FI84839B - Process for separating psicose from other ketoses - Google Patents

Description

1 84839

A method for distinguishing psychosis from other ketosis

The present invention relates to the adsorption separation of ketosis - psychosis - from other ketoses and aldoses by means of a solid layer. More specifically, the present invention relates to a process for separating psychose from a mixture containing psychose, fructose and one or more other ketoses and / or aldoses, and the present process uses an adsorbent consisting of a calcium-exchanging Y-type zeolite to selectively adsorb psychose on feed.

It is known to use aluminosilicates in separation work when not targeting hydrocarbons; for example, it is known to separate specific monosaccharides or monosaccharide groups from carbohydrate feed mixtures. A specific example of group separation is given in U.S. Patent 4,024,331, which discloses the separation of ketoses from a mixture of ketoses and aldoses using an X-type zeolite. Specific monosaccharides, such as glucose and fructose, have been separated from the feed mixture by the same adsorption separation method using X-zeolite disclosed in U.S. Patent 4,442,285.

In particular, the present invention relates to the separation of ketosis - psychosis - from another ketosis - fructose - which is in admixture with aldoses - mannose and glucose. To date, no useful, economical method has been available to separate psychose from other isomerization products of aldose sugars, such as glucose. Therefore, enzymatic means have been used rather than simple isomerization methods to produce fructose, as the enzymatic method of preparation limits the production of psychose to a minimum. A method has now been devised for separating psychose from fructose and other ketoses and aldoses, and this makes it possible to use simpler, more advantageous isomerization methods than have hitherto been used for the preparation of fructose. There are two ways of isomerization to obtain fructose from glucose, namely by the reaction of a weak alkali with glucose and by the reaction of pyridine with glucose. See Chemistry of the Carbohydrates, Pigman et al., Academic Press Inc. NY, NY 194Θ, pages 41, 126-127. Likewise, the isomerization of galactose by either of the above isomerization techniques gives a mixture of aldoses - talose and galactose - and a mixture of ketoses - sorbose and tagatose, and can be separated as disclosed in the present invention. Furthermore, there is considerable interest in various L-sugars which are believed to be low in calories and possibly unmetabolized, which cannot be prepared enzymatically but only by isomerization, such as by the methods mentioned above. The present invention relates to L-sugars as well as D-sugars, and the process according to the invention has been found to be a preferred process for the preparation of L-fructose without Lpsicose as an impurity.

Facts about possible adsorbents suitable for separating mannose from other monosaccharides are disclosed in U.S. Patent 4,471,114. That patent relates to the use of Y-type faujasite, in which calcium cations are exchanged, as an adsorbent for separating mannose from glucose and other monosaccharides.

The separation of mannose from glucose is the subject of GB Patent 1,540,556. In this patent, the adsorbent is a cation exchange resin in salt form, preferably in calcium form.

U.S. Patent 4,340,724 discloses the separation of ketose from aldose by adsorption performed with a Y zeolite in which NH 4, Na, K, Ca, Sr, Ba and combinations thereof are exchanged at the ion exchange site, or an X zeolite in which Ba, Na or Sr or combinations thereof.

3,84839

Separation of psychosis from other ketoses is found in U.S. Patent 4,096,036 to be possible with cation exchange resins which can form complexes with the polyol at a first temperature and which can dissociate the complex at a second temperature in a heat parameter pump device.

It is an object of the present invention to provide a process for the separation of ketose from a feed mixture which is the product of an isomerization of an alkali or pyridine catalyzed aldohexose or aldopentose using a Y-type zeolite in which the calcium cations are located at the cation exchange site.

The present invention therefore relates to a process for separating psychose from an aqueous feed mixture containing psychose and at least one other ketose, and optionally aldose, which process is characterized in that said mixture is contacted under adsorption conditions with an adsorbent consisting of a Y-type zeolite, wherein the calcium cations are at the sites of the variable cations and which selectively adsorb said psychose, the non-absorbable portion of the feed mixture is removed from contact with the adsorbent and then the psychose is recovered by contacting the adsorbent containing psychose with the desorbent under desorption conditions.

The invention further relates to a process for recovering fructose and psychose, which fructose is substantially psycho-unbound, from a feed mixture comprising fructose and psychose and at least one other monosaccharide selected from the group consisting of glucose and mannose, which process is characterized in that said mixture contacting under adsorption conditions with a first adsorbent consisting of a Y-type zeolite having calcium cations at the sites of the exchangeable cations and selectively adsorbing said psychose and said fructose, the non-adsorbable portion of the feed mixture is removed from contact with said adsorbent and fructose is recovered from said first adsorbent and then said extract mixture is contacted with a second adsorbent consisting of a Y-type zeolite in which the calcium cations are located in exchangeable a and which selectively adsorbs said psychose, the non-adsorbed part of said extract mixture is removed from contact with another adsorbent to form essential! a psychose-free, fructose-containing raffinate which is recovered, and the said psychose is recovered by desorbing another adsorbent.

Other objects and embodiments of the present invention pertain to specific feed compositions, adsorbents, desorbent agents, operating conditions, and flow configurations, all of which are set forth in connection with the present invention in a review of the present invention.

Figure 1 relates to a graphical representation of the relative concentration of specific sugars versus retention volume using the method of the present invention.

It is necessary at the outset to present the various terms used in the application to explain the operation, objects and advantages of the present method.

A "feed mixture" is a mixture containing one or more extract components or one or more raffinate components to be separated by the present method. The term "feed stream" means the stream of the feed mixture that enters the adsorbent used in the process.

An "extract component" is a component that is more selectively adsorbed by an adsorbent, while a "raffinate component" is a component that is less selectively adsorbed. The term "desorbent material" generally refers to a material that can desorb an extract component. The term "desorbent stream" or "desorbent feed stream" means the stream through which the deorbent material enters the adsorbent. The term "raffinate stream" or "raffinate effluent stream" means a stream through which a raffinate component is removed from an adsorbent. The composition of the raffinate stream can range from a composition containing substantially 100% desorbent to substantially 100% raffinate components. The term "extract stream" or "extract effluent" refers to a stream through which an extract material desorbed with a desorbent is removed from an adsorbent. Similarly, the composition of the extraction stream may vary from substantially 100% desorbent-containing material to substantially 100% extract components. the extract stream, and preferably at least a portion of the raffinate stream, is separated into equipment, usually fractionation equipment, or a mixture in which at least a portion of the deaorbent material is separated to obtain an extract product and a raffinate product. The terms "extract product" and "raffinate product" means products prepared by a process comprising the extract component and the raffinate component, respectively, in higher concentrations than the extract stream and the raffinate stream.

The adsorbent materials of the present invention contain Y-type crystalline aluminosilicates having calcium cations at the cation exchange sites. Y-type, crystalline aluminosilicates or zeolites can be further classified as faujasilicates. As with all silicates in general, these crystalline compounds are described as a three-dimensional network of basic units containing SiO 2 - centered SiO 2 - and AlO 4 - centered aluminum tetrahedra joined by oxygen atoms at the vertices of both tetrahedra. . The space between the tetrahedra is occupied by water molecules and subsequent water removal or partial water removal produces a crystalline structure that is an interweave of molecular thickness channels. Zeolites are shown in Figure 6,84839 and are further defined in U.S. Patent Nos. 2,332,244 and 3,120,007, respectively, which are incorporated herein by reference. Y zeolites in hydrated and partially hydrated form may be represented in the form of molar oxides as shown below in formula 1, wherein "H" is a cation having a valence of up to 3, "n" is the valence of "M", " w "is a value from 3 to 6 and" y "is a value up to 9, depending on the identity of '' M 'and the degree of hydration of the crystal.

Formula 1

(0.9 + .0.2) M2 / nO: Al2O3 rwSiOj ty ^ O

The electron valence of the tetrahedron is balanced by the cation "M" shown in the above equation, which is located in the zeolite at the sites of the exchangeable cations. These cations, which after initial preparation are predominantly sodium ions, can be replaced by other cations using ion exchange methods familiar to those skilled in the art of crystalline aluminosilicates. Such methods are generally performed by contacting a zeolite or an adsorbent containing a zeolite with an aqueous solution containing an aqueous solution of a cation or a soluble salt of cations attached to the surface of the zeolite. After the desired amount of ion exchange has taken place, the sieves are removed from the aqueous solution, washed and dried to the desired water content. In the present processes, sodium cations as well as cations which are not sodium ions and which may occupy the exchangeable sites of the sodium Y zeolite as an impurity can be partially or practically completely replaced by other cations. The main thing is that the zeolite used in the present invention contains kale cations at the sites of exchangeable cations.

7 84839

The adsorbents normally used in separation methods contain zeolite crystals and an amorphous substance. The zeolite in the ad-sorbent is typically about 75-98% by weight based on the volatile, free composition. The remaining portion of the adsorbent is generally an amorphous material, such as silica or silica-alumina mixtures or compounds, such as clay, which is present in the initial mixtures with the small particles of zeolite material. This amorphous material may be an aid in the zeolite preparation process (e.g., incomplete purification of either zeolite during zeolite preparation), or may be added to a relatively pure zeolite, but in either case is usually intended to aid in the formation of zeolite hard crystalline particles or agglomerates. Usually, the adsorbent is in the form of particles, such as extrudates, aggregates, tablets, macrospheres or granules, having the desired particle size range. The particle size of the adsorbent used in this method is about 16-40 US mesh (420-1190 microns). It has been found that Y-zeolite with calcium cations as well as amorphous authenticates has previously been shown to have selectivity and other necessary properties, and is therefore suitable for use in the present invention.

Certain carbohydrates and so-called simple sugars are classified as monosaccharides. These monosaccharides are hydroxyaldehydes or hydroxyketones containing one ketone or aldehyde unit per molecule and two or more alcoholic hydroxyl groups. Monosaccharides are thus classified as aldoes or ketoses based on their carbonyl unit. Ketoses and aldoses are further classified based on their carbon chain length. Thus, five-carbon and six-carbon monosaccharides are named pentoses and hexoos, respectively. Well known aldohexoses include glucose, mannose and galactose. Arabinose, ribose β 84839 and keylooei are well known aldopentooe. Examples of common ketohexoses are fructose, fear and sorbose. Ribulose and xylulose are common ketopentoses. The present invention relates to a method for distinguishing psychosis from other ketopentoses and ketohexoses.

Thus, the feed compositions that can be used in the method of the present invention consist of a mixture of psychose and at least one other ketose. Possible feed mixtures may be the isomerization products of D-glucose and L-glucose. Such mixtures usually contain significant amounts of monosaccharides, such as psychose, fructose, glucose, mannose, and small amounts of polysaccharides, such as DP 4+, which have a degree of polymerization of four or more, for example. Feed mixtures derived from natural sources or which are isomerization products also contain lesser known monosaccharides. A typical feed grade suitable for the present invention contains psychose, fructose, mannose, glucose and polysaccharides in corresponding proportions, based on the weight percentage of solids, in the range of 0.5 to 90% by weight. In addition, the mixture may contain up to 10% by weight of solids formed by other lesser known sugars.

Although it is not clear which properties of the adsorbent are responsible for the separation described in the context of the present invention, it appears that the separation cannot be solely due to pore size. Since fear and fructose are separated from sugar molecules of the same size, it appears that etheric factors as well as electrostatic attraction play an important role in the separation. Although it is not critical to make the intermolecular interaction responsible for adsorption, one possible explanation is the combination of a cationic interaction that changes the orientation of specific sugar molecules toward the adsorbent pore opening. This altered orientation may provide a configuration corresponding to certain sugar molecules, in particular a structural configuration, which is compatible with the shape of the pore openings of the adsorbent, which has been altered by the presence of specific cations. Therefore, both electrostatic interaction and physical and stoichiochemical considerations may provide a mechanism for this separation.

Although it is possible to prepare very pure products by the process of the present invention, it is fully understood that the adsorbent never completely adsorbs the extract component, and the adsorbent does not completely adsorb the raffinate. Therefore, small amounts of raffinate component may be present in the extract stream, and similarly small amounts of extract component may be present in the raffinate stream. The extract stream and the raffinate stream are then further separated from each other and from the feed mixture by the ratio of the concentrations of the extract component and the specific raffinate component, both present in a given stream. For example, the ratio of the concentration of the more selectively adsorbed substance to the concentration of more selectively adsorbed sugars will be highest in the extract stream, next highest in the feed stream and lowest in the raffinate stream. Similarly, the ratio of less selectively adsorbed sugars to more selectively adsorbed sugars will be highest in the raffinate stream, next highest in the feed mixture and lowest in the extract stream.

The desorbents used in various prior art adsorption separation methods vary depending on such factors as, for example, the type of step used. In a swing bed system, in which a selectively adsorbed feed component is removed from the adsorbent by a cleaning stream, the choice of desorbent is not critical. However, in adsorption-based methods that typically operate as a continuous fraction at constant pressure and at constant temperatures to protect the liquid fraction, the desorbent material must be chosen judiciously in order to meet many criteria. First, the desorbent should displace the extract component from the adsorbent at reasonable mass flow rates without itself adsorbing so strongly that it inappropriately prevents the extract component from displacing the desorbent in the next adsorption cycle. Expressed in terms of selectivity, it is preferred that the adsorbent be more selective for all extract components for the raffinate component than it is for the desorbent material for the raffinate component. Second, the desorbent becomes compatible with a particular adsorbent and a particular feed mixture. More specifically, they must not reduce or destroy the critical selectivity of the adsorbent for the extract component with respect to the raffinate component. In addition, desorbent materials should not react chemically or cause a chemical reaction with the extract component or the raffinate component. Both the extract stream and the raffinate stream are usually removed from the adsorbent by mixing with the desorbent material, and any chemical reaction involving the desorbent agent and the extract component or raffinate component would reduce the purity of the extract product or raffinate product or both. In addition, since both the raffinate stream and the extract stream normally contain a desorbent, the desorbent materials should be materials that can be easily separated from the feed mixture that is circulated into the preparation process. Without the method for separating at least a portion of the desorbent material present in the extract stream and raffinate stream, the concentration of the extract component in the extract product and the concentration of the raffinate component in the raffinate product would not be very high and the desorbent material would not be available for reuse in manufacturing. It is possible that at least part of the desorbent wave is extracted from the extraction and raffinate streams by distillation or evaporation, but other methods such as reverse osmosis li 11 84839 may also be used alone or in combination with distillation or evaporation. As raffinate and extract products are nutrients intended for human consumption, the de-sorbent should also not be toxic. Finally, desorbent waves should also be substances that are readily available and therefore reasonably priced. Desorbents suitable for such separation include water and ethanol or mixtures thereof.

A previous study on this subject has shown that certain properties of adsorbents and desorbents are highly desirable, although not absolutely necessary, for the successful performance of a selective adsorption process. Such methods are equally important to the method in question. Such properties include: adsorption capacity with respect to a certain volume of extract component per volume of adsorbent; selective adsorption of the extracted component with respect to the raffinate component and the desorbent material; and sufficiently rapid adsorption and desorption rates of the extracted component to and from the adsorbent.

It is, of course, necessary that the adsorbent have the capacity to adsorb the specific volume of the extracted component; without such a capacity, the adsorbent is not useful for adsorption-based separation. Further, the higher the capacity of the adsorbent with respect to the extract component, the better the adsorbent. The increased capacity of a particular adsorbent makes it possible to reduce the amount of adsorbent required to separate the extract component of a known concentration contained in a given feed rate of the feed mixture. Reducing the amount of ad-sorbent required for a specific adsorption-based separation lowers the cost of the separation method. It is important that the good initial capacity of the adsorbent during the actual use of the preservative in the separation process is longer®11 than some economically desirable working time requires.

i2 84839

Another necessary property of the adsorbent is the ability of the adsorbent to separate the components of the feed mixture; or, i.e., that the adeorbent selectively adsorbs one component over the other. Relative selectivity can be expressed not only for one feed component over another, but can also be expressed as the difference in selectivity between any component of the feed mixture and the desorbent. Selectivity (B) is defined as the ratio of the two components of the adsorbed phase divided by the ratio of the same two components in the non-adsorbed phase, under equilibrium conditions, as shown in Equation 1 below:

Equation 1

Selectivity = <B> = Zoaino-t C / weight-X D7 »/ weight-X C / weight-X D7 ^ where in the equation C and D are the two components of the feed mixture, expressed as percentages by weight, and subscripts A and U denote adsorbed and non-adsorbed phases, respectively. Equilibrium conditions are defined as conditions in which the feed passing over the adsorbent layer does not change the composition after it has been in contact with the adsorbent layer. Ts. there is no pure transfer of material between the adsorbed nuclei and the adsorbed phases. As the selectivity of the adsorbent for the two components approaches 1.0, the adeorbent does not primarily adsorb one component to the other; both are adsorbed (or not adsorbed) to the same extent relative to each other. When <B> becomes less than or greater than 1.0, the adeorbent adsorbs primarily one component over the other. When comparing the selectivity of the adsorbent with respect to one component C over component D, (B) greater than 1.0 indicates that the adeorbent

II

13,84839 primarily adsorbs component C. (B), which is less than 1.0, indicates that component D is primarily adsorbed, leaving an unadsorbed phase with more component C and an adsorbed phase with more component D. Ideally, in this case, the desorbents should have a selectivity of about 1 or slightly less than 1 for all extract components, so that all of the extract components can be desorbed as a group at reasonable desorbent flow rates, and so that the extract components can displace the desorbent at a later adsorption step. Although it is theoretically possible to separate the extract component from the raffinate component when the selectivity of the adsorbent with respect to the extract component relative to the raffinate component is just slightly greater than 1.0, such selectivity is preferably reasonably greater than 1.0. Like relative volatility, the greater the selectivity, the easier the separation to be performed. Higher selectivities make it possible to use smaller amounts of adsorbent.

A third important property is the rate of exchange of the extract component of the feed mixture material, i.e. the relative rate of desorption of the extract component. This property directly relates to the amount of desorbent material that must be used in the process to recover the extract component from the adsorbent; higher exchange rates reduce the amount of desorbent required to remove the extract component and therefore allow a reduction in the operating costs of the process. At higher exchange rates, less desorbent material must be pumped into the process and separated from the extract stream for reuse in the process.

Adsorption-desorption operations can be performed in a dense solid packing layer which is contacted alternately with the feed mixture and the desorbent material, whereby the process is only semi-continuous. In another embodiment, commonly referred to as an alternating filling system, a combination of two or more static adsorbent benthic filling units may be used, in which the feed mixture may be passed through one or more unit assemblies while desorbent material is drained from one or more other filling units. The feed mixture stream and desorbent material can flow either upward or downward in such layers. Here, any conventional solid-liquid contact system used in static layers can be used.

However, moving bed counterflow systems or simulated moving bed flow systems have a much higher separation power than fixed bed systems and are therefore preferred. In moving bed and simulated moving bed processes, retention and replacement operations take place continuously, which enables both the continuous production of the extract and raffinate stream and the continuous import of the feed and replacement stream. In a preferred embodiment of this process, a simulated moving bed countercurrent system known in the art is used. In such a system, the progressive displacement of several liquid inlet sites down the molecular sieve chamber simulates the upward movement of the adsorbent contained in the chamber.

Only four inlet lines are active at any one time: the inlet flow, the desorbent inlet, the raffinate outflow, and the extract outflow lines. The movement of the liquid occupying the intermediate volume of the packed layer of the adsorbent is simultaneous with the movement simulating the upward movement of the solid adsorbent. In order to maintain countercurrent contact, a liquid flow downstream of the adsorbent chamber is 84839 can be provided by a pump. When the active fluid entry point moves through the cycle, i.e. from the top of the chamber to the bottom, the chamber circulation pump moves through different zones, requiring different flow rates. A programmed flow controller can be used to set and adjust these flow rates.

The active liquid inlets efficiently divide the adsorbent chamber into separate zones, each with a different function. In the particular embodiment of this method, it is generally necessary to have three separate operating zones to perform the method, although in some cases an optional fourth zone may be used.

The adsorption zone, zone 1, is defined as the adsorbent located between the feed inflow and the raffinate outflow. In this zone, the feedstock is in contact with the adsorbent, the extract component is adsorbed and the raffinate stream is withdrawn. Since the main stream passing through zone 1 originates from the feed stream, which becomes a zone to the raffinate stream leaving the zone, the flow in this zone is considered to flow in the direction of the stream when it occurs from the feed inlet streams to the raffinate outlet streams.

Immediately downstream of the liquid flow in zone 1, there is a purification zone, zone 2. A purification zone is defined as an adsorbent located between the extract outlet stream and the feed inlet stream. The basic operations in zone 2 are the displacement of the raffinate transferred to zone 2 with the adsorbent transferred to this zone without the intermediate space of the adsorbent and the selective pore volume or adsorbent of the raffinate material adsorbed on the adsorbent. Purification is accomplished by ie 84839 such that a portion of the extract stream matrix1 leaving zone 3 is discharged into zone 2 at the countercurrent boundary of zone 2, into the extract effluent, to displace the raffinate. In zone 2, the flow of the substance is downstream, from the extract outlet stream to the feed inlet stream.

The desorption zone, i.e. zone 3, is located immediately upstream of the zone with respect to the liquid flowing in zone 2. The desorption zone is defined as an adsorbent located between the desorbent inlet and extract outlet streams. The function of the desorption zone is to allow the desorbent substance that enters this zone to displace the adsorbed substance prepared by the feed contact in zone 1 in the previous round of production. The flow of liquid in zone 3 is substantially parallel to the flow of zones 1 and 2.

In some cases, an optional buffer zone, zone 4, may be used. This zone, defined as the adsorbent located between the raffinate outlet stream and the desorbent inlet stream, if used, is placed immediately upstream of the liquid flow to zone 3. Zone 4 should be utilized to maintain the amount of desorbent utilized in the desorption step, as some of the raffinate stream removed from Zone 1 may move to Zone 4 to displace the desorbent present in this zone from this zone into the desorption zone. Zone 4 will contain sufficient adsorbent so that the raffinate material, which is a raffinate stream exiting Zone 1 to Zone 4, can be prevented from entering Zone 3 and thereby contaminating the extract stream removed from Zone 3. In cases where four the raffinate stream passing from zone 1 to zone 4 must be carefully monitored in order to stop the flow directly from zone 1 to zone 3 when there is a significant amount of raffinate material in the raffinate stream not passing from zone 1 to zone 3.

Cyclic feeding of inlet and outlet streams through the solid bed of adsorbent can be accomplished by utilizing a manifold system with the valves in the tube operating in series to cause the inlet and outlet streams to alternate, thereby allowing liquid to flow countercurrent to the solid adsorbent. Another mode of operation that provides a fixed adsorbent flow countercurrent to the liquid involves the use of a rotary plate valve to connect the inlet and outlet streams to the valve and the lines through which the feed inlet, extract outlet, desorbent inlet and raffinate outlet flow . Both the manifold system and the plate valve are known. More specifically, rotary plate valves that can be utilized in this operation can be found in U.S. Patents 3,040,777 and 3,422,848. Both of the above patents disclose a rotary valve of the rotary type in which the supply of various inlet and outlet streams from fixed sources can be easily achieved. .

In many cases, one zone of action contains a much larger amount of adsorbent than another zone of action. For example, in some operations, the buffer zone may contain a smaller amount of adsorbent compared to the adsorbent required in the adsorption and purification zones. It can also be seen that in cases where a desorbent is used which can easily desorb the extract material from the adsorbent, a relatively small amount of 18,84839 adsorbent is required in the desorption zone compared to the adsorbent required in the buffer zone or adsorption zone or purification zone or all of them. Since it is not required that the adsorbent be located on a single column, the present invention encompasses the use of multiple chambers or sets of chambers.

It is not necessary that all input and output streams be used simultaneously, and in fact, in many cases, some of the streams may be shut off, while others cause material inlet and outlet. The device that can be used to implement the method of the present invention may also include a series of individual layers connected by connecting wires having inlet and outlet taps to which various inlet and outlet streams may be connected and alternatively and periodically transferred to provide continuous operation. In some cases, connecting wires may be connected to transfer taps that, during normal operation, are not used as a wire through which material enters and leaves the process.

It is expected that at least a portion of the effluent will pass to separation devices where at least a portion of the desorbent material can be separated to obtain an extract product in which the concentration of desorbent has decreased. Preferably, but not essential to the operation of the process, at least a portion of the raffinate effluent also passes to separation devices in which at least a portion of the desorbent material can be separated under separation conditions to provide a desorbent stream that can be reused in the process; and a raffinate product with desorbent concentration. The separation device may typically be a fractionation column or an evaporator, both of which are well known in design and operation.

«84839

For a further explanation of the flow diagram of a simulated moving bed countercurrent process, reference may be made to U.S. Patent 2,985,589 and to the publication entitled "Continuous Adsorptive Processing — A New Separation Technique" by D.B. Broughton has presented at Congress the 34th Annual Meeting of the Society of Chemical Engineers, Tokyo, Japan, April 2, 1969 (both are incorporated herein by reference).

Adsorption conditions include a temperature range of about 20-200 ° C, with a temperature range of about 20-200 ° C being preferred, and a pressure range of about atmospheric pressure to about 3445 kPa. , so that the pressure from atmospheric pressure to any pressure required to ensure the stability of the liquid phase is preferred. Desorption conditions are included in the same ranges of temperatures and pressures used in adsorption conditions.

The size of the units used in the process of the present invention can range from the scale of a pilot plant (see, e.g., U.S. Patent 3,706,812) to commercial scale, and their flow rates can range from as little as a few cm 2 per hour to many thousands of dm 2 per hour.

Another embodiment of the invention relating to a simulated moving bed flow system suitable for use in the method of the present invention is the downstream highly efficient simulated moving bed process disclosed in U.S. Patent 4,402,832, which is incorporated herein by reference in its entirety.

The functional test device can be used to study different adsorbents with different feed mixtures and desorbent material, to measure the properties of the adsorbent in terms of adsorption capacity, selectivity and exchange rate. The apparatus comprises an adsorbent chamber having a volume of about 70 cm and having inlet and outlet valves at opposite ends, temperature control devices inside the chamber, and pressure control equipment used to operate the chamber at a predetermined, constant temperature. The chromatographic analysis apparatus may be attached to the outlet line of the chamber and used to qualitatively detect or quantify one or more components of the effluent stream leaving the adeorbent chamber. A heart rate test performed using this apparatus and the following general method is used to selectively select other data. ad sorbent systems 11s. The adsorbent is filled to a certain equilibrium position with desorbent material by allowing the desorbent to pass through the adsorbent chamber. At a given time, a feed impulse containing known concentrations of non-adsorbed polysaccharide tracer, maltrin-DP4 +, aldoses, and other tracer sugars, all diluted in the desorbent, is injected over several minutes. The flow of desorbent is continued and the tracer and aldoses are eluted as in the liquid-solid chromatographic method. The effluent is collected in fractions and analyzed using chromatographic equipment, and tracers for the corresponding component peaks are developed.

Based on the data obtained from the experiment, the efficiency of the adsorbent can be evaluated in the form of the retention volume of the extract or raffinate component, the selectivity for one component over the other component and the rate of desorption of the extract component to the desorbent. The retention volume of the extract or raffinate component can be characterized as the distance between the center of the elution peak of the extract or raffinate component and the center of the elution peak of the tracer component or some other reference point. It is expressed in cm as the volume of desorbent pumped during this period,

II

21 84839 described by the distance between the peaks. The selectivity <B> for the extract component with respect to the raffinate component can be characterized as the ratio of the distance between the center of the elution peak of extract component e and the center of the marker elution peak (or some other reference point) to the corresponding distance between the center of elution peak of the raffinate component and the center of elution peak. The rate of exchange of the extract component with the desorbent can generally be characterized as the width of the elution peak at the peak half-intensity. The smaller the peak width, the higher the desorption rate.

The following examples are intended to further illustrate the method of the present invention and are not intended to limit the scope or spirit of said method. The examples represent test results for various adeorbent and desorbent materials then used in the above-mentioned functional test apparatus.

Example I

In this example, a Y-type zeolite + + pathway with Ca ions at cation exchange sites was used to separate psychose from a carbohydrate mixture containing said sugar. The Y-type zeolite exchanging calcium ions according to the present example was bound to an organic binder and had an average bulk density of 0.71 g / ml. The adeorbent was packed in an 8.4 mm diameter column with a total volume of 70 cm. The feed mixture used 10 ml of a carbohydrate mixture in a solution of 20X solids. The composition of the mixture is given in Table 1.

22 84839

table 1

Dry solids, weight - »_ D-fructoea 28.11 D-mannoo 6.62 D-glucoo 56.06 D-peikoo 4.42 DP4 + 1.59

The experiment was started by passing a water orbent through a column at a flow rate of 3.32 cm / min at a temperature of 65 ° C. At the appropriate time, a 10 ml feed mixture was sprayed onto the column, after which the flow of desorbent was immediately resumed. The figure shows a graphical representation of the retention of psychosis and other sugars in the adsorbent.

Examination of the mean center of each machine entrainment curve indicates that psychosis was well differentiated from the other components of the feed mixture. Fear is by far the most selectively restrained component. Above all, fructose is also selectively retained from the remaining components and can be separated by means of a Ca-Y adsorbent in the same or different adsorbent substrate. From the results obtained in this experiment, the selectivities shown in Table 2 were calculated.

Table 2

B

Psychose / fructose 3.76

Pe ik ooei / mannose 6.33

Psychose / glucose 18.25

Fructose / mannose 1.68

Fructose / glucose 4.84 23 8 4 8 3 9 These selectivities clearly indicate that a high degree of separation is achieved for psychose and a satisfactory degree of separation for fructose with respect to the remaining components of the feed mixture by performing a second treatment with the same Ca-Y adsorbent or another Ca-Y layer. through.

Example II

In this example, the aqueous feed mixture, as described in Table 3, was separated by countercurrent flow using a simulated moving bed system in which the adsorbent was Ca-exchanging Y zeolite and the desorbent was water. The valve cycle was one hour, the temperature was 65 ° C, and the ratio of adsorbent to feed, A / F, was 0.9. The feed was obtained by performing isomerization of L-glucose at 37 ° C and pH 10.6 for 48 hours.

Table 3 - Input analysis

Dry, solid

Feed materials,% by weight L-glucose 62.5 L-fructose 28.3 L-mannose 4.7 L-psychose 3.6

Unknown 0.6

In the first step, most of the fructose and psychose were separated from the remaining components of the extract stream, as can be seen from the following product analysis in Table 4. The extract obtained in the first step, which contained high-purity L-fructose and L-psychose, was then treated in a column containing calcium-exchanging Y-fauja brite. L-psychose, which was much more strongly adsorbed to the zeolite, was in the extract stream, while L-fructose, 24 84839 joeea was not L-peikooe, was 94.6% pure raffinate stream, joeea had small amounts of other components.

Table 4 - Product Analyzer

First stage Second stage

Raf f i- Raf f i

Extract Naate Extract Naate L-glucose 0.5 80.6 0.0 0.5 L-mannol 3.3 5.4 3.4 3.0 L-fructoye 88.9 10.0 73.9 94.6 L -peikooei 3.8 0.6 14.2

Other 3.5 0.1 8.5 1.9 li

Claims (8)

25 84839 A process for separating psychose from an aqueous feed mixture containing psychose and at least one other ketose and optionally aldose, characterized in that said mixture is contacted under adsorption conditions with an adsorbent consisting of a Y-type zeolite in which calcium cations are in place of variable cations and which selectively adsorbing said psychose, the non-adsorbable portion of the feed mixture is removed from contact with the adsorbent and then the psychose is recovered by contacting the adsorbent containing psychose with the desorbent under desorption conditions. A method according to claim 1, characterized in that said feed mixture contains psychose and at least one other ketose selected from the group consisting of fructose, sorbose, tagatose and ribulose. A method according to claim 1 or 2, characterized in that said desorbent consists of water. Method according to claim 1, 2 or 3, characterized in that said separation is performed on a simulated moving bed flow principle. A method according to claim 4, characterized in that countercurrent flow is used in said simulated moving bed principle. A method according to claim 4, characterized in that downstream flow is used in said simulated moving bed principle. 7. A method for recovering fructose and psychose, which fructose is substantially psychose-free, from a feed mixture comprising fructose and psychose and at least one other monosaccharide selected from the group consisting of glucose and mannose, characterized in that said mixture obtains 26 84 839 contacting under adsorption conditions with a first adsorbent consisting of a Y-type zeolite having calcium cations at the sites of exchangeable cations and selectively adsorbing said psychose and said fructose, the non-adsorbable portion of the feed mixture is removed from contact with said adsorbent and said mixture is fructose is recovered from said first adsorbent, and then said extract mixture is contacted with a second adsorbent consisting of a Y-type zeolite in which the calcium cations are located at the sites of the exchangeable cations and which orally adsorbing said psychose, the non-adsorbed portion of said extract mixture is removed from contact with the second adsorbent to form a substantially psychose-free, fructose-containing raffinate which is recovered, and then said psychose is recovered by desorbing the second adsorbent. A method according to claim 7, characterized in that said feed mixture contains L-psychose, L-fructose, L-glucose and L-mannose. 27 8 4 8 3 9