Science in China Series B: Chemistry © 2009

www.scichina.com chem.scichina.com www.springerlink.com

SCIENCE IN CHINA PRESS

Springer

Synthesis and separation performance of silicalite-1 membranes on silica tubes CHEN HongLiang1,2†, LI YanShuo2, ZHU GuangQi2 & YANG WeiShen2 1 2

Shenyang Key Laboratory of Environmental Engineering, Shenyang 110044, China; State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China

High-performance silicalite-1 membranes were synthesized on silica tubes by in-situ hydrothermal synthesis. By using the “solution-filling (SF)” method, the average flux of membranes with the SF method was improved by about 25% compared to that of the membranes without using the SF method; the flux and the separation factor of the membranes prepared with the SF method for an ethanol/water mixture at 60℃ were 0.99 kg/(m2·h) and 73, respectively. It was found that the membranes synthesized on silica tubes exhibited high thermal stability and high reproducibility, and the relatively standard deviations (R.S.D.) of the average flux and separation factor were only 9.6% and 5.6%, respectively, which suggests that the silica support is more suitable than other kinds of supports for preparing highperformance silicalite-1 membranes. silicalite-1 membrane, separation performance, pervaporation, silica tube

Membrane pervaporation has advantages in separating azeotropes, closing-boiling mixtures, and thermally sensitive compounds, and in removing species present at — low concentrations[1 3]. During the pervaporation process, only a fraction of the mixture can pass the membrane as vapor at lower temperature, and the energy consumption is lower than that of the conventional distillation or extraction. Polymeric membranes, have some drawbacks (such as swelling, low chemical resistance, and low thermal stability) which often limit their application, while zeolite membranes show a higher flux and a higher separation factor. Thus great interest has been focused on the preparation and characterization of zeolite membranes, and significant progress has been re— ported in recent years[4 6]. One important application of hydrophobic silicalite-1 membranes is to extract organics from low concentration aqueous solutions[7,8], such as extracting ethanol from fermentation broth[9]. In order to protect the porous supports from the invasion of the synthesis solution during hydrothermal synthesis, Yan et al.[10] impregnated the support with a mixture of furfurylalcohol and tetra-

ethylorthosilicate (TEOS), and the penetration depth of siliceous matter into the supports decreased from ca. 80 μm without pretreatment to ca. 55 μm with pretreatment. In 2002, Hedlund et al.[11] successfully prepared highflux MFI membranes by masking the porous supports with high-melting-point polyethylene wax, and the flux was one or two orders of magnitude higher than those in the literature, but the procedure was complex and not easy to be used for tubular supports. In order to synthesize high-performance silicalite-1 membranes, two factors should be considered: one is the dissolution of alumina supports during hydrothermal synthesis, which will result in less hydrophobic ZSM-5 membranes instead of the desired hydrophobic silicalite-1 membranes; the other is the different thermal expansion coefficients between supports and zeolite layers, which will result in the formation of cracks during Received April 8, 2008; accepted May 11, 2008 doi: 10.1007/s11426-008-0155-8 † Corresponding author (email: [email protected]) Supported by the National Advanced Materials Committee of China (Grant No. 2003AA328010) and the Ministry of Science and Technology of China (Grant No. G2003CB615802)

Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 579-583

calcination. For example, zeolite membranes synthesized on stainless steel can form cracks easily upon thermal stress because the stainless steel has the highest thermal expansion coefficient among Al2O3, SiO2, and stainless steel supports[4]. Therefore, suitable supports are important to the synthesis of crack-free silicalite-1 membranes. In this study, high-performance and highreproducibility silicalite-1 membranes were successfully prepared by using silica tubes instead of alumina supports; the thermal stability of silicalite-1 membrane, and the separation performance towards simulative fermentation broth were also investigated briefly.

1 Experimental

section). 1.3 Pervaporation measurements The separation performance towards ethanol/water mixtures was carried out on a standard pervaporation apparatus. A centrifugal pump circulated the feed through the system to reduce concentration polarization. The permeate and feed concentrations of ethanol and water were measured by off-line gas chromatography (GC, HP 5890). The total flux was calculated by weighing the condensed permeate. The separation factor was determined as αA/W= (YA/YW)/(XA/XW), where XA, XW, YA, and YW denote the mass fractions of components A (ethanol) and W (water) on the feed and permeate sides, respectively.

1.1 Synthesis of silicalite-1 membranes

1.4 Membrane thermal stability

Homemade porous silica tubes were prepared by the casting method. The silica tubes were 11 mm in OD, 7 mm in ID, 90 mm in length, 0.3 μm in pore diameter, and of 45% porosity. Before hydrothermal synthesis, the tubes were polished with sand paper, washed with deionized water in an ultrasonic apparatus, and calcined at 500℃ for 5 h to remove organics adsorbed on the surface. Synthesis solutions were prepared by mixing NaOH, TPABr, silica sol (26 % SiO2), and deionized water at room temperature. The molar composition of synthesis solutions was TPABr : Na2O : SiO2 : H2O=1 : 0.25 : 10 : X (solution 1: X=800; solution 2: X=1000). Before each hydrothermal synthesis, the silica tubes were filled with mixed solution (containing water and glycerol) or without solution-filling, and then sealed with two Teflon caps at both ends. After the synthesis solution was added, the autoclaves were put in an air oven and aged at 75℃ for 12 h, and then the oven temperature was increased to 180℃. The first crystallization with solution 1 was carried out for 14 h and the second crystallization with solution 2 was carried out for 10 h. After each crystallization, the silicalite-1 membranes were washed with deionized water, dried at 80℃ overnight, and then the silicalite-1 membranes were calcined at 500℃ for 12 h to remove the templates.

In order to investigate the thermal stability of silicalite-1 membranes synthesized on the silica tubes, the membrane SS-1 was calcined repeatedly in the air after removing the templates. Each calcination was carried out at 400℃ for 5 h at a calcination rate of 0.2—5℃/min, and the thermal stability of silicalite-1 membranes was evaluated by separating the ethanol/water mixture.

1.2 Membrane characterization The top view and the cross section of silicalite-1 membranes were characterized by scanning electron microscopy (SEM, JEM-1200), and SEM was also used to estimate the thickness of the silicalite-1 membranes (cross 580

2 Results and discussion 2.1 Effects of solution filling Figure 1 shows the SEM images of the silicalite-1 membranes synthesized on silica tubes. By comparing the membranes synthesized by SF method (Figure 1(a) and (b)) with the membranes without SF method (Figure 1(c) and (d)), it can be seen that the surface of membrane with SF method was more compact than that of membrane without SF method, and the thickness of the former was thinner than that of the latter. Therefore, it can be deduced that a higher flux and a higher separation factor should be obtained because of thinner membrane thickness and more compact surface, and these results from the short path for permeated molecules through zeolite channels and the reduction of defects, which was further verified by pervaporation experiments, as shown in Table 1. In order to improve the separation performance of silicalite-1 membranes, a suitable mixed solution composed of high-boiling-point solvent glycerol (b.p.: 290℃) and water was used to fill the silica tubes. As shown in Table 1, membrane (SS-4) with the highest flux of 0.99

CHEN HongLiang et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 579-583

Figure 1 SEM images of silicalite-1 membranes synthesized on silica tubes. (a) Top view and (b) cross section of membrane with solution filling; (c) top view and (d) cross section of membrane without solution filling. Table 1 Separation performances of silicalite-1 membranes synthesized on silica tubesa) EtOH/H2O (3% EtOH)

Membr. No.

Aging temperature (℃)

Crystallization time (h)

Filling

SS-1

75

14+10

Yes

0.83

SS-2

75

14+10

Yes

0.77

SS-3

75

14+10

Yes

0.83

SS-4

75

14+10

Yes

0.99

Flux (kg·m−2·h−1)

RSDb)

Separation factor

RSDb)

63 9.6%

65 67

5.6%

73

SS-5

75

14+10

No

0.68

60

SS-6

75

14+10

No

0.69

64

a) Feed temperature: 60℃; b) RSD was calculated with STDEVP/average.

kg/(m2·h) and separation factor of 73 was obtained by using the SF method, and these results are higher than those reported in the literature[12, 13]; membrane (SS-5) with the lowest flux of 0.68 kg/(m2·h) and separation factor of 60 was obtained without using the SF method. By using the SF method, the average flux and the average separation factor of four membranes were 0.85 kg/(m2·h) and 67, respectively, and the relative standard deviations (R.S.D.) of average total flux and separation factor were only 9.6% and 5.6%, respectively, which exhibit high reproducibility. Without using the SF method, the average flux of membranes SS-5 and SS-6 was 0.68 kg/(m2·h) with an average separation factor of

62. By comparing the pervaporation data, we can see that the average flux can be improved by 25% while the separation factor was kept higher, and this result is similar to that reported by our group previously[14]. During hydrothermal synthesis, the penetration of synthesis solution into support pores was inevitable, but the growth rate of zeolite crystals in the support pores should be reduced because of a low nutrition concentration resulting from the dilution of the mixed filling solution. After the first crystallization, the penetration of synthesis solution into support pores will be further reduced because of the silicalite-1 film formed on the outer surfaces of silica tubes and the protection of vis-

CHEN HongLiang et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 579-583

581

Table 2 Thermal stability of silicalite-1 membrane synthesized on the silica tube a) Membr. No.

0.2

0.5

1.0

2.0

3.0

4.0

5.0

total flux (kg·m ·h )

0.83

0.83

0.81

0.81

0.76

0.70

0.69

separation factor

63

64

63

67

71

72

72

−2

SS-1

Calcination rate (℃· min−1)

EtOH/H2O (3% EtOH) −1

a) Feed temperature: 60℃. Table 3 Separation performance of silicalite-1 membranes for simulative fermentation broth Feed conditions (60℃)a)

Membrane SS-4

3% EtOH

5% EtOH

5% EtOHb)

5% EtOHc)

Permeate side

After calcination

Ethanol concentration (%)

69

75

79

76

75

Water concentration (%)

31

25

21

24

25

Separation factor

73

56

70

58

57

Flux (kg·m−2·h−1)

0.99

1.02

0.85

0.50

0.58

a) K2HPO4, MgSO4·7H2O and CaCl2·H2O: 1.0 g/L, NH4Cl: 3.0 g/L; b) adding K2HPO4, MgSO4·7H2O, CaCl2·H2O, and NH4Cl; c) adding K2HPO4, MgSO4·7H2O, CaCl2·H2O, NH4Cl, and glucose (100 g/L).

cous solution. After hydrothermal synthesis, an evidence to reduce the penetration of synthesis solution was that zeolite crystals were collected within the tubes without using the SF method, while almost no zeolite crystals were collected within the tubes with the SF method. Therefore, it can be deduced that the formation of zeolite crystals in the supports pores will be restrained. Therefore, the SF method is an effective method to decrease the penetration of synthesis solution, which can effectively improve the separation performance of as-synthesized silicalite-1 membranes, as shown in Table 1. 2.2 Thermal stability of silicalite-1 membrane As shown in Table 2, the silicalite-1 membrane (SS-1) showed high thermal stability, and the membrane still showed high separation performance towards ethanol/water mixtures even with a calcination rate of 5℃/min. It was interesting to note that the separation factor of membranes increased slightly after repeated calcination cycles, and this might be caused by the condensation of Si-OH on the crystal surfaces, which improved the hydrophobicity of silicalite-1 membranes. In order to avoid the formation of cracks during calcination, a lower calcination rate was applied by other researchers[15,16]. Work by Holmes et al.[16] showed that the cracks would formed for membranes synthesized on stainless steel supports at a heating rate above 0.25℃/min, so a lower calcination rate was performed. 582

In this work, the silicalite-1 membrane could sustain a calcination rate of 5℃/min, which shows that silica tubes were more suitable for preparing high-performance silicalite-1 membranes. 2.3 Separation performance towards simulative fermentation broth Table 3 shows the separation performance of silicalite-1 membranes towards simulative fermentation broth; the fermentation broth was prepared according to the paper reported by Nomura[9]. After the ethanol concentration changed from 3% to 5% on the feed side, the ethanol concentration on the permeate side was improved from 69% to 75%. These values were much higher than those of polymeric membranes[17], and the changes were similar to those obtained by Nomura et al.[9]. It was interesting to note that the ethanol concentration in the permeate side was improved obviously after the addition of inorganic salts, and this might be caused by the change in the vapor-liquid equilibrium of the feed after the addition of salts, which would decrease the driving strength of water passing membranes. After the pervaporation experiments were carried out for several hours, the flux decreased slowly, but the separation factor was kept almost constantly. After calcining the silicalite-1 membrane at 400℃ for 5 hours again, the separation performance of silicalite-1 membranes recovered, which shows that the reduction of separation performance might be caused by glucose

CHEN HongLiang et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 579-583

molecule absorbed on the surface of silicalite-1 membranes, and the absorbed glucose molecule decreased the hydrophobicity of silicalite-1 membrane and prevented the permeance of ethanol and water. However, a series of experiments was required to verify the above conclusions.

3 Conclusions High-performance and high-reproducibility silicalite-1 membranes were synthesized by using silica tubes in1

Bowen T C, Noble R D, Falconer J L. Fundamentals and applications

stead of alumina tubes, and the silicalite-1 membranes showed high-performance towards simulative fermentation broth. From a simulative fermentation broth with ethanol concentration of 5%, the permeate ethanol concentration was 79%, and this result was higher than the same concentration of a water-ethanol mixture because of the salt effect. After the pervaporation experiments were carried out for several hours, the flux decreased slowly because of the adsorption of glucose molecule on the membrane surface. 10

of pervaporation through zeolite membrane. J Membr Sci, 2004, 245

preparation of supported zeolite ZSM-5 membranes. J Membr Sci,

(1-2): 1－33 2

3

Thongsukmak A, Sirkar K K. Pervaporation membranes highly se-

Cox D, Deckman H, Gijnst W D, de Moor P P, Lai F, McHenry J, Mortier W, Reinoso J, Peters J. High-flux MFI membranes. Microp

Liu F, Liu L, Feng X. Separation of acetone-butanol-ethanol (ABE) 12

2005, 42(3): 273－282 Caro J, Noack M, Kolsch P, Schafer R. Zeolite membranes-state of 13

14

Li Y, Yang W. Microwave synthesis of zeolite membranes: A review. J

8

Matsusuda H, Yanagishita H, Negishi H. Improvement of ethanol

membranes. J Membr Sci, 2001, 187(1-2): 203－212 Chen H, Song C, Yang W. Effects of aging on the synthesis and performance of silicalite membranes on silica tubes without seeding.

15

Micropor Mesopor Mater, 2007, 102(1-3): 249－257 Lin X, Kita H, Okamoto K. A novel method for the synthesis of high

selectivity of silicalite membrane in pervaporation by silicone rubber

performance silicalite membranes. Chem Commun, 2000, 1889－

coating. J Membr Sci, 2002, 210(2): 433－437

1890

Lin X, Kita H, Okamoto K. Silicalie membrane preparation, charac-

16

terization, and separation performance. Ind Eng Chem Res, 2001, Nomura M, Bin T, Nakao S. Selective ethanol extraction from fer-

Holmes S M, Markert C, Plaisted R J, Forrest J O, Agger J R. A novel method for the growth of silicalite membranes on stainless steel

40(19): 4069－4078 9

monolith supports. J Memb Sci, 2003, 215(1-2): 235－247 Nomura M, Yamaguchi T, Nakao S. Transport phenomena through intercrystalline and intracrystallline pathways of silicalite zeolite

Membr Sci, 2008, 316(1-2): 3－17 7

Mesop Mater, 2002, 52 (3): 179－189. Bowen T C, Kalipcilar H, Falconer J L, Noble R D. Pervaporation of organics/water mixtures through B-ZSM-5 zeolite membranes on

38(1): 3－24 Caro J, Noack M, Kolsch P. Zeolite membranes: From the laboratory scale to technical applications. Adsorption, 2005, 11(3-4): 215－227

6

1997, 126 (1) 53－65. Hedlund J, Sterte J, Anthonis M, Bons A J, Carstensen B, Corcoran N,

2007, 302 (1-2): 45－58

their development and perspective. Micropor Mesopor Mater, 2000, 5

11

lective for solvents present in fermentation broths. J Membr Sci,

from dilute aqueous solutions by pervaporation. Sep Purif Technol, 4

Yan Y, Davis M E, Gavalas G R. Use of diffusion barriers in the

17

supports. Chem Mater, 1999, 11(11): 3329－3332 O’Brien D J, Roth L H, McAloon A J. Ethanol production by con-

mentation broth using a silicalite membrane. Sep Purif Technol, 2002,

tinuous fermentation-pervaporation: a preliminary economic analysis.

27(1): 59－66

J Membr Sci, 2000, 166(1): 105－111

CHEN HongLiang et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 579-583

583