Journal of Membrane Science 521 (2017) 53–64

Contents lists available at ScienceDirect

Journal of Membrane Science

journal homepage: www.elsevier.com/locate/memsci

Understanding functionalized silica nanoparticles incorporation in thin

film composite membranes: Interactions and desalination performance
Masoumeh Zargar, Yusak Hartanto, Bo Jin n, Sheng Dai n
School of Chemical Engineering, The University of Adelaide, SA 5005, Australia

art ic l e i nf o a b s t r a c t

Article history: Incorporation of functional nanoparticles in polymer membranes is a promising approach for the de-
Received 29 March 2016 velopment of advanced thin film nanocomposite (TFN) membranes with improved desalination perfor-
Received in revised form mance. This study was aimed to fabricate TFN membranes incorporated with functionalized silica na-
26 August 2016
noparticles (SNs), which were synthesized with different sizes (
50 and
100 nm) and surface func-
Accepted 31 August 2016
Available online 31 August 2016
tionalities (hydroxyl, epoxy and amine). Research focused on getting a better understanding of the in-
terfacial interactions between SNs and polymer matrix and structure-property correlation of the TFN
Keywords: membranes. The physicochemical properties of the functionalized SNs and the TFN membranes were
Thin film nanocomposite membrane characterized using ATR-FTIR, TGA, DLS, SEM, TEM, XPS, tensiometer and streaming potential analysis.
Surface functionalization
The results confirm successful incorporation of the surface functionalized SNs into the polymer mem-
Water flux
branes. The concentration, size and surface functionality of SNs were found to have strong impact on the
Salt rejection
Structure-property correlation surface hydrophilicity, morphology and chemistry of the TFN membranes. The SNs hybridized TFN
members showed increased permeate flux and insignificant change in salt rejection. The surface func-
tionalization of SNs with amine and epoxy moieties could facilitate the chemical interaction between SNs
and polymer monomers, resulting in stabilizing the TFN membranes. Our research outcomes have
provided new insight into the structure-performance correlation of TFN membranes and can be bene-
ficial for fabrication of a wide range of nanoparticle incorporated membranes.
& 2016 Elsevier B.V. All rights reserved.

1. Introduction groups: zero-dimensional metal and metal oxides, one-dimen-

Polymeric thin film composite (TFC) membranes have received
increasing research interests and applications in separation and
purification processes such as desalination and wastewater treat-
ment. TFC membranes demonstrate higher separation perfor-
mance and stability over a wider range of temperature and pH
compared with the conventional cellulose acetate membranes [1].
Although many studies have been offered to the development of
TFC membranes during last decades, there is still an on-going need
to develop technically and economically more feasible TFC mem-
branes with a high separation performance and low energy costs
for their industrial applications [2,3].
Recent studies showed that the integration of nanoparticle (NP)
with the TFC membranes to fabricate thin film nanocomposite
(TFN) membranes can improve the membranes’ physicochemical
properties such as hydrophilicity, mechanical stability, and ther-
mal resistance, as well as their permselectivity [4–6]. The NPs used
as nanofillers in the TFN membranes can be classified into three
n
Corresponding authors.
E-mail addresses: bo.jin@adelaide.edu (B. Jin), s.dai@adelaide.edu.au (S. Dai).
sional nanotubes, and two-dimensional porous NPs [7]. The most
commonly used NPs in TFN membranes include zeolite [5,8,9],
silica [10–14], titanium dioxide [15,16], silver [17,18], graphene
oxide [19–21] and carbon nanotubes [6,22,23].
The TFN membranes can be fabricated by either direct de-
position of NPs on the TFC membranes or encapsulation of NPs
into the TFC membrane during the interfacial polymerization
stage. Surface deposition is a post modification approach and is
mainly used to improve the antifouling capability of the mem-
branes. However, the deposited NPs might leak during the high-
pressure filtration process. Hence, it is necessary to improve the
stability of NPs in the polymer matrix by chemical functionaliza-
tion strategies such as NP surface modification. Further, the gen-
eration of stress weak points due to the possible NP aggregation is
another issue associated with the fabrication of TFN membranes.
These may initiate the formation of defects during the high-
pressure filtration process. Incorporation of NPs into the TFC
membrane during the interfacial polymerization process can be a
promising approach. Jeong et al. fabricated the first TFN mem-
brane by in-situ incorporation of zeolite-A NPs (50–200 nm) into
the polyamide (PA) layer and their zeolite-incorporated TFN
membranes showed an enhanced separation performance and

http://dx.doi.org/10.1016/j.memsci.2016.08.069
0376-7388/& 2016 Elsevier B.V. All rights reserved.
54 M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64
fouling resistance [8]. The sustainable improvement in the struc-
ture and performance of TFN membranes is highly dependent on
the NP stability and distribution in the membrane matrix [4]. The
overall stability of these NPs incorporated into the TFC membrane
through interfacial polymerization is expected to be higher than
those being incorporated via deposition approach. That is due to a
large quantity of NPs which are partially or fully encapsulated
inside the PA thin film layers during the interfacial polymerization
process.
Silica nanoparticles (SNs) have superior biocompatibility and
mechanical stability for the fabrication of TFN membranes, and
their surface can be easily modified with functional chemical
groups [11,24]. Further, SNs show low toxicity as evidenced by
their wide biomedical applications [25–27]. Importantly, SN sur-
face hydrophilic groups can be beneficial for enhancing perme-
ability and fouling resistance of the TFN membranes [4,10]. The
prospect of enhancing surface hydrophilicity and consequently
fouling resistance of the TFC membranes has been well docu-
mented in previous studies [12,28]. For instance, Tiraferri et al.
studied the incorporation of SNs (
14 nm) on the TFC membranes
via the deposition approach, resulting in increasing the fouling
resistance of their TFN membranes [12].
Until now, there are very few reported studies on the in-situ
incorporation of NPs and their interactions within the polymer
matrix of the TFN membranes. Further, the extensive evaluation of
the physicochemical properties of TFN membranes following the
hybridization of functionalized NPs and their performance corre-
lations are missing in the literature. This study aims to address
these fundamental gaps by developing novel TFN membranes in-
corporated with functionalised SNs, which are noted as “SN in-
corporated TFC” (SN-TFC) membranes in this study. Our research
will focus on understanding the structure-property correlation of
the TFN membranes. In particular, we will synthesize SNs with
different sizes (
50 and
100 nm) and functionalize their sur-
faces with the amine or epoxy groups. These functionalities make
the SNs ready to interact with the PA monomers during the in-
terfacial polymerization and can improve the stability of the TFN
membranes. The surface functionalization of the SNs and the
physicochemical properties of the TFN membranes are extensively
characterized. The desalination performance of the SN-TFC mem-
branes is evaluated using a cross-flow filtration unit. The research
outcomes will provide new insights into the influence of different
surface functionality of nanofillers on the overall performance and
physicochemical properties of the resulting TFN membranes.
2. Experimental
2.1. Materials
Polyether sulfone support (PES-MQ-50 kDa) was obtained from
Synder filtration (USA). 1, 3-phenylenediamine (MPD, 499%), 1, 3,
5-benzenetricarbonyl trichloride (TMC, 98%), n-hexane, tetraethyl
orthosilicate (TEOS, 499%), L-arginine, (3-aminopropyl) tri-
methoxysilane (APTMS, 97%) and (3-glycidyloxypropyl) tri-
methoxysilane (GPMS, 498%) were purchased from Sigma-Al-
drich. Sodium chloride (NaCl), aqueous ammonia solution (28%)
and cyclohexane were obtained from VWR International.
2.2. Synthesis and surface modification of silica nanoparticles
2.2.1. Synthesis of silica nanoparticles
SNs with an average size of
100 nm were fabricated accord-
ing to a modified Stöber approach [29]. Briefly, 50 ml absolute
ethanol, 0.9 ml Milli-Q water and 2.04 ml of 28% aqueous ammo-
nia were mixed together and stirred for 10 min at room
temperature. 2.9 ml of TEOS was rapidly injected into the mixture
and the solution was stirred at 25 °C for 24 h. The resulting SNs
were washed 3–4 times with ethanol and then redispersed and
stored in ethanol.
SNs with the average size of
50 nm were synthesized ac-
cording to a modified approach reported by Fouilloux et al. [30].
30 ml of 6 mM L-arginine solution was prepared. 4.65 ml cyclo-
hexane was then added as the solvent for the organic phase. The
mixture was transferred to a pre-heated 60 °C oil bath to warm up,
followed by the dropwise injection of 5.7 ml TEOS to the organic
phase. The stirring rate of the solution was fixed at a low value
such that the aqueous phase was well mixed while the organic
phase was left intact. The reaction was further continued for 72 h
under constant stirring and heating. Finally, bluish silica solution
was washed several times with Milli-Q water and stored in etha-
nol. The obtained NPs possess hydroxyl surface functional groups
and are noted as SN-OH. The two SN-OH NPs with sizes of
50
and
100 nm are named as SN50 and SN100 in this study. Further
chemical modifications were then performed to introduce amine
or epoxy groups to the SNs.
2.2.2. Preparation of amine and epoxy-functionalized silica
nanoparticles
A modified approach from Chang et al. was applied to func-
tionalize the SN surface with amine groups [31]. Typically, 100 mg
of the synthesized SNs were dispersed in 50 ml of ethanol and
degassed with nitrogen for 1 h. This was followed by the increase
of the temperature to 80 °C and the quick injection of 135 mL
APTMS to the reaction vessel. The solution was then refluxed for
6 h. After centrifugation and washing with ethanol and water,
amine-functionalized SNs (SN-NH2) were obtained and finally re-
dispersed in ethanol.
The epoxy group was introduced to the SN surface according to
an established approach [32]. In brief, 100 mg of synthesized SNs
was added into 2.5 ml of 5 wt% GPMS in anhydrous toluene and
then stirred for 12 h at room temperature. The epoxy functiona-
lized SNs (SN-EPX) were centrifuged and washed with anhydrous
toluene and absolute ethanol, respectively, followed by drying
under vacuum. The general schematics for the preparation of SN-
NH2 and SN-EPX are illustrated in Supporting information (Fig.
S1).
2.3. Fabrication of TFC and SN-TFC membranes
To obtain the TFC membranes, an ultra-thin PA layer was casted
on ultrafiltration PES support membranes using the interfacial
polymerization. The PES support was first clamped between an
acrylic plastic plate and an acrylic plastic frame. Then, 2% (w/v)
MPD aqueous solution was poured into the frame and allowed to
penetrate to the PES support for 10 min. Excess MPD solution was
drained and the PES support was rolled with a soft rubber roller
until no visible droplets remained on the surface of the membrane.
This was followed by reassembling the holder and pouring
0.1 (w/v)% TMC in hexane solution into the frame. After 1 min
reaction, excess TMC solution was disposed and the membrane
was air-dried for another 1 min, followed by 10 min curing in a
recirculating oven at 60 °C. The fabricated TFC membranes
were washed with 22 °C distilled water and stored wet at 5 °C for
further characterization and evaluation. The TFC membrane
without any nanoparticle was used as a control membrane in this
study.
SN-TFC membranes were fabricated exactly as described, ex-
cept that different amounts of functionalized SNs (SN-OH, SN-NH2,
and SN-EPX) were dispersed in MPD aqueous solution. To obtain
good SN dispersion, the mixtures of SNs and MPD solution were
sonicated for 1 h at 22 °C exactly before the interfacial
 M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64
polymerization. The resulting membranes incorporated with SN-
OH, SN-NH2 and SN-EPX NPs are noted as SN-OH-TFC, SN-NH2-TFC
and SN-EPX-TFC membranes, respectively.
2.4. Characterization of nanoparticles, TFC and SN-TFC membranes
Attenuated total reflectance Fourier transform infrared spec-
troscopy (ATR-FTIR) was performed using a NICOLET 6700 spec-
trometer equipped with a diamond ATR to evaluate the chemical
structures and functionalities of the fabricated nanoparticles and
membranes. The obtained FTIR spectra are shifted vertically for
clarity.
The surface charges of the synthesized nanoparticles were
measured using a Malvern Zetasizer Nano ZS (Malvern Inst. Ltd., U.
K.) based on the dynamic light scattering (DLS). The samples were
dispersed in milli-Q water and the measurements were performed
at room temperature.
Thermogravimetric analysis (TGA) was performed on a TGA/
DSC 2 STARe system to estimate the surface functionalities, where
5 mg dried nanoparticles were loaded and measurement tem-
perature was from 30 to 800 °C with a heating rate of 10 °C/min
under air. An isotherm step at 120 °C was set for 1 h to remove the
trace amount of adsorbed water. The residual weight of the sam-
ples at the end of this isotherm was considered as the initial
weight and used to calculate the normalized residual weights of
the following stages.
The morphology of fabricated nanoparticles and membranes
were characterized using a scanning electron microscope (FEI
Quanta 450 FEG Environmental SEM (ESEM)) equipped with EDX
(energy dispersive x-ray spectroscopy). The instrument was op-
erated at the voltage of 5–20 kV and the working distance of
10 mm. For the cross-section SEM imaging of the membrane
samples, small pieces of fabric free and wet membranes were
freeze-fractured in liquid nitrogen and mounted vertically on SEM
stands. The prepared samples were dried overnight and sputter
coated with
5 nm layer of platinum using a CRESSINGTON 208
high-resolution sputter coater before imaging.
TEM images were taken on an FEI Tecnai G2 Spirit Transmission
Electron Microscope at an acceleration voltage of 100 kV.
Nanoparticle samples were prepared by placing one drop of
ethanol dispersed nanoparticles on 200 mesh carbon-coated
copper grids. Membrane samples for cross-section TEM imaging
were prepared by gently removing the polyester backing layer to
ensure PA and PSF layers remained together. Small pieces
of fabric free membranes were cut and embedded in Epon resin
and cured in the oven at 50 °C overnight. From the cured samples,
60–70 nm cross sections were cut on a Leica EM UC6 Ultra-
microtome and mounted on formvar and carbon coated copper
grids for imaging.
An Anton Paar SurPASS Electrokinetic Analyser (Anton Paar
GmbH) was used to evaluate the streaming potentials of the fab-
ricated membranes. The wet membranes samples were cut in
rectangular profiles and mounted in the adjustable gap cell ap-
paratus by applying double- sided tape. The measurements were
carried out using 10 mM KCl solution as the electrolyte and the pH
was varied by the addition of 0.1 M HCl or 0.1 M NaOH solutions to
400 ml KCl solution. Multiple rinsing and flow check steps were
carried out prior to the tests to ensure reproducibility. For each
sample, four repeats were recorded and the averages in each pH
were reported. Zeta potentials were obtained from the slope of
streaming potential versus pressure according to the Helmholtz-
Smoluchowski approach.
X-ray photoelectron spectroscopy (XPS) was performed on
the samples using a Kratos Axis Ultra with a monochromatic
aluminium x-ray running at 225 W and the characteristic
55
energy of 1486.6 eV. The area of analysis (Iris aperture) was
0.3 mm  0.7 mm and the electron take off angle was normal to
the sample surface with the analysis depth of approximately
15 nm from the sample surface. The spectra were scanned in the
range of  10 to 1110 eV and the pass energy and step of 160 eV
and 500 meV were used. The charge neutraliser system was used
to reduce the surface charge of the samples. CasaXSP software
package was eventually used to interpret the obtained survey
spectra.
Contact angle measurements were performed on an Attention
Theta Optical tensiometer in sessile drop mode at room tem-
perature. The membranes were attached to clean glass slides using
a double-sided tape and a milli-Q water droplet (1 mL) was auto-
matically dispensed on the surface of the membranes. A FireWire
camera (55 mm focus length) recorded the droplet surface contact
angle equilibrated for 10 s from the time it was dispensed. Curve
fitting and data analysis were then performed using the OneAt-
tension software. The reported contact angles were the average of
at least 8 different positions on each membrane.
2.5. Evaluation of membrane desalination performance
A reverse osmosis cross-flow filtration setup was used to
evaluate the desalination performance of the TFC and SN-TFC
membranes. The effective membrane surface area of the filtration
cell (CF042, Sterlitech) was 42 cm2. Prior to evaluation, the
membrane coupons were compacted with distilled water in the
filtration unit for at least 16 h under the pressure of 15 bar to reach
their equilibrium flux. The system was then operated with 15 L of
2000 mg/L NaCl at a flowrate of 4 L/min under 15 bar and
257 0.1 °C. The concentrate was recycled back to the feed tank
while operating the system, and the permeate was collected in a
container positioned on a digital balance for flux measurement.
The schematic illustration of this lab-scale filtration setup is pre-
sented in Supporting Information (Fig. S2).
The collected permeate weight was recorded using the LabX
Direct Software (Mettler Toledo) and the volumetric permeate flux
(Jw) was calculated using Eq. (1) accordingly:
VP
Jw =
A⋅t (1)
where Jw is the permeate water flux (LMH), VP is the collected
permeate volume (l), A is the active surface area of the membrane
(m2) and t is the operation time (h). Intrinsic water permeability
coefficients of the fabricated membranes can be then calculated
from the obtained Jw values using Eq. (2):
Jw
Pw =
(∆P−∆π ) (2)
2 1 1
where Pw (L m h bar ) is the specific membrane perme-
ability coefficient, ΔP (bar) is the applied transmembrane pressure
and Δπ (bar) is the transmembrane osmotic pressure [9].
In order to evaluate the salt rejection, the conductivity of the
permeate and feed water were measured using a conductometer
(k ¼1.0, AQUA, Cond./pH, TPS, Australia) and were subsequently
converted to the equivalent salt concentrations using a pre-cali-
bration curve which is shown in Fig. S3. Salt rejections (R) were
then calculated using Eq. (3):
⎛ C ⎞
R%=⎜ 1− P ⎟× 100
⎝ CF ⎠ (3)
where R is the salt rejection and CP and CF (g/L) are the equivalent
NaCl concentrations in the collected permeate and feed,
respectively.
56 M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64

The solute diffusive permeation coefficients, B (L m  2 h  1), 3. Results and discussion

were then calculated using Eq. (4):
3.1. Characterization of fabricated silica nanoparticles

B = Jw
( 1−R)
(4)
R
The desalination tests for each TFC and SN-TFC membrane were
performed in three replicates and the average values are reported
for further evaluation and comparison. A homoscedastic (assum-
ing equal variances) two-tailed distribution using student t-test
analysis was finally used to study the statistically significant dif-
ferences (α ¼0.05) between the performance results.
SEM micrographs and TEM images (inset) of the synthesized
SNs in Fig. 1 show that the nanoparticles have uniform spherical
structures with the average diameters of
50 and
100 nm, re-
spectively. SEM and TEM images of the surface functionalized SNs
are presented in Fig. 1. The results confirm that there is no sig-
nificant SN aggregation following their surface functionalization.
The thermogravimetric analysis was performed to study the
surface functionalization of SNs and to estimate the quantity of
their surface functional groups. Physically adsorbed water was

Fig. 1. SEM micrographs and TEM images (inset) of the different sized SN, A) SN50, B) SN100, C) SN50-NH2, D) SN100-NH2, E) SN50-EPX and F) SN100-EPX. Particle size and
morphology are similar as the SN before modification.

removed by heating the dried SNs at 120 °C for 1 h. The tem-
perature was then raised step-wise to 800 °C. The TGA thermo-
grams in Fig. 2A show that SN50 was reduced about 2.5% residual
weight which can be associated with the loss of the chemically
bonded water and organics from the sol-gel fabrication process or
the dihydroxylation of the SNs’ surface hydroxyl groups [33,34].
The residual weight loss is higher for the surface functionalized
SNs due to the decomposition of grafted silane moieties and in-
dicates the successful surface functionalization with amine or
epoxy [34]. The estimated surface functional groups on the SN50-
NH2 and SN50-EPX are determined as
6% and
9%, respectively.
The successful surface modification by amine or epoxy groups
can be further confirmed by zeta potential measurement. Fig. S4
shows the surface zeta potential of different SNs. The amine in-
troduction on the SN-NH2 surface results in changing the zeta
potentials from negative to positive values. On the other hand, the
SN-EPX shows a lower magnitude of surface potential due to the
loss of some surface hydroxyl groups. It should be noted that the
zeta potential values (4 15 mV) indicate that the fabricated SNs
are highly stable in the aqueous medium.
Table 1 summarizes the XPS analysis results for the SN50 with
different surface functionalities and their survey spectra are pre-
sented in Fig. S5. SN50 shows the typical silica peaks at 532.72 eV,
284.72 eV, 153.3 eV and 103.22 eV corresponding to O 1s, C 1s, and
Si 2s, and Si 2p, respectively [35]. For SN50-NH2, N 1s peak at
399.62 eV is emerged with 1.92% composition which confirms the
successful amine functionalization. On the other hand, SN-EPX
shows 30.60% increase in the composition of C 1s which indicates
the successful epoxy functionalization.
3.2. Characterization of TFC and SN-TFC membranes
Fig. 2B shows ATR-FTIR spectra of PES support, TFC and SN-TFC
membranes. Two new peaks at 1541 and 1612 cm  1 are found in
the ATR-FTIR spectra of the TFC and SN-TFC membranes compared
to that of the PES support which is associated with the vibrations
of amide groups: C–N (amide II band) and N–H (aromatic amide)
Table 1
XPS results for SN50 with different surface functionalities.
Sample name Atomic percentage of elements
N (%) O (%)
SN-OH – 59.53
SN-NH2 1.92 56.14
SN-EPX – 54.18
Normalized residual weight loss
1
0.98
0.96
0.94
0.92
0.9
0 100 200 300
Temperature (ºC)
Fig. 2. A) TGA thermograms for SN50 before and after surface functionalization, B) FTIR spectra for PES support, TFC membrane and a typical SN-TFC membrane.
M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64 57
confirming PA fabrication [36]. Since the IR penetration depth
(
0.5  1 mm) exceeds the thickness of the thin PA layer [36,37],
PA, PES and SN peaks are all emerged in these FTIR spectra. The
vibration bands of 955 cm  1 (stretching vibration of Si–OH) and
1090 cm  1 (asymmetric vibration of Si–O–Si) represent the SNs
and the vibration bands of the SN surface functional groups
overlap with the PES and PA bands. To further identify the over-
lapped peaks, the SN, PA and PA-SN thin films extracted from their
PES supports were characterized. The results are reported in Fig.
S6.
Fig. 3 presents the high and low magnification surface and
cross-section SEM micrographs of the synthesized TFC and SN-TFC
membranes incorporated with the 0.05 wt% SNs. All membranes
show a leaf-like structure, which is the expected surface mor-
phology of thin PA layers made from MPD and TMC. The SN50-TFC
membrane represents leafier surface morphology compared to the
SN100-TFC membrane which has some lobe-like surface features
(Fig. 3B and E). These surface morphology and features are re-
sulted from the different sizes of SNs which may influence the
formation of the surface strands. As observed in the high magni-
fication surface images (Fig. 3C, D, F, and G), the membrane surface
features are denser and the polymer strands are more packed for
the SN-NH2-TFC and SN-EPX-TFC membranes compared to the
control TFC and SN-OH-TFC membranes. These compact surface
features are due to the chemical interaction of silane groups with
the PA monomers which may interfere with the interfacial poly-
merization. However, high magnification SEM images of the SN-
TFC membranes indicates that all membranes still represent leaf-
like morphology with some scattered nodule-like features. The
presence of some SNs with the sizes of
50 and
100 nm on the
thin film layer of SN-TFC membranes confirms the successful in-
corporation of the functionalized SNs into the PA layers. Cross-
section SEM images of the TFC and SN-TFC membranes (Fig. 3)
show the structure of the different membranes with maximum

450 nm PA layers. Furthermore, no significant aggregation of
SNs can be observed, indicating the fair distribution of different
SNs into the SN-TFC membranes.
Fig. S7 presents SEM images of the SN50-TFC membranes in-
corporated with SN50 at different concentration. To further elu-
cidate the distribution of nanoparticles, SEM images were cap-
tured from some fractured areas on the surface of SN50-NH2-TFC
and SN100-EPX-TFC membranes as illustrated in Fig. S8. The re-
C (%) Si (%) sults confirm the successful SN incorporation and fair distribution
in these membranes.
21.17 19.30
To extensively analyse the SN distribution in the PA thin film
21.68 20.26
27.65 18.17 layer, cross-section TEM imaging was performed on TFC/SN-TFC
membranes incorporated with functionalized SN50. Three cross-
PES support
TFC membrane
Typical SN-TFC membrane
SN
Transmittance
SN-NH2
1612 1541
SN-EPX
400 500 600 700 800 900 1500 1000 500
Wavenumbers (cm - ¹)
58 M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64

3 µm 500 nm 50 µm 5 µm
)

3 µm 500 nm 50 µm 5 µm

3 µm 500 nm 50 µm 5 µm

3 µm 500 nm 50 µm 5 µm

3 µm 500 nm 50 µm 5 µm

3 µm 500 nm 50 µm 5 µm

3 µm 500 nm 50 µm 5 µm

Fig. 3. SEM micrographs of the surfaces and cross-section of different TFC/SN-TFC membranes: A) TFC, B) SN50-TFC, C) SN50-NH2-TFC, D) SN50-EPX-TFC, E) SN100-TFC, F)
SN100-NH2-TFC, G) SN100-EPX-TFC. For the surface images, the low magnification images were taken at 55000  and the high magnification images were taken at
200000  and for the cross-section images, the low magnification images were at 2000  and the high magnification ones were at 25000  .
 M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64
section samples were prepared from each membrane and their
typical images are shown in Fig. 4. TEM images further confirm the
successful SN distribution into the PA thin film layers. The quantity
of SNs seems to be higher for the SN50-NH2-TFC and SN50-EPX-
TFC membranes compared to the SN50-TFC membranes which can
be attributed to the chemical interaction of SN-NH2 and SN-EPX
with the PA monomers as well as the zeta potentials. These
functionalized SNs could be incoportated to the TFC matrix due to
the faverable interactions, including the electrostatic attraction
between SN-NH2 and negatively charged PES support [38], and the
weaker electrostatic repulsion between SN-EPX and PES support
than those between the SN-OH and PES support. Consistent with
previous studies [8,9,39], thickness of the PA thin film layers varies
in a range of 50  450 nm. The average thicknesses based on TEM
are 255 7178, 154 752, 271 7125, and 102 747 nm for the TFC,
SN50-TFC, SN50-NH2-TFC, and SN50-EPX-TFC membranes, re-
spectively. The higher thickness of the SN50-NH2-TFC membrane
than that of the SN50-EPX-TFC membrane is attributed to their
different amine/acyl chloride molar ratios due to the interaction of
SN-NH2 or SN-EPX with PA monomers. The amine/acyl chloride
ratio is identical to the feed's MPD/TMC ratio for the SN-OH-TFC.
However, this ratio slightly increases for the SN-NH2-TFC due to
the introduction of more –NH2 groups from the NP surface. Con-
versely, SN-EPX can consume some –NH2 groups of MPD mono-
mer, resulting in slightly decreasing the ratio. The interfacial
polymerization of MPD and TMC consists of an extremely fast
stage which forms an ultrathin dense core layer of PA in the oil-
water interface and a slow growth stage, making a thicker layer of
PA by the monomers’ diffusion (mainly MPD) [40]. The decrease of
Fig. 4. Cross-section TEM images of the A) TFC, B) SN50-TFC, C) SN50-NH2-TFC and SN50-EPX-TFC membranes; the scale bars are 100 nm.
59
the MPD/TMC molar ratio is beneficial for forming a denser core
PA layer, while reducing the thickness of the PA layer during the
diffusion stage and vice versa [40–42]. Therefore, the SN-NH2-TFC
would have a less dense core layer but thicker secondary PA layer
formed by the diffusion, while the SN-EPX-TFC would have a
denser core layer but thinner secondary PA layer.
In order to understand the surface electrical double layers of
the fabricated membranes, streaming potential measurements
were performed on the TFC and SN-TFC membranes incorporated
with functionalized SN50. Surface zeta potentials of the mem-
branes can be highly influenced by the unreacted surface func-
tional groups or impurities [43]. Therefore, the surface function-
ality of the introduced SNs can alter the surface potential of the
SN-TFC membranes. Fig. 5 shows the average of four pH depen-
dent zeta potentials for various membranes measured using tan-
gential streaming potential (TSP) method with the Fairbrother and
Mastin approximation. In general, the surfaces of TFC and SN-TFC
membranes are more positively charged at low pH due to the
proton ions’ adsorption and are more negatively charged at high
pH due to the deprotonation of the PA monomers [12,20].
Fig. 5 shows that the isoelectric point (IEP) of the TFC mem-
brane occurs at pH
3.6 which is consistent with the previous
studies [43,44]. However, the zeta potential and IEP of the mod-
ified membranes are slightly changed which are influenced by the
SNs’ surface functional groups. In the case of the SN50-NH2-TFC
membrane, the extra amine groups raise the amine/acyl chloride
molar ratio and make the membrane surface more positively
charged compared to the other membranes increasing the IEP to
pH
4.5. However, the surface potentials and IEPs of the SN50-TFC
60 M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64

20
Table 2
15 XPS results for TFC membrane and TFN membranes incorporated with different
Plain TFC
surface functionalized SN50.
Zeta potential (mV)

10 SN50-TFC
SN50-NH2-TFC Sample name Atomic percentage of elements
5 SN50-EPX-TFC
*
0 N (%) O (%) C (%) Si (%) N/O

-5 TFC 9.89 16.12 73.71 – 0.61

-10 SN-TFC 8.97 16.51 73.92 0.58 0.60
SN-NH2-TFC 9.58 16.85 72.59 0.96 0.67
-15 SN-EPX-TFC 9.19 17.09 72.44 1.26 0.69

-20
2 3 4 5 6 7 8 9 10
pH 80

Contact Angle (Degree)

Fig. 5. Comparison of the streaming potential measurements for the TFC mem- 70
branes with and without SN50 having different surface functionalities. 60
50
and SN50-EPX-TFC membranes become more negative due to the
hydroxyl and/or epoxy functional groups on SNs. It should be 40

noted that all zeta potential curves show amphoteric membrane 30

surfaces or surfaces with both acidic and basic functional groups
[43,45] and their variation can be a strong indicator of the suc-
cessful SN incorporation into the TFC membranes. Further, all
curves appear to be stable at pH over 7, indicating the complete
deprotonation of the carboxylic groups [20].
XPS analysis was performed to estimate the cross-linking
density of the PA layers by comparing their elemental composi-
tions. Normally, a higher N/O ratio indicates a higher cross-linking
density of the PA layers [46–48]. The SNs incorporation introduces
more C, O or N (for SN-NH2 only) in SN-TFC matrix. Thus, the N/O
ratio in SN-TFC membrane needs to be determined for the cross-
linking density evaluation [9,49]. XPS analysis results of SNs in
Table 1 can be used to correct the elemental composition of each
element and calculate the real N/O ratio in the SN-TFC membranes,
which is noted as *N/O. This can be performed based on Si com-
position of each particular TFN membrane. Table 2 shows the in-
itial values of each constituent element and their corresponding
* hydration layer acts as a repulsive barrier against various fouling
N/O ratios.
Results in Table 2 explain that SN50-TFC membrane has a lower
* proteins, etc. [12,28,55].
N/O ratio compared to the control TFC membrane which suggests
a slight decrease of the PA layer's cross-linking density following
the SN50 incorporation, indicating that the unmodified SNs can
alter the bulk PA structure. This finding agrees with the results
reported in previous studies [8,9,48]. The increase in *N/O ratio
after the introduction of SN50-NH2 and SN50-EPX into the PA
layers leads to high cross-linking density. It is due to the chemical
interactions of the functional SNs with the PA monomers and can
be beneficial for enhancing the SNs’ stability in the PA matrix. That
can also explain why more SNs could be detected in the TEM
samples of the SN50-NH2-TFC and SN50-EPX-TFC membranes
compared to the SN50-TFC membrane. These observations are
supported by the FTIR analysis results reported in Fig. S6.
Contact angle is a macroscopic manifestation of the molecular
level surface-water interactions and represents the magnitude of
the affinity towards water or the level of hydrophilicity of the
membranes [8,50–52]. The contact angle was calculated from the
average of the tangent lines to both sides of the dispensed water
droplets on the membrane surface. Fig. 6 shows the average con-
tact angle values of at least 8 random points on the surface of each
membrane. The contact angles decrease with increasing the con-
centration of SNs in the PA thin film layers which is consistent
with the results of previous studies. [35,51–54]. Further, the
equilibrium contact angle decreases for the amine and epoxy
functionalized SN-TFC membranes. Similar contact angle profiles
are observed from all SN-TFC membranes. It is interesting to note
that high hydrophilicity of TFC and TFN membranes corresponds
20
10
0
Fig. 6. Contact angle measurement results for the TFC membranes incorporated
with the different concentrations of SN50 and 0.05 wt% SN-NH2 and SN-EPX.
to a high water flux during the desalination process. Therefore, the
lower contact angle of all SN-TFC membranes compared with the
TFC membrane appears to be beneficial for enhancing their de-
salination performance.
High hydrophilicity of membranes is deemed advantageous for
fouling mitigation, where anti-fouling characteristic comes from
the formation of hydration layers on hydrophilic surfaces. This
matters, such as microorganisms, organic and inorganic materials,
3.3. Desalination performance of TFC and SN-TFC membranes
The desalination performance of TFC and SN-TFC membranes in
terms of water flux and salt rejection were systematically in-
vestigated and calculated using Eqs. (1) and (3). To evaluate the
desalination behavior in a more accurate and independent man-
ner, the intrinsic water and salt permeability coefficients, which
are independent of the operating conditions, were then calculated
from Eqs. (2) and (4).
3.3.1. Effect of nanoparticle concentration
Fig. 7A shows the desalination results of the SN-TFC mem-
branes incorporated with the different concentrations of SN50.
Previous studies on desalination performance of the TFC mem-
branes reported that the permeability and the salt rejection of the
membranes are significantly influenced by the structure and
properties of their supporting layers and fabrication procedure
(e.g. manual/automated coating) [10,37,56,57]. The TFC mem-
branes are usually made by polysulfone ultrafiltration support. In
our study, we used PES support with different structure as shown
in Fig. S9, as well as the manual coating procedure. Thus, our de-
salination data cannot be comparable with those membranes that
are manufactured by polysulfone supports or are automatically
coated. Our TFC membranes perform very much similar salt re-
jections as those reported studies [10,58–61]. This study is to
 M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64
24
Water Flux (LMH) Salt Rejection (%)
22
Water Flux (L/m²h)
20
18
16
14
12
10
TFC membrane 0.05% SN50 0.1% SN50
TFC/TFN Membrane
TFC membrane 0.05% SN50-TFC
26
0.1% SN50-TFC 0.20% SN50-TFC
24
Water Flux (L/m²h)
22
20
18
16
14
12
10
0 100 200 300 400 500 600 700 800 900 1000
Compaction Time (min)
Fig. 7. A) the performance evaluation of the various TFC membranes incorporated
with the different concentrations of SN50, and B) water flux vs. compaction time of
the TFC and SN-TFC membranes incorporated with the different concentrations of
SN50.
evaluate the structure-performance correlation of the TFC mem-
branes incorporated with the surface functionalized SNs. The re-
lative desalination data obtained at the same fabrication and
evaluation procedure are valid rather than the absolute values.
The specific water and salt permeability coefficients of the TFC
membranes incorporated with SN50 at different concentrations
are presented in Table 3. Fig. 7A shows the equivalent water flux
and salt rejection of these membranes. All experiments were
performed in three replicates, and the average values are statisti-
cally assessed. The incorporation of 0.1 wt% SN50 increases the
water permeability by
40% over the TFC membrane. Student t-
test analysis shows statistically significant correlations of water
permeability coefficients of various SN50-TFC membranes and the
TFC membrane with the p-values of 0.0004, 0.0003, and 0.0037 for
the 0.05 wt%, 0.1 wt%, and 0.2 wt% SN50 incorporated TFC mem-
branes, respectively. As the incorporated SN concentration in-
creases from 0.05 to 0.1 wt%, similar high water permeability va-
lues can be observed. The introduction of SNs into the PA layer
may change bulk PA structure and result in the increase of free
volume which contributes to higher permeation rates [9,13,51]. On
the other hand, the hydrophilicity of the TFC membranes increases
after SN incorporation, resulting in further accelerating water so-
lubilization and diffusion [51,62–64].
Although the hydrophilic SNs can enhance the water perme-
ability, there is an optimum concentration to provoke positive
changes in membrane performance. Beyond this optimum level,
the permeability may drop due to the pore blocking effect and/or
the aggregation of excess nanoparticles [14,62,63,65,66]. The low
water flux of the 0.2 wt% SN50-TFC membrane can be attributed to
this. Similar trend has been reported in the previous studies on the
61
Table 3
The average solution diffusion permeability coefficients of the water and salt across
100
the fabricated TFC/SN-TFC membranes. All membranes were incorporated with
98 0.05 wt% SNs unless specified otherwise.
96
Salt Rejection (%)
Sample name Pw (L m  2 h  1 bar) B (L m  2 h  1)
94
92 TFC 1.116 70.045 2.0157 0.551
90 SN50-TFC (0.05 wt%) 1.5047 0.039 2.890 7 0.817
SN50-TFC (0.1 wt%) 1.5167 0.037 2.898 7 0.779
88
SN50-TFC (0.2 wt%) 1.3517 0.049 2.375 70.754
86 SN50-NH2-TFC 1.1537 0.037 0.984 7 0.276
84 SN50-EPX-TFC 1.239 70.091 2.155 70.378
SN100-TFC 1.3687 0.039 2.537 70.378
82
SN100-NH2-TFC 1.1177 0.040 1.2737 0.222
80 SN100-EPX-TFC 1.189 70.049 1.639 70.367
0.2% SN50
incorporation of hydrophilic nanofillers into the TFC membranes
[14,62,63,65,66].
SN-TFC membranes perform a slightly lower salt rejection than
TFC membrane, while the rejection values are found to be im-
proved following the concentration increase. However, this change
appeases insignificant. This observation is in line with those data
reported in the literature, and indicates that no major defects are
formed in the PA structure of the SN-TFC membranes following the
SN concentration up to 0.2 wt% [5,10,65]. Considering the statis-
tically insignificant correlation (p¼ 0.1824–0.5345) of the B values
for TFC membranes incorporated with the 0.05–0.2 wt% SN50, it
can be concluded that the incorporation of SNs at low con-
centrations could significantly improve the water permeability of
the membranes without sacrificing their salt rejections.
EDX mapping of the SN-TFC membranes can provide further
information showing the NP distribution. Fig. S10 presents Si
mappings of the SN-TFC membranes incorporated with different
SN50 concentrations. All SN-TFC membrances show homogeneous
SN distributions across the membranes, while slight aggregation
was found on the SN-TFC membrane with 0.2 wt% SNs. These
observations justifie the desalination results as shown in Fig. 7A,
and agree with those reported data [13,14,62].
0.05 wt% SN50-TFC membrane performs a slightly low (only
0.7% lower), but insignificant (p ¼0.7472) reduction in water flux
and an insignificant (p ¼0.9932) change of salt rejection compared
to the 0.1 wt% SN50-TFC membrane. Considering the statistical
analysis results, 0.05 wt% was determined as the optimum SN
concentration, and thus 0.05 wt% SN50-TFC membrane was cho-
sen for further investigations.
Compaction resistance is an important parameter for evalua-
tion of the TFC membranes for industrial desalination application.
Incorporation of inorganic nanofillers such as SNs into the TFC
membranes is expected to improve the mechanical strength and
eventually compaction resistance of the resulting TFN membranes.
Both ultrafiltration support and PA layer of TFC membranes ex-
perience irreversible compaction under high pressure which re-
sults in the permanent decrease of the water flux [67–69]. Fig. 7B
shows the compaction rates of the SN-TFC membranes in-
corporated with SN50 at different concentrations. The results were
obtained by measuring the water flux as a function of time during
the compaction stage. The compaction rates are 22.8%, 17.6%, 14.2%
and 13.3% for TFC, 0.05 wt%, 0.1 wt% and 0.2 wt% SN50-TFC
membranes, respectively. Student t-test analysis results show
statistically significant improvement of the compaction rates fol-
lowing SN50 concentration increment up to 0.1 wt%. The correla-
tion between the compaction rates and SN50 concentrations of
0.1–0.2 wt% is insignificant (p 40.10). Compaction data for the
functionalized SN hybridized TFC do not show significant corre-
lations with these compaction results obtained from other TFC
membranes. Prendergast et al. observed similar results with im-
proved compaction resistance for their zeolite NPs incorporated
62
TFC membrane and reported the importance of the NP-enhanced
mechanical properties to improve the overall TFC membranes’
performance [69]. Compaction results suggest the NP incorpora-
tion could be beneficial for strengthening TFN membranes which
is another advantage of their application.
3.3.2. Effect of nanoparticle surface functionality
MPD monomer has –NH2 groups while TMC monomer has –
COCl groups. Interfacial interaction of these functional groups in
the interface of MPD aqueous and TMC organic solutions (on the
PES support) results in the formation of TFC membranes with a
thin PA selective layer. To fabricate SN-TFC membranes, SNs can be
either dispersed in the aqueous or organic monomer solutions. If
the SNs can be conjugated with the PA by the chemical interaction
with either of these monomers, the resulting SN-TFC membranes
are expected to have more stable structure and excellent durability
compared to the membranes made through only physical inter-
actions with nanoparticles. Hence, we designed SNs with two
different surface functional groups: epoxy and amine. It is ex-
pected that epoxy groups of SN-EPX can interact with the –NH2
groups of MPD monomer while amine groups of SN-NH2 can in-
teract with the -COCl groups of TMC monomer. After the in-
corporation of the functionalized SNs to the TFC membranes, their
desalination performance was evaluated subject to the SN sizes
and surface functionalities.
Fig. 8A and Table 3 show that all SN-TFC membranes perform a
higher water flux/permeability compared to the control TFC
membrane. The highest water flux is given by the SN-OH-TFC
membranes followed by the SN-EPX-TFC and SN-NH2-TFC mem-
branes, respectively. Systematic data for desalination were ana-
lysed by student t-test comparing the intrinsic water permeability
coefficients (Pw) of the different SN-TFC membranes with the
control TFC membrane. A higher average permeability coefficient
is given by all SN-TFC membranes than the control TFC membrane,
Water Flux (LMH)
25
Water Flux (L/m²h)
20
15
10
5 could explain why both SN-NH2-TFC and SN-EPX-TFC membranes
0 80
TFC SN50
1
Normalized Salt Rejection
0.995
0.99
0.985
SN50-TFC
SN50-NH2-TFC
0.98 SN50-EPX-TFC
SN100-TFC
0.975
SN100-NH2-TFC
0.97 SN100-EPX-TFC
0.965
0.96
0 1
Fig. 8. A) The desalination performance evaluation of the TFC membranes in-
corporated with various SNs having different surface functionalities, and B) The
normalized salt rejection of the fabricated SN-TFC membranes vs. filtration time.
M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64
Statistic results show significant correlation is associated with
desalination performance tests using SN50-OH-TFC (p¼ 0.0004)
and SN100-OH-TFC (p¼ 0.0019) membranes. The average salt re-
jections are slightly higher for the SN-NH2-TFC and SN-EPX-TFC
membranes while the correlations are not significant for all of the
samples. In fact, when comparing the salt permeability coefficients
(B) of the developed membranes, only SN50-NH2-TFC membrane
(p ¼0.0430) and SN100-NH2-TFC membrane (p ¼0.0932) show
significantly lower B values compared to the control TFC mem-
brane. When the SN-TFC membranes are compared against SN-
OH-TFC membranes, the same results are obtained except that the
B values are significantly lower for both of the SN50-NH2-TFC
(p ¼0.0181) and SN100-NH2-TFC (p ¼0.0450) membranes. These
data show that all membranes perform almost similar salt rejec-
tions with slight improvement of the rejection for the amine
functionalized SN-TFC membranes. Salt rejection data and corre-
lations are found to be independent of size variables of the in-
corporated SNs.
The observed performance results can be explained by the
higher crosslinking density of the PA layers in the SN-NH2-TFC and
SN-EPX-TFC membranes compared to those of the SN-OH-TFC and
TFC membranes as evidenced by the XPS analysis results. For the
SN-NH2 and SN-EPX, the SNs can be entrapped into the PA layer by
the covalent bond formation which ensures the higher stability of
the resulting SN-TFC membranes. However, SN-OH has no inter-
action with the PA monomers, making a looser PA structure with
the possibility of the nanoparticles’ leakage after long-time high-
pressure operation.
Another factor that contributes to this performance variation is
the alteration of the amine/acyl chloride molar ratios by the in-
troduction of different SNs. That can influence the formation of the
dense core and the loose secondary PA layers which both play an
important role in the membranes’ filtration behavior [40–42]. As
observed from TEM images, the average thickness of the PA layer
in the SN50-NH2-TFC membranes is higher than the SN50-EPX-TFC
Salt Rejection (%) membranes. Apart from the hydrophilic variation, the thickness
100
variation can also contribute to the slightly lower water flux and
higher salt rejection of the SN-NH2-TFC than the SN-EPX-TFC
Salt Rejection (%)
95 membranes. However, due to the large –NH2/–COCl molar ratio in
the PA fabrication stage and the small ratio of SN/MPD in the MPD
90 monomer solution, the chemical interaction of the functionalized
SNs with the PA monomers could have a larger influence on the
85
overall permeation behavior than the molar ratio variation. That
have almost a similar variation trend in terms of water flux and
SN50-NH2 SN50-EPX SN100 SN100-NH2 SN100-EPX salt rejection as compared to the SN-OH-TFC and control TFC
TFC/ TFN Membrane Samples membranes.
Fig. 8B shows the normalized salt rejection curves of the
membranes during a 6 h operation against 2000 ppm NaCl solu-
tion, where the data are the average of three replicates for each
membrane and each point is the ratio of salt rejection at a certain
time normalized by the initial salt rejection values. Overall, all salt
rejections experience a slight decrease by time. However, the salt
rejection values of the SN-NH2-TFC and SN-EPX-TFC membranes
show an insignificant change when compared with the SN-OH-TFC
membranes. This is due to the formation of strong chemical bonds
between the SN-NH2 or SN-EPX with the PA layers and the com-
pact structures of the SN-TFC membranes. However, the SN-OH
can only physically bond with the PA layer; and hence, might leak
2 3 4 5 6 under high-pressure process and generate some defects on the SN-
Filtration Time (h) OH-TFC membranes. The small salt rejection variability of the SN-
NH2-TFC and SN-EPX-TFC membranes confirms the high perfor-
mance stability of these membranes which is beneficial for their
optimum performance over a longer period of time.
 M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64 63

4. Conclusions functionalizing
SN-EPX Epoxy functionalized silica nanoparticles
In this study, new TFN membranes incorporated with various SN-NH2 Amine functionalized silica nanoparticles
SNs having different sizes (
50 and
100 nm) and surface func- SN-TFC SN incorporated TFC membranes
tionalities (amine, epoxy or hydroxyl groups) are developed. The PA-SN Polyamide-silica nanoparticle hybrids
research focus is to understand the interactions between the dif- APTMS (3-aminopropyl) trimethoxysilane
ferent SNs and the PA thin film layers, and study the SNs’ influ- GPMS (3-glycidyloxypropyl) trimethoxysilane
ences on the physicochemical properties and the desalination TEOS Tetraethyl orthosilicate
performance of the fabricated membranes. Different SNs could be TEM Transmission electron microscopy
successfully incorporated into the thin film PA layers of the SN-TFC SEM Scanning electron microscopy
membranes as evidenced by detailed characterizations (streaming ATR-FTIR Attenuated total reflectance Fourier transform in-
potential, SEM, TEM, XPS and FTIR). The water flux was sig- frared spectroscopy
nificantly increased by approximate 40% using the 0.1 wt% SN50 DLS Dynamic light scattering
incorporated TFC membrane while the salt rejection changes in- TGA Thermogravimetric analysis
significantly. The cross-section TEM images and FTIR and XPS EDX Energy dispersive x-ray spectroscopy
analysis results show that the functionalized SN-NH2 and SN-EPX XPS X-ray photoelectron spectroscopy
could be stably encapsulated into the PA layers. XPS analysis re- Jw Permeate water flux (LMH)
sults revealed that the SN-NH2-TFC and SN-EPX-TFC membranes Pw Intrinsic water permeability coefficient
show more stable salt rejection performance. Desalination per- Vp Collected permeate volume (L)
formance evaluation of the fabricated membranes shows that the A Active membrane surface area (m2)
SN-TFC membranes have a higher average water flux and almost t Operation time (h)
comparable salt rejection values compared to the control TFC R Salt rejection
membrane. The most significant water permeability correlation CP Equivalent NaCl concentrations in collected
coefficient is given by the SN-OH-TFC membranes, while sig- permeate
nificant salt permeability coefficient corresponding to SN-NH2-TFC CF Equivalent NaCl concentrations in feed water
membranes. Overall, the covalent bond formation between the B Intrinsic salt permeability coefficient
SN-NH2 or SN-EPX and the PA structure was found to improve the
incorporation stability of the SNs inside the PA as well as im-
proving the performance stability of the resulting SN-TFC
membranes.
References

Acknowledgements [1] L.Y. Ng, A.W. Mohammad, C.P. Leo, N. Hilal, Polymeric membranes in-
corporated with metal/metal oxide nanoparticles: a review, Desalination 308
(2013) 15–33.
The authors would like to thank our workshop team, Mr. Jason [2] M. Zargar, B. Jin, S. Dai, Development and application of reverse osmosis for
Peak, Mr. Jeffrey Hiorns and Mr. Michael Jung at the University of separation, in: K. Hu, J. Dickson (Eds.), Membrane Processing for Dairy In-
Adelaide for their help in the establishment of the cross-flow fil- gredient Separation, Wiley-Blackwell, 2015, pp. 139–175.
[3] K.P. Lee, T.C. Arnot, D. Mattia, A review of reverse osmosis membrane materials
tration unit. We also sincerely thank Dr. Sara Azari and Ms. Hanaa for desalination- Development to date and future potential, J. Membr. Sci. 370
Hegab for their help to perform the streaming potential mea- (2011) 1–22.
surements. This project was funded by a Category One Initiative [4] Gd Kang, Ym Cao, Development of antifouling reverse osmosis membranes for
water treatment: a review, Water Res. 46 (2012) 584–600.
Research Grant (IRG 13103302) and IPRS Scholarship for Masou-
[5] S.G. Kim, D.H. Hyeon, J.H. Chun, B.H. Chun, S.H. Kim, Nanocomposite poly
meh Zargar at the University of Adelaide, Australia. (arylene ether sulfone) reverse osmosis membrane containing functional
zeolite nanoparticles for seawater desalination, J. Membr. Sci. 443 (2013)
10–18.
[6] H. Zhao, S. Qiu, L. Wu, L. Zhang, H. Chen, C. Gao, Improving the performance of
Appendix A. Supporting information polyamide reverse osmosis membrane by incorporation of modified multi-
walled carbon nanotubes, J. Membr. Sci. 450 (2014) 249–256.
Supplementary data associated with this article can be found in [7] H. Dong, L. Wu, L. Zhang, H. Chen, C. Gao, Clay nanosheets as charged filler
materials for high-performance and fouling-resistant thin film nanocomposite
the online version at http://dx.doi.org/10.1016/j.memsci.2016.08.069. membranes, J. Membr. Sci. 494 (2015) 92–103.
[8] B.H. Jeong, E.M.V. Hoek, Y. Yan, A. Subramani, X. Huang, G. Hurwitz, A.
K. Ghosh, A. Jawor, Interfacial polymerization of thin film nanocomposites: a
new concept for reverse osmosis membranes, J. Membr. Sci. 294 (2007) 1–7.
[9] M.L. Lind, A.K. Ghosh, A. Jawor, X. Huang, W. Hou, Y. Yang, E.M.V. Hoek, In-
Nomenclature fluence of zeolite crystal size on zeolite-polyamide thin film nanocomposite
membranes, Langmuir 25 (2009) 10139–10145.
[10] G.L. Jadav, P.S. Singh, Synthesis of novel silica-polyamide nanocomposite
TFC Thin film composite
membrane with enhanced properties, J. Membr. Sci. 328 (2009) 257–267.
TFN Thin film nanocomposite [11] L. Jin, W. Shi, S. Yu, X. Yi, N. Sun, C. Ma, Y. Liu, Preparation and characterization
PA Polyamide of a novel PA-SiO2 nanofiltration membrane for raw water treatment, Desa-
lination 298 (2012) 34–41.
PES Polyether sulfone
[12] A. Tiraferri, Y. Kang, E.P. Giannelis, M. Elimelech, Highly hydrophilic thin-film
MPD 1, 3-phenylenediamine composite forward osmosis membranes functionalized with surface-tailored
TMC 1,3,5-benzenetricarbonyl trichloride nanoparticles, ACS Appl. Mater. Interfaces 4 (2012) 5044–5053.
SN Silica nanoparticles [13] H. Wu, B. Tang, P. Wu, Optimizing polyamide thin film composite membrane
covalently bonded with modified mesoporous silica nanoparticles, J. Membr.
SN50 Silica nanoparticles with average particle size of Sci. 428 (2013) 341–348.
50 nm [14] J. Yin, E.S. Kim, J. Yang, B. Deng, Fabrication of a novel thin-film nanocomposite
SN100 Silica nanoparticles with average particle size of (TFN) membrane containing MCM-41 silica nanoparticles (NPs) for water
purification, J. Membr. Sci. 423–424 (2012) 238–246.
100 nm [15] R.X. Zhang, L. Braeken, P. Luis, X.L. Wang, B. Van der Bruggen, Novel binding
SN-OH Original silica nanoparticles without further surface procedure of TiO2 nanoparticles to thin film composite membranes via self-
polymerized polydopamine, J. Membr. Sci. 437 (2013) 179–188.
64 M. Zargar et al. / Journal of Membrane Science 521 (2017) 53–64
[16] D. Emadzadeh, W.J. Lau, T. Matsuura, A.F. Ismail, M. Rahbari-Sisakht, Synthesis
and characterization of thin film nanocomposite forward osmosis membrane
with hydrophilic nanocomposite support to reduce internal concentration
polarization, J. Membr. Sci. 449 (2014) 74–85.
[17] J. Yin, Y. Yang, Z. Hu, B. Deng, Attachment of silver nanoparticles (AgNPs) onto
thin-film composite (TFC) membranes through covalent bonding to reduce
membrane biofouling, J. Membr. Sci. 441 (2013) 73–82.
[18] Y. Liu, E. Rosenfield, M. Hu, B. Mi, Direct observation of bacterial deposition on
and detachment from nanocomposite membranes embedded with silver na-
noparticles, Water Res. 47 (2013) 2949–2958.
[19] F. Perreault, M.E. Tousley, M. Elimelech, Thin-film composite polyamide
membranes functionalized with biocidal graphene oxide nanosheets, Environ.
Sci. Technol. Lett. 1 (2014) 71–76.
[20] L. He, L.F. Dumée, C. Feng, L. Velleman, R. Reis, F. She, W. Gao, L. Kong, Pro-
moted water transport across graphene oxide–poly(amide) thin film compo-
site membranes and their antibacterial activity, Desalination 365 (2015)
126–135.
[21] H.R. Chae, J. Lee, C.H. Lee, I.C. Kim, P.K. Park, Graphene oxide-embedded thin-
film composite reverse osmosis membrane with high flux, anti-biofouling, and
chlorine resistance, J. Membr. Sci. 483 (2015) 128–135.
[22] C.F. de Lannoy, E. Soyer, M.R. Wiesner, Optimizing carbon nanotube-re-
inforced polysulfone ultrafiltration membranes through carboxylic acid
functionalization, J. Membr. Sci. 447 (2013) 395–402.
[23] L. Dumée, J. Lee, K. Sears, B. Tardy, M. Duke, S. Gray, Fabrication of thin film
composite poly(amide)-carbon-nanotube supported membranes for enhanced
performance in osmotically driven desalination systems, J. Membr. Sci. 427
(2013) 422–430.
[24] H. Yamada, C. Urata, Y. Aoyama, S. Osada, Y. Yamauchi, K. Kuroda, Preparation
of colloidal mesoporous silica nanoparticles with different diameters and their
unique degradation behavior in static aqueous systems, Chem. Mater. 24
(2012) 1462–1471.
[25] K. Fujiwara, H. Suematsu, E. Kiyomiya, M. Aoki, M. Sato, N. Moritoki, Size-
dependent toxicity of silica nano-particles to Chlorella kessleri, J. Environ. Sci.
Health Part A 43 (2008) 1167–1173.
[26] I.Y. Kim, E. Joachim, H. Choi, K. Kim, Toxicity of silica nanoparticles depends on
size, dose, and cell type, Nanomed.: Nanotechnol. Biol. Med. 11 (2015)
1407–1416.
[27] J.S. Chang, K.L.B. Chang, D.F. Hwang, Z.L. Kong, In vitro cytotoxicitiy of silica
nanoparticles at high concentrations strongly depends on the metabolic ac-
tivity type of the cell line, Environ. Sci. Technol. 41 (2007) 2064–2068.
[28] A. Tiraferri, Y. Kang, E.P. Giannelis, M. Elimelech, Superhydrophilic thin-film
composite forward osmosis membranes for organic fouling control: fouling
behavior and antifouling mechanisms, Environ. Sci. Technol. 46 (2012)
11135–11144.
[29] W. Stöber, A. Fink, J. Colloid Interface Sci. 26 (1968) 62.
[30] S. Fouilloux, A. Désert, O. Taché, O. Spalla, J. Daillant, A. Thill, SAXS exploration
of the synthesis of ultra monodisperse silica nanoparticles and quantitative
nucleation growth modeling, J. Colloid Interface Sci. 346 (2010) 79–86.
[31] B. Chang, D. Chen, Y. Wang, Y. Chen, Y. Jiao, X. Sha, W. Yang, Bioresponsive
controlled drug release based on mesoporous silica nanoparticles coated with
reductively sheddable polymer shell, Chem. Mater. 25 (2013) 574–585.
[32] J. Lu, W. Wei, L. Yin, Y. Pu, S. Liu, Flow injection chemiluminescence im-
munoassay of microcystin-LR by using PEI-modified magnetic beads as cap-
turer and HRP-functionalized silica nanoparticles as signal amplifier, Analyst
138 (2013) 1483–1489.
[33] K. Möller, J. Kobler, T. Bein, Colloidal suspensions of nanometer-sized meso-
porous silica, Adv. Funct. Mater. 17 (2007) 605–612.
[34] Y. Sun, Z. Zhang, C.P. Wong, Study on mono-dispersed nano-size silica by
surface modification for underfill applications, J. Colloid Interface Sci. 292
(2005) 436–444.
[35] B. Deng, J. Yin, Thin-film nano-composite membrane with mesoporous silica
nanoparticles, in, Google Pat. (2014).
[36] V. Freger, J. Gilron, S. Belfer, TFC polyamide membranes modified by grafting
of hydrophilic polymers: an FT-IR/AFM/TEM study, J. Membr. Sci. 209 (2002)
283–292.
[37] A. Prakash Rao, S.V. Joshi, J.J. Trivedi, C.V. Devmurari, V.J. Shah, Structure–
performance correlation of polyamide thin film composite membranes: effect
of coating conditions on film formation, J. Membr. Sci. 211 (2003) 13–24.
[38] S. Salgın, Effects of ionic environments on bovine serum albumin fouling in a
cross-flow ultrafiltration system, Chem. Eng. Technol. 30 (2007) 255–260.
[39] A.K. Ghosh, B.H. Jeong, X. Huang, E.M.V. Hoek, Impacts of reaction and curing
conditions on polyamide composite reverse osmosis membrane properties, J.
Membr. Sci. 311 (2008) 34–45.
[40] V. Freger, Nanoscale Heterogeneity of polyamide membranes formed by in-
terfacial polymerization, Langmuir 19 (2003) 4791–4797.
[41] A.V. Berezkin, A.R. Khokhlov, Mathematical modeling of interfacial poly-
condensation, J. Polym. Sci. Part B: Polym. Phys. 44 (2006) 2698–2724.
[42] W. Xie, G.M. Geise, H.S. Lee, G. Byun, J.E. McGrath, Polyamide interfacial
composite membranes prepared from m-phenylene diamine, trimesoyl
chloride and a new disulfonated diamine, J. Membr. Sci. 403–404 (2012)
152–161.
[43] M. Elimelech, W.H. Chen, J.J. Waypa, Measuring the zeta (electrokinetic) po-
tential of reverse osmosis membranes by a streaming potential analyzer,
Desalination 95 (1994) 269–286.
[44] S. Yu, X. Liu, J. Liu, D. Wu, M. Liu, C. Gao, Surface modification of thin-film
composite polyamide reverse osmosis membranes with thermo-responsive
polymer (TRP) for improved fouling resistance and cleaning efficiency, Sep.
Purif. Technol. 76 (2011) 283–291.
[45] S.S. Deshmukh, A.E. Childress, Zeta potential of commercial RO membranes:
influence of source water type and chemistry, Desalination 140 (2001) 87–95.
[46] S.H. Kim, S.Y. Kwak, T. Suzuki, Positron annihilation spectroscopic evidence to
demonstrate the flux-enhancement mechanism in morphology-controlled
thin-film-composite membrane, Environ. Sci. Technol. 39 (2005) 1764–1770.
[47] J. Duan, Y. Pan, F. Pacheco, E. Litwiller, Z. Lai, I. Pinnau, High-performance
polyamide thin-film-nanocomposite reverse osmosis membranes containing
hydrophobic zeolitic imidazolate framework-8, J. Membr. Sci. 476 (2015)
303–310.
[48] S.Y. Kwak, S.H. Kim, S.S. Kim, Hybrid organic/inorganic reverse osmosis (RO)
Membrane for bactericidal anti-fouling. 1. Preparation and characterization of
TiO2 nanoparticle self-assembled aromatic polyamide thin-film-composite
(TFC) membrane, Environ. Sci. Technol. 35 (2001) 2388–2394.
[49] D. Emadzadeh, W.J. Lau, M. Rahbari-Sisakht, H. Ilbeygi, D. Rana, T. Matsuura, A.
F. Ismail, Synthesis, modification and optimization of titanate nanotubes-
polyamide thin film nanocomposite (TFN) membrane for forward osmosis
(FO) application, Chem. Eng. J. 281 (2015) 243–251.
[50] S. Romero-Vargas Castrillón, X. Lu, D.L. Shaffer, M. Elimelech, Amine enrich-
ment and poly(ethylene glycol) (PEG) surface modification of thin-film com-
posite forward osmosis membranes for organic fouling control, J. Membr. Sci.
450 (2014) 331–339.
[51] N. Niksefat, M. Jahanshahi, A. Rahimpour, The effect of SiO2 nanoparticles on
morphology and performance of thin film composite membranes for forward
osmosis application, Desalination 343 (2014) 140–146.
[52] M. Bao, G. Zhu, L. Wang, M. Wang, C. Gao, Preparation of monodispersed
spherical mesoporous nanosilica–polyamide thin film composite reverse os-
mosis membranes via interfacial polymerization, Desalination 309 (2013)
261–266.
[53] B. Rajaeian, A. Rahimpour, M.O. Tade, S. Liu, Fabrication and characterization
of polyamide thin film nanocomposite (TFN) nanofiltration membrane im-
pregnated with TiO2 nanoparticles, Desalination 313 (2013) 176–188.
[54] M. Zargar, Y. Hartanto, B. Jin, S. Dai, Hollow mesoporous silica nanoparticles: A
peculiar structure for thin film nanocomposite membranes, J. Membr. Sci. 519
(2016) 1–10.
[55] L.W. Zhikang Xu, Xiaojun Huang, Surface Engineering of Polymer Membranes,
Springer, Berlin Heidelberg, 2009.
[56] A. Prakash Rao, N.V. Desai, R. Rangarajan, Interfacially synthesized thin film
composite RO membranes for seawater desalination, J. Membr. Sci. 124 (1997)
263–272.
[57] G.D. Vilakati, M.C.Y. Wong, E.M.V. Hoek, B.B. Mamba, Relating thin film com-
posite membrane performance to support membrane morphology fabricated
using lignin additive, J. Membr. Sci. 469 (2014) 216–224.
[58] A.K. Ghosh, E.M.V. Hoek, Impacts of support membrane structure and chem-
istry on polyamide–polysulfone interfacial composite membranes, J. Membr.
Sci. 336 (2009) 140–148.
[59] S. Hermans, R. Bernstein, A. Volodin, I.F.J. Vankelecom, Study of synthesis
parameters and active layer morphology of interfacially polymerized poly-
amide–polysulfone membranes, React. Funct. Polym. 86 (2015) 199–208.
[60] S.H. Maruf, A.R. Greenberg, Y. Ding, Influence of substrate processing and in-
terfacial polymerization conditions on the surface topography and permse-
lective properties of surface-patterned thin-film composite membranes, J.
Membr. Sci. 512 (2016) 50–60.
[61] M. Son, Hg Choi, L. Liu, E. Celik, H. Park, H. Choi, Efficacy of carbon nanotube
positioning in the polyethersulfone support layer on the performance of thin-
film composite membrane for desalination, Chem. Eng. J. 266 (2015) 376–384.
[62] E.S. Kim, B. Deng, Fabrication of polyamide thin-film nano-composite (PA-TFN)
membrane with hydrophilized ordered mesoporous carbon (H-OMC) for
water purifications, J. Membr. Sci. 375 (2011) 46–54.
[63] A. Peyki, A. Rahimpour, M. Jahanshahi, Preparation and characterization of
thin film composite reverse osmosis membranes incorporated with hydro-
philic SiO2 nanoparticles, Desalination 368 (2015) 152–158.
[64] S. Yu, Z. Lü, Z. Chen, X. Liu, M. Liu, C. Gao, Surface modification of thin-film
composite polyamide reverse osmosis membranes by coating N-iso-
propylacrylamide-co-acrylic acid copolymers for improved membrane prop-
erties, J. Membr. Sci. 371 (2011) 293–306.
[65] M. Safarpour, A. Khataee, V. Vatanpour, Thin film nanocomposite reverse os-
mosis membrane modified by reduced graphene oxide/TiO2 with improved
desalination performance, J. Membr. Sci. 489 (2015) 43–54.
[66] S. Bano, A. Mahmood, S.J. Kim, K.H. Lee, Graphene oxide modified polyamide
nanofiltration membrane with improved flux and antifouling properties, J.
Mater. Chem. A 3 (2015) 2065–2071.
[67] E.M.V. Hoek, A.K. Ghosh, X. Huang, M. Liong, J.I. Zink, Physical–chemical
properties, separation performance, and fouling resistance of mixed-matrix
ultrafiltration membranes, Desalination 283 (2011) 89–99.
[68] M.T.M. Pendergast, J.M. Nygaard, A.K. Ghosh, E.M.V. Hoek, Using nano-
composite materials technology to understand and control reverse osmosis
membrane compaction, Desalination 261 (2010) 255–263.
[69] M.M. Pendergast, A.K. Ghosh, E.M.V. Hoek, Separation performance and in-
terfacial properties of nanocomposite reverse osmosis membranes, Desalina-
tion 308 (2013) 180–185.