- Email: [email protected]
Development of novel composite membranes using quaternized chitosan and Na+-MMT clay for the pervaporation dehydration of isopropanol
Journal of Colloid and Interface Science 338 (2009) 111–120
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Development of novel composite membranes using quaternized chitosan and Na+-MMT clay for the pervaporation dehydration of isopropanol Santosh K. Choudhari, Mahadevappa Y. Kariduraganavar * Center of Excellence in Polymer Science and Department of Chemistry, Karnatak University, Dharwad 580 003, India
a r t i c l e
i n f o
Article history: Received 12 January 2009 Accepted 28 May 2009 Available online 3 June 2009 Keywords: Chitosan Na+-MMT clay Isopropanol Pervaporation Selectivity Activation energy
a b s t r a c t Novel polymer–clay-based composite membranes were prepared by incorporating sodium montmorillonite (Na+-MMT) clay into quaternized chitosan. The resulting membranes were characterized by Fourier transform infrared spectroscopy (FTIR), wide-angle X-ray diffraction (WXAD), and thermogravimetric analysis (TGA). The effect of membrane swelling was studied by varying the water concentration in the feed. The membranes were employed for the pervaporation dehydration of isopropanol in terms of feed composition and Na+-MMT clay loading. The experimental results demonstrated that membrane containing 10 mass% of Na+-MMT clay showed the highest separation selectivity of 14,992 with a flux of 14.23 102 kg/m2 h at 30 °C for 10 mass% of water in the feed. The total flux and flux of water are found to be overlapping each other particularly for clay-incorporated membranes, signifying that the composite membranes developed in the present study involving quaternized chitosan and Na+-MMT clay are highly selective toward water. From the temperature-dependent diffusion and permeation values, the Arrhenius activation parameters were estimated. The resulting activation energy values obtained for water permeation ðEpw Þ are much lower than those of isopropanol permeation ðEpIPA Þ, suggesting that the developed composite membranes have higher separation efficiency for the water–isopropanol system. The estimated Ep and ED values ranged between 8.97 and 11.89, and 7.56 and 9.88 kJ/mol, respectively. The positive heat of sorption ðDHs Þ values were obtained for all the membranes, suggesting that Henry’s mode of sorption is predominant in the process. Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction Pervaporation (PV) is an efficient membrane-based process and has gained acceptance by chemical industries over the years because of its favorable economics, easy maintenance, and simplicity of the process [1–3]. During the last decade, several commercial PV plants have been established all over the world especially for the dehydration of organics. Therefore, much of the attention has now been focused toward the development of new robust membranes with better performance. As an outcome of this, organic–inorganic hybrid materials have recently attracted considerable attention and are popularly termed as ‘‘next generation” membrane materials [4–6]. Such hybrid membranes are typically composed of polymeric and inorganic materials, in which inorganic filler with specific physicochemical characteristics is spatially dispersed into a bulk of the polymeric membranes. This seems to be a facile and feasible solution to have a good flux and high selectivity with desirable mechanical properties of inorganic materials and tunable flexibility with an excellent processability of polymeric materials [7]. Recent studies have dem* Corresponding author. Fax: +91 836 2771275. E-mail address: [email protected] (M.Y. Kariduraganavar). 0021-9797/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2009.05.071
onstrated that the dramatically improved separation properties of organic–inorganic hybrid membranes could be realized by incorporating porous or nonporous inorganic adsorbents such as zeolite, silica, carbon molecular sieve, activated carbon, and clay into the matrices of glassy and rubbery polymers [4–11]. Among these, nano-structured polymer–clay hybrid materials are currently the objects of intensive research because of their unique properties and low cost synthesis as clays are abundantly available in nature. Polymer–clay is the hybrid composite material consisting of polymer matrix with dispersed clay nanoparticles. Nano-clays have also been widely used as an inorganic reinforcement for polymer matrices with nano-scale dispersion of the inorganic phase within the polymer matrix [12–16]. Addition of a small amount of nanoclay into polymer can lead to higher moduli, increased strength, heat resistance [17–19], decreased permeability [20], and flammability [21]. The enhanced properties are presumably due to nanoscale structure effects and the interaction of inorganic–organic materials. Chitosan [poly-bð1 ! 4Þ-D-glucosamine] is an aminopolysaccharide and has attracted a great interest as a basic membrane material due to its high abundance in nature, hydrophilicity, chemical resistance, adequate mechanical strength, good membrane forming capacity, and more importantly its functional groups such
112
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
as amino and hydroxyl, which provide an opportunity for chemical modification. In fact it is an important method for improving the membrane property of pure chitosan. Several modified forms such as carboxylated, sulfated, phosphorylated [22,23], and in particular quaternized chitosan [24–26] have been well tested as membrane material for pervaporation dehydration. The quaternized chitosan showed an excellent water permselectivity for aqueous alcohol solution. However, an excess of swelling results in an increase of both solubility and diffusivity of alcohols, and consequently lowers the water permselectivity. In order to circumvent this difficulty, Uragami et al. [24–26] have tried to crosslink quaternized chitosan with different reagents such as diethylene glycol diglycidyl ether, poly(ethylene oxydiglycolic acid), and tetraethylorthosilicate, and have succeeded to some extent. On the other hand, biopolymer–clay nanocomposites represent an innovative and promising class of new materials [27–30], in which the enhancement of material property largely depends on the spatial distribution, rearrangement of intercalating polymer chains, and the type of interactions existing between polymer and clay. Sodium montmorillonite (2:1 aluminosilicate), which belongs to the general smectite group of bentonite clays, carries a distribution of ve charges in its framework, and thus interacts electrostatically with metal cations such as sodium in its interlayer galleries. It is well established that positively charged polymers, for instance, b-(dimethylamino)ethyl methacrylate hydroacetate can be intercalated in smectite clays following cation-exchange mechanisms [31,32]. Thus, the quaternized chitosan having strong +ve charges on the backbone is an excellent candidate for interaction with Na+-montmorillonite through a cation-exchange process. The opposite charges present on the polymer and clay not only help for better dispersion of clay within the polymer matrix but also establish the strong electrostatic interactions between them. Therefore, the current investigation focuses on the development of new composite hybrid membranes by dispersing different mass% of Na+-MMT clay in the quaternized chitosan. The physicochemical changes in the resulting membranes were investigated using Fourier transform infrared spectroscopy (FTIR), wide-angle X-ray diffraction (WAXD), and thermogravimetric analysis (TGA) techniques, and the membranes were successfully subjected for PV separation of water–isopropanol mixtures. The effects of clay loading, water composition in the feed, and temperature on PV performance were systematically evaluated. From the temperature dependence of permeation flux and diffusion coefficient, the Arrhenius activation parameters were estimated. The results are discussed in terms of PV separation efficiency of the membranes.
a 1-l capacity round-bottom flask fitted with a condenser. To this heterogeneous mixture, 20 ml of methyl iodide was added at once and refluxed for 8 h at 50 °C to yield quaternary iodide ammonium salt. The resulting clear solution was cooled to ambient temperature and then excess sodium chloride was added and stirred overnight to convert iodide ammonium salt to chloride ammonium salt. This solution was poured into an excess quantity of acetone while stirring. The precipitated crude quaternized chitosan was purified by dissolving in water and repeatedly reprecipitating in acetone. The quaternization of chitosan is presented in Fig. 1. The percentage degree of quaternization was estimated by recording a NMR spectrum with D2 O as solvent on a Bruker Avance 300 MHz spectrometer and it was found to be 15%. 2.3. Membrane preparation The quaternized chitosan (3 g) was dissolved in 87 ml of deareated distilled water at room temperature. The solution was filtered using a fritted glass disk filter to remove undissolved residue particles. In the meantime, a known amount of Na+-MMT clay was dispersed in 10 ml of deareated distilled water and kept in an ultrasonic bath at a fixed frequency 38 kHz (Grant XB6, UK) for 30 min to break the possible aggregated particles. The clay dispersion was slowly added into a solution of quaternized chitosan while stirring, and the resulting mixture was further stirred for overnight to obtain uniform membrane material. It was then spread onto a glass plate with the aid of a casting knife in a dust-free atmosphere. The membrane was allowed to dry at room temperature for 2–3 days and then subsequently peeled off. The amount of clay with respect to quaternized chitosan was varied from 0, 5, 10, and 15 mass%, and the membranes thus obtained were designated as M, M-1, M-2, and M-3, respectively. The thickness of these membranes was measured at different points using a Peacock dial thickness gauge (Model G, Ozaki MFG Co. Ltd., Japan) with an accuracy of ±2 lm. The thickness of the membranes was found to be 40 ± 2 lm. 2.4. Ion-exchange capacity (IEC)
2. Experimental
In order to assess the total number of positively charged groups which have electrostatically interacted with clay in the membranes, the ion-exchange capacity of the pure quaternized chitosan and its clay-incorporated composite membranes was determined. The membrane samples were completely dried and their weights were recorded. The dried membrane samples were soaked in 20 ml of 50 vol% alcoholic solution of 0.01 M sodium hydroxide. After 24 h, 10 ml of this solution was titrated against 0.01 M HCl. The ion-exchange capacity of the membranes was calculated as
2.1. Materials
IEC ¼
w 200;000; N-deacetylation degree 75–85%) and Chitosan ðM sodium montmorillonite were obtained from Sigma–Aldrich Chemicals (USA). Methyl iodide and acetone were purchased from Loba Chemicals, Mumbai, India. Isopropanol (IPA) and sodium chloride were purchased from s.d. fine Chemicals Ltd., Mumbai, India. All the chemicals are of reagent grade and were used without further purification. Double-distilled water was used throughout the study.
where B is the amount of HCl used to neutralize the blank solution; P is the amount of HCl used to neutralize the equilibrated solution; m is the weight of membrane sample; 2 is the correction factor.
2.2. Synthesis of quaternized chitosan The quaternization of chitosan was carried out with a prime aim to increase the performance of chitosan membranes. Chitosan (6 g) was dispersed while stirring at room temperature in 760 ml of aqueous-organic mixture (42 vol% methanol), which was taken in
B P MNaOH 2 m
ð1Þ
2.5. Fourier transform infrared spectroscopy The incorporation of different amounts of Na+-MMT clay material and its interaction with quaternized chitosan were confirmed by a FTIR spectrometer (Nicolet, Impact-410, USA). Membrane samples were ground well to make KBr pellets under hydraulic pressure of 400 kg/cm2 and spectra were recorded in the range of 400–4000 cm1. In each scan, the amounts of membrane sample and KBr were kept constant in order to know the changes in the intensities of characteristic peaks with respect to the amount of clay loading.
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
113
Fig. 1. Scheme for the synthesis of quaternized chitosan.
2.6. Wide-angle X-ray diffraction The morphology of the pure quaternized chitosan and its clayloaded membranes was studied at room temperature using a Philips analytical X-ray diffractometer. The X-ray source was nickelfiltered Cu-Ka radiation (40 kV, 30 mA). The dried membranes of uniform thickness ð 40 lmÞ were mounted on a sample holder and scanned in the reflection mode at an angle 2h over a range from 2° to 40° at a speed of 8°/min. 2.7. Thermogravimetric analysis Thermal properties of the quaternized chitosan and its clayincorporated composite membranes were investigated by Perkin– Elmer Diamond thermogravimetric analyzer at a heating rate of 10 °C/min under nitrogen atmosphere. The weight of the samples taken for each record was 5–9 mg. 2.8. Swelling measurements The equilibrium sorption experiments were performed in different compositions of water–isopropanol mixtures using an electronically controlled oven (WTB Binder, Germany). The masses of the dry membranes were first determined and these were equilibrated by soaking in different compositions of the mixtures in a sealed vessel at 30 °C for 24 h. The swollen membranes were weighed as quickly as possible after careful blotting on a digital microbalance (Mettler, B204-S, Toledo, Switzerland) within an accuracy of ±0.01 mg. All the experiments were performed at least three times and the results were averaged. The percentage degree of swelling (DS) was calculated as
DSð%Þ ¼
Ws Wd 100; Wd
ð2Þ
where W s and W d are the masses of the swollen and dry membranes, respectively. 2.9. Pervaporation experiments PV experiments were performed using the in-house designed apparatus reported in our previous articles [6,33]. The effective surface area of the membrane in contact with the feed mixture was 34.23 cm2 and the capacity of the feed compartment was about 250 cm3. The vacuum in the downstream side of the apparatus was maintained [1.333224 103 Pa (10 Torr)] using a twostage vacuum pump (Toshniwal, Chennai, India). The water composition in the feed mixture was varied from 5 to 25 mass%. The test membrane was allowed to equilibrate for about 2 h in the feed compartment with a known volume of feed mixture at the corresponding temperature before the PV experiment was performed. After a steady state was attained, the permeate was collected in a trap immersed in the liquid nitrogen jar on the downstream side at fixed intervals of time. The experiments were carried out at 30, 40, and 50 °C. The flux was calculated by weighing the permeate on a digital microbalance with an accuracy of ±0.01 mg. The compositions of water and isopropanol were estimated by measuring the refractive index of the permeate within an accuracy of ±0.0001 unit, using an Abbe’s refractometer (Atago-3T, Japan), and by comparing it with a standard graph, which was established previously with the known compositions of water–isopropanol mixtures. All the experiments were performed at least three times, and the results were averaged. The results of permeation for water–isopropa-
114
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
nol mixtures during the pervaporation were reproducible within an admissible range. From the PV data, separation performance of the membranes was assessed in terms of total flux (J), separation selectivity ðasep Þ, and pervaporation separation index (PSI). These were respectively calculated using the equations
W J¼ At P =P asep ¼ w IPA F w =F IPA PSI ¼ Jðasep 1Þ;
ð3Þ ð4Þ ð5Þ
where W is the mass of permeate (kg); A, the effective membrane area (m3); t, the permeation time (h); Pw and PIPA are the mass percentage of water and isopropanol in the permeate, respectively; F w and F IPA are the respective mass percentage of water and isopropanol in the feed. 3. Results and discussion 3.1. Membrane characterization 3.1.1. Ion-exchange capacity Residual ionic groups in the resulting membranes were estimated by determining the ion-exchange capacity using the titration method. It is noted that the ion-exchange capacity of quaternized chitosan membrane (M) was found to be 0.342 meq/ g. However, the clay-incorporated quaternized chitosan membranes M-1, M-2, and M-3 showed ion-exchange capacities of 0.233, 0.141, and 0.110 meq/g, respectively. This suggests that the incorporated clay material correspondingly lowers the free ionic groups in the membrane matrix, indicating that both quater-
nary ammonium groups and protonated amino groups of chitosan have strongly interacted with negative sites of clay through an electrostatic force of attraction. The establishment of electrostatic interaction is illustrated in Fig. 2. 3.1.2. FTIR studies Fig. 3 shows the FTIR spectra of quaternized chitosan and its clay-incorporated membranes. The spectrum of quaternized chitosan membrane showed that a characteristic band at around 3400 cm1 was attributed to O–H stretching vibrations and the bands appearing at around 1630 and 1570 cm1 were respectively assigned to amide I and amide II functional groups. The multiple bands appeared between 1200 and 1000 cm1 were assigned to C–O stretching. On incorporating the clay into quaternized chitosan membrane matrix, the intensity of C–O stretching increased and this was predominant with increase of the amount of clay as clearly demonstrated in spectra of M-1 to M-3. This is expected due to Si–O–Si stretching of clay material as it overlaps at the same frequency of C–O. Further, when we observed the spectra carefully, it is noted that the amino bands appearing at 1630 and 1570 cm1 slightly shifted toward a lower frequency with increase of the amount of clay, suggesting the possibility of formation of hydrogen bonds between the clay and the quaternized chitosan [34]. In addition to this, the O–H band also shifted slightly to a lower frequency, which supports the enhanced hydrogen bonding occurring between the clay and the quaternized chitosan. All this evidence confirms the incorporation of clay and its interaction with the quaternized chitosan. 3.1.3. Wide-angle X-ray diffraction studies The dispersion of clay in the quaternized chitosan membrane matrix was confirmed by WAXD, which is the most frequently used
Fig. 2. Scheme for the preparation of composite membranes.
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
115
composite membranes. This might be due to a higher crystallinity or denser packing in the main chain in comparison with neat quaternized chitosan. This is in good agreement with the results reported by Thomas and colleagues [37] and Ren et al. [38] for polymer–clay composites. This could be responsible for better pervaporation performance of composite membranes in comparison with quaternized chitosan membrane. On the other hand, the pattern of Na+-MMT clay showed a reflection peak at around 2h ¼ 6:58, corresponding to basal spacing of 1.34 nm. After incorporating Na+-MMT clay in quaternized chitosan matrix, this peak was shifted to a lower angle. This clearly indicates that quaternized chitosan chains were intercalated into the interlayers of Na+-MMT clay through a cation-exchange mechanism. This is mainly because of the hydrophilic and polycationic nature of polymer, which establishes an excellent miscibility with Na+-MMT. With further increase of the clay content in the quaternized chitosan membrane, the d-spacing values decreased systematically from membranes M-1 to M-3 (2.0, 1.92, and 1.83 nm). The decrease in d-spacing values suggests the intercalation of quaternized chitosan chains into the silicate layers. This is expected as a higher content of clay favors the intercalation. However, the intensity of this peak is very small in all the composite membranes, suggesting that a major part of the clay is in the exfoliated form.
Fig. 3. FTIR spectra of quaternized chitosan and its clay-incorporated composite membranes: (M) 0 mass%; (M-1) 5 mass%; (M-2) 10 mass%; (M-3) 15 mass% of Na+MMT clay.
method to study the structure of clay–polymer composites. Depending on the relative distribution/dispersion of the stacks of clay platelets, three types of polymer-layered silicate nanocomposites (PLSN) can be described [35,36]: intercalated PLSNs, where polymer chains are intercalated into the silicate layers resulting in a well-ordered multilayer morphology built up with alternating polymer and inorganic layers; flocculated PLSNs, where intercalated stacked silicate layers are sometimes flocculated due to the hydroxylated edge–edge interactions; exfoliated/delaminated PLSNs, where silicate layers are completely and homogeneously dispersed in the polymer matrix. Fig. 4 illustrates the WAXD patterns of Na+-MMT clay, quaternized chitosan, and clay-incorporated quaternized chitosan membranes. The pattern of quaternized chitosan membrane exhibited the characteristic crystalline peaks at 2h 12 and 23°. The intensity of these peaks increased with increasing the clay material in the membrane matrix, indicating the enhanced order structure of
3.1.4. Thermogravimetric analysis The thermal stability of quaternized chitosan and its clayloaded composite membranes was investigated by TGA under nitrogen flow. Fig. 5 shows the TGA curves of Na+-MMT clay, quaternized chitosan, and clay-loaded composite membranes. Under nitrogen flow nonoxidative degradation occurred in two stages for all membranes. The weight loss occurring between ambient temperature and 200 °C corresponds to the absorbed water molecules. Such weight loss is about 15% for Na+-MMT clay, whereas quaternized chitosan and its clay-loaded membranes exhibited slightly higher loss ranging from 20 to 25%. This fact indicates that the quaternized chitosan and its clay-loaded membranes have high water retention capacity. However, this was decreased correspondingly with increase of the amount of clay in the membrane. The second stage of decomposition occurred between 200 and 450 °C, and corresponds to weight loss of 50–55%. This is due to degradation and deacetylation of quaternized chitosan [39,40]. If we consider the decomposition temperature at 50% weight loss, the clay-loaded membranes exhibited around 5–30 °C higher temperature than that of quaternized chitosan. This clearly indicates that clay-loaded composite membranes demonstrated higher thermal decomposition temperature compared to quaternized chitosan membrane. This is because of the synergetic effect of clay platelets and quaternized chitosan on the thermal properties. The significant improvement of thermal stability of the quaternized chitosan may be explained as follows [41,42]: good heat barrier properties of clay materials for polymer matrix during formation of chars; formation of carbonaceous layer on the polymer surface at lower temperatures, which could hinder the flux of degradation product and heat flowing into underlying materials; much jammed and conjugated clay–polymer network, or strong interaction between quaternized chitosan nano-fillers could also restrict the polymer motion during heating. 3.2. Effects of feed composition and clay loading on membrane swelling The membrane swelling in certain liquids depends on the chemical composition and microstructure of the polymer, and the incorporated moiety, which strongly influence the sorption mechanism [43]. Therefore, the degree of membrane swelling is
116
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
Fig. 4. Wide-angle X-ray diffraction patterns of Na+-MMT clay, quaternized chitosan, and clay-incorporated composite membranes: (M) 0 mass%; (M-1) 5 mass%; (M-2) 10 mass%; (M-3) 15 mass% of Na+-MMT clay.
of course an important factor in the PV process, which controls the transport of permeating molecules under the driving force of chemical potential gradient. In order to study the effects of feed composition and clay loading on the membrane swelling, the percentage degree of swelling of all the membranes was plotted with respect to water composition in the feed at 30 °C as shown in Fig. 6. It is observed that the degree of swelling increased almost linearly for all the membranes with increasing the mass% of water in the feed. This is due to increased interaction between water molecules and the membrane, owing to the presence of interactive groups ð—Nþ ðCH3 Þ3 ; NHþ 3 ; NH2 , and –OH) in the membrane matrix. This is expected since water is more polar than IPA, which preferentially interacts with membrane resulting in an increased degree of swelling. On the other hand, the degree of swelling decreased with increasing the clay content in the membrane at all feed compositions. This is because, the increase of the clay content in the membrane matrix reduces the free volume due to increased packing density; secondly, it relatively lowers the amount of quaternized chitosan available for interaction. Above all, clay establishes a strong ionic interaction with quaternized chitosan leading to polymer functional groups unavailable for solvent interaction. All these together are responsible for the decreased
degree of swelling with increase of the clay content in the membrane matrix. 3.3. Effect of clay loading on pervaporation In the PV process, the overall performance of the membranes can be assessed based on the flux and selectivity. In order to calculate these parameters the developed membranes were subjected to a pervaporation experiment at 10 mass% of water in the feed. The resulting data of both flux and selectivity are presented in Fig. 7. It is observed that the clay-loaded composite membranes exhibited higher separation selectivity with lower permeation flux compared to quaternized chitosan membrane. As the content of clay was increased in the quaternized chitosan membrane, the selectivity was increased dramatically up to 10 mass% and then decreased drastically. However, the permeation flux was decreased almost linearly with increase of the amount of clay. This complex transport phenomenon can be explained on the basis of decreased free volume and establishment of a tortuous path. In polymer–clay composite membranes, the incorporated clay usually exits in the forms of exfoliated, intercalated, and flocculated states [35,36]. The extent of each state depends on the ratio between polymer
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
117
Fig. 5. Thermogravimetric analysis of Na+-MMT clay, quaternized chitosan, and clay-incorporated composite membranes: (M) 0 mass%; (M-1) 5 mass%; (M-2) 10 mass%; (M-3) 15 mass% of Na+-MMT of clay.
Fig. 7. Variation of pervaporation flux and selectivity with different mass% of Na+MMT clay loading at 10 mass% of water in the feed. Fig. 6. Variation of degree of swelling with different mass% of water in the feed for quaternized chitosan and its Na+-MMT clay-incorporated composite membranes.
and the incorporated clay. In our study, the exfoliated state was predominant at a lower content of clay as evidenced by WAXD data. This favors the molecular level interaction between the polymer and the clay, resulting in decreased free volume. In addition, a high aspect ratio of layered silicates caused them to act as a barrier, offering more resistance to diffusion by creating tortuosity to the diffusion pathway. This has suppressed the diffusion of both water and IPA molecules. However, the diffusion of IPA molecule was affected significantly as the kinetic diameter of IPA molecule is almost four times bigger than that of water molecule [44]. Obviously, transport of IPA molecules was less preferred as compared to water molecules, resulting in increased selectivity and decreased permeation flux. On further increase of the amount of clay
beyond 10 mass% both flux and selectivity were decreased. The decrease in selectivity was due to micro-phase separation occurring between organic and inorganic matrix, which has resulted from the shift of exfoliation state to intercalation followed by flocculation. The decrease in flux accounted for the lower degree of swelling at higher mass% of clay. The pervaporation separation index (PSI) is the product of total permeation flux and separation factor, which characterizes the membrane separation ability. This index can be used as a relative guideline for the design of new membranes for pervaporation separation processes and also to select a membrane with an optimal combination of flux and selectivity. Fig. 8 shows the variation of PSI as a function of mass% of clay loading at 30 °C for 10 mass% of water in the feed. It is observed that there is a dramatic increase in PSI values with increasing the clay from 0 to 10 mass%, but fur-
118
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
and resulting data are presented in Fig. 10. It is observed that the flux was increased with increase of the water composition in the feed. This is mainly because of increased membrane swelling. This is expected due to an establishment of greater interaction between the membrane and the water molecules as membrane contains a large number of hydrophilic groups such as —OH; —NH2 ; —NHþ 3, and —Nþ ðCH3 Þ3 , which preferentially interact with water molecules rather than IPA. On the contrary, the selectivity was decreased exponentially with increasing the mass% of water in the feed. At a higher concentration of water in the feed, a small amount of water dissolves in the membrane, which in turn acts as a plasticizer for the membrane, leading to more flexible polymeric chains in the matrix. In addition, as water is more polar in nature its absorption in the membrane might have weakened the interaction between polymer and clay material, resulting in a somewhat loose structure, which becomes responsible for facilitating the diffusion of some of the IPA molecules in association with water molecules. 3.5. Diffusion coefficient Fig. 8. Variation of pervaporation separation index with different mass% of Na+MMT clay loading at 10 mass% of water in the feed.
ther addition of clay (15 mass%) decreases the PSI value. This is because of a sharp decrease in selectivity arising from the microphase separation between the polymer and the clay in the membrane matrix. In order to assess the efficiency of the membranes, we have plotted the individual fluxes as a function of clay loading at 10 mass% of water in the feed as shown in Fig. 9. From the plot, it is clearly noted that the total flux and flux of water are close to each other for all the clay-loaded composite membranes. However, the closeness is more predominant with increase of the clay. This is supported by the flux of IPA, which is meager and decreased further with increase of the clay, signifying that clay-loaded composite membranes exhibited an excellent separation performance.
In the PV process, mass transport of binary liquid mixtures through a nonporous polymeric membrane is generally described by the solution-diffusion mechanism, which occurs in three steps: sorption, diffusion, and evaporation [45]. Thus, permeation flux and selectivity are governed by the solubility and diffusivity of each component of the feed mixture to be separated. During the process, because of the establishment of fast equilibrium distribution between bulk feed and the upstream surface of a membrane, the diffusion step controls the transport of penetrants [9,46]. Therefore, it is important to estimate the diffusion coefficient ðDi Þ of penetrating molecules to understand the mechanism of transport. From Fick’s law of diffusion, the diffusion flux can be expressed as [47]
J i ¼ Di
dC i dx
;
ð6Þ
In PV, the effect of feed is also an important factor, and thus we have carried out a PV study at different mass% of water in the feed
where J is the permeation flux per unit area (kg/m2 s), D is the diffusion coefficient (m2/s), C is the concentration of permeant (kg/ m3), subscript i stands for water or isopropanol, and x is the diffusion length (m). For simplicity, it is assumed that the concentration profile along the diffusion length is linear. Thus, Di can be calculated with the equation [48]
Fig. 9. Variation of total flux, and fluxes of water and isopropanol with different mass% of Na+-MMT clay loading at 10 mass% of water in the feed.
Fig. 10. Variation of pervaporation flux and selectivity at different feed compositions of water for 10 mass% of Na+-MMT clay-incorporated membrane.
3.4. Effect of feed composition on pervaporation
119
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
Di ¼
Ji d ; Ci
ð7Þ
where d is the membrane thickness. The calculated values of Di at 30 °C for membrane (M-2) containing 10 mass% of Na+-MMT clay are plotted as function of water as shown in Fig. 11. With increasing water composition, the diffusion coefficient of water was decreased while increasing the diffusion coefficient of isopropanol. This is because of a weaker interaction between polymer and clay material, which facilitates the transport of IPA molecules.
Table 1 Pervaporation flux and separation selectivity at different temperatures for different membranes at 10 mass% of water in the feed. Temp. (°C)
30 40 50
J 102 ðkg=m2 hÞ
asep
M
M-1
M-2
M-3
M
M-1
M-2
M-3
16.68 18.41 20.63
15.60 17.49 19.49
14.27 16.37 18.41
11.58 13.42 15.63
1088 817 638
2423 1658 1089
14992 8991 3741
8991 3741 1755
3.6. Effect of temperature on pervaporation The effect of operating temperature on the pervaporation performance was studied for all the membranes at 10 mass% of water in the feed, and the values thus obtained are presented in the Table 1. It is observed that the permeation rate was found to increase from 30 to 50 °C for all the membranes while decreasing the separation selectivity. The increased permeation flux is mainly due to increased driving force at higher temperatures. This is because of the combined effects of both decreased cohesive forces between permeants, and increased vapor pressure at the feed side. However, the interaction between permeants and the membrane decreased at higher temperatures, which led to the permeation of both water and IPA molecules, resulting in decreased selectivity. This effect prompted us to estimate the activation energies for permeation and diffusion using the following Arrhenius-type equation [49],
X ¼ X o exp
Ex ; RT
ð8Þ
where X represents permeation (J) or diffusion (D), X o is a constant representing preexponential factor of Jo or Do . Ex represents activation energy for permeation or diffusion depending on the transport process under consideration and RT is the usual energy term. Arrhenius plots of log J and log D versus 1/T are shown in Figs. 12 and 13 for the temperature dependence of permeation flux and diffusion, respectively. In both cases, a linear behavior was observed, suggesting that both permeability and diffusivity follow an Arrhenius trend. From least-squares fits of these linear plots, the activation energies for permeability ðEp Þ and diffusivity ðED Þ were estimated. Similarly, we have also estimated the activation energies for diffusion of water ðEDw Þ, permeation of water ðEpw Þ, and iso-
Fig. 12. Variation of log J with temperature for different mass% of Na+-MMT clayloaded composite membranes at 10 mass% of water in the feed.
Fig. 13. Variation of log D with temperature for different mass% of Na+-MMT clayloaded composite membranes at 10 mass% of water in the feed.
Fig. 11. Variation of Dw and DIPA at different feed compositions of water for 10 mass% of Na+-MMT clay-incorporated membrane.
propanol ðEpIPA Þ, but the plots are not given to save space. The results thus obtained are presented in Table 2. From Table 2, it is observed that the apparent activation energy values of water ðEpw Þ are much lower than those of isopropanol ðEpIPA Þ, suggesting that membranes have higher separation efficiency toward water. The activation energy values for water permeation ðEpw Þ and total permeation ðEp Þ are almost close to each
120
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
Table 2 Arrhenius activation parameters for permeation and diffusion, and heat of sorption. Parameters (kJ/mol)
M
M-1
M-2
M-3
Ep ED Epw EpIPA EDw DHs
8.97 7.56 8.64 30.00 6.51 1.41
9.47 7.72 9.08 41.33 7.02 1.75
10.91 9.07 10.38 66.41 8.13 1.84
11.89 9.88 11.52 77.49 9.21 2.01
other, signifying that coupled transport of both (water and isopropanol molecules) is minimal as due to a higher selective nature of membranes. The estimated Ep and ED values ranged between 8.97 and 11.89, and 7.56 and 9.88 kJ/mol, respectively. Using these values, we have further calculated the heat of sorption as
DH s ¼ E p E D :
ð9Þ
The resulting DHs values are included in Table 2. The DHs values give additional information about the transport of molecules through the polymer matrix. It is a composite parameter involving contributions of both Henry and Langmuir types of sorption [50]. The Henry type of sorption requires both the formation of a site and the dissolution of chemical species into that site. The formation of a site involves an endothermic contribution to the sorption process. However, the Langmuir type of sorption requires the preexistence of a site in which sorption occurs by a hole-filling mechanism, giving an exothermic contribution. The DHs values obtained in the present study are positive for all the membranes, suggesting that Henry’s sorption is predominant, giving an endothermic contribution. 4. Conclusions Novel composite membranes were prepared using quaternized chitosan and Na+-MMT clay. Thermogravimetric study reveals that composite membranes exhibited better thermal stability compared to quaternized membrane. The performance of membranes was evaluated for the separation of water–IPA mixtures. An increase of Na+-MMT clay content in the quaternized chitosan matrix results in a decrease of permeation flux and an increase of separation selectivity. This was explained on the basis of reduction in free volume and establishment of a tortuous pathway. While assessing the membranes’ efficiency, it is clearly noted that both total flux and flux of water are overlapping each other particularly for clay-incorporated composite membranes, signifying that the composite membranes developed in the present study are highly selective toward water. The PV separation index data also indicated that the clay-incorporated membranes except membrane M-3 demonstrated an excellent PV performance. Among the membranes, the membrane containing 10 mass% of Na+-MMT clay exhibited the highest separation selectivity of 14,992 and with a permeation flux of 14.23 102 kg/m2 h at 30 °C for 10 mass% of water in the feed. With regard to increase in temperature the permeation rate was found to increase while suppressing the selectivity. This was attributed to increased vapor pressure and decreased interaction between permeants and membrane at higher temperatures. A significant difference was noted between Epw and EpIPA values. However, the difference was more significant particularly for the membranes having higher loading of Na+-MMT clay, suggesting
that the composite membranes developed with higher loading of clay exhibited remarkable separation ability toward water. The Ep and ED values ranged between 8.97 and 11.89, and 7.56 and 9.88 kJ/mol. All the membranes exhibited positive DHs values, indicating that sorption is mainly dominated by the Henry mode of sorption, giving an endothermic contribution.
References [1] P.D. Chapman, T. Oliveira, A.G. Livingston, K. Li, J. Membr. Sci. 318 (2008) 5. [2] Z. Huang, Y. Shi, R. Wen, Y.H. Guo, J.F. Su, T. Matsura, Sep. Purif. Technol. 51 (2006) 126. [3] P.D. Chapman, X. Tan, A.G. Livingston, K. Li, T. Oliveira, J. Membr. Sci. 268 (2006) 13. [4] F. Peng, L. Lu, H. Sun, Y. Wang, H. Wu, Z. Jiang, J. Membr. Sci. 275 (2006) 97. [5] T. Uragami, K. Okazaki, H. Matsugi, T. Miyata, Macromolecules 35 (2002) 9156. [6] M.Y. Kariduraganavar, S.S. Kulkarni, A.A. Kittur, J. Membr. Sci. 246 (2005) 83. [7] F. Peng, L. Lu, H. Sun, F. Pan, Z. Jiang, Ind. Eng. Chem. Res. 46 (2007) 2544. [8] Y.L. Liu, C.Y. Hsu, Y.H. Su, J.Y. Lai, Biomacromolecules 6 (2005) 368. [9] A.A. Kittur, M.Y. Kariduraganavar, U.S. Toti, K. Ramesh, T.M. Aminabhavi, J. Appl. Polym. Sci. 90 (2003) 2441. [10] F. Peng, F. Pan, H. Sun, L. Lu, Z. Jiang, J. Membr. Sci. 300 (2007) 13. [11] S. Anilkumar, M.G. Kumaran, S. Thomas, J. Phys. Chem. B 112 (2008) 4009. [12] S. Takahashi, D.R. Paul, Polymer 47 (2006) 7535. [13] R. Stephen, C. Ranganathaiah, S. Varghese, K. Joseph, S. Thomas, Polymer 47 (2006) 858. [14] Z.F. Wang, B. Wang, N. Qi, H.F. Zhang, L.Q. Zhang, Polymer 46 (2005) 719. [15] M.S. Hedenqvist, A. Backman, M. Gallstedt, R.H. Boyd, U.W. Gedde, Compos. Sci. Technol. 66 (2006) 2350. [16] H. Heinz, R.A. Vaia, R. Krishnamoorti, B.L. Farmer, Chem. Mater. 19 (2007) 59. [17] R.A. Vaia, R.K. Teukolsky, E.P. Giannelis, Chem. Mater. 6 (1994) 1017. [18] R.A. Vaia, E.P. Giannelis, Macromolecules 30 (1997) 7990. [19] R.A. Vaia, E.P. Giannelis, Macromolecules 30 (1997) 8000. [20] M.A. Osman, J.E.P. Rupp, U.W. Suter, J. Mater. Chem. 15 (2005) 1298. [21] E. Manias, A. Touny, L. Wu, K. Strawchecker, B. Lu, T.C. Chung, Chem. Mater. 13 (2001) 3516. [22] Y.M. Lee, Desalination 90 (1993) 277. [23] Y.M. Lee, E.M. Shin, J. Membr. Sci. 64 (1994) 145. [24] T. Uragami, M. Takumo, T. Miyata, Macromol. Chem. Phys. 203 (2002) 1162. [25] T. Uragami, S. Yamamoto, T. Miyata, Biomacromolecules 4 (2003) 137. [26] T. Uragami, T. Katayama, T. Miyata, H. Tamura, T. Shiraiwa, A. Higuchi, Biomacromolecules 5 (2004) 1567. [27] M. Darder, M. Colilla, E. Ruiz-Hitzky, Chem. Mater. 15 (2003) 3774. [28] M. Darder, M. Colilla, E. Ruiz-Hitzky, Appl. Clay Sci. 28 (2005) 199. [29] H. Gecol, E. Ergican, P. Miakatsindila, J. Colloid Interface Sci. 292 (2005) 344. [30] H. Gecol, P. Miakatsindila, E. Ergican, S.R. Hiibel, Desalination 197 (2006) 165. [31] R.A. Ruehrwein, D.W. Ward, Soil Sci. 73 (1952) 485. [32] C. Breen, Appl. Clay Sci. 15 (1999) 187. [33] M.Y. Kariduraganavar, A.A. Kittur, S.S. Kulkarni, K. Ramesh, J. Membr. Sci. 238 (2004) 165. [34] C. Tang, N. Chen, Q. Zhang, K. Wang, Q. Fu, X. Zhang, Polym. Degrad. Stab. 94 (2009) 124. [35] M. Alexandre, P. Dubois, Mater. Sci. Eng. R. 28 (2000) 1. [36] S.S. Ray, M. Okamoto, Prog. Polym. Sci. 28 (2003) 1539. [37] Y.H. Ha, Y. Kwon, T. Breiner, E.P. Chen, T. Tzianetopomon, R.E. Cohen, M.C. Boyce, E.L. Thomas, Macromolecules 38 (2005) 5170. [38] J. Ren, A. Silva, R. Krishnamoorti, Macromolecules 33 (2000) 3739. [39] F.A.A. Tirkistani, Polym. Degrad. Stab. 60 (1998) 67. [40] X. Qu, A. Wirsen, A.C. Albertsson, Polymer 41 (2000) 4841. [41] S.F. Wang, L. Shen, Y.J. Tong, L. Chen, I.Y. Phang, P.Q. Lim, T.X. Liu, Polym. Degrad. Stab. 90 (2005) 123. [42] H. Ma, L. Tong, Z. Xu, Z. Fang, Nanotechnology 18 (2007) 375602. [43] A.A. Kittur, S.M. Tambe, S.S. Kulkarni, M.Y. Kariduraganavar, J. Appl. Polym. Sci. 94 (2004) 2101. [44] X. Qiao, T.S. Chung, Ind. Eng. Chem. Res. 44 (2005) 8938. [45] Y.M. Lee, D. Bourgeois, G. Belfort, J. Membr. Sci. 44 (1989) 161. [46] S.T. Hwang, K. Kammermeyer, Membrane in Separation, Wiley-Interscience, New York, 1975. [47] A. Yamasaki, T. Iwatsubo, T. Masuoka, K. Mizoguchi, J. Membr. Sci. 89 (1994) 111. [48] S.P. Kusumocahyo, M. Sudoh, J. Membr. Sci. 161 (1999) 77. [49] R.Y.M. Huang, C.K. Yeom, J. Membr. Sci. 58 (1991) 33. [50] D.H. Weinkauf, D.R. Paul, W.J. Koros (Eds.), in: Barrier Polymers and Structures, ACS Symp. Ser., vol. 423, Washington, DC, 1990, p. 61.
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Development of novel composite membranes using quaternized chitosan and Na+-MMT clay for the pervaporation dehydration of isopropanol Santosh K. Choudhari, Mahadevappa Y. Kariduraganavar * Center of Excellence in Polymer Science and Department of Chemistry, Karnatak University, Dharwad 580 003, India
a r t i c l e
i n f o
Article history: Received 12 January 2009 Accepted 28 May 2009 Available online 3 June 2009 Keywords: Chitosan Na+-MMT clay Isopropanol Pervaporation Selectivity Activation energy
a b s t r a c t Novel polymer–clay-based composite membranes were prepared by incorporating sodium montmorillonite (Na+-MMT) clay into quaternized chitosan. The resulting membranes were characterized by Fourier transform infrared spectroscopy (FTIR), wide-angle X-ray diffraction (WXAD), and thermogravimetric analysis (TGA). The effect of membrane swelling was studied by varying the water concentration in the feed. The membranes were employed for the pervaporation dehydration of isopropanol in terms of feed composition and Na+-MMT clay loading. The experimental results demonstrated that membrane containing 10 mass% of Na+-MMT clay showed the highest separation selectivity of 14,992 with a flux of 14.23 102 kg/m2 h at 30 °C for 10 mass% of water in the feed. The total flux and flux of water are found to be overlapping each other particularly for clay-incorporated membranes, signifying that the composite membranes developed in the present study involving quaternized chitosan and Na+-MMT clay are highly selective toward water. From the temperature-dependent diffusion and permeation values, the Arrhenius activation parameters were estimated. The resulting activation energy values obtained for water permeation ðEpw Þ are much lower than those of isopropanol permeation ðEpIPA Þ, suggesting that the developed composite membranes have higher separation efficiency for the water–isopropanol system. The estimated Ep and ED values ranged between 8.97 and 11.89, and 7.56 and 9.88 kJ/mol, respectively. The positive heat of sorption ðDHs Þ values were obtained for all the membranes, suggesting that Henry’s mode of sorption is predominant in the process. Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction Pervaporation (PV) is an efficient membrane-based process and has gained acceptance by chemical industries over the years because of its favorable economics, easy maintenance, and simplicity of the process [1–3]. During the last decade, several commercial PV plants have been established all over the world especially for the dehydration of organics. Therefore, much of the attention has now been focused toward the development of new robust membranes with better performance. As an outcome of this, organic–inorganic hybrid materials have recently attracted considerable attention and are popularly termed as ‘‘next generation” membrane materials [4–6]. Such hybrid membranes are typically composed of polymeric and inorganic materials, in which inorganic filler with specific physicochemical characteristics is spatially dispersed into a bulk of the polymeric membranes. This seems to be a facile and feasible solution to have a good flux and high selectivity with desirable mechanical properties of inorganic materials and tunable flexibility with an excellent processability of polymeric materials [7]. Recent studies have dem* Corresponding author. Fax: +91 836 2771275. E-mail address: [email protected] (M.Y. Kariduraganavar). 0021-9797/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2009.05.071
onstrated that the dramatically improved separation properties of organic–inorganic hybrid membranes could be realized by incorporating porous or nonporous inorganic adsorbents such as zeolite, silica, carbon molecular sieve, activated carbon, and clay into the matrices of glassy and rubbery polymers [4–11]. Among these, nano-structured polymer–clay hybrid materials are currently the objects of intensive research because of their unique properties and low cost synthesis as clays are abundantly available in nature. Polymer–clay is the hybrid composite material consisting of polymer matrix with dispersed clay nanoparticles. Nano-clays have also been widely used as an inorganic reinforcement for polymer matrices with nano-scale dispersion of the inorganic phase within the polymer matrix [12–16]. Addition of a small amount of nanoclay into polymer can lead to higher moduli, increased strength, heat resistance [17–19], decreased permeability [20], and flammability [21]. The enhanced properties are presumably due to nanoscale structure effects and the interaction of inorganic–organic materials. Chitosan [poly-bð1 ! 4Þ-D-glucosamine] is an aminopolysaccharide and has attracted a great interest as a basic membrane material due to its high abundance in nature, hydrophilicity, chemical resistance, adequate mechanical strength, good membrane forming capacity, and more importantly its functional groups such
112
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
as amino and hydroxyl, which provide an opportunity for chemical modification. In fact it is an important method for improving the membrane property of pure chitosan. Several modified forms such as carboxylated, sulfated, phosphorylated [22,23], and in particular quaternized chitosan [24–26] have been well tested as membrane material for pervaporation dehydration. The quaternized chitosan showed an excellent water permselectivity for aqueous alcohol solution. However, an excess of swelling results in an increase of both solubility and diffusivity of alcohols, and consequently lowers the water permselectivity. In order to circumvent this difficulty, Uragami et al. [24–26] have tried to crosslink quaternized chitosan with different reagents such as diethylene glycol diglycidyl ether, poly(ethylene oxydiglycolic acid), and tetraethylorthosilicate, and have succeeded to some extent. On the other hand, biopolymer–clay nanocomposites represent an innovative and promising class of new materials [27–30], in which the enhancement of material property largely depends on the spatial distribution, rearrangement of intercalating polymer chains, and the type of interactions existing between polymer and clay. Sodium montmorillonite (2:1 aluminosilicate), which belongs to the general smectite group of bentonite clays, carries a distribution of ve charges in its framework, and thus interacts electrostatically with metal cations such as sodium in its interlayer galleries. It is well established that positively charged polymers, for instance, b-(dimethylamino)ethyl methacrylate hydroacetate can be intercalated in smectite clays following cation-exchange mechanisms [31,32]. Thus, the quaternized chitosan having strong +ve charges on the backbone is an excellent candidate for interaction with Na+-montmorillonite through a cation-exchange process. The opposite charges present on the polymer and clay not only help for better dispersion of clay within the polymer matrix but also establish the strong electrostatic interactions between them. Therefore, the current investigation focuses on the development of new composite hybrid membranes by dispersing different mass% of Na+-MMT clay in the quaternized chitosan. The physicochemical changes in the resulting membranes were investigated using Fourier transform infrared spectroscopy (FTIR), wide-angle X-ray diffraction (WAXD), and thermogravimetric analysis (TGA) techniques, and the membranes were successfully subjected for PV separation of water–isopropanol mixtures. The effects of clay loading, water composition in the feed, and temperature on PV performance were systematically evaluated. From the temperature dependence of permeation flux and diffusion coefficient, the Arrhenius activation parameters were estimated. The results are discussed in terms of PV separation efficiency of the membranes.
a 1-l capacity round-bottom flask fitted with a condenser. To this heterogeneous mixture, 20 ml of methyl iodide was added at once and refluxed for 8 h at 50 °C to yield quaternary iodide ammonium salt. The resulting clear solution was cooled to ambient temperature and then excess sodium chloride was added and stirred overnight to convert iodide ammonium salt to chloride ammonium salt. This solution was poured into an excess quantity of acetone while stirring. The precipitated crude quaternized chitosan was purified by dissolving in water and repeatedly reprecipitating in acetone. The quaternization of chitosan is presented in Fig. 1. The percentage degree of quaternization was estimated by recording a NMR spectrum with D2 O as solvent on a Bruker Avance 300 MHz spectrometer and it was found to be 15%. 2.3. Membrane preparation The quaternized chitosan (3 g) was dissolved in 87 ml of deareated distilled water at room temperature. The solution was filtered using a fritted glass disk filter to remove undissolved residue particles. In the meantime, a known amount of Na+-MMT clay was dispersed in 10 ml of deareated distilled water and kept in an ultrasonic bath at a fixed frequency 38 kHz (Grant XB6, UK) for 30 min to break the possible aggregated particles. The clay dispersion was slowly added into a solution of quaternized chitosan while stirring, and the resulting mixture was further stirred for overnight to obtain uniform membrane material. It was then spread onto a glass plate with the aid of a casting knife in a dust-free atmosphere. The membrane was allowed to dry at room temperature for 2–3 days and then subsequently peeled off. The amount of clay with respect to quaternized chitosan was varied from 0, 5, 10, and 15 mass%, and the membranes thus obtained were designated as M, M-1, M-2, and M-3, respectively. The thickness of these membranes was measured at different points using a Peacock dial thickness gauge (Model G, Ozaki MFG Co. Ltd., Japan) with an accuracy of ±2 lm. The thickness of the membranes was found to be 40 ± 2 lm. 2.4. Ion-exchange capacity (IEC)
2. Experimental
In order to assess the total number of positively charged groups which have electrostatically interacted with clay in the membranes, the ion-exchange capacity of the pure quaternized chitosan and its clay-incorporated composite membranes was determined. The membrane samples were completely dried and their weights were recorded. The dried membrane samples were soaked in 20 ml of 50 vol% alcoholic solution of 0.01 M sodium hydroxide. After 24 h, 10 ml of this solution was titrated against 0.01 M HCl. The ion-exchange capacity of the membranes was calculated as
2.1. Materials
IEC ¼
w 200;000; N-deacetylation degree 75–85%) and Chitosan ðM sodium montmorillonite were obtained from Sigma–Aldrich Chemicals (USA). Methyl iodide and acetone were purchased from Loba Chemicals, Mumbai, India. Isopropanol (IPA) and sodium chloride were purchased from s.d. fine Chemicals Ltd., Mumbai, India. All the chemicals are of reagent grade and were used without further purification. Double-distilled water was used throughout the study.
where B is the amount of HCl used to neutralize the blank solution; P is the amount of HCl used to neutralize the equilibrated solution; m is the weight of membrane sample; 2 is the correction factor.
2.2. Synthesis of quaternized chitosan The quaternization of chitosan was carried out with a prime aim to increase the performance of chitosan membranes. Chitosan (6 g) was dispersed while stirring at room temperature in 760 ml of aqueous-organic mixture (42 vol% methanol), which was taken in
B P MNaOH 2 m
ð1Þ
2.5. Fourier transform infrared spectroscopy The incorporation of different amounts of Na+-MMT clay material and its interaction with quaternized chitosan were confirmed by a FTIR spectrometer (Nicolet, Impact-410, USA). Membrane samples were ground well to make KBr pellets under hydraulic pressure of 400 kg/cm2 and spectra were recorded in the range of 400–4000 cm1. In each scan, the amounts of membrane sample and KBr were kept constant in order to know the changes in the intensities of characteristic peaks with respect to the amount of clay loading.
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
113
Fig. 1. Scheme for the synthesis of quaternized chitosan.
2.6. Wide-angle X-ray diffraction The morphology of the pure quaternized chitosan and its clayloaded membranes was studied at room temperature using a Philips analytical X-ray diffractometer. The X-ray source was nickelfiltered Cu-Ka radiation (40 kV, 30 mA). The dried membranes of uniform thickness ð 40 lmÞ were mounted on a sample holder and scanned in the reflection mode at an angle 2h over a range from 2° to 40° at a speed of 8°/min. 2.7. Thermogravimetric analysis Thermal properties of the quaternized chitosan and its clayincorporated composite membranes were investigated by Perkin– Elmer Diamond thermogravimetric analyzer at a heating rate of 10 °C/min under nitrogen atmosphere. The weight of the samples taken for each record was 5–9 mg. 2.8. Swelling measurements The equilibrium sorption experiments were performed in different compositions of water–isopropanol mixtures using an electronically controlled oven (WTB Binder, Germany). The masses of the dry membranes were first determined and these were equilibrated by soaking in different compositions of the mixtures in a sealed vessel at 30 °C for 24 h. The swollen membranes were weighed as quickly as possible after careful blotting on a digital microbalance (Mettler, B204-S, Toledo, Switzerland) within an accuracy of ±0.01 mg. All the experiments were performed at least three times and the results were averaged. The percentage degree of swelling (DS) was calculated as
DSð%Þ ¼
Ws Wd 100; Wd
ð2Þ
where W s and W d are the masses of the swollen and dry membranes, respectively. 2.9. Pervaporation experiments PV experiments were performed using the in-house designed apparatus reported in our previous articles [6,33]. The effective surface area of the membrane in contact with the feed mixture was 34.23 cm2 and the capacity of the feed compartment was about 250 cm3. The vacuum in the downstream side of the apparatus was maintained [1.333224 103 Pa (10 Torr)] using a twostage vacuum pump (Toshniwal, Chennai, India). The water composition in the feed mixture was varied from 5 to 25 mass%. The test membrane was allowed to equilibrate for about 2 h in the feed compartment with a known volume of feed mixture at the corresponding temperature before the PV experiment was performed. After a steady state was attained, the permeate was collected in a trap immersed in the liquid nitrogen jar on the downstream side at fixed intervals of time. The experiments were carried out at 30, 40, and 50 °C. The flux was calculated by weighing the permeate on a digital microbalance with an accuracy of ±0.01 mg. The compositions of water and isopropanol were estimated by measuring the refractive index of the permeate within an accuracy of ±0.0001 unit, using an Abbe’s refractometer (Atago-3T, Japan), and by comparing it with a standard graph, which was established previously with the known compositions of water–isopropanol mixtures. All the experiments were performed at least three times, and the results were averaged. The results of permeation for water–isopropa-
114
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
nol mixtures during the pervaporation were reproducible within an admissible range. From the PV data, separation performance of the membranes was assessed in terms of total flux (J), separation selectivity ðasep Þ, and pervaporation separation index (PSI). These were respectively calculated using the equations
W J¼ At P =P asep ¼ w IPA F w =F IPA PSI ¼ Jðasep 1Þ;
ð3Þ ð4Þ ð5Þ
where W is the mass of permeate (kg); A, the effective membrane area (m3); t, the permeation time (h); Pw and PIPA are the mass percentage of water and isopropanol in the permeate, respectively; F w and F IPA are the respective mass percentage of water and isopropanol in the feed. 3. Results and discussion 3.1. Membrane characterization 3.1.1. Ion-exchange capacity Residual ionic groups in the resulting membranes were estimated by determining the ion-exchange capacity using the titration method. It is noted that the ion-exchange capacity of quaternized chitosan membrane (M) was found to be 0.342 meq/ g. However, the clay-incorporated quaternized chitosan membranes M-1, M-2, and M-3 showed ion-exchange capacities of 0.233, 0.141, and 0.110 meq/g, respectively. This suggests that the incorporated clay material correspondingly lowers the free ionic groups in the membrane matrix, indicating that both quater-
nary ammonium groups and protonated amino groups of chitosan have strongly interacted with negative sites of clay through an electrostatic force of attraction. The establishment of electrostatic interaction is illustrated in Fig. 2. 3.1.2. FTIR studies Fig. 3 shows the FTIR spectra of quaternized chitosan and its clay-incorporated membranes. The spectrum of quaternized chitosan membrane showed that a characteristic band at around 3400 cm1 was attributed to O–H stretching vibrations and the bands appearing at around 1630 and 1570 cm1 were respectively assigned to amide I and amide II functional groups. The multiple bands appeared between 1200 and 1000 cm1 were assigned to C–O stretching. On incorporating the clay into quaternized chitosan membrane matrix, the intensity of C–O stretching increased and this was predominant with increase of the amount of clay as clearly demonstrated in spectra of M-1 to M-3. This is expected due to Si–O–Si stretching of clay material as it overlaps at the same frequency of C–O. Further, when we observed the spectra carefully, it is noted that the amino bands appearing at 1630 and 1570 cm1 slightly shifted toward a lower frequency with increase of the amount of clay, suggesting the possibility of formation of hydrogen bonds between the clay and the quaternized chitosan [34]. In addition to this, the O–H band also shifted slightly to a lower frequency, which supports the enhanced hydrogen bonding occurring between the clay and the quaternized chitosan. All this evidence confirms the incorporation of clay and its interaction with the quaternized chitosan. 3.1.3. Wide-angle X-ray diffraction studies The dispersion of clay in the quaternized chitosan membrane matrix was confirmed by WAXD, which is the most frequently used
Fig. 2. Scheme for the preparation of composite membranes.
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
115
composite membranes. This might be due to a higher crystallinity or denser packing in the main chain in comparison with neat quaternized chitosan. This is in good agreement with the results reported by Thomas and colleagues [37] and Ren et al. [38] for polymer–clay composites. This could be responsible for better pervaporation performance of composite membranes in comparison with quaternized chitosan membrane. On the other hand, the pattern of Na+-MMT clay showed a reflection peak at around 2h ¼ 6:58, corresponding to basal spacing of 1.34 nm. After incorporating Na+-MMT clay in quaternized chitosan matrix, this peak was shifted to a lower angle. This clearly indicates that quaternized chitosan chains were intercalated into the interlayers of Na+-MMT clay through a cation-exchange mechanism. This is mainly because of the hydrophilic and polycationic nature of polymer, which establishes an excellent miscibility with Na+-MMT. With further increase of the clay content in the quaternized chitosan membrane, the d-spacing values decreased systematically from membranes M-1 to M-3 (2.0, 1.92, and 1.83 nm). The decrease in d-spacing values suggests the intercalation of quaternized chitosan chains into the silicate layers. This is expected as a higher content of clay favors the intercalation. However, the intensity of this peak is very small in all the composite membranes, suggesting that a major part of the clay is in the exfoliated form.
Fig. 3. FTIR spectra of quaternized chitosan and its clay-incorporated composite membranes: (M) 0 mass%; (M-1) 5 mass%; (M-2) 10 mass%; (M-3) 15 mass% of Na+MMT clay.
method to study the structure of clay–polymer composites. Depending on the relative distribution/dispersion of the stacks of clay platelets, three types of polymer-layered silicate nanocomposites (PLSN) can be described [35,36]: intercalated PLSNs, where polymer chains are intercalated into the silicate layers resulting in a well-ordered multilayer morphology built up with alternating polymer and inorganic layers; flocculated PLSNs, where intercalated stacked silicate layers are sometimes flocculated due to the hydroxylated edge–edge interactions; exfoliated/delaminated PLSNs, where silicate layers are completely and homogeneously dispersed in the polymer matrix. Fig. 4 illustrates the WAXD patterns of Na+-MMT clay, quaternized chitosan, and clay-incorporated quaternized chitosan membranes. The pattern of quaternized chitosan membrane exhibited the characteristic crystalline peaks at 2h 12 and 23°. The intensity of these peaks increased with increasing the clay material in the membrane matrix, indicating the enhanced order structure of
3.1.4. Thermogravimetric analysis The thermal stability of quaternized chitosan and its clayloaded composite membranes was investigated by TGA under nitrogen flow. Fig. 5 shows the TGA curves of Na+-MMT clay, quaternized chitosan, and clay-loaded composite membranes. Under nitrogen flow nonoxidative degradation occurred in two stages for all membranes. The weight loss occurring between ambient temperature and 200 °C corresponds to the absorbed water molecules. Such weight loss is about 15% for Na+-MMT clay, whereas quaternized chitosan and its clay-loaded membranes exhibited slightly higher loss ranging from 20 to 25%. This fact indicates that the quaternized chitosan and its clay-loaded membranes have high water retention capacity. However, this was decreased correspondingly with increase of the amount of clay in the membrane. The second stage of decomposition occurred between 200 and 450 °C, and corresponds to weight loss of 50–55%. This is due to degradation and deacetylation of quaternized chitosan [39,40]. If we consider the decomposition temperature at 50% weight loss, the clay-loaded membranes exhibited around 5–30 °C higher temperature than that of quaternized chitosan. This clearly indicates that clay-loaded composite membranes demonstrated higher thermal decomposition temperature compared to quaternized chitosan membrane. This is because of the synergetic effect of clay platelets and quaternized chitosan on the thermal properties. The significant improvement of thermal stability of the quaternized chitosan may be explained as follows [41,42]: good heat barrier properties of clay materials for polymer matrix during formation of chars; formation of carbonaceous layer on the polymer surface at lower temperatures, which could hinder the flux of degradation product and heat flowing into underlying materials; much jammed and conjugated clay–polymer network, or strong interaction between quaternized chitosan nano-fillers could also restrict the polymer motion during heating. 3.2. Effects of feed composition and clay loading on membrane swelling The membrane swelling in certain liquids depends on the chemical composition and microstructure of the polymer, and the incorporated moiety, which strongly influence the sorption mechanism [43]. Therefore, the degree of membrane swelling is
116
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
Fig. 4. Wide-angle X-ray diffraction patterns of Na+-MMT clay, quaternized chitosan, and clay-incorporated composite membranes: (M) 0 mass%; (M-1) 5 mass%; (M-2) 10 mass%; (M-3) 15 mass% of Na+-MMT clay.
of course an important factor in the PV process, which controls the transport of permeating molecules under the driving force of chemical potential gradient. In order to study the effects of feed composition and clay loading on the membrane swelling, the percentage degree of swelling of all the membranes was plotted with respect to water composition in the feed at 30 °C as shown in Fig. 6. It is observed that the degree of swelling increased almost linearly for all the membranes with increasing the mass% of water in the feed. This is due to increased interaction between water molecules and the membrane, owing to the presence of interactive groups ð—Nþ ðCH3 Þ3 ; NHþ 3 ; NH2 , and –OH) in the membrane matrix. This is expected since water is more polar than IPA, which preferentially interacts with membrane resulting in an increased degree of swelling. On the other hand, the degree of swelling decreased with increasing the clay content in the membrane at all feed compositions. This is because, the increase of the clay content in the membrane matrix reduces the free volume due to increased packing density; secondly, it relatively lowers the amount of quaternized chitosan available for interaction. Above all, clay establishes a strong ionic interaction with quaternized chitosan leading to polymer functional groups unavailable for solvent interaction. All these together are responsible for the decreased
degree of swelling with increase of the clay content in the membrane matrix. 3.3. Effect of clay loading on pervaporation In the PV process, the overall performance of the membranes can be assessed based on the flux and selectivity. In order to calculate these parameters the developed membranes were subjected to a pervaporation experiment at 10 mass% of water in the feed. The resulting data of both flux and selectivity are presented in Fig. 7. It is observed that the clay-loaded composite membranes exhibited higher separation selectivity with lower permeation flux compared to quaternized chitosan membrane. As the content of clay was increased in the quaternized chitosan membrane, the selectivity was increased dramatically up to 10 mass% and then decreased drastically. However, the permeation flux was decreased almost linearly with increase of the amount of clay. This complex transport phenomenon can be explained on the basis of decreased free volume and establishment of a tortuous path. In polymer–clay composite membranes, the incorporated clay usually exits in the forms of exfoliated, intercalated, and flocculated states [35,36]. The extent of each state depends on the ratio between polymer
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
117
Fig. 5. Thermogravimetric analysis of Na+-MMT clay, quaternized chitosan, and clay-incorporated composite membranes: (M) 0 mass%; (M-1) 5 mass%; (M-2) 10 mass%; (M-3) 15 mass% of Na+-MMT of clay.
Fig. 7. Variation of pervaporation flux and selectivity with different mass% of Na+MMT clay loading at 10 mass% of water in the feed. Fig. 6. Variation of degree of swelling with different mass% of water in the feed for quaternized chitosan and its Na+-MMT clay-incorporated composite membranes.
and the incorporated clay. In our study, the exfoliated state was predominant at a lower content of clay as evidenced by WAXD data. This favors the molecular level interaction between the polymer and the clay, resulting in decreased free volume. In addition, a high aspect ratio of layered silicates caused them to act as a barrier, offering more resistance to diffusion by creating tortuosity to the diffusion pathway. This has suppressed the diffusion of both water and IPA molecules. However, the diffusion of IPA molecule was affected significantly as the kinetic diameter of IPA molecule is almost four times bigger than that of water molecule [44]. Obviously, transport of IPA molecules was less preferred as compared to water molecules, resulting in increased selectivity and decreased permeation flux. On further increase of the amount of clay
beyond 10 mass% both flux and selectivity were decreased. The decrease in selectivity was due to micro-phase separation occurring between organic and inorganic matrix, which has resulted from the shift of exfoliation state to intercalation followed by flocculation. The decrease in flux accounted for the lower degree of swelling at higher mass% of clay. The pervaporation separation index (PSI) is the product of total permeation flux and separation factor, which characterizes the membrane separation ability. This index can be used as a relative guideline for the design of new membranes for pervaporation separation processes and also to select a membrane with an optimal combination of flux and selectivity. Fig. 8 shows the variation of PSI as a function of mass% of clay loading at 30 °C for 10 mass% of water in the feed. It is observed that there is a dramatic increase in PSI values with increasing the clay from 0 to 10 mass%, but fur-
118
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
and resulting data are presented in Fig. 10. It is observed that the flux was increased with increase of the water composition in the feed. This is mainly because of increased membrane swelling. This is expected due to an establishment of greater interaction between the membrane and the water molecules as membrane contains a large number of hydrophilic groups such as —OH; —NH2 ; —NHþ 3, and —Nþ ðCH3 Þ3 , which preferentially interact with water molecules rather than IPA. On the contrary, the selectivity was decreased exponentially with increasing the mass% of water in the feed. At a higher concentration of water in the feed, a small amount of water dissolves in the membrane, which in turn acts as a plasticizer for the membrane, leading to more flexible polymeric chains in the matrix. In addition, as water is more polar in nature its absorption in the membrane might have weakened the interaction between polymer and clay material, resulting in a somewhat loose structure, which becomes responsible for facilitating the diffusion of some of the IPA molecules in association with water molecules. 3.5. Diffusion coefficient Fig. 8. Variation of pervaporation separation index with different mass% of Na+MMT clay loading at 10 mass% of water in the feed.
ther addition of clay (15 mass%) decreases the PSI value. This is because of a sharp decrease in selectivity arising from the microphase separation between the polymer and the clay in the membrane matrix. In order to assess the efficiency of the membranes, we have plotted the individual fluxes as a function of clay loading at 10 mass% of water in the feed as shown in Fig. 9. From the plot, it is clearly noted that the total flux and flux of water are close to each other for all the clay-loaded composite membranes. However, the closeness is more predominant with increase of the clay. This is supported by the flux of IPA, which is meager and decreased further with increase of the clay, signifying that clay-loaded composite membranes exhibited an excellent separation performance.
In the PV process, mass transport of binary liquid mixtures through a nonporous polymeric membrane is generally described by the solution-diffusion mechanism, which occurs in three steps: sorption, diffusion, and evaporation [45]. Thus, permeation flux and selectivity are governed by the solubility and diffusivity of each component of the feed mixture to be separated. During the process, because of the establishment of fast equilibrium distribution between bulk feed and the upstream surface of a membrane, the diffusion step controls the transport of penetrants [9,46]. Therefore, it is important to estimate the diffusion coefficient ðDi Þ of penetrating molecules to understand the mechanism of transport. From Fick’s law of diffusion, the diffusion flux can be expressed as [47]
J i ¼ Di
dC i dx
;
ð6Þ
In PV, the effect of feed is also an important factor, and thus we have carried out a PV study at different mass% of water in the feed
where J is the permeation flux per unit area (kg/m2 s), D is the diffusion coefficient (m2/s), C is the concentration of permeant (kg/ m3), subscript i stands for water or isopropanol, and x is the diffusion length (m). For simplicity, it is assumed that the concentration profile along the diffusion length is linear. Thus, Di can be calculated with the equation [48]
Fig. 9. Variation of total flux, and fluxes of water and isopropanol with different mass% of Na+-MMT clay loading at 10 mass% of water in the feed.
Fig. 10. Variation of pervaporation flux and selectivity at different feed compositions of water for 10 mass% of Na+-MMT clay-incorporated membrane.
3.4. Effect of feed composition on pervaporation
119
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
Di ¼
Ji d ; Ci
ð7Þ
where d is the membrane thickness. The calculated values of Di at 30 °C for membrane (M-2) containing 10 mass% of Na+-MMT clay are plotted as function of water as shown in Fig. 11. With increasing water composition, the diffusion coefficient of water was decreased while increasing the diffusion coefficient of isopropanol. This is because of a weaker interaction between polymer and clay material, which facilitates the transport of IPA molecules.
Table 1 Pervaporation flux and separation selectivity at different temperatures for different membranes at 10 mass% of water in the feed. Temp. (°C)
30 40 50
J 102 ðkg=m2 hÞ
asep
M
M-1
M-2
M-3
M
M-1
M-2
M-3
16.68 18.41 20.63
15.60 17.49 19.49
14.27 16.37 18.41
11.58 13.42 15.63
1088 817 638
2423 1658 1089
14992 8991 3741
8991 3741 1755
3.6. Effect of temperature on pervaporation The effect of operating temperature on the pervaporation performance was studied for all the membranes at 10 mass% of water in the feed, and the values thus obtained are presented in the Table 1. It is observed that the permeation rate was found to increase from 30 to 50 °C for all the membranes while decreasing the separation selectivity. The increased permeation flux is mainly due to increased driving force at higher temperatures. This is because of the combined effects of both decreased cohesive forces between permeants, and increased vapor pressure at the feed side. However, the interaction between permeants and the membrane decreased at higher temperatures, which led to the permeation of both water and IPA molecules, resulting in decreased selectivity. This effect prompted us to estimate the activation energies for permeation and diffusion using the following Arrhenius-type equation [49],
X ¼ X o exp
Ex ; RT
ð8Þ
where X represents permeation (J) or diffusion (D), X o is a constant representing preexponential factor of Jo or Do . Ex represents activation energy for permeation or diffusion depending on the transport process under consideration and RT is the usual energy term. Arrhenius plots of log J and log D versus 1/T are shown in Figs. 12 and 13 for the temperature dependence of permeation flux and diffusion, respectively. In both cases, a linear behavior was observed, suggesting that both permeability and diffusivity follow an Arrhenius trend. From least-squares fits of these linear plots, the activation energies for permeability ðEp Þ and diffusivity ðED Þ were estimated. Similarly, we have also estimated the activation energies for diffusion of water ðEDw Þ, permeation of water ðEpw Þ, and iso-
Fig. 12. Variation of log J with temperature for different mass% of Na+-MMT clayloaded composite membranes at 10 mass% of water in the feed.
Fig. 13. Variation of log D with temperature for different mass% of Na+-MMT clayloaded composite membranes at 10 mass% of water in the feed.
Fig. 11. Variation of Dw and DIPA at different feed compositions of water for 10 mass% of Na+-MMT clay-incorporated membrane.
propanol ðEpIPA Þ, but the plots are not given to save space. The results thus obtained are presented in Table 2. From Table 2, it is observed that the apparent activation energy values of water ðEpw Þ are much lower than those of isopropanol ðEpIPA Þ, suggesting that membranes have higher separation efficiency toward water. The activation energy values for water permeation ðEpw Þ and total permeation ðEp Þ are almost close to each
120
S.K. Choudhari, M.Y. Kariduraganavar / Journal of Colloid and Interface Science 338 (2009) 111–120
Table 2 Arrhenius activation parameters for permeation and diffusion, and heat of sorption. Parameters (kJ/mol)
M
M-1
M-2
M-3
Ep ED Epw EpIPA EDw DHs
8.97 7.56 8.64 30.00 6.51 1.41
9.47 7.72 9.08 41.33 7.02 1.75
10.91 9.07 10.38 66.41 8.13 1.84
11.89 9.88 11.52 77.49 9.21 2.01
other, signifying that coupled transport of both (water and isopropanol molecules) is minimal as due to a higher selective nature of membranes. The estimated Ep and ED values ranged between 8.97 and 11.89, and 7.56 and 9.88 kJ/mol, respectively. Using these values, we have further calculated the heat of sorption as
DH s ¼ E p E D :
ð9Þ
The resulting DHs values are included in Table 2. The DHs values give additional information about the transport of molecules through the polymer matrix. It is a composite parameter involving contributions of both Henry and Langmuir types of sorption [50]. The Henry type of sorption requires both the formation of a site and the dissolution of chemical species into that site. The formation of a site involves an endothermic contribution to the sorption process. However, the Langmuir type of sorption requires the preexistence of a site in which sorption occurs by a hole-filling mechanism, giving an exothermic contribution. The DHs values obtained in the present study are positive for all the membranes, suggesting that Henry’s sorption is predominant, giving an endothermic contribution. 4. Conclusions Novel composite membranes were prepared using quaternized chitosan and Na+-MMT clay. Thermogravimetric study reveals that composite membranes exhibited better thermal stability compared to quaternized membrane. The performance of membranes was evaluated for the separation of water–IPA mixtures. An increase of Na+-MMT clay content in the quaternized chitosan matrix results in a decrease of permeation flux and an increase of separation selectivity. This was explained on the basis of reduction in free volume and establishment of a tortuous pathway. While assessing the membranes’ efficiency, it is clearly noted that both total flux and flux of water are overlapping each other particularly for clay-incorporated composite membranes, signifying that the composite membranes developed in the present study are highly selective toward water. The PV separation index data also indicated that the clay-incorporated membranes except membrane M-3 demonstrated an excellent PV performance. Among the membranes, the membrane containing 10 mass% of Na+-MMT clay exhibited the highest separation selectivity of 14,992 and with a permeation flux of 14.23 102 kg/m2 h at 30 °C for 10 mass% of water in the feed. With regard to increase in temperature the permeation rate was found to increase while suppressing the selectivity. This was attributed to increased vapor pressure and decreased interaction between permeants and membrane at higher temperatures. A significant difference was noted between Epw and EpIPA values. However, the difference was more significant particularly for the membranes having higher loading of Na+-MMT clay, suggesting
that the composite membranes developed with higher loading of clay exhibited remarkable separation ability toward water. The Ep and ED values ranged between 8.97 and 11.89, and 7.56 and 9.88 kJ/mol. All the membranes exhibited positive DHs values, indicating that sorption is mainly dominated by the Henry mode of sorption, giving an endothermic contribution.
References [1] P.D. Chapman, T. Oliveira, A.G. Livingston, K. Li, J. Membr. Sci. 318 (2008) 5. [2] Z. Huang, Y. Shi, R. Wen, Y.H. Guo, J.F. Su, T. Matsura, Sep. Purif. Technol. 51 (2006) 126. [3] P.D. Chapman, X. Tan, A.G. Livingston, K. Li, T. Oliveira, J. Membr. Sci. 268 (2006) 13. [4] F. Peng, L. Lu, H. Sun, Y. Wang, H. Wu, Z. Jiang, J. Membr. Sci. 275 (2006) 97. [5] T. Uragami, K. Okazaki, H. Matsugi, T. Miyata, Macromolecules 35 (2002) 9156. [6] M.Y. Kariduraganavar, S.S. Kulkarni, A.A. Kittur, J. Membr. Sci. 246 (2005) 83. [7] F. Peng, L. Lu, H. Sun, F. Pan, Z. Jiang, Ind. Eng. Chem. Res. 46 (2007) 2544. [8] Y.L. Liu, C.Y. Hsu, Y.H. Su, J.Y. Lai, Biomacromolecules 6 (2005) 368. [9] A.A. Kittur, M.Y. Kariduraganavar, U.S. Toti, K. Ramesh, T.M. Aminabhavi, J. Appl. Polym. Sci. 90 (2003) 2441. [10] F. Peng, F. Pan, H. Sun, L. Lu, Z. Jiang, J. Membr. Sci. 300 (2007) 13. [11] S. Anilkumar, M.G. Kumaran, S. Thomas, J. Phys. Chem. B 112 (2008) 4009. [12] S. Takahashi, D.R. Paul, Polymer 47 (2006) 7535. [13] R. Stephen, C. Ranganathaiah, S. Varghese, K. Joseph, S. Thomas, Polymer 47 (2006) 858. [14] Z.F. Wang, B. Wang, N. Qi, H.F. Zhang, L.Q. Zhang, Polymer 46 (2005) 719. [15] M.S. Hedenqvist, A. Backman, M. Gallstedt, R.H. Boyd, U.W. Gedde, Compos. Sci. Technol. 66 (2006) 2350. [16] H. Heinz, R.A. Vaia, R. Krishnamoorti, B.L. Farmer, Chem. Mater. 19 (2007) 59. [17] R.A. Vaia, R.K. Teukolsky, E.P. Giannelis, Chem. Mater. 6 (1994) 1017. [18] R.A. Vaia, E.P. Giannelis, Macromolecules 30 (1997) 7990. [19] R.A. Vaia, E.P. Giannelis, Macromolecules 30 (1997) 8000. [20] M.A. Osman, J.E.P. Rupp, U.W. Suter, J. Mater. Chem. 15 (2005) 1298. [21] E. Manias, A. Touny, L. Wu, K. Strawchecker, B. Lu, T.C. Chung, Chem. Mater. 13 (2001) 3516. [22] Y.M. Lee, Desalination 90 (1993) 277. [23] Y.M. Lee, E.M. Shin, J. Membr. Sci. 64 (1994) 145. [24] T. Uragami, M. Takumo, T. Miyata, Macromol. Chem. Phys. 203 (2002) 1162. [25] T. Uragami, S. Yamamoto, T. Miyata, Biomacromolecules 4 (2003) 137. [26] T. Uragami, T. Katayama, T. Miyata, H. Tamura, T. Shiraiwa, A. Higuchi, Biomacromolecules 5 (2004) 1567. [27] M. Darder, M. Colilla, E. Ruiz-Hitzky, Chem. Mater. 15 (2003) 3774. [28] M. Darder, M. Colilla, E. Ruiz-Hitzky, Appl. Clay Sci. 28 (2005) 199. [29] H. Gecol, E. Ergican, P. Miakatsindila, J. Colloid Interface Sci. 292 (2005) 344. [30] H. Gecol, P. Miakatsindila, E. Ergican, S.R. Hiibel, Desalination 197 (2006) 165. [31] R.A. Ruehrwein, D.W. Ward, Soil Sci. 73 (1952) 485. [32] C. Breen, Appl. Clay Sci. 15 (1999) 187. [33] M.Y. Kariduraganavar, A.A. Kittur, S.S. Kulkarni, K. Ramesh, J. Membr. Sci. 238 (2004) 165. [34] C. Tang, N. Chen, Q. Zhang, K. Wang, Q. Fu, X. Zhang, Polym. Degrad. Stab. 94 (2009) 124. [35] M. Alexandre, P. Dubois, Mater. Sci. Eng. R. 28 (2000) 1. [36] S.S. Ray, M. Okamoto, Prog. Polym. Sci. 28 (2003) 1539. [37] Y.H. Ha, Y. Kwon, T. Breiner, E.P. Chen, T. Tzianetopomon, R.E. Cohen, M.C. Boyce, E.L. Thomas, Macromolecules 38 (2005) 5170. [38] J. Ren, A. Silva, R. Krishnamoorti, Macromolecules 33 (2000) 3739. [39] F.A.A. Tirkistani, Polym. Degrad. Stab. 60 (1998) 67. [40] X. Qu, A. Wirsen, A.C. Albertsson, Polymer 41 (2000) 4841. [41] S.F. Wang, L. Shen, Y.J. Tong, L. Chen, I.Y. Phang, P.Q. Lim, T.X. Liu, Polym. Degrad. Stab. 90 (2005) 123. [42] H. Ma, L. Tong, Z. Xu, Z. Fang, Nanotechnology 18 (2007) 375602. [43] A.A. Kittur, S.M. Tambe, S.S. Kulkarni, M.Y. Kariduraganavar, J. Appl. Polym. Sci. 94 (2004) 2101. [44] X. Qiao, T.S. Chung, Ind. Eng. Chem. Res. 44 (2005) 8938. [45] Y.M. Lee, D. Bourgeois, G. Belfort, J. Membr. Sci. 44 (1989) 161. [46] S.T. Hwang, K. Kammermeyer, Membrane in Separation, Wiley-Interscience, New York, 1975. [47] A. Yamasaki, T. Iwatsubo, T. Masuoka, K. Mizoguchi, J. Membr. Sci. 89 (1994) 111. [48] S.P. Kusumocahyo, M. Sudoh, J. Membr. Sci. 161 (1999) 77. [49] R.Y.M. Huang, C.K. Yeom, J. Membr. Sci. 58 (1991) 33. [50] D.H. Weinkauf, D.R. Paul, W.J. Koros (Eds.), in: Barrier Polymers and Structures, ACS Symp. Ser., vol. 423, Washington, DC, 1990, p. 61.