Forsberg, Holger K. (Kotka, SF)
Hamalainen, Lauri (Kantvik, SF)
Melaja, Asko J. (Kantvik, SF)
Virtanen, Jouko Johannes (Kantvik, SF)

05/437224

05/13/1975

01/28/1974

Download PDF 3883365       PDF help

Click for automatic bibliography generation

Suomen, Sokeri Osakeyhtio (Helsinki, SF)

127/60

127/30, 127/61, 127/58, 127/15

C13K11/00; C13F1/02; C13K9/00

127/30,58,60

3704168

November 1972

Hard

Wolk, Morris O.

Marantz, Sidney

Brumbaugh, Graves, Donohue & Raymond

This application is a continuation-in-part of our copending application Ser. No. 215,391, filed Jan. 4, 1972, now abandoned.

We claim

1. A method for crystallizing fructose from a water solution thereof containing glucose as an impurity which comprises the steps of

2. The method of claim 1, wherein the mass in step (e) is maintained at a supersaturation in the range of 1.1 to 1.2 with respect to fructose.

3. The method of claim 1, wherein, after the formation of a substantial quantity of crystals in the crystallization step (e), an additional quantity of the aqueous fructose solution of step (b) is added, the temperature and quantity of added solution being controlled to provide a solution which is at the saturation point for fructose, a second crystallization step is conducted by again lowering the temperature of the mass at a controlled rate to increase the size of the fructose crystals therein without substantial formation of new fructose crystals, and the fructose crystals are separated from the mass when the crystal size is in the range of about 200 to 500 μm.

4. The method of claim 3, wherein the step (e), and the subsequent second crystallization, the mass is maintained at a supersaturation of 1.1 to 1.2 with respect to fructose.

5. A method according to claim 3, characterized in that the crystallization is carried out as not less than two steps, where for each step a separate crystallization device is used, and where the volume of the mass is increased from step to step.

6. A method according to claim 3, wherein a crystal quantity of 45 to 55 percent by weight of the dry-substance quantity of the solution is reached in step (e) before a quantity of fresh solution is added whose temperature is such that it together with the mass forms a mixture in which the mass is approximately at the saturation point in respect of fructose.

7. A method according to claim 1, characterized in that the temperature of the mass is controlled in step (e) in a crystallizer to provide optimum conditions for growth of the original seed crystals without substantial formation of new crystals, and, while maintaining this temperature constant, fresh solution at the same temperature as or above the temperature of the mass is added to the mass at a controlled accelerating rate corresponding to the crystallization rate until the crystallizer is filled, the supersaturation of the mass with respect to fructose being maintained within the range of 1.1 to 1.2.

8. The method of claim 7, wherein, after the step of adding fructose solution, the temperature of the mass is lowered at a rate controlled to maintain supersaturation with respect to fructose within the range of 1.1 to 1.2.

9. A method according to claim 1, characterized in that a crystal mass that has been taken from a step of a previous crystallization is used as seed crystals.

10. A method according to claim 1, characterized in that during crystallization dry warm air is blown onto the surface of the mass to evaporate water therefrom.

11. The method of claim 1, wherein the method is conducted in a continuous manner.

12. The method of claim 1, wherein the aqueous fructose solution contains a low molecular weight organic solvent selected from the group consisting of methanol, ethanol and isopropanol at a level from about 2 to about 4% by weight of the sugar solids present therein, the presence of the solvent providing an increase in the rate of crystallization of fructose.

13. A method for crystallizing fructose from a water solution thereof containing glucose as an impurity which comprises the steps of

Crystallization of fructose from water presents a number of technical difficulties because of the high solubility of fructose in water. In addition, conventional crystallization using evaporation techniques is out of the question, due to the poor thermal stability of fructose.

Since the need for fructose has been extensively increasing in recent years, it involves considerable progress in this field to be able to prepare, by means of simple crystallization from a water solution, highly pure, large-crystal, and consequently mechanically easy-to-handle fructose in an economically advantageous way. All of these advantages are obtained by following the process of the present invention.

In accordance with the present invention, the fructose is crystallized from the water solution by adding to the solution, as crystal elements, a small quantity of fructose crystals of as homogeneous size as possible and by allowing these seed crystals to grow in size. The formation of new crystals is prevented by keeping the distance between seed crystals suitably short and by precisely controlling the temperature during the entire crystallization. The distance between seed crystals is determined by the ratio of seed crystal volume to the total volume of the solution. The volume of the solution is increased either continuously or stepwise while the crystallization process proceeds, in order to maintain the desired distance between seed crystals.

The invention is concerned with crystallization of fructose from water solution in order to obtain crystals whose average crystal size is 200 to 600 μm. In particular it is concerned with the crystallization of fructose from a water solution whose dry-substance content is not lower than about 90 percent by weight, whereby the dry substance contains not less than 90 percent by weight of fructose and the rest of glucose. Water solutions of fructose whose dry-substance content is within the range of 90 to 94 percent by weight and where the purity of said fructose is within the range of 90 to 99 percent by weight, are obtained by separating fructose from glucose by means of the method according to U.S. Pat. No. 3,692,582 of Asko J. Melaja.

Until now fructose has been crystallized from methanol. On the other hand, when fructose is crystallized from water, the difficulties are great because of the high solubility of fructose and because of its low thermic stability. The viscosity of concentrated aqueous solutions of fructose is very high, and the viscosity cannot be lowered by elevating the temperature because the fructose is unstable at the elevated temperature needed to substantially reduce viscosity. Under these circumstances, very small crystals result whose separation out of the solution is uneconomical. Moreover, the drying of the mass of small crystals is difficult. Because of all the difficulties described above, experts in this field have considered the crystallization of fructose from water solution as practically impossible on a commercially attractive basis.

Due to the poor thermic stability of fructose, conventional crystallization by means of evaporation is out of the question, and the crystallization is carried out according to the invention by lowering the temperature of the solution and, in addition, possibly by means of evaporation.

One object of the present invention is to make fructose crystallize out of water solution as large crystals that can be easily separated from the solution by means of centrifugation. A further object of the invention is to succeed in crystallizing fructose as crystals of sufficiently large dimensions that they are free-flowing, in which case their handling, such as weighing and packing, is simplified. Another object of the invention is to crystallize fructose that is of a highly homogeneous crystal size, and moreover, very pure.

Fructose crystallized from methanol has so far been used mainly by the pharmaceutical industry, which has been content with very small crystals with sizes clearly under 0.15 μm. The large-crystal fructose crystallized in accordance with the invention is, as a matter of fact, a novel product. Since the need for fructose has been extensively increasing in recent years, it involves considerable progress in this field to be able to prepare, by means of simple crystallization out of a water solution, highly pure, large-crystal, and consequently mechanically easy-to-handle fructose in an economically advantageous way.

As compared with the crystallization from methanol, an additional advantage is that the obtained product is completely free of methanol, which is a poisonous substance.

The above objects of the invention are according to the invention achieved by first providing a water solution of fructose whose dry-substance content is not lower than about 90 percent by weight, and where the dry substance contains from about 90 to about 99 percent fructose. The temperature of the fructose solution at this initial stage is that at which the solution is saturated with respect to the fructose, e.g., about 58° to 65°C. The fructose is then crystallized from the water solution by adding to the solution, as crystal elements, a small quantity of fructose crystals, of as homogeneous a size as possible, and by allowing these seed crystals to grow in size while, at the same time, preventing the formation of new crystals. This is achieved by keeping the distances of the seed crystals from each other as suitably short and by precisely controlling the temperature during the entire crystallization to maintain an optimal degree of supersaturation in respect to fructose of the range of 1.1 to 1.2. It has been ascertained that as the average distance between the crystals increases, the risk of formation of new crystals also increases. Similarly, as the degree of supersaturation increases above the optimal range, the risk of formation of new crystals increases.

It is further essentially characteristic of the method according to the invention that the volume of the solution be increased, either continuously or stepwise, while the crystallization proceeds. When the crystallization is carried out as several steps, i.e., as two or more subsequent steps, the velocity of formation of small crystals is substantially reduced.

When the crystallization is carried out as two or more steps, each step may be carried out in a separaate crystallizer, in which case the volume of the crystallizer is increased from step to step because of the increase in the volume of the solution. Alternatively, a crystallization as several steps may be carried out in the same crystallizer whose volume is equal to the solution volume of the last step.

The addition of 2-4 percent on the sugar mass of a low molecular weight organic solvent, such as methanol, ethanol or isopropanol, reduces the viscosity of the solution as well as the solubility of the fructose. Thus the crystallization and the centrifugation of the fructose crystals can be improved.

It is a further essential characteristic of the process of the present invention that the pH of the fructose solution be adjusted to within the range of 4.5 to 5.5 before evaporation, if that occurs, and before crystallization.

In the literature of the prior art, it appears that little attention has been paid to the importance of a close control of the pH during fructose crystallization procedures, and suggested pH ranges of the prior art processes vary widely. For example, the method of the Jackson patent, U.S. Pat. No. 2,007,971 suggests a pH range of 6-7. This range has been found to be too high because of the rapid formation at that pH of colored deterioration products and the epimerization of fructose to glucose and mannose which takes place in neutral or alkaline fructose solutions.

The Kusch et al. method in Austrian Pat. No. 489,610 suggests a wide pH range of 3.5 to 8.0. J. Murtaugh et al., U.S. Pat. No. 2,949,389, suggests that the most favorable pH range is below 4.0. This proposal agrees with applicants' findings obtained with dilute fructose solutions.

During the development of the method of the present invention, it was observed that crystal yields in some cases were much smaller than should be expected. From investigations, it has been found that the pH range of 3.0 to 4.0, which had originally been considered to be the most favorable in order to minimize the degradation and epimerization of fructose at high temperatures, based on applicants' study of the prior art, was not a favorable pH range for crystallization of solutions having a high fructose concentration. From a review of the literature, including A, Sapranov: Sacharnaja Promyslennost 9 (1968) p. 19 and Kato et al: Agr. Biol. Chem. 33 (1969) p. 939, applicants had drawn the conclusion that the degradation of fructose at high temperatures is at a minimum if a pH in the range of 3.2-3.6 is maintained. According to R. Jackson, U.S. Bureau of Standards Research Paper RP (1933), the stability of fructose passes a maximum at pH 3.3. R. Jackson also came to the conclusion that the concentration of the solution had no influence on the degradation of fructose, based on his studies of solutions with concentrations of 2-10 grams per 100 ml. Actually, fructose is rapidly destroyed in neutral or alkaline solutions (pH over 6), and at low pH, below 3, difructoses and their anhydrides are formed together with other deterioration products.

It has now been found, unexpectedly, that a pH range of 4.5-5.5 provides the most favorable pH range for crystallizing fructose from concentrated aqueous solutions. During the course of the studies, it became evident that the influence of pH on the formation of undesirable difructoses and difructose anhydrides is increased in fructose solutions of high concentrations. It was also found that the relative amounts of the seven major deterioration products vary with a change in the pH, that water is liberated when the difructose anhydrides are formed and that when a concentrated fructose solution (over 90% solids) is kept at a temperature of 60°C at a pH of 3, more than 10% of the fructose is converted to difructoses within 10 hours.

Furthermore, it is believed that difructoses and difructose anhydrides are actual crystallization inhibitors.

By conducting the process of the present invention at a pH of 4.5 to 5.5, preferably 5.0, a minimum of fructose destruction is experienced. This results in an increased yield of fructose crystals of the desired large dimension of between 200 and 600 μm within a shorter crystallization time. Total crystallization time was decreased, for example, from 180 hours to 120 hours and the yield of crystalline fructose was increased to 45-50% of the dry substance in the solution as compared with 30-35% of the dry substance at a pH in the range of 3-4.

A detailed description of the invention is given below. A two-step crystallization is described which is carried out in two separate crystallizers. The task is the crystallization of fructose out of a water solution whose dry-substance content is about 90 to 94 percent by weight and purity in respect of fructose about 90 to 99 percent, and where the impurity is glucose. As a result, crystals are obtained whose size, as determined by the Screen Test defined hereinafter, is 300 to 500 μm and crystal quantity 45 to 55 percent by weight of the dry-substance quantity. The purity of the finished product is higher than 99.5 percent with respect to fructose.

The next step of the process involves adjustment of pH of the water solution described in the paragraph above. The pH of the solution should be adjusted to within the range of 4.5 to 5.5 and preferably to pH 5.0. This may be done for example by the addition of an aqueous solution of Na ₂ CO ₃ . A suitable alternative procedure is to adjust the pH by means of an anion exchanger. The latter process can be conducted conveniently where the fructose solution is obtained by the method of the Melaja patent, U.S. Pat. No. 3,692,582, by using an anion exchanger before the fructose is separated from glucose.

Step I

a. Crystallizer No. 1 is filled with the water solution of fructose described in the preceding paragraph, whose temperature is controlled to such a level that the solution is saturated in respect of fructose (t = 58° to 65°C).

b. Into the solution a small quantity of fructose crystals of maximum uniformity of size are added as crystal elements, either crystals of 5 to 10 μm as suspended in isopropanol or larger crystals, for example, 80 to 100 μm, as dry. The seed crystal quantity (m _s ) depends on the size of the seed crystals (d _s ), on the quantity of the finished crystals (M), and on the desired crystal size (D) in accordance with the following equation:

m _s [tons] = (d _s /D) ³ M [tons]

c. Hereupon the supersaturation of the solution in respect of fructose is increased by lowering the temperature, and by means of a programmed controlling of the temperature, maximum crystallization velocity is produced without formation of disturbing new crystal elements. The temperature program depends on the purity and dry-substance content of the solution used for crystallization, and these programs are compiled on experimental basis for different cases. By means of samples taken with specified intervals, out of which the supersaturation of the mother liquid is determined, the correctness of the program is checked and, if required, it can be changed during the crystallization. The optimal supersaturation in respect of fructose has been found to be in the range of about 1.1 to 1.2. The cooling programs are preferably calculated to maintain the supersaturation within this range during the crystallization.

d. The crystallization takes about 50 hours, after which the temperature of the mass is 25° to 35°C, this also depending on the type of the solution used. At the end of the crystallization step the crystal quantity is approximately 50 percent by weight of the dry-substance content of the mass.

Step 2

Crystallizer No. II is filled simultaneously with the ready mass obtained from the preceding step and with a fresh solution whose temperature has before feeding been controlled to such a level that it together with the mass forms a mixture in which the mother liquid is saturated or slightly undersaturated in respect of fructose. After filling, a fine adjustment of the temperature is carried out.

Hereupon the same operations are carried out as the operations (c) and (d) of step 1. On termination of the process of the fructose crystals are separated from the liquor by means of centrifugation. The most suitable centrifugal machines are those of the type used in sugar recovery. Because of the high viscosity of the mass, a large centrifugal force is necessary. A preferred apparatus has a drum diameter of 42 to 48 inches and a rotation rate of 1400 to 1800 r/min. In a typical operation, a 42 inch centrifugal machine is filled with 120 to 250 kg of the fructose crystal mass. The crystals are washed by water (1-2 liters/batch). As the fructose crystals leave the centrifuge, their remaining water content is about 1.5%. Centrifuging time for each batch, including filling, centrifuging, and discharge, is 10 to 14 min. The capacity of a centrifugal machine of this type (diameter 42 inches, height of basket 600 mm) is 250 to 500 kg/h.

Horizontal cylindrical tanks whose diameter is 2 to 2.7 m. are used as crystallizers I and II. Externally, they are well insulated in order to prevent heat losses. They are fitted with a through axle on which helical refrigerating pipes are mounted. The rotation speed of the axle is 0.75 to 1.5 rpm. The refrigeration area is about 2.5 m ² /m ³ . The volume (length) of the crystallizers increases from step to step depending on the final crystal size of the step as follows:

V _II /V _I = (d _II /d _I ) ³

wherein V _I is the volume of crystallizer I and V _II the volume of crystallizer II, and d _I and d _II the final crystal size in steps 1 and 2, respectively. This arrangement provides for indirect temperature control; the temperature of the water circulating in the refrigeration pipes is controlled, and because of the large refrigeration area, a small difference in temperature between the mass and the refrigeration water is obtained (2° to 7°C). A direct temperature control procedure may be used, if desired.

Crystallization can also be promoted by evaporating water out of the mass, for example, by blowing warm dry air onto its surface.

In the following table of values are given for effective crystallization.

Table 1 ______________________________________ Step 1 Step 2 ______________________________________ Volume of crystallizer 10 m ³ 31 m ³ Total fructose quantity 13 t 40 t Crystallization time 50 h 60 h Crystal size 290 μm 500 μm Crystal quantity 6.3 t Yield of centrifugation 20 - 21 t ______________________________________

Instead of two separate crystallizers, I and II, it is possible to use a single crystallizer. Crystallization can also be carried out as more steps than two, for example, as three or four steps or as even more steps, whereby a number of crystallizing devices corresponding to the number of steps or only one crystallizing device may be used. Also, it is possible, as an alternative, to use a combination of the two alternatives presented above, for example, a three-step method, whereby the steps 1 and 2 are carried out in the same crystallizing device and step 3 in a separate crystallizing device. When one crystallizer is used for several steps, its volume must correspond to the solution volume of the last step to be carried out in it. In such cases the crystallizer can, for example, be fitted with a vertical axle on which helical refrigeration pipes are mounted. In each of these alternative procedures, the pH of the fructose solution which is introduced into the crystallizer should be in the range of pH 4 to pH 6, preferably 4.5 to 5.5.

The attached drawings give a graphical description of a two-step crystallization in two crystallizers. The graphical presentation has been obtained from a fractory-scale crystallization of fructose from a solution containing 87.5 percent fructose, 4.5 percent glucose and 8 percent water.

In the drawings

FIG. 1 presents the total quantity of fructose and the crystallized quantity of fructose in tons in the system as a function of time, the total crystallization time being 110 hours,

FIG. 2 presents the temperature of the fructose solution in °C as a function of time,

FIG. 3 presents the crystallized quantity of fructose in the system in percent as a function of time,

FIG. 4 presents the crystal size of the fructose in the system in μm as a function of time,

FIGS. 5A through 5H give a graphical description of the conditions used during one preferred embodiment of the process of the present invention,

FIG. 6 is a continuous crystallizer,

FIGS. 7A and 7B show graphically the formation of difructoses and water at 60°C in fructose solutions containing 90% solids during storage over a 12 hour period at a pH varying from 1.9 to 5.0,

FIG. 8 is a graph showing the improved yield obtained according to the present invention,

FIG. 9 is a diagram of a typical chromatogram of deterioration products of fructose obtained at a pH below about 4 in concentrated fructose solutions stored at elevated temperatures,

FIG. 10 is a graph showing the correlation between pH, difructose content, and yield of fructose crystals, and

FIG. 11 is another graph showing the influence of pH on diffructose level and on yield of fructose crystals.

The graphical presentation of FIGS. 1 through 4 has been divided into five sections: filling of crystallizer I; crystallization in crystallizer I (called pre-crystallization); filling of crystallizer II; crystallization in crystallizer II; and centrifuging. In FIG. I curve A presents the total quantity of fructose in the system and curve B the quantity of crystallized fructose. The O moment marked in the figure is the moment on which the solution starts being fed into the crystallizer I.

The fifth section of the graphical presentation in FIGS. 1 through 4 is concerned with the centrifuging of the solution containing the crystals. As indicated above, the centrifuging time for one filling of a typical centrifugal machine requires from 10 to 14 min. for a complete cycle of filling, centrifuging and discharge. In the procedure illustrated by the drawings, two centrifugal machines, each having a capacity of 250 to 500 kg/h, were used continuously until the crystals from the complete batch were separated. The total crystal quantity was 20 tons. It is for this reason that each of the drawings shows a time span of approximately 20 hours. for the centrifuging operation. With regard to FIGS. 3 and 4, the curves show a slight reduction in quantity of fructose crystals and crystal size, which takes place at the onset of centrifuging, and this is a reflection of losses which take place during the centrifuging step, and particularly during the washing of the crystals to remove the solution adhering thereto.

In connection with the transition from step 1 (crystallizer I) to step 2 (crystallizer II), at which time fresh solution of the same nature as the starting solution is added, a little part of the fructose crystals dissolve (FIG. 1), the temperature rises (FIG. 2), the crystal quantity calculated as percent of the entire mass is reduced substantially (FIG. 3), and the average crystal size becomes slightly smaller (FIG. 4).

The method described above, in which the addition of the fresh solution took place on transition from step 1 to step 2, can be modified so that fresh solution is being added continuously during step 1 until its end, in which case no further addition takes place on transition to step 2. This method is carried out as two steps but in one crystallizer as follows:

STEP 1

Into the crystallizer a solution, containing 87.5 percent fructose, 4.5 percent glucose and 8 percent water, is introduced whose temperature is so high (65°C) that it is undersaturated in respect of fructose. The quantity of the solution is such that it, together with the fructose crystals to be added as crystal elements, constitutes a mass in which about 15 percent by weight of the dry-substance content are as crystals. The solution is cooled down to the saturation point, and the necessary quantity of crystal elements, whose crystal size is about 100 μm, is added. For example, if the quantity of the solution is 900 kg, wherein there is 830 kg of dry substance, 150 kg of crystal elements of 100 μm size are added. The temperature is lowered to such a value (50° to 55°C) that optimum circumstances, including the optimum degree of supersaturation within the range of 1.1 to 1.2, are reached for the growth of crystals and the temperature is maintained constant. Since due to crystallization fructose is all the time leaving the solution, more solution must now be constantly brought to the mass in order that the circumstances remained unchanged. The degree of supersaturation must remain within the optimal range. When the crystals grow, the crystal area also grows and so does the quantity of fructose being crystallized within a unit of time. This is why the rate of addition of the solution is continuously accelerated in accordance with a program which has been established in advance. When the crystallizer becomes full, the addition of solution is stopped.

Step 2

This step is a refrigeration crystallization of the same type as the one described in step (c) of the method described first above. Hereby the mass is cooled from 50 to 55°C until the temperature has been reached at which the crystal quantity is about 50 percent by weight of the total dry-substance quantity of the mass.

Reference to FIG. 5 will serve to explain in more detail the considerations described in steps 1 and 2 above. The changes in mass weight in terms of tons of dry substance, temperature, the proportion of crystals to the entire mass, the crystal size, the crystallization rate, the mother liquid purity, the degree of supersaturation and the dry substance content of the mother liquid are each described in terms of curves following the complete process. It will be noted that the crystallizer becomes completely filled at the conclusion of step 1 and that the optimum crystallization conditions are maintained during step 2 by a reduction in the temperature of the mass to maintain the supersaturation with respect to fructose within the range of 1.1 to 1.2.

The crystallization device is a tank of the same type as that described in connection with the first method. For adding the solution a pump with adjustable output is used, to which a program mechanism has been coupled to produce an addition of solution at a rate that is varied in accordance with the desired program.

The following table gives values for an effected crystallization.

Table 2 ______________________________________ Volume of crystallizer 30 m ³ Mass 29 m ³ = 42.9 t 39.5 t dry substance Size of the crystals obtained 100 μm to 500 μm Time 120 hours Crystal quantity recovered 19 t = 49% by wt. of the fructose contained in the original solution ______________________________________

As seed crystals in the crystallization according to the invention may be used crystal mass taken out of a purposeful step of previous crystallization. As seed crystals may be used, for example, the crystal mass remaining after centrifugation in the method of this invention.

Where tons, sometimes abbreviated hereinbefore as tn, and t, are used in the specification and claims of this application, metric tons are intended.

Where crystal size of fructose is stated in the specification above or in the appended claims, the crystal size was determined from dry samples and from the final product by means of the Screen (Grist) Tests which have been tentatively adopted by the International Commission for Uniform Methods of Sugar Analysis as a standard method for determining sugar crystal size. This method is described on Pages 94, 95 and 96 of De Whalley (Ed) ICUMSA Methods of Sugar Analysis, Elsevier, N.Y., 1964.

During the crystallization procedures, the crystal size of the crystals in the syrup was determined by microscopic examination.

The term "supersaturation," as used herein and in the appended claims, is defined by the formula of Claasen and Holven (Honig P.: Principles of sugar technology Vol. II, Elsevier, New York 1959, p. 232) as the "supersaturation coefficient" which is expressed by the formula:

supersaturation = (S/W/S ₁ /W ₁ ) P,T

S/W = sugar/water ratio for a solution of purity P and temperature T.

S ₁ /W ₁ = sugar/water ratio for a saturated solution of purity P and temperature T.

(sugar = fructose)

The continuous crystallization can be carried out in a crystallizer which is similar to that used in batch crystallization. It is preferred, however, in the continuous crystallization process, to divide the crystallizer into sections by means of walls and each section is provided with a separate device for temperature control and for the addition of fructose solution. In the dividing walls there are holes through which the mass flows continuously from one section to another. Each section can also be a separate apparatus in which case the mass flows from one crystallizer to another. The crystallizers can be of the same size or of different sizes.

A typical continuous crystallizer is shown in FIG. 6 and functions as follows:

From a feed tank, which is equipped with a stirrer, a mass of seed crystals suspended in saturated fructose solution is added continuously or in portions to the first section of the crystallizer. Simultaneously fructose solution, the temperature of which in regard to fructose is so high that it is saturated or slightly unsaturated, is added continuously at a controlled speed to the first section of the crystallizer. In this section the mass is cooled to a certain temperature so that the crystals grow at maximal speed without any notable formation of new crystals.

To the mass, which has flowed to the second section, similar fructose solution is added at a controlled speed and the temperature is lowered again in the same way as in the first section. Seed crystals are added neither in this section nor in the following ones.

The number of the successive sections may vary and the operational range of each section depends on the number and size of the sections. The characteristic figures following the operation of a continuous crystallizer, divided in five sections, where the delay in each step is similar, 20-30 hours, are given in Table 3 below. The capacity of this crystallizer is about 140 kg/h of fructose crystals.

Table 3 ____________________________________________________________ ______________ Continuous Crystallizer ____________________________________________________________ ______________ Stage I II III IV V ____________________________________________________________ ______________ Volume m ³ 0.3 0.75 1.5 2.8 5.0 Seed crystals 100 μm 0.9 -- -- -- -- kg/h Fructose solution 92% d.s. 1/h 12 18 30 52 88 Crystal growth μm 100 - 180 180 - 240 240 - 320 320 - 380 380 - 440 Temp. change °C 50 - 45 45 - 40 40 - 35 35 - 30 30 - 25 Cooling surface m ² 2 3 4.5 9 18 ____________________________________________________________ ______________

FIGS. 7A and 7B show the influence of changes in pH within the range of pH 1.9 to pH 5.0 on the amount of difructoses and the amount of water formed during storage of concentrated fructose solutions at a temperature of 60°C. The difructoses were analyzed by thin layer chromatogram while the water was found by the Karl Fischer method.

It is seen that fructose is rapidly changed to difructoses and their anhydrides at pH 1.9, while at a pH of 5.0 the amount of difructoses formed is at a minimum. Similarly, substantial quantities of water are liberated when the difructoses anhydrides are formed and this is shown in FIG. 7B.

FIG. 8 summarizes yield data from a number of crystallization procedures carried out at pH ranging from 3.5 to 5.0. The six determinations shown on the graph in the range of 3.5 to 4.0 were taken from commercial production runs and this shows the wide variation and relatively low level of yields. The four determinations in the range of 4.5 to 5.0 show less fluctuation and higher yields ranging from about 40 to about 45%.

FIG. 9 shows a typical thin layer chromatogram where the unidentified deterioration products of fructose are shown as spots Nos. 1, 2, 4, 5, 6, 7 and 8. Spot No. 3 is fructose. The rapid formation of difructoses decreases the purity and the concentration of the fructose in the solution and thus tends to inhibit fructose crystallization and to decrease crystal yield. In addition, there appears to be a separate strong influence on crystallization caused by the presence of the difructoses and their anhydrides which appears to cause actual inhibition of crystallization.

FIG. 10 gives data collected from 53 samples, which data show the improvement in yields obtained by adjusting the pH of the fructose solution to about pH 5. The total yields before centrifuging drying and screening are 2-3% higher. The upper graph gives for each sample the yield as a percent of the original fructose content. It will be noted that from sample 16 on the pH of the samples was adjusted to about pH 5.

The lower graph of FIG. 10 indicates the percent of difructoses in each respective sample, the level of difructoses having been determined by thin layer chromatography.

FIG. 11 gives data which shows the influence of the level of difructoses upon the yield of crystalline fructose. As can be seen from the data, there is a correlation between a high difructose content and a low yield of fructose crystals. Where the pH of the fructose solution is adjusted to about pH 5 before the crystallization procedure is commenced, the difructose content of the liquors is maintained below 3% and usually below 1.5%.

Where reference to pH of the fructose solutions is made hereinabove and hereinafter, the pH of the solution was measured after dilution of an aliquot of the sample to about 50% dry substance.

EXAMPLE I

Fructose solution was obtained by the method of Melaja (U.S. Pat. No. 3,692,582) and crystallized in accordance with the two stage crystallization described above, the results of which are summarized in Table 1 above. The solution before adjustment of pH contained 75% by weight of dry substance. 98% of the dry substance was fructose and the solution had a pH of 3.6.

The pH of the solution was adjusted, before the evaporation step, by the addition of 3.5 kg of Na ₂ CO ₃ as a water solution. The adjusted pH was 5.0 (when measured after dilution of an aliquot to about 50% dry substance).

After evaporation to 92.5% by weight of dry substance, 31 m ³ of the solution was crystallized, the total crystallization time being 110 hours (step 1 = 50 hours and step 2 = 60 hours). The yield was 21 t, which equalled 49% of dry substance as fructose crystals.

EXAMPLE II

A fructose solution was prepared and crystallized as described in Example I. The pH was adjusted to 5.1 prior to evaporation by the addition of 4.9 kg of Na ₂ CO ₃ . The solution was evaporated to 92.5% dry substance. The fructose content was 97.0% of dry substance. The crystallization was carried out in two steps the time being 50 hours + 70 hours. The final temperature was 38.9 °C and the yield was 21 t or 49% of dry substance as fructose crystals.

EXAMPLE III

A fructose solution was prepared and crystallized as in Examples I and II. The pH was adjusted to 4.9 and the solution evaporated to 92.5% dry substance. The fructose content was 97% of the dry substance. The crystallization was carried out in two steps in 60 hours + 70 hours. The final temperature was 35.5 °C. The yield was 20 t or 48% of dry substance as crystalline fructose.

EXAMPLE IV

A fructose solution was prepared and crystallized as described in Examples I-III. The pH was adjusted to 4.5 and the solution evaporated to 92.5%. The fructose content was 96%. The crystallization was carried out in 60 hours + 70 hours and the yield was 44% of the dry substance as fructose crystals.

EXAMPLE V

A fructose solution was prepared and crystallized as described in the previous examples. The pH was adjusted to 5.5 and the solution was evaporated to 92.4%. The fructose content was 96%. The crystallization was carried out in 60 hours + 70 hours and the yield 43% of the dry substance as fructose crystals.