Dehydration of Alcohol-Water Mixtures Through Composite

required may have limited the application of reverse osmosis membrane technology (2^) . ... Before the membranes performance was measured, the prelimi...
0 downloads 0 Views 942KB Size
35 Dehydration of Alcohol-Water Mixtures Through Composite Membranes by Pervaporation Downloaded by STANFORD UNIV GREEN LIBR on July 1, 2012 | http://pubs.acs.org Publication Date: January 1, 1985 | doi: 10.1021/bk-1985-0281.ch035

YUTAKA TAKETANI and HIROYOSHI MINEMATSU Central Research Laboratories, Teijin Ltd., 4-3-2 Asahigaoka, Hino City, Tokyo 191, Japan

Three types of composite membranes from polyethyleneimine (PEI) with isophthaloyl chloride (IPC), trimesoyl chloride (TMC) and 5-chlorosulfonyl isophthaloyl chloride (CSIPC) were studied for dehydration of aqueous azeotropes of ethanol or isopropanol by pervaporation. The water selectivities of the composite membranes from PEI and CSIPC were the best of the three, and were 365 and 2069 for aqueous azeotropes of ethanol and isopropanol, respectively. The water selectivities decreased in the order of CSIPC≥TMC>IPC. The membranes from the PEIs modified with monochloroacetic acid or polyacrylic acid, and crosslinked with IPC improved the water selectivity at a cost of decreased flux rate. The performance of the composite membranes was analyzed and discussed from a chemical structural view point. The reverse osmosis membrane process is well established both in the science and technology for the desalination of sea water and brackish water. The wide applications of the reverse osmosis membrane for the various processes are well studied (1) . The high osmotic pressure required may have limited the application of reverse osmosis membrane technology (2^) . To avoid this, the pervaporation process is thought to be one of the alternatives which separates the liquid mixtures through membranes, where there are some variations in the ways of reducing the permeate pressure (£—6J. The pervaporation is one membrane process that may provide an economical alternative to the distillation process. To be so, the membranes should have excellent selectivities and flux rates. Several works were done for the separation of an ethanol and water mixture by the pervaporation with various membranes (7^9_) . The recent work showed the results of an application of reverse osmosis composite membranes to the pervaporation process (10). In case of alcohol azeotropes such as an ethanol-water and isopropanol-water mixtures, the water is the minor component and, therefore, it would be more practical to remove

0097-6156/85/0281-0479$06.00/0 © 1985 American Chemical Society

In Reverse Osmosis and Ultrafiltration; Sourirajan, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1985.

Downloaded by STANFORD UNIV GREEN LIBR on July 1, 2012 | http://pubs.acs.org Publication Date: January 1, 1985 | doi: 10.1021/bk-1985-0281.ch035

480

REVERSE OSMOSIS AND ULTRAFILTRATION

water than to extract alcohols from the mixture. For this purpose, pervaporation membranes should have high water selectivities. In order to prepare membranes of this type, one may apply the concept of preferentially absorbed water proposed by Sourirajan for the reverse osmosis membranes (1). If a membrane has the preferentially absorbed water layer at the surface of the membrane, then this layer of water might repel organic components, resulting in higher selectivity of water even in the pervaporation process. In this report, three types of composite membranes were prepared by dipping microporous polysulfone membranes into a polyethyleneimine (PEI) solution, and then crosslinking the PEI with three different aromatic acid halides. The pervaporation performance of the three membranes was examined in the dehydration of ethanol or isopropanol azeotropic mixtures. The performance characteristics of the membranes were discussed with regard to their chemical structures. Experimental Pervaporation. The pervaporation apparatus consists of a constant temperature bath and a pump that circulates the feed solution through a radial flow cell (membrane surface area 11.1 cm ) at a rate of about 1.2 1/min. and with the bath temperature controlled to ±0.2°C. The downstream compartment consists of two parallel pumping station that allows alternate sampling from the cold traps of ethanol and dry ice. The membrane was supported by a sintered metal plate. The composite membrane was put into the cell so that the thin layer of the composite membrane faced the feed solution. Before the membranes performance was measured, the preliminary sampling was made in the first two or three hours. The feed and permeate concentrations were determined by gas chromatography using Hitachi model 163 with stainless steel Porapak Q columns (oven temperature of 170°C and 190°C for ethanol and isopropanol, respectively, under a He stream of 20 ml/min.) equipped with Shimadzu CR-1 recorder. The concentration was determined by the calibration curves prepared with use of standard solutions. The flux was determined from the amount of material captured in the trap, and selectivity was calculated by the following equation.

where P and F denote the permeate and the feed weight fractions, respectively. Materials. PEI was supplied in a 30% aqueous solution, manufactured by Nippon Shokubai Chemicals. Isophthaloyl chloride (IPC) and trimesoyl chloride (TMC) were of reagent grade from Wako Chemicals. 5-Chlorosulfonyl-isophthaloyl chloride (CSIPC) was prepared from the sodium 3,5-dicarboxy benzene sulfonate by the reaction with thionyl chloride and a small amount of DMF. CSIPC was recrystallized from CCli+ and CHCI3. All other materials were of reagent grade, and used without further purification. Modification of PEI. Aqueous PEI solution was prepared by diluting 30% PEI solution (20 g) with 18.68 g of water. Monochloro-

In Reverse Osmosis and Ultrafiltration; Sourirajan, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1985.

35.

TAKETANI AND MINEM ATSU

Dehydration of Alcohol- Water Mixtures

481

Downloaded by STANFORD UNIV GREEN LIBR on July 1, 2012 | http://pubs.acs.org Publication Date: January 1, 1985 | doi: 10.1021/bk-1985-0281.ch035

acetic acid (1.32 g) was added to the solution and the mixture was stirred for 3 h at room temperature. The solution thus obtained was diluted to 2% with water, and was used for the composite membranes preparations. Composite membranes. The composite membranes were prepared by the procedure described elsewhere (ll^r12) . Throughout the experiment, polysulfone microporous membranes (pure water permeability 3.0—7.0xl05 kg/m 2 s kPa) were used as the support for the composite membranes. In order to isolate the active layer of the membrane, the composite membranes thus formed were immersed in CHCI3 for 15—20 min. The polysulfone microporous membranes were dissolved, and crosslinked thin layers were floated off. An individual piece of the thin film was scooped with a preweighed filter paper, and was transferred to another CHCI3 bath to rinse off the polysulfone completely. After this procedure, the membranes with filter papers were dried and weighed. The weight was approximately between 2—6 mg for a 5 cm diameter disk sample. These samples were divided into two parts and were used for chemical analyses. Titration of composite membranes. The weighed thin layer films were immersed in deionized water overnight with stirring. The rinsed and filtered polymers were immersed again in deionized water. This procedure was repeated three times. The rinsed and dried films were then immersed in a 1/50 N NaOH aqueous solution and the mixture was stirred overnight. The titration of an aliquot of 10 ml of the filtrate with AgNC>3 aqueous solution gave the Cl content of B meq/g dried polymer, which would be equal to the amount of the salt of amines hydrochloride. The remaining filtrate was then titrated for the NaOH consumption measurement with a 1/100 N HC1 aqueous solution to give A meq/g polymer. The value A meq/g would determine the sum of all acidic components such as -C00H, -SO3H and amines hydrochloride. The other run was conducted using the second portion of the weighed thin films. Instead of a 1/50 N NaOH solution, a 1/50 N HC1 aqueous solution was used for the immersion, and the value of C meq/g polymer, was obtained by the measurement of the consumed amount of HC1 with a 1/100 N NaOH aqueous solution. The C meq/g polymer was thought to be equivalent to the amount of free amine in the composite membranes. The blank experiment was also done by using the filter papers. Instruments. The Nippon Denshi JIR-40X and JESCA-4 were used for the FT-IR and XPA measurements. Results and Discussion Pervaporation performance of composite membranes. The pervaporation experiments were done for the separation of azeotropic mixtures of two alcohols, that is, ethanol and isopropanol, in which water is the minor component, containing 5 wt% and 12 wt%, respectively. Three types of crosslinking reagents, IPC, TMC and CSIPC, were used to obtain in situ composite membranes on polysulfone microporous supports. The performances features are shown in Table I. It was reported that a high molecular weight PEI itself affords the insoluble thin layer, which showed rather high salt rejection

In Reverse Osmosis and Ultrafiltration; Sourirajan, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1985.

482

REVERSE OSMOSIS AND ULTRAFILTRATION

O

ffi

K

«

CM O

8

1

^ • iH H

CTi



in rH

o

o





CN

VD

O

^r

sr

m



CNJ

CO

u

,

0

o

CD ID •H

^ - N

vD

^

XCM

3 g\

H

VD rH 1

fa CP

»

in

in 00



"Nf

ro ro

ro ro





rH

ro

U 0

H

00 vD

ro

ro

ro

H



^ O

J*

o o o o r^

co CD

§ u

qui

0

rH rd

O

U o O CO

CD

-P •H CO

CT> VD CN

r-

O

•P

ro

•a 0

vD

r-

CO

H >i

rC

6

VD

^

in

fa &>

CN rH

£

H

W

^ rH

VD O rH

r* ro

CT> H

o

H

r-i

^r

ro ro

rH

rH

C

0

4H rH

3 CO

0 U 0

PH

6 O ffi CNJ O

1

8

CM

ro

r-

vO rH

0^ rH

CN

r^

6

00 rH

U

o O VD

1 in

3.0

&

s

6.4

H

0

7.3

m

PH

PH

H W

U

£

X CNJ

c

0

3

•H

g

•H

+J