Novel Swelling-Resistant Sodium Alginate Membrane Branching

Sep 28, 2016 - (9) The carbohydrate chains are composed of sugar moieties, which contain large amounts of hydroxyl groups and carboxyl groups, endowin...
0 downloads 9 Views 3MB Size
Research Article www.acsami.org

Novel Swelling-Resistant Sodium Alginate Membrane Branching Modified by Glycogen for Highly Aqueous Ethanol Solution Pervaporation Chen-Hao Ji, Shuang-Mei Xue, and Zhen-Liang Xu* State Key Laboratory of Chemical Engineering, Membrane Science and Engineering R&D Center, Chemical Engineering Research Center, East China University of Science and Technology (ECUST), 130 Meilong Road, Shanghai 200237, China ABSTRACT: A novel carbohydrate chain cross-linking method of sodium alginate (SA) is proposed in which glycogen with the branched-chain structure is utilized to cross-link with SA matrix by the bridging of glutaraldehyde (GA). The active layer of SA composite ceramic membrane modified by glycogen and GA for pervaporation (PV) demonstrates great advantages. The branched structure increases the chain density of the active layer, which compresses the free volume between the carbohydrate chains of SA. Large amounts of hydroxyl groups are consumed during the reaction with GA, which reduces the hydrogen bond formation between water molecules and the polysaccharide matrix. The two factors benefit the active layer with great improvement in swelling resistance, promoting the potential of the active layer for the dehydration of an ethanol−water solution containing high water content. Meanwhile, the modified active layer is loaded on the rigid α-Al2O3 ceramic membrane by dipcoating method with the enhancement of anti-deformation and controllable thickness of the active layer. Characterization techniques such as SEM, AFM, XRD, FTIR, XPS, and water contact angle are utilized to observe the composite structure and surface morphology of the composite membrane, to probe the free volume variation, and to determine the chemical composition and hydrophilicity difference of the active layer caused by the different glycogen additive amounts. The membrane containing 3% glycogen in the selective layer demonstrates the flux at 1250 g m−2 h−1 coupled with the separation factor of 187 in the 25 wt % water content feed solution at the operating temperature of 75 °C, reflecting superior pervaporation processing capacity compared with the general organic PV membranes in the same condition. KEYWORDS: sodium alginate, glycogen, branched carbohydrate chain, swelling resistance, high-flux organic pervaporation membrane water flux under the mild operation conditions.8 When the water content and the operating temperature simultaneously increase to a high level, the excessive swelling will deprive the overall separating capacity of the membrane despite the increase in flux. Therefore, extensive efforts have been made in this work to explore the potential of organic PV membrane under harsh operating conditions. Sodium alginate (SA) is a low-cost and easily available natural polysaccharide that possesses excellent membraneforming ability benefiting from its linear and long carbohydrate chain structure.9 The carbohydrate chains are composed of sugar moieties, which contain large amounts of hydroxyl groups and carboxyl groups, endowing the polysaccharide with excellent hydrophilicity and water permselective nanochannels. However, the loose stacking and less interaction of the linear chains will generate overly large free volumes during the swelling of SA membrane under high water content feed

1. INTRODUCTION Pervaporation (PV) has been considered as an efficient and environmentally friendly separation technology in the dehydration of various organic intermediates with great industrial importance, mostly represented by ethanol.1 PV membranes, serving as the core element in the PV process, demonstrate distinct performance derived from different membrane materials, membrane structures, and nanomaterials doped in the membranes.2−6 All of these regulation methods are aiming at the same goal, enhancing the separation selectivity and promoting the flux of the desired permeate so as to further improve the pervaporation performance.7 PV membranes based on organic active layers possess stable long-term performance but always face the dilemma in which the increase of the desired permeate flux and the separation factor always requires the sacrifice of each other. Taking the water permselective organic PV membrane as an example, it has advantages in longterm operation (related to industrial application value) that can maintain a high separation factor under harsh operating conditions (turbulence flow and impact of impurities) compared to inorganic ones but coupled with relatively low © 2016 American Chemical Society

Received: August 10, 2016 Accepted: September 19, 2016 Published: September 28, 2016 27243

DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253

Research Article

ACS Applied Materials & Interfaces

Figure 1. Specific mechanism and purpose of the chemical modification in the active layer.

optimized to promote the swelling resistance under high water content feed solution and high operating temperature as well as anti-deformation ability under ultrahigh vacuum degree.31,32 Substantial characterizations were conducted to comprehensively investigate the mechanism of performance improvement benefited from the composite structure and branched-chain structure in the active layer.33 As a consequence, the dual optimization of the ceramic supported hybrid polysaccharide membrane endows the membrane with great potential to overcome the trade-off phenomenon and high processing capacity, especially under the high water content feed solution and operating temperature, because the appropriate swelling promotes the flux coupled with the slight loss of separation factor benefited from the branched cross-linking. Figure 1 well describes the mechanism and purpose of the chemical modification in the active layer of the composite membrane. The difference lying in the PV performance of the polysaccharide active layer with or without the branching modification is clearly depicted in the schematic diagram.

solution and high operating temperature, leading to inferior separation factor and the permeation of undesired components.10 As a result, the adjustment of composite structure and the chemical modification of the selective layer give solutions for the optimization of high-permeate flux as well as acceptable separation factors for PV membranes. In the adjustment of composite structure, selecting an appropriate supporting membrane and enhancing the compatibility between the active and supporting layer can help enhance the separation performance and resist deformation in the PV process by restricting the swelling dimensionality and improving the rigid structure of the membrane under high vacuum degree, respectively.11 In the chemical modification of the active layer, the incorporation of specific nanofillers such as modified carbon nanotubes,12,13 graphenes,14,15 metal−organic frameworks (MOFs), 16 zeolitic imidazolate frameworks (ZIFs),17,18 nanoparticles,19−21 and heteropolyacids22,23 can help further increase the hydrophilicity and create particular permselective nanochannels for water molecules through the polymer matrix of the active layer.24 Besides, grafting modification and polymer blending25,26 are also efficient ways to improve the separation performance of PV membranes. However, these traditional modification methods may be accompanied by aggravated swelling degree, especially in high water content feed solutions, which decreases the operational flexibility. Cross-linking with an organic cross-linker27,28 and ionic cross-linking29 are both popular modifications for SA. The former is based on the combination of particular functional groups on the cross-linkers and SA matrix, whereas the latter is based on the chelation between the carboxyl groups on SA and the ions represented by Ca2+; the essence of ionic cross-linking is to close the distance between the linear chains of SA, but it lacks the subdivision of the free volumes in the active layer matrix.30 On the other hand, the effect of cross-linking with organic cross-linkers depends on the chemical properties of the cross-linker and the enhancement in chemical structure of the active layer. In this work, a novel hybrid polysaccharide PV active layer was fabricated by cross-linking branched polysaccharide (glycogen) with traditional linear polysaccharide matrix (sodium alginate) followed by exterior dip-coating on the rigid α-Al2O3 ceramic membrane. Both the membrane structure and the chemical modification of the active layer were

2. EXPERIMENTAL SECTION 2.1. Materials. Sodium alginate (SA), glycogen (MW = 2.7 × 105− 3.5 × 106), and 25 wt % glutaraldehyde (GA) aqueous solution were purchased from Sinopharm (China). The α-Al2O3 ceramic membrane with a mean pore size of 20 nm as the supporting membrane was fabricated in our laboratory.27 The superfine titania nanoparticle with a diameter around 5 nm was obtained from Xuan Cheng Jin Rui New Material Co., Ltd. (China). Absolute ethyl alcohol was purchased from Sinopharm (China), and the deionized water was prepared in our laboratory. 2.2. Fabrication of the Composite Ceramic PV Membrane. To better investigate the swelling behavior of the active layer coated on the rigid α-Al2O3 ceramic hollow fiber membrane instead of investigating the modified polysaccharide active layer separated from the supporting layer, the fabrications of a composite membrane and its testing module were conducted simultaneously. The advantage of such a method is that it accurately obtains the weight of the active layer for subsequent calculation of the in situ swelling degree. The specific preparation process of a single membrane testing module is as follows: a 60 mm section of the α-Al2O3 ceramic hollow fiber membrane was connected to a specially made stainless steel tube, the joint between the ceramic membrane and stainless steel tube as well as the other terminus of the ceramic membrane were sealed by epoxy sealant. After the epoxy sealant solidified thoroughly, surface treatment was 27244

DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253

Research Article

ACS Applied Materials & Interfaces

Figure 2. FTIR spectra of the composite membranes. conducted on the outer surface of the ceramic membrane to fill the pores so as to provide a much smoother interface for the formation of defect-free active layer. The specific process of the interface treatment of the ceramic membrane is dip-coating the ceramic membrane four times with the preprepared solution, which contains 0.5% superfine titania (5 nm) and 0.2% sodium alginate. The superfine titania was used to fill the pores, whereas SA was used to improve the viscous force between the titania nanoparticles and the ceramic supporting membrane. After the interface treatment of the ceramic membrane was finished, the entire membrane module was weighed, including the weight of the ceramic membrane, the stainless steel tube, the epoxy sealant, and the pore filler. The total weight of the membrane module was noted as m1. The ceramic membrane was dip-coated twice with a hybrid polysaccharide solution of sodium alginate and glycogen in different concentrations and then placed upside down during the drying process in ambient temperature to ensure homogeneous distribution of the active layer along the axial direction of the hollow fiber membrane. After the drying process, the composite membranes were immersed in 0.1 wt % GA solution for 4 h of cross-linking followed by thorough drying and preservation in a dust-free environment. The weight of the resulting membrane module was measured and noted as m2. Moreover, the weight of polysaccharide active layer easily obtained by the difference between m2 and m1 was noted as mac for convenient measurement of the swelling degree. The modified polysaccharide solution contains 2 wt % sodium alginate with the addition of 1, 2, 3, and 5 wt % glycogen according to the weight of the SA, respectively. For convenience, these composite membranes were denoted PVMCGSA-1, PVMCGSA-2, PVMCGSA-3, and PVMCGSA-5, whereas the one only cross-linked with glutaraldehyde without any addition of glycogen was denoted PVMCGSA-0. The capital letters G and SA represent glycogen and sodium alginate, respectively. 2.3. Characterizations. Field emission scanning electron microscopy (Nova NanoSEM 450, USA) was used to observe the outer surface and cross-section morphology of the composite membrane. Meanwhile, the modified polysaccharide active layer, that is, the outer surface of the composite membrane, was probed by atomic force microscopy (AFM, Veeco/DI) to visualize the surface morphology of the active layer and obtain the surface roughness. The chemical composition of the composite membrane was analyzed by Fourier transform infrared spectroscopy (FTIR, Nicolet-6700) combined with X-ray photoelectron spectroscopy (XPS, ESCALAB 25) to investigate the combination mode between SA and glycogen. The free volume variation of the polysaccharide active layer was indirectly detected by X-ray diffraction (XRD, D8 Advance Vinci) through the measurement of the SA interchain spacing diffraction intensity. Hydrophilicities of composite membranes’ separating layers expressed in dynamic water contact angle were characterized by the water contact angle using the

contact angle meter (JC2000A, Shanghai Zhong Cheng Digital Equipment Co., Ltd., China) at 25 °C. 2.4. Pervaporation and Membrane Swelling Synchronous Experiment. The PV performance of the composite hollow fiber membrane was conducted in the dehydration of ethanol−water solution at operating temperatures of 45, 55, 65, and 75 °C. The feed solution was 1 L of ethanol−water solution with the water content varying from 10 to 25 wt % to systematically investigate the separation performance and swelling behavior of the composite membrane. The downstream pressure was 0.8 kPa controlled by a vacuum pump to provide the chemical potential difference for permeation. The calculations of the flux (J) and separation factor (α) refer to the following formulas:

J=

Q A×t

(1)

α=

Yw /Ya X w /Xa

(2)

Q represents the total weight of the permeate; A and t represent the effective area and operating time of the PV process, respectively; Y and X serve as symbols for the weight fraction of a single component in the permeate and feed side, respectively; and the lower case letters “w” and “a” are marks for water and alcohol, respectively. Because the separation performance was affected by swelling degree as well as the relatively low PV flux of organic-based active layer, the duration of a single PV test was set as 2 h for accuracy, and each membrane was tested at least three times under a particular operating temperature and water content to evaluate the separation performance by the average data. After the last PV testing process was finished, which meant the composite membrane had undergone 6 h of swelling in the practical PV operating condition, the membrane was taken from the water− ethanol system and placed in ambient atmosphere for 30 s to thoroughly remove the moisture in the pores of the supporting αAl2O3 membrane by the vacuum degree inside the lumen, which blew away the water−ethanol on the outer surface of the membrane by compressed air. The membrane module was separated from the testing equipment and weighed again, which was denoted m3. The weight of the entire membrane module (m2) was mentioned in the preparation process above, so that the weight of absorbed solution denoted msw equals m3 minus m2. The swelling degree (ω) of the polysaccharide active layer can be accurately obtained by the following formula: ω= 27245

m − m2 msw = 3 mac m2 − m1

(3) DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253

Research Article

ACS Applied Materials & Interfaces

Figure 3. Molecular structural formula of SA and glycogen and the cross-linking mechanism by GA.

3. RESULTS AND DISCUSSION 3.1. Chemical Composition. The chemical composition of the membranes was analyzed by the FTIR spectrum combined with the XPS technique to investigate the influence caused by the addition of glycogen in the active layer. From the FTIR spectra (Figure 2) of the composite membranes, it can be observed that these adsorption peaks related to the metal oxides (Al2O3 and TiO2)34,35 are distributed between the wavenumbers of 648 and 460 cm−1 with approximately equal transmittance, but the intensity of the absorption peaks at wavenumbers of 3440 and 1637 cm−1 decrease gradually with the increase in glycogen additive amount. The two peaks above are related to the stretching vibration and the bending vibration of hydroxyl groups, respectively. As a consequence, the decrease in the hydroxyl group’s intensity reveals the promoted cross-linking between carbohydrate chains of SA and glycogen by GA under the same cross-linking condition. The consumption of hydroxyl groups accompanied by the addition of glycogen will certainly cause the relative content decrease of the hydroxyl groups as reflected in the FTIR spectra. SA and glycogen are both representative polysaccharides but composed of different sugar moieties; the former one is a linear copolymer composed of β-D-mannuronic (M) and α-L-guluronic (G) residues of various sequences,36 and the latter one is composed solely of glucose by the linkage of α-1,4-glucosidic bond and branched by the α-1,6-glucosidic bond. The molecular structural formula of these two polysaccharides and the crosslinking mechanism with GA37 are displayed in Figure 3. The linear carbohydrate chain structure and high molecular weight of SA endow it with better membrane-forming ability compared to glycogen, which possesses a branched carbohydrate chain

and a relatively low molecular weight. However, the linear, rigid, and overlong carbohydrate chain of SA also generates high steric hindrance for combination of two different carbohydrate chains with GA. Through the fitting calculation of the C 1s region in the XPS result (Figure 4) it can be found that the CO content in PVMCGSA-0 (17.60%) is nearly 2% higher than that of PVMCGSA-5 (15.64%). Because the CO content of PVMCGSA-0 and PVMCGSA-5 originating from the −COONa group of SA is on the same level, the excessive CO content of PVMCGSA-0 is mainly derived from the incomplete cross-linking reaction between the glutaraldehyde and the polysaccharide matrix. The incomplete cross-linking reaction of glutaraldehyde refers to the cross-linking reaction in which only one of the formyl groups has been cross-linked with two hydroxyl groups on a single sugar moiety of SA, but the other formyl group on GA is still unreacted. However, in the cross-linking process of PVMCGSA-5, the addition of glycogen, which possesses the branching carbohydrate chain structure (higher opportunity to approach the sugar moiety of SA), chemical structure similarity with sodium alginate (both composed of sugar moiety) and better carbohydrate chain mobility (shorter chain length and lower molecular weight compared to SA) make it possible for glutaraldehyde to crosslink these two polysaccharides together by consuming both of the two formyl groups on the two sides of this molecule, thus leading to the result that the CO content in PVMCGSA-0 (17.60%) is nearly 2% higher than that of PVMCGSA-5 (15.64%) and the CO content in PVMCGSA-0 (60.31%) is approximately 4% lower than that of PVMCGSA-5 (64.53%) because the break of a CO double bond in GA is accompanied by the generation of two CO single bonds during the cross-linking process. Meanwhile, intramolecular 27246

DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253

Research Article

ACS Applied Materials & Interfaces

Figure 4. XPS spectra: (a-1, b-1) wide scan of PVMCGSA-0 and PVMCGSA-5; (a-2, b-2) C 1s region of PVMCGSA-0 and PVMCGSA-5.

respectively. Through comparison of the AFM images it can be observed that the surface roughness of the polysaccharide active layers decreases notably with the increase in glycogen additive amount. The roughness parameters are listed in Table 2 in which Ra and Rm represent for the arithmetic mean roughness and maximum roughness, respectively. In consideration that the preparation processes of all the composite membranes are exactly the same without any difference, the surface structure difference in the membranes’ active layers may mainly originate from the chemical composition of the active layer. The active layer of the composite membrane forms through the solvent evaporation of the dip-coated polysaccharide solution. Therefore, the evaporation process of water will evidently influence the surface morphology. As proved in FTIR spectra, the branched polysaccharide active layer demonstrates the lower relative content of free hydroxyl groups due to better crosslinking between carbohydrate chains of SA and glycogen. The cross-linking of these hydroxyl groups homogenizes the electron cloud density of the original hydroxyl groups, leading to a decrease in water affinity, which will eventually cause a decline in water-holding capacity of the hybrid polysaccharide active layer. On the other hand, the free volume in carbohydrate chains of the polysaccharide layer is compressed by the formation of branched structure with the addition of glycogen, which also decreases the water-holding capacity. As a result, the evaporation rate and capacity of water in the polysaccharide chains of PVMCGSA-0 are much faster and larger than those in PVMCGSA-1, PVMCGSA-3, and PVMCGSA-5 under the same ambient atmosphere (at a

dehydration during the cross-linking process also leads to an increase of the carbon−oxygen ratio, which can be observed from the elementary composition (listed in Table 1) of Table 1. Elementary Composition of PVMCGSA-0 and PVMCGSA-5 code

O 1s (%)

Al 2p (%)

C 1s (%)

Ti 2p (%)

PVMCGSA-0 PVMCGSA-5

55.20 52.61

33.28 33.01

10.25 13.13

1.27 1.25

PVMCGSA-0 (10.25−55.20%) and PVMCGSA-5 (13.13− 52.61%) because of the loss of oxygen element in the dehydration process. Such a cross-linking process between the carbohydrate chains of SA and the branched polysaccharide (glycogen) can significantly optimize the swelling degree of the active layer by structure improvement. The moderate decrease in hydroxyl groups relative content is closely related to the water uptake capacity of the active layer. These two factors based on chemical composition optimization will further benefit this kind of hybrid polysaccharide composite membrane in the potential PV operation under highly aqueous ethanol solution at high temperature, which promotes the processing capacity of the PV process. 3.2. Surface Morphology and Composite Structure. Figure 5 displays the surface morphology of the composite membrane by the observation of SEM combined with the AFM technique. The lower case letters “a”, “b”, and “c” represent the membranes PVMCGSA-0, PVMCGSA-3, and PVMCGSA-5, 27247

DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253

Research Article

ACS Applied Materials & Interfaces

Figure 5. SEM and AFM images of the surface morphology of the composite membranes: (a-1, b-1, c-1) SEM surface morphology of PVMCGSA-0, PVMCGSA-3, and PVMCGSA-5 with the magnification factor of 10000×; (a-2, b-2, c-2) AFM images of PCMCGSA-0, PVMCGSA-3, and PVMCGSA-5 with the scanning area of 5 μm × 5 μm.

interlayer can be observed, which well supports the dense polysaccharide active layer around 5.0 ± 0.6 μm. The thickness of the active layer is positively correlated with the concentration of polysaccharide solution utilized for dip-coating. The specific reason can be ascribed to the following aspects. In a single dipcoating process, the weight of the polysaccharide solution (including the weight of solute and solvent) adhered to the supporting membrane is constant under the action of gravity, so the increase in concentration of polysaccharide solution signifies the enhanced adhesion amount of polysaccharide, which leads to a thicker active layer after thorough evaporation of the solvent (water). 3.3. Effect of SA Concentration on Cross-Section Morphology. In the dip-coating process, the concentration of polysaccharide solution will influence the structure and morphology of the active layer to a large extent. Figure 7 illustrates that the excess concentration of polysaccharide solution (4 wt % SA aqueous solution) will lead to an overly thick and dense active layer of the composite membrane around 13 ± 0.8 μm. This will bring several disadvantages for the composite membrane such as an increase in mass transfer resistance. In Figure 7a-2, an obvious crack continuously

Table 2. Outer Surface Roughness Values of the Composite Membranes code

Ra (nm)

Rm (nm)

PVMCGSA-0 PVMCGSA-3 PVMCGSA-5

12.4 9.87 8.83

211 112 105

temperature of 25 °C and a humidity of 50%) during the active layer forming stage. The slow and uniform water evaporation from uniformly distributed free volumes contributes to the smoother surface of membrane with branched carbohydrate chains than that with linear ones. The cross sections of PVMCGSA-3 can be clearly observed to illustrate the composite structure by SEM images in Figure 6. It can be observed that the supporting α-Al2O3 ceramic hollow fiber membrane with inner and outer diameters of 3 and 4 mm, respectively, demonstrates uniform hollow fiber shape as well as unexceptionable rigidity. This rigid structure can well support the polysaccharide layer to overcome the deformation under high vacuum degree inside the lumen of the hollow fiber membrane in the PV process. From Figure 6b,c, a distinct 27248

DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253

Research Article

ACS Applied Materials & Interfaces

Figure 6. Cross-section morphology of PVMCGSA-3 with magnification factors of (a) 30×, (b) 2000×, and (c) 10000×.

Figure 7. SEM images of the composite membranes with the obvious defect caused by the inappropriate concentration of the polysaccharide solution used for dip-coating: (a-1, a-2, a-3) cross section of the membrane in which the dip-coating solution contains 4 wt % sodium alginate; (b-1, b-2, b-3) surface morphology and cross section of the membrane in which the dip-coating solution contains only 1 wt % sodium alginate.

distributed on the whole cross-section can be observed between the supporting membrane and the pore filler adhering to the active layer. The generation of this crack is due to the lowtemperature brittle fracture of the composite membrane for the SEM sample in the liquid nitrogen. Under rapid cooling condition, the materials are fragile, especially for the titania nanoparticle layer with limited cohesion of small quantities of SA. The excessively dense active layer will generate great tension during its thermal contraction in the liquid nitrogen. This stress concentrates on the most fragile region on the cross section of the membrane, which exactly locates at the interfacial region between the supporting membrane and the pore fillers. After the active layer separates from the supporting membrane, the composite membrane undergoes structural failure and loses

its significance, such as the function of restricting the swelling dimensionality of the active layer. On the other hand, if the SA concentration of the polysaccharide solution is reduced to 1 wt %, the active layer will fill into the pores of the supported membrane as observed in Figure 7b-2,b-3 and leads to large numbers of defects on the surface of the active layer as illustrated in Figure 7b-1. Therefore, an inappropriate concentration of the dip-coating solution will negatively affect the formation of the active layer, so the optimization of the interlayer between the supporting membrane and the active layer as well as the concentration of dip-coating solution are of great importance to the formation of a defect-free active layer. In this work, the combination of the particular α-Al2O3 ceramic membrane, the TiO2 filling pore interlayer treatment, and the 27249

DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253

Research Article

ACS Applied Materials & Interfaces dip-coating solution with 2 wt % sodium alginate effectively contribute to the formation of a defect-free and homogeneous active layer with the precise thickness observed in Figure 6. 3.4. Free Volume Variation in Active Layer. XRD patterns of the polysaccharide active layer are displayed in Figure 8. The diffraction peaks at 2-θ of 13.6° and 21.7° are the

Figure 9. Dynamic water contact angle of the membranes.

3.6. Pervaporation and Swelling Performance. Figure 10 illustrates in detail the effect of water content (10 and 25 wt %), temperature (from 45 to 75 °C), and glycogen additive amount on the PV performance of PVMCGSA-3 and PVMCGSA-0. When the water content in the feed solution is relatively low (10 wt % in this work), both the flux and the separation factor of PVMCGSA-3 and PVMCGSA-0 present near-linear regularity at the operating temperature from 45 to 75 °C. This is because the polysaccharide layer of composite membranes will not fully swell in a feed solution that has a low water content. Thus, the free volume between the carbohydrate chains expands mildly with the increasing temperature, which ensures the separation factor of the PV membrane, but when the water content is changed to a high level (25 wt % in this work), the change regularity of flux and separation factor demonstrates the distinct difference between PVMCGSA-3 and PVMCGSA-0 in Figure 10c,d. Although the separation factors of PVMCGSA-3 and PVMCGSA-0 both decrease with increasing temperature, the decrease rate of the separation factor in PVMCGSA-3 slows while the one in PVMCGSA-0 accelerates, respectively. The difference in downtrend is attributed to intrinsic properties of polysaccharide active layer, and the incorporation of glycogen is beneficial to PVMCGSA-3 under a high water content situation for several reasons. The branched carbohydrate chain structure of glycogen compresses the free volume and increases chain density, which prevents the active layer from excessive swelling under harsh operating conditions as proved above. However, when it comes to PVMCGSA-0 with a single SA active layer cross-linked with GA, the incomplete cross-linking of linear carbohydrate chain is unable to resist excessive swelling under harsh operating conditions. Meanwhile, the expansion of free volume with the rising temperature will further aggravate the swelling of PVMCGSA-0. When the operating temperature reaches 75 °C, although the flux is as high as approximately 2200 ± 57 g m−2 h−1, the reduced separation factor around 10 ± 1 is unfavorable for organic PV membrane. On the contrary, PVMCGSA-3 demonstrates outstanding PV performance of flux at 1250 ± 21 g m−2 h−1 coupled with the separation factor of 187 ± 8. The effect of glycogen additive amount on PV performance under 75 °C and 25 wt % water is presented in Figure 10e. SA membrane with a small addition of glycogen has significant improvement in the separation ability with a slight

Figure 8. XRD pattern of the polysaccharide active layer.

characteristic peaks of sodium alginate.38,39 From the figure it can be found that the intensity of these peaks decreases regularly with the increasing glycogen additive amount, and it can be proved that the total amount of the d-spacing related to linear polymer chains of SA decreases because the branched carbohydrate chains of glycogen will increase the chain density to cover a certain amount of d-spacing in SA matrix. Because the active layer is completely composed of carbohydrate chains, the free volume is definitely correlated with the d-spacing variation. As the chain density of SA increased with the crosslinking of branched chains, the free volume existing between these linear chains of SA will absolutely be compressed. Therefore, the decrease in free volume and the increase in carbohydrate chain density will further promote the swelling resistance of the active layer, which will improve the separation performance of the active layer under harsh operating conditions. 3.5. Active Layer Hydrophilicity. The hydrophilicity of the composite membranes’ active layer expressed in dynamic water contact angle is shown in Figure 9. It can be observed that both the dynamic and the initial water contact angles enlarge with the increase in glycogen additive amount, which indicates the membrane hydrophilicity declines with the increase in glycogen additive amount. These phenomena are in accordance with the FTIR and XRD results mentioned above, that the incorporation of glycogen leads to a higher cross-linking degree in a hybrid polysaccharide PV membrane than in a single SA membrane. The increased cross-linking degree results in an apparent decrease in the relative amount of hydroxyl groups, which is closely related to the hydrophilicity. In addition, the branched carbohydrate chain structure in the hybrid polysaccharide PV membrane makes the chain denser, which further reduces the water absorption rate of the active layer according to the curves of dynamic water contact angle in Figure 9. 27250

DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253

Research Article

ACS Applied Materials & Interfaces

Figure 10. Pervaporation performance of (a, c) PVMCGSA-3 under 10 and 25 wt % water, respectively, (b, d) PVMCGSA-0 under 10 and 25 wt % water, respectively, under the temperature range from 45 to 75 °C and (e, f) PV performance and swelling degree according to the glycogen additive amount.

Table 3. PV Performance Comparison with the Alginate-Based PV Membrane in the Literature EtOH/H2O (wt %) 90:10 96:4 90:10 90:10 75:25

supporting membrane PAN PAN α-Al2O3 ceramic membrane

active layer

modification (cross-linking)

flux (kg m−2 h−1)

separation factor

temperature (°C)

reference

SA SA/PVP SA SA SA

zwitterionic graphene oxides phosphoric acid attapulgite nanorods zeolite 4A glycogen

2.140 0.500 1.356 0.106 1.250

1370 364 2030 396 187

77 30 76 25 75

40 28 39 41 this work

27251

DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253

Research Article

ACS Applied Materials & Interfaces

of Chemical Engineering (SKL-ChE-14C03) for giving financial support.

decrease in water flux because the hybrid polysaccharide membrane with compacted chain density still preserves a vast amount of free volumes between the carbohydrate chains for the generation of nanochannels. On the other hand, despite the consumption of hydroxyl groups cross-linked with GA for connection of different carbohydrate chains, there exist large amounts of free hydroxyl groups and carboxyl groups on sugar moieties of hybrid polysaccharide, which provides enough water affinity to promote the transportation of water molecules through the nanochannels in the active layer. Table 3 provides PV performance comparison between the alginate-based membrane in the literature and this work. From a comparison of the PV performance of PVMCGSA-3 with those in the representative literature in Table 3, it can be learned that PVMCGSA-3 has advantages in harsh operating conditions benefiting from the swell-resistant property. Due to the moderate swelling of the modified polysaccharide active layer under high temperature and high water content feed solution, the flux of PVMCGSA-3 is higher with the appropriate separation factor. Although the alginate-based membranes modified by particular nanofillers demonstrate much better PV performance, especially the graphene-doped one, PVMCGSA-3 still possesses the potential for further modification to improve the performance because the active layer of PVMCGSA-3 is still nearly pure polysaccharide in chemical structure (composed of sugar moieties) but endowed with the swelling-resistant property, which has advantages to serve as membrane matrix to load particular nanofillers compared to the original sodium alginate.



4. CONCLUSIONS The ceramic-supported and glycogen cross-linked sodium alginate membrane demonstrates excellent swelling resistance originates from optimization of the composite structure and the intrinsic properties of the chemically modified active layer. The rigid ceramic membrane well supports the polysaccharide layer, whereas the branched-chain structure compresses the free volume, and the cross-linking between different carbohydrate chains consumes the excess free hydroxyl groups. The composite membrane with 3 wt % glycogen cross-linked in the active layer demonstrates controllable PV performance in the ethanol−water solution containing high water content (25 wt %) at 45−75 °C. The flux increases from 770 to 1250 g m−2 h−1 while the separation factor decreases from 374 to 187, and the in situ swelling degree of the polysaccharide active layer is only 18.6% under 75 °C, which endows this kind of membrane with great potential for the dehydration of ethanol−water solution containing high water content.



REFERENCES

(1) Chapman, P. D.; Oliveira, T.; Livingston, A. G.; Li, K. Membranes for the Dehydration of Solvents by Pervaporation. J. Membr. Sci. 2008, 318 (1−2), 5−37. (2) Liu, Q.; Huang, B.; Huang, A. Polydopamine-based Superhydrophobic Membranes for Biofuel Recovery. J. Mater. Chem. A 2013, 1 (38), 11970. (3) Wang, X.-S.; Ji, Y.-L.; Zheng, P.-Y.; An, Q.-F.; Zhao, Q.; Lee, K.R.; Qian, J.-W.; Gao, C.-J. Engineering Novel Polyelectrolyte Complex Membranes with Improved Mechanical Properties and Separation Performance. J. Mater. Chem. A 2015, 3 (14), 7296−7303. (4) Peng, Y.; Lu, H.; Wang, Z.; Yan, Y. Microstructural Optimization of MFI-Type Zeolite Membranes for Ethanol−Water Separation. J. Mater. Chem. A 2014, 2 (38), 16093−16100. (5) Ngamou, P. H. T.; Overbeek, J. P.; Kreiter, R.; van Veen, H. M.; Vente, J. F.; Wienk, I. M.; Cuperus, P. F.; Creatore, M. PlasmaDeposited Hybrid Silica Membranes with a Controlled Retention of Organic Bridges. J. Mater. Chem. A 2013, 1 (18), 5567. (6) Zuo, J.; Shi, G. M.; Wei, S.; Chung, T. S. The Development of Novel Nexar Block Copolymer/Ultem Composite Membranes for C2C4 Alcohols Dehydration via Pervaporation. ACS Appl. Mater. Interfaces 2014, 6 (16), 13874−13883. (7) 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 (34), 9814. (8) Wee, S.-L.; Tye, C.-T.; Bhatia, S. Membrane Separation ProcessPervaporation through Zeolite Membrane. Sep. Purif. Technol. 2008, 63 (3), 500−516. (9) Toti, U. S.; Aminabhavi, T. M. Different Viscosity Grade Sodium Alginate and Modified Sodium Alginate Membranes in Pervaporation Separation of Water + Acetic Acid and Water + Isopropanol Mixtures. J. Membr. Sci. 2004, 228 (2), 199−208. (10) Badiger, H.; Shukla, S.; Kalyani, S.; Sridhar, S. Thin Film Composite Sodium Alginate Membranes for Dehydration of Acetic Acid and Isobutanol. Appl. Polym. Sci. 2014, 131 (6), 40018. (11) Zhang, W.; Xu, Y.; Yu, Z.; Lu, S.; Wang, X. Separation of Acetic Acid/Water Mixtures by Pervaporation with Composite Membranes of Sodium Alginate Active Layer and Microporous Polypropylene Substrate. J. Membr. Sci. 2014, 451, 135−147. (12) Xue, S. M.; Xu, Z. L.; Tang, Y. J.; Ji, C. H. Polypiperazine-Amide Nanofiltration Membrane Modified by Different Functionalized Multiwalled Carbon Nanotubes (MWCNTs). ACS Appl. Mater. Interfaces 2016, 8 (29), 19135−19144. (13) Sajjan, A. M.; Jeevan Kumar, B. K.; Kittur, A. A.; Kariduraganavar, M. Y. Novel Approach for the Development of Pervaporation Membranes Using Sodium Alginate and Chitosanwrapped Multiwalled Carbon Nanotubes for the Dehydration of Isopropanol. J. Membr. Sci. 2013, 425−426, 77−88. (14) Yeh, T.-M.; Wang, Z.; Mahajan, D.; Hsiao, B. S.; Chu, B. High Flux Ethanol Dehydration Using Nanofibrous Membranes Containing Graphene Oxide Barrier Layers. J. Mater. Chem. A 2013, 1 (41), 12998. (15) Liang, B.; Zhan, W.; Qi, G.; Lin, S.; Nan, Q.; Liu, Y.; Cao, B.; Pan, K. High Performance Graphene Oxide/Polyacrylonitrile Composite Pervaporation Membranes for Desalination Applications. J. Mater. Chem. A 2015, 3 (9), 5140−5147. (16) Su, Z.; Chen, J. H.; Sun, X.; Huang, Y.; Dong, X. AmineFunctionalized Metal Organic Framework (NH2-MIL-125(Ti)) Incorporated Sodium Alginate Mixed Matrix Membranes for Dehydration of Acetic Acid by Pervaporation. RSC Adv. 2015, 5 (120), 99008−99017. (17) Fan, H.; Wang, N.; Ji, S.; Yan, H.; Zhang, G. Nanodisperse ZIF8/PDMS Hybrid Membranes for Biobutanol Permselective Pervaporation. J. Mater. Chem. A 2014, 2 (48), 20947−20957.

AUTHOR INFORMATION

Corresponding Author

*(Z.-L.X.) E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We gratefully acknowledge the National Natural Science Foundation of China (20076009, 21176067, 21276075, and 21406060), the Shanghai Sailing Program (14YF1404800), the Fundamental Research Funds for the Central Universities (WA1514305), and the Open Project of State Key Laboratory 27252

DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253

Research Article

ACS Applied Materials & Interfaces

(35) Violi, I. L.; Zelcer, A.; Bruno, M. M.; Luca, V.; Soler-Illia, G. J. Gold Nanoparticles Supported in Zirconia-Ceria Mesoporous Thin Films: a Highly Active Reusable Heterogeneous Nanocatalyst. ACS Appl. Mater. Interfaces 2015, 7 (2), 1114−1121. (36) Hecht, H.; Srebnik, S. Structural Characterization of Sodium Alginate and Calcium Alginate. Biomacromolecules 2016, 17 (6), 2160−2167. (37) Chan, A. W.; Whitney, R. A.; Neufeld, R. J. Kinetic Controlled Synthesis of PH-Responsive Network Alginate. Biomacromolecules 2008, 9, 2536−2545. (38) Li, Y.; Jia, H.; Pan, F.; Jiang, Z.; Cheng, Q. Enhanced AntiSwelling Property and Dehumidification Performance by Sodium Alginate−Poly(Vinyl Alcohol)/Polysulfone Composite Hollow Fiber Membranes. J. Membr. Sci. 2012, 407−408, 211−220. (39) Xing, R.; Pan, F.; Zhao, J.; Cao, K.; Gao, C.; Yang, S.; Liu, G.; Wu, H.; Jiang, Z. Enhancing the Permeation Selectivity of Sodium Alginate Membrane by Incorporating Attapulgite Nanorods for Ethanol Dehydration. RSC Adv. 2016, 6 (17), 14381−14392. (40) Zhao, J.; Zhu, Y.; He, G.; Xing, R.; Pan, F.; Jiang, Z.; Zhang, P.; Cao, X.; Wang, B. Incorporating Zwitterionic Graphene Oxides into Sodium Alginate Membrane for Efficient Water/Alcohol Separation. ACS Appl. Mater. Interfaces 2016, 8 (3), 2097−2103. (41) Nigiz, F. U.; Dogan, H.; Hilmioglu, N. D. Pervaporation of Ethanol/Water Mixtures Using Clinoptilolite and 4A Filled Sodium Alginate Membranes. Desalination 2012, 300, 24−31.

(18) Li, Y.; Wee, L. H.; Martens, J. A.; Vankelecom, I. F. J. ZIF-71 as a Potential Filler to Prepare Pervaporation Membranes for Bio-Alcohol Recovery. J. Mater. Chem. A 2014, 2 (26), 10034−10040. (19) Ji, C.-H.; Xue, S.-M.; Xu, Z.-L.; Ma, X.-H. Fabrication and Characterization of Novel Hollow Fiber Catalytic Packing of PFSA− PES−ZrO2 (shell)−TiO2 (core) Solid Superacid via Wet-Spinning Method. Chem. Eng. Sci. 2016, 152, 1−11. (20) He, J.; Chen, J.; Ren, L.; Wang, Y.; Teng, C.; Hong, M.; Zhao, J.; Jiang, B. Fabrication of Monodisperse Porous Zirconia Microspheres and Their Phosphorylation for Friedel-Crafts Alkylation of Indoles. ACS Appl. Mater. Interfaces 2014, 6 (4), 2718−2725. (21) Lu, P. P.; Xu, Z. L.; Yang, H.; Wei, Y. M. Processing-StructureProperty Correlations of Polyethersulfone/Perfluorosulfonic Acid Nanofibers Fabricated via Electrospinning from Polymer-Nanoparticle Suspensions. ACS Appl. Mater. Interfaces 2012, 4 (3), 1716−1723. (22) Teli, S. B.; Gokavi, G. S.; Sairam, M.; Aminabhavi, T. M. Highly Water Selective Silicotungstic Acid (H4SiW12O40) Incorporated Novel Sodium Alginate Hybrid Composite Membranes for Pervaporation Dehydration of Acetic Acid. Sep. Purif. Technol. 2007, 54 (2), 178− 186. (23) Magalad, V. T.; Supale, A. R.; Maradur, S. P.; Gokavi, G. S.; Aminabhavi, T. M. Preyssler Type Heteropolyacid-Incorporated Highly Water-selective Sodium Alginate-based Inorganic−Organic Hybrid Membranes for Pervaporation Dehydration of Ethanol. Chem. Eng. J. 2010, 159 (1−3), 75−83. (24) Liu, G.; Hung, W.-S.; Shen, J.; Li, Q.; Huang, Y.-H.; Jin, W.; Lee, K.-R.; Lai, J.-Y. Mixed Matrix Membranes with Molecular-InteractionDriven Tunable Free Volumes for Efficient Bio-Fuel Recovery. J. Mater. Chem. A 2015, 3 (8), 4510−4521. (25) Rao, P. S.; Krishnaiah, A.; Smitha, B.; Sridhar, S. Separation of Acetic Acid/Water Mixtures by Pervaporation through Poly(Vinyl Alcohol)−Sodium Alginate Blend Membranes. Sep. Sci. Technol. 2006, 41 (5), 979−999. (26) Chen, J. H.; Liu, Q. L.; Zhu, A. M.; Zhang, Q. G. Dehydration of Acetic Acid by Pervaporation Using SPEK-C/PVA Blend Membranes. J. Membr. Sci. 2008, 320, 416−422. (27) Rachipudi, P. S.; Kittur, A. A.; Sajjan, A. M.; Kamble, R. R.; Kariduraganavar, M. Y. Solving the Trade-off Phenomenon in Separation of Water−Dioxan Mixtures by Pervaporation through Crosslinked Sodium−Alginate Membranes with Polystyrene Sulfonic Acid-co-Maleic Acid. Chem. Eng. Sci. 2013, 94, 84−92. (28) Kalyani, S.; Smitha, B.; Sridhar, S.; Krishnaiah, A. Separation of Ethanol-Water Mixtures by Pervaporation Using Sodium Alginate/ Poly(Vinyl Pyrrolidone) Blend Membrane Crosslinked with Phosphoric Acid. Ind. Eng. Chem. Res. 2006, 45, 9088−9095. (29) Veerapur, R. S.; Gudasi, K. B.; Sairam, M.; Shenoy, R. V.; Netaji, M.; Raju, K. V. S. N.; Sreedhar, B.; Aminabhavi, T. M. Novel Sodium Alginate Composite Membranes Prepared by Incorporating Cobalt (III) Complex Particles Used in Pervaporation Separation of Water− Acetic Acid Mixtures at Different Temperatures. J. Mater. Sci. 2007, 42 (12), 4406−4417. (30) Huang, R. Y. M.; Pal, R.; Moon, G. Y. Characteristics of Sodium Alginate Membranes for the Pervaporation Dehydration of EthanolWater and Isopropanol Mixtures. J. Membr. Sci. 1999, 160, 101−113. (31) Han, L.-F.; Xu, Z.-L.; Cao, Y.; Wei, Y.-M.; Xu, H.-T. Preparation, Characterization and Permeation Property of Al2O3, Al2O3−SiO2 and Al2O3−Kaolin Hollow Fiber Membranes. J. Membr. Sci. 2011, 372 (1−2), 154−164. (32) Peters, T. A.; Benes, N. E.; Keurentjes, J. T. F. Hybrid CeramicSupported Thin PVA Pervaporation Membranes: Long-Term Performance and Thermal Stability in the Dehydration of Alcohols. J. Membr. Sci. 2008, 311 (1−2), 7−11. (33) Jadav, G. L.; Aswal, V. K.; Singh, P. S. Preparation of Bifunctional Poly(Dimethylsiloxane) Membrane by Dual X-Linking. J. Mater. Chem. A 2013, 1 (15), 4893. (34) Ko, Y. N.; Choi, S. H.; Kang, Y. C.; Park, S. B. Electrochemical Properties of ZrO2-Doped V2O5 Amorphous Powders with Spherical Shape and Fine Size. ACS Appl. Mater. Interfaces 2013, 5 (8), 3234− 3240. 27253

DOI: 10.1021/acsami.6b10053 ACS Appl. Mater. Interfaces 2016, 8, 27243−27253