A novel strategy for enhancing the electrospun PVDF support layer of thin-film composite forward osmosis membranes

Abstract

A simple and novel treatment methodology is introduced to produce PVDF-based thin-film composite forward osmosis (TFC-FO) electrospun membranes for enhanced desalination performance. The proposed treatment strategy is based on improving the surface properties of the PVDF electrospun nanofiber support layer using triethylamine (TEA). The results indicated that this strategy enhanced the interfacial polymerization step by overcoming the hydrophobicity feature dilemma of the PVDF support layer. As an FO membrane, the characterization of the modified membrane shows a distinct decrease in the structure parameter of the support layer by 67%, which mitigates the adverse effect of the internal concentration polarization (ICP) by 45%. Moreover, the performance of the modified TFC-FO membrane exhibited a high water flux, approximately 68 LMH and low reverse salt flux, about 2 g m⁻² h⁻¹ at 2 M NaCl draw solution, with >99.5% salt rejection. Overall, the introduced modification technique has the advantages of being inexpensive, easy to implement, and appropriate for commercial membranes.

---

1. Introduction

Forward osmosis (FO) technology, as a promising desalination technology, has attracted considerable attention due to its simplicity and energy saving. Typically, in FO, the desalination process depends only on osmotic pressure as a driving force for water flux through a semi-permeable membrane, which saves the high-energy requirement for conventional techniques. In addition to energy conservation, FO has more advantages over reverse osmosis (RO) and thermal distillation processes such as a lower fouling tendency and relatively higher recovery.¹ Moreover, besides desalination of sea and brackish water, FO is more attractive for other applications such as the treatment of wastewater (i.e. the removal of toxic ions), power regeneration, concentration of pharmaceutical products and liquid food processing.^2–5

Basically, the FO system consists of three main parts: a draw solution (DS, high osmotic pressure), a feed solution (FS, low osmotic pressure) and a semipermeable membrane. The membrane is the heart of this process; therefore, it is responsible for the selectivity, water flux and salt rejection. Accordingly, many efforts were exerted to produce effective membranes. Several membranes have been introduced such as cellulose acetate (CA) or cellulose triacetate (CTA),^6,7 polyamide and composite polyamide or thin film composite (TFC).^8–10 Among the reported FO membranes, polyamide and TFC membranes exhibit excellent salt rejection, high water flux and distinct stability over a broad pH range (2–12).¹¹ TFC membranes are composed of a thin film polyamide selective layer (PA) seated on the top of a support layer. Recently, this class of membrane has a wide prevalence as an effective FO membrane.¹²

The desired characteristics for optimum TFC membranes are a defect-free ultrathin selective layer with excellent permeability (high water flux) and selectivity (high salt rejection), a high porous thin support layer with small structural parameter (S) to decrease the internal concentration polarization (ICP), excellent mechanical properties, and excellent hydrophilicity to augment water flux and reduce fouling.¹³ The TFC membrane performance can enhanced by modifying the support layer and/or the active selective layer. For instance, to modify the active selective layer, incorporation of hydrophilic nanoparticles into a polyamide layer (PA) such as hydrophilic functionalized titanate nanotubes (TNTs),¹⁴ zeolite nanoparticles,^15,16 halloysite nanotubes (HNTs)¹⁷ and TiO₂ NPs¹⁸ has been carried out. In addition, the enhancement of the support layer was achieved using many polymers to fabricate the TFC membranes, such as PVDF,¹⁹ polyethersulfone,²⁰ polytriazole-co-polyoxadiazole,²¹ polysulfone,²² cellulose ester,²³ polyketone²⁴ and PAN/cellulose acetate (CA) blend,²⁵ in order to improve the support layer characteristics and augment the FO desalination process.

Among the polymers utilized for the support layer, PVDF, a semi-crystalline polymer, has emerged as an attractive polymer used for support layer fabrication due to its excellent properties such as thermal stability, chemical resistance, and mechanical properties. Unfortunately, the hydrophobicity of PVDF is a big dilemma constraining its wide application in FO desalination process.^26,27 Typically, the hydrophobicity feature leads to a discrete layer of the active polyamide film because the first step in the interfacial polymerization process uses an MPD aqueous solution.

Besides the nature of the used polymer, the fabrication technique of the membrane has a vital role in the support layer performance.²⁸ When compared to the phase inversion technique, used for membrane fabrication, electrospinning is a versatile technique to produce a class of nanofiber materials that reveals intrinsically high porosity with an interconnected pore structure, which can be used as a promising candidate for the TFC membrane support layer. Moreover, the electrospun support layer has many advantages over conventional layers fabricated using a phase inversion method such as high porosity, low structure parameter (low tortuosity, which leads to a hindered ICP) and low thickness (controlled by time or electrospun solution volume).²⁸

Herein, to modify the wettability, the PVDF electrospun nanofiber mat was improved using a simple and highly effective modification process, involving treatment with TEA before the interfacial polymerization process. Simply, the PVDF electrospun nanofiber support layer was immersed in TEA for 1 min, and then the aqueous phase was poured onto the support surface. The ​treatment purpose is to achieve two objectives: (1) to create cross-linking on the support layer and (2) to enhance the support layer wettability, which allows the MPD aqueous phase to wet the support layer and consequently, fabricate a continuous and smooth PA layer. The results revealed a distinct enhancement in the TFC properties as well as a highly improved performance in the FO unit.

2. Experimental

2.1 Materials

Poly(vinylidene fluoride) (PVDF) (average M_w ∼ 275 000, Sigma-Aldrich) and N,N-dimethylformamide (DMF, >99.5%, Sigma-Aldrich) were purchased from Sigma-Aldrich and used to prepare the electrospun nanofiber PVDF mats. Triethylamine (TEA, Samchun) was utilized for the modification of the membrane. m-Phenylenediamine (MPD, >98%, Alfa Aesar) and 1,3,5-benzenetricarbonyl chloride (TMC, 98%, Alfa Aesar) were used in the synthesis of the polyamide layer on the surface of the electrospun nanofiber membrane. Unless specified, all of the chemicals and reagents used in this study were of analytical grade and used as received without further treatment. Furthermore, sodium chloride (NaCl, Sigma-Aldrich), magnesium chloride (MgCl₂, Sigma-Aldrich) and distilled water were used for the FO experiments.

2.2 Preparation of the electrospun PVDF support layer of the FO membrane

Four grams of PVDF was dissolved in 16 g of DMF at 60 ± 2 °C to prepare a 20% PVDF solution. The mixture was stirred for 12 h using a hot plate magnetic stirrer. Then, the viscous solution was transferred to a 5 mL plastic syringe to be electrospun using a laboratory scale electrospinning apparatus. An 18 ± 0.1 kV voltage was applied between the spinneret (plastic syringe) and the rotating stainless steel (170 rpm) collector covered with a polyethylene sheet to fabricate the PVDF electrospun nanofiber membrane wherein the distance was kept at 15 ± 0.5 cm. The electrospinning process was carried at room temperature (23 ± 3 °C) and 55% ± 7% humidity. The tip diameter was 0.45 mm. Finally, the produced PVDF membranes were dried at 70 °C in a vacuum oven overnight before use.

2.3 Modification of the support layer and the interfacial polymerization step

An effective, easy, fast and low-cost modification process was conducted to enhance the properties of the support layer, such as the wettability, and form cross-linking between the nanofibers. The fabricated electrospun support layer was immersed in TEA for 1.0 min. For comparison, the polyamide layer was fabricated on the surface of the pristine support layer (PVDF) and the modified support layer (PVDF–TEA) (as shown in Scheme 1). The polyamide layer was synthesized as follows: the support layer was immersed in 2 wt% aqueous MPD solution for 2 min to allow good penetration of the MPD monomers. Then, the membrane was dried using filter paper and was left for 1 min before immersing into 0.15 wt% TMC organic solution (in hexane) for 1 min. Finally, the membrane was washed three times with hexane and distilled water to remove the unreacted monomer and by products such as hydrochloric acid (HCl) and dried for 5 min in a drying oven at 80 °C. Finally, the TFC membranes (the treated; TFC–PVDF–TEA and untreated; TFC–PVDF) were stored in deionized water. It can be noted that in the case of pristine PVDF, before the synthesis of the polyamide layer, the membrane was first washed with 20 wt% 2-propanol to overcome the membrane hydrophobicity and enhance the surface wettability before immersing into the aqueous MPD solution. Moreover, before performing the polymerization, the pristine PVDF membrane was washed 5 times with water to ensure the removal of propanol.

Scheme 1 Interfacial polymerization of m-phenylene diamine (MPD) with trimesoyl chloride (TMC).²⁹

2.4 Characterization of the membrane

The surface morphology of the membranes was studied using a scanning electron microscope equipped with an EDX analysis tool (SEM and FE-SEM, Hitachi S-7400; Japan). The wettability, surface energy and the work of adhesion of the membranes were measured using a ramé-hart Contact Angle Goniometer, which was used to measure the water contact angle (WCA) of the membrane. Then, the average for the three samples was calculated and used in the results section. The chemical structure of the membranes was investigated using FTIR spectroscopy (Mode: FTLA 2000 series, ABB) and Raman Spectroscopy (FT-Raman spectroscopy, RFS-100S, Bruker, Germany). Furthermore, the water flux and the reverse salt flux of the TFC membranes were determined using the FO system in AL-FS mode and the salt rejection of the membrane was measured based on the conductivity data of the feed and draw solution using LabQuest 2 connected to the data logger.

2.5 Forward osmosis system and evaluation of the membrane

All FO tests were conducted in the FO cell designed and implemented locally in South Korea wherein the feed solution was faced to the polyamide layer of the membranes (AL-FS mode). Moreover, the FO system was operated in counter-current flow and both the feed and draw solutions were maintained at room temperature and enter into the system with a 0.1 L min⁻¹ volumetric flow rate; the system setup is presented in Fig. 1.

Fig. 1 Schematic of the FO setup.

Four different concentrations of NaCl (0.5, 1, 1.5 and 2 M) were used as a draw solution (DS) to investigate the membranes performance. Distilled water was used as a feed solution (FS). A digital balance was used to measure the decrease in the mass of the feed solution in order to compute the water flux (J_v, LMH) through the FO membrane. Moreover, the balance was connected to a PC to record the decrease in FS with time for 1 h. Conversely, the reverse salt flux (J_s, g m⁻² h⁻¹) was calculated by measuring the conductivity of the FS at the beginning and the end of the test using a conductivity meter (LabQuest 2). Numerically, eqn (1) and (2) were used to calculate the water flux and reverse salt flux, respectively.³⁰ The effective area of the membrane was 5 cm² and all the experiments were conducted at room temperature.

where Δm (g) is the mass of the water passing through the membrane with an effective surface area A (m²) at interval time Δt (h) and ρ is the water density (997 g L⁻¹).V_t and C_t are the volume (L) and the salt concentration (g L⁻¹) of the feed at the end of the experiment, respectively. V₀ and C₀ are related to the volume and salt concentration at the beginning of the test (t = 0), respectively. The water permeability coefficient (A, L m⁻² h⁻¹ bar⁻¹), the salt permeability coefficient (B, L m⁻² h⁻¹) and the structural parameter (S, μm) were determined using an Excel-based algorithm established by Alberto Tiraferri et al.³¹ Herein, four different concentrations of NaCl draw solutions were used: 0.5, 1, 1.5 and 2 M. Furthermore, the salt rejection was carried out in the FO system in which a 200 ppm NaCl solution was used as a feed solution and 2 M MgCl₂ was used as a draw solution. The membrane salt rejection (R, %) was calculated using eqn (3)where C_ft and C_f0 are the concentration of NaCl in the feed solution at time t and 0, respectively, and C_Ds0 is the initial concentration of NaCl in the draw solution.³²

3. Results and discussion

3.1 Characterization of the membrane

3.1.1 Chemistry and structure of the membranes. The FT-IR spectra of the membranes before and after activation and Raman spectra are presented in Fig. 2. Fig. 2A displays the full spectra of the support layer of the pristine (PVDF) and the treated support layers (PVDF-TEA). A strong peak at 1179 cm⁻¹ was assigned to the C–F stretching that can be observed over a wide range, 1400–1000 cm⁻¹. The peak at 1401 cm⁻¹ indicates the presence of –CH₂– groups, whereas the peak in the region of 830–520 cm⁻¹ refers to the C–F deformation vibrations.³³ Moreover, the peaks at 2905 and 2976 cm⁻¹ observed for the PVDF and PVDF–TEA membranes were attributed to C–H stretching vibrations.³⁴

Fig. 2 The FT-IR full spectra of the support layers (A), specific spectra for the support layers in the range from 1500 to 400 cm⁻¹ (B), the Raman shifts for the support layers (C) and the full FT-IR spectra of the TFC membranes (D).

Furthermore, the phases of pristine PVDF and treated PVDF are shown in Fig. 2B and S1.† It is known that PVDF can be found in different phase structures. Herein, the α-phase and β-phase were found wherein the α-phase is distinguished at 762 cm⁻¹ (CF₂ bending) and 795 cm⁻¹ (CH₂ rocking), whereas the β-phase is displayed at 840 (CH₂ rocking), 1276 and 1431 cm⁻¹.^35,36 As shown, the intensity of the peaks of the β-phase was decreased in the PVDF–TEA membrane due to the crosslinking in PVDF as a result of the treatment with TEA.

Moreover, the calculation method used to determine the β-phase content has been explained in our previous work.¹⁹ In brief, the intensity of the absorbance peak at 840 cm⁻¹ for the β-phase and at 762 cm⁻¹ for the α-phase was calculated and used in eqn (4) to estimate the amount of β-phase.^37,38

where (β%) is the percentage of β-phase content; A_β and A_α are the infrared absorbance at 840 and 762 cm⁻¹, characteristic of the β-and α-phases, respectively; and K_β (7.7 × 10⁴ cm² mol⁻¹) and K_α (6.1 × 10⁴ cm² mol⁻¹) are the absorption coefficient at the respective wavenumber.³⁷

Numerically, as shown in Fig. S1,† the β-phase content decreased from 86.5% ± 3.3% in pristine PVDF to 79.4% ± 1.25% in the treated membrane due to the decrease in the absorbance intensity of the β peak at 840 cm⁻¹, as shown in Fig. 2B.

The Raman results of the pristine PVDF support layer, in Fig. 2C, show that the peaks at ∼839 cm⁻¹ and ∼1432 cm⁻¹ are assigned to the β and γ bands, which are attributed to the rocking CH₂ and asymmetric stretching CF₂ vibrations (∼839 cm⁻¹) and bending CH₂ vibrations (∼1432 cm⁻¹).³⁹ Moreover, the band at 2974 cm⁻¹ was assigned to the CH₂ symmetric stretching.⁴⁰ In contrast, treatment of the PVDF support layer with TEA introduced changes in the crystal structure wherein the band at ∼839 cm⁻¹ disappeared, which is typical for the β-crystal phase. In addition, the bands at ∼2974 cm⁻¹ and ∼1432 cm⁻¹ also disappeared for the PVDF–TEA support layer owing to a decrease in the intensities of the crystal peaks of PVDF.³⁹ Moreover, the band at 1696 cm⁻¹ was attributed to C=C str.³³ Overall, the Raman results confirmed the change in the structure and crystallinity of the modified support layer (PVDF–TEA), which proved the formation of cross-linking in the material.

The FT-IR spectra for the TFC membranes covered with a polyamide layer are presented in Fig. 2D. As shown, the peaks at 1544 and 1616 cm⁻¹ can be assigned to the amide II band (N–H in-plane bending and N–C stretching vibration of a –CO–NH– group) and aromatic amide (N–H) deformation vibrations,⁴¹ respectively. The stretching vibrational band at 1665 cm⁻¹ was assigned to the C=O stretching of carboxylic acids, referred to the amide I band, which was attributed to the polyamide layer.^41–43 Moreover, a broad peak, appearing at 3140–3600 cm⁻¹ and centered at 3340 cm⁻¹, represents the N–H stretching of the primary and secondary amines.

3.1.2 Morphology of the membranes. The surface morphology is an important factor used to characterize the support layer of the FO membrane. Therefore, SEM and FE-SEM were used to analyze the surface morphology. Fig. 3 reveals the surface FE-SEM images and EDX mapping of the PVDF and PVDF treated (PVDF–TEA) membranes. Fig. 3A and C display the FE-SEM image and EDX mapping, of the pristine PVDF electrospun nanofibers support layer, respectively, wherein smooth and bead-free nanofibers are observed. The PVDF–TEA morphology is introduced in Fig. 3B in which the cross-linking between the nanofibers is clearly observed and EDX (Fig. 3D) was used to estimate the element percentage. Moreover, the cross-linking was related to the effect of TEA on the electrospun PVDF nanofibers. On the other hand, the big balls in Fig. 3B may be related to the agglomeration of short electrospun nanofibers and this pretension was supported by the EDX results (Fig. 3D and S2†). However, further explanation is provided in the ESI based on Fig. 3 and S2.†

Fig. 3 FE-SEM images and corresponding EDX mapping of the electrospun support layer: (A and C) PVDF and (B and D) PVDF–TEA.

Based on the aforementioned characterization, the cross-linking formed can be explained as follows: first, double bonds are formed within the polymer chain due to the elimination of the HF from the polymer. Later on, crosslinking is formed by the conjunction of these double bonds, as shown in Scheme 2, which displays a conceptual illustration of the expected mechanism. From the EDX results in Fig. 3C and D, it can be noticed that the fluorine weight percentage was reduced by 31% for PVDF–TEA, which confirmed the elimination of HF from PVDF in the presence of TEA. In addition, the changes in the crystallinity and phase structure of PVDF–TEA based on the Raman and FT-IR analysis support this finding.

Scheme 2 The proposed mechanism of cross-linking in the PVDF–TEA support layer.^39,44,46

It can be noted that the mechanism of dehydrofluorination and crosslinking of the fluorinated polymers in the presence of an amine, diamine and also ammonia has been previously reported.^44–46

Moreover, the polyamide layer morphology was characterized by recording the SEM and FE-SEM images of the surface of the membranes after the interfacial polymerization process. From Fig. 4A and B, it can be observed that the polyamide layer was formed in both the pristine and the treated PVDF electrospun support layers. Moreover, the top surface of the polyamide layer in the case of the pristine PVDF membrane was thinner than the PVDF–TEA membrane, as shown in Fig. 4C and D. In addition, the nanofibers can be seen in the event of the pristine membrane (Fig. 4C), while the PA layer was denser in the case of the treated material. Formation of a relatively thick and continuous PA layer upon treatment with TEA was assigned to the improvement of the membrane's wettability, which will be discussed later. EDX mapping indicated that the nitrogen and oxygen contents were increased in the treated TFC membrane, as shown in the lower panels of Fig. 4C and D.

Fig. 4 FE-SEM images and EDX mapping of the TFC membranes: (A) the top surface of the TFC–PVDF membrane (the inset is the high magnification image), (B) the top surface of the TFC–PVDF–TEA membrane (the inset is the high magnification image), (C) EDX of the TFC–PVDF membrane and (D) EDX of the TFC–PVDF–TEA membrane.

3.1.3 Wettability, work of adhesion and surface energy of the TFC-FO membrane. The wettability of a membrane is considered a vital factor. In particular, the hydrophilicity of the support layer enhances the interfacial polymerization as well as decreases the internal concentration polymerization (ICP). The water contact angle, work of adhesion and surface energy of the TFC membranes and support layers were determined to characterize the membranes. The WCA of the electrospun PVDF support layer was measured before and after treatment with TEA and the results (shown in Fig. 5 and S3, ESI†) indicate that the WCA for the pristine support layer (PVDF) was 137°, which decreased to approximately 0 after wetting with TEA (Fig. S3†). The instantaneous super-hydrophilicity of the treated membrane enhanced the wetting of the modified surface by the MPD aqueous solution as well as enhanced the distribution of the aqueous phase on the support layer.

Fig. 5 The wettability (A) and work of adhesion and the surface energy (B) of the modified and unmodified membranes before and after the activation.

Moreover, the wettability of the top (PA layer) and bottom (support layer) surfaces of the TFC membranes was measured and plotted in Fig. 5 and S3.† The results indicate that the hydrophilicity of the modified TFC–PVDF–TEA membrane was increased for both surfaces. Numerically, the WCA was 39.56° and 81.5° for the top and support layer, respectively. On the other hand, the top surface (active layer) of the pristine TFC–PVDF membrane showed a high WCA (about 87.9°) and the bottom surface (support layer) exhibits a hydrophobic surface; the WCA was 109.3°. The enhancement of the hydrophilicity of the modified membrane was attributed to the increase in the surface energy and the work of adhesion, as shown in Fig. 5, as a result of the increase in the nitrogen and oxygen contents on the surface (Fig. 4C and D).

Furthermore, the PVDF support layer revealed a minuscule value for the surface energy as well as the work of adhesion; this was related to the fluorine atoms, which have a low surface energy. Instead, the surface energies of the top surfaces were 29.8 and 52.9 mJ m⁻² for the pristine TFC–PVDF and the modified TFC–PVDF–TEA, respectively. The increase of the latter was associated with the increase in the nitrogen contents as discussed beforehand. Alternatively, the bottom surface (support layer) of the TFC–PVDF membrane displayed a high surface energy when compared to the support layer before the polymerization step; this increase was attributed to the formation of PA at some point on the bottom side of the membrane. However, the increase in the surface energy of the bottom surface (support layer) for TFC–PVDF–TEA by approximately 7 times was related to the reduction of the fluorine percentage, as shown in Fig. 3D, as well as the formation of PA at some point on the bottom. Moreover, the work of adhesion between the water and the membrane surface was measured; the results supported the surface energy data wherein the same trend was observed in the work of adhesion. Overall, the modified TFC–PVDF–TEA membrane exhibited an excellent hydrophilicity for both the polyamide and support layers. The surface energy and the work of adhesion results support and confirm this finding.

3.2 Evaluation the membrane in FO experiments

The big challenge in FO membrane fabrication is the ICP, which is mainly affected by the support layer structure parameter (S, μm). Determination of the water permeability coefficient (A, L (m² h bar)⁻¹) and the salt permeability coefficient (B, L (m² h)⁻¹) is necessary to characterize the membrane performance in FO. The transport parameters were estimated using an algorithm (Microsoft Visual Basic within Microsoft Excel) that was developed by other researchers.³¹ In brief, the experimental data of the water and reverse salt flux from the FO experiments, using 4 different draw solution concentrations, were used to estimate the intrinsic transport parameters, A, B, and S, and the results are reported in Table 1 (more explanation in ESI,† Section 3.2). Moreover, the results show that the coefficient of variation of J_w/J_s (CV) was less than the recommended value (10%). Furthermore, the coefficients of determination of the water (R_w²) and salt (R_s²) fluxes were >0.99% and >95% for the TFC–PVDF–TEA and pristine TFC–PVDF membranes, respectively. These values achieved the second factor for the reliability of the results, which must be greater than 0.95.

Table 1 The properties and structural parameters of the TFC–PVDF and TFC–PVDF–TEA membranes

	A (L m^−2 h^−1 bar^−1)	B (L m^−2 h^−1)	S (μm)	R^2 − Jw	R^2 − Js	CV%
TFC–PVDF	0.98	0.16	89.7	0.959	0.975	9.59
TFC–PVDF–TEA	0.96	0.03	29.5	0.993	0.993	8.2

---

From Table 1, the modified PVDF support layer (PVDF–TEA) has the lowest structural parameter (S), which enhanced the membrane performance by decreasing the ICP. Correspondingly, the salt permeability (B, L m⁻² h⁻¹) was decreased for the modified TFC membrane by 5.3 times when compared to the pristine TFC–PVDF, as shown in Table 1. On the other hand, a negligible change in the water permeability (A, L m⁻² h⁻¹ bar⁻¹) was observed (Table 1). Overall, from these characterizations, it can be concluded that the modified TFC membrane was more suitable for FO than the pristine one.

On the other hand, the water and reverse salt fluxes (Fig. 6) were determined to evaluate the membranes using a laboratory scale FO system (Fig. 1) wherein eqn (1) and (2) were used to estimate the water and reverse salt flux values, respectively. The water flux for the pristine and modified TFC membranes are presented in Fig. 6A at various osmotic pressures (concentrations) of the draw solutions. It was observed that the modified TFC membrane has a higher water flux than the pristine TFC membrane at all the concentrations of the draw solutions studied. The water flux was increased by about 27%, 37% and 58% for the modified TFC membrane at DS concentrations of 0.5, 1 and 2 M NaCl, respectively. The high increase in the rate of water flux was related to the smallest value of ICP in the modified support layer, which will be discussed later. Likewise, the excellent water flux of the modified TFC membrane was attributed to the increase in the hydrophilicity of the support layer of the treated membrane as well as the decrease in the structural parameter. Alternatively, the reverse salt flux (Fig. 6B) observed for the modified TFC membrane was very small when compared to the pristine sample; this was due to the low salt permeability of the modified membrane, which was related to the high rejection layer. On the other hand, the reverse salt flux of both membranes was in an acceptable range.

Fig. 6 Water flux (A) and reverse salt flux (B) of the pristine TFC–PVDF membrane and modified TFC–PVDF–TEA membrane.

Moreover, the reverse flux selectivity (J_w/J_s), which depends on both the water and salt permeability of the rejection layer, whilst independent of the draw solution concentration,⁴⁷ was determined and the results presented in Fig. 7. It was notable that the modified membrane showed an excellent reverse flux selectivity wherein the average was 34 L g⁻¹ compared to 5 L g⁻¹ found for the pristine sample.

Fig. 7 Reverse flux selectivity of the pristine and modified TFC membranes at different draw solution osmotic pressures.

3.3 Influence of the modification on the CP

In general, a decrease in the water flux and the reverse flux selectivity is usually related to the CP. However, to understand the aforementioned results of TFC–PVDF–TEA and the cause of the excellent behavior of the modified membrane, the numerical value of the CP was calculated based on Changwon Suh and Seockheon Lee model.⁴⁸ The internal concentration polymerization (ICP) and external concentration polymerization (ECP) were estimated. Moreover, in the FO process, water passes from the FS side to the DS side through the active layer and simultaneously the salt transfer from the bulk (C_Db) of the DS to the surface of the support layer (C_Dm). Then, it passes through the support layer porous until the interface of the active layer (C_i) to mix with that water passed through the active layer (Scheme 3, left).

Scheme 3 The profile of draw solute concentration in the TFC-FO membrane (left) and the profile of the TFC–PVDF and TFC–PVDF–TEA membranes with the numerical osmotic pressure profile with a 2 M draw solution and DW as the feed solution (right).

Herein, C_i, C_Dm and C_Fm were estimated using eqn (S1)–(S3) in the ESI† based on the Suh and Seockheon report.⁴⁸ Sequentially, the drop in the osmotic pressure difference was calculated based on the difference in C_i and C_Dm divided by the theoretical osmotic pressure difference to estimate ICP and the results are presented in Fig. 8.

Fig. 8 (A) The drop in the osmotic pressure difference due to ICP and (B) the theoretical, experimental and calculated (based on the ICP) water flux.

From the results in Fig. 8A, it was observed that the difference in the osmotic pressure decreased by 25.5% for the TFC–PVDF membrane at 0.5 M of the DS and the percentage of the drop increases by increasing the draw solution concentration wherein it reaches to 50% at 2 M DS. In contrast, TFC–PVDF–TEA shows a smaller drop in the osmotic pressure difference than the pristine sample. Moreover, an increase in the draw solution concentration for TFC–PVDF–TEA also increases the ​drop percentage of the pressure ​difference; numerically decreased by 11.7% and 30.6% for the 0.5 M and 2 M draw solutions, respectively.

However, the ICP has the significant effect of the diminution of the pressure difference, whereas the external concentration polarization (ECP) has a negligible value in this case because the feed was distilled water. Therefore, no concentrative ECP and a slight dilutive ECP (0.7% and 1.1% for PVDF and PVDF–TEA, respectively) was carried out due to the low cross velocity.

The osmotic pressure profile is plotted in Scheme 3 (right) and used to determine the ICP effect on the osmotic pressure. The osmotic pressure of 2 M NaCl DS decreased by 51.2% at the interface; numerically from 99.11 to 48.4 bar for the TFC–PVDF membrane, while the drop was 31.7% for the TFC–PVDF–TEA membrane.

The theoretical water flux, based on the product of the water permeability, and the difference in the theoretical osmotic pressure, were calculated and compared with the experimental results as well as the results computed based on the concentration polarization CP (using the difference between C_i and C_Fm), as shown in Fig. 8B. The results of the experimental and those calculated based on the CP were almost identical and less than the theoretical results. This demonstrates the influence of the CP on the decrease of the water flux and the performance of the FO membrane. Furthermore, a drop from 24% to 54% (for 0.5 M and 2 M, respectively) in the water flux was observed for TFC–PVDF. However, the TFC–PVDF–TEA membrane shows less water flux drop than the pristine sample numerically from 5% to 30% at 0.5 M and 2 M, respectively; this was related to the decrease in the ICP due to the smallest structural parameter. Overall, the decrease in the structural parameter has a strong effect on the ICP as well as the water flux wherein the increase or decrease in the structural parameter shows an increase or decrease in the actual path through the membrane support layer (tortuosity), which has a negative or positive effect, respectively.

Moreover, an increase in the hydrophilicity of the support layer enhanced the passing of water. The hydrophilic support layer allows the draw solution to wet its surface easily and this leads to passing the draw solution through the support layer, making an almost real osmotic pressure at the inner surface of the polyamide layer, which correspondingly enhances the passing of water from the feed to the draw solution.

Overall, both the TFC–PVDF and TFC–PVDF–TEA have a low reverse salt flux value, which points out that electrospinning is a very successful process to prepare the FO membrane. Moreover, the modified TFC membrane has the highest water flux value and reveals an acceptable range of reverse salt flux. Therefore, the two membranes exhibited an excellent salt rejection (>99%) as calculated based on eqn (3); this was related to the polyamide layer on the membrane surface wherein it is known that the polyamide layer has an excellent salt rejection.

Based on the obtained results, it can be concluded that the PVDF–TEA support layer is optimal to synthesize a TFC-FO electrospun membrane with promising performance, including a high water flux, excellent salt rejection, and low reverse salt flux.

Likewise, in order to evaluate the introduced membrane properly, a comparison with other reported FO membranes is presented in Table 2. From the table, the modified TFC–TEA membrane achieved the highest water flux as well as the lowest (J_s/J_v) ratio when compared to the other membranes. Overall, the introduced study reports a new technique that can be used to enhance FO membrane performance as well as opening new opportunities for using the electrospinning technique for the preparation of effective electrospun membranes used in FO desalination technology.

Table 2 Performance of the FO membrane using a 2 M NaCl draw solution and DI water as the feed solution

Membrane	Water flux Jv (LMH), FO	Reverse salt flux Js (gMH)	Js/Jv (g L^−1)	Ref.
TFC32	25.4	58	∼2.3	49
CAB_M	12	5.1	0.42	50
TFC-FO on sPPSU-5 (5 mol% sDCDPS)	62.8	14.9–35	0.23–0.56	51
GOT-0.25 TFC-FO	∼36	∼8	0.22	22
CN/rGO-M-0.5	41.4	∼9.6	0.23	52
TFC–PVDF	43	9.3	0.22	This work
TFC–PVDF–TEA	68	2	0.03

---

In summary, a novel treatment technique for PVDF via a chemical modification method using triethylamine has been successfully introduced in this work. Moreover, the results of the modified TFC-FO membrane reveal a very high water flux with excellent salt rejection and very low reverse salt flux, which results in the lowest structure parameter reported to date, high water permeability, low salt permeability as well as low ICP.

Acknowledgements

M. Obaid thanks the Egyptian government for the financial support of his PhD studies including this research work. The authors extend their appreciation to International Scientific Collaboration Program ISCP at King Saud University for funding this research work through ISCP-002.

References

1. B. Mi and M. Elimelech, J. Membr. Sci., 2010, 348, 337–345.
2. T.-S. Chung, L. Luo, C. F. Wan, Y. Cui and G. Amy, Sep. Purif. Technol., 2015, 156(2), 856–860.
3. T.-S. Chung, X. Li, R. C. Ong, Q. Ge, H. Wang and G. Han, Curr. Opin. Chem. Eng., 2012, 1, 246–257.
4. Q. She, X. Jin and C. Y. Tang, J. Membr. Sci., 2012, 401–402, 262–273.
5. E. M. Garcia-Castello and J. R. McCutcheon, J. Membr. Sci., 2011, 372, 97–101.
6. J. Su, Q. Yang, J. F. Teo and T.-S. Chung, J. Membr. Sci., 2010, 355, 36–44.
7. M. Shibuya, M. Yasukawa, T. Takahashi, T. Miyoshi, M. Higa and H. Matsuyama, Desalination, 2015, 362, 34–42.
8. N. Y. Yip, A. Tiraferri, W. A. Phillip, J. D. Schiffman and M. Elimelech, Environ. Sci. Technol., 2010, 44, 3812–3818.
9. X. Lu, L. H. Arias Chavez, S. Romero-Vargas Castrillón, J. Ma and M. Elimelech, Environ. Sci. Technol., 2015, 49, 1436–1444.
10. P. Lu, S. Liang, T. Zhou, X. Mei, Y. Zhang, C. Zhang, A. Umar and Q. Wang, RSC Adv., 2016, 6, 56599–56609.
11. J. Kucera, Reverse Osmosis: Industrial applications and processes, Wiley, 2011.
12. A. Tiraferri, N. Y. Yip, W. A. Phillip, J. D. Schiffman and M. Elimelech, J. Membr. Sci., 2011, 367, 340–352.
13. N. Widjojo, T.-S. Chung, M. Weber, C. Maletzko and V. Warzelhan, J. Membr. Sci., 2011, 383, 214–223.
14. D. Emadzadeh, W. J. Lau, M. Rahbari-Sisakht, H. Ilbeygi, D. Rana, T. Matsuura and A. F. Ismail, Chem. Eng. J., 2015, 281, 243–251.
15. M. L. Lind, A. K. Ghosh, A. Jawor, X. Huang, W. Hou, Y. Yang and E. M. Hoek, Langmuir, 2009, 25, 10139–10145.
16. N. Ma, J. Wei, R. Liao and C. Y. Tang, J. Membr. Sci., 2012, 405–406, 149–157.
17. M. Ghanbari, D. Emadzadeh, W. J. Lau, S. O. Lai, T. Matsuura and A. F. Ismail, Desalination, 2015, 358, 33–41.
18. M. Amini, A. Rahimpour and M. Jahanshahi, Desalin. Water Treat., 2015, 1–11.
19. M. Obaid, Z. K. Ghouri, O. A. Fadali, K. A. Khalil, A. A. Almajid and N. A. Barakat, ACS Appl. Mater. Interfaces, 2016, 8, 4561–4574.
20. P. Sukitpaneenit and T.-S. Chung, Environ. Sci. Technol., 2012, 46, 7358–7365.
21. P. H. H. Duong, S. Chisca, P.-Y. Hong, H. Cheng, S. P. Nunes and T.-S. Chung, ACS Appl. Mater. Interfaces, 2015, 7, 3960–3973.
22. M. J. Park, S. Phuntsho, T. He, G. M. Nisola, L. D. Tijing, X.-M. Li, G. Chen, W.-J. Chung and H. K. Shon, J. Membr. Sci., 2015, 493, 496–507.
23. R. C. Ong, T.-S. Chung, J. S. de Wit and B. J. Helmer, J. Membr. Sci., 2015, 473, 63–71.
24. M. Yasukawa, S. Mishima, M. Shibuya, D. Saeki, T. Takahashi, T. Miyoshi and H. Matsuyama, J. Membr. Sci., 2015, 487, 51–59.
25. N.-N. Bui and J. R. McCutcheon, Environ. Sci. Technol., 2013, 47, 1761–1769.
26. L. Huang, J. T. Arena and J. R. McCutcheon, J. Membr. Sci., 2016, 499, 352–360.
27. M. Baghbanzadeh, D. Rana, C. Q. Lan and T. Matsuura, Sep. Purif. Technol., 2016, 157, 60–71.
28. X. Song, Z. Liu and D. D. Sun, Adv. Mater., 2011, 23, 3256–3260.
29. G. Han, T.-S. Chung, M. Toriida and S. Tamai, J. Membr. Sci., 2012, 423, 543–555.
30. P. Duong, S. Chisca, P. Hong, H. Cheng, S. P. Nunes and T.-S. Chung, ACS Appl. Mater. Interfaces, 2015, 7, 3960–3973.
31. A. Tiraferri, N. Y. Yip, A. P. Straub, S. R.-V. Castrillon and M. Elimelech, J. Membr. Sci., 2013, 444, 523–538.
32. V. Parida and H. Y. Ng, Desalination, 2013, 312, 88–98.
33. G. Socrates, Infrared and Raman characteristic group frequencies: table and charts, 2001.
34. M. A. Muñoz, C. Carmona and M. Balón, Chem. Phys., 2007, 335, 37–42.
35. N. Shi, J. Duan, J. Su, F. Huang, W. Xue, C. Zheng, Y. Qian, S. Chen, L. Xie and W. Huang, J. Appl. Polym. Sci., 2013, 128, 1004–1010.
36. M. Sharma, G. Madras and S. Bose, Cryst. Growth Des., 2015, 15, 3345–3355.
37. C. Ribeiro, V. Sencadas, J. L. G. Ribelles and S. Lanceros-Méndez, Soft Mater., 2010, 8, 274–287.
38. S. Lanceros-Mendez, J. Mano, A. Costa and V. Schmidt, J. Macromol. Sci., Part B: Phys., 2001, 40, 517–527.
39. T. G. Mofokeng, A. S. Luyt, V. P. Pavlović, V. B. Pavlović, D. Dudić, B. Vlahović and V. Djoković, J. Appl. Phys., 2014, 115, 084109.
40. P. Nallasamy and S. Mohan, Indian J. Pure Appl. Phys., 2005, 43, 821–827.
41. C. Y. Tang, Y.-N. Kwon and J. O. Leckie, Desalination, 2009, 242, 149–167.
42. Y.-N. Kwon and J. O. Leckie, J. Membr. Sci., 2006, 282, 456–464.
43. S. Wu, G. Zheng, H. Lian, J. Xing and L. Shen, J. Appl. Polym. Sci., 1996, 61, 415–420.
44. D. Thomas, J. Appl. Polym. Sci., 1964, 8, 1415–1427.
45. W. Zhang, Z. Shi, F. Zhang, X. Liu, J. Jin and L. Jiang, Adv. Mater., 2013, 25, 2071–2076.
46. A. Taguet, B. Ameduri and B. Boutevin, in Crosslinking in Materials Science, Springer, 2005, pp. 127–211.
47. W. A. Phillip, J. S. Yong and M. Elimelech, Environ. Sci. Technol., 2010, 44, 5170–5176.
48. C. Suh and S. Lee, J. Membr. Sci., 2013, 427, 365–374.
49. Z. Liu, H. Yu, G. Kang, X. Jie, Y. Jin and Y. Cao, J. Membr. Sci., 2016, 497, 485–493.
50. R. C. Ong, T.-S. Chung, B. J. Helmer and J. S. de Wit, Ind. Eng. Chem. Res., 2012, 51, 16135–16145.
51. N. Widjojo, T.-S. Chung, M. Weber, C. Maletzko and V. Warzelhan, Chem. Eng. J., 2013, 220, 15–23.
52. Y. Wang, R. Ou, H. Wang and T. Xu, J. Membr. Sci., 2015, 475, 281–289.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra18153h

---