Separation of acetic acid-water mixtures by pervaporation through silicalite membrane

Separation of acetic acid-water mixtures by pervaporation through silicalite membrane

j o u r n a l of MEMBRANE SCIENCE ELSEVIER Journal of Membrane Science 123 (1997) 225-233 Separation of acetic acid-water mixtures by pervaporation ...

662KB Sizes 6 Downloads 293 Views

j o u r n a l of MEMBRANE SCIENCE ELSEVIER

Journal of Membrane Science 123 (1997) 225-233

Separation of acetic acid-water mixtures by pervaporation through silicalite membrane Tsuneji Sanoa,*, Shigeyuki Ejiri a, Kiyoshi Yamada a, Yusuke Kawakami a Hiroshi Yanagishita b a Japan Advanced Institute of Science and Technology, Tatsunokuchi, lshikawa 923-12, Japan b National Institute of Materials and Chemical Research, Tsukuba, Ibaraki 305, Japan

Received 19 April 1996; revised 15 July 1996; accepted 16 July 1996

Abstract Polycrystalline silicalite membranes were prepared on two kinds of porous supports by hydrothermal synthesis. The pervaporation performance of the silicalite membrane obtained was investigated using an acetic acid-water mixture as a feed. The silicalite membrane on the sintered stainless steel support selectively permeates acetic acid in the concentration of the feed acetic acid in the region of 5 to 40 vol%. However, the membrane on the porous alumina support showed no separation for the aqueous acetic acid solution. From the fact that the top layer of the membrane on the alumina support was not composed of pure silicalite but ZSM-5 zeolite crystals, which contained BrOnsted acidic sites (Si(OH)A1) in the framework, it was suggested that the acidic sites associated with the framework aluminums play an important role in the separation of the acetic acid-water mixture. A long-term test of the pervaporation was also carried out to clarify the stability of the membrane. Keywords: Pervaporation; Acetic acid; Adsorption; Zeolite; Silicalite membrane

1. Introduction Pervaporation is a membrane separation process involving l i q u i d - g a s systems applicable to a variety of azeotropic liquid mixtures. There are a lot of articles and patents related to many kinds of organic membranes for separation of acetic a c i d - w a t e r mixtures by pervaporation, as acetic acid is one of the most important intermediates in the chemical and the food industries. Most papers deal with studies on the dehydration o f acetic a c i d - w a t e r mixtures [1]. Very

* Corresponding author.

recently, the preferential pervaporation of acetic acid through a few organic membranes was found [2-4]. However, scarcely any studies have dealt with inorganic membranes. Zeolite has been focused on as one of the materials for the inorganic membranes because of its molecular sieving property, thermal resistance and chemical stability [5-12]. W e have studied the preparation of pure zeolite membrane such as silicalite, ZSM-5 and SAPO-5 from the standpoint of understanding the growth process of zeolite crystals [13-15]. As the A1 distribution and the change in the morphology along the width of membrane reflects the time course of the crystal growth, we can get

0376-7388/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved. PI1 S03 76-7 3 8 8(96)00224-4

226

T. Sano et al. / Journal of Membrane Science 123 (1997) 225-233

much information regarding the crystal growth process by analyzing the membrane obtained. In the course of studies of the preparation of pure zeolite membranes, we found that the preferential pervaporation of organic compounds such as ethanol, 1-propanol and acetic acid takes place through a polycrystalline silicalite membrane [ 16-18]. In this paper, we carried out further investigation on the separation of acetic acid-water mixture through the silicalite membrane on a porous support.

2. Experimental Silicalite and ZSM-5 (SiO2/AI203 ratio of 300) membranes on porous supports were prepared according to a procedure described previously [16-18]. Sintered stainless steel and alumina disc (5 cm diameter) with an average pore diameter of ca. 0.5-2 Izm were used as a porous support. The zeolite membrane crystallized on the support was washed with deionized water and dried at 100°C. After that, the membrane was calcined at 400°C for 20 h in order to decompose the organic amine occluded in the zeolite framework. The membrane was not disintegrated by this process. Powdery silicalite and ZSM-5 zeolite crystals were also prepared in a similar manner. The identification of the polycrystalline silicalite and ZSM-5 membranes was achieved by X-ray diffraction (Rigaku Rint 2100). The surface and cross section of the membrane was characterized by scanning electron microscopy (Hitachi S-4100) with energy-dispersive X-ray analysis system (EDX). The chemical compositions of ZSM-5 zeolite was measured by X-ray fluorescence spectroscopy (Philips PW 2400). Diffuse reflectance infrared (FT-IR) spectra were measured at room temperature using a JEOL JIR7000 spectrometer equipped with an evacuable heatable chamber. The spectra were taken at 4 cm -l resolution for 500 scans. The powdered zeolite was placed in a thin-walled ampule and then evacuated to ca. 10 -5 Torr at 400°C for 2 h. The adsorption of acetic acid or ethanol was carried out by equilibrating silicalite or ZSM-5 (ca. 1 g) with an aqueous solution of acetic acid or ethanol (2.5-10 vol%, 10 ml) at 30°C. The samples were left overnight to ensure equilibrium. The amounts of

acetic acid or ethanol in the solutions initially and at equilibrium were measured by gas chromatography. The depletion of acetic acid or ethanol in solution was assumed to be due only to adsorption by zeolite. Nitrogen adsorption isotherms at - 1 9 6 ° C were measured using a conventional volumetric apparatus (Bell Japan BELSORP 28SA). Prior to adsorption measurements, the powdered zeolite (ca. 0.1 g) was evacuated at 400°C for 12 h. The pervaporation measurements using an aqueous acetic acid or ethanol solution as a feed were performed using a standard pervaporation cell [17]. The down stream pressure was maintained below 2 Torr. Liquid nitrogen was used as a cooling agent for the cold trap. The compositions of the feed and the permeate were determined by gas chromatography. The pervaporation performance was evaluated by the flux and the acetic acid (ethanol) concentration of the permeate.

3. Results and discussion 3.1. Pervaporation performance of silicalite membrane prepared on stainless steel support We have already reported that even if silicalite membranes are prepared on the stainless steel support with the same method, a slight difference in the pervaporation performance is observed [17]. Therefore, the pervaporation performance of the silicalite membrane used in this study was at first checked using an aqueous ethanol solution of 5 vol% at 30°C. As shown in Table 1, the silicalite membrane on the stainless steel support exhibited a high separation factor ot(EtOH/H20) of 59 combined with a flux of 0.22 k g / m 2 h. Fig. 1 shows the effect of the acetic acid concentration of the feed on the total flux and the acetic acid concentration of the permeate for the silicalite membrane. The total flux decreases rapidly by adding acetic acid to water and then becomes constant. The acetic acid concentrations of the permeates are higher than those of the feeds. Taking into account that the acetic acid concentration of the permeate was higher than that calculated from the vapor-liquid equilibrium data, it was indicated that the silicalite membrane shows the acetic acid permselectivity in the

227

T. Sano et a l . / Journal of Membrane Science 123 (1997) 225-233

Table 1 Pervaporation performance of silicalite membranes prepared on stainless steel and alumina supports Membrane

Ethanol/water

Silicalite/stainless steel support Silicalite (ZSM-5)/alumina support

Acetic acid/water

Flux (kg/m 2 h)

Ethanol conc. of permeate (vol%)

a(EtOH/ H20)

Flux (kg/m 2 h)

Acetic acid conc. of permeate (vol%)

a(CHsCOOH/ H20)

0.22 0.19

74 20

59 4.2

0.038 0.94

33 15

2.6 1

Feed ethanol concentration: ca. 5 vol%. Feed acetic acid concentration: ca. 15 vol%. Feed temperature: 30°C.

concentration of the feed acetic acid in the region of 5 to 40 vol%. So far, there is no report concerning to inorganic membranes which permeates acetic acid preferentially. As shown in Table 1 and Fig. 1, however, acetic acid in the permeate was not highly concentrated as compared to the separation of the e t h a n o l - w a t e r mixture. It is well known that the transport mechanism on pervaporation through polymeric membranes can be described by a solution-diffusion model [19]. The membrane performance depends upon (1) sorption of permeants in the feed side and (2) their diffusion through the membrane. As the molecular sizes of acetic acid (ca. 0.4 nm), ethanol (ca. 0.4 nm) and water (ca. 0.28 nm) are below the intrinsic pore size of silicalite (ca. 0.6 nm), the sorption process seems to play an important role in pervaporation. To clarify

60

0.7 0.6 ( ~" o.5 E 20.4 .--- 0.3 ~0.2 o.1

/l: Z

/

& 50

E 40 ~.

the adsorption property of the silicalite membrane, the adsorption experiments was carried out using aqueous acetic acid and ethanol solutions. The result obtained are summarized in Table 2. The amount of acetic acid adsorbed was larger than that of ethanol, suggesting the high separation potential of the silicalite membrane. The silicalite membrane prepared is polycrystalline and the small pores with ca. 1 nm diameter originated from silicalite grains other than that intrinsic zeolitic pores of ca. 0.6 nm exist within the membrane [20]. Therefore, it has been considered that the separation of acetic a c i d - w a t e r mixtures takes place mainly through the small pores of ca. 1 nm diameter. As shown in Table 2, the intrinsic zeolitic pores show the high adsorption ability of o, acetic acid due to the high hydrophobic property of silicalite crystals. The carboxyl groups of acetic acid molecules adsorbed into the intrinsic zeolitic pores seems to change the surface property of the pores from silicalite grains to slightly hydrophilic. Consequently, the high concentration of acetic acid through the membrane, as expected from the adsorption data, seems not to be achieved.

~J

2o ~

§

10o

< lO 20 30 40 50 60 Acetic acid concentration of feed (vol%)

Fig. 1. Effect of acetic acid concentration of feed on pervaporation performance of silicalite membrane prepared on stainless steel support. Feed temperature: 30°C.

Table 2 Adsorption of acetic acid and ethanol from aqueous solutions by silicalite membrane detached from stainless steel support at 30°C Initial concentration Adsorption amounts of acetic acid or Acetic acid Ethanol ethanol (vol%) (mg(mmol)/g) (mg(mmol)/g) 2.5 5.0 10.0

145 (2.42) 164 (2.73) 186 (3.10)

70 (1.52) 89 (1.93) 87 (1.89)

228

T. Sano et al. / Journal of Membrane Science 123 (1997) 225-233 40

0.16

2" 0.12 E

~810/Of

a0 "6

=

20

0.08

E

- I&/

o

E 0.04 o 0

/

u/

ooo.

c

10 § "5 t~ ._o

0 20

30 40 Temperature

50

0 60

<

(~3)

Fig. 2. Component flux and acetic acid concentration of the permeate as a function of the feed temperature. Feed acetic acid concentration: 15 vol%. Fig. 2 shows an effect of the feed temperature on the pervaporation performance of the membrane. The pervaporation measurements were conducted at 30°C using the aqueous acetic acid solution of 15 vol% as a feed. The component flux increased linearly with the feed temperature, while the acetic acid concentration of the permeate was hardly dependent on the temperature. This indicates that the adsorption of acetic acid onto the pore surface from silicalite grains is not affected by the feed temperature. Table 3 shows the apparent activation energies of the pervaporation rates of acetic acid and water calculated from the data in Fig. 2. The apparent activation energies of ethanol and water are also presented. The apparent activation energy of the pervaporation rate of water in aqueous solutions of acetic acid and ethanol are considerably higher than that of pure water, indicating that the transport of water was strongly suppressed in the case of the pervaporation of acetic acid-water mixtures. Although the exact reason why water permeation is suppressed, it is not clarified at present due to limited data, the preferential adsorption of ethanol or acetic acid molecules onto the pore surface seems to hinder water permeation. 3.2. Pervaporation performance o f silicalite membrane prepared on alumina support

Next, the separation performance of the silicalite membrane prepared on the porous alumina support

was investigated. As shown in Table 1, the separation of acetic acid-water mixture did not take place through the membrane although the separation of ethanol-water mixture OCCUlTed. It is well known that the hydrophobicity of ZSM-5 zeolite decreases when the A1 content in the zeolite framework increases [21,22]. The synthesis mixture prepared for silicalite membrane does not contain the A1 source. However, it is reasonable to consider that a part of the surface of alumina support is dissolved due to the high pH of the synthesis mixture. The surface composition of the membrane was measured by EDX. From the fact that the surface SiO2/A1203 ratio was ca. 300-400, it was indicated that the membrane obtained was not composed of pure silicalite but ZSM-5 zeolite crystals. As the membrane is soaked in the aqueous acetic acid solution during the pervaporation experiment, it seems that the acidic bridged hydroxyl groups of Si(OH)A1 associated with framework aluminums are formed within the membrane. To confirm this, the as-synthesized ZSM-5 zeolite with a SiO2/A1203 ratio of 200 was treated with an acetic acid solution of 15 vol% at 30°C for 72 h. Fig. 3 shows the F T - I R spectra of hydroxyl groups of the ZSM-5 zeolites before and after the treatment. In the IR spectrum of the H-ZSM-5 zeolite treated with the aqueous acetic acid solution, the peak assigned to the acidic bridged hydroxyl groups of Si(OH)A1, BrCnsted sites, was observed at 3605 cm -1. The peak at 3740 c m - J is assigned to terminal silanol groups located on the external and internal surfaces of the silicalite crystals. The broad peak at near 3500 cm -~ is attributable to hydrogen bonding adjacent hydroxyl groups. The peak intensity at 3605 cm-1 was almost the same as that in the spectrum of ZSM-5 zeolite treated with 0.6 N HC1 solution, a conventional protonation condition of ZSM-5 zeolite. From the Table 3 Apparent activation energies of pervaporation rates of acetic acid, ethanol and water Mixture AE (kcal/mol) Acetic acid Acetic acid/water Ethanol/water Water

Ethanol

Water

8.7

12.9 10.0 6.3

13.8

T. Sano et a l . / Journal of Membrane Science 123 (1997) 225-233

229

through the alumina supported silicalite (actually ZSM-5 containing BrCnsted acidic sites) membrane seems to attribute to the enhancement of adsorption of acetic acid molecules into the intrinsic zeolitic pores by acidic sites presented, namely an increase in the hydrophilicity of the pore surface originated from zeolite grains. 3.3. Long-term test on stainless steel supported silicalite membrane

4000

3800 3600 3400 3200 Wavenumbers (cm -1 )

3000

Fig. 3. F F - I R spectra of ZSM-5 zeolites before and after treatment with aqueous acidic acid solution. (A) As-synthesized ZSM-5 zeolite, (B) treated with aqueous acidic acid solution of 15 vol% at 30°C for 72 h, (C) treated with 0.6 N HC1.

fact that the Na/A1 ratios of ZSM-5 zeolites before and after the treatment were 0.88 and 0.06, respectively, the formation of acidic side by H - ion exchange was also indicated. The influence of acidic sites of zeolite on the adsorption behavior was investigated using the protonated ZSM-5 zeolite membrane with a SIO2/A1203 of 300 detached from the support. The adsorption of ethanol was also carried out as a reference. The result obtained was listed in Table 4. The amount of acetic acid adsorbed on the protonated ZSM-5 zeolite membrane was considerably larger than that on the silicalite membrane as shown in Table 2. No difference in the amount of ethanol adsorbed was observed between protonated ZSM-5 and silicalite membranes. The exact reason why the adsorption of acetic acid is enhanced by the presence of acidic sites is not clear at present. However, no separation Table 4 Adsorption of acetic acid and ethanol from aqueous solutions by protonated ZSM-5 (SiO 2 / A I 2 0 3 ratio of 300) membrane at 30°C Initial concentration

To investigate the stability of the separation performance of the membrane, a long-term test of the pervaporation was carried out for one month using the fresh silicalite membrane prepared on the stainless steel support. Although the acetic acid permselectivity of the membrane was not so high, the acetic acid concentration of the permeate was kept nearly constant (ca. 25%) (Fig. 4A). However, the acetic acid concentration after breaking the pervaporation experiment fell to below 20% (C). In the case of the separation of ethanol-water mixtures, the decrease in the ethanol permselectivity, c~(EtOH/H20) of 57, was not observed. To clarify the reason for the considerable decrease in the acetic acid permselectivity, the silicalite membrane was treated with a pure ethanol as ethanol molecules selectively permeate through the membrane. The permselectivity of the membrane was completely restored by repeating the treatment with pure ethanol (D). This indicates that the considerable decrease in the acetic acid permselectivity cannot attribute to the structural destruction of the membrane. (A) 3O

I

25 n~ o~e°_.~°

~

20

oe

15

oe

I

10

300000

Acetic acid (mg(mmol)/g)

Ethanol (mg(mmol)/g)

2.5 5.0 10.0

180 (3.00) 202 (3.37) 203 (3.38)

70 (1.52) 93 (2.02) 92 (2.00)

(C)

.

Break

Adsorption amounts

of acetic acid or ethanol (vol%)

- (B)

~'~ 0

0

P

"

" ~0 0.2 04~"

I I lOgO I io



° o

o

OO00011

L

I

5

10

I I

'

(D)

Ethanol treatment I~1 !0.25

I

I oil I n I i I II

ooe

=

'

15 20 Time (day)

I.el I" I O01 I I i I I o I I Olool o I IO

i

i

I

I

'!

!

25

0.15 0.1

v zx ~_ "~

0.05 ~_ 0

30

Fig. 4. Long-term test of pervaporation performance of silicalite membrane prepared on stainless steel support. Feed acetic acid concentration: 15 vol%. Feed temperature: 30°C.

T. Sano et al. / Journal of Membrane Science 123 (1997) 225-233

230

Next, to elucidate the influence of the preadsorption of water or acetic acid on the pervaporation performance of the silicalite membrane, the membrane was soaked in water or an aqueous acetic acid solution of 15 vol% before the pervaporation experiment. The decrease in the acetic acid permselectivity was not observed for the membrane soaked in water, while the considerable decrease in the permselectivity was observed for the membrane soaked in the aqueous acetic acid solution. This seems to indicate that adsorption of acetic acid molecules into intrinsic zeolitic por~s of silicalite crystals causes a decrease in the permselectivity of the membrane. Based on the above results, we have considered the considerable decrease in the pervaporation performance after break and the restoration of the permselectivity by the ethanol treatment as follows (Fig. 5). As acetic acid molecules selectively adsorb onto the pore surface originated from silicalite grains, the silicalite membrane shows high acetic acid permselectivity (Fig. 5A). However, when the pervaporation experiment is broken, acetic acid molecules enter into intrinsic zeolitic pores of silicalite crystals (Fig. 5B). The carboxyl groups of acetic acid molecules adsorbed into intrinsic zeolitic pores change the surface property of the pores originated from silicalite grains to more hydrophilic, resulting the decrease in the acetic acid permselectivity (Fig. 5C). When the membrane is treated with pure ethanol, the acetic acid molecules in pores are removed by the treatment. Therefore, the permselectivity of the membrane is restored (Fig. 5D). In the case of the membrane soaked in the feed

Acetic acid

(A)

Water

OCOOH

",,b/ • obod'

i;i;i:F:;:i~i::i:~il}!i:0i;:i 0

:~~: }i!:i;:!.......... i:!;:!~ OH0 , 0

Ooo

HO

(B) o O o

0

i!ii!~!i!i!ii!i:~!~ili~i~0i • ~::!i:!:!!:!:!i!i¢]::i ~!!i; , . ,~ .H .O ...... H O O COOH~• 0 : E[(YOD .+.... :il;i;i:i::i:i!ii:E:ii:ii:.Ei

....;!i}::qi#.::i::ii::ii;::ii~i

---r~oo.

0 .,,os-O-'--

........... o.'6o.ol#iiiiii@i,,i i

aqueous acetic acid solution for more than one month, however, the pervaporation performance of the silicalite membrane was not restored even by repetition of the ethanol treatment. To investigate the potential of the structural destruction of the membrane, the membrane was soaked in an aqueous acetic acid solution of 50 vol% at 30°C for 200 h. The micropore volumes of the silicalite membrane before and after the treatment was measured by the DubininRadushkevich ( D - R ) analysis of adsorption isotherms of nitrogen at - 196°C [23]. Fig. 6 shows the adsorption isotherms of nitrogen and their D - R plots on the silicalite membrane before and after the treatment. The micropore volumes W0(N2) before and after the treatment obtained from the intercept on the y-axis of the linear plots are 0.202 and 0.200 ml(liquid)/g, respectively. This indicates that there is no structural destruction of the membrane. Table 5 shows the amount of acetic acid adsorbed on the membrane after the acetic acid treatment. A comparison of the value with that in Table 2 indicates clearly an increase in the amount of acetic acid adsorbed. As the stainless steel supported silicalite membrane did not contain the acidic bridged hydroxyl groups of Si(OH)A1, this indicates formation of new adsorption sites of acetic acid. Fig. 7 shows the F T - I R spectra of hydroxyl groups of the silicalite membranes before and after the acetic acid treatment. In the spectrum of the membrane after the treatment, the small peak near 3690 c m - ' as well as two well-defined peaks at 3740 and near 3500 c m - ] are observed. The peak intensity at 3500 cm ' was slightly increased by the treatment. The peak at 3690

/0/° C o~o 0 .o

:O;:o

oH • U

HO

Ethanol

(D) O © 0 © © ~60i:i::iiiiii:#i oH ....::i@ :i::::i::#} Ho ~:::~i~i~{~:~i~:~i~i~:i~:~}: o.@_0.o 0II q

1t0

Oo UO~

Fig. 5. Plausible mechanism of restoration of permselectivity by ethanol treatment. A, B, C and D correspond to Fig. 4.

T. Sano et al. / Journal of Membrane Science 123 (1997) 225-233 200

~

~

231

(B)

(A)

.~-0.5,

150

~O

~ tWiii~Ea~ qi



100 q

0

~-1.5

O4

z

0

I

0

0.2

I

I

0.4 0.6 P/Ps

I

0.8

"2

I

0

0.2

I

I

I

0.4 0.6 0.8 (log (Ps/P)) 2

Fig. 6. Adsorptionisotherms of nitrogenat - 196°C (A) on silicalitemembranesdetached from stainlesssteel support before (©) and after ( 0 ) treatmentwith 50 vol% acetic acid solution and their D-R plots (B).

c m - 1 may be produced by the hydrolysis of S i - O - S i bonds of the framework, indicating a partial structural destruction of silicalite crystals. Taking into account the fact that the protonated ZSM-5 zeolite membrane containing the acidic bridged hydroxyl groups gives an increase in the amount of acetic acid adsorbed, the hydroxyl groups produced by the hydrolysis of S i - O - S i bonds seems to enhance the adsorption of acetic acid, changing the pores from silicalite grains to more hydrophilic.

Table 6 Concentrationof vinegar using silicalitemembrane

3.4. Condensation o f vinegar and ethanol produced by fermentation

Feed temperature: 30°C. nd: not detect.

Component

Feed

Permeate

Acetic acid (vol%) Alcohol (vol%) Pyroglutamicacid (mg/100 ml) Lactic acid (rag/100 ml) Pyruvic acid (mg/100 ml) Citric acid (rag/100 ml) Succinic acid (mg/100 ml) ct-Ketoglutaricacid (mg/100 ml) Malic acid (rag/100 ml)

14.0 0.32 0.68 0.13 0.12 4.05 3.99 0.20 0.02

35.3 0.49 nd 1.16 nd nd nd nd nd

As the pervaporation process is much more attractive for the extraction of a minor component in a solvent, the silicalite membrane seems to be applicable to the condensation process in the production of vinegar and alcohol produced by fermentation of biomasses. The transport property of the membrane was tested against vinegar and ethanol containing small amounts of components such as alcohol, carTable 5 Amounts of acetic acid adsorbed on silicalite membrane treated with 50 vol% acetic acid solution at 30°C for 120 h Initial concentration of acetic acid (vol%)

Amountof acetic acid adsorbed (mg(mmol)/g)

2.5 5.0 10.0

198 (3.30) 199 (3.32) 206 (3.43)

Adsorption temperature: 30°C.

O tO <

4ooo' 3doo' 3e'oo ' J o o ' 32'oo ' 3ooo Wavenumbers (cm-1) Fig. 7. FT-IR spectra of silicalitemembranesbefore (A) and after (B) treatment with 50 vol% acetic acid solutionat 30°C for 120 h.

232

T. Sano et a l . / Journal of Membrane Science 123 (1997) 225-233

Table 7 Concentrationof fermentationethanol using silicalitemembrane Component

Feed

Permeate

Ethanol (vol%) Citrulline(ppm) Proline (ppm) Glutamic acid (ppm) Alanine(ppm) Leucine (ppm) Glycine(ppm) Phenylanaline(ppm) Asparagine(ppm) Tyrosine (ppm) Serine (ppm)

13.9 6.2 5.2 5.0 4.4 3.6 2.8 2.0 1.8 1.8 1.6

90.4 nd nd nd 6.7 nd nd nd nd nd nd

suggested that the acidic sites associated with the framework aluminums play an important role in the separation of acetic acid-water mixture. The partial structural destruction of the silicalite crystals originated from the hydrolysis of S i - O - S i bond in the framework by acetic acid was also suggested.

Acknowledgements The authors gratefully acknowledge Nakano Central Research Institute for the supply of vinegar and ethanol produced by fermentation.

Feed temperature: 30°C. nd: not detect.

References boxylic acids and amino acids, which were supplied by Nakano Central Research Institute. The representative results are presented in Tables 6 and 7. The permeate rate of vinegar was 0.020 k g / m 2 h. The solution was concentrated to over twice the original concentration. Except for alcohol and lactic acid, the organic acids were hardy detected in the concentrated solution. On the other hand, the permeate rate of fermentation ethanol was 0.27 k g / m 2 h. The ethanol concentration of the permeate was more than 90 vol%. Except for alanine, however, the other amino acids were hardy detected in the concentrated solution. No permeation of carboxylic acids and amino acids seems to be attributable to their high boiling temperatures as the driving force for the permeation is the difference in partial pressure applied across the membrane.

4. Conclusions It was concluded that although the silicalite membrane prepared on the sintered stainless steel support is polycrystalline, the membrane permeates acetic acid in preference to water from an aqueous acetic acid solution. However, when the porous alumina disc was used as the support, the separation of acetic acid from aqueous acetic acid solution did not take place. From the fact that the top layer of the membrane is not composed of pure silicalite crystals but ZSM-5 zeolite crystals, which contain Br~nsted acidic sites (Si(OH)A1) in the framework, it was

[1] M. Yoshikawa, T. Yukoshi, K. Sanui and N. Ogata, Separation of water-acetic acid mixture by pervaporation through synthetic polymer membranescontainingpendant carboxylic acid, Maku (Membrane), 10 (1985) 247; T.Q. Nguyen, A. Essamri, R. Clement and J. Neel, Synthesis of membranes for the dehydration of water-acetic acid mixtures by pervaporation, 1. Polymer material selection, Makromol. Chem., 188 (1987) 1973; R.Y.M. Huang, A. Moreira, R. Notarfonzo and Y.F. Xu, Pervaporation separation of acetic acid-water mixtures using modified membranes. I. Blended polyacrylic acid (PAA)-Nylon 6 membranes, J. Appl. Polym. Sci., 35 (1988) 1191; G.H. Koops, J.A.M. Nolten, M.H.V. Mulder and C.A. Smolders, Poly(vinyl chloride)polyacrylonitrile composite membranes for the dehydration of acetic acid, J. Membrane Sci., 81 (1993) 57. [2] J. Bai, A.E. Fouda, T. Matsuura and J.D. Hazlett, A study on the preparation and performance of polydimethylsiloxanecoated polyetherimidemembranesin pervaporation,J. Appl. Polym. Sci., 48 (1993) 999. [3] M. Yoshikawa, S. Kuno, T. Wano and T. Kitano, Specialty polymeric membranes. 4. Pervaporation separation of acetic acid/water mixtures through modified polybutadienemembranes, Polym. Bull., 31 (1993) 607. [4] S. Deng, S. Sourirajan and T. Matsuura, A study of polydimethylsiloxane/aromaticpolyamide laminatedmembranes for separation of acetic acid/water mixtures by pervaporation process, Sep. Sci. Technol., 29 (1994) 1209. [5] J.G. Tsikoyiannisand W.O. Haag, Synthesis and characterization of a pure zeolitic membrane,Zeolites, 12 (1992) 126. [6] E.R. Geus, H. van Bekkum, W.J.W. Bakker and A. Moulijn, High-temperature stainless steel supported zeolite (MFI) membranes: preparation, module construction, and permeation experiments, MicroporousMater., 1 (1993)131. [7] M.D. Jia, K.V. Peinemannand R.D. Behling, Ceramic zeolite composite membranes. Preparation, characterization and gas permeation,J. MembraneSci., 82 (1993) 15. [8] P. Kolsch, D. Venzke, M. Noack, P. Toussaint and J. Caro,

T. Sano et a l . / Journal of Membrane Science 123 (1997) 225-233

[9]

[10]

[11] [12]

[13]

[14]

[15]

[16]

Zeolite-in-metal membranes: Preparing and testing, J. Chem. Soc., Chem. Commun., (1994) 2491. Y. Yan, M. Tsapatsis, G.R. Gavals and M.E. Davis, Zeolite ZSM-5 membranes grown on porous ~x-Al20 3, J. Chem. Soc., Chem. Commun., (1995) 227. H. Kita, K. Horii, Y. Ohtoshi, K. Tanaka and K. Okamoto, Synthesis of a zeolite NaA membrane for pervaporation of water/organic liquid mixtures, J. Mater. Sci. Lett., 14 (1995) 206. R.D. Noble and J.L. Falconer, Silicalite-1 zeolite composite membranes, Catal. Today, 25 (1995) 209. N. Nishiyama, K. Ueyama and M. Matsukata, A defect-free mordenite membrane synthesized by vapor-phase transport method, J. Chem. Soc., Chem. Commun., (1995) 1967. T. Sano, K. Kiyozumi, M. Kawamura, F. Mizukami, H. Takaya, T. Mouri, W. Inaoka, Y. Toida, M. Watabnabe and K. Toyoda, Preparation and characterization of ZSM-5 zeolite film, Zeolites, 11 (1991) 842. T. Sano, Y. Kiyozumi, K. Maeda, M. Toba, S. Niwa and F. Mizukami, New preparation method for highly siliceous zeolite films, J. Mater, Chem., 2 (1992) 141. T. Sano, Y. Kiyozumi, K. Maeda, M. Toba, S. Niwa and F. Mizukami, Synthesis and characterization of polycrystalline SAPO-5 film, J. Mol. Catal., 77 (1992) L19. T. Sano, H. Yanagishita, Y. Kiyozumi, D. Kitamoto and F.

[17]

[18]

[19]

[20]

[21]

[22]

[23]

233

Mizukami, Separation of ethanol/water mixture by silicalite membrane, Chem. Lett., (1992) 2413. T. Sano, H. Yanagishita, Y. Kiyozumi, F. Mizukami and K. Haraya, Separation of ethanol/water mixture by silicalite membrane on pervaporation, J. Membrane Sci., 95 (1994) 221. T. Sano, S. Ejiri, H. Hasegawa, Y. Kawakami, N. Enomoto, Y. Tamai and H. Yanagishita, Silicalite membrane for separation of acetic acid/water mixture, Chem. Lett., (1995) 153. M.H.V. Mulder and C.A. Smolders, On the mechanism of separation of ethanol/water mixtures by pervaporation. I. Calculation of concentration profiles, J. Membrane Sci., 17 (1984) 289. T. Sano, M. Hasegawa, Y. Kawakami, Y. Kiyozumi, H. Yanagishita, D. Kitamoto and F. Mizukami, Potentials of silicalite membranes for the separation of alcohol/water mixtures, Stud. Surf. Sci. Catal., 84 (1994) 1175. R. Le Van Mao, O. Pilati, A. Marzi, G. Leofanti, A. Villa and V. Ragaini, Adsorption characteristics of some zeolites, React. Kinet. Catal. Lett., 15 (1980) 293. D.H. Olson, W.O. Haag and R.M. Lago, Chemical and physical properties of the ZSM-5 substitutional series, J. Catal., 61 (1980) 390. S.J. Gergg and K.S.W. Sing, Adsorption, Surface Area and Porosity, Academic Press, New York, 1982.