Biomacromolecules 2004, 5, 1567-1574
1567
Dehydration of an Ethanol/Water Azeotrope by Novel Organic-Inorganic Hybrid Membranes Based on Quaternized Chitosan and Tetraethoxysilane Tadashi Uragami,*,† Takuya Katayama,† Takashi Miyata,† Hiroshi Tamura,† Tadashi Shiraiwa,† and Akon Higuchi‡ Faculty of Engineering and High Technology Research Center, Kansai University, Suita, Osaka 564-8680, Japan, and Department of Industrial Chemistry, Faculty of Engineering, Seikei University, Musashino, Tokyo 180-8633, Japan Received February 23, 2004; Revised Manuscript Received April 13, 2004
To control swelling of quaternized chitosan (q-Chito) membranes, mixtures of q-Chito as an organic component and tetraethoxysilane (TEOS) as an inorganic component were prepared using the sol-gel reaction, and novel q-Chito/TEOS hybrid membranes were formed. In the separation of an ethanol/water azeotrope by pervaporation, the effect of TEOS content on the water/ethanol selectivity of q-Chito/TEOS hybrid membranes was investigated. Hybrid membranes containing up to 45 mol % TEOS exhibited higher water/ ethanol selectivity than the q-Chito membrane. This resulted from depressed swelling of the membranes by formation of a cross-linked structure. However, introduction of excess TEOS led to greater swelling of the hybrid membranes. Therefore, the water/ethanol selectivity of the hybrid membranes containing more than 45 mol % TEOS was lower than that of the q-Chito membrane. The relationship between the structure of q-Chito/TEOS hybrid membranes and their permeation and separation characteristics during pervaporation of an ethanol/water azeotrope is discussed in detail. Introduction Membrane separation techniques have attracted considerable attention for industrial applications because they can be highly efficient and low cost. Pervaporation is a type of membrane separation technology used to treat multicomponent liquid solutions.1-5 It is especially promising for separation of organic liquid mixtures, such as azeotropic5,6 or close-boiling point mixtures.7,8 Many polymeric materials have been investigated as pervaporation membranes for the separation of the ethanol/ water azeotrope in high-performance dehydration systems.9-11 Chitosan, an amino polysaccharide, is a promising hydrophilic membrane material that contains many reactive amino and hydroxyl groups. Previously, we have reported that chemical modifications of chitosan, such as cross-linking and N-alkylation, can enhance the water/ethanol selectivity of the membranes.11-13 In particular, quaternized chitosan (qChito) membranes prepared by N-alkylation of chitosan showed high water/ethanol selectivity during pervaporation. However, excess quaternization leads to increased hydrophilicity, and, therefore, increased swelling of the q-Chito membranes. As a result, these membranes showed reduced water/ethanol selectivity. Therefore, control of the physical structure of the q-Chito membranes is required to improve their water/ethanol selectivity during pervaporation. Organic-inorganic hybrid materials are viewed as next generation materials in many applications because they have * To whom correspondence showed be addressed. † Kansai University. ‡ Seikei University.
both the film-forming properties of a polymer and the stability of an inorganic compound. Previously, we prepared organic-inorganic hybrid membranes by the sol-gel reaction of poly (vinyl alcohol) and tetraethoxysilane (TEOS) to improve the selectivity of polymer membranes.14 In this study, we prepared novel organic-inorganic hybrid membranes again using the sol-gel reaction. It is well-known that quaternized chitosan membranes are highly water/ethanol selective during pervaporation of aqueous ethanol solutions.2 However, swelling of the q-Chito membrane in an aqueous ethanol solution results in an increase in both solubility and diffusivity of ethanol, which leads to lower water/ethanol selectivity.2 In this work, to control swelling of q-Chito membranes, mixtures of q-Chito and TEOS were prepared by the sol-gel reaction and fabricated into q-Chito/TEOS hybrid pervaporation membranes. The relationship between the structure of the q-Chito/TEOS hybrid membranes and their permeation and separation characteristics for an azeotropic mixture of ethanol/water by pervaporation is discussed in detail. Experimental Section Materials. Chitosan, with 100% deacetylation and an average molecular weight of (3-4) × 105 g/mol, was supplied by Koyo Chemical Co. Ltd., Japan. The synthesis of quaternized chitosan (q-Chito) was performed according to a previously described procedure.15 The purified chitosan powder (6.0 g) was dispersed into an aqueous solution of 42 wt % methanol (760 mL) at room temperature, and a
10.1021/bm0498880 CCC: $27.50 © 2004 American Chemical Society Published on Web 05/22/2004
1568
Biomacromolecules, Vol. 5, No. 4, 2004
Uragami et al.
Scheme 1. Sol-Gel Reaction of q-Chito with TEOS
desired amount of methyl iodide was added to this heterogeneous mixture and stirred at 50 °C. The quaternization of chitosan was then carried out at 50 °C for 6 h. Excess sodium chloride was then added to the reaction mixture to convert the iodide ammonium salt to chloride ammonium salt. This solution was poured into excess acetone to precipitate the q-Chito. The crude q-Chito was dissolved in pure water and purified by repeatedly precipitating the polymer in acetone. The degree of quaternization of the purified q-Chito was determined by its 270 MHz 1H nuclear magnetic resonance (1H NMR) (JEOL; EX-270) spectra. In this study, q-Chito with 18% quaternization was synthesized. All other reagents and solvents used in this study were supplied by commercial sources. Preparation of q-Chito/TEOS Hybrid Membranes. The organic-inorganic hybrid membranes were prepared by the sol-gel reaction of q-Chito with TEOS (Scheme 1) as follows: q-Chito was dissolved in DMSO by stirring for 12 h at 80 °C, after which TEOS was added to the q-Chito solution at room temperature and the mixture was then stirred for 30 h. An aqueous HCl solution as a catalyst was homogeneously mixed with the above solution containing q-Chito and TEOS. The q-Chito/TEOS hybrid membranes were prepared by pouring the casting solution onto Teflon plates, and then allowing the solvent to evaporate completely at 80 °C for 40 h. The resulting membranes were transparent and defect-free. Permeation Measurements. Pervaporation experiments were carried out using the apparatus described in previous studies16-25 under the following conditions: permeation temperature, 40 °C; pressure on the permeate side, 1 × 10-2 Torr. The effective membrane area was 13.8 cm2. An aqueous solution of 96.5 wt % ethanol was used as the feed solution. The permeate was condensed in a U-tube under liquid nitrogen temperature. The permeation rate (kg/(m2 h)) for an azeotrope of ethanol/water by pervaporation was determined from the weight (kg) of permeate collected in a cold trap, the permeation time (h), and the effective membrane area (m2). To provide a comparison of the permeation rates of membranes with different thickness, the normalized permeation rate (kg m/(m2 h)), which is the product of the permeation rate and the membrane thickness, was used. The permeation results for aqueous ethanol feed solution by pervaporation were reproducible, and the errors inherent in these permeation measurements ranged within a few percent. Permeation results reported in this study were measured after reaching steady-state flux. The compositions
of the feed solution and permeate were determined by a gas chromatograph (Shimadzu GC-9A) equipped with a flame ionization detector (FID) and a capillary column (Shimadzu Co. Ltd.; Shimalite F) heated to 200 °C. The separation factor, Rsep H2O/EtOH, during pervaporation was calculated from eq 1 Rsep H2O/EtOH ) (PH2O/PEtOH)/(FH2O/FEtOH)
(1)
where FH2O and FEtOH and PH2O and PEtOH are the weight fractions of water and ethanol in feed solution and the permeate, respectively. Composition of the Solution Absorbed in q-Chito/ TEOS Hybrid Membranes. The q-Chito/TEOS hybrid membranes were dried completely under reduced pressure at room temperature and weighed. The dried membranes were immersed into an azeotropic mixture of ethanol/water in a sealed vessel at 40 °C until equilibrium was reached. A large amount of the swollen q-Chito/TEOS hybrid membranes was placed in a container, as shown in Figure 1. The system in Figure 1 was evacuated, and the container with the swollen membranes was heated. The solution sorbed into the swollen membranes was completely desorbed under reduced pressure and was collected in a U-tube cooled with liquid nitrogen. The composition of the solutions sorbed in the q-Chito/TEOS hybrid membranes was then determined by measuring the ethanol concentration in the collected solution by gas chromatography (Shimazu GC-14A). The ethanol/water composition in the q-Chito/TEOS hybrid
Figure 1. Diagram of the apparatus to determine the composition of the solution absorbed in q-Chito/TEOS hybrid membranes.
Biomacromolecules, Vol. 5, No. 4, 2004 1569
Dehydration of an Ethanol/Water Azeotrope
membranes and that in the feed solution yielded the sorption selectivity, Rsorp H2O/EtOH, as expressed in eq 2 Rsorp H2O/EtOH ) (MH2O/MEtOH)/(FH2O/FEtOH)
(2)
where FH2O and FEtOH are the weight fractions of water and ethanol in the feed solution and MH2O and MEtOH are those sorbed in the membrane, respectively. Measurement of Contact Angle. The contact angles of chloroform on the surfaces of the q-Chito/TEOS hybrid membranes were measured by a contact angle meter (Erma, model G-1) at 25 °C. The contact angle, θ, was calculated from eq 326,27 θ ) cos-1[(cos θa + cos θr)/2]
(3)
where θa and θr are the advancing contact angle and the receding contact angle, respectively. Degree of Membrane Swelling. The dried q-Chito/TEOS hybrid membranes were weighed and immersed into an azeotrope mixture of ethanol/water in a sealed vessel at 40 °C until equilibrium was reached. The membranes were then taken out of the vessel, wiped quickly with a filter paper, and weighed. The degree of swelling (DS) of the q-Chito/ TEOS hybrid membranes was then determined from eq 4 DS ) Ws/Wd
(4)
where Ws is the weight of the membrane swollen in an azeotropic mixture of ethanol/water and Wd is the weight of the dried membrane. Membrane Density. The density of the q-Chito/TEOS hybrid membranes was determined by measuring their weights in air and n-hexane with an electric gravity meter (Mirage Boeki, SD-120L) at 25 °C. FT-IR Measurements. The IR spectra of the q-Chito/ TEOS hybrid membranes were measured by FT-IR (PerkinElmer, FT-IR spectrometer PARAGON 1000). Differential Scanning Calorimetry (DSC). DSC measurements were carried out by means of DSC equipment (Seiko Instruments, DSC 6220). The specimens were heated from about -150 to +300 °C at heating rate of 5 °C/min. Cross-Link Density of q-Chito/TEOS Hybrid Membranes. The cross-link density, F, of the q-Chito/TEOS hybrid membranes was calculated from the network theory of rubber elasticity given in eq 528-30 F ) E′/3dφRT
(5)
where the modulus, E,′ was determined from measurements with a dynamic mechanical analyzer (Rheogel-E4000 F3, U.B.M) under the following conditions: frequency, 1, 2, 4, 10 Hz; temperature, 40 °C; d is membrane density, φ is the front factor (where φ ) 1), R is the gas constant, and T is the absolute temperature. Transmission Electron Micrographs (TEM). The qChito/TEOS hybrid membranes were embedded in epoxy resin and sliced into thin films (thickness ≈50 nm) with a microtome (Leica; Reichert Ultracut E). The structure of the
Figure 2. Effects of the TEOS content on the normalized permeation rate (O) and the ethanol concentration in the permeate (b) during pervaporation of an azeotrope (96.5 wt %) of ethanol/water through q-Chito/TEOS hybrid membranes.
membranes was observed by a transmission electron microscope (TEM) (JEOL JEM-1210) at an accelerating voltage of 80 kV. Results and Discussion Effect of TEOS Content on Pervaporation Characteristics of q-Chito/TEOS Hybrid Membranes. We previously reported the effects of quaternization of chitosan membranes on the permeation and separation characteristics for an azeotropic mixture of ethanol/water by pervaporation.2 With increasing degree of quaternization, the permeation rate decreased; however, the water/ethanol selectivity increased. On the other hand, the degree of swelling of the q-Chito membrane increased slightly, whereas the density of the q-Chito membrane decreased with an increased degree of quaternization. From these results, we concluded that quaternization of chitosan can lead to an increase in the water/ ethanol selectivity for an ethanol/water azeotrope.2 However, the water/ethanol separation factor was not particularly high. To enhance the water/ethanol selectivity of the q-Chito membrane, swelling of the membrane must be depressed. In this study, our strategy was to depress swelling of the q-Chito membrane by introduction of the inorganic component, TEOS. Figure 2 shows the effects of TEOS content in the q-Chito/ TEOS hybrid membranes on the normalized permeation rate and the ethanol concentration in the permeate for an azeotrope of ethanol/water during pervaporation. The TEOS content in mol % in the figure is for the pyranose ring in the q-Chito molecule. The permeate ethanol concentrations of all q-Chito/TEOS hybrid membranes were very low compared to that in the feed solution. These results suggest that the q-Chito/TEOS hybrid membranes showed high water/ethanol selectivity. High water/ethanol selectivity was observed in the q-Chito/TEOS hybrid membranes containing up to 45 mol % TEOS; however, it’s selectivity decreased when the TEOS content was further increased. The best performance was observed for the q-Chito/TEOS hybrid
1570
Biomacromolecules, Vol. 5, No. 4, 2004
Scheme 2. Hydrolysis and Condensation Reaction for TEOS
membrane with 20 mol % TEOS content where ethanol could not be detected in the permeate by gas chromatography. This result suggests that ethanol could be rejected almost perfectly by this q-Chito/TEOS hybrid membrane. The normalized permeation rate gradually decreased by increasing the TEOS content up to 45 mol % but then increased at higher TEOS contents. These results can be attributed to both the formation of a cross-linked structure and a difference in the state of the cross-linked network. Formation of q-Chito/TEOS Hybrid Membranes. Scheme 2 shows the process of the polycondensation reaction of TEOS.31 In the first step of the q-Chito/TEOS hybrid membrane preparation process, TEOS was hydrolyzed in the presence of an acid catalyst (A), which leads to formation of silanol groups. The resulting silanol groups yielded siloxane bonds due to a dehydration or dealcoholysis reaction (B) with other silanol groups or ethoxy groups during drying of the membrane. These reactions led to the formation of cohesive siloxane domains in the membrane. Because these siloxane domains were dispersed in the membrane, the silanol groups in the siloxane domains and the hydroxyl groups in q-Chito formed hydrogen and covalent bonds, which are the cross-link points, as illustrated in Scheme 3. In Figure 3, the FT-IR spectra of q-Chito and q-Chito/ TEOS hybrid membranes are shown. As seen in Figure 3,
Uragami et al.
the peak (1100-1210 cm-1) for the OH group in the q-Chito molecule decreased, but the peak (800-840 cm-1) of the Si-O-Si bond increased with increasing TEOS content. These results suggest the formation of a cross-linked structure due to the sol-gel reaction, in which the silanol groups in the tetrasilanol silane (TSS), produced by the hydrization of TEOS, reacted with the hydroxyl groups in the q-Chito molecule. Chemical Structure of q-Chito/TEOS Hybrid Membrane. In general, the selectivity of polymer membranes is dependent on their chemical and physical structures, which influence the solubility of permeants in the membrane and the diffusivity of permeants through the membrane. Therefore, we investigated the chemical and physical structures of the q-Chito/TEOS hybrid membranes to elucidate their permeation and separation characteristics, as shown in Figure 2. Figure 4 shows the ethanol concentration sorbed into q-Chito/TEOS hybrid membranes immersed in an aqueous solution of 96.5 wt % ethanol and the contact angle for chloroform as a function of the TEOS content in the hybrid membrane. The ethanol concentration sorbed into the qChito/TEOS hybrid membranes decreased with increasing TEOS content up to 45 mol % and then increased at higher TEOS content. The contact angle for chloroform increased with increasing TEOS content. This increase in the contact angle for chloroform, which is a hydrophobic solvent, suggests that the membrane surface of the q-Chito/TEOS hybrid membranes became more hydrophilic with an increase in the TEOS content. The decrease in the ethanol concentration sorbed in the q-Chito/TEOS hybrid membranes with increasing TEOS content up to 45 mol % is due to an increase in the hydrophilicity of the membrane surface. This suggests that the solubility of water in the q-Chito/TEOS hybrid membrane was improved by introducing TEOS into the q-Chito molecule. This analysis is supported by the
Figure 3. FT-IR spectra of the q-Chito/TEOS hybrid membranes. Scheme 3. Tentative Illustration for the Interaction between q-Chito and TEOS
Biomacromolecules, Vol. 5, No. 4, 2004 1571
Dehydration of an Ethanol/Water Azeotrope
Figure 4. Effects of the TEOS content on the ethanol concentration in q-Chito/TEOS hybrid membranes (b) immersed in 96.5 wt % ethanol at 40 °C, and the contact angle for CHCl3 (O) on the surface of these membranes.
Figure 5. Relationship between the TEOS content and Tg in the q-Chito/TEOS hybrid membranes.
Table 1. Separation Factor, Sorption Selectivity, and Diffusion Selectivity of q-Chito/TEOS Hybrid Membranes TEOS content (mol %)
Rsep. H2O/EtOH
Rsorp. H2O/EtOH
Rdiff. H2O/EtOH
0 10 20 45 60
726 3098 extremely high 35480 30
442 1061 1489 1754 70
1.6 2.9 considerably high 20 0.4
sorption selectivity, Rsorp H2O/EtOH, shown in Table 1, and determined using eq 2 and the data in Figure 4. As can be seen in Table 1, the water sorption selectivity of the q-Chito/TEOS hybrid membranes increased with TEOS content up to 45 mol % and then decreased remarkably at higher TEOS content. The diffusion selectivity was calculated from the separation factor (determined using eq 1 and data of Figure 2), and the sorption selectivity via eq 6 Rdiff
H2O/EtOH )
Rsep H2O/EtOH/Rsorp H2O/EtOH
(6)
The water/ethanol diffusion selectivity of the q-Chito/TEOS hybrid membranes was lower than the water/ethanol sorption selectivity, except for the hybrid membrane containing 20 mol % TEOS. Because the separation factor of the membrane with 20 mol % TEOS could not be determined, it was impossible to calculate the diffusion selectivity. However, we suggest that the diffusion selectivity of this hybrid membrane was very high, because the separation factor of this membrane is extremely high. This presumption suggests that the chemical and physical structures of the q-Chito/ TEOS hybrid membrane containing 20 mol % TEOS were optimized for the dehydration of an azeotropic mixture of ethanol/water by pervaporation. Physical Structure of the q-Chito/TEOS Hybrid Membrane. To characterize the cross-linked structure of the q-Chito/TEOS hybrid membranes, their glass-transition temperatures (Tg’s) were measured by DSC. Figure 5 shows the relationship between the TEOS content in the q-Chito/TEOS hybrid membranes and Tg’s of these membranes. Tg of the q-Chito/TEOS hybrid membranes increased linearly with increasing TEOS content. The increase in Tg suggests that the formation of cross-linked structure in the hybrid mem-
Figure 6. Effects of the TEOS content on the membrane density (O) and the degree of swelling (b) of q-Chito/TEOS hybrid membranes immersed in 96.5 wt % ethanol at 40 °C.
branes decreases the molecular chain mobility of q-Chito. Figure 6 shows the effects of the TEOS content on the membrane density and the degree of swelling of the q-Chito/ TEOS hybrid membranes. The density of q-Chito/TEOS hybrid membranes increased with increasing TEOS content. The degree of swelling decreased for hybrid membranes containing up to 45 mol % TEOS but increased at higher TEOS content. This increase in the degree of swelling of the hybrid membrane with excess TEOS may be attributed to a lower cross-link density of the q-Chito/TEOS hybrid membrane based on the formation of large cohesive TSS domains. To determine the cross-link density of the q-Chito/TEOS hybrid membranes, the elastic modulus of the membranes was measured by using a dynamic mechanical analyzer. In Figure 7, the relationship between the elastic modulus of the q-Chito/TEOS hybrid membrane and the TEOS content in the hybrid membrane is shown. The elastic modulus deceased with increasing TEOS content. The cross-link density of the q-Chito/TEOS hybrid membranes expressed in eq 5 was determined using the elastic modulus obtained from measurement of the dynamic viscoelasticity and the membrane density measured by the electric gravity meter. In Figure 7, the effect of TEOS content in q-Chito/TEOS hybrid membranes on the cross-link density of the membranes is also shown. The cross-link density also decreased with an increase in TEOS content. The decrease in the cross-link density with increasing TEOS content can
1572
Biomacromolecules, Vol. 5, No. 4, 2004
Uragami et al.
Figure 7. Effects of the TEOS content on the elastic modulus (O) and the cross-link density (b) of q-Chito/TEOS hybrid membranes.
be attributed to the fact that the formation of cohesive domains due to reactions between the TSS molecules was more predominant than the formation of cross-linking between q-Chito and TSS molecules. Therefore, when excess TEOS was added to the casting solution, the cross-link reaction is reduced, the elastic modulus decreases, and consequently the cross-link density also decreased. On the basis of the above results, we found that both the decrease in the water/ethanol selectivity and the increase in the normalized permeation rate at high TEOS content (see Figure 2) are related to the increase in swelling of the hybrid membranes due to a decrease in cross-link density. Performance of q-Chito/TEOS Hybrid Membranes. In Table 2, the permeation and separation characteristics of various water-selective polymer membranes when treating an aqueous ethanol solution by pervaporation are summarized. The separation factors for the q-Chito/TEOS hybrid membranes are significantly higher than those of other polymer membranes. However, the permeation rates of the q-Chito/TEOS hybrid membranes are lower. These results suggest that the introduction of TEOS to q-Chito through sol-gel hybridization formed advanced, highly selective membranes that may be useful for the dehydration of the ethanol/water azeotrope. However, in practice, it will be necessary to improve the permeation rate. If more permeable
Figure 8. TEM image of q-Chito/TEOS hybrid membranes containing 45 and 60 mol % of TEOS.
forms of the q-Chito and TEOS hybrid membrane are developed, and an ultrathin membrane is prepared, q-Chito/ TEOS hybrid membranes with both high permeation rate and high separation factor can be obtained. This study suggests that the preparation of organicinorganic hybrid membranes for the dehydration of an azeotropic mixture of ethanol/water is possible by minimizing swelling of the membranes, which is accomplished by controlling the TEOS content in the hybrid materials. Structure, Permeation, and Separation Characteristics of q-Chito/TEOS Hybrid Membranes. In Figure 8, TEM photographs of the q-Chito/TEOS hybrid membranes containing 45 and 60 mol % of TEOS are shown. In the TEM pictures, the TEOS phase appears dark. In the q-Chito/TEOS
Table 2. Permeation and Separation Characteristics of Various Polymer Membranes during Pervaporation of an Aqueous Ethanol Solution membrane
feed (wt %)
temp (°C)
permeation rate (kg/cm2 h)
separation factor (Rsep H2O/EtOH)
ref
cellophane cellulose acetate poly(tetrafluoroethylene)-g-poly(vinylpyrrolidone Nafion-H+-(CH2)2NH+ polyacrylonitrile-poly(vinylpyrrolidone)blend polystyrene poly(vinyl chloride) alginic acid chitosan chitsan acetate salt glutaraldehyde cross-linked chitosan PVA/TEOS hybrida PVA/TEOS hybridb q-Chito/TEOS hybridc q-Chito/TEOS hybridd
75.6 95.6 95.6 95.6 95.6 95.6 95.6 95.6 95.6 95.6 95.6 85 85 96.5 96.5
60 60 25 70 20 40 40 40 40 40 40 40 40 40 40
6 0.2 2.2 5 2.2 0.005 0.003 0.048 0.065 0.074 0.033 0.005 0.004 0.008 0.007
5 5.9 2.9 2.5 3.2 101 63 8.8 17 20 390 329 893 ∞ 35480
32 33 34 35 36 39 40 37 38 41 41 14 14 this study this study
a TEOS content 25 wt %, annealed at 160 °C, 8 h. b TEOS content 25 wt %, annealed at 130 °C, 24 h. c TEOS content 20 mol %, membrane thickness 66.8 mm. d TEOS content 45 mol %, membrane thickness 76.7 mm.
Dehydration of an Ethanol/Water Azeotrope
Biomacromolecules, Vol. 5, No. 4, 2004 1573
selectivity was improved remarkably as shown in Figure 2. In the q-Chito/TEOS hybrid membrane containing excess TEOS (>45 mol %), however, the polycondensation reaction between the TEOS molecules proceeds preferentially. Therefore, the cross-link reaction between q-Chito and TEOS was less favored. The resulting cohesion TEOS domains break the hydrogen bonds between the q-Chito molecules, and consequently, the elastic modulus and the cross-link density of the q-Chito/TEOS membrane decreased. These results influence on the diffusion path and, consequently, on the permeation rates and the selectivity. Therefore, an increase in the permeation rate and a decrease in the water/ethanol selectivity were observed. Conclusions
Figure 9. Change of cross-link structure in q-Chito/TEOS hybrid membranes.
hybrid membranes with 45 mol % of TEOS, a contrast between the q-Chito and TEOS could not be observed. Thus, it is suggested that the q-Chito and TEOS domains were randomly dispersed at the molecular level. On the other hand, in the q-Chito/TEOS membrane with 60 mol % of TEOS, a cohesive TEOS domain was observed in the hybrid membrane. These TEOS domains were spheres of less than 10 nm. This result suggests that in the presence of excess TEOS the polycondensation reaction between TEOS species progressed until silica sol particles were formed. From these results, the hypothetical structure of q-Chito membranes cross-linked with TEOS is shown in Figure 9. In the q-Chito/TEOS hybrid membranes containing 45 mol % of TEOS or less, the cross-linked structure was formed by smooth reactions between the q-Chito and TEOS and between the TEOS molecules. In this case, the elastic modulus and the cross-link density of the q-Chito/TEOS hybrid membranes decreased with increasing TEOS content. Consequently, although the normalized permeation rate decreased with increasing TEOS content, the water/ethanol
Organic-inorganic hybrid membranes with various TEOS contents were prepared from q-Chito and TEOS using the sol-gel reaction. The permeation and separation characteristics of these hybrid membranes for treatment of an azeotrope of ethanol/water by pervaporation were studied. The hybrid membranes containing up to 45 mol % TEOS showed higher water/ethanol selectivity than the unmodified q-Chito membrane. This is because the swelling of the hybrid membranes was depressed by the formation of the crosslinked structure between q-Chito and TEOS. However, the introduction of excess TEOS caused swelling of the hybrid membranes, and hence, the water/ethanol selectivity of the hybrid membrane with a TEOS content of 60 mol % was lowered remarkably. TEM images of the q-Chito/TEOS hybrid membranes demonstrated that TEOS was homogeneously hybridized in the membrane containing up to 45 mol % TEOS but was heterogeneously aggregated in the hybrid membrane with higher TEOS content. This suggests that, in the formation of the former membranes, the condensation reaction between q-Chito and TEOS was dominant. However, for the latter membrane, the self-condensation reaction among TEOSs was more predominant than the reaction between q-Chito and TEOS. The addition of excess TEOS yielded heterogeneous structures and caused swelling of the hybrid membranes. In this study, we demonstrated that membranes with high water/ethanol selectivity can be designed by hybridization of q-Chito as an organic component and TEOS as an inorganic component using the sol-gel reaction, and q-Chito offers potential for selective separation of water/ ethanol. Acknowledgment. This research was financially supported by a Grant-in-Aid for Scientific Research on Priority Areas (B) “Novel Smart Membranes Containing Controlled Molecular Cavity” from the Ministry of Education and Science, Culture, Sports, Science, and the Kansai University Special Research Found, 2003. References and Notes (1) Huang, R. Y. M. PerVaporation Membrane Separation Process; Elsevier: Amsterdam, 1991. (2) Uragami, T.; Takuno, M.; Miyata, T. Macromol. Chem. Phys. 2002, 203, 1162. (3) Yamaguchi, T.; Nakao, S.; Kimura, S. Macromolecules 1991, 24, 5522.
1574
Biomacromolecules, Vol. 5, No. 4, 2004
(4) Feng, X. S.; Huang, R. Y. M. Ind. Eng. Chem. Res. 1997, 36, 1048. (5) Baker, R. W. In Membrane Separation System, Recent DeVelopments and Future Directions; Baker, R. W., Cusssler, E. L., Eykamp, W., Koros, W. J., Riley, R. L., Strathmann, H., Eds.; Noyes Data Corp.: Park Ridge, NJ, 1991; p 151. (6) Psaume, R.; Aurell, Y.; Mora, C. J.; Bersillon, J. L. J. Membr. Sci. 1988, 36, 373. (7) Blume, I.; Wijmans, J. G.; Baker, R. W. J. Membr. Sci. 1990, 49, 253. (8) Fang, Y.; Pham, V. A.; Matuura, T.; Santerre, J. P.; Narbaitz, R. M. J. Appl. Polym. Sci. 1994, 54, 1937. (9) Huang, R. Y. M.; Yeom, C. K. J. Membr. Sci. 1990, 51, 273. (10) Liang, L.; Ruckenstein, E. J. Membr. Sci. 1995, 106, 167. (11) Uragami, T.; Yamamoto, S.; Miyata, T. Biomacromolecules 2003, 4, 137. (12) Uragami, T.; Matsuda, T.; Okuno, H.; Miyata, T. J. Membr. Sci. 1994, 88, 243. (13) Uragami, T.; Kato, S.; Miyata, T. J. Membr. Sci. 1997, 124, 203. (14) Uragami, T.; Okazaki, K.; Matsugi, H.; Miyata, T. Macromolecules 2002, 35, 9156. (15) Uragami, T.; Aketa, T.; Gobodani, S.; Sugihara, M. Polym. Bull. 1986, 15, 101. (16) Miyata, T.; Takagi, T.; Kadota, T.; Uragami, T. Macromol. Chem. Phys. 1995, 196, 1211. (17) Miyata, T.; Higuchi, J.; Okuno, H.; Uragami, T. J. Appl. Polym. Sci. 1996, 61, 1315. (18) Miyata, T.; Takagi, T.; Uragami, T. Macromolecules 1996, 29, 7787. (19) Miyata, T.; Nakanishi, Y.; Uragami, T. Macromolecules 1997, 30, 5563; ACS Symp. Ser. 1999, 733, 280. (20) Uragami, T.; Doio, T.; Miyata, T. ACS Symp. Ser. 1999, 733, 263. (21) Miyata, T.; Takagi, T.; Higuchi, J.; Uragami, T. J. Polym. Sci., Polym. Phys. 1999, 37, 1545.
Uragami et al. (22) Miyata, T.; Obata, S.; Uragami, T. J. Polym. Sci., Polym. Phys. 2000, 38, 584; Macromolecules 1999, 32, 3712; 1999, 32, 8465. (23) Uragami, T.; Yamada, H.; Miyata, T. Trans. Mater. Res. Soc. Jpn. 1999, 24, 165; J. Membr. Sci. 2001, 187, 255. (24) Uragami, T.; Meotoiwa, T.; Miyata, T. Macromolecules 2001, 34, 6806. (25) Miyata, T.; Yamada, H.; Uragami, T. Macromolecules 2001, 34, 8026. (26) Iked, Y. Kobunshi Hyomen, Kiso, Oyo, 1986; Chapter X, p 231. (27) Uragami, T.; Tsukamoto, K.; Inui, K.; Miyata, T. Macromol. Chem. Phys. 1998, 199, 49. (28) Ogata, M.; Kawata, T.; Kinjo, N. Kobunshi Ronbunshu 1987, 44, 193. (29) Tobolsky, A. V.; Carlson, D. W.; Indictor, N. J. Polym. Sci. 1961, 54, 175. (30) Flory, P. J. Principle of Polymer Chemistry; Cornell University Press: Ithaca, NY, 1953. (31) Sakka, S.; Ito, S.; Kamiya, K. J. Non-Cryst. Solids 1985, 71, 311. (32) Heisler, E. G.; Hunter, A. S.; Sicilians, J.; Treadway, R. H. Science 1956, 124, 77. (33) Aptel, P.; Cunny, J.; Josefowicz, J.; Morel, G.; Neel, J. J. Appl. Polym. Sci. 1974, 18, 351. (34) Aptel, P.; Challard, N.; Cunny, J.; Neel, J. J. Membr. Sci. 1982, 11, 3. (35) Ishikawa, T. Chem. Eng. 1984, 29, 19. (36) Ngnyen, Q. T.; Blanc, L. L.; Neel, J. J. Membr. Sci. 1985, 22, 245. (37) Uragami, T.; Saito, M. Sep. Sci. Technol. 1989, 24, 541. (38) Uragami, T.; Saito, M.; Takigawa, K. Macromol. Chem. Rapid. Commun. 1988, 9, 361. (39) Uragami, T.; Morikawa, T. Makromol. Chem. 1989, 190, 329. (40) Uragami, T.; Morikawa, T.; Okuno, H. Polymer 1989, 30, 329. (41) Uragami, T.; Takigawa, K. Polymer 1990, 31, 668.
BM0498880