- Email: [email protected]
DMC mixtures
Journal of Colloid and Interface Science 316 (2007) 580–588 www.elsevier.com/locate/jcis
Composite hybrid membrane of chitosan–silica in pervaporation separation of MeOH/DMC mixtures Jian Hua Chen, Qing Lin Liu ∗ , Jun Fang, Ai Mei Zhu, Qiu Gen Zhang Department of Chemical and Biochemical Engineering, College of Chemistry & Chemical Engineering, Xiamen University, Xiamen 361005, China Received 14 May 2007; accepted 9 September 2007 Available online 14 September 2007
Abstract Chitosan–silica hybrid membranes (CSHMs) were prepared by cross-linking chitosan (CS) with 3-aminopropyl-triethoxysilane (APTEOS). The dynamic behaviors of the CS membrane and the CSHM were investigated in pervaporation (PV) of methanol/dimethyl carbonate (MeOH/DMC) mixtures. The membranes were characterized by X-ray diffraction (XRD), contact angle meter, scanning electron microscopy (SEM) and differential scanning calorimetry (DSC). The transition state of PV processes were studied. During the PV processes, the amorphous region of the membranes increases and the contact angle between MeOH and the membrane decreases within a range of operating time and then remains almost constant implying a reconstruction occurred on the membrane surface. The silica is well distributed in the CSHM matrix and the thermal stability of the CSHM is enhanced. The time for a PV process to reach a steady state decreases with increasing MeOH concentration or feed temperature, and it is longer for the CSHM than the CS membrane under the same operating condition. Swelling experiments show that the degree of swelling (DS) is greatly depressed by cross-linking CS with APTEOS. Sorption data indicate that the selectivity of solubility and diffusion of the CSHM are greatly improved over the CS membrane. The CSHM presents superior separation behaviors over other membranes with a flux of 1265 g/(h m2 ) and separation factor of 30.1 in PV separation of 70 wt% MeOH in feed at 50 ◦ C. © 2007 Elsevier Inc. All rights reserved. Keywords: Chitosan; Hybrid membranes; Surface reconstruction; MeOH; Dimethyl carbonate
1. Introduction Recently, separation of organic/organic mixtures using PV techniques is being investigated extensively owing to its great importance in chemical and petrochemical industries. The separation of organic/organic mixtures, which may roughly be classified as follows: (1) separation of polar/non-polar solvent mixtures, (2) separation of aromatic/aliphatic mixtures, (3) separation of aliphatic hydrocarbons, and (4) separation of aromatic isomers. Due to very similar chemical and physical properties between the organic components, the organic mixture was difficult to be separated by distillation or other membrane processes. Dimethyl carbonate (DMC) is an environmental benign chemical compound and unique intermediate with versatile * Corresponding author. Fax: +86 592 2184822.
E-mail address: [email protected] (Q.L. Liu). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.09.022
chemical reactivity which has attracted increasing interest from both practical and fundamental points of view in recent years [1,2]. It can be used as an intermediate to substitute for highly toxic phosgene and dimethyl sulfate in many chemical processes owing to the presence of two methyl groups and one carbonyl group in its molecule. Because of its high oxygen content [3], low toxicity, rapid biodegradability and low impact on air quality, DMC has been proposed as a replacement for methyl tert-butyl ether (MTBE) as a fuel additive. Currently, the popular synthesis of DMC based on the oxycarbonylation of MeOH [4,5] has been successfully developed up to the industrial scale: (1) the oxidative carbonylation of MeOH with carbon monoxide and oxygen catalyzed by cuprous chloride, and (2) an oxidative carbonylation process using a palladium catalyst and methyl. Indeed, excess MeOH is used for a high conversion yield. The use of excess MeOH, however, causes a purification problem because MeOH forms an azeotrope with DMC at different MeOH content under different
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
pressure. Separation is an important step in DMC manufacturing. Various separation processes (e.g., extractive distillation, pressure swing distillation, liquid–liquid extraction, adsorption on zeolites, and low temperature crystallization) have been proposed to separate and purify DMC; however, few literatures about the separation of MeOH/DMC mixtures by PV are available. Hydrophobic polymers such as polyethylene and polypropylene [6] were used mainly as membrane materials. However, these hydrophobic materials did not show a high selectivity for organic mixtures, because they have slight differences in the interaction with one component from the other in the mixtures. As for MeOH/DMC mixtures, MeOH is more hydrophilic and polar than DMC due to dipole moment of hydroxyl group. As a result, the interaction between the polar material and MeOH should be stronger than DMC, which results in larger affinity of MeOH for polar or hydrophilic polymer materials. Therefore, it should be possible to improve MeOH selectivity by using a hydrophilic membrane. Furthermore, MeOH is more permeable than DMC due to the difference in the molecular size. Accordingly, the molecular size and polarity difference should be a key factor to obtain a higher permselectivity. CS is widely used in membrane applications because of its high hydrophilicity, good film-forming character and excellent chemical-resistant properties. CS is also a cyclo-aliphatic polymer. It contains active amino groups and hydroxyl groups that can be chemically modified by several chemicals such as glutaraldehyde [7], epichlorohydrin [8], sulfuric acid [9], GPTMS [10,11] and dialdehyde starch [12], etc. CS-based membranes have been extensively studied not only for the dehydration of alcohols [13], other industrial solvents like tetrahydrofuran, 1,4-dioxane, isopropanol and ethylene glycol [14,15], but also for the separation of organic/organic mixtures MeOH/MTBE [16,17], alcohol/toluene [18], benzene/cyclohexane [19] and MeOH/DMC [20,21]. A PV process consists of three consecutive steps: (I) feed components sorption into the membranes at the feed side, (II) permeant diffusion through the membranes, and (III) permeant desorption into a vapor phase at the permeate side. Desorption is very quick since the pressure at the permeate side is very low, so a PV process is controlled mainly by the first two steps. Therefore, study on the surface properties of the membrane during PV processes is extremely important, since PV performance is greatly dependent on the surface properties of the membrane. For this purpose, a novel CSHM with APTEOS content of 10 wt% (this content proved to be the optimum in our previous work [22]) were prepared by cross-linking APTEOS with CS so as to enhance both flux and permselectivity simultaneously. The surface properties such as hydrophilicity and crystallinity of the CS membrane and the CSHM coupled with permeation properties during PV processes were investigated by using a contact angle meter and X-ray diffraction (XRD) method. Moreover, the effects of feed concentration and temperature on the membrane dynamic behaviors were also studied.
581
2. Experimental 2.1. Materials CS with a degree of deacetylation of approximately 92% (Yuhuan Ocean Biochemical Co. Ltd., China) and 3-aminopropyl-triethoxysilane (APTEOS) (Shanghai Yaohua Chemical Plant, China), were used without further purification. The other solvents and reagents, of analytical grade and used without further purification, were purchased from Sinophatm Chemical Reagent Co. Ltd. 2.2. Membrane preparation Chitosan was dissolved in a 2.0 wt% acetic acid aqueous solution to form a solution of 2.0 wt% chitosan. The solution was stirred at room temperature for 2 h, filtered, and then a certain amount of APTEOS as well as concentrated HCl as a catalyst were added to initiate the cross-linking reaction. The mixture was stirred at 50 ◦ C for about 12 h. The resulted homogeneous solution was cast onto a glass plate with a casting knife, and then dried in an oven at 50 ± 1 ◦ C with relative humidity 65 ± 2% for about 4 h. The obtained membranes were peeled off gently, and were further dried at 70 ◦ C for 8 h. The thickness of membrane, measured by SEM, is 20 ± 0.5 µm. The mass ratio of APTEOS to CS was 10 wt% and designated as CSHM. 2.3. Membrane characterization Crystal structure characterization was carried out by XRD (Panalytical X’pert Philip, Holland) using CuKα radiation. The diffraction was operated at 40 kV and 30 mA at a 2θ range of 5–40◦ , using a step size of 0.0167◦ and a counting time of 10 s per step. The morphology of surface and thickness of the membranes were measured by using SEM (LEO 1530, Oxford Instruments), which was operated at EHT = 20 kV. DSC was performed using a Netzsch DSC 204 Phoenix. Accurately weighted 5 mg samples were placed into aluminum cups and heated from 25 to 400 ◦ C at a constant heating rate of 10 ◦ C/min under constant nitrogen purging at 20 mL/min. Static contact angles between MeOH and the CS membranes and/or the CSHMs were measured by the pendant drop method using a contact angle meter (SL200B, SOLON TECH, Shanghai, China) at 23 ± 1 ◦ C with 65 ± 2% relative humidity. All reported values are the average of eight measurements taken at different location on the membrane surface. The errors are less than 4%. 2.4. Pervaporation measurement PV experiments were performed on the PERVAP 2201 (SULZER CHEMTECH, Germany) at 30, 40 and 50 ◦ C, respectively. The effective membrane area was 62.21 cm2 . The membrane was supported on a filter paper over a porous sintered steel disk of 6.4 cm in diameter. The vacuum at the permeate side was maintained 300 Pa using a vacuum pump, and
582
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
the permeate was collected in a liquid nitrogen trap. The concentrations of both the feed and the permeate were measured by gas chromatography (GC-950 Haixin, Shanghai, China) equipped with a thermal conductivity detector (TCD) and a column packed with GDX102. The results for PV separation of MeOH/DMC mixtures were reproducible, and the errors in the PV measurements are less than 1.5%. The separation performances of the membranes can be evaluated on the basis of total flux (J ) and separation factor (a). Q , A·t PM /PD aPV = , FM /FD Jp =
H2 NCH2 CH2 Si–(OC2 H5 )3 + XH2 O → H2 NCH2 CH2 Si–(OC2 H5 )3−x (OH)x + xC2 H5 OH Scheme 1. Hydrolysis reaction for APTEOS.
R O
Si
R OR1 +
O
R
R Δ
−→
O
OH
O
O
(1)
Si
Si
O
Si
O
+ R1 OH
(2) O
where Q is the weight of the permeate collected within time t, and A is the effective membrane area, PM and PD are the mass fraction of MeOH and DMC in the permeate, FM and FD are that of MeOH and DMC in the feed, respectively.
O
Scheme 2. Condensation reaction for APTEOS. R = –CH2 CH2 NH2 and R1 = –H or –CH2 –CH3 .
2.5. Swelling measurements and sorption experiments Before each experiment the pre-dried CS membrane and CSHM were weighed and immersed in MeOH/DMC mixtures at a desired temperature for 48 h to reach an equilibrium swelling. The sample was taken out at appropriate intervals (about 3 h), wiped with tissue paper carefully to remove the surface solvent, and weighed as quickly as possible, then dipped again into the liquids. The experiments continued until the weight of the sample keeping approximately constant. All experiments were repeated at least for three times, and the results were averaged. The errors are no more than 5%. The DS was calculated by MW − MD × 100%, DS (%) = MD
3. Results and discussion (3) 3.1. Chemical structure of APTEOS/CS hybrid membranes
where MD and MW denote the mass of the dried and swollen membranes, respectively. For sorption in a binary solution, the sorbed liquid was collected in a liquid nitrogen trap by desorbing the equilibrated sample in the purge-and-trap apparatus, and the concentration of the collected liquid was measured by gas chromatography. The solubility selectivity, as , is expressed by αs =
XM /XD , LM /LD
(4)
where XM and XD are the mass fraction of MeOH and DMC in the membrane; LM and LD are the mass fraction of MeOH and DMC in the binary solution. According to the solution–diffusion mechanism, diffusion selectivity, aD , was calculated by aD =
aPV . aS
Scheme 3. Chemical structure and reaction model for cross-linking chitosan.
(5)
The chemical structures of chitosan and APTEOS and their cross-linking reaction are schematically illustrated in Schemes 1, 2 and 3. In preparing the hybrid membranes, APTEOS was first hydrolyzed in the presence of an acid catalyst (HCl), leading to the formation of silanol groups. Then, the hydroxyl groups in one silanol formed siloxane bonds with those in another silanol or with the amine groups in CS through a dehydration or dealcoholysis reaction during the membrane drying. Also, the hydroxyl groups and amine groups in the silanol can form hydrogen bonds with the dissociative hydroxyl groups or amine groups in the CS amorphous region, and these hydrogen and siloxane bonds are the cross-linking spots in the hybrid membranes. 3.2. Membrane characterization The change of the crystal structure of the membrane during PV processes was investigated. Figs. 1 and 2 show the XRD
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
583
(a)
Fig. 1. XRD patterns of the CS membranes immersed in MeOH/DMC mixtures at 50 ◦ C with various MeOH mass fractions: (a) without being immersed, (b) 30, (c) 50 and (d) 70 wt%.
(b)
(c) Fig. 3. SEM images of (a) the CS membrane, (b) the CSHM and (c) the cross-sectional of CSHM (×2000). Fig. 2. XRD patterns of the CSHMs immersed in MeOH/DMC mixtures at 50 ◦ C with various MeOH mass fractions: (a) without being immersed, (b) 30, (c) 50 and (d) 70 wt%.
spectra for the CS membrane and the CSHM with or without being immersed in MeOH/DMC solutions, respectively. Comparing Fig. 2 with Fig. 1, we find that the diffraction peaks located at 11.5 and 18.3◦ vanished with APTEOS being introduced. There is only one intense diffraction peak at 21.2◦ appeared for the CSHM. This is attributed to the arrangement of the CS chains that is disturbed by the introduced APTEOS, and then decreased the crystallinity of the CSHM. Figs. 1 and 2 suggest that the higher the feed MeOH concentration is, the less the crystallinity of the CS membranes and the CSHMs. This is because more MeOH molecules interact with the polymer matrix with MeOH concentration increasing, some of the formed hydrogen-bonding, between the –OH and/or –NH2 groups of intra-molecular or inter-molecular of CS, which caused the CS chains to become more compact, was broken. This further caused the polymer chain less compact. The SEM images of the CS membrane and the CSHM are reflected in Figs. 3a and 3b and the image of the cross-section of
Fig. 4. DSC thermo-grams of the CS membrane and the CSHM.
the CSHM is shown in Fig. 3c. It indicates that the silica phase dispersed well into the CSHM matrix, and the thickness of the membrane is 20.4 µm. Fig. 4 presents the thermo-grams of
584
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
(a)
(a)
(b) (b) Fig. 5. Contact angle between MeOH and the CS membranes as a function of time immersed in MeOH/DMC mixtures with (a) MeOH mass fraction range of 30 to 70 wt% at 30 ◦ C. (b) 30 wt% MeOH mass fraction at the temperature range of 30 to 50 ◦ C.
the CS membrane and the CSHM, showing initial endothermic peaks at 55.5 and 50.4 ◦ C and higher exothermic peaks at 280.3 and 291.8 ◦ C, respectively. The endothermic peaks are corresponding to the loss of water associated with the hydrophilic groups of polymers while the exothermic peaks result from the degradation of the CS chains. It suggests that the CSHM became more thermally stable than that of the CS membrane. The contact angle between MeOH and the membrane was measured to investigate the dynamic properties of the membrane during PV processes. The samples coated onto the glass slide were submersed in MeOH/DMC mixtures at 30 ◦ C and the concentration of MeOH was changed from 30 to 70 wt%; otherwise, when MeOH concentration was kept at 30 wt%, the temperature of solution was changed from 30 to 50 ◦ C for a desired time (such as 0, 10, 20, 30 min, etc.). The sample was removed from the MeOH solution and dried with nitrogen stream prior to contact angle measurement. It can be seen from Fig. 5a that
Fig. 6. Contact angle between MeOH and the CSHMs as a function of time immersed in MeOH/DMC mixtures with (a) MeOH mass fraction range of 30–70 wt% at 30 ◦ C. (b) 30 wt% MeOH mass fraction at the temperature range of 30–50 ◦ C.
the time for the contact angle to reach a steady value decreased from 60 to 20 min, and the steady contact angle decreased from 75.13 to 69.24◦ , when the MeOH concentration varied from 30 to 70 wt%. These are attributed to the surface reconstruction of the membrane when it contacts with MeOH/DMC mixtures. With increasing MeOH concentration, more MeOH molecules would touch the membrane and induce more hydroxyl groups of the membrane to move toward the feed side, thus the hydrophilicity of the membrane was improved. Moreover, some of the formed hydrogen-bonding, between the –OH and/or –NH2 groups of intra-molecular or inter-molecular of CS, was broken because of the interaction of MeOH molecules with the polymer matrix in the membrane. This resulted in a less compact of polymer matrix, which increased the mobility of the CS chains, and thus reduced the time for both hydroxyl groups to reconstruct and the contact angle to reach a steady value. The effect of temperature on the surface reconstruction is shown in Fig. 5b. The time for the contact angle to reach a steady value
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
decreased with increasing feed temperature. This is attributed to the increase of the rate of reconstruction resulting from the improvement of mobility of the CS chains as the temperature increased. The CSHM was also studied to investigate the effect of cross-linking on the surface reconstruction, as presented in Fig. 6. Comparing with Fig. 5a, Fig. 6a shows the contact angle between MeOH and the CSHM shared the same changing trend as the CS membrane. However, the initial value of the contact angle between MeOH and the CSHM (73.5◦ ) is lower than that of the CS membrane (80.35◦ ) and the steady contact angle for the CSHM is lower than that for the CS membrane under the same condition. It may be attributed to the increase of the hydrophilic groups, e.g., –OH groups introduced when the APTEOS is hydrolyzed. Upon comparing Fig. 6 with Fig. 5, we find that the time for the contact angle for the CSHM to reach a steady value is longer than that for the CS membrane under the same condition. This is attributed to the decrease of the rate of reconstruction resulting from the decrease of the mobility of the CS chains when the CS is cross-linked with APTEOS.
585
(a)
3.3. Effects of feed concentration and temperature on the CS and CSHM swelling Generally speaking, cross-linking limits the mobility of polymer chains and thus excessive membrane swelling can be suppressed. Fig. 7a shows the effect of feed MeOH concentration on the DS of the CS membrane and the CSHM at 30 ◦ C. The DS of the CS membrane and the CSHM decreased with MeOH concentration decreasing. This is attributed to the hydrophilicity nature of the CS-based membranes (the properties of MeOH is similar to that of water). Fig. 7b shows the DS of the CS membrane and the CSHM in 30 wt% MeOH solution vs. the temperature. The DS of the CS membrane and the CHSM increased with temperature increasing. This is owing to the increase of the mobility of the CS chains with temperature increasing, which leads to an increase of the free volume in the polymer network, and results in an increase in permeant sorption. Fig. 7 also shows that under the same condition, the DS of the CSHM is less than the CS. This is due to the cross-linking between the CS and APTEOS which tends to make the amorphous region of the CSHM more compact. 3.4. Sorption and diffusion properties Based on the experimental sorption data, the selectivity of solubility and diffusion was calculated from Eqs. (4) and (5). The dependence of the sorption behavior of the CS membrane and the CSHM on feed MeOH concentration was investigated, as shown in Fig. 8a. It shows that both the solubility and diffusion selectivities decreased almost linearly with increasing MeOH concentration in the range of 30–70 wt%, and both the solubility and diffusion selectivities of the CSHM is higher than those of the CS membrane. The decrease in the solubility and diffusion selectivities is due to the increase in the DS. The solubility selectivity of the CSHM is higher than that of the CS membrane owing to the higher hydrophilicity of the CSHM; and the diffusion selectivity of the CSHM is higher than that
(b) Fig. 7. (a) The effect of MeOH concentration on the DS of the CS membranes and the CSHMs at 30 ◦ C. (b) The effect of feed temperature on the DS of the CS membranes and the CSHMs in 30 wt% MeOH solution.
of the CS membrane attributing to the more compact amorphous regions of the CSHM. The dependence of the solubility and diffusion selectivities on feed temperature is also plotted in Fig. 8b. Both the solubility and diffusion selectivities of the CS membrane and the CSHM decrease with increasing the temperature. This is attributed to an increase in the DS with the feed temperature increasing. 3.5. Pervaporation characteristics 3.5.1. Effect of feed concentration on time-dependent separation property during PV processes Feed concentration is one of the important factors during PV processes. PV performance of the CS membrane as a function of operating time is studied for MeOH concentration range of 30–70 wt% at 30 ◦ C. The experimental results are illustrated in Fig. 9a. It is found that, taking MeOH concentration of 30 wt% as an example, permeation flux increases from 150 to 170 g/(h m2 ) and separation factor increases from 25.3 to 28
586
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
(a)
(a)
(b) (b) Fig. 8. (a) The effect of MeOH concentration on the solubility and diffusion selectivities of the CS membrane and the CSHM at 30 ◦ C. (b) The effect of feed temperature on the solubility and diffusion selectivities of the CS membrane and the CSHM for 30 wt% MeOH in feed.
correspondingly, in about 60 min, and then both remain almost unchanged within our experimental time. The increase of the flux is attributed to that the crystallinity of the CS membrane decreased during PV processes, which is proved by XRD. Whereas, the improvement of separation factor is attributed to the increase of CS membrane hydrophilicity which has been proved by the contact angle experiments. Fig. 9a also indicates that the flux increases and the separation factor decreases with increasing MeOH concentration. The increase in the flux is due to the DS increase; while the decrease in the separation factor is attributed to the decrease of the solubility and diffusion selectivities. We can also observe that the time for a PV process to reach a steady state decreases with MeOH concentration increasing. This is similar to the effect of MeOH concentration on the time for the contact angle to reach a steady value mentioned above. A further investigation of such PV behaviors was performed for the CSHM, as shown in Fig. 9b. It shows that MeOH concentration dependence of PV performance of the CSHM with operating time is similar to that of the CS membrane. However, still refer to MeOH concentration of 30 wt%, the time (about 70 min) for the flux and separation factor of the CSHM
Fig. 9. (a) Permeation flux and separation factor of the CS membrane as function of time in PV of MeOH concentration range of 30–70 wt% at 30 ◦ C. (b) Permeation flux and separation factor of the CSHM as function of time in PV of MeOH concentration range of 30–70 wt% at 30 ◦ C.
to reach an equilibrium value (from 850 to 900 g/(h m2 ) for the flux and 46 to 49 for the separation factor, correspondingly) is longer than that of the CS membrane (about 60 min). This is because the cross-linking between APTEOS and CS makes the CS chains more compact, slowed down the mobility of the CSHM chains, hence raised the time for surface reconstruction. Meanwhile, the permeation flux and separation factor of the CSHM is higher than those of the CS membrane. This is attributed to the increase of the amorphous region and hydrophilicity when the CS is cross-linked with APTEOS, which could be proved by the XRD, contact angle and sorption experiments mentioned previously. 3.5.2. Effect of feed temperature on time-dependent separation property during PV processes Temperature has a substantial effect on permeation flux and separation factor. Temperature dependence of permeation properties of the CS membrane and the CSHM with operating time was obtained in PV separation of 30 wt% MeOH concentration within temperature range of 30 to 50 ◦ C, as illustrated in Figs. 10a and 10b. It shows that the PV process reach a steady state more quickly with increasing operating temperature. This
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
(a)
587
(b)
Fig. 10. (a) Permeation flux and separation factor of the CS membrane as function of time in PV of 30 wt% MeOH in feed at operating temperature range of 30–50 ◦ C. (b) Permeation flux and separation factor of the CSHM as function of time in PV of 30 wt% MeOH in feed at operating temperature range of 30–50 ◦ C.
Table 1 Comparison of PV performance of the membranes reported in the available literatures for MeOH/DMC mixtures Membrane
Thickness (± µm)
Temperature (◦ C)
Mass % of MeOH in feed
Chitosan
– – – – – –
25 35 45 55 55 55
23 23 23 23 53 70
Chitosan crosslinked with H2 SO4
– – – – – –
25 35 45 55 55 55
Chitosan
20 20 20 20 20
Chitosan crosslinked with APTEOS
20 20 20 20 20
Separation factor
Reference
34 52 98 130 248 291
30 24 20 17 11 8
[20] [20] [20] [20] [20] [20]
29 29 29 29 52 70
81 122 170 215 370 498
38 31 22 19 13 9
[21] [21] [21] [21] [21] [21]
30 40 50 50 50
30 30 30 50 70
170 220 259 252 300
28 22 16.6 15.8 13.5
Present work Present work Present work Present work Present work
30 40 50 50 50
30 30 30 50 70
890 1050 1158 1225 1275
49 45.3 41.4 39.4 29.8
Present work Present work Present work Present work Present work
is attributed to the increase of the rate of surface reconstruction resulting from the improvement of the mobility of the CS chains with temperature increasing. Fig. 10 also shows that the flux increases and the separation factor decreases with temperature increasing. Generally speaking, the mobility of the CS chains increases with feed temperature increasing, which leads to an increase in both the free volume of the membrane and the motion rate of the permeant molecules. The feed vapor pressure also increases with increasing feed temperature, whereas, the permeate side pressure remains constant. These lead to an increase of the driving force for mass transfer. All these result in an increase of permeation flux. The decrease of separation fac-
Flux (g/(h m2 ))
tor is attributed to the increase of the DS and the decrease of solubility and diffusion selectivities with increasing temperature. Upon comparing Fig. 10a with Fig. 10b, we find that the time for the PV process to reach the steady state decreases with temperature increasing, and the time for the CSHM is longer than that of the CS membrane. These are similar to the contact angle experiments and the reasons are the same. 3.6. Comparison of present membranes with literature data A comparison of the present PV results with the published results is important to assess the superiority of the membranes.
588
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
The present data are compared in Table 1 with the available literature data. It shows that the performance of the membranes in the present study is significantly improved.
gram of Higher Education (no. 2005038401) in preparation of this article is gratefully acknowledged. References
4. Conclusions The CSHM were prepared by cross-linking CS with APTEOS, and the thermal stability of the CSHM is improved. The results of PV separation MeOH/DMC show that the CSHM exhibits a higher permeation flux and separation factor than those of the CS membrane under the same operating condition, and the surface reconstruction occurred to both the CS membrane and the CSHM during PV processes. The time for the surface reconstruction of the CSHM is longer than that of the CS. Under the same operating condition, the effects of feed MeOH concentration and temperature on the rate of surface reconstruction of the CS membrane and the CSHM is similar. The rate of surface reconstruction increases with increasing MeOH concentration or temperature. The feed concentration affects not only the hydrophilicity of the membrane, but also the rate of surface reconstruction. However, under the same feed concentration, the feed temperature only affects the rate of surface reconstruction. Acknowledgments The support of National Nature Science Foundation of China Grant no. 50573063, the Program for New Century Excellent Talents in University and the research fund for the Doctoral Pro-
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]
I.J. Drake, K.L. Fujdala, A.T. Bell, et al., J. Catal. 230 (2005) 14–27. Z. Li, K. Xie, R.C.T. Slade, Appl. Catal. A Gen. 205 (2001) 85–92. M. Richter, M.J.G. Fait, R. Eckelt, et al., J. Catal. 245 (2007) 11–24. J.-C. Hu, Y. Cao, P. Yang, et al., J. Mol. Catal. A Chem. 185 (2002) 1–9. Y. Zhang, I.J. Drake, D.N. Briggs, et al., J. Catal. 244 (2006) 219–229. B. Smitha, D. Suhanya, S. Sridhar, et al., J. Membr. Sci. 241 (2004) 1–21. P. Srinivasa Rao, B. Smitha, S. Sridhar, et al., Sep. Purif. Technol. 48 (2006) 244–254. F.-L. Mi, C.-Y. Kuan, S.-S. Shyu, et al., Carbohydr. Polym. 41 (2000) 389– 396. P. Mukoma, B.R. Jooste, H.C.M. Vosloo, J. Power Sources 136 (2004) 16–23. Y.L. Liu, Y.H. Su, K.R. Lee, et al., J. Membr. Sci. 251 (2005) 233–238. Y. Shirosaki, K. Tsuru, S. Hayakawa, et al., Biomaterials 26 (2005) 485– 493. C.E. Schmidt, J.M. Baier, Biomaterials 21 (2000) 2215–2231. B.-B. Li, Z.-L. Xu, F. Alsalhy Qusay, et al., Desalination 193 (2006) 171– 181. A. Svang-Ariyaskul, R.Y.M. Huang, P.L. Douglas, et al., J. Membr. Sci. 280 (2006) 815–823. B. Smitha, G. Dhanuja, S. Sridhar, Carbohydr. Polym. 66 (2006) 463–472. S.Y. Nam, Y.M. Lee, J. Membr. Sci. 157 (1999) 63–71. S. Cao, Y. Shi, G. Chen, J. Appl. Polym. Sci. 74 (1999) 1452–1458. G.Y. Moon, R. Pal, R.Y.M. Huang, J. Membr. Sci. 176 (2000) 223–231. K. Inui, K. Tsukamoto, T. Miyata, et al., J. Membr. Sci. 138 (1998) 67–75. W. Won, X. Feng, D. Lawless, J. Membr. Sci. 209 (2002) 493–508. W. Won, X. Feng, D. Lawless, Sep. Purif. Technol. 31 (2003) 129–140. J.H. Chen, Q.L. Liu, X.H. Zhang, et al., J. Membr. Sci. 292 (2007) 125– 132.
Composite hybrid membrane of chitosan–silica in pervaporation separation of MeOH/DMC mixtures Jian Hua Chen, Qing Lin Liu ∗ , Jun Fang, Ai Mei Zhu, Qiu Gen Zhang Department of Chemical and Biochemical Engineering, College of Chemistry & Chemical Engineering, Xiamen University, Xiamen 361005, China Received 14 May 2007; accepted 9 September 2007 Available online 14 September 2007
Abstract Chitosan–silica hybrid membranes (CSHMs) were prepared by cross-linking chitosan (CS) with 3-aminopropyl-triethoxysilane (APTEOS). The dynamic behaviors of the CS membrane and the CSHM were investigated in pervaporation (PV) of methanol/dimethyl carbonate (MeOH/DMC) mixtures. The membranes were characterized by X-ray diffraction (XRD), contact angle meter, scanning electron microscopy (SEM) and differential scanning calorimetry (DSC). The transition state of PV processes were studied. During the PV processes, the amorphous region of the membranes increases and the contact angle between MeOH and the membrane decreases within a range of operating time and then remains almost constant implying a reconstruction occurred on the membrane surface. The silica is well distributed in the CSHM matrix and the thermal stability of the CSHM is enhanced. The time for a PV process to reach a steady state decreases with increasing MeOH concentration or feed temperature, and it is longer for the CSHM than the CS membrane under the same operating condition. Swelling experiments show that the degree of swelling (DS) is greatly depressed by cross-linking CS with APTEOS. Sorption data indicate that the selectivity of solubility and diffusion of the CSHM are greatly improved over the CS membrane. The CSHM presents superior separation behaviors over other membranes with a flux of 1265 g/(h m2 ) and separation factor of 30.1 in PV separation of 70 wt% MeOH in feed at 50 ◦ C. © 2007 Elsevier Inc. All rights reserved. Keywords: Chitosan; Hybrid membranes; Surface reconstruction; MeOH; Dimethyl carbonate
1. Introduction Recently, separation of organic/organic mixtures using PV techniques is being investigated extensively owing to its great importance in chemical and petrochemical industries. The separation of organic/organic mixtures, which may roughly be classified as follows: (1) separation of polar/non-polar solvent mixtures, (2) separation of aromatic/aliphatic mixtures, (3) separation of aliphatic hydrocarbons, and (4) separation of aromatic isomers. Due to very similar chemical and physical properties between the organic components, the organic mixture was difficult to be separated by distillation or other membrane processes. Dimethyl carbonate (DMC) is an environmental benign chemical compound and unique intermediate with versatile * Corresponding author. Fax: +86 592 2184822.
E-mail address: [email protected] (Q.L. Liu). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.09.022
chemical reactivity which has attracted increasing interest from both practical and fundamental points of view in recent years [1,2]. It can be used as an intermediate to substitute for highly toxic phosgene and dimethyl sulfate in many chemical processes owing to the presence of two methyl groups and one carbonyl group in its molecule. Because of its high oxygen content [3], low toxicity, rapid biodegradability and low impact on air quality, DMC has been proposed as a replacement for methyl tert-butyl ether (MTBE) as a fuel additive. Currently, the popular synthesis of DMC based on the oxycarbonylation of MeOH [4,5] has been successfully developed up to the industrial scale: (1) the oxidative carbonylation of MeOH with carbon monoxide and oxygen catalyzed by cuprous chloride, and (2) an oxidative carbonylation process using a palladium catalyst and methyl. Indeed, excess MeOH is used for a high conversion yield. The use of excess MeOH, however, causes a purification problem because MeOH forms an azeotrope with DMC at different MeOH content under different
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
pressure. Separation is an important step in DMC manufacturing. Various separation processes (e.g., extractive distillation, pressure swing distillation, liquid–liquid extraction, adsorption on zeolites, and low temperature crystallization) have been proposed to separate and purify DMC; however, few literatures about the separation of MeOH/DMC mixtures by PV are available. Hydrophobic polymers such as polyethylene and polypropylene [6] were used mainly as membrane materials. However, these hydrophobic materials did not show a high selectivity for organic mixtures, because they have slight differences in the interaction with one component from the other in the mixtures. As for MeOH/DMC mixtures, MeOH is more hydrophilic and polar than DMC due to dipole moment of hydroxyl group. As a result, the interaction between the polar material and MeOH should be stronger than DMC, which results in larger affinity of MeOH for polar or hydrophilic polymer materials. Therefore, it should be possible to improve MeOH selectivity by using a hydrophilic membrane. Furthermore, MeOH is more permeable than DMC due to the difference in the molecular size. Accordingly, the molecular size and polarity difference should be a key factor to obtain a higher permselectivity. CS is widely used in membrane applications because of its high hydrophilicity, good film-forming character and excellent chemical-resistant properties. CS is also a cyclo-aliphatic polymer. It contains active amino groups and hydroxyl groups that can be chemically modified by several chemicals such as glutaraldehyde [7], epichlorohydrin [8], sulfuric acid [9], GPTMS [10,11] and dialdehyde starch [12], etc. CS-based membranes have been extensively studied not only for the dehydration of alcohols [13], other industrial solvents like tetrahydrofuran, 1,4-dioxane, isopropanol and ethylene glycol [14,15], but also for the separation of organic/organic mixtures MeOH/MTBE [16,17], alcohol/toluene [18], benzene/cyclohexane [19] and MeOH/DMC [20,21]. A PV process consists of three consecutive steps: (I) feed components sorption into the membranes at the feed side, (II) permeant diffusion through the membranes, and (III) permeant desorption into a vapor phase at the permeate side. Desorption is very quick since the pressure at the permeate side is very low, so a PV process is controlled mainly by the first two steps. Therefore, study on the surface properties of the membrane during PV processes is extremely important, since PV performance is greatly dependent on the surface properties of the membrane. For this purpose, a novel CSHM with APTEOS content of 10 wt% (this content proved to be the optimum in our previous work [22]) were prepared by cross-linking APTEOS with CS so as to enhance both flux and permselectivity simultaneously. The surface properties such as hydrophilicity and crystallinity of the CS membrane and the CSHM coupled with permeation properties during PV processes were investigated by using a contact angle meter and X-ray diffraction (XRD) method. Moreover, the effects of feed concentration and temperature on the membrane dynamic behaviors were also studied.
581
2. Experimental 2.1. Materials CS with a degree of deacetylation of approximately 92% (Yuhuan Ocean Biochemical Co. Ltd., China) and 3-aminopropyl-triethoxysilane (APTEOS) (Shanghai Yaohua Chemical Plant, China), were used without further purification. The other solvents and reagents, of analytical grade and used without further purification, were purchased from Sinophatm Chemical Reagent Co. Ltd. 2.2. Membrane preparation Chitosan was dissolved in a 2.0 wt% acetic acid aqueous solution to form a solution of 2.0 wt% chitosan. The solution was stirred at room temperature for 2 h, filtered, and then a certain amount of APTEOS as well as concentrated HCl as a catalyst were added to initiate the cross-linking reaction. The mixture was stirred at 50 ◦ C for about 12 h. The resulted homogeneous solution was cast onto a glass plate with a casting knife, and then dried in an oven at 50 ± 1 ◦ C with relative humidity 65 ± 2% for about 4 h. The obtained membranes were peeled off gently, and were further dried at 70 ◦ C for 8 h. The thickness of membrane, measured by SEM, is 20 ± 0.5 µm. The mass ratio of APTEOS to CS was 10 wt% and designated as CSHM. 2.3. Membrane characterization Crystal structure characterization was carried out by XRD (Panalytical X’pert Philip, Holland) using CuKα radiation. The diffraction was operated at 40 kV and 30 mA at a 2θ range of 5–40◦ , using a step size of 0.0167◦ and a counting time of 10 s per step. The morphology of surface and thickness of the membranes were measured by using SEM (LEO 1530, Oxford Instruments), which was operated at EHT = 20 kV. DSC was performed using a Netzsch DSC 204 Phoenix. Accurately weighted 5 mg samples were placed into aluminum cups and heated from 25 to 400 ◦ C at a constant heating rate of 10 ◦ C/min under constant nitrogen purging at 20 mL/min. Static contact angles between MeOH and the CS membranes and/or the CSHMs were measured by the pendant drop method using a contact angle meter (SL200B, SOLON TECH, Shanghai, China) at 23 ± 1 ◦ C with 65 ± 2% relative humidity. All reported values are the average of eight measurements taken at different location on the membrane surface. The errors are less than 4%. 2.4. Pervaporation measurement PV experiments were performed on the PERVAP 2201 (SULZER CHEMTECH, Germany) at 30, 40 and 50 ◦ C, respectively. The effective membrane area was 62.21 cm2 . The membrane was supported on a filter paper over a porous sintered steel disk of 6.4 cm in diameter. The vacuum at the permeate side was maintained 300 Pa using a vacuum pump, and
582
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
the permeate was collected in a liquid nitrogen trap. The concentrations of both the feed and the permeate were measured by gas chromatography (GC-950 Haixin, Shanghai, China) equipped with a thermal conductivity detector (TCD) and a column packed with GDX102. The results for PV separation of MeOH/DMC mixtures were reproducible, and the errors in the PV measurements are less than 1.5%. The separation performances of the membranes can be evaluated on the basis of total flux (J ) and separation factor (a). Q , A·t PM /PD aPV = , FM /FD Jp =
H2 NCH2 CH2 Si–(OC2 H5 )3 + XH2 O → H2 NCH2 CH2 Si–(OC2 H5 )3−x (OH)x + xC2 H5 OH Scheme 1. Hydrolysis reaction for APTEOS.
R O
Si
R OR1 +
O
R
R Δ
−→
O
OH
O
O
(1)
Si
Si
O
Si
O
+ R1 OH
(2) O
where Q is the weight of the permeate collected within time t, and A is the effective membrane area, PM and PD are the mass fraction of MeOH and DMC in the permeate, FM and FD are that of MeOH and DMC in the feed, respectively.
O
Scheme 2. Condensation reaction for APTEOS. R = –CH2 CH2 NH2 and R1 = –H or –CH2 –CH3 .
2.5. Swelling measurements and sorption experiments Before each experiment the pre-dried CS membrane and CSHM were weighed and immersed in MeOH/DMC mixtures at a desired temperature for 48 h to reach an equilibrium swelling. The sample was taken out at appropriate intervals (about 3 h), wiped with tissue paper carefully to remove the surface solvent, and weighed as quickly as possible, then dipped again into the liquids. The experiments continued until the weight of the sample keeping approximately constant. All experiments were repeated at least for three times, and the results were averaged. The errors are no more than 5%. The DS was calculated by MW − MD × 100%, DS (%) = MD
3. Results and discussion (3) 3.1. Chemical structure of APTEOS/CS hybrid membranes
where MD and MW denote the mass of the dried and swollen membranes, respectively. For sorption in a binary solution, the sorbed liquid was collected in a liquid nitrogen trap by desorbing the equilibrated sample in the purge-and-trap apparatus, and the concentration of the collected liquid was measured by gas chromatography. The solubility selectivity, as , is expressed by αs =
XM /XD , LM /LD
(4)
where XM and XD are the mass fraction of MeOH and DMC in the membrane; LM and LD are the mass fraction of MeOH and DMC in the binary solution. According to the solution–diffusion mechanism, diffusion selectivity, aD , was calculated by aD =
aPV . aS
Scheme 3. Chemical structure and reaction model for cross-linking chitosan.
(5)
The chemical structures of chitosan and APTEOS and their cross-linking reaction are schematically illustrated in Schemes 1, 2 and 3. In preparing the hybrid membranes, APTEOS was first hydrolyzed in the presence of an acid catalyst (HCl), leading to the formation of silanol groups. Then, the hydroxyl groups in one silanol formed siloxane bonds with those in another silanol or with the amine groups in CS through a dehydration or dealcoholysis reaction during the membrane drying. Also, the hydroxyl groups and amine groups in the silanol can form hydrogen bonds with the dissociative hydroxyl groups or amine groups in the CS amorphous region, and these hydrogen and siloxane bonds are the cross-linking spots in the hybrid membranes. 3.2. Membrane characterization The change of the crystal structure of the membrane during PV processes was investigated. Figs. 1 and 2 show the XRD
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
583
(a)
Fig. 1. XRD patterns of the CS membranes immersed in MeOH/DMC mixtures at 50 ◦ C with various MeOH mass fractions: (a) without being immersed, (b) 30, (c) 50 and (d) 70 wt%.
(b)
(c) Fig. 3. SEM images of (a) the CS membrane, (b) the CSHM and (c) the cross-sectional of CSHM (×2000). Fig. 2. XRD patterns of the CSHMs immersed in MeOH/DMC mixtures at 50 ◦ C with various MeOH mass fractions: (a) without being immersed, (b) 30, (c) 50 and (d) 70 wt%.
spectra for the CS membrane and the CSHM with or without being immersed in MeOH/DMC solutions, respectively. Comparing Fig. 2 with Fig. 1, we find that the diffraction peaks located at 11.5 and 18.3◦ vanished with APTEOS being introduced. There is only one intense diffraction peak at 21.2◦ appeared for the CSHM. This is attributed to the arrangement of the CS chains that is disturbed by the introduced APTEOS, and then decreased the crystallinity of the CSHM. Figs. 1 and 2 suggest that the higher the feed MeOH concentration is, the less the crystallinity of the CS membranes and the CSHMs. This is because more MeOH molecules interact with the polymer matrix with MeOH concentration increasing, some of the formed hydrogen-bonding, between the –OH and/or –NH2 groups of intra-molecular or inter-molecular of CS, which caused the CS chains to become more compact, was broken. This further caused the polymer chain less compact. The SEM images of the CS membrane and the CSHM are reflected in Figs. 3a and 3b and the image of the cross-section of
Fig. 4. DSC thermo-grams of the CS membrane and the CSHM.
the CSHM is shown in Fig. 3c. It indicates that the silica phase dispersed well into the CSHM matrix, and the thickness of the membrane is 20.4 µm. Fig. 4 presents the thermo-grams of
584
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
(a)
(a)
(b) (b) Fig. 5. Contact angle between MeOH and the CS membranes as a function of time immersed in MeOH/DMC mixtures with (a) MeOH mass fraction range of 30 to 70 wt% at 30 ◦ C. (b) 30 wt% MeOH mass fraction at the temperature range of 30 to 50 ◦ C.
the CS membrane and the CSHM, showing initial endothermic peaks at 55.5 and 50.4 ◦ C and higher exothermic peaks at 280.3 and 291.8 ◦ C, respectively. The endothermic peaks are corresponding to the loss of water associated with the hydrophilic groups of polymers while the exothermic peaks result from the degradation of the CS chains. It suggests that the CSHM became more thermally stable than that of the CS membrane. The contact angle between MeOH and the membrane was measured to investigate the dynamic properties of the membrane during PV processes. The samples coated onto the glass slide were submersed in MeOH/DMC mixtures at 30 ◦ C and the concentration of MeOH was changed from 30 to 70 wt%; otherwise, when MeOH concentration was kept at 30 wt%, the temperature of solution was changed from 30 to 50 ◦ C for a desired time (such as 0, 10, 20, 30 min, etc.). The sample was removed from the MeOH solution and dried with nitrogen stream prior to contact angle measurement. It can be seen from Fig. 5a that
Fig. 6. Contact angle between MeOH and the CSHMs as a function of time immersed in MeOH/DMC mixtures with (a) MeOH mass fraction range of 30–70 wt% at 30 ◦ C. (b) 30 wt% MeOH mass fraction at the temperature range of 30–50 ◦ C.
the time for the contact angle to reach a steady value decreased from 60 to 20 min, and the steady contact angle decreased from 75.13 to 69.24◦ , when the MeOH concentration varied from 30 to 70 wt%. These are attributed to the surface reconstruction of the membrane when it contacts with MeOH/DMC mixtures. With increasing MeOH concentration, more MeOH molecules would touch the membrane and induce more hydroxyl groups of the membrane to move toward the feed side, thus the hydrophilicity of the membrane was improved. Moreover, some of the formed hydrogen-bonding, between the –OH and/or –NH2 groups of intra-molecular or inter-molecular of CS, was broken because of the interaction of MeOH molecules with the polymer matrix in the membrane. This resulted in a less compact of polymer matrix, which increased the mobility of the CS chains, and thus reduced the time for both hydroxyl groups to reconstruct and the contact angle to reach a steady value. The effect of temperature on the surface reconstruction is shown in Fig. 5b. The time for the contact angle to reach a steady value
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
decreased with increasing feed temperature. This is attributed to the increase of the rate of reconstruction resulting from the improvement of mobility of the CS chains as the temperature increased. The CSHM was also studied to investigate the effect of cross-linking on the surface reconstruction, as presented in Fig. 6. Comparing with Fig. 5a, Fig. 6a shows the contact angle between MeOH and the CSHM shared the same changing trend as the CS membrane. However, the initial value of the contact angle between MeOH and the CSHM (73.5◦ ) is lower than that of the CS membrane (80.35◦ ) and the steady contact angle for the CSHM is lower than that for the CS membrane under the same condition. It may be attributed to the increase of the hydrophilic groups, e.g., –OH groups introduced when the APTEOS is hydrolyzed. Upon comparing Fig. 6 with Fig. 5, we find that the time for the contact angle for the CSHM to reach a steady value is longer than that for the CS membrane under the same condition. This is attributed to the decrease of the rate of reconstruction resulting from the decrease of the mobility of the CS chains when the CS is cross-linked with APTEOS.
585
(a)
3.3. Effects of feed concentration and temperature on the CS and CSHM swelling Generally speaking, cross-linking limits the mobility of polymer chains and thus excessive membrane swelling can be suppressed. Fig. 7a shows the effect of feed MeOH concentration on the DS of the CS membrane and the CSHM at 30 ◦ C. The DS of the CS membrane and the CSHM decreased with MeOH concentration decreasing. This is attributed to the hydrophilicity nature of the CS-based membranes (the properties of MeOH is similar to that of water). Fig. 7b shows the DS of the CS membrane and the CSHM in 30 wt% MeOH solution vs. the temperature. The DS of the CS membrane and the CHSM increased with temperature increasing. This is owing to the increase of the mobility of the CS chains with temperature increasing, which leads to an increase of the free volume in the polymer network, and results in an increase in permeant sorption. Fig. 7 also shows that under the same condition, the DS of the CSHM is less than the CS. This is due to the cross-linking between the CS and APTEOS which tends to make the amorphous region of the CSHM more compact. 3.4. Sorption and diffusion properties Based on the experimental sorption data, the selectivity of solubility and diffusion was calculated from Eqs. (4) and (5). The dependence of the sorption behavior of the CS membrane and the CSHM on feed MeOH concentration was investigated, as shown in Fig. 8a. It shows that both the solubility and diffusion selectivities decreased almost linearly with increasing MeOH concentration in the range of 30–70 wt%, and both the solubility and diffusion selectivities of the CSHM is higher than those of the CS membrane. The decrease in the solubility and diffusion selectivities is due to the increase in the DS. The solubility selectivity of the CSHM is higher than that of the CS membrane owing to the higher hydrophilicity of the CSHM; and the diffusion selectivity of the CSHM is higher than that
(b) Fig. 7. (a) The effect of MeOH concentration on the DS of the CS membranes and the CSHMs at 30 ◦ C. (b) The effect of feed temperature on the DS of the CS membranes and the CSHMs in 30 wt% MeOH solution.
of the CS membrane attributing to the more compact amorphous regions of the CSHM. The dependence of the solubility and diffusion selectivities on feed temperature is also plotted in Fig. 8b. Both the solubility and diffusion selectivities of the CS membrane and the CSHM decrease with increasing the temperature. This is attributed to an increase in the DS with the feed temperature increasing. 3.5. Pervaporation characteristics 3.5.1. Effect of feed concentration on time-dependent separation property during PV processes Feed concentration is one of the important factors during PV processes. PV performance of the CS membrane as a function of operating time is studied for MeOH concentration range of 30–70 wt% at 30 ◦ C. The experimental results are illustrated in Fig. 9a. It is found that, taking MeOH concentration of 30 wt% as an example, permeation flux increases from 150 to 170 g/(h m2 ) and separation factor increases from 25.3 to 28
586
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
(a)
(a)
(b) (b) Fig. 8. (a) The effect of MeOH concentration on the solubility and diffusion selectivities of the CS membrane and the CSHM at 30 ◦ C. (b) The effect of feed temperature on the solubility and diffusion selectivities of the CS membrane and the CSHM for 30 wt% MeOH in feed.
correspondingly, in about 60 min, and then both remain almost unchanged within our experimental time. The increase of the flux is attributed to that the crystallinity of the CS membrane decreased during PV processes, which is proved by XRD. Whereas, the improvement of separation factor is attributed to the increase of CS membrane hydrophilicity which has been proved by the contact angle experiments. Fig. 9a also indicates that the flux increases and the separation factor decreases with increasing MeOH concentration. The increase in the flux is due to the DS increase; while the decrease in the separation factor is attributed to the decrease of the solubility and diffusion selectivities. We can also observe that the time for a PV process to reach a steady state decreases with MeOH concentration increasing. This is similar to the effect of MeOH concentration on the time for the contact angle to reach a steady value mentioned above. A further investigation of such PV behaviors was performed for the CSHM, as shown in Fig. 9b. It shows that MeOH concentration dependence of PV performance of the CSHM with operating time is similar to that of the CS membrane. However, still refer to MeOH concentration of 30 wt%, the time (about 70 min) for the flux and separation factor of the CSHM
Fig. 9. (a) Permeation flux and separation factor of the CS membrane as function of time in PV of MeOH concentration range of 30–70 wt% at 30 ◦ C. (b) Permeation flux and separation factor of the CSHM as function of time in PV of MeOH concentration range of 30–70 wt% at 30 ◦ C.
to reach an equilibrium value (from 850 to 900 g/(h m2 ) for the flux and 46 to 49 for the separation factor, correspondingly) is longer than that of the CS membrane (about 60 min). This is because the cross-linking between APTEOS and CS makes the CS chains more compact, slowed down the mobility of the CSHM chains, hence raised the time for surface reconstruction. Meanwhile, the permeation flux and separation factor of the CSHM is higher than those of the CS membrane. This is attributed to the increase of the amorphous region and hydrophilicity when the CS is cross-linked with APTEOS, which could be proved by the XRD, contact angle and sorption experiments mentioned previously. 3.5.2. Effect of feed temperature on time-dependent separation property during PV processes Temperature has a substantial effect on permeation flux and separation factor. Temperature dependence of permeation properties of the CS membrane and the CSHM with operating time was obtained in PV separation of 30 wt% MeOH concentration within temperature range of 30 to 50 ◦ C, as illustrated in Figs. 10a and 10b. It shows that the PV process reach a steady state more quickly with increasing operating temperature. This
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
(a)
587
(b)
Fig. 10. (a) Permeation flux and separation factor of the CS membrane as function of time in PV of 30 wt% MeOH in feed at operating temperature range of 30–50 ◦ C. (b) Permeation flux and separation factor of the CSHM as function of time in PV of 30 wt% MeOH in feed at operating temperature range of 30–50 ◦ C.
Table 1 Comparison of PV performance of the membranes reported in the available literatures for MeOH/DMC mixtures Membrane
Thickness (± µm)
Temperature (◦ C)
Mass % of MeOH in feed
Chitosan
– – – – – –
25 35 45 55 55 55
23 23 23 23 53 70
Chitosan crosslinked with H2 SO4
– – – – – –
25 35 45 55 55 55
Chitosan
20 20 20 20 20
Chitosan crosslinked with APTEOS
20 20 20 20 20
Separation factor
Reference
34 52 98 130 248 291
30 24 20 17 11 8
[20] [20] [20] [20] [20] [20]
29 29 29 29 52 70
81 122 170 215 370 498
38 31 22 19 13 9
[21] [21] [21] [21] [21] [21]
30 40 50 50 50
30 30 30 50 70
170 220 259 252 300
28 22 16.6 15.8 13.5
Present work Present work Present work Present work Present work
30 40 50 50 50
30 30 30 50 70
890 1050 1158 1225 1275
49 45.3 41.4 39.4 29.8
Present work Present work Present work Present work Present work
is attributed to the increase of the rate of surface reconstruction resulting from the improvement of the mobility of the CS chains with temperature increasing. Fig. 10 also shows that the flux increases and the separation factor decreases with temperature increasing. Generally speaking, the mobility of the CS chains increases with feed temperature increasing, which leads to an increase in both the free volume of the membrane and the motion rate of the permeant molecules. The feed vapor pressure also increases with increasing feed temperature, whereas, the permeate side pressure remains constant. These lead to an increase of the driving force for mass transfer. All these result in an increase of permeation flux. The decrease of separation fac-
Flux (g/(h m2 ))
tor is attributed to the increase of the DS and the decrease of solubility and diffusion selectivities with increasing temperature. Upon comparing Fig. 10a with Fig. 10b, we find that the time for the PV process to reach the steady state decreases with temperature increasing, and the time for the CSHM is longer than that of the CS membrane. These are similar to the contact angle experiments and the reasons are the same. 3.6. Comparison of present membranes with literature data A comparison of the present PV results with the published results is important to assess the superiority of the membranes.
588
J.H. Chen et al. / Journal of Colloid and Interface Science 316 (2007) 580–588
The present data are compared in Table 1 with the available literature data. It shows that the performance of the membranes in the present study is significantly improved.
gram of Higher Education (no. 2005038401) in preparation of this article is gratefully acknowledged. References
4. Conclusions The CSHM were prepared by cross-linking CS with APTEOS, and the thermal stability of the CSHM is improved. The results of PV separation MeOH/DMC show that the CSHM exhibits a higher permeation flux and separation factor than those of the CS membrane under the same operating condition, and the surface reconstruction occurred to both the CS membrane and the CSHM during PV processes. The time for the surface reconstruction of the CSHM is longer than that of the CS. Under the same operating condition, the effects of feed MeOH concentration and temperature on the rate of surface reconstruction of the CS membrane and the CSHM is similar. The rate of surface reconstruction increases with increasing MeOH concentration or temperature. The feed concentration affects not only the hydrophilicity of the membrane, but also the rate of surface reconstruction. However, under the same feed concentration, the feed temperature only affects the rate of surface reconstruction. Acknowledgments The support of National Nature Science Foundation of China Grant no. 50573063, the Program for New Century Excellent Talents in University and the research fund for the Doctoral Pro-
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]
I.J. Drake, K.L. Fujdala, A.T. Bell, et al., J. Catal. 230 (2005) 14–27. Z. Li, K. Xie, R.C.T. Slade, Appl. Catal. A Gen. 205 (2001) 85–92. M. Richter, M.J.G. Fait, R. Eckelt, et al., J. Catal. 245 (2007) 11–24. J.-C. Hu, Y. Cao, P. Yang, et al., J. Mol. Catal. A Chem. 185 (2002) 1–9. Y. Zhang, I.J. Drake, D.N. Briggs, et al., J. Catal. 244 (2006) 219–229. B. Smitha, D. Suhanya, S. Sridhar, et al., J. Membr. Sci. 241 (2004) 1–21. P. Srinivasa Rao, B. Smitha, S. Sridhar, et al., Sep. Purif. Technol. 48 (2006) 244–254. F.-L. Mi, C.-Y. Kuan, S.-S. Shyu, et al., Carbohydr. Polym. 41 (2000) 389– 396. P. Mukoma, B.R. Jooste, H.C.M. Vosloo, J. Power Sources 136 (2004) 16–23. Y.L. Liu, Y.H. Su, K.R. Lee, et al., J. Membr. Sci. 251 (2005) 233–238. Y. Shirosaki, K. Tsuru, S. Hayakawa, et al., Biomaterials 26 (2005) 485– 493. C.E. Schmidt, J.M. Baier, Biomaterials 21 (2000) 2215–2231. B.-B. Li, Z.-L. Xu, F. Alsalhy Qusay, et al., Desalination 193 (2006) 171– 181. A. Svang-Ariyaskul, R.Y.M. Huang, P.L. Douglas, et al., J. Membr. Sci. 280 (2006) 815–823. B. Smitha, G. Dhanuja, S. Sridhar, Carbohydr. Polym. 66 (2006) 463–472. S.Y. Nam, Y.M. Lee, J. Membr. Sci. 157 (1999) 63–71. S. Cao, Y. Shi, G. Chen, J. Appl. Polym. Sci. 74 (1999) 1452–1458. G.Y. Moon, R. Pal, R.Y.M. Huang, J. Membr. Sci. 176 (2000) 223–231. K. Inui, K. Tsukamoto, T. Miyata, et al., J. Membr. Sci. 138 (1998) 67–75. W. Won, X. Feng, D. Lawless, J. Membr. Sci. 209 (2002) 493–508. W. Won, X. Feng, D. Lawless, Sep. Purif. Technol. 31 (2003) 129–140. J.H. Chen, Q.L. Liu, X.H. Zhang, et al., J. Membr. Sci. 292 (2007) 125– 132.