Synthesis of zeolite NaA membranes with high performance and high reproducibility on coarse macroporous supports

Synthesis of zeolite NaA membranes with high performance and high reproducibility on coarse macroporous supports

Journal of Membrane Science 444 (2013) 513–522 Contents lists available at SciVerse ScienceDirect Journal of Membrane Science journal homepage: www...

3MB Sizes 4 Downloads 120 Views

Journal of Membrane Science 444 (2013) 513–522

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

Synthesis of zeolite NaA membranes with high performance and high reproducibility on coarse macroporous supports Huazheng Li a, Jinqu Wang a, Jing Xu a, Xiangdi Meng a, Bo Xu a, Jianhua Yang a,n, Shiyang Li a, Jinming Lu a, Yan Zhang a, Xiaolan He b, Dehong Yin a a b

State Key Laboratory of Fine Chemicals, Institute of Adsorption and Inorganic Membrane, Dalian University of Technology, Dalian, Liaoning 116024, China Chongqing Chemical Industry Vocational College, Chongqing 400020, China

art ic l e i nf o

a b s t r a c t

Article history: Received 18 January 2013 Received in revised form 8 April 2013 Accepted 9 April 2013 Available online 22 April 2013

Pilot-scale zeolite NaA membranes with high PV performance supported on cheap coarse macroporous supports were prepared by one single secondary growth using varying temperature hot dip-coating seeding method (VTHD). Through the VTHD method, a thin, dense and pinhole-free asymmetric NaA seed layer composed of large and small NaA seeds could be manipulated onto the surface of a coarse macroporous support. The large NaA seeds mainly acted as fillers to reduce the pore sizes of the support while the small NaA seeds acted as nuclei to provide sites for NaA crystal growth. The effects of the seed suspension concentrations, the seed sizes and coating temperature on the morphology of the seed layer and the separation performance of the resulting NaA membranes were investigated. The reproducibility of the VTHD method was as high as 70%. The zeolite NaA membrane prepared by the VTHD method showed a water flux of 2.85 kg m−2 h−1 with a separation factor over 10,000 in dehydrating the 90 wt% ethanol/10 wt% water mixture at 343 K. The use of a cheap macroporous support and the high reproducibility of the VTHD method provide the feasibility for large-scale commercial production of low cost zeolite NaA membranes, promoting the broad application of zeolite NaA membranes. & 2013 Elsevier B.V. All rights reserved.

Keywords: Pilot-scale zeolite NaA membrane Varying temperature hot dip-coating method Pervaporation performance High reproducibility Dehydration

1. Introduction Zeolites have been widely used as catalysts [1,2], ion exchangers [3] and adsorbents [4]. Due to their ordered structures, uniform pore sizes, good thermal stability, mechanical strength and chemical resistance, they have shown a great potential for membrane-based applications. In the past two decades, a variety of zeolite membranes, such as LTA [5–10], MFI [11–13], DDR [14], FAU [15] membranes, have been intensively studied as separators [16], reactors [17] and sensors [18]. Among these zeolite membranes, zeolite LTA membranes proved ideally suited for the dehydration of organics by pervaporation or vapor permeation because of their high hydrophilicity and small crystal-graphic pore size. In 1999, Mitsui Engineering &Shipbuilding Co. firstly put zeolite LTA membranes into industrilization for dehydration of alcohol/water mixtures [19]. Recently, the Nano-Research Institute Inc., a 100% subsidiary of Mitsui &Co., has installed vapor permeation units in Brazil (3000 l/d) and India (30,000 l/d) at an operation temperature of 403 K for the dewatering of bioethanol using LTA membranes [20]. Besides, many efforts were focused on development of zeolite NaA membranes with high

n

Corresponding author. Tel.: +86 411 84986147; fax: +86 411 84986147. E-mail addresses: [email protected], [email protected] (J. Yang).

0376-7388/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.memsci.2013.04.030

performance to extend implementation of zeolite NaA membranes into industrial applications [21–23]. Particularly, hollow ceramic fiber supported NaA membranes exhibit an extremely high water flux of about 9.0 kg m−2 h−1 with a selectivity of 10,000 for dehydrating the 90 wt% ethanol/10 wt% water mixture at 348 K were reported [5]. Secondary growth method, coating zeolite crystals on a support surface before hydrothermal synthesis, is widely accepted as a highly efficient method to fabricate high quality zeolite membranes due to its decoupling nucleation and crystal growth process. The seeds act as nuclei for the growth of zeolite crystals, thus the properties of a seed layer including the seed size, continuity, density and thickness are crucial to the properties and the separation performance of the resulting zeolite membranes. Therefore, one of the key steps of secondary growth method is the formation of uniform and continuous seed layer on the outer surface of a support. Many approaches, such as dip-coating [24], rub-coating [25], spin coating [26], vacuum seeding [27] dip-coating-wiping [5], cationic polymer treatment [28] etc., have been explored to coat zeolite seeds on a substrate surface and pinhole-free zeolite membranes were prepared successfully by these methods. However, the supports were generally of low surface roughness and small pore sizes (smaller than 1 μm [5,8,10,27–31]), and they are comfortable for the formation of good quality seed layer. Compared to polymer membranes, the high cost of a zeolite NaA membrane is the main obstacle for its broad application in

514

H. Li et al. / Journal of Membrane Science 444 (2013) 513–522

industry [20,32]. Caro et al. [20] pointed out that the price of a zeolite membrane was governed by the production cost of supports (about 70%) rather than by the zeolite layer. Cheap macroporous supports are desirable for industrial applications due to the low cost and the potential high flux of the resulting zeolite membrane. However, it may be more difficult to gain a high performance zeolite membrane on such a macroporous support because the larger the pore size of the support is, the higher the risk of forming defects will be. In our previous work [33,34], we developed a varying temperature hot dip-coating method (VTHD) for the preparation of high performance zeolite T and MFI membranes on cheap coarse macroporous supports. The VTHD method is mainly composed of three steps: (1) hot dip-coating of seeds at higher temperature; (2) rubbing off the excessive seeds; and (3) hot dip-coating of seeds at lower temperature. This method is found flexible and effective for combined control over the seed suspension concentration, seeding temperature and seed size. In the present work, high performance zeolite NaA membranes in pilot-scale were prepared on low cost coarse macroporous α-Al2O3 tubes with a mean pore size of 2∼3 μm and some large defect pores up to 20 μm using the VTHD seeding method. The use of coarse macroporous supports aimed at reducing the production cost of zeolite NaA membranes. Combined manipulation over the seed size, thickness, coverage and defect of the seed layer were carried out using VTHD seeding method to obtain a thin, compact and pinhole-free NaA seed layer that was proven significant for the achievement of a high performance zeolite NaA membrane. Besides, the reproducibility of the VTHD method was examined.

(498 wt%, Tianjin Kermel Chemical Reagent Co., Ltd), tetramethylammomium hydroxide (TMAOH, 499 wt%, RudongZhenfengYiyang Chemical Co., Ltd), aluminum isopropoxide (499.5 wt%, Tianjin Chem reagent Institute), colloidal silica (SiO2: 25 wt%, Na2O: 0.3 wt %, Qingdao Haiyang Chemical Co., Ltd), sodium aluminate (Al2O3: 41 wt%, Na2O: 24.92 wt%, Sinopharm Chemical Reagent Co., Ltd) were obtained commercially as reagent chemicals and used as received without purification. The deionized water was home-made.

2. Experimental

2.3. Synthesis of zeolite NaA membranes

2.1. Materials

2.3.1. Preparation of seed layers on macroporous substrates The SEM images of the support (Fig. 1b and c) revealed that the pore size of the support is quite large, of 2–3 μm even with some big defects of about 20 μm as marked with red circles in Fig. 1b, and the surface of the support is quite rough. In order to form a continuous, compact and pinhole-free zeolite NaA seed layer on such a coarse support, the VTHD seeding method was applied. The deposition

As shown in Fig. 1a two types (denoted as A and B) of coarse macroporous alumina tubes with an average pore size of 2–3 μm (OD:13 mm, ID: 9 mm, length: 500 mm, Foshan Ceramics Research Institute, China) were used as substrates. If not stated otherwise, type A tubes were used as the supports. Sodium hydroxide

2.2. Synthesis of submicron-size NaA crystal seeds 0.4 μm NaA crystals were hydrothermally prepared according to the previously published procedure [35] with molar composition of 0.32 Na2O: 3.4 SiO2: Al2O3: 8.4 TMAOH: 257H2O. The synthesis solution was prepared by adding 0.544 g of sodium hydroxide and 71.668 g of tetramethylammomium hydroxide into 152.252 g of deionized water, followed by adding 19.108 g of aluminum isopropoxide into the above solution at 313 K. Then 38.112 g of colloidal silica was added dropwise until the mixture was clear. The resultant synthesis solution was sealed into a Teflon-lined stainless steel autoclave at 373 K for 12 h after aging at room temperature for 12 h under vigorous stirring. After the products were recovered by repeated centrifugation and washing with deionized water, dried at 393 K overnight, and finally calcined in air at 773 K for 6 h to remove the templates. The pure phase of the zeolite NaA crystals was confirmed by XRD, and the size and shape were obtained by SEM observation (not shown in this article).

Fig. 1. Optical photo (a) and SEM (b and c) images of the coarse macroporous substrate (inserted red circles: large defect pores). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

H. Li et al. / Journal of Membrane Science 444 (2013) 513–522

515

Table 1 Seeding conditions for zeolite NaA membranes. Seeding method Membr. number Seeding condition First coating step

Second coating step

Seed concentration Seed size Coating temp. Treatment of seed Seed concentration Seed size Coating temp. Treatment of (wt%) (μm) T(K) layer (wt%) (μm) T(K) seed layer HD HD HD HD HD HD VTHD VTHD-S

M1 M2 M3 M4 M5 M6 M7 M8

0.5 1 2 – – – 2 2

2 2 2 – – – 2 2

423 423 423 – – – 423 423

procedures to form the zeolite NaA seed layer were typically carried out as follows: (1) the pre-heated (453 K) tube with two ends sealed with Teflon plugs was dipped into well-dispersed 2 μm NaA seed solution for about 20 s, and followed by drying at 323 K for 3 h and then at 453 K for 3 h; (2) the outer surface of the dried support was rubbed carefully with cotton to remove the superfluous seeds that loosely packed on the surface. Subsequently, the support was heated in an oven at 353 K; (3) the pre-heated support was seeded with well-dispersed submicron-sized seeds. If only step (1) or step (3) is carried out, this coating process was defined as hot dip-coating (HD) method. The effects of the seed concentration and seed size on the seed layer together with the membrane performance were investigated by HD method (M1–M6). For comparison, the supports were also seeded by the VTHD seeding method with the same NaA crystal size in step (1) and (3) (denote as VTHD-S), respectively. Detailed seeding conditions are listed in Table 1. 2.3.2. Hydrothermal synthesis of zeolite NaA membranes The clear synthesis solution was prepared by mixing the sodium aluminate, colloidal silica, sodium hydroxide and distilled water. The molar ratio of the resultant gel mixture was Al2O3: 5SiO2: 50Na2O: 1000H2O. First, 9.852 g of sodium aluminate and 162.195 g of sodium hydroxide were dissolved in 674.08 g of deionized water. When the solution was clear, 47.52 g of colloidal silica was added into the mixture dropwise. The resultant solution was aged at room temperature for 6 h. The seeded supports were placed vertically with a Teflon holder in a stainless steel autoclave to avoid any precipitation during the membrane synthesis. The synthesis gel was poured into autoclave carefully. The crystallization was carried out in an air-circled oven at 353 K for 5 h. After crystallization, the membranes were washed several times with deionized water until the PH value of the water became neutral, then was dried overnight at room temperature.

No rubbing No rubbing No rubbing – – – Rubbing Rubbing

– – – 0.5 1 2 0.5 0.5

– – – 0.4 0.4 0.4 0.4 2

– – – 353 353 353 353 353

– – – No No No No No

rubbing rubbing rubbing rubbing rubbing

Fig. 2. Scheme of pervaporation setup: (1) feed tank, (2) charging pump, (3) pressure gauge, (4) adjusting valve, (5) mass flowmeter, (6)electric furnace, (7) temperature controller, (8) membrane module, (9) cold trap with liquid nitrogen, (10) vacuum gauge, (11) vacuum pump.

Fig. 2. A membrane was mounted in a membrane module (8). The temperature of the feed mixture was controlled at a specified constant value by a temperature-controller (7) using an electric furnace (6) and a thermocouple. The permeate was collected using a cold trap (9) cooled by liquid nitrogen and the permeate pressure of inside membrane was kept below 100 Pa by an oil rotary vacuum pump (11). The removed water through a membrane was compensated by adding the corresponding amount water into feed in an interval to maintain the constant feed composition. The permeation flux was obtained from the weight of the condensate in a specified time interval (several minutes–half an hour). Meanwhile the compositions of the permeate and feed mixtures were determined by a gas chromatograph. The water/alcohol separation factor was defined in Eq. (1). Ya=Yb Xa=Xb

2.4. Characterization of zeolite NaA membranes

α¼

The zeolite NaA membranes were characterized by X-ray diffraction (XRD) with Cu Kα radiation and by scanning electron microscope (SEM) using a KYKY2800B at an acceleration voltage of 15 kV after coating with gold.

where X and Y is the weight fractions of the water components in the feed mixture and in the permeate, respectively. Subscripts a and b refers to water and alcohol components, respectively.

2.5. Pervaporation measurements of the synthesized NaA membranes

3. Results and discussions

ð1Þ

3.1. Synthesis of zeolite NaA membranes by VTHD seeding method The pervaporation performance of the as-synthesized zeolite NaA membranes were tested at 343 K using a 90 wt% ethanol/ 10 wt% water mixture. Schematic diagram of the experimental apparatus used for pervaporation measurements is shown in

3.1.1. Synthesis of zeolite NaA membranes by hot dip-coating It is well recognized that the seed size and the thickness of a seed layer largely affect the morphology and performance of the

516

H. Li et al. / Journal of Membrane Science 444 (2013) 513–522

resulting membrane [36–38] because the seeds act as nuclei provide sites for zeolite crystal growth. The effects of NaA crystal size and the thickness of the NaA seed layer on the morphology and separation performance of NaA membranes were first examined by varying the seed sizes and the concentrations of seed suspension solution. Fig. 3 shows the typical SEM images of the seed layer on the supports which were hot dip-coated using various concentrations of the large (2 μm) seed suspension solution at temperature 423 K, and Fig. 4 is those of the small (0.4 μm) seed suspension solution at 353 K. The seed concentration largely affected the density, coverage and the thickness of the resulting seed layer. For the dilute seed solution of 0.5 wt%, it can be clearly seen that the surface of the substrate was partly covered by NaA seeds with some naked region (Fig. 3a). The thickness of the seed layer was about 2–3 μm (Fig. 3b). In case that the seed concentration was increased to 1 wt%, as shown in Fig. 3c, the seed layer was fully and continuously covered with the NaA crystals, but some pinholes were observed as marked with red circles. This was mainly because the seed concentration was not high enough to cover the large defect pores with NaA crystals. The thickness of the seed layer was 5–6 μm (Fig. 3d). When the concentration of the seed suspension was 2 wt%, the surface of the substrate was densely and fully covered with NaA seeds as shown in Fig. 3e. The thickness of the seed layer was 6–7 μm (Fig. 3f) which was slightly thicker than the thickness of the seed layer of M2 (Fig. 3d).

This pinhole-free seed layer was expected to result in a pinholefree zeolite NaA membrane. When the seed size was reduced to 0.4 μm, for all of the seed concentrations (0.5 wt%, 1 wt%, 2 wt%), large amount of the seeds penetrated into the depth of the support pores (Fig. 4b, d, f), leaving a discontinuous and uncompact seed layer formed on the surface with large amount of defects and pinholes presenting (Fig. 4a, c, e). The seed layers were much thinner than those of the M1–M3. After seeding, all the seeded supports were subjected to the crystallization reaction under the same condition to grow zeolite NaA membranes. XRD patterns for each of the membranes M1–M6 (not shown here) represented the characteristic peaks of LTA zeolite together with the peaks of the support, indicating that pure zeolite NaA membranes were formed on the macroporous supports. Figs. 5 and 6 show the SEM images of the prepared membranes of M1–M3 and M4–M6, respectively. For all the membranes that coated with either large or small NaA seeds with various seed concentrations, the support surfaces were fully covered with intergrown cubic-shape zeolite NaA crystals. However, either intercrystalline grain boundary defects or cracks were observed for all the membranes as marked with red circles. The defects reasonably resulted from the pinhole defects and the lower seed coverage in the seed layer as stated above. Cracks were also observed for the membranes M1–M3, resulting from the thick zeolite layer [39] of M1, M2 and M3, which was 10 μm, 11 μm and 14 μm, respectively.

Fig. 3. SEM images of the seed layer of M1 (a and b), M2 (c and d) and M3 (e and f) (inserted red circle: large defect pores). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

H. Li et al. / Journal of Membrane Science 444 (2013) 513–522

517

Fig. 4. SEM images of the seed layer of M4 (a and b), M5 (c and d) and M6 (e and f) (inserted red circle: defect sites). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

It seems that the roughness and grain boundary defects increased in the order of M1 4M24 M3 because of the increasing defects and decreasing coverage of the seed layer. It is reported that the small size seeds, due to their smaller weight, smaller sizes and larger specific surface area [36,38], are more suitable for preparing denser and thinner zeolite membranes compared with micron-size seeds. Fig. 6 shows the SEM images of the prepared membranes of M4–M6 coated with small NaA seeds of 0.4 μm with various seed concentrations. Compared to the membranes M1–M3, the surfaces of the membranes M4–M6 were fully covered with much smaller cubic NaA crystals because of the former small seeds. From Fig. 6b, d and f, the zeolite layers were hardly distinguished from the substrates due to the penetration of the small seeds into the depth of the pores. The thickness of the M4–M6 was about 4 μm, 5 μm, 6 μm, respectively, which was much thinner than that of the corresponding M1, M2 and M3 due to the much samller seed size. On the other hand, larger boundary defects were observed compared to the membranes M1–M3, resulting from much lower coverage and larger defects of the seed layer of M4–M6 as stated above. The PV performance of the prepared zeolite NaA membranes were tested using pilot-scale PV equipment. The feed was a 90 wt% ethanol/10 wt% water mixture, and the temperature was fixed at 343 K. The results obtained from the different membranes are summarized in Table 2. The fluxes of M1, M2 and M3 were 3.02 kg m−2 h−1, 2.87 kg −2 −1 m h and 2.01 kg m−2 h−1, with separation factors of 379, 490 and

1120, respectively. For the membranes coated with 0.4 μm NaA seeds, the fluxes of M4, M5 and M6 were 3.35 kg m−2 h−1, 3.11 kg m−2 h−1, 2.86 kg m−2 h−1, respectively, and the separation factors were as low as 13, 24 and 309, respectively. It is clear that the flux decreased while separation factor increased with increasing seed concentration and seed size. It is because the thickness of the seed layer increased with both increasing seed concentration and seed size. It is well known that thick seed layer lead to thick zeolite membrane. Instead, the coverage/density increased and the large defects or pinholes decreased with increasing seed concentration and seed size as observed above, leading to increasing separation factor.

3.1.2. Synthesis of zeolite NaA membrane using NaA seeds with different crystal size by the VTHD method As discussed above, compared to the large NaA seeds of 2 μm, the small NaA seeds of 0.4 μm led to a less continuous and compact seed layer with larger amount of defects and pinholes exsiting, forming NaA membranes with higher fluxes but much lower separation selectivities. The much lower separation selectivities are due to the presence of much larger defects in the seed layer that resulted from the small seeds penetrating into the depth of the pores. In our previous work, the VTHD seeding method was found feasible for combined control over the seed suspension concentrations, the seed sizes and the thickness. A dense, continuous, thin and pinhole-free seed layer could be formed on the surface of a coarse macroporous

518

H. Li et al. / Journal of Membrane Science 444 (2013) 513–522

Fig. 5. Typical SEM images of M1, M2 and M3 (a, c, e for the top surface and b, d, f for the cross-section, inserted red circles: visible cracks or pinholes). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

support by rationally constructing an asymmetric structure seed layer that is composed of large and small seeds. The large seeds mainly act as fillers to smooth the surface and reduce the pore/defect sizes of the support for providing a comfort support structure to deposit the small size seeds. The small size crystals mainly serve as nuclei to provide advantageous sites that are beneficial for the formation of polycrystalline zeolite membranes. In order to form a continuous, dense, pinholefree seed layer on our coarse macroporous support using the small NaA seeds, in this work the VTHD method was applied to deposit NaA crystal seeds on the macroporous supports to prepare pilot-scale zeolite NaA membranes with high performance. The VTHD seeding procedures to deposit NaA seeds are as follows: (1) hot dip-coating of 2 wt% large NaA seed (2 μm) suspension at high temperature of 423 K; (2) rubbing off loose and superfluous NaA seed crystals on the support surface; (3) hot dip-coating of 0.5 wt% small seed (0.4 μm) suspension at low temperature of 353 K. Fig. 7a and b gives the top and cross section of SEM images of the support after step 1 and step 2, respectively. It can be seen from Fig. 7a, that the pores, especially the large defect pores, were filled with seeds after wiping the superfluous seeds. Seeds randomly distributed in the pits on the support instead of forming a thick, dense and continuous seed layer before wiping. In this way, the roughness of the substrate was decreased. The large seeds remained on the substrate acted as fillers at this stage, and made it comfortable for the coating of the small seeds in the next step. After hot dip-coated the small NaA seeds onto

the rubbed support (step3), instead of a discontinuous seed layer as in Fig. 4a, c, e, a dense, compact and continuous seed layer was formed on the support without pinholes/cracks visible. The seed layer was thin, about 1–2 μm. For comparison, instead of depositing the small NaA seeds in step (3), the large NaA seeds (2 μm) was hot dip-coated onto the rubbed support for formation of seed layer of M8 while keeping all the procedures same as that of M7. A seed layer similar to that of M3 was formed. However, the thickness of the M8 was about 11–12 μm, larger than that of M3, even dilute seed suspension solution was coated. As discussed in our previous reports [33,34], the drive force of the VTHD method for moving seeds onto the surface of the support includes both of the capillary force and the temperaturedriven pressure difference. The temperature-driven pressure difference particularly contributes to drive the large zeolite seeds into pinholes or crackes in which the capillary force was weak because of the large pore/defect sizes, eliminating the negative influence on the formation of pinhole-free seed layer. On the other hand, the reduced effective pore sizes of the support after step (1) and (2) were more favorable for driving the seeds onto the surface of the support due to the increased capillary force. This was the reason why a compact, dense, continuous, pinhole-free seed layer could be formed on the support of M7 and the thickness of seed layer of M8 was larger than that of M3. Besides the peaks of the support, only LTA characteristic peaks were detected on the XRD patterns (not shown here) for the

H. Li et al. / Journal of Membrane Science 444 (2013) 513–522

519

Fig. 6. SEM images of the zeolite membrane of M4 (a and b), M5 (c and d) and M6 (e and f).

Table 2 The PV performances of synthesized zeolite NaA membranes for dehydrating 90 wt% ethanol/10 wt% water mixture at 343 K. Sample no.

Pervaporation temperature (K)

Water flux (kg m−2 h−1)

Separation factor (α)

M1 M2 M3 M4 M5 M6 M7 M8

343 343 343 343 343 343 343 343

3.02 2.87 2.01 3.35 3.11 2.86 2.43 2.03

379 490 1120 13 24 309 4 10,000 1230

membrane M7 and M8. The compact, continuous and pinhole-free seed layer of M7 led to the formation of the compact and pinholefree zeolite NaA membrane as shown in Fig. 8a. The thickness of the membrane M7 was 4–5 μm, much thinner than that of M3. Instead, the membrane M8 was composed of continuous but less inter-grown larger cubic-shape NaA crystals. Inter-crystalline voids could be seen from the cross-section view (Fig. 8d) similar to that of M3.The thickness of M8 was about 14 μm. The pervaporation separation through the membrane M7 and M8 was conducted under the same conditions as those of the M1–M6. The membrane M7 showed a flux of 2.43 kg m−2 h−1 and an excellent selectivity with separation factor larger than 10,000. Instead, the

membrane M8 showed a similar separation performance to that of M3, with the flux and separation factor of 2.03 kg m−2 h−1 and 1230 kg m−2 h−1, respectively. The high performance of M7 was due to the absence of pinhole defects in the thinner membrane as mentioned above. To our surprise, the thickness of the membrane M8 was about two times as much as that of membrane M7 while its flux was only slightly smaller than that of the M7, and didn't linearly decrease with the membrane thickness. One possible reason was that the difference in the microstructures of the two membranes, including the grain boundaries and inter-crystalline voids, surely affected the permeation of water molecules. The larger inter-crystalline voids in the membrane M8 might allow more water molecules permeating through the membrane. The above results indicated that the VTHD method is flexible and effective for combined control of the seed suspension concentration, seeding temperature and seed size. The large NaA crystals remained in the pores of support after rubbing mainly act as fillers, preventing the penetration of the small NaA seeds into the depth of the support and allowing the formation of a dense, continuous and pinhole-free seed layer on the support. It is also indicated that the VTHD method is effective for successful synthesis of high performance zeolite NaA membranes on cheap coarse macroporous supports. 3.2. The reproducibility of VTHD seeding method Besides the feasibility on the successful formation of pinholefree zeolite NaA membranes on a coarse symmetric macroporous

520

H. Li et al. / Journal of Membrane Science 444 (2013) 513–522

Fig. 7. SEM images of the seed layer of M7 after step 2 (a and b), and step3 (c and d) and the seed layer of M8 after step 3 (e and f).

Fig. 8. SEM images of the top and cross-section surface of the prepared zeolite NaA membranes M7 (a and b) and M8 (c and d).

H. Li et al. / Journal of Membrane Science 444 (2013) 513–522

support, the reproducibility, a crucial factor affecting the industrial mass production, must be taken into account. To investigate the reproducibility of the VTHD seeding method, a series of zeolite NaA membranes were produced on the supports of 500 mm in length. Six batches of zeolite NaA membranes were synthesized on two types (denoted as A and B, with porosity of 35% and 40%, respectively) of the macroporous supports, the conditions of both seed deposition and membrane crystallization reaction were the same as those of M7. Table 3 listed the PV test results of those zeolite NaA membranes. An interesting finding was that, all of 12 membranes exhibited almost the same separation selectivities and the morphologies as shown in Table 3 and Fig. 9, respectively. However, the NaA membranes supported on the B type tubes showed water fluxes of 2.83 70.1 kg m−2 h−1 that were higher than the NaA membranes on the A type tubes of 2.43 7 0.1 kg m−2 h−1, and are competitively high to the reported ones of 0.52–2.15 kg m−2 h−1 [10,16,19,22,27,28,30]. This flux difference is probably due to the difference in the porosity of the two type

521

supports, and similar result was reported to the NaA membranes supported on the hollow ceramic fibers [5]. The PV test results suggested that the reproducibility of the VTHD seeding method was high about 70%. Considering that the cost of the support used in this work is from one fifth to one third to that of the Japan mullite tube and the support with pore size of 1 μm or smaller available in China, the cost of the support in this manuscript is about 40% of the total cost of the membrane. The cost of our zeolite NaA membranes can be decreased by 40%. Therefore, the VTHD seeding method can be employed for the production of industry-scale zeolite NaA membranes on cheap macroporous substrates.

4. Conclusions In conclusion, a simple seeding method, namely, VTHD seeding method was applied to deposit NaA seeds onto a coarse macroporous α-Al2O3 tubular support with large defect pores that are

Table 3 The PV performances of synthesized zeolite NaA membranes on 500 mm length macroporous supports by VTHD seeding method in dehydrating 90 wt% ethanol/10 wt% water mixture at 343 K. Batch

First Second Third Fourth Fifth Sixth

a

Membrane

Aa-1-1 A-1-2 A-2-1 A-2-2 A-3-1 A-3-2 Ba-4-1 B-4-2 B-5-1 B-5-2 B-6-1 B-6-2

PV performance Water flux (kg m−2 h−1)

Separation factor (α)

2.45 2.51 2.42 2.37 2.38 2.43 2.85 2.79 2.87 2.89 2.78 2.83

4 10,000 8,079 4 10,000 9,237 1,321 9,890 4 10,000 9,180 2,796 9,789 4 10,000 8,523

Porosity of A and B type of the support is 35% and 40%, respectively.

Fig. 9. Typical SEM images of zeolite layer of A-2-2 (a and b) and B-5-2 (c and d).

522

H. Li et al. / Journal of Membrane Science 444 (2013) 513–522

harmful to the formation of a dense and pinhole-free NaA seed layer. After hot dip-coating the large (2 μm) seeds, followed by rubbing the loose and superfluous NaA seeds on the support, the large seeds could effectively plug the pores/defects of the low cost supports, meanwhile, made it comfortable for the coating of the small (0.4 μm) NaA seeds in the next step. A thin, compact, continuous and asymmetric zeolite NaA seed layer composed of the large and small NaA crystal seeds can be formed on coarse macroporous tubular α-Al2O3 supports by VTHD method. After one single secondary growth, NaA membranes with high pervaporation performance were obtained for dehydration of ethanol. The flux was 2.85 kg m−2 h−1 with separation factor over 10,000 for the 90 wt% ethanol/10 wt% water mixture at 343 K. The VTHD reproducibility was high, possesing a promising potential in mass production of NaA membranes for industrilization applications.

[16]

[17]

[18]

[19]

[20] [21]

[22]

Acknowledgements We are grateful to the financial support from National Natural Science Foundation of China (No. 21076029) and Program for New Century Excellent Talents in University (NCET-10-0286) and National University Student Innovation Program (No. 201110141038). References [1] B. Chan, Leo Radom, Zeolite-catalyzed hydrogenation of carbon dioxide and ethane, J. Am. Chem. Soc. 130 (2008) 9790–9799. [2] P. Kuhn, P. Pale, J. Sommer, B. Louis, Probing Cu-USY zeolite reactivity: design of a green catalyst for the synthesis of diynes, J. Phys. Chem. C 113 (2009) 2903–2910. [3] S.W Carr, 29Si27Al and 1H solid-state NMR study of the surface of zeolite MAP, Chem. Mater. 9 (1997) 1927–1932. [4] C. M. Lew, R. Cai, Y.S. Yan, Zeolite thin films: from computer chips to space stations, Acc. Chem. Res. 43 (2) 210–219. [5] Z.B. Wang, Q.Q. Ge, J. Shao, Y.S. Yan, High performance zeolite LTA pervaporation membrane on ceramic hollow fiber by dip coating-wiping seed deposition, J. Am. Chem. Soc. 131 (2009) 6910-01911. [6] Q.Q. Ge, Z.B. Wang, Y.S. Yan, High-performance zeolite NaA membrane on polymer–zeolite composite hollow fiber supports, J. Am. Chem. Soc. 131 (2009) 17056–17057. [7] S.Y. Guo, C.L. Yu, X.H. Gu, W.Q. Jin, J. Zhong, C.L. Chen, Simulation of adsorption, diffusion, and permeability of water and ethanol in NaA zeolite membranes, J. Membr. Sci. 376 (2011) 40–49. [8] W.H. Yuan, H.R.. Chen, R..R.. Chang, L. Li, Synthesis and characterization of high performance NaA zeolite–polyimide composite membranes on a ceramic hollow fiber by dip-coating deposition, Desalination 273 (2011) 343–351. [9] A.S. Huang, J. Caro, Cationic polymer used to capture zeolite precursor particles for the facile synthesis of oriented zeolite LTA molecular sieve membrane, Chem. Mater. 22 (2010) 4353–4355. [10] C.H. Cho, K.Y. Oh, S.K. Kim, J.G. Yeo, Y.M. Lee, Improvement in thermal stability of NaA zeolite composite membrane by control of intermediate layer structure, J. Membr. Sci. 366 (2011) 229–236. [11] Y. Liu, Y.S. Li, W.S. Yang, Phase-segregation-induced self-assembly of anisotropic MFI microbuilding blocks into compact and highly b-oriented monolayers, Langmuir 27 (2011) 2327–2333. [12] W. Xiao, J.H. Yang, J.M.. Lu, J.Q. Wang, Preparation and characterization of silicalite-1 membrane by counter-diffusion secondary growth, J. Membr. Sci. 345 (2009) 183–190. [13] G. Li, E. Kikuchi, M. Matsukata, A study on the pervaporation of water-acetic acid mixtures through ZSM-5zeolite membranes, J. Membr. Sci. 218 (2003) 185–194. [14] J. Kuhn, K.. Yajima, T. Tomita, J. Gross, F. Kapteijn, Dehydration performance of a hydrophobic DD3R zeolite membrane, J. Membr. Sci. 321 (2008) 344–349. [15] K. Sato, K. Sugimoto, Y. Sekine, M. Takada, M. Matsukata, T. Nakane, Application of FAU-type zeolite membranes to vapor/gas separation under high

[23]

[24]

[25]

[26]

[27]

[28] [29]

[30]

[31]

[32]

[33]

[34]

[35] [36]

[37]

[38]

[39]

pressure and high temperature up to 5 MPa and 180 1C, Microporous Mesoporous Mater. 101 (2007) 312–318. Y.S. Li, H.L. Chen, J. Liu, H.B. Li, W.S. Yang, Pervaporation and vapor permeation dehydration of Fischer–Tropsch mixed-alcohols by LTA zeolite membranes, Sep. Purif. Technol. 57 (2007) 140–146. B.H. Jeong, K.I. Sotowa, K. Kusakae, Modeling of an FAU-type zeolite membrane reactor for the catalytic dehydrogenation of cyclohexane, Chem. Eng. J. 103 (2004) 69–75. M. Arvand, M.. Kermanian, M..A.. Zanjanchi, Direct determination of aluminum in foods and pharmaceutical preparations by potentiometry using an AlMCM41 modified polymeric membrane sensor, Electrochim. Acta 55 (2010) 6949–6952. Y. Morigami, M.. Kondo, J. Abe, H. Kita, K.. Okamoto, The first large-scale pervaporation plant using tubular-type module with zeolite NaA membrane, Sep. Purif. Technol. 25 (2001) 251–260. J. Caro, M.. Noack, P. KÖlsch, Zeolite membranes: from the laboratory scale to technical applications, Adsorption 11 (2005) 215–227. K. Sato, T. Nakane, A high reproducible fabrication method for industrial production of high flux NaA zeolite membrane, J. Membr. Sci. 301 (2007) 151–161. Y.S. Li, H.L. Chen, J. Liu, W.S. Yang, Microwave synthesis of LTA zeolite membranes without seeding, J. Membr. Sci. 277 (2006) 230–239. C.L. Yu, C. Zhong, Y.M. Liu, X.H. Gu, G. Yang, W.H. Xing, N.P. Xu, Pervaporation dehydration of ethylene glycol by NaA zeolite membranes, IChemE 90 (2012) 1372–1380. G.L. Shao, J.H. Yang, X.F. Zhang, G. Zhu, J.Q. Wang, C. Liu, Seeded growth of beta zeolite membranes using zeolite structure-directing agent, Mater. Lett. 61 (2007) 1443–1445. K. Kusakabe, T. Kuroda, A. Murata, S. Morooka, Formation of a Y-type zeolite membrane on a porous α-alumina tube for gas separation, Ind. Eng. Chem. Res. 36 (1997) 649–655. L. Tosheva, B. Mihailova, L.H. Wee, B. Gasharova, K. Garbev, A.M. Doyle, Indirect observation of structured incipient zeolite nanoparticles in clear precursor solutions, Angew. Chem. Int. Ed. 47 (2008) 8650–8653. A.S. Huang, W.S. Yang, J. Liu, Synthesis and pervaporation properties of NaA zeolite membranes prepared with vacuum-assisted method, Sep. Purif. Technol. 56 (2007) 158–167. A.S. Huang, W.S. Yang, Enhancement of NaA zeolite membrane properties through organic cation addition, Sep. Purif. Technol. 61 (2008) 175–181. Y.M. Liu, Z.Z. Yang, C.L. Yu, X.H. Gu, N.P. Xu, Effect of seeding methods on growth of NaA zeolite membranes, Microporous Mesoporous Mater. 143 (2011) 348–356. H. Ahn, H. Lee, S.B. Lee, Y. Lee, Dehydration of TFEA/water mixture through hydrophilic zeolite membrane by pervaporation, J. Membr. Sci. 291 (2007) 46–52. K. Sato, K. Aoki, K. Sugimoto, K. Izumi, S. Inoue, J. Saito, S. Ikeda, T. Nakane, Dehydrating performance of commercial LTA zeolite membranes and application to fuel grade bio-ethanol production by hybrid distillation/vapor permeation process, Microporous Mesoporous Mater.. 115 (2008) 184–188. Z.B. Wang, Q.Q. Ge, J.S. Gao, J. Shao, C.J. Liu, Y.S. Yan, High-performance zeolite membranes on inexpensive large-pore supports: highly reproducible synthesis using a seed paste, ChemSusChem 4 (2011) 1570–1573. W. Xiao, Z. Chen, L. Zhou, J.H. Yang, J.M. Lu, J.Q. Wang, A simple seeding method for MFI zeolite membrane synthesis on macroporous support by microwave heating, Microporous Mesoporous Mater. 142 (2011) 154–160. X.X. Chen, J.Q. Wang, D.H. Yin, J.H. Yang, J.M. Lu, Y. Zhang, Z. Chen, Highperformance zeolite T membrane for dehydration of organics by a new varying temperature hot-dip coating method, AlChE J. 59 (2013) 936–947. S. Mintova, N.H. Olson, V. Valtchev, T. Bein, Mechanism of zeolite A nanocrystal growth from colloids at room temperature, Science 283 (1999) 958–960. J. Shao, Q.Q. Ge, L.J. Shan, Z.B. Wang, Y.S. Yan, Influences of seeds on the properties of zeolite NaA membrane on alumina hollow fibers, Ind. Eng. Chem. Res. 50 (2011) 9718–9726. T.A. Kuzniatsova, M.L. Mottern, W.V. Chiu, Y. Kim, P.K. Dutta, H. Verweij, Synthesis of thin, oriented zeolite NaA membranes on a macroporous support, Adv. Funct. Mater. 18 (2008) 952–958. Z.Z. Yang, Y.M. Liu, C.L. Yu, X.H. Gu, N.P. Xu, Ball-milled NaA zeolite seeds with submicron size for growth of NaA zeolite membranes, J. Membr. Sci. 392–393 (2012) 18–28. L.L. Lai, J. Shao, Q.Q. Ge, Z.B. Wang, Y.S. Yan, The preparation of zeolite NaA membranes on the inner surface of hollow fiber supports, J. Membr. Sci. 409– 410 (2012) 318–328.