The present invention relates to a method for fractionating a solution into two or more fractions en- riched with different components. In particular, the invention relates to a method for fractionating a solution by a chromatographic simulated moving bed (SMB) process in which the liquid flow is effected in a system comprising one or more columns containing one or more partial packed beds. The dissolved substances present in the feedstock are separated in the partial packed beds, and a separation

(i.e., dry solids) profile is formed. The columns/partial packed beds form one or more loops. The novel separation system comprises at least two separation profiles in the same loop.

A separation profile is formed by the feed solu¬ tion and the dry substance recirculated. The separation profile is a complete or an essentially complete dry solids profile. The novel chromatographic SMB method is capable of essentially improving the SMB separation capacity.

Fractionation of a solution comprising many dis¬ solved substances into fractions enriched with different components is often necessary in order to recover the desired components optimally pure. The method of the invention can be employed to carry out such fractionation. A sulphite cooking liquor, for instance, can be fraction¬ ated by the method so as to give a fraction rich in mono- saccharides and/or a fraction rich in lignosulphonates . Furthermore, molasses or vinasse can be fractionated in this way to obtain fractions rich in a sugar, such as sucrose, and/or betaine, for instance.

The method of the invention is particularly well suitable for separating substances that are difficult to separate from mixtures containing them. Such mixtures in-

elude fructose/glucose syrups, beet-derived juices, invert sugar mixtures, starch hydrolysates, wood hydrolysates, milk whey solutions and other lactose-containing solu¬ tions, solutions containing amino acids, fermentation broths containing various organic acids, such as citric acid, bagasse hydrolysates, and particularly solutions containing inositol, mannitol, sorbitol, xylitol, erythritol, glutamic acid and/or glycerol .

Continuously operated chromatographic separation processes presently commonly employ the simulated moving bed method, which is used in a variety of different applications. The simulated moving bed method may be sequ¬ ential or continuous or comprise a combination of a con¬ tinuous and a sequential method. In a continuous simulated moving bed process, all fluid streams typically flow con¬ tinuously. The streams are: the supply of feed solution and eluent, the circulating of the liquid mixture, and the withdrawal of products. The flow rate for these flows may be adjusted in accordance with _ the separation goals (yield, purity, capacity) . Normally 8 to 20 partial packed beds are combined into a loop. The eluent and feed supply and product ^" withdrawal points are shifted cyclically in the downstream direction in the packing material bed. On account of the supply of eluent and feed solution, the withdrawal of products, and the flow through the packing material bed, a dry solids profile is formed in the pack¬ ing material bed. Constituents having a lower migration rate in the packed bed are concentrated in the back slope of the separation profile, i.e. dry solids profile, while constituents having a higher migration rate are con¬ centrated in the front slope. The points of introduction of the feed solution and eluent and the withdrawal points of the product or products are shifted cyclically at sub ^¬ stantially the same rate at which the dry solids profile moves in the packing material bed. The eluent and feed

supply and product withdrawal points are shifted cyclic¬ ally by using feed and product valves located along the packing material bed, typically at the upstream and down¬ stream end of each partial packed bed. If product frac- tions of very high purity are desired, short cycle times and multiple partial packed beds must be employed (the apparatus has the requisite valves and feed and withdrawal equipment)

In the sequential simulated moving bed process, some of the fluid streams do not flow continuously. The streams are: the supply of feed solution and eluent, the circulating of the liquid mixture, and the withdrawal of products (eluting phase; two to four or more products) . The flow rate and the volumes of the different feeds and product fractions may be adjusted in accordance with the separation goals (yield, purity, capacity) . The process commonly comprises three basic phases: feeding, elution and circulation During the feeding phase, a feed solution, and possibly also an eluent during a simul- taneous eluting phase, is introduced into predetermined partial packed beds, and simultaneously a product fraction or fractions are withdrawn. During the eluting phase, eluent is introduced into a predetermined partial packed bed or predetermined partial packed beds, and during these phases two, three or even four product fractions are with¬ drawn. During the circulating phase, no feed solution or eluent is supplied to the partial packed beds and no products are withdrawn.

The method of the invention may be either continu- ous or sequential, or a combination of a continuous and a sequential process.

The continuous simulated moving bed process has been disclosed in U. S. Patent 2 985 589 (Broughton et al . ) , for example. In accordance with this process, the mixture to be fractionated is introduced into one partial

packed bed and eluent is introduced into another partial packed bed, and two product fractions are withdrawn substantially simultaneously. There are at least four par¬ tial packed beds, forming a single loop with continuous circulation, and the feed and product withdrawal points are shifted cyclically in the downstream direction in the packing material bed. A similar method is described in U. S. Patent 4 412 866 (Schoenrock et al . ) .

Sequential simulated moving bed processes are described in British application 2 240 053 and U. S. Pat¬ ents 4 332 623 (Ando et al . ) , 4 379 751 (Yoritomi et al . ) and 4 970 002 (Ando et al . ) , for instance. A sequential simulated moving bed process applied to the recovery of betaine and sucrose from beet molasses is described in the Applicants' Finnish Patent 86 416 (U. S. Patent 5 127 957) . In these methods, only one complete or essen¬ tially complete dry solids profile is circulated in the partial packing material loop. Also the Applicants' co- pending Finnish applications 930 321 (filing date January 26, 1993) and 932 108 (filing date May 19, 1993) relate to a sequential simulated moving bed method, the first applied to the fractionation of molasses and the latter tro the fractionation of sulphite cooking liquor. As is described in these applications, the simulated moving bed method may include multiple loops; yet a single dry solids profile is circulated in each loop.

Finnish Patent 86 416 (U. S. Patent 5 127 957) referred to above discloses a method for recovering beta¬ ine and sucrose from beet molasses employing a sequential simulated moving bed process. The chromatographic system comprises at least 3 chromatographic partial packed beds in series. In the method, betaine and sucrose are sep¬ arated during the same sequence comprising a molasses feeding phase wherein the molasses feedstock is supplied to one of said partial packed beds and eluent water is

supplied substantially simultaneously to another of said partial packed beds, an eluent feeding phase, and a cir¬ culating phase. These are repeated either once or several times during the sequence. In the method disclosed in the above-stated

Finnish application 930 321, the liquid flow is effected in a system comprising at least two partial packed beds, and the product or products are recovered during a multi- step sequence. _A sequence comprises feeding, eluting and circulating phases. During the circulating phase, the liquid present in the partial packed beds with its dry solids profile is circulated in two or more loops compris¬ ing one, two or more partial packed beds. A loop may be closed or "open", in other words, when liquid is circu- lated in one loop, eluent can be introduced into the other loop and a product fraction can be withdrawn therefrom. During the feed and elution, the flow through the packing material beds may take place between successive loops, wherein the flows carry material from one loop to another. During the circulating pha-se, the loop is closed and sep¬ arated from the other loops. Only one dry solids profile is circulated in each loop.

The Applicants' Finnish application 941 866 dis¬ closes a simulated moving bed method for the continuous fractionation of solutions, employing ion exchange resins of two or more different ionic forms, so that the dry solids profile formed upon passage of the solution through a chromatographic packing material having a first ionic form is passed to a chromatographic packing material hav- ing a second ionic form without the partially separated components being remixed, and/or that the concentration and pumping stages of the solution, included in the prior art methods for fractionating solutions with packed beds of two different ionic forms, can be avoided. U. S. Patent 5 198 120 (Masuda et al . ) discloses

a method for fractionating a ternary or multi-component solution by a simulated moving bed method comprising a series of several columns. A circulation shut-off valve is "in between" the column series. The solution to be frac- tionated is supplied to the column located immediately after the shut-off valve in the downstream direction, and simultaneously one or more product fractions are withdrawn from a column located upstream. During a combined eluting and circulating phase, the solution is circulated in a loop comprising the entire column series.

European application 663 224 (applicant Mitsu¬ bishi) discloses a method for fractionating a ternary or multi-component solution by a simulated moving bed method comprising a series of four columns. In this method, a loop may comprise two to four columns; yet only one dry solids profile is circulated in a loop.

The present invention relates to a method for fractionating a solution by a chromatographic simulated moving bed process in which the liquid flow is effected in a system comprising one or more columns containing one or more partial packed beds. The dissolved substances present in the feedstock are separated in the partial packed beds, and a separation (i.e.,- dry solids) profile is formed. The columns/partial packed beds form one or more loops. The novel method is characterized in that the separation sys¬ tem comprises at least two separation profiles in the same loop.

A separation profile is formed by the feed solu¬ tion and the dry substance recirculated. The separation profile comprises all constituents present in the feed¬ stock, i.e. constituents having a low migration rate, con¬ stituents having an intermediate migration rate, and con¬ stituents having a high migration rate. Accordingly, the separation profile is a complete or an essentially com- plete dry solids profile. Preferably, part of the con-

stituent having the highest migration rate is withdrawn prior to the circulation phase.

The novel chromatographic SMB method is capable of essentially improving the SMB separation capacity. Compared with the earlier chromatographic SMB processes, for example, the novel SMB method has afforded a separa¬ tion capacity improved by several dozen per cent for ins¬ tance in molasses separation, while the product fraction performance characteristics have remained substantially. the same.

The chromatographic apparatus employed comprises one column or several columns connected in series, fluid conduits connecting the columns, solution and eluent con¬ tainers, feed and eluent conduits, recycle and feed pumps, heat exchangers, product fraction withdrawal conduits, and valves, flow and pressure regulators and on-line concen¬ tration, density, optical activity and conductivity meters. The process is proceeded with in a state of equi¬ librium, and the progress of the separation process is monitored with a density meter. The separation is con¬ trolled by a microprocessor, which controls the volume flow rates and volumes of feeds, employing quantity/volume measuring devices, temperature controllers, valves and pumps. The number of columns is 1 to 20, preferably 2 to

8. Preferably, a column comprises one or several separate partial packed beds.

A strong acid ion exchange resin, such as Finex V09C (manufacturer Finex Oy) , Finex V13C (manufacturer Finex Oy) , Finex CS 11 GC (manufacturer Finex Oy) or Purolite PCR 651 (manufacturer Purolite Co) is preferably used as a column packing material .

The eluent employed is preferably water.

The flow rate may range from 0.5 to 15 m ³ /hr./m ² ; preferably the flow rate is 3 to 10 mVhr./m ² .

Example 1

Two-profile separation from molasses

The column series employed for the separation com¬ prised eight separate partial packed beds, four of which formed a column in each case, i.e. there were two columns. Product fractions were withdrawn from column 1 and column 2, residual fractions were withdrawn from columns 1 and 2, but a sucrose and betaine fraction were withdrawn from column 2 only, as were_ the recycle fractions. The test conditions are shown in Table 1A.

Table 1A

Test conditions

Separation temperature 85°C

Resin bed height 14 m (7 m/column)

Column diameter 11.1 cm

A two-profile separation sequence was produced in the column series, wherein two separate separation pro¬ files, i.e. essentially complete dry solids profiles, were simultaneously circulated in the separation process. The fractionation was performed in an eight-step sequence. The sequence had a cycle length of 38 minutes and comprised the following steps:

Step 1: 1.5 litres of feed solution was supplied to column 1 at a volumetric flow rate of 50 0/hr., and a corresponding amount of a recycle fraction was eluted from column 2.

Step 2: 5.5 litres of feed solution (continuation of feeding phase) was supplied to column 1 at a volumetric flow rate of 50 {/hr., and a corresponding amount of a sucrose fraction was eluted from column 2. Step 3: 3.2 litres of eluent water was supplied to

column 1 at a volumetric flow rate of 60 {/hr., and a cor¬ responding amount of the remaining part of sucrose frac¬ tion was withdrawn (from column 2) .

Step 4 : The supply of eluent water to column 1 was continued (3.3 litres; 60 {/hr.) , and a corresponding amount of a recycle fraction was eluted from column 2.

Step 5 : The supply of eluent water to column 1 was continued (5.0 litres; 60 {/hr.) , and a corresponding amount of a betaine fraction was eluted from column 2. Step 6 : The supply of eluent water to column 1 was continued (13.0 litres; 68 {/hr.) , and a corresponding amount of a residual fraction was eluted from the same column. Simultaneously eluent water was also supplied to column 2 (10.0 litres; 73 { /hr. ) , and a corresponding amount of the remaining part of betaine fraction was with¬ drawn from the same column.

Step 7 : The supply of eluent water to column 2 was continued (2.5 litres; 73 { /hr. ) , and a corresponding amount of a residual fraction was eluted from the same column.

Step 8: Circulation in the loop formed by columns 1 and 2 (6.0 litres; 75 {/hr.) .

The sequence containing these steps was repeated eight times to equilibrate the system, whereafter the method proceeded in a state of equilibrium. In the state of equilibrium, samples were collected during one sequence .

The composition of the raw material of the feed solution is shown in Table IB.

Table IB

Analysis of raw material

Composition, % of dry solids

Tri- Sucrose Monosaccha- Betaine Calcium saccha- rides rides - 2.8 57.6 0.6 7.6 < 0.04

Dry solids content g/100 g 52.7 pH 9.4

Conductivity mS/cm 13.7

The separation resin employed as a packing mater¬ ial was Finex V09C, which was initially regenerated with sodium chloride (during the fractionation it was equi¬ librated by cations from the molasses) . The analysis of the separation resin is shown in Table 1C.

Table 1C

Analysis of separation resin

Make Finex V09C

DVB % 5.5

Capacity equiv/{ 1.56 Mean bead size mm 0.360

In equilibrium, the combined residual fraction had a dry solids content of 6.7 g/100 g, and the sucrose purity was 12.4% by weight. The sucrose fraction had a dry solids content of 19.1 g/100 g, and the sucrose purity was 90.1% by weigh. The betaine fraction had a dry solids

content of 4.4 g/100 g, and the betaine purity was 43.1% by weight .

Example 2

Two-profile separation from vinasse

The column series employed for the separation com¬ prised three separate separation columns. Residual frac¬ tions were withdrawn from all columns, and a betaine fraction was withdrawn from column 3 only.

The test conditions are shown in Table 2A.

Table 2A

Test conditions

Separation temperature 80°C

Resin bed height 10.5 m (3.5 m/column) Column diameter 20 cm

A two-profile separation sequence was produced in the column series, wherein two separate separation pro¬ files, i.e. essentially complete dry solids profiles were simultaneously circulated in the separation process. The fractionation was performed in a 7-step sequence. The sequence had a cycle length of about 58 minutes and com- prised the following steps:

Step 1: 7 litres of feed solution was supplied to column 1 at a volumetric flow rate of 150 {/hr., and a corresponding amount of a residual fraction was eluted from column 2. Simultaneously 7 litres of eluent water was introduced into column 3 at a volumetric flow rate of 150 {/hr., and a corresponding amount of a betaine fraction was eluted from the same column.

Step 2: 5 litres of feed solution (continuation of feeding phase) was supplied to column 1 at a volumetric flow rate 150 {/hr., and a corresponding amount of a

betaine fraction was eluted from column 3.

Step 3: 29 litres of feed solution was supplied to column 1, and a corresponding amount of residual fraction was eluted from the same column. Simultaneously 17 litres of eluent water was introduced into column 2 at a volu¬ metric flow rate of 90 { /hr . , and a corresponding amount of the remaining part of betaine fraction was withdrawn from column 3.

Step 4 : The supply of eluent water to column 2 was continued (6 litres; 80 { /hr . ) and a corresponding amount of residual fraction was eluted from column 1.

Step 5: Circulation in the loop formed by columns 1-3 (10 litres; 100 { /hr . ) .

Step 6 : Eluent water was introduced into column 1 (33 litres; 100 {/hr.) , and a corresponding amount of a residual fraction was eluted from column 3.

Step 7 : Eluent water was introduced into column 3 (28 litres; 150 {/hr.) , and a corresponding amount of a residual fraction was eluted from column 2. A sequence comprising these steps was repeated eight times to equilibrate the system, whereafter the method proceeded in a state of equilibrium. In the state of equilibrium, samples - were collected during one sequence . The composition of the raw material of the feed solution is shown in Table 2B.

Table 2B

Analysis of raw material

Composition, % of dry solids

Betaine 13.8

Dry solids content g/100 g 53.8 pH 7.1

Conductivity mS/cm 46.5

The separation-resin employed as a packing material was Finex V13C, which was initially regenerated with sodium chloride (during the fractionation it was equi¬ librated by cations from the vinasse) . The analysis of the separation resin is shown in Table 2C.

Table 2C Analysis of separation resin

Make Finex V13C

DVB % 8.0 Capacity equiv/{ 1.8

Mean bead size mm 0.34

In equilibrium, the betaine fraction had a dry solids content of 22.4%, and the betaine purity was 48.3% by weight. The combined residual fraction had a dry solids content of 18.4%, and the betaine purity was 2.3% by weight .

Example 3

Two-profile separation from glucose/fructose mixture

The column series employed for the separation com- prised four separate separation columns. Product fractions were withdrawn from all columns.

The test conditions are shown in Table 3A.

Table 3A Test conditions

Separation temperature 65°C

Resin bed height 11.2 m (2.8 m/column) Column diameter 20 cm

A two-profile separation sequence was produced in the column series, wherein two separate separation pro¬ files, i.e. essentially complete dry solids profiles were simultaneously circulated in the separation process. The fractionation was performed in a four-step sequence. The sequence had a cycle- length of 74 minutes and comprised the following steps:

Step 1: 18 litres of a feed solution was supplied to columns 1 and 3 at a volumetric flow rate of 120 {/hr., and a corresponding amount of a glucose fraction was eluted from the same columns.

Step 2: 8 litres of eluent water was supplied to columns 2 and 4 at a volumetric flow rate of 120 { /hr. , and a corresponding amount of a glucose fraction was eluted from columns 3 and 1.

Step 3: 30 litres of eluent water was supplied to columns 2 and 4 , and a corresponding amount of a fructose fraction was eluted from the same columns. Step 4: Circulation in the loop formed by columns

1-4 (106 litres; 130 {/hr.) .

The sequence containing these steps was repeated eight times to equilibrate the system, whereafter the method proceeded in a state of equilibrium. In the state of equilibrium, samples were collected during one sequence.

The composition of the raw material of the feed solution is shown in Table 3B.

Table 3B

Analysis of raw material

Composition, % of dry solids

Glucose 49.2 Fructose 49.5

Dry solids content g/100 g 50

The separation resin employed as a packing material was Finex CS 11 GC, which was regenerated with calcium chloride. The analysis of the separation resin is shown in Table 3C.

Table 3C

Analysis of separation resin

Make Finex CS 11 GC DVB % 5.5

Capacity equiv/{ 1.5

Mean bead size mm 0.310

In equilibrium, the combined glucose fraction had a dry solids content of 23.3 g/100 g, and the glucose purity

was 96.7% by weight. The combined fructose fraction had a dry solids content of 20.3 g/100 g, and the fructose purity was 99.7%.

Example 4

Three-profile separation from molasses

The column series employed for the separation com¬ prised three separate separation columns. Residual frac¬ tions were withdrawn from all columns, and sucrose, re- cycle and betaine fractions were withdrawn from column 1. The test conditions are shown in Table 4A.

Table 4A

Test conditions

Separation temperature 80°C

Resin bed height 15.3 m (5.1 m/column)

Column diameter 20 cm

A three-profile separation sequence was produced in the column series, wherein three separate separation pro¬ files, i.e. essentially complete dry solids profiles were simultaneously circulated in the separation process. The fractionation was performed in an eight-step sequence. The sequence had a cycle length of 43 minutes and comprised the following steps :

Step 1: 9 litres of a feed solution was supplied to column 2 at a volumetric flow rate of 110 { /hr . , and a corresponding amount of a recycle fraction was withdrawn from column 1.

Step 2: 11 litres of a feed solution (continuation of feeding phase) was supplied to column 2 at a volumetric flow rate of 110 {/hr. , and a corresponding amount of a sucrose fraction was eluted from column 1.

Step 3 : 16 litres of eluent water was supplied to column 2 at a volumetric flow rate of 110 { /hr. , and a corresponding amount of sucrose fraction was eluted from column 1. Step 4: 5.5 litres of eluent water was supplied to column 2 at a volumetric flow rate of 110 {/hr., and a corresponding amount of a recycle fraction was withdrawn from column 1.

Step 5: The_supply of eluent water to column 2 was continued (9 litres; 110 {/hr.) , and a corresponding amount of a betaine fraction was eluted from column 1.

Step 6: Eluent water was introduced into columns 1,

2 and 3 (5 litres into each at a volumetric flow rate of

120 {/hr.) ,- a corresponding amount of a betaine fraction was eluted from column 1 and a corresponding amount of residual fractions was eluted from columns 2 and 3.

Step 7: The supply of eluent water to columns 1, 2 and 3 was continued (15 litres at 100 {/hr. to column 1, 21 litres at 140 {/hr. to column 2, 21 litres at 140 {/hr. to column 3) , and a corresponding amount of residual frac¬ tions was eluted from columns 1, 2 and 3.

Step 8 :- Circulation in the loop formed by columns 1-3 (4 litres; 120 {/hr.) .

The sequence containing these steps was repeated eight times to equilibrate the system, whereafter the method proceeded in a state of equilibrium. In the state of equilibrium, samples were collected during one sequence.

The composition of the raw material of the feed solution is shown in Table 4B.

Table 4B

Analysis of raw material

Composition, % of raw material

Tri- Sucrose Betaine Calcium saccha- rides 3.9 60.4 5.1 < 0.04

Dry solids content g/100 g 58.5 Solids < 0.1 vol. % pH 9.0 Conductivity mS/cm 23.1

The separation resin employed as a packing material was Finex V09C, which was initially regenerated with sodium chloride (during the fractionation it was equi- librated by cations from the molasses) . The analysis of the separation resin is shown in Table 4C.

Table 4C

Analysis of separation resin

Make Finex V09C

DVB % 5.5

Capacity equiv/{ 1.5 Mean bead size mm 0.310

In equilibrium, the combined sucrose fraction had a dry solids content of 22 g/100 ml, and the sucrose purity was 92.7% by weight. The combined betaine fraction had a dry solids content of 5.2 g/100 ml, and the betaine purity

was 36.6% by weight. The combined residual fraction had a dry solids content of 8.2 g/100 ml.

Example 5

Two-profile continuous separation from molasses The column used for the separation comprised 14 separate partial packed beds between which a liquid per¬ vious but resin-retaining exchange plate was provided; a feeding and withdrawal apparatus was disposed on top of the exchange plate. The column had a diameter of 0.2 m, and the bed had a height of 14 m (1 m per partial packed bed) . The step length was 310 s, which for the feed was divided into two parts, i.e. feeding 280 s and rinsing of the pipe system and the feed and withdrawal apparatus 30 s. The cycle length was 4 340 s. The separation resin employed as a packing material was Purolite PCR 651, which was initially regenerated with sodium chloride (during the fractionation it was equi¬ librated by cations from the molasses) . The analysis of the separation resin is shown in Table 5A.

Table 5A

Analysis of separation resin

Make Purolite PCR 651

DVB % 5.5

Capacity equiv/{ 1.5

Mean bead size mm 0.340

The composition of the raw material of the feed solution is shown in Table 5B.

Table 5B

Analysis of raw material

Composition, % of dry solids Sucrose 59

Dry solids content g/100 g 59 pH 9.2 Conductivity mS/cm 13.8

The sequence comprised 14 steps. The situation dur¬ ing one step was the following:

Sucrose fractions were withdrawn from partial packed beds 2 and 9, and residual fractions were withdrawn from partial packed beds 6 and 13.

Feed solution and rinsing solution were introduced into partial packed beds 5 and 12, and eluent water was introduced into partial packed beds 1 and 8. Partial packed beds 1 and 2 (and 8 and 9 respect¬ ively) of the column formed a sucrose elution zone where the volumetric flow rate was 214 {/hr. The sucrose frac¬ tion was withdrawn at a volumetric flow rate of 19.6 {/hr. Partial packed beds 3 and 4 (and 10 and 11 respect- ively) formed an extraction zone where the volumetric flow rate was 194 {/hr. Partial packed beds 5 and 6 (and 12 and 13 respectively) formed an ion exclusion zone where the volumetric flow rate was 206 { /hr . The feed solution was introduced between these zones (volumetric flow rate 12.5 {/hr.), as was the rinsing solution (volumetric flow rate 12.5 { /hr. ) .

Partial packed bed 7 (and 14 respectively) formed a transition zone where the volumetric flow rate was 140 {/hr. The residual fraction was withdrawn from between

the ion exclusion zone and the transition zone (volumetric flow rate 66 { /hr. ) .

The feed and withdrawal points were shifted cycli¬ cally at intervals of 310 seconds for one partial packed bed in the flow direction (correspondingly, the elution, extraction, ion exclusion and transition zones were shifted for one partial packed bed) .

In equilibrium, the sucrose fraction had a dry solids content of 27.1% and a sucrose purity of 87.8%. -The residual fraction had a dry solids content of 5.5% and a sucrose purity of 16.4%.

Example 6

Two-profile separation from xylitol run-off

The column series employed for the separation com- prised three columns. Product fractions were withdrawn from column 1 and residual fractions from columns 1, 2 and 3.

The test conditions are shown in Table 6A.

Table 6A

Test conditions

Separation temperature 70°C

Resin bed height 11.1 m (3.8 m/column) Column diameter 20 cm

Number of columns 3

A two-profile separation sequence was produced in the column series, wherein two separate separation pro- files, i.e. essentially complete dry solids profiles, were simultaneously circulated in the separation process. The fractionation was performed in an eight-step sequence. The sequence had a cycle length of 87 minutes and comprised the following steps: Step 1: 15.0 litres of feed solution was supplied

to the top of column 1 at a volumetric flow rate of 52 {/hr., and a corresponding amount of a recycle fraction was eluted from column 2. Simultaneously eluent water was supplied to column 3 (15.0 litres; 160 {/hr.) and a corresponding amount of the xylitol fraction from the second profile was withdrawn from the same column.

Step 2: 10.0 litres of feed solution (continuation of feeding phase) was supplied to column 1 at a volumetric flow rate of 125 { /hr. , and a corresponding amount of a xylitol fraction was eluted from column 3.

Step 3 : Circulation in the loop formed by columns 1 to 3 (15.0 litres; 125 {/hr.) .

Step 4: Eluent water was supplied to column 2 (15.0 litres; 125 {/hr.) , and a corresponding amount of a recycle fraction was eluted from column 1.

Step 5 : The supply of eluent water to column 2 was continued (40.0 litres; 140 {/hr.) , and a corresponding amount of a recycle fraction was eluted from column 3. Simultaneously eluent water was supplied to column 1 (15.0 litres; 55 { /hr. ) and a corresponding amount of a recycle fraction was eluted from the same column.

Step 6 : The supply of eluent water to column 1 was continued (15.0 litres; 125 {/hr.) , and a corresponding amount of a residual fraction was eluted from column 3. Step 7: Circulation in the loop formed by columns 1 to 3 (18.0 litres; 125 {/hr.) .

Step 8: Eluent water was supplied to column 3 (20.0 litres; 125 {/hr.) , and a corresponding amount of a residual fraction was eluted from column 2. The sequence containing these steps was repeated eight times to equilibrate the system, whereafter the method proceeded in a state of equilibrium. In the state of equilibrium, samples were collected during one sequence . The composition of the raw material of the feed

solution is shown in Table 6B . Before the separation, the raw material was filtered and the concentration of the feed solution was adjusted to 50 g/100 g.

Table 6B

Analysis of raw material

Composition, % of dry solids

Gly- Mannitol Rhamnitol Xylitol Sorbitol Others cerol

2.5 18.2 1.1 43.0 7.8 28.5

Dry solids content g/100 g 50.6 pH 5.2

Conductivity mS/cm 0.3

The separation resin employed as a packing_material was Finex CS 13 GC (a polystyrene matrix cross-linked with divinylbenzene; manufacturer Finex Oy, Finland) , which was initially regenerated with calcium chloride (during the fractionation it was equilibrated by cations from the run¬ off) . The analysis of the separation resin is shown in Table 6C

Table 6C

Analysis of separation resin

Make Finex CS 13 GC

DVB % 6.5

Capacity equiv/{ 1.65

Mean bead size mm 0.41

In equilibrium, the combined residual fraction had a dry solids content of 7.7 g/100 g, and the xylitol concentration was 19.1% by weight. The xylitol fraction had a dry solids content of 16.5 g/100 g, and the xylitol purity was 66.2% by weight and the xylitol yield 80.6% by weight .