6% weight loss in equilibrium with 5 M H2SO4. Solvent
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 458–467 | 463

View Article Online
Paper
RSC Advances
Fig. 4 AFM 3D images of membrane upper surface for (a) TS-30a and (b) TS-60a.
DMSO . DMF . THF . acetic acid (Fig. 11, ESI ). Also,
membrane flux was dependent on silica content in the
membrane matrix.
similar variation trend with increasing cross-linking time.
Thus the observation indicates that both higher curing
temperature and longer curing time results in a higher degree
of cross-linking density and a dense membrane. Curing time
and temperature data revealed that 3 h cross-linking time and
60 uC temperatures to give a satisfactory NF membrane.
3
Effect of TS content on membrane performance
Fig. 5 shows the improvement in membrane performance with
TS content, and TS-60a membrane exhibited best separation
performance. At high organosiloxane content (.60%), mem-
branes turned unstable and brittle in nature. With increase in
silica content, membrane flux decreased (from 60.73 to 54.09 l
Effect of membrane-forming solution pH and membrane
thickness on their performance
Previous analysis in our lab concluded that acid-catalyzed sol–
gel of siloxane results a linear structure while base catalysis
forms a cluster-like structure.39 Here, we the investigated
effect of pH (membrane-casting solution) in the acidic region
(2.0–6.0) on membrane performance. The obtained results
indicate that flux reduced (64.7–52.08 l m22 h21) with lowered
m22
h
21), while salt rejection increased (from 0.63 to 0.28).
According to the ‘‘Space Charge’’ model,38 the surface charge
density and the pore radius are two important parameters for
NF membranes. The amount of phosphoric acid groups
increases with increasing TS concentration, which increase
the effective membrane surface charge density. On the other
hand, degree of cross-linking has been directly related to TS
content. Thus, salt rejection increases with increasing TS
concentration and vice versa for water flux.
Effect of formaldehyde concentration on nanofiltration
performance
The effect of cross-linking agent (formaldehyde) concentration
(1.0–2.5 wt%) on membrane performance was investigated for
TS-60 membrane (Fig. S12, ESI ). Flux decreased (from 62.05 to
3
54.09 l m22 21) while rejection increased (from 0.63 to 0.37)
h
with formaldehyde concentration. Above 2.5 wt% concentra-
tion of formaldehyde, no significant effect was observed on
membrane filtration performance.
Effect of cross-linking temperature and time
h
Membrane flux reduced (from 61.52 to 53.3 l m22 21), while
rejection increased (from 0.63 to 0.38) with increase in cross-
linking temperature (Fig. S13a, ESI ), which may be due to
Fig. 5 Performance of NF membranes at different TC content. Feed solution:
1.0 g L21 NaCl; applied pressure: 1.2 MPa; temperature: 25 uC.
3
enhanced cross-linking density. Fig. S13b (ESI ) shows a
3
464 | RSC Adv., 2013, 3, 458–467
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
pH, while rejection was enhanced (0.35–0.63). Beyond pH 3.0,
filtration performance was unaffected (Fig. S14, ESI ), because
3
of complete hydrolysis of siloxane.
The membrane thickness showed a significant effect on its
performance (Fig. S15, ESI ). With increasing membrane
3
thickness membrane flux decreased (54.03–12.96 l m22 h21
)
with minor alteration in rejection (from 0.66 to 0.63).
Membrane thickness analysis showed that 100 mm thick
membranes were suitable for NF processes.
Effect of feed concentration on nanofiltration performance
Membrane filtration experiments were performed at 25 uC and
1.2 MPa operating pressure with varied concentration of NaCl
solution (1–4 g L21) (Fig. 6). Increase in feed concentration
resulted in a decrease in membrane flux and rejection.40 This
observation was generally addressed by charged NF mem-
brane. Decrease in flux (Fig. 6) may be attributed to an
increase in osmotic pressure with NaCl concentration, which
further reduced the real NF driving force of TS-60a membrane.
A negatively charged membrane surface (phosphoric acid
functionalization) exchanges H+ with Na+ and forms an
electrical double layer at the membrane/solution interfacial
zone. The resulting Donnan potential increased with NaCl feed
concentration and thus deteriorates the salt rejection of TS-60a
membrane.
Fig. 7 Performance evaluation of TS-60a membrane for prolonged period
(other experimental conditions similar to Fig. 5).
performance. Organic probe molecules (sucrose and PEG)
showed relatively high rejection because of large molecule size
and absence of any interaction with membrane matrix.
Membrane anti-fouling behaviour
Filtration performance of TS-60a membrane was studied for
feed NaCl solution (1.0 gL21) at constant pressure (1.2 MPa)
(Fig. 7), and negligible alteration (less than 1%) was observed
after 300 h operation. The chlorine tolerant nature of the
membrane was assessed by deterioration in membrane
performance after 24 h treatment in NaOCl solution (5000
ppm) at 30 uC. Membrane performance was unaffected, maybe
due to acetylated PVA and chemically inert silica. A short-term
stability test for the prepared hybrid membranes was
performed at 1.2 MPa operating pressure with E. coli broth
solution, as reported earlier.43 The membrane maintained
performance after 50 h operation (Fig. 8a). 1 cm2 pieces of
Separation performance of different substances
Different probe molecules (inorganic electrolytes and organic
molecules) were used to study the membrane rejection
performance (Fig. S16, ESI ) and followed the trend: NaCl ,
3
MgSO4 , Na2SO4 , sucrose , PEG 2000. The same trend was
also observed for a commercial NF membrane grafted with
carboxylic acid groups (NTR 7450).41 TS-60 membrane showed
significantly higher rejection for divalent anions (SO422) in
comparison with monovalent anions (Cl2). This observation is
in accordance with Donnan exclusion theory,42 with strong
22
22
electrostatic repulsion between SO4
and –PO3
groups
(fixed to the membrane matrix). The charge and size of the
molecule/ion is also an important factor to control membrane
Fig. 8 Anti-biofouling performance of different NF membranes. (a) Variation of
membrane flux for TC-60a and TC-30 with time in bacterial environment (LB
media; feed solution 1.000 ¡ 0.001 OD at 600 nm) (other experimental
conditions similar to Fig. 5). (b) Optical image for 50 h used TC-60a membrane;
(c) optical image for 50 h used TC-30 membrane.
Fig. 6 Performance of TS-60a membrane for NaCl solutions with different
concentrations (other experimental conditions similar to Fig. 5).
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 458–467 | 465

View Article Online
Paper
RSC Advances
Table 2 Comparative performance of NF membranes
Membrane
RNaCl
JNaCl (l m22 h21
)
Testing conditions
Ref.
NF-7
PPO
QAPPESK
MO-6
0.37
0.36
0.31
0.55
0.68
0.63
21.3
63
52
52.13
14.88
54.09
1 g L21 NaCl; 25 uC; 6.0 bar
1 g L21 NaCl; 25 uC; 3.5 bar
1 g L21 NaCl; 25 uC; 4.0 bar
1 g L21 NaCl; 25 uC; 9.0 bar
2 g L21 NaCl; 25 uC; 11.0 bar
1 g L21 NaCl; 25 uC; 12.0 bar
44
45
6
20
OA-6
27
TS-60aa
This work
a
Maximum error is ¡0.01%.
used membrane (50 h) was immersed in 100 ml distilled water
at 300 rpm for 3 h, then 1 ml solution was dispersed in LB
Media (nutritionally enriched media for bacterial growth) and
the obtained pictures are presented in Fig. 8(b&c).
Antibacterial properties of hybrid NF membranes were tested
against Escherichia coli. MIC value for TS-60 hybrid membrane
showed 7000 mg ml21 MIC value (lower than for poly(vinyl
pyridine)).32 To explore bacteriostatic or bactericidal nature of
hybrid membranes, different weighted amounts of samples
(10–10 000 mg ml21) were incubated with E. coli in aqueous LB
broth for 24 h. 100 mL aliquots were placed on nutrient soft
agar plates followed by incubation at 37 uC for 24 h, and
bacterial colonies were counted. Hybrid membranes showed
bacteriostatic properties. For drinking water purposes, we
analysed the filtered water with the TOC analyzer and found
no measurable carbon, indicating that organosiloxane (TS)
does not leach out from membrane matrix. TS-60a membrane
showed minimum bacterial fouling because of small surface
roughness and high amount of TS content. TS-60a membrane
separation performance compared with previously reported
membranes (Table 2), confirming that this membrane is
suitable for NF filtration.
membranes retained their permeability after 30 days operation
in E. coli broth solution, which confirmed their effective anti-
biofilm forming activity.
Water-soluble TS constitutes a new family of materials that
are applicable to preparing charged NF membranes with good
anti-biofouling and stable nanofiltration performances.
Acknowledgements
Ajay K. Singh is thankful to Council of Scientific and Industrial
Research (CSIR), New Delhi, for providing Senior Research
Fellowship (Ext). Instrumental support received from
Analytical Science Division, CSMCRI, is also gratefully
acknowledged. The authors are thankful to a reviewer for very
relevant and supportive suggestions to improve the quality of
the manuscript.
References
1 M. Elimelech and W.A. Phillip, Science, 2011, 333, 712.
2 R. F. Service, Science, 2006, 313, 1088.
3 M. A. Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis,
B. J. Marinas and A. M. Mayes, Nature, 2008, 452, 301.
4 K.-S. Jang, H. J. Kim, J. R. Johnson, W. G. Kim, W. J. Koros,
C. W. Jones and S. Nair, Chem. Mater., 2011, 23, 3025.
5 J. He, X.M. Lin, H. Chan, L. Vukovi, P. Kral and
H.M. Jaeger, Nano Lett., 2011, 11, 2430.
6 Y. Ji, Q. An, Q. Zhao, H. Chen, J. Qian and C. Gao, J. Membr.
Sci., 2010, 357, 80.
7 R. B. Schoch, A. Bertsch and P. Renaud, Nano Lett., 2006, 6,
543.
8 M. R. Teixeira, M. J. Rosa and M. Nystrom, J. Membr. Sci.,
2005, 265, 160.
9 S. Verissimo, K. V. Peinemann and J. Bordado, J. Membr.
Sci., 2006, 279, 266.
10 A. R. Roudman and F. A. DiGiano, J. Membr. Sci., 2000, 175,
61.
11 K. Kim, G. Chowdhury and T. Matsuura, J. Membr. Sci.,
2004, 179, 43.
12 I.-C. Kim, H.-G. Yoon and K.-H. Lee, J. Appl. Polym. Sci.,
2002, 84, 1300.
13 D. A. Musale and A. Kumar, Sep. Purif. Technol., 2000, 21,
27.
14 A. Akbari, S. Desclaux, J. C. Remigy and P. Aptel,
Desalination, 2002, 149, 101.
Conclusions
Herein, we report preparation methodology of organosiloxane
(TS) monomer and anti-fouling stable NF membranes via sol–
gel reaction followed by formal cross-linking. The resultant
phosphonic acid-grafted NF membranes were uniform, trans-
parent, and flexible. It was possible to tailor the thermal,
chemical, and mechanical stabilities of these hybrid NF
membranes by incorporating organosiloxane (TS) in the
organic segment. The biofouling nature of these membranes
was confirmed by their unaffected filtration performance (TS-
60a) under
a bacterial environment. The representative
membrane (TS-60a) showed 54.09 l m22 h21 flux and 0.64
rejections for NaCl (1 g l21) feed solution. Prepared mem-
branes (TS-60a) exhibited good thermal, mechanical, and
chemical (solvent and chlorine) stability. Membrane charge
density and cross-linking density controlled the membrane
performance. We believe the prepared NF membranes will
have an impact as solvent-resistant membranes for the
petroleum industry. The reported method is also extendable
to develop other hybrid anti-biofouling NF membranes with
desired properties without silver nanoparticles. In addition,
466 | RSC Adv., 2013, 3, 458–467
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
15 Z. P. Zhao, J. Li, D. X. Zhang and C. X. Chen, J. Membr. Sci.,
2004, 232, 1.
16 G. Decher, Science, 1997, 277, 1232.
31 M. Zhou, P. R. Nemade, X. Lu, X. Zeng, E. S. Hatakeyama,
R. D. Noble and D. L. Gin, J. Am. Chem. Soc., 2007, 129,
9574.
17 H. B. Park, B. D. Freeman, Z. B. Zhang, M. Sankir and J.
E. McGrath, Angew. Chem., Int. Ed., 2008, 47, 6019.
18 R. P. Schneider, L. M. Ferreria, P. Binder, E. M. Bejarana, K.
P. Goes, E. Slongo, C. R. Machado and G. M. Z. Rosa, J.
Membr. Sci., 2005, 266, 18.
19 J. Mansouri, S. Harrisson and V. Chen, J. Mater. Chem.,
2010, 20, 4567.
20 A. K. Singh, P. Singh, S. Mishra and V. K. Shahi, J. Mater.
Chem., 2012, 22, 1834.
32 V. Sambhy, M. M. MacBride, B. R. Peterson and A. Sen, J.
Am. Chem. Soc., 2006, 128, 9798.
33 A. K. Singh, O. P. Pandey and S. K. Sengupta, Spectrochim.
Acta, Part A, 2012, 85, 1.
34 B. P. Tripathi and V. K. Shahi, J. Phys. Chem. B, 2008, 112,
15678.
35 Z. Xie, E. J. Henderson, O. D. W. Wang, J. E. Lofgreen,
C. Kube, T. Scherer, P. M. Brodersen, Z. Z. Gu and G.
A. Ozin, J. Am. Chem. Soc., 2011, 133, 5094.
36 B. P. Tripathi and V. K. Shahi, ACS Appl. Mater. Interfaces,
2009, 1, 1002.
37 S. Singh, A. Jasti, M. Kumar and V.K. Shahi, Polym. Chem.,
2010, 1, 1302.
38 F. A. Morrison Jr. and J. F. Osterle, J. Chem. Phys., 1965, 43,
2111.
39 R. K. Nagarale, G. S. Gohil, V. K. Shahi and R. Rangarajan,
Macromolecules, 2004, 37, 10023.
40 Y. Ji, Q. An, Q. Zhao, H. Chen and C. Gao, J. Membr. Sci.,
2011, 376, 254.
41 J. Schaep, B. Van der Bruggen, C. Vandecasteele and
D. Wilms, Sep. Purif. Technol., 1998, 14, 155.
42 S. Sarkar and A. K. Sengupta, Environ. Sci. Technol., 2010,
44, 1161.
43 Y. Baek, J. Yu, S.H. Kim, S. Lee and J. Yoon, J. Membr. Sci.,
2011, 382, 91.
21 Horii and K. Kannan, Arch. Environ. Contam. Toxicol., 2008,
55, 701.
22 N. Suzuki, S. Kiba, Y. Kamachi, N. Miyamotoc and
Y. Yamauchi, J. Mater. Chem., 2011, 21, 5338.
23 S. Egli, M. G. Nussbaumer, V. Balasubramanian, M. Chami,
N. Bruns, C. Palivan and W. Meier, J. Am. Chem. Soc., 2011,
133, 4476.
24 A. Kugel, B. Chisholm, S. Ebert, M. Jepperson, L. Jarabek
and S. Stafslien, Polym. Chem., 2010, 1, 442.
25 R. K. Nagarale, J.M. Lee and W. Shin, Electrochim. Acta,
2009, 54, 6508.
26 H. M. Colquhoun, D. Chappell, A. L. Lewis, D. F. Lewis, G.
T. Finlan and P. J. Williams, J. Mater. Chem., 2010, 20, 4629.
27 A. K. Singh, S. Prakash, V. Kulshrestha and V. K. Shahi, ACS
Appl. Mater. Interfaces, 2012, 4, 1683.
28 J. Gowda, A.M.A. Khader, B. Kalluraya, Padma Shree and
A.R. Shabaraya, Eur. J. Med. Chem., 2011, 46, 4100.
29 B. P. Tripathi, A. Saxena and V. K. Shahi, J. Membr. Sci.,
2008, 318, 288.
44 X. Z. Wei, L. P. Zhu, H. Y. Deng, Y. Y. Xu, B. K. Zhu and Z.
M. Huang, J. Membr. Sci., 2008, 323, 278.
45 C. Yan, S. H. Zhang, D. L. Yang and X. G. Jian, J. Appl.
Polym. Sci., 2008, 107, 1809.
30 B. Bolto, T. Tran, M. Hoang and Z. Xie, Prog. Polym. Sci.,
2009, 34, 969.
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 458–467 | 467