- Email: [email protected]
water mixture by pervaporation using NaA zeolite membrane synthesized by vacuum seeding method
Vacuum 92 (2013) 70e76
Contents lists available at SciVerse ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
High performance dehydration of ethyl acetate/water mixture by pervaporation using NaA zeolite membrane synthesized by vacuum seeding method M. Moheb Shahrestani a, A. Moheb a, *, M. Ghiaci b a b
Department of Chemical Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran Department of Chemistry, Isfahan University of Technology, Isfahan 84156-83111, Iran
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 April 2012 Received in revised form 9 November 2012 Accepted 30 November 2012
High quality NaA zeolite membranes were prepared by vacuum assisted secondary growth. At the first stage in which NaA powder was synthesized, increasing the aging time led to formation of impure and smaller crystals. However, at the aging time of 48 h, pure NaA zeolite particles with average particle size of 1.5 mm were obtained. In the second stage, the outer surface of porous a-Al2O3 tubular supports were seeded by vacuum method using 1.5 mm NaA particles. The most stable and uniform seeded layer was obtained at seeding time and suspension concentration of 90 s and 5 g L1, respectively. Then, 6 h of secondary growth of the zeolite layer on the seeded surface was carried out at 373 K three times. The XRD and SEM results showed the formation of a uniform and dense layer of pure NaA zeolite with an average thickness of 45 mm. Dehydration experiments were conducted on ethyl acetate/water mixtures with 2 wt% water content. The average total fluxes were 0.147, 0.208, and 0.315 kg m2 h1 at 303, 313, and 323 K, respectively. The separation factor was 163,000. This parameter did not change with temperature and it was due to very close activation energies of ethyl acetate and water. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Aging time Ethyl acetate NaA zeolite membrane Pervaporation Vacuum seeding
1. Introduction Pervaporation is an effective and energy-efficient membrane process which can be used for some liquideliquid separation applications. This process is a promising candidate for cases in which separation encounters difficulties due to liquid physical and chemical properties where it is difficult to achieve desirable results by conventional separation processes such as distillation and extraction [1e3]. During pervaporation process, a liquid feed mixture is contacted with the active side of a dense membrane and one component passes through the membrane to the other side. This is due to the selectivity of the membrane under the act of vacuum. Implementing vacuum force causes the passing species to be evaporated, leaving the membrane body as vapor phase to be later condensed in a cold trap [3e5]. This environmentally clean and economic technology can be used for many applications such as dehydration of aqueouseorganic mixtures, removal of volatile organic compounds from aqueous solutions, and organiceorganic separations [3e8]. Polymeric membranes are widely used for many pervaporation applications [1e3,9]. However, one major
* Corresponding author. Tel.: þ98 311 3915618; fax: þ98 311 3912677. E-mail address: [email protected] (A. Moheb). 0042-207X/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.vacuum.2012.11.019
drawback of the polymeric membranes which limits their practical applications is their insufficient chemical and thermal stability [1e 3,9e11]. Furthermore, swelling of the polymeric membranes generally leads to higher permeability and lower selectivity [1,12]. On the other hand, being free from swelling effect and having good thermal and chemical stabilities of inorganic membranes make them proper candidates for a broad range of separation applications [1,5,12]. Among them, zeolite membranes take advantage of the unique properties of zeolites such asehighly crystalline ordered structure and having molecular-sized pores [3,13]. These traits are very suitable for separating liquid-phase mixtures via pervaporation process. Basically, molecular transport and separation of different molecules through zeolite pores are based on differences in their adsorption and diffusion properties [13]. Various synthesis strategies and methods have been investigated for synthesis of the zeolite membranes [14]. Coating of the zeolite seeds on the surface of a porous support before hydrothermal synthesis, known as secondary growth method, is an effective approach to synthesizing a high quality zeolite membrane [15e19]. In general, synthesis with seeds gives a better control of membrane formation process [15,20,21]. In addition, the secondary growth method facilitates the growth of the zeolite crystals on the support surface (by heterogeneous nucleation) rather than in the bulk of synthesis solution, which in turn prevents transformation of
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76
the zeolite phase into other undesired phases [15]. However, preparing zeolite suspensions and coating the zeolite seeds on a porous support to obtain a uniform seeding layer is still a challenge [15,19]. Several seeding methods such aserub coating, dip coating, spray and spin coating, and slip castingeare being used for coating the supports. Also, vacuum seeding is a simple and effective method for support seeding. In this method a uniform seeding layer is formed on the support by adjusting the seeding parameters such as vacuum intensity, seeding time and suspension concentration [15,17]. Due to its low toxicity, ethyl acetate is widely used in industrial processes. Therefore, ethyl acetate purification and dehydration during its production process is an important issue [21]. Zeolite membranes, especially NaA membranes (due to their high hydrophilic nature) are good candidates for dehydration of organic solvents [7]. To our knowledge, there are few published reports on dehydration of highly concentrated ethyl acetate/water mixtures. Dehydration of ethyl actetate/water mixtures by pervaporation process using poly (vinyl alcohol) membranes cross-linked with tartaric acid was studied by Salt and co-workers. Their experiments were conducted at 30, 40, and 50 C with the feed mixtures containing 1e2.5 wt% water. According to their results, water and ethyl acetate fluxes varied in range of 0.008e0.065 and 0.0001e 0.0015 kg m2 h1, respectively. The separation factor was in order of 103e104 [4]. Tian and Jiang used solvent-cast poly (vinylidene fluoride-co-hexafluoropropene) membranes in a pervaporation system for ethyl acetate separation from its aqueous solution. In this work, the concentration of ethyl acetate in the feed mixture was in range of 1.5e3.0 wt% and their membranes showed relatively high flux for this species, ranging from 0.2 to 2 kg m2 h1. However, the reported separation factor of about 102 was low [22]. Zhu et al. used Poly (VDF-co-HFP)/PP composite membranes to separate ethyl acetate from aqueous solution. Their membranes showed low separation factors varying from 50 to 200 [23]. Also, Shah et al. separated ethyl acetate from water solution containing 5 wt% ethyl acetate by using commercial NaA zeolite membranes (Mitsui Engineering and Shipbuilding, Japan) at 60 C. The water flux was very high (2.65 kg m2 h1) and the separation factor was in the moderate range of 200e500 [24]. Recently, dehydration of ethyl acetate/water mixtures using PVA/ceramic composite pervaporation membrane has been reported by Xia and co-workers. Using mixtures containing 5.1 wt% water, they obtained the maximum separation factor of 625 [25]. It is seen that in most cases the separation factor for dehydration of ethyl acetate has been in range of a few hundreds, indicating that making membranes with improved performance is an essential task. In this study, the NaA zeolite membrane was synthesized by vacuum assisted secondary growth method and used for dehydration of an ethyl acetate/water mixture containing 98 wt% of ethyl acetate. The main purpose of this work was to achieve simultaneous desirable separation factor and permeate flux. To this aim, the micrometer size NaA powder was synthesized and then the zeolite layer was grown on a porous alumina support. During the experiments, the effects of some parameters esuch as seeding conditions and number of zeolite layers grown on the porous supporteon the performance of the membranes were investigated. Also, the effect of temperature on the fluxes of water and ethyl acetate was studied. 2. Experimental section 2.1. Synthesis of NaA zeolite seeds NaA zeolite particles were synthesized by conventional heating method in an oven under air atmosphere. The initial aqueous gel for
71
zeolite synthesis was prepared with molar composition of 1.0 Al2O3: 1.92 SiO2: 3.16 Na2O: 134 H2O. Sodium silicate solution (25e 28% SiO2, 7e8%Na2O, 63% H2O, MERCK), aluminium hydroxide (>99%, MERCK), and sodium hydroxide pellets (>98%, MERCK) were used as source materials for Si, Al, and Na elements, respectively. The alumina solution was prepared by dissolving an adequate amount of aluminium hydroxide powder in hot aqueous sodium hydroxide solution at 373 K. Almost half of the total amount of water was used in this stage. On the other hand, the silica solution was prepared by dissolving the necessary amount of sodium silicate in the remaining water. Then, the Si solution was added to the Al solution in a polypropylene bottle and the mixtures were agitated for 15 min to obtain homogeneous gels. Prior to hydrothermal synthesis, the synthesis gels were aged at room temperature for different times. After aging, the NaA powder was synthesized by placing the gels in the oven at 373 K for 3 h. The synthesized powders were thoroughly washed with distilled water followed by centrifuging for several times till the pH of the resultant mixture was lower than 9. 2.2. Synthesis of the zeolite membranes Homemade porous a-Al2O3 tubes (12 mm O.D. and 6 mm I.D.) with average surface pore size of 0.5 mm and porosity of about 43% were used as supports. Before seeding, the surface of each support was polished with sand-paper and cleaned with deionized water in an ultrasonic cleaner bath for 10 min. Finally the clean supports were calcinated at 673 K for 3 h. To proceed with the membrane synthesis process, each pre-treated support was vertically fixed in a Teflon holder. The bottom end of the support was sealed and the upper was connected to a vacuum pump. Then, the support was immersed in a beaker containing the suspension of zeolite NaA seeds. At this stage, the vacuum was applied for a given time to coat the surface of the support with NaA particles as initial seeds for zeolite membrane synthesis. A simple schematic of the system is shown in Fig. 1(a). Due to vacuuming, the suspension of seeds was continuously driven to the surface of support so that a uniform seeding layer was formed on the outer surface of support. After seeding, the support was dried at 383 K at low heating and cooling rate of one K min1 to protect the seed layer from cracking. The membrane synthesis was completed by conventional hydrothermal synthesis method in an oven. For this purpose, at first a synthesis gel with the molar composition of 1 Al2O3: 2 SiO2: 3 Na2O: 200 H2O was prepared. Then, the two ends of the support were sealed by Teflon caps to prevent the contact of the gel with the inner surface of the tubular support. The support was placed vertically in a stainless-steel Teflon-lined autoclave and the synthesis gel was carefully poured into the autoclave. The autoclave was placed in the oven for 6 h at 378 K. After synthesis, the membrane was washed with distilled water until neutral, followed by drying at 383 K for 3 h at a heating and cooling rate of one K min1. 2.3. Pervaporation experiments Pervaporation experiments were conducted on 2 wt% water containing ethyl acetate/water binary mixtures at 303, 313, and 323 K. Analytical grade ethyl acetate was purchased from Arman Sina Co. (Iran). Pervaporation experiments were carried out using a simple system shown in Fig. 1(b). The membrane was positioned in a Teflon holder. The upper part of the holder was connected to the vacuum line via a glass tube and the whole set was placed in a sealed glass flask containing 400 ml of the feed mixture. The pressure at the permeate side was adjusted at 0.8 mbar. The condensed permeate was collected in a U-shape tube placed in a cold trap. Since the total volume of the collected permeate was
72
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76 Table 1 Aging time of the synthesized NaA zeolite powders. Sample no.
Powder
Aging time (h)
1 2 3 4
P1 P2 P3 P4
4 48 72 0
3. Results and discussion 3.1. Powder synthesis results
Fig. 1. Schematics of: (a) apparatus for support seeding by vacuum method; (b) apparatus used for pervaporation experiments.
very small, it was assumed that the concentration of the feed did not change during each experiment. The duration for each experiment was adjusted to 3 h to collect enough amount of permeate for sampling and further analyses. By weighing the permeate, the average permeation flux (kg m2 h1) was determined by Eq. (1):
J ¼ W=Dt$A
Four samples with different aging times were prepared for this part of the work. The aging times and XRD patterns of these samples are presented in Table 1 and Fig. 2, respectively. The XRD pattern of the standard NaA zeolite is also shown in Fig. 2 [26]. As seen in Fig. 2, NaA zeolite phase formation was strongly influenced by aging time. For example, broad peaks of XRD pattern for sample P4, which was prepared with no aging followed by 3 h hydrothermal synthesis, revealed that the sample was amorphous and NaA zeolite was not formed. However, 4 h of aging led to the appearance of a few weak peaks for sample P1. These peaks could be attributed to the formation of NaA zeolite phase, but these peaks were still broad in comparison with those of the standard NaA zeolite. This was due to the existence of an amorphous phase. Hence, it can be concluded that 4 h of aging time was not enough for complete nucleation and only a few numbers of nuclei were formed. On the other hand, since the hydrothermal growth period was short, the formed nuclei did not have enough time to grow and this sample (P1) was amorphous. The XRD patterns of P2 and P3 samples show sharp and intensive peaks. Comparison of XRD patterns of these samples with the pattern for the standard NaA zeolite confirmed that NaA zeolite phase was formed after 48 and 72 h of aging. However, the angle position of some diffraction peaks in XRD pattern of the sample P3, which are marked with (*) in Fig. 2, differed from those of standard
(1)
where W, Dt, and A are total weight of permeate, experiment duration, and the effective surface area of the membrane, respectively. 2.4. Characterization studies Crystalline structure of the membranes was characterized by using X-ray diffraction (XRD) method (Phillips) with Cu Ka radiation. The XRD analyses were conducted at 40 kV and 30 mA with 2q varying from 5 to 50 , step size of 0.05 , and counting time of 1 s per step. The morphology of the synthesized powders and membranes was studied by scanning electron microscopy (SEM) method (Philips, XL30). The composition of the permeate was determined by gas chromatography (GC) method. For this purpose, a Schimadzu GC-20M equipped with a PEG column was used and the carrier gas was helium. The selectivity, awater/ethyl acetate was determined by using Eq. (2):
awater=EAc ¼ ðYwater =YEAc Þ=ðXwater =XEAc Þ;
(2)
where Y is the weight fraction in the permeate and X is that of the feed.
Fig. 2. XRD patterns of NaA zeolite powders: (a) without aging, (b) 4 h aging, (c) 48 h aging, (d) 72 h aging, and (e) standard NaA zeolite.
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76
zeolite. This means that the prolonged aging time was not in favor of forming pure NaA zeolite phase and led to formation of some impure phases. SEM images of the P2 and P3 sample powders are shown in Fig. 3. It has been reported by other researchers that long aging of NaA zeolite at a low temperature had led to production of small particles with the average size up to 0.4 mm [27]. Fig. 3 revealed the same result. The average particle size of sample P2 was about 1.5 mm whereas that of sample P3 was about 1 mm. This proved that increasing the aging time from 48 to 72 h led to particle size reduction. The activation energy of nucleation (z15 kJ mol1) is much less than the energy of crystal growth (z60 kJ mol1). Lower temperatures are more favorable for nucleation and a greater number of nuclei can be formed. On the other hand, at elevated temperatures, the rate of crystal growth is increased [28]. However, aging the reaction mixtures for synthesis of zeolites is often carried out at fairly low temperatures [27]. In the present work, by aging the initial gels at room temperature, the rate of nucleation was higher than that of crystal growth. As a result, small crystals were achieved in the second stage of synthesis process, which was hydrothermal crystal growth. However, as previously pointed out, increasing aging time led to the formation of impure phases. Therefore, aging time of 48 h was chosen for the rest of the work to obtain pure NaA zeolite particles, and sample P2 was used for seeding and membrane synthesis.
Fig. 3. SEM images of synthesized NaA powders with: (a) 48 h aging, (b) 72 h aging.
73
Table 2 Seeding time and concentration of seeding suspensions of as-synthesized membranes. Sample
Seeding time (s)
Suspension concentration (g/L)
M1 M2 M3 M4 M5
90 180 300 90 90
5 5 5 7 10
3.2. Results of membranes synthesis Using the zeolite powder obtained after 48 h of aging (sample P2), five different membranes were synthesized. The conditions of synthesis process for these membranes are listed in Table 2. It was observed that when the seeding time was longer than 90 s (samples M2 and M3) or the concentration of the seeding suspension was more than 5 g L1 (samples M4 and M5), the seeding layer was easily taken off the porous alumina support right after the seeded supports or synthesized membranes were dried. In addition, the aggregation of zeolite particles was observed in some parts of the supports, which revealed the formation of a non-uniform zeolite layer on the support. Therefore, the seeding time and suspension concentration of 90 s and 5 g L1, respectively, were chosen as the most proper conditions for the membrane synthesis (sample M1) from macroscopic point of view. To gain a preliminary idea about the microscopic structure of the selected membrane, some pervaporation experiments were done at 303 K for dehydration of water/ethanol mixture containing 90 wt% of ethanol. A very small separation factor (a ¼ 1.63) along with a high ethanol flux proved that the synthesized zeolite layer on the membrane surface had some cracks or non-zeolite pores. To overcome this problem, the synthesis stage of the zeolite layer was repeated for two more times and the resultant membrane was named M6. Repeating the pervaporation test on the aforementioned water/ethanol mixture in the same conditions by using membrane M6 revealed that repeating synthesis stage was quite effective and the separation factor increased from 1.63 to 29,000. Also, the total permeate flux was 0.592 kg m2 h1. However, increasing the number of synthesized zeolite layer had relatively significant negative effect on the flux. This was due to increased mass transfer resistance by increased zeolite layer thickness. It is shown in Fig. 4 which presents the variation of total permeation flux versus membrane thickness.
Fig. 4. Variation of total permeation flux vs. membrane thickness for M6 membrane.
74
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76
3.3. Ethyl acetate dehydration results The membrane synthesized at the optimized conditions (sample M6) was used for dehydration of ethyl acetate containing 2 wt% water at 303, 313, and 323 K. The membrane showed average total flux of 0.315, 0.208, and 0.147 kg m2 h1 at 323, 313, and 303 K, respectively. Using GC analysis for the feed mixture and permeated product, a separation factor value of about 163,000 was obtained for all temperatures. The very high separation factor found in this work could be justified by considering the nature of the species and the zeolite. The kinetic diameter of ethyl acetate (0.52 nm) is much larger than the pore size of NaA zeolite (0.41 nm). It means that in general it is almost impossible for ethyl acetate molecules to pass through zeolitic pores. On the other hand, both the partial and total permeation fluxes increase in the presence of non-zeolitic pores. Normally non-zeolitic pores in a zeolite layer are due to nonuniformity of the layer and the existence of defects. In this work, it was observed that despite high concentrations of ethyl acetate in the feed mixture, the total flux was in an acceptable range even though the separation factor was very high. This proved that repeating zeolite layer synthesis stage for three times yielded an effective and promising NaA zeolite membrane for dehydration of ethyl acetate. The XRD pattern and SEM images of the membrane are shown in Figs. 5 and 6, respectively. In Fig. 5, the peaks denoted with (*) are related to a-alumina support and other peaks implicate the formation of a pure NaA zeolite layer. Fig. 6(a) shows that a dense layer of NaA zeolite was formed on the support surface and the membrane surface with inter-grown crystals had no appreciable defect, which verified good membrane performance and its high selectivity (a ¼ 163,000). Also, Fig. 6(b) revealed that a uniform zeolite layer with average thickness of 45 mm formed on the support layer. This relatively thick layer increased the mass transfer resistance of the membrane and led to relatively low permeation flux of the synthesized membrane in comparison with the findings by others for permeation flux of water through NaA zeolites [18,24]. However, the flux could be well justified by considering the very high separation factor. One interesting result of this work was that separation factor showed no variation with temperature, as shown in Fig. 7. This can be explained by considering the effect of activation energy. It has been reported that permeation of water through NaA zeolite membranes has activated nature and follows Arrhenius relation [3,24]. The partial fluxes of water and ethyl acetate were calculated using Eq. (3) and the logarithms of these fluxes were plotted against 1/T in Fig. 8.
Fig. 5. XRD pattern of the membrane used for ethyl acetate dehydration (sample M6).
Fig. 6. SEM images of M6 membrane: (a) top view, (b) cross section view.
Ji ¼ Jtotal $Yi
(3)
The average activation energies for water and ethyl acetate permeation through NaA zeolite membrane were determined from slopes of the linear plots of Fig.8 (a) and (b), respectively. The activation energy for water was about 31.4 kJ mol1 which
Fig. 7. Variation of separation factor vs. temperature for M6 membrane.
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76
75
conducted at the temperature range of 30e80 C with the feed mixtures containing 5 wt% of water. However, they did not report any data about the values of water and isopropanol apparent activation energy in their report [29]. In another work, Sommer and Melin [9] investigated the performance of the same membranes as the ones used by Verkerk et al. for the same separation purpose. The conducted their experiments at the temperature range of 60e 120 C and found that the apparent activation energies of water and isopropanol were 32.2 and 52.3 kJ mol1, respectively. They concluded that, due to the smaller activation energy for water compared to isopropanol, the separation factor decreased slightly with increasing temperature. In the case of water/ethanol mixtures, the apparent activation energies for these components were 33.2 and 31.6 kJ mol1, respectively and they reported that the separation factors for these mixtures remained constant with increasing temperature [9]. This was similar to the results obtained in the present work in which water and ethyl acetate apparent activation energies were found to be the same and the separation factor did not change significantly with temperature. 4. Conclusions
Fig. 8. Logarithm of partial permeation flux vs. 1/T for M6 membrane: (a) water, (b) ethyl acetate.
concurred with the findings of other researchers for pervaporation of water/ethanol mixtures through NaA zeolite membranes. Okamoto et al. reported the activation energy of 35 kJ mol1 when the water concentration was in the range of 10wt% [2]. This energy was found to be 41e43 and 51e52 kJ mol1 for the mixtures containing 2.5e59 wt% and 10e100 wt% of water, respectively [3,24]. It should be pointed out that the variation of the activation energy is usual and depends on the purity of the zeolite phase and water concentration. Therefore, it can be said that the aforementioned findings, including the one related to this work, concurred with each other. To examiner’s knowledge, there is no report on the activation energy of ethyl acetate permeation through zeolite membranes. From Fig. 8(b) this value (E z 31.6 kJ mol1) was found very close to that of water. It meant the temperature dependence of the fluxes of both species were almost the same. This was the reason why the separation factor did not change with the variation of temperature. However, it is worth mentioning that different temperature dependence of separation factor has been observed and reported by the researchers. For example, Verkerk et al. reported increasing separation factor by increase in the operational temperature in their work on dehydration of water/isopropanol mixture by using ceramic membranes. Their experiments were
Micron size zeolite NaA powder was synthesized and used for membrane synthesis on a porous tubular alumina support. The support was seeded by the vacuum method and the zeolite layer was made by conventional hydrothermal route. The results showed that aging of the synthesis gel before hydrothermal growth had an important effect on the zeolite formation. The purest NaA zeolite phase was obtained by 48 h of aging at room temperature. Longer aging times led to the formation of smaller particles, but impure phases were formed along with zeolite phase. It was seen that both parameters of seeding time and suspension concentration had strong effects on the result and quality of the seeded layers made by vacuum seeding method. The most stable and uniform layer was obtained at 90 s for seeding time and 5 g L1 suspension concentration. However, high quality membrane was achieved after repeating the synthesis stage three times. The membrane prepared by three stage synthesis process showed very high separation factor value (a ¼ 163,000) and fairly good permeation flux (0.315 kg m2 h1) in comparison with the membranes used by other researchers for the same purpose. However, repeating synthesis process significantly increased the zeolite layer thickness and its mass transfer resistance. Also, it was found that the separation factor did not change with temperature which was due to very close average activation energies of water and ethyl acetate for permeation through the synthesized membrane. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]
Wee S, Ching-Thian T, Subhash B. Sep Purif Technol 2008;63:500e16. Okamoto K, Kita H, Horii K, Tanaka K. Ind Eng Chem Res 2001;40:163e75. Pera-Titus M, Llorens J, Tejero J, Cunill F. Catal Today 2006;118:73e84. lu A, Salt Í, Keles¸r S, Özkan S, Dinçer S. Vacuum 2005;79:215e20. Salt Y, Hasanog Chapman PD, Oliveira T, Livingston AG, Li K. J Memb Sci 2008;318:5e37. Ahn H, Lee H, Lee SB, Lee Y. Desalination 2006;193:244e51. Ahn H, Lee H, Lee SB, Lee Y. J Membr Sci 2007;291:46e52. Smitha B, Suhanya D, Sridhar S, Ramakrishna M. J Membr Sci 2004;241:1e21. Sommer S, Melin T. Chem Eng Sci 2005;60:4509e23. Kazemimoghadam M, Mohammadi T. Desalination 2010;255:196e200. Urtiaga A, Gorri ED, Casado C, Ortiz I. Sep Purif Technol 2003;32:207e13. Li Y, Zhou H, Zhu G, Liu J, Yang W. J Membr Sci 2007;297:10e5. Bowen TC, Noble Richard D, Falconer John L. J Membr Sci 2004;245:1e33. Li Y, Chen H, Liu J, Yang W. J Membr Sci 2006;277:230e9. Huang A, Lin YS, Yang W. J Membr Sci 2004;245:41e51. Das N, Kundu D, Chatterjee M. J Coat Technol Res 2010;7:383e90. Wang Z, Ge Q, Shao J, Yan Y. J Am Chem Soc 2009;131:6910e1. Pera-Titus M, Bausach M, Llorens J, Cunill F. Sep Purif Technol 2008;59: 141e50. [19] Huang A, Yang W. Mater Res Bull 2007;42:657e65. [20] Huang A, Yang W. Sep Purif Technol 2008;61:175e81.
76
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76
[21] Zhang XH, Liu QL, Xiong Y, Zhu AM, Chen Y, Zhang QG. J Membr Sci 2009;327: 274e80. [22] Tian X, Jiang X. J Hazard Mater 2008;153:128e35. [23] Zhu BK, Tian XZ, Xu YY. Desalination 2005;184:71e8. [24] Shah D, Kissick K, Ghorpade A, Hannah R, Bhattacharyya D. J Membr Sci 2000; 179:185e205.
[25] Xia S, Dong X, Zhu Y, Wei W, Xiangli F, Jin W. Sep Purif Technol 2011;77:53e9. [26] Treacy MMJ, Higgins JB. Collection of simulated XRD powder patterns for zeolites. Elsevier; 2001. [27] Cundy CS, Cox PA. Microporous Mesoporous Mater 2005;82:1e78. [28] Brar T, France P, Smirniotis PG. Ind Eng Chem Res 2001;40:1133e9. [29] Verkerk AW. J Membr Sci 2001;193:227e38.
Contents lists available at SciVerse ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
High performance dehydration of ethyl acetate/water mixture by pervaporation using NaA zeolite membrane synthesized by vacuum seeding method M. Moheb Shahrestani a, A. Moheb a, *, M. Ghiaci b a b
Department of Chemical Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran Department of Chemistry, Isfahan University of Technology, Isfahan 84156-83111, Iran
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 April 2012 Received in revised form 9 November 2012 Accepted 30 November 2012
High quality NaA zeolite membranes were prepared by vacuum assisted secondary growth. At the first stage in which NaA powder was synthesized, increasing the aging time led to formation of impure and smaller crystals. However, at the aging time of 48 h, pure NaA zeolite particles with average particle size of 1.5 mm were obtained. In the second stage, the outer surface of porous a-Al2O3 tubular supports were seeded by vacuum method using 1.5 mm NaA particles. The most stable and uniform seeded layer was obtained at seeding time and suspension concentration of 90 s and 5 g L1, respectively. Then, 6 h of secondary growth of the zeolite layer on the seeded surface was carried out at 373 K three times. The XRD and SEM results showed the formation of a uniform and dense layer of pure NaA zeolite with an average thickness of 45 mm. Dehydration experiments were conducted on ethyl acetate/water mixtures with 2 wt% water content. The average total fluxes were 0.147, 0.208, and 0.315 kg m2 h1 at 303, 313, and 323 K, respectively. The separation factor was 163,000. This parameter did not change with temperature and it was due to very close activation energies of ethyl acetate and water. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Aging time Ethyl acetate NaA zeolite membrane Pervaporation Vacuum seeding
1. Introduction Pervaporation is an effective and energy-efficient membrane process which can be used for some liquideliquid separation applications. This process is a promising candidate for cases in which separation encounters difficulties due to liquid physical and chemical properties where it is difficult to achieve desirable results by conventional separation processes such as distillation and extraction [1e3]. During pervaporation process, a liquid feed mixture is contacted with the active side of a dense membrane and one component passes through the membrane to the other side. This is due to the selectivity of the membrane under the act of vacuum. Implementing vacuum force causes the passing species to be evaporated, leaving the membrane body as vapor phase to be later condensed in a cold trap [3e5]. This environmentally clean and economic technology can be used for many applications such as dehydration of aqueouseorganic mixtures, removal of volatile organic compounds from aqueous solutions, and organiceorganic separations [3e8]. Polymeric membranes are widely used for many pervaporation applications [1e3,9]. However, one major
* Corresponding author. Tel.: þ98 311 3915618; fax: þ98 311 3912677. E-mail address: [email protected] (A. Moheb). 0042-207X/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.vacuum.2012.11.019
drawback of the polymeric membranes which limits their practical applications is their insufficient chemical and thermal stability [1e 3,9e11]. Furthermore, swelling of the polymeric membranes generally leads to higher permeability and lower selectivity [1,12]. On the other hand, being free from swelling effect and having good thermal and chemical stabilities of inorganic membranes make them proper candidates for a broad range of separation applications [1,5,12]. Among them, zeolite membranes take advantage of the unique properties of zeolites such asehighly crystalline ordered structure and having molecular-sized pores [3,13]. These traits are very suitable for separating liquid-phase mixtures via pervaporation process. Basically, molecular transport and separation of different molecules through zeolite pores are based on differences in their adsorption and diffusion properties [13]. Various synthesis strategies and methods have been investigated for synthesis of the zeolite membranes [14]. Coating of the zeolite seeds on the surface of a porous support before hydrothermal synthesis, known as secondary growth method, is an effective approach to synthesizing a high quality zeolite membrane [15e19]. In general, synthesis with seeds gives a better control of membrane formation process [15,20,21]. In addition, the secondary growth method facilitates the growth of the zeolite crystals on the support surface (by heterogeneous nucleation) rather than in the bulk of synthesis solution, which in turn prevents transformation of
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76
the zeolite phase into other undesired phases [15]. However, preparing zeolite suspensions and coating the zeolite seeds on a porous support to obtain a uniform seeding layer is still a challenge [15,19]. Several seeding methods such aserub coating, dip coating, spray and spin coating, and slip castingeare being used for coating the supports. Also, vacuum seeding is a simple and effective method for support seeding. In this method a uniform seeding layer is formed on the support by adjusting the seeding parameters such as vacuum intensity, seeding time and suspension concentration [15,17]. Due to its low toxicity, ethyl acetate is widely used in industrial processes. Therefore, ethyl acetate purification and dehydration during its production process is an important issue [21]. Zeolite membranes, especially NaA membranes (due to their high hydrophilic nature) are good candidates for dehydration of organic solvents [7]. To our knowledge, there are few published reports on dehydration of highly concentrated ethyl acetate/water mixtures. Dehydration of ethyl actetate/water mixtures by pervaporation process using poly (vinyl alcohol) membranes cross-linked with tartaric acid was studied by Salt and co-workers. Their experiments were conducted at 30, 40, and 50 C with the feed mixtures containing 1e2.5 wt% water. According to their results, water and ethyl acetate fluxes varied in range of 0.008e0.065 and 0.0001e 0.0015 kg m2 h1, respectively. The separation factor was in order of 103e104 [4]. Tian and Jiang used solvent-cast poly (vinylidene fluoride-co-hexafluoropropene) membranes in a pervaporation system for ethyl acetate separation from its aqueous solution. In this work, the concentration of ethyl acetate in the feed mixture was in range of 1.5e3.0 wt% and their membranes showed relatively high flux for this species, ranging from 0.2 to 2 kg m2 h1. However, the reported separation factor of about 102 was low [22]. Zhu et al. used Poly (VDF-co-HFP)/PP composite membranes to separate ethyl acetate from aqueous solution. Their membranes showed low separation factors varying from 50 to 200 [23]. Also, Shah et al. separated ethyl acetate from water solution containing 5 wt% ethyl acetate by using commercial NaA zeolite membranes (Mitsui Engineering and Shipbuilding, Japan) at 60 C. The water flux was very high (2.65 kg m2 h1) and the separation factor was in the moderate range of 200e500 [24]. Recently, dehydration of ethyl acetate/water mixtures using PVA/ceramic composite pervaporation membrane has been reported by Xia and co-workers. Using mixtures containing 5.1 wt% water, they obtained the maximum separation factor of 625 [25]. It is seen that in most cases the separation factor for dehydration of ethyl acetate has been in range of a few hundreds, indicating that making membranes with improved performance is an essential task. In this study, the NaA zeolite membrane was synthesized by vacuum assisted secondary growth method and used for dehydration of an ethyl acetate/water mixture containing 98 wt% of ethyl acetate. The main purpose of this work was to achieve simultaneous desirable separation factor and permeate flux. To this aim, the micrometer size NaA powder was synthesized and then the zeolite layer was grown on a porous alumina support. During the experiments, the effects of some parameters esuch as seeding conditions and number of zeolite layers grown on the porous supporteon the performance of the membranes were investigated. Also, the effect of temperature on the fluxes of water and ethyl acetate was studied. 2. Experimental section 2.1. Synthesis of NaA zeolite seeds NaA zeolite particles were synthesized by conventional heating method in an oven under air atmosphere. The initial aqueous gel for
71
zeolite synthesis was prepared with molar composition of 1.0 Al2O3: 1.92 SiO2: 3.16 Na2O: 134 H2O. Sodium silicate solution (25e 28% SiO2, 7e8%Na2O, 63% H2O, MERCK), aluminium hydroxide (>99%, MERCK), and sodium hydroxide pellets (>98%, MERCK) were used as source materials for Si, Al, and Na elements, respectively. The alumina solution was prepared by dissolving an adequate amount of aluminium hydroxide powder in hot aqueous sodium hydroxide solution at 373 K. Almost half of the total amount of water was used in this stage. On the other hand, the silica solution was prepared by dissolving the necessary amount of sodium silicate in the remaining water. Then, the Si solution was added to the Al solution in a polypropylene bottle and the mixtures were agitated for 15 min to obtain homogeneous gels. Prior to hydrothermal synthesis, the synthesis gels were aged at room temperature for different times. After aging, the NaA powder was synthesized by placing the gels in the oven at 373 K for 3 h. The synthesized powders were thoroughly washed with distilled water followed by centrifuging for several times till the pH of the resultant mixture was lower than 9. 2.2. Synthesis of the zeolite membranes Homemade porous a-Al2O3 tubes (12 mm O.D. and 6 mm I.D.) with average surface pore size of 0.5 mm and porosity of about 43% were used as supports. Before seeding, the surface of each support was polished with sand-paper and cleaned with deionized water in an ultrasonic cleaner bath for 10 min. Finally the clean supports were calcinated at 673 K for 3 h. To proceed with the membrane synthesis process, each pre-treated support was vertically fixed in a Teflon holder. The bottom end of the support was sealed and the upper was connected to a vacuum pump. Then, the support was immersed in a beaker containing the suspension of zeolite NaA seeds. At this stage, the vacuum was applied for a given time to coat the surface of the support with NaA particles as initial seeds for zeolite membrane synthesis. A simple schematic of the system is shown in Fig. 1(a). Due to vacuuming, the suspension of seeds was continuously driven to the surface of support so that a uniform seeding layer was formed on the outer surface of support. After seeding, the support was dried at 383 K at low heating and cooling rate of one K min1 to protect the seed layer from cracking. The membrane synthesis was completed by conventional hydrothermal synthesis method in an oven. For this purpose, at first a synthesis gel with the molar composition of 1 Al2O3: 2 SiO2: 3 Na2O: 200 H2O was prepared. Then, the two ends of the support were sealed by Teflon caps to prevent the contact of the gel with the inner surface of the tubular support. The support was placed vertically in a stainless-steel Teflon-lined autoclave and the synthesis gel was carefully poured into the autoclave. The autoclave was placed in the oven for 6 h at 378 K. After synthesis, the membrane was washed with distilled water until neutral, followed by drying at 383 K for 3 h at a heating and cooling rate of one K min1. 2.3. Pervaporation experiments Pervaporation experiments were conducted on 2 wt% water containing ethyl acetate/water binary mixtures at 303, 313, and 323 K. Analytical grade ethyl acetate was purchased from Arman Sina Co. (Iran). Pervaporation experiments were carried out using a simple system shown in Fig. 1(b). The membrane was positioned in a Teflon holder. The upper part of the holder was connected to the vacuum line via a glass tube and the whole set was placed in a sealed glass flask containing 400 ml of the feed mixture. The pressure at the permeate side was adjusted at 0.8 mbar. The condensed permeate was collected in a U-shape tube placed in a cold trap. Since the total volume of the collected permeate was
72
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76 Table 1 Aging time of the synthesized NaA zeolite powders. Sample no.
Powder
Aging time (h)
1 2 3 4
P1 P2 P3 P4
4 48 72 0
3. Results and discussion 3.1. Powder synthesis results
Fig. 1. Schematics of: (a) apparatus for support seeding by vacuum method; (b) apparatus used for pervaporation experiments.
very small, it was assumed that the concentration of the feed did not change during each experiment. The duration for each experiment was adjusted to 3 h to collect enough amount of permeate for sampling and further analyses. By weighing the permeate, the average permeation flux (kg m2 h1) was determined by Eq. (1):
J ¼ W=Dt$A
Four samples with different aging times were prepared for this part of the work. The aging times and XRD patterns of these samples are presented in Table 1 and Fig. 2, respectively. The XRD pattern of the standard NaA zeolite is also shown in Fig. 2 [26]. As seen in Fig. 2, NaA zeolite phase formation was strongly influenced by aging time. For example, broad peaks of XRD pattern for sample P4, which was prepared with no aging followed by 3 h hydrothermal synthesis, revealed that the sample was amorphous and NaA zeolite was not formed. However, 4 h of aging led to the appearance of a few weak peaks for sample P1. These peaks could be attributed to the formation of NaA zeolite phase, but these peaks were still broad in comparison with those of the standard NaA zeolite. This was due to the existence of an amorphous phase. Hence, it can be concluded that 4 h of aging time was not enough for complete nucleation and only a few numbers of nuclei were formed. On the other hand, since the hydrothermal growth period was short, the formed nuclei did not have enough time to grow and this sample (P1) was amorphous. The XRD patterns of P2 and P3 samples show sharp and intensive peaks. Comparison of XRD patterns of these samples with the pattern for the standard NaA zeolite confirmed that NaA zeolite phase was formed after 48 and 72 h of aging. However, the angle position of some diffraction peaks in XRD pattern of the sample P3, which are marked with (*) in Fig. 2, differed from those of standard
(1)
where W, Dt, and A are total weight of permeate, experiment duration, and the effective surface area of the membrane, respectively. 2.4. Characterization studies Crystalline structure of the membranes was characterized by using X-ray diffraction (XRD) method (Phillips) with Cu Ka radiation. The XRD analyses were conducted at 40 kV and 30 mA with 2q varying from 5 to 50 , step size of 0.05 , and counting time of 1 s per step. The morphology of the synthesized powders and membranes was studied by scanning electron microscopy (SEM) method (Philips, XL30). The composition of the permeate was determined by gas chromatography (GC) method. For this purpose, a Schimadzu GC-20M equipped with a PEG column was used and the carrier gas was helium. The selectivity, awater/ethyl acetate was determined by using Eq. (2):
awater=EAc ¼ ðYwater =YEAc Þ=ðXwater =XEAc Þ;
(2)
where Y is the weight fraction in the permeate and X is that of the feed.
Fig. 2. XRD patterns of NaA zeolite powders: (a) without aging, (b) 4 h aging, (c) 48 h aging, (d) 72 h aging, and (e) standard NaA zeolite.
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76
zeolite. This means that the prolonged aging time was not in favor of forming pure NaA zeolite phase and led to formation of some impure phases. SEM images of the P2 and P3 sample powders are shown in Fig. 3. It has been reported by other researchers that long aging of NaA zeolite at a low temperature had led to production of small particles with the average size up to 0.4 mm [27]. Fig. 3 revealed the same result. The average particle size of sample P2 was about 1.5 mm whereas that of sample P3 was about 1 mm. This proved that increasing the aging time from 48 to 72 h led to particle size reduction. The activation energy of nucleation (z15 kJ mol1) is much less than the energy of crystal growth (z60 kJ mol1). Lower temperatures are more favorable for nucleation and a greater number of nuclei can be formed. On the other hand, at elevated temperatures, the rate of crystal growth is increased [28]. However, aging the reaction mixtures for synthesis of zeolites is often carried out at fairly low temperatures [27]. In the present work, by aging the initial gels at room temperature, the rate of nucleation was higher than that of crystal growth. As a result, small crystals were achieved in the second stage of synthesis process, which was hydrothermal crystal growth. However, as previously pointed out, increasing aging time led to the formation of impure phases. Therefore, aging time of 48 h was chosen for the rest of the work to obtain pure NaA zeolite particles, and sample P2 was used for seeding and membrane synthesis.
Fig. 3. SEM images of synthesized NaA powders with: (a) 48 h aging, (b) 72 h aging.
73
Table 2 Seeding time and concentration of seeding suspensions of as-synthesized membranes. Sample
Seeding time (s)
Suspension concentration (g/L)
M1 M2 M3 M4 M5
90 180 300 90 90
5 5 5 7 10
3.2. Results of membranes synthesis Using the zeolite powder obtained after 48 h of aging (sample P2), five different membranes were synthesized. The conditions of synthesis process for these membranes are listed in Table 2. It was observed that when the seeding time was longer than 90 s (samples M2 and M3) or the concentration of the seeding suspension was more than 5 g L1 (samples M4 and M5), the seeding layer was easily taken off the porous alumina support right after the seeded supports or synthesized membranes were dried. In addition, the aggregation of zeolite particles was observed in some parts of the supports, which revealed the formation of a non-uniform zeolite layer on the support. Therefore, the seeding time and suspension concentration of 90 s and 5 g L1, respectively, were chosen as the most proper conditions for the membrane synthesis (sample M1) from macroscopic point of view. To gain a preliminary idea about the microscopic structure of the selected membrane, some pervaporation experiments were done at 303 K for dehydration of water/ethanol mixture containing 90 wt% of ethanol. A very small separation factor (a ¼ 1.63) along with a high ethanol flux proved that the synthesized zeolite layer on the membrane surface had some cracks or non-zeolite pores. To overcome this problem, the synthesis stage of the zeolite layer was repeated for two more times and the resultant membrane was named M6. Repeating the pervaporation test on the aforementioned water/ethanol mixture in the same conditions by using membrane M6 revealed that repeating synthesis stage was quite effective and the separation factor increased from 1.63 to 29,000. Also, the total permeate flux was 0.592 kg m2 h1. However, increasing the number of synthesized zeolite layer had relatively significant negative effect on the flux. This was due to increased mass transfer resistance by increased zeolite layer thickness. It is shown in Fig. 4 which presents the variation of total permeation flux versus membrane thickness.
Fig. 4. Variation of total permeation flux vs. membrane thickness for M6 membrane.
74
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76
3.3. Ethyl acetate dehydration results The membrane synthesized at the optimized conditions (sample M6) was used for dehydration of ethyl acetate containing 2 wt% water at 303, 313, and 323 K. The membrane showed average total flux of 0.315, 0.208, and 0.147 kg m2 h1 at 323, 313, and 303 K, respectively. Using GC analysis for the feed mixture and permeated product, a separation factor value of about 163,000 was obtained for all temperatures. The very high separation factor found in this work could be justified by considering the nature of the species and the zeolite. The kinetic diameter of ethyl acetate (0.52 nm) is much larger than the pore size of NaA zeolite (0.41 nm). It means that in general it is almost impossible for ethyl acetate molecules to pass through zeolitic pores. On the other hand, both the partial and total permeation fluxes increase in the presence of non-zeolitic pores. Normally non-zeolitic pores in a zeolite layer are due to nonuniformity of the layer and the existence of defects. In this work, it was observed that despite high concentrations of ethyl acetate in the feed mixture, the total flux was in an acceptable range even though the separation factor was very high. This proved that repeating zeolite layer synthesis stage for three times yielded an effective and promising NaA zeolite membrane for dehydration of ethyl acetate. The XRD pattern and SEM images of the membrane are shown in Figs. 5 and 6, respectively. In Fig. 5, the peaks denoted with (*) are related to a-alumina support and other peaks implicate the formation of a pure NaA zeolite layer. Fig. 6(a) shows that a dense layer of NaA zeolite was formed on the support surface and the membrane surface with inter-grown crystals had no appreciable defect, which verified good membrane performance and its high selectivity (a ¼ 163,000). Also, Fig. 6(b) revealed that a uniform zeolite layer with average thickness of 45 mm formed on the support layer. This relatively thick layer increased the mass transfer resistance of the membrane and led to relatively low permeation flux of the synthesized membrane in comparison with the findings by others for permeation flux of water through NaA zeolites [18,24]. However, the flux could be well justified by considering the very high separation factor. One interesting result of this work was that separation factor showed no variation with temperature, as shown in Fig. 7. This can be explained by considering the effect of activation energy. It has been reported that permeation of water through NaA zeolite membranes has activated nature and follows Arrhenius relation [3,24]. The partial fluxes of water and ethyl acetate were calculated using Eq. (3) and the logarithms of these fluxes were plotted against 1/T in Fig. 8.
Fig. 5. XRD pattern of the membrane used for ethyl acetate dehydration (sample M6).
Fig. 6. SEM images of M6 membrane: (a) top view, (b) cross section view.
Ji ¼ Jtotal $Yi
(3)
The average activation energies for water and ethyl acetate permeation through NaA zeolite membrane were determined from slopes of the linear plots of Fig.8 (a) and (b), respectively. The activation energy for water was about 31.4 kJ mol1 which
Fig. 7. Variation of separation factor vs. temperature for M6 membrane.
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76
75
conducted at the temperature range of 30e80 C with the feed mixtures containing 5 wt% of water. However, they did not report any data about the values of water and isopropanol apparent activation energy in their report [29]. In another work, Sommer and Melin [9] investigated the performance of the same membranes as the ones used by Verkerk et al. for the same separation purpose. The conducted their experiments at the temperature range of 60e 120 C and found that the apparent activation energies of water and isopropanol were 32.2 and 52.3 kJ mol1, respectively. They concluded that, due to the smaller activation energy for water compared to isopropanol, the separation factor decreased slightly with increasing temperature. In the case of water/ethanol mixtures, the apparent activation energies for these components were 33.2 and 31.6 kJ mol1, respectively and they reported that the separation factors for these mixtures remained constant with increasing temperature [9]. This was similar to the results obtained in the present work in which water and ethyl acetate apparent activation energies were found to be the same and the separation factor did not change significantly with temperature. 4. Conclusions
Fig. 8. Logarithm of partial permeation flux vs. 1/T for M6 membrane: (a) water, (b) ethyl acetate.
concurred with the findings of other researchers for pervaporation of water/ethanol mixtures through NaA zeolite membranes. Okamoto et al. reported the activation energy of 35 kJ mol1 when the water concentration was in the range of 10wt% [2]. This energy was found to be 41e43 and 51e52 kJ mol1 for the mixtures containing 2.5e59 wt% and 10e100 wt% of water, respectively [3,24]. It should be pointed out that the variation of the activation energy is usual and depends on the purity of the zeolite phase and water concentration. Therefore, it can be said that the aforementioned findings, including the one related to this work, concurred with each other. To examiner’s knowledge, there is no report on the activation energy of ethyl acetate permeation through zeolite membranes. From Fig. 8(b) this value (E z 31.6 kJ mol1) was found very close to that of water. It meant the temperature dependence of the fluxes of both species were almost the same. This was the reason why the separation factor did not change with the variation of temperature. However, it is worth mentioning that different temperature dependence of separation factor has been observed and reported by the researchers. For example, Verkerk et al. reported increasing separation factor by increase in the operational temperature in their work on dehydration of water/isopropanol mixture by using ceramic membranes. Their experiments were
Micron size zeolite NaA powder was synthesized and used for membrane synthesis on a porous tubular alumina support. The support was seeded by the vacuum method and the zeolite layer was made by conventional hydrothermal route. The results showed that aging of the synthesis gel before hydrothermal growth had an important effect on the zeolite formation. The purest NaA zeolite phase was obtained by 48 h of aging at room temperature. Longer aging times led to the formation of smaller particles, but impure phases were formed along with zeolite phase. It was seen that both parameters of seeding time and suspension concentration had strong effects on the result and quality of the seeded layers made by vacuum seeding method. The most stable and uniform layer was obtained at 90 s for seeding time and 5 g L1 suspension concentration. However, high quality membrane was achieved after repeating the synthesis stage three times. The membrane prepared by three stage synthesis process showed very high separation factor value (a ¼ 163,000) and fairly good permeation flux (0.315 kg m2 h1) in comparison with the membranes used by other researchers for the same purpose. However, repeating synthesis process significantly increased the zeolite layer thickness and its mass transfer resistance. Also, it was found that the separation factor did not change with temperature which was due to very close average activation energies of water and ethyl acetate for permeation through the synthesized membrane. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]
Wee S, Ching-Thian T, Subhash B. Sep Purif Technol 2008;63:500e16. Okamoto K, Kita H, Horii K, Tanaka K. Ind Eng Chem Res 2001;40:163e75. Pera-Titus M, Llorens J, Tejero J, Cunill F. Catal Today 2006;118:73e84. lu A, Salt Í, Keles¸r S, Özkan S, Dinçer S. Vacuum 2005;79:215e20. Salt Y, Hasanog Chapman PD, Oliveira T, Livingston AG, Li K. J Memb Sci 2008;318:5e37. Ahn H, Lee H, Lee SB, Lee Y. Desalination 2006;193:244e51. Ahn H, Lee H, Lee SB, Lee Y. J Membr Sci 2007;291:46e52. Smitha B, Suhanya D, Sridhar S, Ramakrishna M. J Membr Sci 2004;241:1e21. Sommer S, Melin T. Chem Eng Sci 2005;60:4509e23. Kazemimoghadam M, Mohammadi T. Desalination 2010;255:196e200. Urtiaga A, Gorri ED, Casado C, Ortiz I. Sep Purif Technol 2003;32:207e13. Li Y, Zhou H, Zhu G, Liu J, Yang W. J Membr Sci 2007;297:10e5. Bowen TC, Noble Richard D, Falconer John L. J Membr Sci 2004;245:1e33. Li Y, Chen H, Liu J, Yang W. J Membr Sci 2006;277:230e9. Huang A, Lin YS, Yang W. J Membr Sci 2004;245:41e51. Das N, Kundu D, Chatterjee M. J Coat Technol Res 2010;7:383e90. Wang Z, Ge Q, Shao J, Yan Y. J Am Chem Soc 2009;131:6910e1. Pera-Titus M, Bausach M, Llorens J, Cunill F. Sep Purif Technol 2008;59: 141e50. [19] Huang A, Yang W. Mater Res Bull 2007;42:657e65. [20] Huang A, Yang W. Sep Purif Technol 2008;61:175e81.
76
M. Moheb Shahrestani et al. / Vacuum 92 (2013) 70e76
[21] Zhang XH, Liu QL, Xiong Y, Zhu AM, Chen Y, Zhang QG. J Membr Sci 2009;327: 274e80. [22] Tian X, Jiang X. J Hazard Mater 2008;153:128e35. [23] Zhu BK, Tian XZ, Xu YY. Desalination 2005;184:71e8. [24] Shah D, Kissick K, Ghorpade A, Hannah R, Bhattacharyya D. J Membr Sci 2000; 179:185e205.
[25] Xia S, Dong X, Zhu Y, Wei W, Xiangli F, Jin W. Sep Purif Technol 2011;77:53e9. [26] Treacy MMJ, Higgins JB. Collection of simulated XRD powder patterns for zeolites. Elsevier; 2001. [27] Cundy CS, Cox PA. Microporous Mesoporous Mater 2005;82:1e78. [28] Brar T, France P, Smirniotis PG. Ind Eng Chem Res 2001;40:1133e9. [29] Verkerk AW. J Membr Sci 2001;193:227e38.