Carboxylated carbon nanofibers as hydrophilic porous material to modification of cellulosic membranes for forward osmosis desalination

Carboxylated carbon nanofibers as hydrophilic porous material to modification of cellulosic membranes for forward osmosis desalination

Accepted Manuscript Title: Carboxylated carbon nanofibers as hydrophilic porous material to modification of cellulosic membranes for forward osmosis d...

1MB Sizes 4 Downloads 67 Views

Accepted Manuscript Title: Carboxylated carbon nanofibers as hydrophilic porous material to modification of cellulosic membranes for forward osmosis desalination Author: Zoheir Dabaghian Ahmad Rahimpour PII: DOI: Reference:

S0263-8762(15)00382-2 http://dx.doi.org/doi:10.1016/j.cherd.2015.10.008 CHERD 2041

To appear in: Received date: Revised date: Accepted date:

8-4-2015 1-7-2015 7-10-2015

Please cite this article as: Dabaghian, Z., Rahimpour, A.,Carboxylated carbon nanofibers as hydrophilic porous material to modification of cellulosic membranes for forward osmosis desalination, Chemical Engineering Research and Design (2015), http://dx.doi.org/10.1016/j.cherd.2015.10.008 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Highlights ● Novel carboxylated CNFs embedded CTA forward osmosis membranes were fabricated. ● Carbon nanofibers can modify the structural parameters of membrane. ● CTA FO membrane containing hydrophilic CNFs shows excellent FO performance.

Ac ce

pt

ed

M

an

us

cr

ip t

● Mechanical property of synthesized membrane was enhanced by the CNFs.

1 Page 1 of 33

Carboxylated carbon nanofibers as hydrophilic porous material to modification of cellulosic membranes for forward osmosis

cr

us

Zoheir Dabaghian, Ahmad Rahimpour*

ip t

desalination

Membrane Research Laboratory, School of Chemical Engineering, Babol University of

‫٭‬

Ac ce

pt

ed

M

an

Technology, Babol, Iran

Corresponding author

Ph: +98 111 3220342

Fax: +98 111 3220342 Email: [email protected] [email protected]

2 Page 2 of 33

Abstract In this work, cellulose triacetate (CTA) membranes containing carboxylated carbon nanofibers (CNFs) were synthesized via phase inversion method for the forward osmosis (FO) application. At the first, CNFs were functionalized using carboxyl groups (COOH) in

ip t

order to make a hydrophilic property. Fourier transform infrared spectroscopy (FTIR) accredited the formation or existence of chemical functional groups on CNFs. Then, the

cr

different contents of carboxylated CNFs (0.25, 0.5 and 1 wt%) were added in the casting

us

solution as additive to improve the FO performance of membrane. The synthesized FO membranes were characterized in terms of surface properties, structure of membranes,

an

intrinsic separation properties and as well as FO performance and subsequently compared with commercial membrane. The membrane surface hydrophilicity was enhanced with

M

increasing carboxylated CNFs content in the casting solution. The FO experiments were performed by 10 mM NaCl solution as a feed solution and 1 M NaCl solution as a draw

ed

solution in both orientation of membrane. The application of these FO membranes was also investigated for sea water desalination. The modified membrane exhibited superior

pt

performance in term of high water flux and low solute diffusion. The water flux of the

Ac ce

prepared membrane reached to a maximum of 15.6 L/m2 h which was nearly 2 times as high as that of the neat CTA membrane. The tensile strength measurement confirmed that this parameter of modified FO membrane is greater than that of the unmodified CTA.

Keywords: Nanocomposite; Cellulosic

membrane; Forward osmosis;

Desalination;

Carboxylated CNFs.

3 Page 3 of 33

1. Introduction Fresh potable water is a vital human need and thus looming water shortages threaten the world’s population live [1, 2]. To meet this challenge more effort has been put into evaluating the potential methods such as membrane-based process to recover fresh water from seawater

ip t

and brackish water [3, 4]. Reverse osmosis (RO) is currently a common technology to water purification and seawater desalination. RO has dominated for several decades but the fouling

cr

problems and intensive energy consumption are huge challenge in this process [5-7].

us

FO is considered to be new and green alternative process due to its low energy requirements. Compared to RO this process does not require hydraulic pressure. The FO process as a novel

an

membrane process employs the natural osmotic pressure as the driving force between draw solution with high osmotic pressure and feed solution to move clean water across the

M

semipermeable membrane [8, 9]. Hydration Technologies Inc. designed and produced commercial asymmetric FO membranes which are made from cellulose material [10].

ed

Recently the world's first forward osmosis seawater desalination plant have successfully implemented and operated at Al Khaluf in Oman [11]. However, internal concentration

pt

polarization (ICP) is a serious problem in FO process that significantly reduces the driving

Ac ce

force for the water transport [12-15].

In theory, an ideal FO membrane is expected to have an active layer with high water permeability and low solute permeability along with a thinner support layer with smaller structural parameter and highly porous in order to minimize internal concentration polarization [9, 16]. Especially, precious effects of nanoparticle mixed matrix membranes (MMMs) on the mitigation of membrane ICP and improvement of structure have been reported by many researchers newly. Previous studies showed the hydrophilic bulk and surface modification with different nanofiller such as (carbon nanotube) CNT [17, 18], zeolite [19] and titanium dioxide [20] may significantly enhance the membrane properties

4 Page 4 of 33

and overcome the low water flux and ICP problems of FO membranes. M. Amini et al. [17] investigated the effect of functionalized multi-walled carbon nanotubes in rejection layer of thin film nano-composite (TFN) membranes prepared by interfacial polymerization. The TFN membranes exhibited high water permeability and acceptable salt rejection in comparison with thin film composite (TFC) membrane. Wang et al. [18] prepared nano-composite membranes by dispersing

ip t

carboxylated multiwall carbon nanotubes within polyethersulfone substrate for forward osmosis

cr

desalination. The performance of these nanocomposite FO membranes with appropriate amounts of nanofiller was better than that of the commercial membrane. This was mainly due to the much porous

us

structure and smoother selective layer resulted in less ICP and higher osmotic water flux.

Among the many types of active materials, carbon nanofibers (CNFs) are of great practical

an

importance due to their superior chemical, electrical, flexibility and high aspect ratio (above 106) in combination with their unique nanostructures [21]. Also CNFs have superior

M

separation capability as well as excellent physical properties including high tensile moduli

ed

and strength, which can be used as potential fillers in the fabrication of polymeric composites such as asymmetric membranes [22-24]. However unlike related carbon materials such as

pt

multiwall carbon nanotubes, the sidewalls of CNFs are more chemically reactive and existent more options for chemical functionalization [25]. Fibers with exposed edge planes along the

Ac ce

entire interior and exterior surfaces of the nanofibers have a hollow core that is surrounded by a cylindrical fiber which makes a unique carbon nano-structure. These edge sites are reactive and facilitate chemical modification of the fiber surface improve compatibility with polymer matrix [26]. The low dispersion and chemical inertness of carbon nanostructures are the main processing limitations for preparation of mixed matrix membranes. Furthermore the uniform dispersion of CNFs in a polymer matrix has been identified as a major issue which must be addressed in the preparation of high-performance membranes [17, 18, 27]. Carboxylation is the most common and effective functionalization method through which COOH functional

5 Page 5 of 33

groups attached on CNFs surfaces and carbon nanofibers found excellent dispersibility in water and many common polar solvents [18, 28]. Herein, we report our attempt for modifying structure of CTA membrane by addition of functionalized CNFs containing carboxylic functional groups in the casting solution to

ip t

prepare the desirable membranes for FO application. It is worthwhile to note that no study has yet been reported on the possibility of FO membranes comprising CNFs in the membrane

cr

structure. The effect of carboxylated CNFs on the structure, hydrophilicity, tensile strength

us

and water flux of these membranes was studied. The capability of these novel FO membranes

an

to seawater desalination was investigated.

2. Material and method

M

2.1. Material

CNFs with average pore diameter of 8-12 nm purchased from Sigma-Aldrich were used for

ed

the preparing the carboxylated CNFs. The typical properties of this type of carbon nanofibers are listed in Table 1. CTA (43–49 wt% acetyl), cellulose acetate (CA) (39.8 wt% acetyl,

pt

average Mn~30,000) as base polymer and 1,4-dioxane were provided from Sigma-Aldrich

Ac ce

(MO, USA). Lactic acid, acetone, sodium chloride (NaCl) and methanol were purchased from Merck. Concentrated sulfuric acid and nitric acid from Scharlau were used to functionalize the CNFs. Caspian sea water with the properties provided in Table 2 was used as practical feed.

2.2. Chemical functionalization of CNFs Oxidized CNFs (carboxylated) was prepared following the traditional nitric and sulfuric oxidation process [29]. At the first, 1 g of pristine CNFs powder was added in 150 ml of mixture of concentrated H2SO4/HNO3 (3:1 v/v). The resulting nanofibers-acid mixture

6 Page 6 of 33

dispersed via sonication for 1 h (frequency 40 kHz, Model i7300 TD, Korea) at 40-45 °C. Then the solution was refluxed using magnetic stirrer quipped with reflux condenser and oil bath. This reaction consisted was kept at 70 °C for 4 h. After carboxylation the dilute slurry was filtered and back solid was washed with deionized water until neutral pH to remove

ip t

excess acid in CNFs structure. At the final stage, functionalized CNFs were dried in drying

cr

oven overnight at 90 °C. The functionalization procedure is schematically shown in Fig. 1.

us

2.3. Preparation of mixed matrix membrane

Cellulose triacetate flat sheet membranes were fabricated using phase inversion method [30].

an

At the first, the different content of carboxylated CNFs (0, 0.25, 0.5 and 1 wt%) were added to the solvents with 3:1 ratio of acetone/dioxin and stirred for 15 min. The mixture was

M

sonicated for 1h to ensure the homogeneous spread of the carbon nanofibers. After that, CTA (1 wt%) and CA (14.3 wt%) with lactic acid (5 wt%) and methanol (5.7 wt%) were added to

ed

mixture, respectively. The homogenies dope solution was obtained under magnetic stirring for 24 h at room temperature. After that the dope solutions were cast on a clean glass plate

pt

with a thickness of 75 µm using a film applicator in a room temperature (25 °C) and relative

Ac ce

humidity of about 65%. After partial evaporation time of 60 s, the cast films were subsequently immersed in tap water coagulation bath with temperature of 5 °C for 12 h to complete the phase separation process. Finally, the membranes were annealed in deionized water at 85 °C for 15 min to remove the excess solvents.

2.4. Membrane Characterization The cross-sectional images of the unmodified and modified CTA membranes were observed using by a field emission scanning electron microscope (FESEM: Mira 3-XMU). Each membrane sample was fractured under liquid nitrogen to give a consistent and clean cut.

7 Page 7 of 33

Then, all the samples were dried at room temperature for 24 h and the membranes were then sputter-coated with a thin layer of gold to minimize sample charging for observation by a EMI tech machine (model: K450E, England). The surface functional groups of CNFs were characterized by FTIR (BRUKER, model:

ip t

TENSOR 27, Germany). Sample discs were prepared by mixing 1 mg of the samples with 300 mg of KBr and scanned in the range of 400–4000 cm-1. Also the chemical composition of

cr

prepared membrane was investigated by attenuated total reflection-Fourier transform infrared

us

spectra (ATR-IR) (BRUKER, model: EQUINOX 55, Germany). The measurements were performed at room temperature.

an

In order to examine the variations in the surface hydrophilicity characteristics of the membranes, water contact angle was conducted using a contact angle meter (OCA 15 plus) at

minimize the experimental error.

M

room temperature. The value was averaged from 4 random location of each sample to

ed

The surface morphologies and roughness of the membranes were evaluated by Atomic force microscopy (AFM) (AFM model: Nanosurf easyScan 2 Flex). A tapping mode was

pt

conducted to scan the membrane surface and the scan size was 5 μm×5μm.

Ac ce

The mechanical strength of MMMs was characterized by CT3 Texture Analyzer tensile testing equipment (Brookfield engineering). The flat sheet membranes with 80 µm thickness were cut into stripes with 10 mm width and 80 mm. The average overall porosity of fabricated FO membranes was obtained by gravimetric method which equation (1) was employed to calculate the volume porosity of membrane. (1) are

and

is the polymer density. Also

is the density of water and

Where

the weight of wet and dry membrane, respectively.

8 Page 8 of 33

2.5. Measurement of membrane intrinsic separation properties Synthesized FO membranes were tested in a lab-scale cross-flow RO filtration mode with effective membrane area of 29.6 cm2 under a trans-membrane pressure of 3 bar. The desired

ip t

cross flow rate at 2 cm/s. Pure water permeability (A) was obtained by dividing the DI water

membrane area and

is the volume difference of feed solution,

(2)

(3) is the

an

is the FO water flux,

us

cr

permeate rate by the membrane effective area.

is the determined time of the test.

M

The salt rejection (R) and salt permeability (B) were measured based on conductivity

(4) (5)

pt

A (∆p - ∆π)

ed

measurement using the following equations [12, 16] :

Ac ce

Where ∆p is pressure difference and is ∆π osmotic pressure difference across the membrane.

2.6. Forward osmosis evaluation The FO performance of the fabricated membranes was evaluated by a cross flow lab-scale set-up. The detailed explanation and schematic of FO setup are given elsewhere [17]. The flows velocities of both feed and draw solution were kept at 800 ml/min during the FO tests which flowed counter-currently along the membrane. Also the temperatures of the feed and draw solutions were fixed at 25 0C during the FO test. 10 mM NaCl solution and seawater were used as a feed solution. The osmotic water flux from the feed solution to draw solution

9 Page 9 of 33

was calculated from the weight change of the feed solution using a digital mass balance (EK4100i, A&D Company Ltd., Japan): (6) and

predetermined time (

are the weight changes and volume changes respectively over a .

and

are the density of feed solution and effective

ip t

Where

cr

membrane area, respectively.

The reverse solute flux from draw solution to the feed side was calculated from the increase

us

of the feed conductivity based on conductivity measurement by the following equation [31].

and

an

and

and

are the volume of the feed measured at the beginning and the end of the test are the salt concentration of feed measured at the beginning and the end of the

M

Where

(7)

reported.

ed

test. To minimize errors, every experiment was carried out twice and the average value was

The S value can be evaluated in accordance to the classical ICP model using Eq. (8) and Eq. (9)

Ac ce

pt

for AL-FS and AL-DS, respectively [32, 33]: AL-FS

(8)

AL-DS

(9)

In addition to S parameter as a factor for evaluating the ICP phenomena, the FO membrane was tested in semi-long term duration with seawater to investigate the membrane behavior in practical. Caspian seawater was used to evaluate the FO performance in semi-long term tests. For all experiments, three different of the membrane sample were selected and the average of obtained data was reported as the final data.

10 Page 10 of 33

3. Result and discussion 3.1. Characterizations of carboxylated CNFs Fig. 2 exhibits the FTIR spectra of the pristine and modified carbon nanofibers. The sharp peak at 2372 cm-1 in the original CNFs spectrum indicated C–H stretching bands in CNFs

ip t

structure. The intensity of this peak decreased in the modified sample. This could be represented the reduction of surface C-H bonds in modified sample. Also functionalized

cr

CNFs showed a relative increase in the O-H groups stretching vibration band at 3460 cm-1.

us

This increment of the intensity can be represented the overlap of acidic hydroxylic groups obtained from COOH groups with water molecules which somehow justified the presence of

an

COOH groups in the modified CNFs structure [34]. Functionalized CNFs exhibited a new band at 1716 cm-1 which can be attributed to stretching vibrations of carbonyl groups present

M

in carboxylic acids.

Fig. 3 shows the dispersion of modified and unmodified CNFs in water. It was clearly

ed

observed that the carboxylated CNFs dispersed very well in water. This is due to presence of

pt

hydrophilic COOH functional groups in the modified CNFs.

Ac ce

3.2. Surface properties and morphological studies of CNFs FO membranes Fig. 4 shows the ATR spectra of neat and modified CTA membranes. The peak around 1740 cm–1 is due to the carbonyl vibration in the –COOR (acetyl) groups of CA/CTA (Fig. 4 a). This peak shifts a little from 1740 cm–1 to 1733cm–1 in the modified membrane due to the – COOH groups of carboxylated CNFs [35, 36]. However, the spectra of blend membranes are different from that of the original CTA membrane, showing conjugation of C=O bond of carboxyl groups with C=C bonds around 1656 cm−1 by the added CNFs [37]. The contact angle values of neat and modified membranes are given in Table 3. The unmodified CTA membrane had the highest water contact angle of 71◦. The surface 11 Page 11 of 33

hydrophilicity of the modified membranes was enhanced with an increase of functionalized CNFs amount. The contact angle of the membranes decreased from 71° to 57° with increasing carboxylated carbon nanofibers from 0 to 1 wt%. This result can be explained by the fact that the hydrophilic CNFs migrate spontaneously to the interface of cast film and

ip t

water during phase inversion to reduce the interface energy. Due to migration of the hydrophilic nanofibers to the surface of membrane, the nanofillers induce their hydrophilic

cr

properties to the membrane and lead to more adsorption of water molecules which results in

us

improve osmotic water flux [38, 39] . Owing to black color of the carbon nanofibers, the membrane top surface became darker than bottom surface of membrane indicating the

an

migration of carboxylated CNFs to the top-layer surface of membrane. This behavior was also observed for carbon nano-tubes (CNT) in the membrane structure [38, 40].

M

The cross-sectional and surface FESEM images of prepared membranes are shown in Fig. 5. FESEM images indicated that all fabricated membranes had an asymmetric structure with

ed

thickness about 60-80 µm. It can be clearly observed that a thin dense layer was formed on the surface of membranes due to rapid evaporation of the solvents from casting film.

pt

Interestingly, existence of dense bottom layer was confirmed by the high magnification of the cross

Ac ce

section image, as shown in Fig. 5. The dense bottom layer refers to the different phase inversion

process where occurs near the bottom surface. Hydrophilic CTA and CA molecules are in direct contact with the hydrophilic glass plate and would adhere upon the glass surface and aggregate to form the bottom skin during the membrane formation process [5]. Therefore the membranes are consisting of dense layer structure in both top and bottom skins and a porous middle layer. According to FESEM images, the structure of the membranes was changed significantly with addition of different amount of carboxylated CNFs. The neat CTA membrane has relatively dense asymmetric structure in cross section. Nevertheless, the addition of any functionalized CNFs into the casting solution resulted in significantly higher

12 Page 12 of 33

porosity with sponge-like nanoporous network structure (as shown in Table 3). This difference can be explained due to presence the large number of functional carboxyl groups in functionalized CNFs structure. The hydrophilic property of casting solution was increased by using hydrophilic CNF in the casting solution which increased the membrane porosity due

ip t

to more facilitated phase inversion during membrane coagulation [38, 41]. As seen in Table 3, the overall porosity of all modified membranes was higher than of all commercial CTA

cr

membranes.

us

Some macro-voids can be observed in the structure of neat membrane in Fig. 5(a). The macro-voids were disappeared and the sponge-like structure was formed when the CNFs

an

were added to the casting solution. The sponge-like structure can be explained by the good interconnection between polymer chain and CNFs and changes in phase inversion mechanism

M

[40, 42].

The surface AFM images of membranes were shown in Fig. 6. The surface roughness of

ed

fabricated membranes was examined by AFM images. The roughness of the membrane surface was decreased by incorporation of CNFs in the casting solution (as shown in Table

pt

4). With increment of hydrophilic property of dope solution due to addition of the hydrophilic

Ac ce

nanofiller, the diffusion mechanism will be fast enough during formation of the surface dense layer which leads to make a smoother surface for membrane. Also, the modified membrane with superior homogeneity in the surface can mitigate the surface roughness in compared unmodified membrane [38, 43, 44]. The S value for all synthesized FO membranes and three commercial FO membranes are given in Table 3. S value is the ratio of the tortuosity and thickness to porosity (S=lτ/ε) [31]. Based on the classical model of ICP expanded by Loeb et al., S is an exponential function of ICP. Thus the small value of S results in lower ICP and unavoidably leads to better FO

13 Page 13 of 33

performance [31, 32, 45]. It can be clearly seen that the structural values of modified FO membranes are low compared to the neat CTA and commercial FO membranes.

3.3. Effect of carboxylated CNFs loading on membrane separation properties

ip t

The pure water, salt permeability and NaCl rejection of prepared membranes are given in Table 5. The properties of three commercial CTA forward osmosis membranes developed by

cr

HTI were also pointed in this Table. The water permeability of the modified membranes is

us

higher than that of the neat and commercial CTA membrane. As can be seen, the neat CTA membrane showed the water permeability of 1.2 L/m2h bar and salt rejection of 80% when it

an

was tested with 20 mM NaCl solution under a trans-membrane pressure of 3 bar. The water permeability of 0.5 wt% CNFs modified membrane was 2.1 L/m2 h bar, which was about

M

75% higher than that of neat CTA membrane. This can be explained by the enhanced hydrophilicity of membrane and consequently the increased bulk porosity as well as sponge-

ed

like structure upon the addition of carboxylated CNFs [40].

The water permeability

decreased to 1.45 L/m2 h bar at 1 wt% CNFs membrane. Also, the salt rejection of 1 wt%

pt

CNFs membrane slightly decreased to 81%. It can be concluded that the increasing the higher

Ac ce

amount of CNFs in the casting solution had negatively affected the separation efficiency. It is caused by the localized agglomeration in membrane structure due to weak dispersion throughout the polymer solutions, which can cause reduction in the membrane surface layer integrity.

The salt permeability/water permeability (B/A) ratio was shown Table 5. This value has direct correlation with the FO membrane selectivity. The smaller B/A ratio led to lower solute reverse diffusion from the draw solution into the feed solution during FO process [17]. It should be noted that, all modified membranes prepared in this work showed smaller B/A values in comparison to the neat CTA membranes.

14 Page 14 of 33

3.4. Forward osmosis evaluation The performance of fabricated membranes in terms of water flux in both AL-FS (active layer facing feed solution) and AL-DS (active layer facing draw solution) modes were investigated.

ip t

The effect of CNFs loading on FO membrane performance using 10 mM NaCl as the feed solution and 1 M NaCl as the draw solution in both AL-DS and Al-FS membrane orientations

cr

is shown in Fig. 7. The FO water flux increased from 8.7 L/m2 h to 15.6 L/m2 h in the AL –

us

FS mode as the CNFs content increased from 0 to 1 wt%. The improved water flux for modified membranes could be attributed by three reasons. (i) Increase in the surface

an

hydrophilicity of the membrane (ii) Increase in membrane porosity (iii) Hydrophilic carbon nanofibers make hydrophilic passage due to existence of much hydrophilic COOH functional

M

groups in the membrane structure. This concept has been used to explain the water passage in the membrane structure. In this way, water molecules that coming from membrane surface

ed

easily pass through waterways in the membrane structure [46]. The comparison of FO water flux between AL-DS and Al-FS modes indicated that the

pt

orientation of membranes had no significant influence in water flux. Moreover, the reverse

Ac ce

salt flux in FO mode was very low as around 1 g/m2 h. These results can be attributed to the existence a relatively dense layer at the bottom of membrane and fully dense layer on top surface [14].

The S values for unmodified and modified membranes are provided in Table 3. The S parameter implied how severe the ICP phenomena effect in which it should have minimum values to maximize the water flux. The best modified membrane had a much smaller S value (0.6 mm) than neat CTA membrane (1.05 mm). This indicated that the CNFs incorporation in the membrane had greatly improved the mass transfer efficiency of the membrane. Higher

15 Page 15 of 33

porosity and smoother surface morphology resulted in less structure parameter and high water flux for modified membranes. The semi-long term seawater desalination test was performed using 1 M NaCl solution as draw solution and Caspian seawater as feed solution in order to investigate the effect of

ip t

membrane behavior against fouling and ICP phenomena. The obtained results are shown in Fig. 8. The osmotic flux of the neat membrane was high at initial time but the flux was

cr

decreased over time to 5.5 L/m2 h due to fouling and ICP phenomena. However, there is no

us

strong decline in water flux of modified membranes special at 0.5 wt% CNFs.

an

3.5. Mechanical property of the membranes

The effect of different CNFs loadings on the tensile properties of membranes was

M

investigated. Fig. 9 showed the stress–strain curves that obtained from tensile test. It can be observed the tensile strength of membranes increased with increasing the content of CNFs.

ed

For instance, the strength of the modified membrane with 1 wt% of CNFs loading was two times higher than that of the neat CTA membrane.

pt

As result, the incorporation of CNFs as a reinforcing agent in cellulosic membrane is helpful

Ac ce

for fabricating self-supporting membrane and tends to have higher strength. However a balance must be made between the strength and the FO flux in practical applications [18]. Therefore, it can be concluded that the membrane with 0.5 wt% of CNFs loading is desirable in term of mechanical strength and forward osmosis performance.

4. Conclusion Cellulose triacetate forward osmosis membrane were successfully modified using incorporating different amount of carboxylated carbon nanofibers ranging from zero to 1 wt% in the membrane matrix. Hydrophilic carbon nanofibers in the polymer matrix due to

16 Page 16 of 33

hydrophilic functional group can significantly improve the membrane performance in water and seawater desalination. This confirms positive effect of functionalized CNFs on producing more porous and hydrophilic membrane which results in higher osmotic flux and lower internal concentration polarization. The water flux increased to 15.5 L/m2h with increase in

ip t

content of CNFs to 0.5 wt% and then decreased in 1 wt% due to agglomeration of CNFs. Also, the formation of button dense layer due to the hydrophilic glass plate led to low solute

cr

flux in FO membrane, especially for modified membrane. The results of seawater semi-long

us

time test show that the ICP was significant reduced by addition of low content of hydrophilic nanofiller in the membrane structure. The tensile strength of the FO membranes was

an

improved by loading CNFs. More importantly, the current work demonstrates the CNFs as a key material for green technology in membrane structure are very promising for the

Acknowledgments

ed

M

development of high performance FO membranes for practicable applications.

The authors gratefully acknowledge the support given to this work by Prof. Mohsen

Ac ce

Technology.

pt

Jahanshahi, the head of Nanotechnology Research Institute of Babol University of

17 Page 17 of 33

Nomenclature water permeability coefficient (L/m2 h bar)

Aeff

effective membrane surface area (m2)

B

salt permeability coefficient (L/m2 h; m/s)

Cf

solute concentration in feed solution (mol/m3)

Cp

solute concentration in permeate solution (mol/ m3)

D

solute diffusion coefficient (m2/s)

l

thickness of membrane (mm)

Jwater

pure water flux in the RO testing mode (L/m2 h)

Js

volumetric flux of salt (L/m2 h)

Jv

volumetric flux of water (L/m2 h)

mwet

wet mass of membrane (g)

mdry

mass of membrane (g)

∆mfeed

weight change of feed solution (g)

R

salt rejection determined using a feed water containing 20 mM NaCl

Ra

mean roughness

S

membrane structural parameter (m)

∆t

measuring flux time (s) membrane porosity

τ

tortuosity

πdraw πfeed

cr

us

an

Ac ce

ε

ρwater

M

ed

pt

Greek letters

ρp

ip t

A

polymer density (g/ml) water density (g/ml)

density of draw solution (g/ml)

osmotic pressures (atm)

18 Page 18 of 33

References

Ac ce

pt

ed

M

an

us

cr

ip t

[1] A. Antony, J.H. Low, S. Gray, A.E. Childress, P. Le-Clech, G. Leslie, Scale formation and control in high pressure membrane water treatment systems: A review, Journal of Membrane Science 383 (2011) 1-16. [2] C. Feng, K. Khulbe, T. Matsuura, R. Gopal, S. Kaur, S. Ramakrishna, M. Khayet, Production of drinking water from saline water by air-gap membrane distillation using polyvinylidene fluoride nanofiber membrane, Journal of Membrane Science 311 (2008) 1-6. [3] Q. Ge, J. Su, G.L. Amy, T.-S. Chung, Exploration of polyelectrolytes as draw solutes in forward osmosis processes, Water research 46 (2012) 1318-1326. [4] K.P. Lee, T.C. Arnot, D. Mattia, A review of reverse osmosis membrane materials for desalination—development to date and future potential, Journal of Membrane Science 370 (2011) 1-22. [5] R.C. Ong, T.-S. Chung, Fabrication and positron annihilation spectroscopy (PAS) characterization of cellulose triacetate membranes for forward osmosis, Journal of Membrane Science 394 (2012) 230-240. [6] A. Peyki, A. Rahimpour, M. Jahanshahi, Preparation and characterization of thin film composite reverse osmosis membranes incorporated with hydrophilic SiO2 nanoparticles, Desalination, doi:10.1016/j.desal.2014.05.025. [7] H. Choi, J. Park, T. Tak, Y.-N. Kwon, Surface modification of seawater reverse osmosis (SWRO) membrane using methyl methacrylate-hydroxy poly (oxyethylene) methacrylate (MMA-HPOEM) comb-polymer and its performance, Desalination 291 (2012) 1-7. [8] M.F. Flanagan, I.C. Escobar, Novel charged and hydrophilized polybenzimidazole (PBI) membranes for forward osmosis, Journal of Membrane Science 434 (2013) 85-92. [9] R. Hausman, B. Digman, I.C. Escobar, M. Coleman, T.-S. Chung, Functionalization of polybenzimidizole membranes to impart negative charge and hydrophilicity, Journal of membrane science 363 (2010) 195-203. [10] J. Herron, Asymmetric forward osmosis membranes, Google Patents, 2008. [11] A. Bennett, Desalination and water reuse: What's the future for forward osmosis?, Filtration+ Separation 50 (2013) 28-34. [12] C.Y. Tang, Q. She, W.C. Lay, R. Wang, A.G. Fane, Coupled effects of internal concentration polarization and fouling on flux behavior of forward osmosis membranes during humic acid filtration, Journal of Membrane Science 354 (2010) 123-133. [13] S. You, C. Tang, C. Yu, X. Wang, J. Zhang, J. Han, Y. Gan, N. Ren, Forward osmosis with a novel thin-film inorganic membrane, Environ. Sci. Technol. 47 (2013) 8733-8742. [14] D. Emadzadeh, W. Lau, T. Matsuura, A. Ismail, M. Rahbari-Sisakht, Synthesis and characterization of thin film nanocomposite forward osmosis membrane with hydrophilic nanocomposite support to reduce internal concentration polarization, Journal of Membrane Science 449 (2014) 74-85. [15] Z. Zhou, J.Y. Lee, T.-S. Chung, Thin film composite forward-osmosis membranes with enhanced internal osmotic pressure for internal concentration polarization reduction, Chemical Engineering Journal 249 (2014) 236-245. [16] N. Niksefat, M. Jahanshahi, A. Rahimpour, The effect of SiO< sub> 2 nanoparticles on morphology and performance of thin film composite membranes for forward osmosis application, Desalination 343 (2014) 140-146. [17] M. Amini, M. Jahanshahi, A. Rahimpour, Synthesis of novel thin film nanocomposite (TFN) forward osmosis membranes using functionalized multi-walled carbon nanotubes, Journal of Membrane Science 435 (2013) 233-241.

19 Page 19 of 33

Ac ce

pt

ed

M

an

us

cr

ip t

[18] Y. Wang, R. Ou, Q. Ge, H. Wang, T. Xu, Preparation of polyethersulfone/carbon nanotube substrate for high-performance forward osmosis membrane, Desalination 330 (2013) 70-78. [19] N. Ma, J. Wei, S. Qi, Y. Zhao, Y. Gao, C.Y. Tang, Nanocomposite substrates for controlling internal concentration polarization in forward osmosis membranes, Journal of Membrane Science 441 (2013) 54-62. [20] D. Emadzadeh, W. Lau, T. Matsuura, M. Rahbari-Sisakht, A. Ismail, A novel thin film composite forward osmosis membrane prepared from PSf–TiO< sub> 2 nanocomposite substrate for water desalination, Chemical Engineering Journal 237 (2014) 70-80. [21] L. Zhang, A. Aboagye, A. Kelkar, C. Lai, H. Fong, A review: carbon nanofibers from electrospun polyacrylonitrile and their applications, Journal of Materials Science 49 (2014) 463-480. [22] Y.-K. Choi, Y. Gotoh, K.-i. Sugimoto, S.-M. Song, T. Yanagisawa, M. Endo, Processing and characterization of epoxy nanocomposites reinforced by cup-stacked carbon nanotubes, Polymer 46 (2005) 11489-11498. [23] G.J. Ehlert, Y. Lin, H.A. Sodano, Carboxyl functionalization of carbon fibers through a grafting reaction that preserves fiber tensile strength, Carbon 49 (2011) 4246-4255. [24] J.A. Mapkar, G. Iyer, M.R. Coleman, Functionalization of carbon nanofibers with elastomeric block copolymer using carbodiimide chemistry, Applied Surface Science 255 (2009) 4806-4813. [25] L. Zhang, A. Melechko, V. Merkulov, M. Guillorn, M. Simpson, D. Lowndes, M. Doktycz, Controlled transport of latex beads through vertically aligned carbon nanofiber membranes, Applied physics letters 81 (2002) 135-137. [26] L. Guadagno, M. Raimondo, V. Vittoria, L. Vertuccio, K. Lafdi, B. De Vivo, P. Lamberti, G. Spinelli, V. Tucci, The role of carbon nanofiber defects on the electrical and mechanical properties of CNF-based resins, Nanotechnology 24 (2013) 305704. [27] K. Goh, L. Setiawan, L. Wei, W. Jiang, R. Wang, Y. Chen, Fabrication of novel functionalized multi-walled carbon nanotube immobilized hollow fiber membranes for enhanced performance in forward osmosis process, Journal of Membrane Science 446 (2013) 244-254. [28] C.-F. de Lannoy, E. Soyer, M.R. Wiesner, Optimizing carbon nanotube-reinforced polysulfone ultrafiltration membranes through carboxylic acid functionalization, Journal of Membrane Science 447 (2013) 395-402. [29] J. Liu, A.G. Rinzler, H. Dai, J.H. Hafner, R.K. Bradley, P.J. Boul, A. Lu, T. Iverson, K. Shelimov, C.B. Huffman, Fullerene pipes, Science 280 (1998) 1253-1256. [30] T.P.N. Nguyen, E.-T. Yun, I.-C. Kim, Y.-N. Kwon, Preparation of cellulose triacetate/cellulose acetate (CTA/CA)-based membranes for forward osmosis, Journal of Membrane Science 433 (2013) 49-59. [31] J. Wei, C. Qiu, C.Y. Tang, R. Wang, A.G. Fane, Synthesis and characterization of flatsheet thin film composite forward osmosis membranes, Journal of Membrane Science 372 (2011) 292-302. [32] S. Loeb, L. Titelman, E. Korngold, J. Freiman, Effect of porous support fabric on osmosis through a Loeb-Sourirajan type asymmetric membrane, Journal of Membrane Science 129 (1997) 243-249. [33] G.D. Mehta, S. Loeb, Internal polarization in the porous substructure of a semipermeable membrane under pressure-retarded osmosis, Journal of Membrane Science 4 (1979) 261-265. [34] O. Netskina, O. Komova, E. Tayban, G. Oderova, S. Mukha, G. Kuvshinov, V. Simagina, The influence of acid treatment of carbon nanofibers on the activity of palladium

20 Page 20 of 33

Ac ce

pt

ed

M

an

us

cr

ip t

catalysts in the liquid-phase hydrodechlorination of dichlorobenzene, Applied Catalysis A: General 467 (2013) 386-393. [35] Y. Tian, M. Wu, R. Liu, Y. Li, D. Wang, J. Tan, R. Wu, Y. Huang, Electrospun membrane of cellulose acetate for heavy metal ion adsorption in water treatment, Carbohydrate polymers 83 (2011) 743-748. [36] Z. Ma, M. Kotaki, S. Ramakrishna, Electrospun cellulose nanofiber as affinity membrane, Journal of membrane science 265 (2005) 115-123. [37] J.-H. Choi, J. Jegal, W.-N. Kim, Fabrication and characterization of multi-walled carbon nanotubes/polymer blend membranes, Journal of Membrane Science 284 (2006) 406-415. [38] V. Vatanpour, S.S. Madaeni, R. Moradian, S. Zinadini, B. Astinchap, Fabrication and characterization of novel antifouling nanofiltration membrane prepared from oxidized multiwalled carbon nanotube/polyethersulfone nanocomposite, Journal of Membrane Science 375 (2011) 284-294. [39] E. Celik, H. Park, H. Choi, H. Choi, Carbon nanotube blended polyethersulfone membranes for fouling control in water treatment, Water research 45 (2011) 274-282. [40] V. Vatanpour, M. Esmaeili, M.H.D.A. Farahani, Fouling reduction and retention increment of polyethersulfone nanofiltration membranes embedded by amine-functionalized multi-walled carbon nanotubes, Journal of Membrane Science 466 (2014) 70-81. [41] A. Rahimpour, S. Madaeni, S. Mehdipour-Ataei, Synthesis of a novel poly (amideimide)(PAI) and preparation and characterization of PAI blended polyethersulfone (PES) membranes, Journal of membrane science 311 (2008) 349-359. [42] S. Majeed, D. Fierro, K. Buhr, J. Wind, B. Du, A. Boschetti-de-Fierro, V. Abetz, Multiwalled carbon nanotubes (MWCNTs) mixed polyacrylonitrile (PAN) ultrafiltration membranes, Journal of Membrane Science 403 (2012) 101-109. [43] M. Peyravi, A. Rahimpour, M. Jahanshahi, Thin film composite membranes with modified polysulfone supports for organic solvent nanofiltration, Journal of Membrane Science 423 (2012) 225-237. [44] M. Peyravi, A. Rahimpour, M. Jahanshahi, A. Javadi, A. Shockravi, Tailoring the surface properties of PES ultrafiltration membranes to reduce the fouling resistance using synthesized hydrophilic copolymer, Microporous and Mesoporous Materials 160 (2012) 114125. [45] W.C. Lay, J. Zhang, C. Tang, R. Wang, Y. Liu, A.G. Fane, Factors affecting flux performance of forward osmosis systems, Journal of Membrane Science 394 (2012) 151-168. [46] J.R. McCutcheon, M. Elimelech, Influence of membrane support layer hydrophobicity on water flux in osmotically driven membrane processes, Journal of Membrane Science 318 (2008) 458-466.

21 Page 21 of 33

Table 1 Typical physical properties of carbon nanofibers. Nanofiber density (including hollow core)

(lb/ft3)

1.2 – 3.0

3

1.4 - 1.6

3

(g/cm )

ip t

Average bulk density

(g/cm )

2.0 - 2.1

Average catalyst (Iron) content

(ppm)

< 100

Average outer diameter

(nm)

Average inner diameter

(nm)

125 - 150 50-70

us

2

cr

Nanofiber wall density

m /g

Total pore volume

(cm3/g)

0.075

Average pore diameter

(angstroms Å)

123.99

µm

50-100

an

Average specific surface area

ed

M

Average lengths

20 - 30

Sulfate 1500 ppm

Magnesium 500 ppm

Calcium 160 ppm

Potassium 100 ppm

Sodium 4470 ppm

Chloride 5515 ppm

pH 6.5

Total dissolved solid 6000 ppm

Ac ce

Conductivity 13.5 µs

pt

Table 2 The properties of Caspian seawater

22 Page 22 of 33

Table 3 Structural properties of synthesized FO membranes. S valueb (mm)

Contact angle Reference (deg.)

CTA

0

69 ± 2

1.05 ± 0.15

71± 3

Current work

CNF-0.25

0.25

79 ± 1

0.66 ± 0.1

61± 2

Current work

CNF-0.5

0.5

83 ± 1

0.60 ± 0.05

59.5± 2

1

80 ± 1

0.74 ± 0.1

57± 2

0

64 ± 1

0.72 ± 0.15

63 ± 3

0

46 ± 1

1.00 ± 0.54

73 ± 2

[31]

0

50 ± 2

1.38 ± 0.26

64 ± 2

[31]

CTA-Wd CTA-NW

e

Current work

[31]

cr

CTA-HW

c

Current work

us

CNF-1

ip t

Membranes Concentration Porositya (wt%) (%)

Determined by gravimetric measurement and water as a wetting solvent.

b

Determined from FO water flux results.

c

Cellulose triacetate (CTA) with polyester highly woven fabric.

d

Cellulose triacetate (CTA) with polyester woven fabric.

e

Cellulose triacetate (CTA) with polyester non-woven fabric.

ed

M

an

a

Table 4 Surface roughness parameters of unmodified and modified membranes. Membranes

Sa (nm)

CTA CNF-0.25

Sz (nm)

21.37

31.40

353.66

15.79

24.84

204.11

Ac ce

pt

Sq (nm)

CNF-0.5

7.45

11.85

120.38

CNF-1

5.93

8.61

128.05

23 Page 23 of 33

Table 5 Separation properties of synthesized FO membranesa Pure water permeabilityb

NaClc

Salt permeabilityd

B/A

(l/

Rejection (%)

(

Kpa

h bar)

(

m/s Pa)

m/s)

Reference

ip t

Membranes

1.2 ± 0.2

3.4 ± 0.4

80 ± 2

25 ± 2

73 ± 4

Current work

CNF-0.25

1.9 ± 0.15

5.3 ± 0.6

84 ± 3

30.2 ± 1.5

57 ± 3

Current work

CNF-0.5

2.1 ± 0.15

5.9 ± 0.6

87 ± 1

26.1 ± 1.5

1.45 ± 0.1

4.1 ± 0.4

81 ± 2

CTA-W

e

CTA-NW

1.19 ± 0.19 0.33 ± 0.04

e

0.46 ± 0.07

3.3 ± 0.5 0.9 ± 0.1 1.3 ± 0.2

44 ± 3

us

CTA-HW

e

Current work

28.4 ± 1

69 ± 2

Current work

78.5

f

25.6 ± 1.4

84 ± 8

[31]

81.9

f

4 ± 0.9

47 ± 12

[31]

92.4

f

22 ± 3

[31]

an

CNF-1

cr

CTA

2.7 ± 0.2

All experimental data are reported as the average of at least three repeated measurements.

b

Were measured in the RO testing mode over an applied pressure of 3 bar and DI water as feed solution.

c

Were measured in the RO testing mode over an applied pressure of 3 bar and 20 mM NaCl as feed solution.

d

Were measured in the RO testing mode over an applied pressure of 3 bar and 20 mM NaCl as feed solution.

M

a

e

pt

Were measured in the RO testing mode over an applied pressure of 375 kPa and 20 mM NaCl as feed solution.

Ac ce

f

ed

All of these data are reported by Tang and co-workers in the RO testing mode and an applied pressure of 100500 kPa.

24 Page 24 of 33

ip t cr us an M

Ac ce

pt

ed

Fig. 1. Procedure for the preparation of carboxylated CNFs.

25 Page 25 of 33

ip t cr us an M ed pt

Ac ce

Fig. 2. FTIR spectra of (a) pristine CNFs, (b) carboxylated CNFs.

26 Page 26 of 33

ip t cr us an M ed

Fig. 3. Dispersion of pristine and functionalized carbon nanofibers in water: (a) pristine CNFs (b)

Ac ce

pt

Functionalized CNFs.

27 Page 27 of 33

ip t cr us an M ed

Ac ce

pt

Fig. 4. ATR-IR spectra of (a) neat CTA membrane, (b) carboxylated CNFs (0.5 wt%).

28 Page 28 of 33

ip t cr us an M ed pt Ac ce Fig. 5. Cross-sectional FESEM microphotographs of all fabricated membranes with different CNF loadings: (a) CTA, (b) CNF-0.25, (c) CNF-0.5 and (d) CNF-1.

29 Page 29 of 33

ip t cr us an M ed pt Ac ce

Fig.6. Surface AFM images of the CNF/CTA membranes with different concentrations of oxidized carbon nanofibers (a) CTA only, (b) 0.25%, (c) 0.5% and (d) 1 % carboxylated CNFs.

30 Page 30 of 33

ip t cr us an M

Ac ce

pt

ed

Fig. 7. The water flux (in both AL-FS and AL-DS modes) and reverse salt flux (in AL-FS mode)

31 Page 31 of 33

ip t cr us an M ed

Ac ce

pt

Fig. 8. The seawater FO flux during semi- long term test.

32 Page 32 of 33

ip t cr us an M ed

Ac ce

pt

Fig. 9. Stress–strain profiles of unmodified and modified membranes.

33 Page 33 of 33