HK1019568A - Method for separating a catalyst by membrane electrodialysis - Google Patents

Description

Method for separating catalyst by membrane electrodialysis

The invention relates to a method for separating a catalyst from a solution containing the catalyst by membrane electrodialysis.

More specifically, the present invention relates to a process for the separation of a catalyst for a homogeneous molecular oxygen oxidation reaction.

There are a number of homogeneous catalytic oxidation processes whereby the oxidation of cycloalkanes to the corresponding diacids can be carried out using soluble heavy metal (e.g. cobalt or manganese) salts.

Us patent 2223493 published 12.1940 discloses the oxidation of cyclic hydrocarbons to the corresponding diacids by means of an oxygen-containing gas in the presence of an oxidation catalyst such as a cobalt compound, typically in a liquid phase containing acetic acid, at a temperature of at least 60 ℃. This patent envisages isolation of the adipic acid formed by crystallization, but does not describe a method by which the catalyst can be recycled to a new oxidation operation, let alone relating to the activity of the catalyst after one or more cycles.

Patent FR- A-2722783 discloses A method for separating and recycling the cobalt catalyst used in the oxidation reaction of cyclohexane to adipic acid, after separation of the main reaction product and at least part of the acetic acid solvent from the reaction mixture. The process essentially comprises extracting the majority of the catalyst with the aid of cyclohexane or a mixture of cyclohexane and acetic acid. The process is very efficient, the recycled catalyst does not lose its activity; however, this process involves a large amount of solvent and requires several successive operations.

It follows that there is a need for a process which makes it possible to separate the homogeneous catalyst dissolved in the reaction mixture, which process should be equally effective when applied in a simpler manner.

Patent FR- A-1591176 discloses A process for recovering A metal catalyst and nitric acid present in A mother liquor obtained from the separation of reactants obtained from the nitric oxidation of cyclohexanol and/or cyclohexanone, which process comprises passing part of the mother liquor containing metal salts, nitric acid and organic acids through an electrodialysis cell. The metal catalyst used is a copper salt or a vanadium salt.

Patent FR- A-2026288 discloses A process for recovering large quantities of nitric acid and of metal cations from an acidic residual liquid formed in the preparation of adipic acid by liquid-phase oxidation of cyclohexanone or cyclohexanol, which comprises introducing this liquid into an electrodialysis unit comprising one or more electrodialysers, and recovering the nitric acid and the metal cations in A recovery liquid which may be water or A dilute solution of nitric acid. The metal catalyst used is a copper salt or a vanadium salt.

The two processes are very similar, or even identical, using concentrated nitric acid solutions. This feature greatly facilitates the separation of the metal salt present as nitrate from the undissociated carboxylic acid.

The invention relates to the separation of a homogeneous catalyst for the oxidation of cyclohexane with oxygen, which catalyst is free of nitric acid.

More precisely, the invention relates to a method for separating a homogeneous catalyst dissolved in a mixture also containing at least one fatty diacid, characterized in that the catalyst comprises cobalt and the separation is carried out by membrane electrodialysis.

Homogeneous catalysts are metal compounds commonly used in reactions for oxidizing cycloalkanes to fatty diacids. They are more specific catalysts containing cobalt alone or in admixture with other metals such as manganese, copper, iron, vanadium or cesium or mixtures thereof. These metals are present in the form of compounds which are soluble in the reaction mixture of the oxidation of cycloalkanes. Such compounds are hydroxides, oxides and organic or inorganic salts. Preferred compounds are cobalt salts, which may be present alone or in admixture with compounds based on the following metals: manganese and/or copper and/or iron and/or cesium and/or vanadium.

Examples of such cobalt salts which may be mentioned include cobalt chloride, cobalt bromide, cobalt nitrate and cobalt carboxylates, such as cobalt acetate, cobalt propionate, cobalt adipate, cobalt glutarate or cobalt succinate. Cobalt acetate tetrahydrate is particularly preferred since one of the most common solvents used in the oxidation of cycloalkanes is acetic acid.

The mixture subjected to membrane electrodialysis contains at least one diacid formed in the oxidation of cycloalkanes and usually one or more other diacids formed as by-products; the mixture may also contain all by-products of the reaction. When the catalyst is used in the oxidation of cyclohexane, adipic acid is mainly obtained, but glutaric and succinic acids and more or less cyclohexanol, cyclohexanone, cyclohexyl esters, lactones and hydroxycarboxylic acids may also be formed.

Although the oxidation is generally carried out in an organic solvent or without a solvent under appropriate conditions, acetic acid is preferred in the case of cyclohexane oxidation, but it is preferred that the mixture subjected to electrodialysis contains water.

It follows that the mixture in which the homogeneous catalyst is present preferably comprises water; the solvent which has optionally been used in the production of the solution to be treated can be replaced in whole or in part by water before the electrodialysis. The water may generally be from 10% to 100% of the solvent mixture of the solution subjected to electrodialysis, preferably from 50% to 100% of such solvent mixture.

In general, electrodialysis is a process of extracting ionized species in a solution to be treated by migration through an ion exchange membrane under the action of a dc electric field.

The electrodialysis device used comprises different chambers separated by cation membranes and anion membranes alternately. These compartments can be divided into a dilution compartment (D), in which the compound to be separated, i.e. the catalyst of the process of the invention, is present in a small amount, and a concentration compartment (C), in contrast, which is rich in the compound to be separated.

In fact, under the action of the electric field, the cations in the solution to be treated migrate towards the cathode through the cation exchange membrane (cation membrane), leaving the chamber (D). When they move to the next chamber (C), they will not be able to leave that chamber due to the presence of the next anion exchange membrane (anion membrane). At the same time, the anions migrate towards the anode, pass through the anion membrane, entering the adjacent chamber (C), and therefore cannot leave due to the presence of the next cation membrane.

Two adjacent compartments (C) and (D) form an electrodialysis cell. The electrodialyser comprises a stack of membranes with a large number of cells. The number of channels per electrodialyser should generally be as large as possible, for example, the number advantageously varying between 10 and 500 channels.

In fact, the anionic and cationic membranes are placed alternately in a filter-press type system.

The inorganic films used in the method of the present invention can be classified into two major groups according to the preparation method thereof.

First, they are heterogeneous membranes, which can be made from ion exchange resins mixed with binders (e.g., polyvinyl chloride, polyethylene, etc.). The composition thus formed is coated on a web of, for example, polyester or polyacrylonitrile fibers.

They may also be homogeneous membranes, obtainable by introducing functional groups into an inert matrix by chemical or radiochemical grafting. More widely used chemical methods generally involve functionalizing polymer latexes containing aromatic nuclei (e.g., styrene/divinylbenzene or styrene/butadiene). The scrim was coated with the latex so functionalized, similar to the case of heterogeneous films. Radiochemical methods generally involve grafting an aromatic compound (e.g., styrene) into an inert substrate (e.g., polyethylene or polytetrafluoroethylene film) under irradiation conditions. The aromatic nucleus is subsequently functionalized analogously to the chemical method.

Cation exchange membranes (cation membranes) contain strongly acidic groups (in most cases sulfonate groups) or weakly acidic groups (frequently carboxylate groups). The acidic group is rarely PO₃ ^2-、HPO₂ ^-、AsO₃ ^2-And SeO₃ ^-。

The anion exchange membrane (anion membrane) contains strongly basic groups (in most cases quaternary ammonium groups) or weakly basic groups (in most cases amino groups). The basic groups are rarely quaternary phosphonium or sulfonium groups.

In the method of the present invention, the cationic membrane preferably comprises strongly acidic groups, preferably sulfonate groups; the anionic membrane preferably comprises strongly basic groups, preferably quaternary ammonium groups.

Besides the membranes, the electrodialyser of course also comprises a cathode and an anode. The anode may be composed of materials commonly used for electrodialysis, such as graphite, titanium coated with noble metals or noble metal oxides, in particular titanium coated with platinum. The cathode may also be composed of materials commonly used for electrodialysis, such as graphite, stainless steel or nickel.

The solution to be treated is fed to an electrodialyser, said solution being at least partially an aqueous solution. It is necessary to circulate the anolyte solution at the anode and the catholyte solution at the cathode. These solutions often form a single electrolyte solution. In the process of the invention, it is very suitable to use a single electrolyte circuit. The function of the electrolyte solution is to ensure sufficient conductivity. The conductivity is preferably equal to or higher than 20mS/cm, this lower limit not being critical for the implementation of the process of the invention.

The electrolyte used is an ionizable compound, such as a salt, an acid or a base. The electrolyte is preferably selected from non-electroactive compounds, and therefore, it is preferred not to use chloride on an industrial scale, since chloride may generate chlorine at the anode.

Examples of electrolytes that may be mentioned include neutral salts, such as sulfates; acids such as sulfamic acid, water-soluble carboxylic acids and sulfuric acid. A catalyst metal salt, more particularly a cobalt salt (e.g., cobalt acetate), may also be used as the electrolyte.

The process according to the invention should avoid the use of electrolyte solutions whose pH may lead to the precipitation of metal compounds which one wishes to separate by means of electrodialysis, which is why the use of acidic electrolytes is preferred.

The voltage applied to the electrodialysers must avoid polarizing the system, i.e. dissociating the water under the action of a very strong electric field. Generally suitable voltages are from 0.5 to 2.5 volts per cell, preferably from 0.5 to 1.5 volts per cell. Turbulence of the liquid can be increased by using thin slots and a separate frame, thereby reducing polarization. The groove thickness is preferably 0.5mm to 2mm, more preferably 0.75mm to 1.5 mm.

The temperature at which the process of the invention is carried out should be maintained within a range compatible with the stability of the membrane. In fact, it is in principle advantageous to increase the temperature, which increases the fluidity of the electrolyte and reduces the viscosity of the solution to be treated, but increasing the temperature shortens the life of the membrane. Therefore, the operation is preferably carried out at a temperature lower than or equal to 50 ℃, more particularly from 20 ℃ to 40 ℃.

The electrodialyser can be operated in different ways. Firstly, the continuous operation (straight-through operation) can be carried out, and the solution to be treated continuously passes through the membrane stack; if the degree of treatment to be achieved so requires, a plurality of stages of individual electrodialysers can be arranged in series. It is also possible to carry out a batch operation (recirculation operation) in which the solution to be treated is recirculated in the tank until the desired degree of treatment is obtained. Finally, a partial recirculation pass-through operation is performed.

As previously mentioned, the reaction mixture containing the diacids in the presence of the homogeneous catalyst to be separated originates essentially from the oxidation of cycloalkanes to the corresponding diacids. Briefly, in the following description, the oxidation of cyclohexane to adipic acid is generally considered to produce smaller but still appreciable amounts of glutaric and succinic acids.

Before the treatment of such mixtures by electrodialysis, it is generally advantageous to carry out operations which make it possible in particular to isolate the majority of the adipic acid, the compound of interest of the preparation.

The isolation can be carried out by known methods, such as precipitation of adipic acid by cooling the mixture.

In order to be able to carry out the electrodialysis according to the invention, the remaining solution can be collected with water, as the case may be, after partial or complete removal of the organic solvent which may be present.

The solution subjected to electrodialysis may generally contain from 0.0001 to 1mol of catalyst per kg of solution, from 0.001 to 1mol of glutaric acid per kg of solution, from 0.001 to 1mol of succinic acid per kg of solution and from 0.001 to 1mol of adipic acid per kg of solution.

The examples which follow illustrate the invention. Example 1

The electrodialyser used had an active surface of 2dm²Each comprising a chamber into which a solution to be treated is introduced (chamber D, the solution being diluted in a catalyst) and a chamber C which receives the catalyst during electrodialysis.

The membranes separating each compartment D from the adjacent compartment C are:

-a Neosepta AMX brand anionic membrane containing quaternary ammonium groups,

-a sulfonate-containing Neosepta CMX brand cationic membrane.

The electrolyte consists of an aqueous solution of sulfamic acid, the conductivity of which at 20 ℃ is 20 mS/cm. The solution was circulated at a flow rate of 400l/h and a volume of 2 l.

The volume of the aqueous solution to be treated was 1.6l, this solution initially containing:

0.18mol/kg of adipic acid

Glutaric acid 0.87mol/kg

0.36mol/kg succinic acid

0.27mol/kg of cobalt in the form of cobalt acetate.

A batch operation method (recycling operation) is used.

The circulation flow rate of this solution in chamber D was 180 l/h.

The solution flowing in the chamber C and intended to receive the cobalt salt is initially an aqueous solution of sodium chloride at a concentration of 5g/l, a volume of 1.6l and a flow rate of 180 l/h.

The initial conductivity of each solution in compartment C was 10 mS/cm.

The electrodialysis was carried out under the application of 18V.

To follow the progress of the operation, different solution samples were taken periodically. The amount of cobalt was determined by atomic absorption spectroscopy, the amount of diacid by gas chromatography, followed by determination of the pH, conductivity and volume of the solution.

The results of the solution pH, conductivity and volume measurements are given in table 1 below.

The concentration results for the different samples in mol/kg are given in table 2 below. By definition, L1 denotes the chamber D solution, also referred to as the feed solution; l2 denotes a cell C solution, also called concentrated solution; l3 denotes an electrolyte solution.

From the results of table 1, it can be determined that the amount of water in the feed solution is reduced. This can be explained on the basis of the following facts: the migrating ions are hydrated; at the end of the test, the conductivity of the feed solution and the concentrated solution differed greatly (dialysis phenomenon).

To achieve accurate results, the values in table 2 take into account the volume change. The results of Table 2 are expressed on the basis of 1Kg of the initial fluid to be treated.

The symbol "-" in the table means that it was not measured.

TABLE 1

Time min	Sample solution	Temperature of	pH	Conductivity mS/cm	Volume ml
						0	L1	20.0	3.5	10.7	1600
0	L2	20.0	4.6	10.1	1600
						0	L3	20.0	-	20.0	2000
18	L1	-	3.5	9.6	1567
						20	L2	-	4.7	14.1	1639
38	L1	-	3.2	7.6	1530
						40	L2	27.2	4.7	16.4	1677
57	L1	28.4	2.9	5.1	1495
						58	L2	-	4.9	17.8	1712
80	L1	-	2.1	1.9	1416
						80	L2	-	4.8	18.8	1793
86	L1	30.2	2.0	1.8	1405
						86	L2	-	4.8	18.9	1805

TABLE 2

Time min	Sample solution	Comol/kg	Adipic acid mol/kg	Glutaric acid mol/kg	Succinic acid mol/kg
						0	L1	0.267	0.183	0.869	0.355
0	L2	0	0	0	0
						0	L3	0	0	0	0
0	All are	0.267	0.183	0.869	0.355
						18	L1	0.192	0.184	0.867	0.340
20	L2	0.066	0	0	0.002
						20	All are	0.258	0.184	0.867	0.342
38	L1	0.137	0.175	0.822	0.316
						40	L2	0.127	0.005	0.050	0.036
40	All are	0.264	0.180	0.872	0.352
						57	L1	0.064	0.174	0.827	0.306
58	L2	0.193	0	0.044	0.076
						58	All are	0.257	0.174	0.871	0.382
80	L1	0.005	0.149	0.753	0.271
						80	L2	0.251	0	0.082	0.080
80	All are	0.256	0.149	0.835	0.351
						86	L1	0.0005	0.163	0.774	0.269
86	L2	0.244	0	0.081	0.078
						86	L3	0.001	0	0	0.004
86	All are	0.245	0.163	0.855	0.351

EXAMPLE 2 Oxidation of cyclohexane to adipic acid

The object of this example is to prepare a solution for use in a process for the electrodialysis separation of a catalyst.

After first purging with nitrogen at ambient temperature, a 1.5 liter jacketed autoclave was charged with the following composition (the autoclave was made of titanium, fitted with turbines and various openings for introducing reactants and fluids or for discharging reaction products or fluids):

-cobalt acetate tetrahydrate: 4.0g

-acetic acid: 359g

-cyclohexane: 289.7g

-acetaldehyde: 1.2g

After the autoclave was closed, the nitrogen pressure was increased to 20bar (2MPa), stirring was started (800 revolutions per minute) and the temperature was increased to 120 ℃ in 29 minutes. The nitrogen was then replaced by 20bar of air containing 4% oxygen and the flow rate of the outlet gas was controlled at 250 l/h.

After an induction time of 10 minutes, at which time no oxygen was consumed, the temperature rose abruptly to 106 ℃ and oxygen consumption began. The oxygen content of the air at the inlet was raised to 16% by means of a flow meter system. The oxygen content at the autoclave outlet was maintained below 5% throughout the experiment, and the average temperature of the autoclave was maintained at 106-.

After 50 liters of oxygen had been consumed (corresponding to a cyclohexane conversion of about 20%), cyclohexane (4.3ml/min) and an acetic acid solution containing 1.1% by weight of cobalt acetate tetrahydrate (flow rate of 3.9ml/min) were continuously injected.

The above injection addition was continued until a reaction mixture comprising about 5700g of adipic acid and comprising 35.3kg of an acetic acid phase and 10.3kg of a cyclohexane phase was obtained.

The liquid content in the autoclave was kept constant by means of a level probe. The reaction mixture was recovered in a glass vessel heated to 70 ℃ using a servo-controlled pneumatic bottom valve.

The two-phase separation of the reaction mixture obtained is carried out at 70 ℃.

The acetic acid phase was concentrated to give about 19kg of material. Adipic acid was crystallized, separated by filtration and recrystallized in water to give 4.2kg of purified adipic acid.

The mixture of acetic acid and the aqueous solution resulting from the crystallization and recrystallization of adipic acid was about 11.5 kg. The mixture was concentrated to about 50% of the original mass and then diluted with about twice the amount of water. Removing part of cyclohexanone, cyclohexanol and ester compounds by settling and separation.

This gave a dilute acetic acid solution containing the following components: cobalt 0.4485% (weight ratio), acetic acid 193g/kg solution, water 626g/kg solution, adipic acid 41g/kg solution, glutaric acid 27.9g/kg solution, succinic acid 13.3g/kg solution, hydroxycaproic acid 4.8g/kg solution, hydroxyadipic acid 9.4g/kg solution, cyclohexanone 10.6g/kg solution, cyclohexanol 5.7g/kg solution, cyclohexyl acetate 3.4g/kg solution, butyrolactone 6.4g/kg solution, valerolactone 0.8g/kg solution, various cyclohexyl esters 41.2mmol/kg solutions. Example 3

The electrodialyser used had an active surface of 2dm²Each comprising a compartment (compartment D) into which the dilute acetic acid solution to be treated prepared in example 2 is introduced, said solution being diluted in a catalyst, and a compartment C which receives the catalyst during the electrodialysis.

The membranes separating each compartment D from the adjacent compartment C are:

-a Neosepta AMX brand anionic membrane containing quaternary ammonium groups,

-a sulfonate-containing Neosepta CMX brand cationic membrane.

The electrolyte consists of an aqueous solution of sulfamic acid, the conductivity of which at 20 ℃ is 20 mS/cm. The solution was circulated at a flow rate of 400l/h and a volume of 2 l.

The volume of the dilute acetic acid solution to be treated was 1.6l (composition as shown in example 2).

A batch operation method (recycling operation) is used.

The circulation flow rate of this solution in chamber D was 180 l/h.

The solution flowing in the chamber C and receiving the cobalt salt is initially an aqueous solution of cobalt acetate tetrahydrate having a concentration of 10g/l, an initial conductivity of from 3.5 to 4mS/cm, a volume of 1.6l and a flow rate of 180 l/h.

The electrodialysis was carried out under the application of 18V.

To follow the progress of the operation, different solution samples were taken periodically. The amount of cobalt was determined by atomic absorption spectroscopy, and the amount of diacid and the amount of other organic compounds were determined by gas chromatography. The pH, conductivity and volume of the solution were then determined.

As already reported in table 1, the amount of water in the feed solution was reduced. This can be explained on the basis of the following facts: the migrating ions are hydrated; at the end of the test, the conductivity of the feed solution and the concentrated solution differed greatly (dialysis phenomenon).

To achieve accurate results, the values in tables 3 and 4 take into account the change in volume. These results are expressed on the basis of 1kg of the fluid to be treated.

The results for the Co catalyst are given in table 3. Table 4 gives the results for the organic compounds present in the solution to be treated.

TABLE 3

Time min	Sample solution	Comol/kg	Initial Co in Co%/L1
				0	L1	0.070	100
0	L2	0.026	0
				0	L3	0	0
0	All are	0.096
				8	L1	0.037	53
9	L2	0.043	24.5
				17	L1	0.017	24.5
18	L2	0.061	50
				25	L1	0.004	6
26	L2	0.075	70
				28	L1	0.002	3
28	L2	0.088	88.5
				28	L3	0.006	8.5
28	All are	0.096

Testing of the mixture derived from the reaction test for oxidation of cyclohexane to adipic acid confirms the effectiveness of the Co catalyst separation by electrodialysis.

TABLE 4

Compound (I)	Relative to the remainder of the initial solution L1%
				Time 0min	Time period of 17min	The time is 28min
	Adipic acid	100	97.5				98.5
Glutaric acid				100	95	93.5
	Succinic acid	100	91.5				88
Hydroxyhexanoic acid				100	100	100
	Hydroxy adipic acid	100	89				88
Cyclohexanol				100	96	100
	Acetic acid cyclohexyl ester	100	91				97.5
Cyclohexanone				100	97	98
	Butyrolactone	100	91				79
Other cyclohexyl esters				100	100	100
	Acetic acid	100	96				96

The organic compounds remain predominantly in the feed solution. Example 4

Using the same electrodialyser as in example 3, a dilute acetic acid solution to be treated was prepared as in example 2.

The solution which flowed in the cell C and received the cobalt salt was initially a portion of the solution L2 obtained in the solution electrodialysis test from example 2, from which the electrolyte solution L3 also originated.

All application conditions were the application conditions in example 3.

The purpose of this test is to show that the amount of Co in the concentrated solution in the C compartment can be increased continuously.

As already reported in example 1, the amount of water in the feed solution is reduced. This can be explained on the basis of the following facts: the migrating ions are hydrated; at the end of the test, the conductivity of the feed solution and the concentrated solution differed greatly (dialysis phenomenon).

To achieve accurate results, the values in tables 5 and 6 take into account the change in volume. These results are expressed on the basis of 1kg of the fluid to be treated.

Table 5 gives the results for the Co catalyst; table 6 gives the results for the organic compounds present in the solution to be treated.

TABLE 5

Time min	Sample solution	Comol/kg	Initial Co in Co%/L1
				0	L1	0.075	100
0	L2	0.086	0
				0	L3	0.010	0
0	All are	0.171
				7	L1	0.045	60
8	L2	0.098	15.5
				16	L1	0.032	43
17	L2	0.120	46
				25	L1	0.012	16
26	L2	0.133	63
				30	L1	0.005	7
30	L2	0.153	89
				30	L3	0.017	9
30	All are	0.175

This example using a solution of Co that has been concentrated in cell C confirms the effectiveness of the Co catalyst separation by electrodialysis and the possibility of obtaining a more Co-concentrated solution.

TABLE 6

Compound (I)	Relative to the remainder of the initial solution L1%
				Time 0min	For a period of 16min	For a period of 30min
	Adipic acid	100	104				95
Glutaric acid				100	99.5	91.5
	Succinic acid	100	95				83.5
Hydroxyhexanoic acid				100	100	99
	Hydroxy adipic acid	100	100				88
Cyclohexanol				100	99	100
	Acetic acid cyclohexyl ester	100	92.5				100
Cyclohexanone				100	97	100
	Butyrolactone	100	100				100
Other cyclohexyl esters				100	100	93
	Acetic acid	100	95				95

Claims (10)

1. Method for separating a homogeneous metal catalyst dissolved in a mixture also containing at least one fatty diacid, characterized in that the catalyst comprises cobalt and the separation is carried out by membrane electrodialysis.

2. Process according to claim 1, characterized in that the catalyst is selected from cobalt-containing catalysts, the cobalt being present alone or in admixture with other metals, such as manganese, copper, iron, vanadium or cesium or mixtures thereof.

3. The process according to claim 2, characterized in that the metal is present in the form of compounds dissolved in the oxidation reaction mixture of the cycloalkane, such as hydroxides, oxides and organic or inorganic salts.

4. A process according to any one of claims 1 to 3, characterised in that the catalyst is selected from cobalt salts, which may be present alone or in admixture with other compounds of metals, such as manganese and/or copper and/or iron and/or caesium and/or vanadium.

5. Process according to any one of claims 1 to 4, characterized in that the mixture subjected to membrane electrodialysis contains at least one diacid formed in the oxidation of cycloalkanes and one or more other diacids formed as by-products, preferably adipic acid and glutaric and succinic acids.

6. Process according to any one of claims 1 to 5, characterized in that the mixture in which the homogeneous catalyst is present comprises water, optionally a solvent which has been used in the formation of the solution to be treated and which can be replaced in its entirety or in part by water before the electrodialysis is carried out.

7. Process according to claim 6, characterized in that the water is 10% to 100% of the solvent mixture of the solution subjected to electrodialysis, preferably 50% to 100% of the solvent mixture.

8. Process according to any one of claims 1 to 7, characterized in that the electrodialysis device used comprises a stack of a large number of cells, each cell consisting of two adjacent concentrating compartments (C) and diluting compartments (D), said compartments (C) and (D) being alternately joined by cationic and anionic membranes.

9. Method according to any one of claims 1 to 8, characterized in that the membrane consists of a matrix grafted with functional groups which in the case of cationic membranes are anionic groups, such as sulphonate groups, and in the case of anionic membranes are cationic groups, such as quaternary ammonium groups.

10. Process according to any one of claims 1 to 9, characterized in that the solution subjected to electrodialysis contains from 0.0001 to 1mol of catalyst per kg of solution, from 0.001 to 1mol of glutaric acid per kg of solution, from 0.001 to 1mol of succinic acid per kg of solution and from 0.001 to 1mol of adipic acid per kg of solution.