Enhancing Dehydration Performance of Isopropanol by Introducing Intermediate Layer into Sodium Alginate Nanofibrous Composite Pervaporation Membrane

  • Peiyun Li
  • Cheng Cheng
  • Ke Shen
  • Tonghui Zhang
  • Xuefen WangEmail author
  • Benjamin S. Hsiao
Research Article


A novel three-tier composite membrane based on highly porous nanofibrous substrate was demonstrated for efficient isopropanol dehydration by pervaporation. Here, polyethyleneimine (PEI) modified graphene oxide (GO) sheets were vacuum-assistant assembled onto porous electrospun polyacrylonitrile (PAN) nanofibrous substrate to achieve a smooth, hydrophilic and compact PEI-GO intermediate layer. The introduction of PEI chains endowed GO interlayer with sufficient interaction for bonding adjacent GO nanosheets to enhance stability in water/isopropanol mixture and also with the ascended interlamellar space to improve the water-sorption ability due to the abundant active amino groups. Benefiting from PEI-GO layer, a defect-free sodium alginate (SA) skin layer could be facilely manufactured with elaborately controlled thickness as thin as possible in order to reduce mass transfer resistant and enhance permeability maximally. Meanwhile, the interlayer would also contribute to enhance interfacial adhesion to promote the structure integrity of three-tier thin-film nanofibrous composite (TFNC) membrane in pervaporation dehydration process. After fine-tuning of membrane preparation process, the SA/PEI(75)-GO-60/PAN TFNC membrane exhibited competitive pervaporation performance with the permeate flux of 2009 g/m2 h and the separation factor of 1276 operated at 70 °C for dehydration of 90 wt% isopropanol solution. The unique three-tier composite membrane structure suggested an effective and facile approach to design novel membrane structure for further improvement of pervaporation performance.

Graphic Abstract


Polyethyleneimine modified graphene oxide Intermediate layer Thin-film nanofibrous composite membrane Pervaporation membrane 


Pervaporation (PV), as a highly-efficient and newly-developing membrane process for dehydration of various organic mixture, has aroused extensive attention of researchers. Compared with traditional distillation separation process, pervaporation has demonstrated advantages of relatively simple and convenient operation, environmental benignity and low energy consumption, especially in separating azeotropic mixtures and purifying heat-sensitive hazardous compounds [1, 2, 3]. Recently, many regulation strategies were attempted to enhance the PV membrane selectivity and promote permeation flux at the same time, such as exploiting new membrane materials [4], designing suitable membrane structures [5] and introducing novel nanofillers into polymer matrix [6, 7]. Compared with mixed matrix membranes, PV composite membranes commonly consist of a porous substrate and an ultra-thin selective layer. By selecting suitable membrane materials and optimizing preparation technology of substrate and skin layer separately, the PV separation performance of composite membrane can be further improved. Electrospun nanofibrous membrane has become an ideal candidate as the alternative porous substrate of composite membrane in membrane separation fields such as ultrafiltration, nanofiltration and pervaporation [8, 9, 10]. The unique structure properties of electrospun mat such as high porosity, excellent permeability and fully interconnected pore structure could distinctly reduce mass transfer resistance for composite membrane [11]. However, the fluctuant surface and micro-scale interstitial pores of nonwoven nanofibrous scaffold brought new challenges on preparing a dense, ultrathin and defect-free selective layer above the porous substrate because the coating solution was inclined to penetrate into the substrate’s three-dimensional channels, and nonselective defects would be generated out of control. To overcome the disadvantage, some creative methods of preparing nanofibrous composite membrane have been reported in literatures, such as pre-crosslinking casting solution before coating [12], taking substrates presoaked in crosslinking agent [13], elecrospraying combined with hot-pressing post-treatment [14] and incorporating the spin-coating technique into interfacial polymerization [15].

Sodium alginate (SA), as the widely used hydrophilic material for pervaporation membrane, was preferential water sorption and diffusion through the membrane on account of the abundant carboxyl and hydroxyl groups, and it could maintain the stable gel network in water/alcohol mixture after crosslinking by divalent cations [16, 17]. Theoretically, according to solution-diffusion mass transport mechanism [2], an ultrathin, hydrophilic and defect-free selective layer with low water transfer resistance was required for composite PV membrane in dehydration process. However, it was difficult to achieve an intact SA layer on the electrospun nanofibrous substrate due to its high permeability and sub-micro pore structure by conventional coating method in practice. Jiang et al. have ever reported a mussel-inspired intermediate layer by co-deposition dopamine and poly(ethyleneimine)(PEI) on porous support layer [18]. The hydrophilic biomimetic adhesive layer introduced multiple interfacial interaction between SA top layer and PAN support layer which remarkably enhanced swelling resistance and separation performance of the PV membrane. Besides, a lot of hydrophilic materials have been applied to fabricate the interlayer for adjusting the pore size and surface wettability of support layer, such as carbopol (CP) [19], polyamide [20] and SiO2 sol–gel [21]. The existence of intermediate layer avoided the collapse of membrane structure and enhanced the stability of composite membrane in organic solvent/water mixture. Hence, introducing a smooth and hydrophilic intermediate layer would be a facile way to effectively manipulate membrane structure endowed PV membranes with high selectivity and great stability simultaneously.

As an ideal membrane material, graphene oxide (GO) has tripped research wave due to its unique two-dimension lamellar structure, amazing flexibility and controllable surface chemistry [22, 23]. In addition, GO lamellae were easy to assemble on porous substrate by vacuum-assisted filtration to obtain an ultrathin hydrophilic layer with highly-permeable inner nanochannels [24, 25]. More interestingly, Nair et al. demonstrated that GO membrane had a particular mass transferring characteristic for water molecules, namely, GO membranes would provide rapidly permeating channels for water vapor while completely impede the permeation of other organic solvent liquids and vapor [26]. Hence, the existence of unique laminar nanochannels endowed GO membrane broad application prospect in pervaporation dehydration. Based on our previous experience, the merely GO layer on nanofibrous substrate could not be qualified for pervaporation dehydration [27]. Here, it was expected to combine the advantages from GO, SA and nanofibrous substrate for manufacturing a high performance three-tier composite membrane with SA selective skin layer and GO intermediate layer. Unfortunately, strong electrostatic repulsive force was existed between GO membrane and SA layer because GO sheets and SA had the same negative charges. SA coating solution cannot uniformly spread onto GO interlayer to form a defect-free selective layer. Additionally, the tight stacking of GO lamellae often results in high mass transfer resistance. In recent years, several chemicals such as aldehydes [28] and diamines [29] were introduced into the gallery between GO sheets to modify the properties of interlamellar nanochannels. Actually, the connecting bridge between GO sheets fixed by small molecular modifier was not strong enough and the stretching of d-spacing was not efficient due to the short chain of modifier and large space of adjacent GO sheets. Hydrophilic soft polymer was the more suitable modifier to adjust the property of interspacing and enhance interfacial adhesion [27].

In this work, in consideration of the surface wettability, interfacial adhesion, the size and water affinity of membrane inner nanochannels, a positive charged hydrophilic polymer polyethyleneimine (PEI) will be chosen as a modifier to fabricate modified GO lamellae for preparation of functionalized intermediate layer. PEI could easily graft onto the GO sheets via chemical bonds between amino and oxygen-containing groups to provide sufficient interlamellar reaction so that avoided the destroy of lamellae stacking structure [30]. The introduction of PEI could stretch the interlayer nanochannels and decorate a large amount of hydrophilic amino groups on the surface [31], which would notably improve the membrane water–adsorption and alcohol-repulsion ability and effectively decrease the mass transfer resistance of GO acting as an intermediate layer. As shown in Scheme 1, PEI-GO lamellae were easily assembled on the polyacrylonitrile (PAN) nanofibrous substrate by vacuum-assisted filtration and the thickness of PEI-GO layer could be well-controlled by adjusting the volume of PEI-GO dispersion. The introduction of hydrophilic intermediate layer provided multiple interactions (electrostatic interaction and hydrogen bonding) between PEI-GO sheets and SA to ensure the integrity of the thin-film nanofibrous composite (TFNC) membrane and avoided the detachment of top layer and collapse of membrane structure in pervaporation process. As a potential facile way to fabricate high-performance TFNC pervaporation membrane, the three-tier membrane structure containing intermediate layer will play an important role in pervaporation developing.
Scheme 1

The fabrication process and structures of SA/PEI-GO/PAN TFNC membrane

Experimental Section


PAN (Mw = 120,000 g/mol) was purchased directly from Shanghai Jinshan Co., Ltd. (China). Graphite (325 mesh) was kindly supplied from Shanghai Yifan Graphite Co., Ltd. (China). Sodium alginate (SA) and polyethylenimine with different molecular weights (PEI, Mw = 10,000 Da, 25,000 Da, 750,000 Da) were procured from Sigma-Aldrich (USA) and Alfa Aesar. N,N′-dimethylformamide (DMF), calcium chloride anhydrous (CaCl2, ≥ 96 wt%) and isopropanol (IPA  ≥ 99.7 wt%) and were purchased by Shanghai Lingfeng Chemical Reagent Co., Ltd. (China). All chemicals were of analytical grade and were used as received without further purification.

Preparation of Electrospun PAN Nanofibrous Substrate

According to our previous work [27], 8 wt% PAN solution was fabricated by dissolving a certain amount of PAN powders in DMF at 50 °C under mild stirring for 24 h. Then the homogeneous solution was transferred through a syringe needle. The parameters of electrospinning were controlled as follows: the flow rate was settled at 1 mL/h, the distance between the spinneret and the rotating drum collector was 15 cm and the applied voltage was 20 kV. The freshly obtained electrospun PAN mats were dried in the vacuum oven at 60 °C overnight. In order to obtain a smooth surface for next-step fabrication and improve the mechanical strength of substrates, the PAN nanofibrous scaffolds were cut into sizes of 7 cm × 7 cm and then cold-pressed under 10 MPa for 30 s at ambient temperatures.

Preparation of TFNC Membrane

The synthesis of PEI/GO Dispersion

Graphene oxide was prepared by chemical exfoliation according to the modified Hummers method reported in previous literature [32], The resulting GO dispersion was purified by dialysis and then diluted to 0.5 mg/mL for use. Because PEI chains and GO sheets possessed opposite charges in aqueous dispersion, adjacent GO sheets were easily connected by positive PEI molecules via strong interactions (e.g. electrostatic attraction and chemical bonding). Therewith a lot of modified GO sheets were stacked and aggregated, PEI-GO flocculation were generated. To obtain the homogenous PEI/GO dispersion without any flocculation, the mass ratio of PEI and GO was fixed at 20:1 [33]. The excess PEI molecules maintained positive zeta potential of dispersion and weakened the influence of adding negative charge to keep the dispersion stable and uniform. When tiny GO was added into PEI solution, it was rapidly dispersed and had little impact on dispersion stability. Hence, 10 mL GO dispersion (0.5 mg/mL) was dropwise added into 100 mL PEI solution (1 mg/mL) of different molecular weights (Mw = 750,000, 25,000, 10,000 Da) under vigorously sustained stirring. After dropwise adding the yellow–brown GO dispersion into PEI solution, the mixture became dark-brown. And then the mixture was continuous stirring for 12 h at room temperature for sufficient reaction. The modified GO dispersion and were denoted as PEI(X)-GO, where X represented the molecular weight of PEI (× 10,000 Da).

Preparation of SA/PEI-GO/PAN TFNC Membranes

The as-prepared PEI/GO dispersion was vacuum filtrated through the electrospun PAN nanofibrous substrate to form a smooth and hydrophilic layer. The amount of modified GO deposited on the PAN nanofibrous support layer varied relying on the volume of the PEI/GO dispersion. Excessive unreacted PEI was isolated through the nanofibrous mat into the filtrate solution during the process. The PEI-GO intermediate layer was prepared by air-drying and stored at room temperature for later use. Then, 2 wt% SA aqueous solution was cast onto the PEI-GO intermediate layer using a 30 μm casting knife to form a defect-free hydrogel layer and then the as-prepared membrane was immersed into a CaCl2 crosslinking bath and dried in the atmosphere before testing. The as-prepared PEI-GO membranes and resulting three-tier TFNC membranes were denoted as PEI(X)-GO-Y/PAN and SA/PEI(X)-GO-Y/PAN, where X represented the molecular weight of PEI (× 10,000 Da) and Y represented the deposition amount of GO (mg/m2).


Atomic force microscopy (AFM, Digital Instruments Multi-Mode, DI-NS3a, USA) were utilized to characterize graphene oxide. The surface and cross-section morphology and microstructure of the composite membrane were characterized by field emission scanning electron microscopy (FE-SEM, S-4800, Hitachi, Japan). The surface composition of the composite membrane was investigated by Nicolet 8700 FT-IR spectrometer (attenuated total reflectance mode, Thermometer, USA). Raman spectra of GO/PAN and PEI-GO/PAN composite membrane were obtained on a Nanofinder 30 (Tokyo instrument Co., Osaka, Japan) to confirm the structural changes. The surface wettability was performed by a dynamic contact angle testing instrument (OCA40, Dataphysics, Filderstadt, Germany) equipped with a dynamic image capture camera.

Pervaporation Separation Experiment

The pervaporation dehydration performance of the membranes was evaluated by utilizing a 90 wt% isopropanol solution as a model system. Pervaporation testing was performed on a lab-made experimental setup as shown in Fig. 1. A glass container of 3 L was selected as a feed tank for the IPA/water feed solution. The feed tank should be equipped with heating system which could control feed solution at different temperatures when required and effectively prevent heat loss and evaporation at high temperature. The down-stream pressure of the membrane cell was maintained below 0.1 kPa using a vacuum pump and the preheated feed solution was continuously circulated at a flow rate of 30 L/h with the help of a diaphragm laboratory pump. Before pervaporation testing, the system should be cycled and stabilized for 2 h. A cold trap immersed in liquid nitrogen was settled at down-stream to collect permeation. The tested membrane with an effective membrane area of 14 cm2 was installed in the measuring cell. Each permeation sample was collected per hour, weighted and then measured the composition by gas chromatography (GC, Chromatography 112A, China). Every permeate sample should be collected at least three times and their average was calculated to present the pervaporation performance. The permeation flux (J, g/m2h), the separation factor of water/isopropanol (α) and pervaporation separation index (PSI) were calculated from the following equations [34]:
Fig. 1

Schematic diagram of the pervaporation setup: 1—temperature controlling system, 2—feed solution, 3—flow meter, 4—membrane cell, 5—membrane, 6—porous support plate, 7—diaphragm laboratory pump, 8—vacuum gauge, 9—liquid nitrogen trap, 10—buffer tank and 11—vacuum pump

$$\it {\text{J}} = \frac{Q}{A \times t}$$
$${{\alpha = }}\frac{{{{Y}_{{\rm w}}} /{{Y}_{{\rm iso}}}}}{{{{X}_{{\rm w}}} /{{X}_{{\rm iso}}}}}$$
$${{{\rm PSI} = J\; \times (\alpha \; - 1)}}$$

where Q is the mass of collected permeation during the sampling time t and A is the effective membrane area in membrane cell. XW and Xiso are the respective weight fractions of water and isopropanol on the feed side. YW and Yiso are the respective weight fractions of water and isopropanol on the permeate side.

The relationship of feed temperature and partial permeation flux (Ji, g/m2h) of each component is determined by the Arrhenius equation as following:
$${{\text{J}}_{\text{i}}}{\text{ = A exp}}\left( {{ - }\frac{{{{\text{E}}_{\text{p,i}}}}}{\text{RT}}} \right)$$

where A is the pre-exponential factor; R (kJ mol−1 K−1) is the gas constant; T (K) is the absolute temperature of feed solution and Ep,i (kJ mol−1) is the apparent activation energy for permeation, which can be calculated by plotting ln (Ji) vs 1000/T.

Results and Discussion

Optimization of The Modification Towards GO Interlayer

The intermediate layer was utilized to prevent the penetration of coating solution and enhance the interfacial adhesion between the support and separation layer. Therefore, the intermediate layer should provide good adhesion with substrates and possess a lot of reactive sites to offer a functional and hydrophilic flat for next-step SA layer fabrication. PEI was a hydrophilic polymer with abundant amino groups, which would easily interact with the oxygen containing groups of GO via covalent bonds and electrostatic interactions to enlarge the interlamellar spacing and provide abundant reactive sites for SA chains [7]. The successful modification of GO by PEI were verified by Fourier transform infrared (FTIR). Figure 2a showed the FTIR spectra of the GO/PAN and three PEI-GO/PAN membranes under the transmission mode to analyze the chemical structures of the intermediate layer. The GO/PAN membrane had the following characteristic peaks corresponding to oxygen functional groups of GO sheets: –OH stretching of hydroxyl groups at 3405 cm−1, –C=O stretching of carboxyl group at 1732 cm−1, C–C stretching of aromatic at 1634 cm−1, and C–O–C stretching of alkoxy and epoxy group at 1080 cm−1 and 1250 cm−1 [35]. As to three PEI-GO/PAN membranes, the peak intensity of carboxyl group at 1732 cm−1 was decreased greatly due to the partial reduction of GO. And the peak of C–O epoxy group at 1250 cm−1 disappeared caused by the reaction of amino group of PEI and the epoxy groups of GO. Meanwhile, three new peaks appeared at 1650 cm−1, 1567 cm−1, and 1467 cm−1, which representatives of the carbonyl group in amide group, N–Hx groups in PEI chains and new C–N bonds, respectively [30, 33]. In addition, for PEI-GO, the peak at 3405 cm−1 became broader and shifted to lower wave numbers (at around 3273 cm−1, red shift), attributing to hydrogen bonding interactions between the N–H bonds of PEI and –OH groups of GO sheets [36]. These results demonstrated that PEI chains were successfully incorporated into GO sheets and fixed on the surface of PEI-GO intermediate layer.
Fig. 2


FT-IR spectra and b Raman spectra of GO, PEI(1)-GO, PEI(2.5)-GO and PEI(75)-GO

Each individual PEI (Mw = 750,000, 25,000, 10,000 Da) altered GO sheets at different degrees and had various influence on the changes in d-spacing. Raman spectroscopy was selected to further investigate the extent of interaction between GO sheets and with different molecular weights. As shown in Fig. 2a, with the increase of PEI molecular weight, the peak intensity of C–N bond in PEI-GO at 1467 cm−1 was slightly increased but the intensity of carbonyl group at 1650 cm−1 was decreased. Correspondingly, in Fig. 2b, Raman spectrum of GO/PAN showed two prominent peaks located at 1351 and 1593 cm−1 which was equivalent to D and G bands, respectively. The D band represented the structural disorder and defects associated with the functional groups on GO sheets, while the G band was ascribed to the first order scattering of the E2g phonon of the sp2 carbon–carbon bond. After modification by PEI, the PEI-GO/PAN spectra exhibited two peaks at the same location of GO, but interestingly, the intensity ratio of the D and G band (ID/IG) increased from 1.25 for GO to 1.60 for PEI(75)-GO. Moreover, with the augmented molecular weight of PEI, the ID/IG ratio increased [1.29 for PEI(1)-GO, 1.34 for PEI(2.5)-GO and 1.60 for PEI(75)-GO] [37, 38]. During the reaction process, the –NH2 moieties of PEI were nucleophilic added to the epoxy groups of GO, which caused the decrease in average size of domains and the increase the number of aromatic domains with smaller average size. The increase of C–N bond peak intensity and ID/IG ratio demonstrated that more amino groups on higher molecular weight PEI chains reacted with oxygen-containing groups of GO sheets. High molecular weight PEI would facilitate the reaction with SA molecules at the interfacial zone. Meanwhile, the introduction of PEI provided GO interlayer spacing with more hydrophilic groups which let these nanochannels show good water affinity and alcohol repulsion.

To directly investigate of the different enlargement of interlamellar space among three intermediate layers, a series of PEI-GO/PAN membranes were fabricated by depositing 100 mg/m2 PEI-GO on PAN nanofibrous membrane. The SEM images of surface and cross-section morphologies of PEI-GO-100/PAN TFNC membranes with different PEI molecular weight were shown in Fig. 3. In Fig. 3a–c, the nanofibrous substrates were fully covered by PEI-GO layers which exhibited peak-to-valley morphology, the surface became smoother and the contour of underlying nanofibrous mat turned just slight prominent with the increase of PEI molecular weight. Correspondingly, the average thickness of PEI-GO layer was increased from 55 nm of PEI(1)-GO to 170 nm of PEI(75)-GO layer from the cross-section images of composite membranes in Fig. 3d–f. The increase of membrane thickness was attribute to the changing of PEI molecular steric dimensions. The PEI(75)-GO layer had the maximum d-spacing enlargement of GO sheets after modification, the multiple interaction between high molecular PEI chains and GO sheets guaranteed the stability of intermediate layer in practical application. Moreover, there was no obvious boundary between PEI(75)-GO intermediate layer and substrate because the higher molecular weight PEI provided better compatibility with long polymer chains of PAN for great attachment onto the PAN nanofibrous substrate. Additionally, the flexible lamellar PEI-GO nanosheets could be sucked easily and adhered tightly onto the surface of the nanofibrous substrate due to its microporous structure and fluctuant morphology by the vacuum suction during the initial vacuum filtration process.
Fig. 3

Surface and cross-section morphologies of PEI(1)-GO-100/PAN (a, d), PEI(2.5)-GO-100/PAN (b, e) and PEI(75)-GO-100/PAN (c, f) TFNC membrane

Among three PEI-GO intermediate layers, PEI(1)-GO layer possessed the most fluctuated surface which would be hard to prepare an even SA hydrogel layer. The minute amount of lowest molecular weight PEI introducing into gallery between GO sheets led to weak interaction between SA layer and the intermediate layer which cannot effectively decrease the interfacial stress and would induce the generation of detachments. Compared with PEI(2.5), high Mw PEI(75) had the larger molecular steric dimensions and better compatibility with SA chains. Also the intercalation of PEI(75) would stretch the interlamellar spacing maximally and reduce the mass transfer resistance effectively. In addition, the PEI(75) chains fixed on the surface of PEI(75)-GO layer would be benefit for the film-forming of SA solution enhancing the interfacial adhesion to ensure the integrity of the three-tier membrane. Therefore, SA/PEI(75)-GO/PAN should reveal the highest permeate flux and best separation factor. To attest the deduction, the tentatively investigation of the pervaporation performance for SA/PEI-GO/PAN composite membrane were performed. Three membranes with different PEI-GO intermediate layer were fabricated and the dehydration performance was evaluated by home-made experiment module in separating 90 wt% aqueous isopropanol solution at 60 °C. As shown in Fig. 4, SA/PEI(1)-GO/PAN almost failed to separate isopropanol and water, which showed much low separation factor. In addition, the whole SA layer fell off the intermediate layer after testing. It confirmed that PEI(1) cannot effectively connect two layer and maintain the integrity of three-tier composite membrane. Compared with SA/PEI(2.5)-GO/PAN membrane, SA/PEI(75)-GO/PAN membrane showed higher permeation flux and better separation ability. PEI(75) endowed nanochannels with more hydrophilic groups, which would facilitate the water sorption into the interlamellar space and accelerate the water transport rate through membrane. The nonselective defects and interfacial voids were effectively sealed by SA layer adhered firmly onto the intermediate layer, resulting in high separation factor and efficiency. Consequently, 750,000 Da PEI was chosen as the optimal modification agent for GO to prepare the PEI-GO intermediate layer.
Fig. 4

Pervaporation performance of SA/PEI-GO-100/PAN nanofibrous composite membranes varied with different molecular weight of PEI in separating 90 wt% aqueous isopropanol solution at 60 °C

Optimization of the Deposition Amount of PEI-GO on PAN Nanofibrous Substrate

As for thin-film composite membrane, the property of selective layer played dominant role in pervaporation separation process. Manipulating the microstructure of selective layer would be the effective way to achieve the desired separation permeability and selectivity. More importantly, the selective layer should be fabricated as thin as possible because a thinner selective layer has a higher water permeability and a lower hydraulic resistance. With the assistant of intermediate layer, it would be easier and controllable to prepare an ultrathin and defect-free skin layer on nanofibrous substrate has become easier and controllable. In consideration of this dominated role, as building blocks of the layer, the deposition amount of PEI-GO should be firstly optimized.

Figures 5 and 6 showed the surface and cross-section morphologies of PEI(75)-GO/PAN composite membrane with different deposition amount of PEI(75)-GO, respectively. In Fig. 5a, PAN nanofibrous substrate showed scaffold-like fibers network with uniform fiber diameter and interconnected pores. Figure 5b–f showed the typical surface morphologies of two-tier composite membranes. The PAN scaffold was gradually covered with the increase of deposition amount of PEI(75)-GO and the profile of undulated nanofibers slowly became indistinct. At the deposition loading of 40 mg/m2, the GO layer was uneven and some pits were generated above the pores of PAN substrate. An intact well-packed intermediate layer was formed when the loading amount reached 60 mg/m2. The cross-section morphology and thickness of the intact PEI(75)-GO layer were investigated from the cross-section FESEM images as shown in Fig. 6. With the deposition amount increased from 40 to 100 mg/m2, the relation between average thickness and deposition amount of PEI(75)-GO sheets was nearly fitted linear, while the mean thickness was from 65 nm at 40 mg/m2 increased to 170 nm at 100 mg/m2. There was no PEI(75)-GO sheet and aggregates infusing into the substrate and on the surface caused by the vacuum-assisted filtration preparation method. Although the size of some PEI(75)-GO sheets were smaller than the pores (Fig. S1), the sheets preferred to depositing on the PAN nanofibrous scaffold because the filtration-assisted assembly method would let the larger PEI(75)-GO sheets block the substrate pores immediately [27]. So, the thickness of intermediate layer was well-controlled by adjusting the volumes of filtered PEI(75)-GO suspension. The intact, ultrathin and hydrophilic intermediate layer was important for fabricating an integrated SA top layer because it was conductive to increasing the interfacial compatibility and decreasing the interfacial stress between support and top layer [18]. Considering the membrane uniformity, stability and mass transfer resistance of intermediate layer, 60 mg/m2 was selected as the optimized deposition amount of PEI(75)-GO for further fabrication of SA skin layer.
Fig. 5

Surface morphologies of PAN nanofibrous substrate with different deposition amount of modified GO: a 0 mg/m2, b 20 mg/m2, c 40 mg/m2, d 60 mg/m2, e 80 mg/m2 and f 100 mg/m2

Fig. 6

Cross-section morphologies of PEI(75)-GO/PAN nanofibrous substrate with different deposition amount of PEI(75)-GO: a 40 mg/m2, b 60 mg/m2, c 80 mg/m2, d 100 mg/m2

Fabrication and Characterization of SA/PEI-GO/PAN TFNC Membrane for Pervaporation Dehydration

Like simple GO/PAN TFNC membrane [27], PEI(75)-GO/PAN TFNC membrane also could not meet the expectant demand for pervaporation performance (Fig. S2) which has low selectivity of isopropanol/water mixture caused by the loose stacking of PEI(75)-GO sheets ascribed to much larger steric hindrance of PEI(75). For achieving high pervaporation performance, SA top layer was fabricated by casting SA aqueous solution onto the optimized PEI-GO intermediate layer and was crosslinked by calcium ions. Figure 7a was the digital photo of the as-prepared SA/PEI(75)-GO/PAN TFNC membrane. The resultant three-tier composite membrane has a very hydrophilic surface (water contact angle was 16.2 ± 1.5° from the inset image) due to the substantial carboxyl group came from SA chains, which could be partially ionized into carboxylate anion in aqueous solution. SA/PEI(75)-GO/PAN was crosslinked by Ca2+ to enhance the stability in feed solutions for long-time pervaporation operation [39]. After coating SA onto the intermediate layer, the sufficient interaction bonded SA top layer and PEI-GO interlayer tightly. The dense structure at interfacial zone and improved surface hydrophilicity increased the membrane selectivity. SA top layer played the decisive role of pervaporation separation. Figure 7b–c showed the surface and cross-section morphologies of SA/PEI(75)-GO-60/PAN TFNC membrane. In Fig. 7b, the surface SEM image demonstrated that the PEI(75)-GO layer was completely covered with a smooth and crack-free SA top layer and the undulation of the underlying nanofibrous support cannot be observed. The cross-sectional images in Fig. 7c indicated that each layer of the resulting three-tier composite membrane had uniform thickness. The thickness of whole separation layer (including SA top layer and PEI(75)-GO intermediate layer) was around 860 nm. SA barrier layer was firmly adhered on PEI(75)-GO intermediate layer while no obvious void and gap was observed within the interfacial boundary. The existence of PEI(75)-GO intermediate layer provided a functional bridge platform to avoid the SA aqueous solution penetrating into the substrate and contribute to fabricating a uniform top layer for preventing the membrane from disassembly and invalidation in pervaporation testing.
Fig. 7


A digital photo of SA/PEI(75)-GO-60/PAN TFNC pervaporation membrane. The surface (b) and cross-sectional (c) FE-SEM images of SA/PEI(75)-GO-60/PAN TFNC pervaporation membrane. And the inset showed the photograph of water contact angle on the membrane surface

Pervaporation Dehydration Performance of SA/PEI-GO/PAN TFNC Membrane

Effect of PEI-GO Deposition Amount on Pervaporation Performance

The pervaporation dehydration performances of the SA/PEI(75)-GO/PAN TFNC membranes were evaluated by permeation flux and separation factor using 90 wt% aqueous isopropanol as feed solution at 60 °C. As shown in Fig. 8, SA/PEI(75)-GO-20/PAN TFNC membrane possessed rather high permeation flux (approach 4000 g/m2 h) but lowest selectivity of these composite membranes. The phenomenon should be attributed to much lower deposition amount of PEI(75)-GO (only 20 mg/m2) which cannot cover all pores of PAN nanofibrous substrate. The cavities of PEI(75)-GO layer would tempt the SA casting solution to penetrate through the incomplete intermediate layer into PAN substrate. Therefore, many non-selective defects would be produced in SA top layer which made water and isopropanol molecules passing through the TFNC membrane without sieving separation. With the sustained increase of deposition amount (from 40 to 100 mg/m2), the permeation flux of TFNC membranes was visibly decreased, which was mainly caused by the increased mass transfer resistance on account of the thicker PEI(75)-GO layer. While the variation of selectivity was different from the trend of permeation flux. The water content in permeate was highest at the loading of 60 mg/m2 and then became slightly decreased with higher deposition amount of PEI(75)-GO. To further intuitively demonstrate the variation tendency of selectivity, separation factors were calculated from Eq. (2) which would replace water content in permeate to comprehensively evaluate the membrane selectivity. As shown in Fig. 8b, when the deposition amount of PEI(75)-GO reached 40 mg/m2, the separation factor was still relatively low. It may be attribute to the poor distribution of SA solution on the uneven PEI(75)-GO-40 layer which would facilitate the formation of local defects in SA layer even if the pores of PAN nanofibers substrate were almost covered by PEI(75)-GO sheets without any cavitation. During pervaporation process, the site of defects in SA layer would suffer much more stress and evolve into fractures due to the low pressure at the downstream of membrane. So SA/PEI(75)-GO-40/PAN TFNC membrane has rather high permeate flux but relatively poor selectivity. The best separation selectivity was occurred at loading of 60 mg/m2 and then separation factor became lower with the further increase of PEI(75)-GO loading amount. The decreased trend of selectivity may attribute to the variation of surface hydrophilicity and stability of the PEI(75)-GO layer. On one hand, with the increase of loading amount, the decrease of hydrophilicity would result in uneven spreading of SA solution on intermediate layer. On the other hand, after introducing of many PEI chains into GO sheets interlayer, the interlamellar spacing was expanded because of PEI(75) with larger molecular steric dimension [30]. Actually, when the thickness of interlayer increased too much, SA molecules were difficult to diffuse through the dense crosslinking zone at the upper section of interlayer into bottom section to achieve effective interfacial bonding. The stability of intermediate layer became poor and the sieving selectivity was decreased. Hence, SA/PEI(75)-GO-60/PAN TFNC membrane showed the optimum separation performance with a permeation flux of 1892 g/m2 h and separation factor of 1491 at 60 °C. As a comprehensively evaluated index of composite membranes, PSI values were calculated to overall characteristic the separation performance [34]. As shown in Fig. 8b, the SA/PEI(75)-GO-60/PAN TFNC membrane has the highest PSI value up to 2.82 × 106.
Fig. 8

Pervaporation performance of SA/PEI(75)-GO/PAN nanofibrous composite membranes varied with deposition amount of PEI(75)-GO in separating 90 wt% aqueous isopropanol solution at 60 °C a permeation flux and water concentration in permeate, b separation factor and PSI value

Effect of Water Content in Feed on Pervaporation Performance

After a series of optimizations, SA/PEI(75)-GO-60/PAN TFNC membrane was selected as a representative to investigate the effect of feed composition (water content from 10 to 30 wt%) on membrane pervaporation dehydration performance at 60 °C. Figure 9a demonstrated that the permeation flux increased and the separation factor decreased continuously with the augment of water content in feed. The water flux and isopropanol flux were calculated, respectively for further investigation. As shown in Fig. 9b, the total flux and water flux were synchronously increased while the isopropanol flux was increased slightly. The increased water content had the most direct effect on the enhancement of activity and partial pressure of water, which intensify the driving force of water to make water molecules transferring quickly through the selective layer. Meanwhile, the augment of water content in feed would exacerbate the swelling degree and consequent larger free volume cavities of SA polymeric chains. The enhancement of driving force and cavities facilitated water molecules permeation, which resulted in the increase of total and water flux at the same time [40]. Besides, it is interesting that the coupling effect between the water molecules and isopropanol molecules facilitated the formation of molecular clusters. The severer influence of swelling degree and free volume cavities was occurred at higher water concentration. Therefore, with the increase of water content, more clusters combining isopropanol molecular permeated to the downstream of composite membrane and these molecular clusters inhibited water molecules permeation by blocking the interaction between water and membrane material, which imparted the membrane with higher isopropanol permeability and lower diffusion selectivity at higher water content in feed [7, 41]. The variation trend of permeate flux and selectivity is caused by the synergistic effect consist of above three aspects and similar results have been reported for other SA-based membranes [40, 42]. SA/PEI(75)-GO-60/PAN TFNC membrane showed a desirable pervaporation performance at 25% water content, the total flux was up to 2980 g/m2 h while the water content in permeation maintained higher than 98.5 wt%.
Fig. 9

Effect of feed concentration on pervaporation performance of SA/PEI(75)-GO-60/PAN TFNC membrane in separating 90 wt% aqueous isopropanol solution at 60 °C: a permeate flux and separation factor; b water flux and isopropanol flux

Stability of SA/PEI-GO/PAN TFNC Pervaporation Membrane

The stability of the composite membrane is vital to evaluate the membrane pervaporation performance. So the separation property was measured under different operation temperature and testing time to investigate the stability of the three-tier composite membrane in pervaporation application. Figure 10a showed the pervaporation flux and selectivity of SA/PEI(75)-GO-60/PAN TFNC membrane conducted at different operating temperature (from 30 to 70 °C) in separating 90 wt% aqueous isopropanol solution. With rising of temperature, the permeate flux exhibited a dramatically increasing trend but the separation factor was decreased slightly versus the increasing trend. There are three aspects of effect factors which can illustrate impacts of operation temperature on pervaporation dehydration: the driving force, the membrane structure and the interactions between permeate molecules and membrane. Firstly, when the operation temperature increased, the thermal promoted molecule motion enhanced the partial pressure in the upstream side, which resulted in the increased mass transfer driving force. The enhanced force was in favor of accelerating the diffusion rates which facilitated water and isopropanol molecules clusters passing through the membrane quickly to the vacuum downstream side [42]. Secondly, the increased mobility of SA chains resulted in more free volume and loosen chemical structure at high temperature which was in favor of clusters transferring. So the total flux grew continuously with the rising temperature. However, caused by the promoted thermal molecular motion, the multiple interactions between intermediate layer and top barrier layer were weakened. The loosen membrane structure was obtained under the dual influence of weakened interaction and increased chain mobility. Moreover, due to quicker molecular motion with rising temperature, the interaction between permeate molecules and membrane weakened, the molecule adsorption on membrane surface was reduced subsequently. Hence the selectivity of SA layer reduced and the destruction of the three-tier composite membrane structure become more severe, which led to the declination of membrane separation factor. To quantitatively measure the effects of temperature on pervaporation, the partial flux and apparent activation energy (Ep) were calculated as showed in Fig. 10b–c. In Fig. 10b, the water flux and isopropanol flux was increasing with the rising temperature and the water flux was much higher than isopropanol. Figure 10c demonstrated the apparent activation energy for water and isopropanol were 4.47 kJ/mol and 22.4 kJ/mol, respectively. Lower water activation energy indicated that water molecules required less energy to diffuse through the membrane so the TFNC membrane possessed much higher water flux. While the much higher isopropanol activation energy suggested that the isopropanol flux was more sensitive towards temperature. That is to say the increase rate of isopropanol flux was higher than water. With the augment of temperature, more isopropanol molecules permeated through the TFNC membrane, resulting in the decreased selectivity. These quantitative measurements has well confirmed above assumptions and explanation.
Fig. 10

Effect of the operating temperature on pervaporation performance of SA/PEI(75)-GO-60/PAN TFNC membrane in separating 90 wt% aqueous isopropanol solution a total flux and separation factor; b flux of each component and c apparent activation energy of each component

Comparison of Pervaporation Performance with Literatures

Table 1 summaries some pervaporation performance of SA-based mixed matrix membranes (MMM) and composite membranes under similar operation condition reported in recent literatures. The SA/PEI(75)-GO-60/PAN TFNC membrane exhibited superior permeation and selectivity for isopropanol dehydration, and the PSI was calculated to comprehensively compare membrane pervaporation performance. At low feed temperatures (below 30 °C), compared with the other SA-based membranes, the TFNC membrane we fabricated exhibited much higher permeation flux of 1713 g/m2 h and comparable separation factor of 2991. Furthermore, the PSI value still maintained at high level, which was larger than 8 times of that of SA-HPA MMMs. At high temperature (above 70 °C), the accelerated thermal molecule motion, intensive mass transfer driving force and the increased mobility of SA chains enhanced the permeate flux but reduced the separation flux. The stability and interfacial adhesion of each layer in SA/PEI(75)-GO-60/PAN TFNC membrane may be slightly weakened. But it should be mentioned that the overall performance still maintained the high level in pervaporation. The PSI value was much higher than some MMMs and composite membrane, which was about five and two times higher than 5.95 of SA-ZIF-8 MMM and 11.0 of HA/SA/mPAN composite membrane at 76 °C, respectively. PAN electrospun substrate brought in high permeate flux and low performance resistance due to its high porosity and interconnected pore structure. The three-tier structure with functional intermediate layer overcame the shortcoming such as the leakage of coating liquids into substrate, the defects caused by fluctuated surface of nanofibrous substrate, and poor structure integrity of thin-film nanofibrous composite membrane. In brief, the TFNC membrane we fabricated in this study was competitive in practical application.
Table 1

Comparison of separation performance with various membranes in the literature


Water content in feed (wt%)

Temp. (°C)

Flux (g/m2 h)

Separation factor

PSI (× 105)

























































































This study






aHydrolysed PAN ultrafiltration membranes

bElectrospun PAN nanofibrous substrates


In this work, the facile and novel design of composite membrane showed great potential for further enhancing the pervaporation dehydration separation effectivity of organic aqueous mixtures. PEI-GO intermediate layer provided a functional flat in order to overcome the easy penetration into nanofibrous mat and was beneficial for uniform distribution of casting solution to obtain a defect-free top layer. Meanwhile, it was in favor of enhancing interfacial adhesion between nanofibrous substrate and hydrophilic top layer to improve the integrity and stability of TFNC pervaporation membrane. The intercalation of PEI, not only modified the membrane channel structure and expanded d-spacing of GO lamellae to enhance the water affinity and alcohol repulsion, but also improved the interaction of GO sheets to ensure the stability of GO-based layer in application. The optimized SA/PEI(75)-GO-60/PAN TFNC membrane exhibited competitive pervaporation performance with the permeation flux of 2009 g/m2 h and separation factor of 1276 operated at 70 °C with 90 wt% isopropanol solution and the novel three-tier composite membrane showed the great prospect for developing high quality pervaporation membrane.



This work was supported by Natural Science Foundation of Shanghai (19ZR1401300) and Program for Innovative Research Team in University of Ministry of Education of China (IRT_16R13).


Funding has been received from Natural Science Foundation of Shanghai with Grand No. 19ZR1401300 and Program for Innovative Research Team in University of Ministry of Education of China with Grand No. IRT_16R13.

Supplementary material

42765_2019_5_MOESM1_ESM.docx (438 kb)
Supplementary material 1 (DOCX 438 kb)


  1. 1.
    Semenova SI, Ohya H, Soontarapa K. Hydrophilic membranes for pervaporation: an analytical review. Desalination. 1997;110:251.CrossRefGoogle Scholar
  2. 2.
    Ong YK, Shi GM, Le NL, Tang YP, Zuo J, Nunes SP, Chung T-S. Recent membrane development for pervaporation processes. Prog Polym Sci. 2016;57:1.CrossRefGoogle Scholar
  3. 3.
    Chapman PD, Oliveira T, Livingston AG, Li K. Membranes for the dehydration of solvents by pervaporation. J Membr Sci. 2008;318:5.CrossRefGoogle Scholar
  4. 4.
    Zuo J, Chung T-S. Design and synthesis of a fluoro-silane amine monomer for novel thin film composite membranes to dehydrate ethanol via pervaporation. J. Mater. Chem. A. 2013;1:9814.CrossRefGoogle Scholar
  5. 5.
    Liu J, Bernstein R. High-flux thin-film composite polyelectrolyte hydrogel membranes for ethanol dehydration by pervaporation. J Membr Sci. 2017;534:83.CrossRefGoogle Scholar
  6. 6.
    Cheng X, Jiang Z, Cheng X, Guo S, Tang L, Yang H, Wu H, Pan F, Zhang P, Cao X. Bimetallic metal-organic frameworks nanocages as multi-functional fillers for water-selective membranes. J Membr Sci. 2018;545:19.CrossRefGoogle Scholar
  7. 7.
    Yang H, Cheng X, Cheng X, Pan F, Wu H, Liu G, Song Y, Cao X, Jiang Z. Highly water-selective membranes based on hollow covalent organic frameworks with fast transport pathways. J Membr Sci. 2018;565:331.CrossRefGoogle Scholar
  8. 8.
    Wang Z, Ma H, Hsiao BS, Chu B. Nanofibrous ultrafiltration membranes containing cross-linked poly (ethylene glycol) and cellulose nanofiber composite barrier layer. Polymer. 2014;55:366.CrossRefGoogle Scholar
  9. 9.
    Wang X, Fang D, Hsiao BS, Chu B. Nanofiltration membranes based on thin-film nanofibrous composites. J Membr Sci. 2014;469:188.CrossRefGoogle Scholar
  10. 10.
    Wang J, Zhang P, Liang B, Liu Y, Xu T, Wang L, Cao B, Pan K. Graphene oxide as an effective barrier on a porous nanofibrous membrane for water treatment. ACS Appl Mater Interfaces. 2016;8:6211.CrossRefGoogle Scholar
  11. 11.
    Huang ZM, Zhang YZ, Kotaki M, Ramakrishna S. A review on polymer nanofibers by electrospinning and their applications in nanocomposites. Compos Sci Technol. 2003;63:2223.CrossRefGoogle Scholar
  12. 12.
    Yu X, Shen L, Zhu Y, Li X, Yang Y, Wang X, Zhu M, Hsiao BS. High performance thin-film nanofibrous composite hemodialysis membranes with efficient middle-molecule uremic toxin removal. J Membr Sci. 2017;523:173.CrossRefGoogle Scholar
  13. 13.
    Yoon K, Kim K, Wang X, Fang D, Hsiao BS, Chu B. High flux ultrafiltration membranes based on electrospun nanofibrous PAN scaffolds and chitosan coating. Polymer. 2006;47:2434.CrossRefGoogle Scholar
  14. 14.
    Shen L, Yu X, Cheng C, Song C, Wang X, Zhu M, Hsiao BS. High filtration performance thin film nanofibrous composite membrane prepared by electrospraying technique and hot-pressing treatment. J Membr Sci. 2016;499:470.CrossRefGoogle Scholar
  15. 15.
    Hung W-S, Lai C-L, An Q, De Guzman M, Shen T-J, Huang Y-H, Chang K-C, Tsou C-H, Hu C-C, Lee K-R. A study on high-performance composite membranes comprising heterogeneous polyamide layers on an electrospun substrate for ethanol dehydration. J Membr Sci. 2014;470:513.CrossRefGoogle Scholar
  16. 16.
    Zhang M, Lin H, Shen L, Liao BQ, Wu X, Li R. Effect of calcium ions on fouling properties of alginate solution and its mechanisms. J Membr Sci. 2017;525:320.CrossRefGoogle Scholar
  17. 17.
    Dudek G, Krasowska M, Turczyn R, Gnus M, Strzelewicz A. Structure, morphology and separation efficiency of hybrid Alg/Fe3O4 membranes in pervaporative dehydration of ethanol. Sep Purif Technol. 2017;182:101.CrossRefGoogle Scholar
  18. 18.
    Zhao J, Fang C, Zhu Y, He G, Pan F, Jiang Z, Zhang P, Cao X, Wang B. Manipulating the interfacial interactions of composite membranes via a mussel-inspired approach for enhanced separation selectivity. J. Mater. Chem. A. 2015;3:19980.CrossRefGoogle Scholar
  19. 19.
    Ma J, Zhang M, Wu H, Yin X, Chen J, Jiang Z. Mussel-inspired fabrication of structurally stable chitosan/polyacrylonitrile composite membrane for pervaporation dehydration. J Membr Sci. 2010;348:150.CrossRefGoogle Scholar
  20. 20.
    Xu J, Gao C, Feng X. Thin-film-composite membranes comprising of self-assembled polyelectrolytes for separation of water from ethylene glycol by pervaporation. J Membr Sci. 2010;352:197.CrossRefGoogle Scholar
  21. 21.
    Chen Y, Xiangli F, Jin W, Xu N. Organic–inorganic composite pervaporation membranes prepared by self-assembly of polyelectrolyte multilayers on macroporous ceramic supports. J Membr Sci. 2007;302:78.CrossRefGoogle Scholar
  22. 22.
    Mi B. Graphene oxide membranes for ionic and molecular sieving. Science. 2014;343:740.CrossRefGoogle Scholar
  23. 23.
    Huang K, Liu G, Lou Y, Dong Z, Shen J, Jin W. A graphene oxide membrane with highly selective molecular separation of aqueous organic solution. Angew Chem Int Ed. 2014;126:7049.CrossRefGoogle Scholar
  24. 24.
    Han Y, Xu Z, Gao C. Ultrathin graphene nanofiltration membrane for water purification. Adv Funct Mater. 2013;23:3693.CrossRefGoogle Scholar
  25. 25.
    Xu WL, Fang C, Zhou F, Song Z, Liu Q, Qiao R, Yu M. Self-assembly: a facile way of forming ultrathin, high-performance graphene oxide membranes for water purification. Nano Lett. 2017;17:2928.CrossRefGoogle Scholar
  26. 26.
    Nair R, Wu H, Jayaram P, Grigorieva I, Geim A. Unimpeded permeation of water through helium-leak–tight graphene-based membranes. Science. 2012;335:442.CrossRefGoogle Scholar
  27. 27.
    Cheng C, Shen L, Yu X, Yang Y, Li X, Wang X. Robust construction of a graphene oxide barrier layer on a nanofibrous substrate assisted by the flexible poly(vinylalcohol) for efficient pervaporation desalination. J. Mater. Chem. A. 2017;5:3558.CrossRefGoogle Scholar
  28. 28.
    Hua D, Rai RK, Zhang Y, Chung T-S. Aldehyde functionalized graphene oxide frameworks as robust membrane materials for pervaporative alcohol dehydration. Chem Eng Sci. 2017;161:341.CrossRefGoogle Scholar
  29. 29.
    Hung W-S, Tsou C-H, De Guzman M, An Q-F, Liu Y-L, Zhang Y-M, Hu C-C, Lee K-R, Lai J-Y. Cross-linking with diamine monomers to prepare composite graphene oxide-framework membranes with varying d-spacing. Chem Mater. 2014;26:2983.CrossRefGoogle Scholar
  30. 30.
    Lu J, Gu Y, Chen Y, Yan X, Guo Y, Lang W. Ultrahigh permeability of graphene-based membranes by adjusting D-spacing with poly (ethylene imine) for the separation of dye wastewater. Sep Purif Technol. 2019;210:737.CrossRefGoogle Scholar
  31. 31.
    Zhang L, Chen B, Ghaffar A, Zhu X. Nanocomposite membrane with polyethylenimine-grafted graphene oxide as a novel additive to enhance pollutant filtration performance. Environ Sci Technol. 2018;52:5920.CrossRefGoogle Scholar
  32. 32.
    Xu Y, Bai H, Lu G, Li C, Shi G. Flexible graphene films via the filtration of water-soluble noncovalent functionalized graphene sheets. J Am Chem Soc. 2008;130:5856.CrossRefGoogle Scholar
  33. 33.
    Park WB, Bandyopadhyay P, Nguyen TT, Kuila T, Kim NH, Lee JH. Effect of high molecular weight polyethyleneimine functionalized graphene oxide coated polyethylene terephthalate film on the hydrogen gas barrier properties. Compos. Part B: Eng. 2016;106:316.CrossRefGoogle Scholar
  34. 34.
    Liu G, Jiang Z, Chen C, Hou L, Gao B, Yang H, Wu H, Pan F, Zhang P, Cao X. Preparation of ultrathin, robust membranes through reactive layer-by-layer (LbL) assembly for pervaporation dehydration. J Membr Sci. 2017;537:229.CrossRefGoogle Scholar
  35. 35.
    Zhang Y, Zhang S, Chung T-S. Nanometric graphene oxide framework membranes with enhanced heavy metal removal via nanofiltration. Environ Sci Technol. 2015;49:10235.CrossRefGoogle Scholar
  36. 36.
    Liu C, Liu H, Lu C, Tang K, Zhang Y. Polyethyleneimine-modified graphene oxide/PNIPAm thermoresponsive hydrogels with rapid swelling/deswelling and improved mechanical properties. J Mater Sci. 2017;52:11715.CrossRefGoogle Scholar
  37. 37.
    Liu H, Kuila T, Kim NH, Ku B-C, Lee JH. In situ synthesis of the reduced graphene oxide–polyethyleneimine composite and its gas barrier properties. J. Mater. Chem. A. 2013;1:3739.CrossRefGoogle Scholar
  38. 38.
    Roy S, Tang X, Das T, Zhang L, Li Y, Ting S, Hu X, Yue C. Enhanced molecular level dispersion and interface bonding at low loading of modified graphene oxide to fabricate super nylon 12 composites. ACS Appl Mater Interfaces. 2015;7:3142.CrossRefGoogle Scholar
  39. 39.
    Huang K, Liu G, Shen J, Chu Z, Zhou H, Gu X, Jin W, Xu N. High-efficiency water-transport channels using the synergistic effect of a hydrophilic polymer and graphene oxide laminates. Adv Funct Mater. 2015;25:5809.CrossRefGoogle Scholar
  40. 40.
    Li Q, Liu Q, Zhao J, Hua Y, Sun J, Duan J, Jin W. High efficient water/ethanol separation by a mixed matrix membrane incorporating MOF filler with high water adsorption capacity. J Membr Sci. 2017;544:68.CrossRefGoogle Scholar
  41. 41.
    Wang M, Pan F, Yang L, Song Y, Wu H, Cheng X, Liu G, Yang H, Wang H, Jiang Z. Graphene oxide quantum dots incorporated nanocomposite membranes with high water flux for pervaporative dehydration. J Membr Sci. 2018;563:903.CrossRefGoogle Scholar
  42. 42.
    Yang H, Wu H, Pan F, Li Z, Ding H, Liu G, Jiang Z, Zhang P, Cao X, Wang B. Highly water-permeable and stable hybrid membrane with asymmetric covalent organic framework distribution. J Membr Sci. 2016;520:583.CrossRefGoogle Scholar
  43. 43.
    Kalyani S, Smitha B, Sridhar S, Krishnaiah A. Pervaporation separation of ethanol–water mixtures through sodium alginate membranes. Desalination. 2008;229:68.CrossRefGoogle Scholar
  44. 44.
    Kariduraganavar M, Kittur A, Kulkarni S, Ramesh K. Development of novel pervaporation membranes for the separation of water–isopropanol mixtures using sodium alginate and NaY zeolite. J Membr Sci. 2004;238:165.CrossRefGoogle Scholar
  45. 45.
    Nigiz FU, Dogan H, Hilmioglu ND. Pervaporation of ethanol/water mixtures using clinoptilolite and 4A filled sodium alginate membranes. Desalination. 2012;300:24.CrossRefGoogle Scholar
  46. 46.
    Magalad VT, Supale AR, Maradur SP, Gokavi GS, Aminabhavi TM. Preyssler type heteropolyacid-incorporated highly water-selective sodium alginate-based inorganic–organic hybrid membranes for pervaporation dehydration of ethanol. Chem Eng J. 2010;159:75.CrossRefGoogle Scholar
  47. 47.
    Adoor SG, Rajineekanth V, Nadagouda MN, Rao KC, Dionysiou DD, Aminabhavi TM. Exploration of nanocomposite membranes composed of phosphotungstic acid in sodium alginate for separation of aqueous–organic mixtures by pervaporation. Sep Purif Technol. 2013;113:64.CrossRefGoogle Scholar
  48. 48.
    Liu G, Jiang Z, Cao K, Nair S, Cheng X, Zhao J, Gomaa H, Wu H, Pan F. Pervaporation performance comparison of hybrid membranes filled with two-dimensional ZIF-L nanosheets and zero-dimensional ZIF-8 nanoparticles. J Membr Sci. 2017;523:185.CrossRefGoogle Scholar
  49. 49.
    Gao C, Zhang M, Ding J, Pan F, Jiang Z, Li Y, Zhao J. Pervaporation dehydration of ethanol by hyaluronic acid/sodium alginate two-active-layer composite membranes. Carbohydr Polym. 2014;99:158.CrossRefGoogle Scholar
  50. 50.
    Ji C, Xue S, Xu Z. Novel swelling-resistant sodium alginate membrane branching modified by glycogen for highly aqueous ethanol solution pervaporation. ACS Appl Mater Interfaces. 2016;8:27243.CrossRefGoogle Scholar

Copyright information

© Donghua University, Shanghai, China 2019

Authors and Affiliations

  1. 1.State Key Lab for Modification of Chemical Fibers and Polymer MaterialsDonghua UniversityShanghaiPeople’s Republic of China
  2. 2.Department of ChemistryStony Brook UniversityStony BrookUSA

Personalised recommendations