SE500579C2 - Regulating the compsns. of permeates during pervaporation of mixts. - in order to enrich the concn. of a particular component by altering the hydrodynamics in the feed soln. e.g. beverages, dairy products - Google Patents

Abstract

Method for regulating the relative permeate compsns. during liq. sepn. of multi-component mixts. by a pervaporation process, comprises altering the hydrodynamics in the feed in order to regulate the concns. of the different components to a pre-determined ratio of the permeate.

Description

2 full or partial vacuum or by sweeping an inert gas on the permeate side of the membrane. The permselective membrane will be more permeable to certain kinds of substances compared to other kinds of substances, ie the composition of the permeate will be different from the composition of the feed solution. The preferentially permeated substances will thus be enriched in relation to the non-preferentially permeated substances. The selectivity of the process for a component called i can be described by the enrichment factor, B ir si = PJ (1) Cf, i and Cfii are the volumetric concentrations of in which Cp'i component i in permeate and feed solution, respectively. A preferentially permeated component will have an enrichment factor greater than 1 while a non-preferentially permeated component will have an enrichment factor of less than 1. Whether a component will be permeated preferentially or non-preferentially depends on the membrane material. In general, hydrophilic membranes will preferentially permeate hydrophilic substances over hydrophobic substances, and vice versa.

Examples of membrane materials that could be used to make hydrophobic membranes are elastomeric polymers such as poly [dimethylsiloxane], poly [1-trimethylsilyl-1-propyl fl] and its derivatives, polyether block amides, nitrile butadiene rubber and polyhirethane]. Materials that could be used for the manufacture of hydrophilic membranes are polymers that are below the glass transition temperature such as cellulose acetate, poly [vinyl alcohol, poly [sulfone] and poly [vinyl chloride] [Pervaporation membrane separation processes, R.Y.M. Huang (Ed.), 1991].

Previous technology An article in Journal of Food Engineering 10 (1989) 65-71, Recovery and concentration of apple juice aroma compounds by pervaporation, by E. Bengtsson, G. Trägårdh and B. Hallström described recycling and concentration of flavorings from apple juice, 3 ie a complex multicomponent mixture. Separation took place as a result of the difference in permeability of the various components through the membrane, but no effect on the feed solution occurred. The investigation showed that the enrichment factors for the different flavoring substances were noticeably different for different substances.

In an article by S.W. Zhang and T. Matsuura in Journal of Food Process Engineering 14 (1991) 291-296 described the recovery and concentration of flavoring substances in apple licenses by means of pervaporation. The feed solution was not affected and it was only shown that flavorings in the apple essence could be concentrated by means of the pervaporation process.

I Proceedings of Third International Conference on Pervaporation Processes in the Chemical Industries, Bakish Materials Corporation, Englewood, New Jersey, 1988, 429-438 described 'A.

Voilley, B. Schmidt, D. Simatos and S. Baudron extraction of two flavors, which were used separately, by the pervaporation process. It was shown that the flow rate of the feed solution played a role as a factor in the concentration of the flavor. sid.

In Journal of Membrane Science 36 (1988) 373-384, R. described

Psaume, P. Aptel, Y. J.C. Bersillon extraction of trichlorethylene in dilute aqueous solutions with perforation. It was shown that the concentration of trichlorethylene in the feed solution did not affect the flow of water through the membrane while Aurelle, Mora. and J.L. the flow of trichlorethylene increased with the concentration in the feed as well as with the Reynolds number.

In Proceedings of Third International Conference on Pervaporation Processes in the Chemical Industries, Bakish Materials Corporation, Englewood, New Jersey, 1988, p. 449-462, P. Cöté and C. Lipski described pervaporation of wastewater containing trichlorethylene, carbon tetrachloride and phenol. High flows and separation factors were obtained for the removal of trichlorethylene and carbon tetrachloride for which performance was a function of the Reynolds number while performance for removal of phenol was found to be ¿_, _ f, __.:_T ,. 4 a) I J '4 independent of the Reynolds number.

In Proceedings of Fifth International Conference on Pervaporation Processes in the Chemical Industries, Bakish Materials Corporation, Englewood, New Jersey, 1991, p. 110-122, T. Lamer and A. Voilley described how different parameters affect the pervaporation of flavorings. Four different flavors were studied, though not in a mixture but separately. It was shown that the flux of the flavor increased with the flow rate.

In Separation Science and Technology 27 (4), p. 467-491, 1992 described R. Gref, Q.T. Nguyen and J. Néel influence of membrane properties on system performance during pervaporation during concentration polarization. The study used octanol solutions of water, which had a maximum water content of 4.7%.

Only one organic substance was used, namely octanol. It was shown that the selectivity of the pervaporation process can in a complex way depend on the mass transport in the boundary layer due to a possible concentration-dependent coupling of the flux in the membrane.

Pervaporation has been used to dewater mixtures of water and organic substances [U.S. Pat. 2,913,507, 2,953,502 and 2,981,660, Binning et al] and for the separation of hydrocarbon mixtures [U.S. Pat. 3,930,990, Brun et al., And U.S. Pat. 3,776,970, Strazik et al]. Pervaporation has also been used for the extraction of organic substances from aqueous solutions [U.S. Pat. 4,952,751, Blume et al., And U.S. Pat. 4,591,440, Higashimura et al]. In those cases, the organic substances found in the permeate will be enriched several times, but depending on the selectivity of the membrane, different substances will be enriched differently. Thus, a multicomponent mixture of organic substances in an aqueous solution used as a feed solution will give rise to a permeate enriched in organic substances, but with a different relative composition of the organic substances. If the relative composition of the organic substances in the feed solution is highly desirable, this change in the relative composition will be a disadvantage of the process. As mentioned above, the mass transfer from the bulk of the feed solution to the permeate can be considered to take place in four successive steps, ie a mass transport in the feed solution, sorption of penetrants to the membrane, mass transport through the membrane and desorption to the permeate. The flux, Ji, ie the mass transfer rate through the membrane (m3 / (m2s)), is affected by several parameters such as temperature, permeate pressure, membrane thickness, membrane material and concentration in the feed solution [Pervaporation membrane separation processes, R.Y.M. Huang (Ed.), 1991]. In general, increased temperatures, reduced permeate pressures and reduced membrane thicknesses lead to increased flux.

As mentioned above, it has also been observed that the conditions in the feed solution affect the flux through the membrane. Increased cross-flow rates, or increased agitation rates, lead to increased flux through the membrane. This has been considered to be due to mass transport in an area of the feed solution which is close to the membrane, the so-called the boundary layer, which can limit the flux through the membrane. The flux through the boundary layer has been described with equations such as Eq. (2), J: = kni (Cf.i 'CbLi) <2) in which kf'i is a mass transfer coefficient for the boundary layer, which can be obtained from Sherwood correlations. For cross flow perporation, these can generally be written as Eq. (3), d Sh = a ~ Re ° -Sc ° i (s) L da kgd 'r SfI - = ._'_' f_ _ (4) l: d Re = z "(m V Sh = V ï (6) in and for agitated cells, Sh == a mRe ° -Sc 'where Sh = k fi lr z Re: (Or2 V Sk = .ÄÄ D.- in which a Constant b Constant C Constant d Constant dh Hydraulic diameter Mass transfer coefficient Radius in stirred cell Diffusion coefficient Flux Length Re Reynolds number Sc Schmidt number Sh Sherwood number SEK C Volumetric concentration DJL Examples of the coefficients a, [_ [_ [_ [_ (7) (8) (9) (10) [-]] v Kinematic viscosity [ m2 / s]] B Enrichment factor] m Stirring speed [row / s] 1 [m] Index [m / s] f Feed solution [m] in Permeable component in [-] org Organic [m2 / s] p Permeat [m / s ] rel Relative [N] -1 -1 1 | _ "If" 'Il ”-1 b, c and d are given in Table 1.

In summary, it can be said that it is known that pervaporation enriches multicomponent mixtures several times, but the process 7 has different selectivity for different substances. This in turn will lead to an enriched permeate whose composition with respect to relative concentration for the mutual substances is different from the composition of the feed solution. It has also been found that pervaporation is affected by mass transport in the boundary layer of the feed solution, but this has only been used in optimization studies of pervaporation for one substance dissolved in another.

Drawings In the drawings, Figure 1 shows a pervaporation plant used in Example 1 for carrying out the method of the present invention.

Figure 2 is a graphical representation of organic fluxes plotted against the Reynolds number at a constant hydraulic diameter.

Figure 3 is a graphical representation of organic fluxes plotted against the hydraulic diameter at a constant Reynolds number.

Figure 4 is a graphical representation of normalized fluxes plotted against the Reynolds number at a constant hydraulic diameter, which shows how the feed flow affects the permeate composition.

Figure 5 is a graphical representation of normalized fluxes plotted against the hydraulic diameter, which shows how the feed flow at constant Reynolds numbers affects the permeate composition.

Figure 6 is a graphical representation of relative permeate concentration plotted against feed concentration and Reynolds number at a constant 'hydraulic diameter.

Figure 7 is a graphical representation of combinations of feed concentrations and Reynolds numbers resulting in the desired relative permeate concentration described in Example 3.

Concentration of beverages is of great importance for the industry, e.g. in the manufacture of concentrated juices to be diluted with water - ___ f.

LH \ u æ - RJ 8 before consumption. A mild and good method for concentration is with the help of reverse osmosis, but due to fundamental causes of the process, this method can not concentrate beverages to commercially acceptable concentrations and is therefore of little importance [Bengtsson, Trägårdh & Hallström J. Food Soi ., 1989]. Evaporation is instead an alternative that is often used when concentrating beverages, but when evaporating beverages, flavorings are often lost together with the leaving water vapors. This loss can be reduced in two ways. The flavors can be extracted directly from the beverage before concentration and then returned after concentration. This is an interesting method if the flavorings are temperature sensitive and could be affected by the heat treatment during concentration. Flavorings that remain in the beverage during concentration can, however, be lost during concentration. The second alternative is to condense the evaporator vapors and then extract the flavorings from it. The advantages of this are that the flavorings in the evaporator vapors are often enriched and that all the flavorings that are lost during the evaporation are subjected to extraction. In both cases, the extracted flavors are present in low concentrations and pervaporation has been shown to be a suitable technique [cf. Bengtsson et al. ibid.]. However, previous pervaporation studies have indicated that the pervaporation process changes the flavor composition in such a way that the permeate has a different composition compared to the feed solution.

SUMMARY OF THE INVENTION An object of the invention is to provide a method by means of which in a pervaporation process of a multicomponent mixture the relative concentration of one or more components in the permeate can be intentionally changed to adjust its relative concentration in comparison with the other components in the invention. the multicomponent mixture to a concentration suitable for obtaining a desired concentration ratio of the various compounds.

Another object of the present invention is to provide: u! 9 provide a method that can be used to minimize the difference between relative permeate concentration and relative feed concentration for a multicomponent feed and its permeate pervaporation.

The above object is achieved by manipulating the hydrodynamics in the feed flow by adjusting one or more parameters which affect the hydrodynamics in the feed, so that the desired relative composition of the permeate is obtained.

The method of the present invention is useful for treating various multicomponent mixtures. Multicomponent mixtures for which the method of the present invention is useful are e.g. plant extracts as aroma and fragrance materials, beverages selected from the group consisting of fruit juices, vegetable juices, tea, coffee, alcoholic beverages such as but not limited to, beer, wine, cider and including alcoholic beverages that are completely or partially de-alcoholized, dairy products and mixtures of the above groups.

The present invention is suitable in any process in which a multicomponent mixture is treated with pervaporation and where the composition of the permeate is important. Such a process is, for example, a process in which fermentation products are harvested from a fermenter. If only one of the fermentation products is desired while the other fermentation products are considered as by-products, then the method can be used to increase the efficiency of the process by adjusting the feed hydrodynamics in such a way as to increase the relative concentration in the permeate of the desired component.

Another example where the invention is suitable is e.g. if the pervaporation process is used to produce a permeate which must have a given composition. Such a process can e.g. found in the beverage industry. Suppose e.g. that a beverage with a unique flavor composition should be concentrated using pervaporation. Pervaporation of the beverage would result in an enriched permeate having a relative composition in the permeate that is different from that of the beverage. This could be fatal in such a way that the taste profile of the beverage is lost and that the original taste of the beverage cannot be regained if the concentrated beverage is diluted again to obtain a drinkable product. By adjusting the hydrodynamic conditions in the feed solution, this difference can be minimized.

By adjusting the hydrodynamic conditions in the feed solution, by manipulating crossflow speeds / stirring speeds, the geometric profile of the modules, etc., the mass transport in the feed boundary layer can be affected and thus also the total mass transport in the pervaporation process. This is used to drive the pervaporation process at a given hydrodynamic state in the feed which results in the mass transfer coefficients in the feed affecting the overall mass transport in such a way that the desired concentration ratio in the permeate is obtained.

As mentioned above, different types of membranes have different selectivity for different substances, ie. some substances are permeated more easily than others through a given membrane. A substance that is easily permeated through the membrane will also be strongly affected by the hydrodynamic conditions in the feed solution, while a substance that is permeated less easily will be less affected, ie. a substance that is only slightly susceptible to the method can be made susceptible by a change of membrane material.

The invention is illustrated by the following examples without, however, limiting it.

Example 1 Pervaporation experiments were performed with equipment schematically shown in Figure 1. A multicomponent flavor model, the composition of which is given in Table 2, consisted of five typical flavorings in an aqueous solution. Ten liters of the multi-component aroma model: recycled, with different crossflow speeds through four plate modules.

The modules had a length of 230 mm, a width of 20 mm and a height of 0.78 mm, which corresponds to a hydraulic diameter of 1.50 mm.

The membranes mounted in the modules were poly [dimethylsiloxane] -11 composite membranes (Type 1060, Deutsche Carbone, Geschäftseinheit GFT, Neuenkirchen-Heinitz, Germany) with a active layer thickness of 10 μm. During the experiments, the temperature was kept at 25 ° C and the permeate pressure was 2 mbar. The system was allowed to stabilize for three hours before measurements were performed during the fourth hour. The permeate was condensed with liquid nitrogen, and a Varian Vista 6000 equipped with an NB-351 column (25 m, 0.32 mm) was used to determine the composition of feed and permeate. Each experiment was repeated twice so that each measurement consisted of three subsamples. These experiments correspond to an increase in the feed cross-flow rate, here expressed as the dimensionless Reynolds number, which resulted in increased total mass transfer coefficients for the model components, but as mentioned above, the increase for the different components may not be the same for all components and if the relative the concentrations in the permeate of the model components are compared with each other as in Figure 4 where, Ji Ci, rel, p = (11) z Jorg this phenomenon is clarified. The cross-flow rate, which in this case is expressed as the dimensionless Reynolds number, can thus be used to regulate the relative composition of organic substances in the permeate. If the relative composition of the permeate, Ci rel pi figure 4 is compared with the relative composition of the feed, Cfi Ci, rel, f = (12) 2 Cf, org see table 2, it is found, in this particular experiment, that operation of the equipment at Reynolds at approx. 100 gives a permeate whose difference in composition compared to the feed is minimized.

Example 2 Pervarporation experiments were performed as in Example 1 but now the hydraulic diameter was varied instead while the Reynolds number was kept constant at 1000. Effects on flux and relative concentrations in the permeate are presented in Figures 3 and 5, respectively. 12 Example 3 This example illustrates how the method can be used to regulate the relative concentration in the permeate for a component that has a variable concentration in the feed.

The process can be described by the following equations Ji = kov, icf, i (13) 1 = 1 + 1 (14) kov, in kf, in kmembr, i 'where kf'i can be obtained from equations (3) - (10) and where kmembr i is a membrane constant for the membrane used in Example 1, see Table 2. The model's ability to describe the process is presented in Figures 2 and 3 where squares represent experimental measurement data and where solid lines represent model calculations. The flux can then be converted to relative concentrations by equation (11).

Assume that an aqueous solution contains 1-penten-3-ol, in a concentration of 10 ppm, trans-2-hexenal in a concentration of 20 ppm, and linalool whose concentration can vary between 35 and 55 ppm. Assume also that pervaporation 'of the mixture should give a permeate in which the relative concentration of linalool must be 0.55¿0.05. Simulated results on linalool's relative concentration in the permeate are presented in Figure 6 as a function of linalool concentration in the feed and crossflow velocity, expressed as the dimensionless Reynolds number, for a pervaporation plant described in Example 1, but with a hydraulic diameter of 2.00 mm. As shown in Figure 6, certain combinations of feed concentration and Reynolds number give rise to undesirable relative concentrations of linalool in the permeate. Consequently, the crossflow rate at high feed concentrations would be adjusted so that the Reynolds numbers become large so as to obtain the desired relative concentration of linalool in the permeate. Similarly, the crossflow rate must be reduced so that low Reynolds numbers are achieved if the feed concentration is low. The combinations of feed concentrations and Reynolds numbers that give rise to the desired relative concentration of linalool in the permeate are presented in Figure 7, which illustrates the ability of the method to be used to compensate for fluctuations in the feed concentration.

As will be apparent to those skilled in the art, many modifications and variations can be made in this invention without departing from the spirit and scope thereof.

Table l. _ Some examples of the coefficients in the Sherwood correlation.

Flow - bcde Pervaporations Reference regime system Laminar __.- 1.62 0.33 0.33 0.33 Cross fl ow * RQ <2000 1.85 0.33 0.33 0.33 cross fl ow ”r 10 'Re> 105 0.023 0.8 0.33 0 Cross fl ow in Larninär 0.29 0.57 0.33 - Stirred cell 1 'Tufbulenz 0.044 0.75 0.33 - Stirred cell T * E. Matthiasson and B. Sivik, Concentration polarization and fouling, Desalination 35 (1980) 59-103.

T H.H. Nijhuis, Removal of Trace organics from water by pervaporation - A technical and economic analysis. Ph.D. thesis, University of Twente, Enschede, The Netherlands, 1990. i V. Gekas och B. Hallström, Mass transfer in the membrane concentration polarization layer under turbulent cross flow.

I. Critical literature review and adaptation of existing Sherwood correlations to membrane operations, Membrane Science, 30 (1987) 153-170.

Journal of 14 Table 2.

Properties of the multi-component model. 2-methyl-2-methyl-1-penten-trans-2-lina-propanal butanal-3-ol hexenal lool 9 7 10 20 50 Conc. (ppm) Ci, error 0.09 0.07 0/10 0.21 0.52 CORPRODUCT 64.6 92.2 115 146 198 (° C) olekyl weight 72.11 86.14 86.13 98.14 154,25 Di 2 QBZIUN S fi ßdüm Bßllüm 7f & 1Um 6§210w (m / S) fgfïggr. i 4, s7-1o '° 1,19-1o- “132-10 * Los-w * s, 1s-1o' °

Claims (8)

15 PATENT REQUIREMENTS A method for controlling the relative permeate composition in liquid separation of multicomponent mixtures by pervaporation process, characterized by manipulating the hydrodynamics in the feed to control the concentration of the various components to a predetermined ratio in the permeate. Method according to claim 1, characterized in that the manipulation of the hydrodynamics in the feed takes place by adjusting the feed's crossflow speed. Method according to claim 1, characterized in that the manipulation of the hydrodynamics in the feed takes place by adjusting the stirrer speed in a stirred cell. A method according to claim 1, characterized in that the manipulation of the hydrodynamics in the feed takes place by adjusting the hydraulic diameter again crossflow pervaporation plant. 5. A method according to claim 1, characterized in that the manipulation of the hydrodynamics in the feed is achieved by adjusting the radius of a stirred cell. Method according to claim 1, characterized in that the manipulation of the hydrodynamics in the feed is achieved by adjusting the viscosity of the feed. Method according to claim 1, characterized in that the multicomponent mixture is a beverage selected from the group consisting of fruit juices, vegetable juices, tea, coffee, alcoholic beverages such as beer, wine, cider and comprising fully or partially alcoholic alcoholic beverages, dairy products and mixtures of the abovementioned beverages. - ^. \ --_-. \,. . . _ \ .- Ä .. ~ '\ .f' i / 16 Method according to claim 1, characterized in that the multicomponent mixture is a material from plant extracts such as e.g. aroma and fragrance materials.