GB2324257A

GB2324257A - Pervaporation of ethanol/water mixtures

Info

Publication number: GB2324257A
Application number: GB9807820A
Authority: GB
Inventors: Virender Nath Malhotra; Parish Bhasi
Original assignee: EAST LONDON, University of
Current assignee: EAST LONDON, University of
Priority date: 1997-04-14
Filing date: 1998-04-14
Publication date: 1998-10-21
Also published as: GB9707501D0; GB9807820D0

Abstract

Ethanol/water mixtures of at least 80% ethanol content are subjected to pervaporation using a water-selective vulcanised natural rubber latex membrane 6 to provide a retentate enriched in ethanol. The membrane may be used in a jacketed cell 20 containing a steel bellows 5 connected to a source of high-pressure nitrogen and to which the mixture is supplied via conduits 11,12. A vacuum is drawn at the outlet 13 and a heating medium to the jacket 23. The membrane is supported by a steel plate 8.

Description

PERVAPORATION OF ETHANOL/WATER MIXTURES This invention relates to the pervaporation of mixtures of ethanol and water to remove water therefrom.

Pervaporation is a known technique for separating or concentrating one liquid component present in a mixture thereof with at least one other miscible liquid. In this technique, the mixture is placed in contact with a liquid impervious membrane, and the vapour phase permeate (which will be richer than the mixture in one of the components) is condensed and collected on the other side of the membrane, where a vacuum or inert gas flow is provided.

In recent years, pervaporation has been tried in the separation (by which we include component concentration) of various azeotropic liquid mixtures. In Polvmer International, 30 (1993), 123-128, Huang and Rhim describe attempts to use the technique with ethanol-water mixtures.

The article lists various polymer membranes which have been used, and specifically describes the use of certain poly(vinyl alcohol) (PVA) membranes modified with maleic acid.

In general, the membranes previously used or suggested for pervaporation of ethanol-water mixtures have not been entirely satisfactory. In many, the selectivity is too low to be of much practical use. In others, the flux rate (i.e. the throughput) is too low to be very useful. In some, the structure of the membrane appears to change in use so that it can only be used for a short time before the performance becomes unacceptable. Others are not useful at high temperatures.

We have now found that membranes of excellent quality, durability and utility for ethanol-water mixtures can be made of vulcanised natural rubber latex.

In one aspect, therefore, the invention provides a method of removing water from a mixture of ethanol and water containing at least 80% by weight ethanol, which method comprises subjecting the mixture to pervaporation using a water-selective vulcanised natural rubber latex membrane.

We have found that it is possible to make vulcanised natural rubber latex membranes for pervaporation of ethanol-water mixtures, which membranes give excellent selectivity, have a high flux rate, and do not deteriorate quickly in use even at high temperatures.

Whilst the method of the invention is useful for increasing the ethanol concentration in mixtures thereof with water, containing at least 80% ethanol, it is especially useful for treating azeotropic mixtures which contain 94.6% ethanol. By the method of the invention, it is possible to obtain mixtures of increased ethanol content, and even ethanol which is substantially free of water.

The formation of thin films of vulcanised natural rubber latex is generally well known and is widely used, for example, in the manufacture of rubber gloves, condoms and other thin-walled rubber articles. In general, the natural rubber latex is compounded with a vulcanising agent (normally sulphur although other compounds can be used instead of sulphur) and an accelerator, and then after forming the film, the mixture is heated to effect vulcanisation. Alternatively, the mixture can be heated to cause prevulcanisation, and then formed into a film which can subsequently be further heated to complete the vulcanisation process.

The thin films which are used as pervaporation membranes in accordance with the present invention can be made in the same way. We have found, however, that their properties and qualities for use as waterselective pervaporation membranes will vary in dependence on a number of factors.

One factor is the degree of vulcanisation (cross-linking).

Unvulcanised natural rubber latex films, whilst giving some selectivity as pervaporation membranes, are not used in accordance with the invention since they tend to be tacky and difficult to handle. Some vulcanisation is needed to avoid these problems and to prevent (or reduce) swelling of the membrane in use. However, the extent of vulcanisation must not be so great as to seriously affect deleteriously the steady state flux or the selectivity of the membrane in use. Over-vulcanisation of natural latex rubber results in a relatively hard, non-elastic material, and this is of course to be avoided. The degree of vulcanisation should be sufficient to provide a strong elastic membrane. For example, films made from natural latex rubber containing 0.2 parts by weight sulphur per hundred parts by weight rubber (0.2 phr), and heated for 12 minutes at 1200C gave good water-selectivity membranes. The films were of a wet thickness of about 75 micrometres. These films were considered to be fully vulcanised in the sense that all the sulphur had reacted. They were flexible elastic films of good strength (having regard to their thickness), and perforation-free. As a general matter, the amount of sulphur to be included will usually be up to about 1.5 phr, preferably about 0.2 phr, but other amounts can be used.

For the purposes of the present invention where the amount of crosslinking is generally low, we prefer to use swelling index as a measure of the crosslinking. Swelling index (S.I.) measures the change in area of a rubber test-piece on immersion in toluene to equilibrium at 25"C. The lower the crosslink density, the higher the degree of swelling and truly uncrosslinked rubber will dissolve in toluene. To measure S.I., we stamp circular discs of 40mm diameter from a thin rubber sheet, mark the edge of the disc with non-soluble ink and immerse it in toluene in a petri dish.

With thin discs, equilibrium is usually reached in 30 minutes. A microscope slide is then placed on the swollen disc to hold it flat and the petri dish placed over millimetre-squared graph paper. The swollen diameter is measured in two orthogonal directions and averaged. Since area is proportional to the square of the diameter (d), S.I. is defined as d2 swollen - d 2 initial. With discs of diameter 40mm, d2 (initial) is 1600 d2 initial mm2. For the present invention, the S.I. of the vulcanised rubber membranes should preferably be from 2.7 to 6.4, more preferably between 3.15 and 4.9.

Another factor which can have some influence on the selectivity and other properties of the membrane is the protein content of the natural rubber latex. As is well known, there is a small amount of protein present in natural latex rubber, largely in the water phase and on or around the surface of (but not within) the rubber particles. We have found that it is advantageous to increase the amount of protein in the latex such as by adding a protein. For example, skim latex can advantageously be used as can other proteins such as casein and lysozyme. Furthermore, we have found that the same or similar improvements are obtained by adding polyacrylamide and, generally, a material containing basic nitrogen groups.

The amount of nitrogen-containing additive is not critical but will be no more than the amount of natural rubber latex with which it is blended, and will usually be very much less.

In the method of the invention, an increase in operating temperature increases the flux rate but decreases permselectivity. It is, therefore, a matter of compromise to select an acceptable flux rate whilst maintaining adequate permselectivity. Generally, the flux rate increases as the pressure on the exit side of the membrane is reduced. The effect of this on permselectivity is relatively small. The flux rate increases as the membrane thickness is reduced but the effect is not pronounced within the usual range of operation. Permselectivity decreases as thickness is reduced. Increase in the degree of sulphur crosslinking of the membrane decreases the flux rate and generally decreases the water permselectivity of the membrane. Uncrosslinked (unvulcanised) membranes were more water-permselective than cross linked membranes, but as previously noted they suffered from the disadvantage of being sticky and hence not practicable.

The nature of the natural rubber latex is not critical, and any of the well known natural rubber latices can be used. However, we have found that high-ammonia (HA) latex is excellent for the purpose. The skim latex additive referred to above is obtained as a by-product in the preparation of HA latex by the centrifugation of the latex.

In general, the thickness of the vulcanised natural latex rubber membranes for use in the present invention will be from about 25 to 225 micrometres, most preferably around 75 to 100 micrometres. As will be understood, whilst thinner membranes can give a higher flux rate, they are generally weaker and so have a shorter life. The thicker membranes, being stronger, can have a longer life but may give a significantly lower flux rate. The thicknesses referred to are wet thicknesses, i.e. the thickness of the wet latex film as cast and before drying. Dry thickness is generally a half to two-thirds of the wet thickness.

The invention will be further described with reference to the accompanying drawings, in which: FIG. 1 is a schematic illustration of laboratory equipment for carrying out the method of the invention; and FIG. 2 is a vertical sectional view of one example of a pervaporation cell.

Referring to the drawings, Fig. 1 shows a chiller unit A the top of which is connected by line T to the top of a pervaporation cell H, and the bottom of the unit is connected by line U to the bottom of the cell H. In line U is a cold liquid circulating pump B. Lines T and U are interconnected by line V which connects T and U to a hot water bath D equipped with a stirrer and thermostate. Line V includes a hot water circulating pump E. Isolating valves C are provided in lines T, U and V where indicated.

A high pressure oxygen-free nitrogen cylinder L, provided with a pressure regulator K, is connected to cell H, with a pressure gauge G adjacent the cell. Cell H is provided with a drain valve I.

Line W connects cell H via metal-metal coupling J to two liquid nitrogen traps M each containing liquid nitrogen N in a flask 0.

Vacuum pump P evacuates line W and a vacuum regulating valve Q is provided, downstream of which is vacustat gauge R and vacuum gauge S.

The cell of Fig. 2 is an example of the type of cell H shown in Fig. 1. The cell comprises a hollow body 20 with caps 21, 22 closing each end. The body contains a glandless stainless steel expandable bellows 5 connected via a passageway 6 in cap 21 to an oxygen-free nitrogen supply line 2 (with an associated pressure gauge 1). At the bottom of body 20 is the membrane 6 under test, held at its periphery to body 20 by O-rings 10 and 15 in cap 22. A porous stainless steel support 8 for the membrane is provided. Cap 22 includes a port 13 for conducting permeate vapour out of the cell (and, in Fig. 1, to the cold traps M). Around body 20 and caps 21, 22 is a jacket 23 with connections 14 and 4 for receiving and exiting a heat medium to maintain the desired temperature.

The body 20 is connected to ducts 11 and 12 which are, respectively, inlet and outlet ducts for continuous supply of, or for filling with, the water/ethanol mix to be treated.

In the operation of the cell, membrane 6 is mounted in position and the body 20 is filled with ethanol/water mixture using ducts 11 and 12. Heat medium is supplied to the jacket 23 and nitrogen to the bellow 5. Referring to Fig. 1, heat medium is stored in A and conducted in lines T and U, with optional passage as derived through water bath D. the nitrogen gas is supplied from L. A vacuum is applied by pump P to draw pervaporate out of outlet 13 of the cell through line W to the cold traps M where it is condensed. No further detailed description will be given since the method of pervaporation is well known.

In order that the invention may be more fully understood, the following Examples are given by way of illustration only.

Example 1 Membrane DreDaration a) Mixtures were made up of natural latex rubber. sulphur and an accelerator, the sulphur contents being 0.2, 0.5 and 1.5 phr. The mixtures contained (in parts per hundred dry rubber by weight) antioxidant (1 phr), zinc oxide (1.5 phr), accelerator (1.0 phr), the natural rubber latex containing 60% rubber solids. Films of varying thicknesses were formed immediately from these mixtures by applying a small quantity of each mixture to a glass plate and wiping with an applicator to spread the mixture. to an even thickness. The wet thicknesses were 25, 50, 75 and 100 micrometres. Aftei drying at room temperature for ten minutes. the films were sprinkled with talc and then carefully lifted from the glae plate and placed in an oven for curing. A curing temperature of 25it, 60'C or 120 C wasusedforl2minutes.

After washing off the talc with water and/or isopropyl alcohol, pinholefree membranes were cut from the films for use in pervaporation.

b) Membranes were also made, for comparison only, from the natural latex rubber alone (with no sulphur).

Example 2 Pervaporation a) The apparatus shown schematically in Fig. 1 of the accompanying drawings was used, the pervaporation cell used being shown in section in Fig. 2.

b) Each of the membranes employed was tested for any inherent defects (e.g. pinholes) and then (if satisfactory used for pervaporation of an ethanol-water mixture.

Pervaporation flux measurements were conducted a a constant temperature within the range of 5sC to 80it.

Before the start of an experiment, a selected temperature was obtained by circulating either the heating or cooling liquid through the outer jacket of the pervaporation cell for a duration of not less than half hour.

An Edwards vacuum pump was used to evacuate the cold traps. and the vacuum line. The vacuum line was connected to the downstream compartment and the cold traps which were immersed in liquid nitrogen simultaneously. A vacustat was used to measure the vacuum in the line. A nitrogen pressure of 0.3 to 0.4 mPa gauge was applied indirectly to the feed solution via the stainless steel bellows. The cold traps were changed approximately once every hour in order to collect the condensed permeate vapour. The inlet and outlet ports of the cold trap were immediately sealed with rubber bellows to eliminate the loss of condensed permeate due to evaporation during thawing process. The cold traps were weighed to determine the amount of permeate flux and transferred into the small sample bottle and placed in a refrigerator for chemical analysis. The pervaporation experiment was carried out for a steady state period of approximately seven hours every day. In some cases, the test was continued the following day or for the rest of the following week. Hence the pervaporation flux measurement was carried out for a maximum of 120 hours. Pervaporation experiments were also conducted with membranes which had already been used. The results (i.e. separation, flux) were found to be the same as with freshly made membranes.

The feed and permeate were analysed by gas chromatography or by refractive index measurement.

c) Results The results are set out in the following Tables 1 and 2. In the Tables, the membrane thicknesses are of the wet membranes. Drying reduces the thickness by about 50%.

TABLE 1 Steady state values of permeate flux rate and composition of pervaporate at 500C of various water ethanol mixtures through 75 micrometre thick fully crosslinked (90 minutes at 100"C) rubber latex membranes with 0.2 phr sulphur, made from compounded HA latex described more fully hereafter in Examples 3 to 11.

Wt% EtOH in Flux rate Wt% of EtOH feed (g/m2h) in permeate 0 182 0 5 104 14.3 20 104 36.2 50 156 57.8 96 343 90.2 100 166 100 As can be seen, these membranes are ethanol selective up to 50% ethanol feed, but at 96% are water selective. The change occurs between at 70% and 80% ethanol feed.

TABLE 2 Pervaporation at 25"C through 75 micrometre thick "fully crosslinked" rubber latex membrane with different levels of incorporated sulphur.

Concentration of 90 Wt% 97.4 Wt% EtOH in upstream EtOH EtOH feed Rubber membrane with 0.2 phr sulphur Total permeate 593 593 flux (g/m2hr) Water flux in 565 424.8 permeate (g/m2hr) Wt% of EtOH 4.7% 5.8% in permeate Rubber membrane with 1.00 phr sulphur Total permeate 280 189 flux (g/m2hr) Water flux in 151 74.8 permeate (g/m2hr) Wt% of EtOH 46.7% 60.47% in permeate Rubber membrane with 1.5 phr sulphur Total permeate 190 88.8 flux (g/m2hr) Water flux in 90 19.8 permeate (g/m2hr) Wt% of EtOH 51.4% 77.5% in permeate Examples 3-12 A series of natural rubber latex membranes were made and each used for pervaporation of a mixture of ethanol and water containing 94.0 or 94.5% ethanol. The results are set out in Table 3. In the Table, "steady state" means the average value reached after two hours.

All the membranes were made from a compounded HA latex blended with various additives as described. The compounded HA latex consisted of HA latex (100 parts), potassium hydroxide (0.10 parts), zinc diethyldithiocarbamate (1.0 part), zinc mercaptobenzothiazole (1.0 part), sulphur (0.20 parts) and zinc oxide (1.0 part) (the parts are by weight).

Unless otherwise noted, vulcanisation was for 12 minutes at 1200C.

The additives used were: polyacrylamide; casein; polyvinylpyrrolidone; Sephadex G-25 and G-25-150 (beaded gels prepared by crosslinking dextran with epichlorohydrin under alkaline conditions, sold by Sigma-Aldrich Co. Ltd.); lysozyme (from chicken egg white); and skim latex. Skim latex is obtained as a by-product in the preparation of high-ammonia (HA) latex by the centrifugation of field latex. It contains the bulk of the non-rubber substances present in field latex and is especially high in protein. In general, films from HA latex contain about 3 to 4% protein; those from skim latex contain about 10 to 15% protein (percentages are by weight).

TABLE 3

Example Initial Opera. Steady Steady Steady Steady No. Concen. Temp. "C State State State State %wt Flux Concen. Separation P.S.L* EtOH g/m2h %wt EtOB Factor 3 94.5 16 69 86.6 2.66 182 50 118 89.1 2.12 249 4 94.5 18 45 81.1 3.86 172 50 104 89.3 2.06 214 78 171 90.3 1.85 316 5 94.0 15 13 81.0 3.68 48 78 280 85.2 272 763 6 94.0 IS 71 83.0 3.28 231 78 470 89.0 1.94 910 7 94.0 78 184 87.8 2.27 418 78 218 87.5 2.24 488 8 94.0 15 44 82.1 3.42 151 78 366 90.0 1.74 637 78 493 92.5 1.27 627 9 94.0 15 49 85.2 2.72 134 78 405 92.0 1.36 552 78 104 90.6 1.63 169 78 139 91.2 1.51 209 10 94.0 15 151 85.5 2.66 ~ 400 11 94.0 15 44 86.5 2.45 108 78 163 89.2 1.90 309 12 94.0 18 134 72.7 - * Pervaporative separation index which is given by multiplying the flux (in g/m2h) by the separation factor.

Details of each membrane are as follows: Example 3: Compounded HA latex + 0.1 1% of polyacrylamide by weight.

Three coatings were used to prevent pin holes in the membrane.

Dry film thickness: not greater than 150 micrometres.

Example 4: Compounded HA latex + 0. 18% by wt of casein.

Three coatings were used to prevent pin holes in the membrane.

Dry film thickness: not greater than 150 micrometres.

Example 5: Compounded HA latex (lOg) + skim latex (lOg) and zinc oxide (0.6g) Triple coating was employed to prevent pin holes in the membrane.

After the first two coatings, the film was dried inside an oven for 10 minutes at 1200C.

After the third coating, the film was again vulcanised for 12 mins at 1200C.

Dry film thickness: not greater than 150 micrometres.

Example 6: Compounded HA latex (lOg) + skim latex (lOg) + zinc oxide (0.6g) Four coatings were used to prevent pin holes in the membrane. These were built up of a first coating of the mixture, then a coating of the skim latex alone, then a second coating of the mixture and finally a top coating of the skim latex.

Each coating was dried at 1200C for 10 mins before the application of the next coating.

Finally, the film was again vulcanised for 12 mins at 1200C.

Dry film thickness: not greater than 150 micrometres.

Example 7: Compounded HA latex (lOg) + skim latex (lOg) + Sephadex G-25 (0.2g) and zinc oxide (0.6g).

Triple coating was used to prevent pinholes in the membrane.

After the first two coatings, the film was dried inside the oven for 12 min at 1200.

After the third coating, the film was vulcanised for 12 mins at 1200C.

Dry film thickness: not greater than 150 micrometres.

Example 8: Compounded HA latex (lOg) + skim latex (14.6g) and zinc oxide (0.6g). The 14.6g of skim latex was obtained from drying 20g skim latex. The casting method was similar to Example 6.

Example 9: Compounded HA latex (lOg) + skim latex (9.lg) and zinc oxide (0.6g). The 9.lg of skim latex was obtained from drying 20g of skim latex.

Four coatings were used to prevent pinholes in the membrane. These were built up of a first coating of the compounded HA latex, then a first coating of the skim latex, then a second coating of the compounded latex and finally a second top coating of the skim latex. Each coating was dried at 1200C for 10 mins before the application of the next coating.

Example 10: Compounded HA latex (lOg) + skim latex (5.5g) + lysozyme (0.3g) and zinc oxide (0.6g).

The 5.5g of skim latex was obtained from drying 10g of skim latex. The casting method was similar to Example 8.

Example 11: Compounded HA latex (lOg) + skim latex (2.6g) + Sephadex G-25-150 (0.9g) and zinc oxide (0.3g) The 2.6g of skim latex was obtained from drying lOg of skim latex. The casting method was similar to Example 7.

The pervaporation conditions were: Upstream Pressure: 50 psi Downstream Pressure: 3.0 mmHg Duration of each test: 6.5 hrs Method of analysis: Refractive index Example 12 Compounded HA latex + 6.5% by weight of polyvinylpyrrolidone.

Dry film thickness: not greater than 100 micrometres.

Comparative Example In this Example, the need for the presence in the rubber membrane of a basic nitrogen component is illustrated by comparing the performance of a membrane prepared from an uncompounded, unvulcanised synthetic polyisoprene latex (Shell Chemicals' IR-309 latex), which contains no protein or other basic nitrogen material, with that of a similar membrane prepared from uncompounded, unvulcanised natural rubber latex (HA latex) in the pervaporative separation of a 94.0 wt % ethanol feed at 180C.

Membrane Steady State Flux Steady State Concentration (g/m2.h) (wt % EtOH) Synthetic polyisoprene 277 91.5 HA latex 167 82.5

Claims

CLAIMS: 1. A method of removing water from a mixture of ethanol and water containing at least 80% by weight ethanol, which method comprises subjecting the mixture to pervaporation using a water-selective vulcanised natural rubber latex membrane.
2. A method according to claim 1, wherein the mixture is an azeotropic mixture of ethanol and water.
3. A method according to claim 1 or 2, wherein the membrane is a blend of natural rubber and a material including basic nitrogen groups.
4. A method according to claim 3, wherein the membrane is a blend of natural rubber and a protein or skim latex.
5. A method according to claim 3, wherein the membrane is a blend of natural rubber and polyacrylamide.
6. A method according to any of claims 1 to 5, wherein the natural rubber latex is a high ammonia latex.
7. A method according to any of claims 1 to 6, wherein the degree of vulcanisation of the natural rubber latex in the membrane is such that the swelling index is from 2.7 to 6.4.
8. A method according to any of claims 1 to 7, wherein the membrane is cast as a wet latex film and then dried, and wherein the thickness of the wet latex film is from 25 to 225 micrometres.
9. A method of increasing the concentration of ethanol in a mixture of ethanol and water containing at least 80% by weight of ethanol, substantially as herein described with reference to any of the Examples.
10. A mixture of ethanol and water which has been concentrated by the method of any of claims 1 to 9.
11. Substantially water-free ethanol produced by the process of any of claims 1 to 9.