US2315699A - Crystallization of sorbitol - Google Patents

Description

Patented Apr. 6, 1943 2,315,699 CRYSTALLIZATION OF sormrror.

Rudolph Max Goepp, In, New Castle, Del., as-

signor to Atlas Powder Company, Wilmington, Del., a corporation of Delaware No Drawing. Application July 27, 1939, Serial No. 286,893

Claims.

The present invention relates to a method for crystallizing sorbitol. More particularly the invention is concerned with the recovery of sorbitol crystals from the technical process liquors obtained by the catalytic hydrogenation or by the electrolytic reduction of monosaccharides and the like.

The principal object of the invention is to control the crystallization of sorbitol from these process liquors so as to produce a pure uniform quality product which is dry and free flowing.

The method comprises first initiating crystal growth in the liquors; then controlling the progress of crystallization to produce the maximum quantity of crystals of the largest possible size and with the greatest possible amount of freedom from one another so as to give a readily separable mass of crystals; and lastly, separating the crystals so produced from the solvent or mother liquor.

An object of the invention is to prepare 'seed crystals upon which the sorbitol will later be deposited.

A further object is to prepare a magma or mass of seed crystals in syrup startin from dry, solid sorbitol crystals and syrup by intimate mixing of the crystals and syrup accompanied by violent agitation to break up the crystals into a large number of smaller ones.

Another object is to work up an initial batch of magma by repeated dilutions and violent intermixings, until the necessary volume is attained.

Another object is the production of a magma consisting of crystals and mother liquor which can be separated by filtration on a filter press or in a centrifuge and freed of most of the mother liquor (which contains most of the impurities) by washing. A particular object is to control the crystallization as to crystal size and as to the amount of solid phase in the crystallizing magma. The invention embraces the method of growing crystals from either aqueous or alcoholic solutions.

Another object is to provide a large mass of seed crystals and to keep the syrup and seed in constant slow motion so that crystal growth will consist largely in increasing the size and thickness of the seed crystals and so that only a minor part of the crystallization will arise from spontaneous nucleation. In this connection the method contemplates, in one aspect, a semicontinuous procedure in which a large portion of a crystallized mass of syrup is left to seed the next portion of uncrystallized syrup, only a part of the crystallized mass being withdrawn at a time for separation.

One of the objects of the invention is to grow the crystals isothermally at a moderate economical temperature. The method may be carried out as a batch procedure or it may be operated as a semi-continuous or continuous crystallization.

Another object is to decrease false grain, refuse solid phase and increase the dissolved phase by recycling, that is, withdrawing portions of the crystallizing mass, heating to melt and dissolve solids and returning to the mass from which it was withdrawn.

A further object consists in separating the crystals from the mother liquor by filtration which removes both the mother liquor and the dissolved impurities.

As part of the separation of the crystals from the liquor, the cake obtained by filtering is dried to remove remaining traces of solvent.

Finally, the dried crystals can be ground or otherwise prepared for use.

A further object of the invention is the conversion of purified sorbitol dissolved in volatile solvents to a dry free-running powder without recourse to spray drying.

Still another object is the stripping of sorbitol from the process liquors from monosaccharide reductions to produce liquors containing mixtures of sorbitol and related polyhydroxylic compounds, which mixtures are suitable as an ingredient in the preparation of mixtures of polyhydroxylic compounds distinguished by their high solubility in water, resistance to crystallization therefrom and suitability for conditioning purposes and'other uses.

The above and other objects will be better understood by reference to the following detailed description of the preferred embodiments of the invention.

In the following description several terms are repeatedly employed and it is here deemed advisable to set out definitely the meaning of these terms as employd in this specification and the claims appended hereto.

Magma refers to the mixture of crystals and liquor.

Concentration refers to the fraction by weight of total solids in the magma, whether in crystal form or in solution. Concentration may be determined by known direct methods such as drying on sand or filter paper or, more conveniently, a secondary standard may be employed, e. g. refractive index which may be converted quickly to concentration by reading the corresponding value on a standardized refractive index-concentration curve. For the satisfactory working of the process it is necessary to know only approximately the purity and absolute concentration of the syrup. Analytical errors in determining these factors can readily be compensated by making slight changes in the liquid content of the replacement. The liquid content is determined by a reproducible, if not absolutely accurate, test such as refractive index. For methanolic magmas the concentration can be controlled synthetically, that is, in the mixing ofithe magma, o1? analytically by a combined refractive index-density-composition chart.

Total solids includes all constituents of the magma or solution other than the solvent used in the process. This is essentially the non-volatile portion.

False grain refers to crystals of sorbitol distinctly smaller than those of maximum size obtainable under normal conditions, and which small crystals result from the breaking of larger crystals or from spontaneously produced nuclei.

Overload is a measure of supersaturation of a solution and is defined as the theoretical yield of solids, based on original total dissolved solids, obtained in passing from a higher concentration S1 to a lower concentration S2. It is defined mathematically as where S1 is the initial concentration of the solution being measured and S2 is the concentration of a saturated solution at the temperature under consideration.

Purgeability is the ease and completeness of removal of mother liquor from a filter cake, and by extension, the potential filterability of crystals in a magma.

Purity is the percentage of sorbitol in the total solids.

Percentage separated solids as used herein, is the percentage of crystals separated, based on total solids. This is mathematically equal to Cm- Cl 100 X of) where Cm=concentration of total solids of the magma and Cr =concentration of total solids of the filtrate after removing the crystals from the magma.

for sorbitol since no other polyhydric material,

such as sugar, mannitol, etc., exhibits the same behavior with pyridine. The pyridine number is the weight of sorbitol crystallized from anhydrous pyridine as above multiplied by 100, and divided by the weight of the sample (ash, moisture and sugar free). The pyridine number for pure sorbitol is about 95. The preparation of the sorbitol pyridine complex and its treatment to free sorbitol therefrom is described by Strain in J. Am. Chem. Soc. vol. 56, page 1757 (1934). The pyridine number of a sorbitol-containing product is an index of its crystallizing tendency from relatively highly concentrated aqueous solutions. The higher the pyridine number the greater the crystallizing tendency. The greater the complexity of the sorbitol-containing prodnot the less its crystallizing tendency and vice versa.

Replacement is the fresh liquor added to replace the vvithdrawal, in the process of continuous or semi-continuous crystallization.

Withdrawal is the removal of a portion of the crystallizing batch for further processing, and also the portion withdrawn.

Sorbitol has been one of the most difficultly crystallizable of the hexahydrc alcohols. Commercial sorbitol produced by catalytic hydrogenation or electrolytic reduction of monosaccharides has generally been marketed in the form of 50% aqueous solutions or lumps of solidified semi-translucent gel containing approximately 3 to 5% of water, or as a spraydried powder containing all impurities originally present. So far as is known, no one has yet disclosed a practicable process for the large scale crystallization of impure sorbitol before the present invention.

It is known that, under certain hitherto imperfectly understood conditions, sorbitol crystallizes from certain types of concentrated aqueous solution, particularly extracted sorb apple juice, as long needles, which may be freed of mother liquor by trituration with alcohol, and then recrystallized, in small yield, from ethyl alcohol. Even a few percent of impurities are suficient to change the crystallizing behavior of sorbitol profoundly, so that, instead of separating from ethyl alcohol as individual crystals, it comes out as small droplets of sorbitol-rich solution. The droplets then crystallize very imperfectly to give clusters of microcrystals enclosing considerable mother liquor. If methyl alcohol is used as the solvent, pronounced supersaturation occurs, and separation, when finally induced by the customary methods of seeding and scratching, is in the form of gelatinous masses. Similarly, if water is used as the solvent, a supersaturated solution of pure or impure sorbitol, at a concentration of 85% or more, at 25 C., either sets up to a stiff, translucent gel, or, if the concentration is not too great, a few very fine crystals are produced spontaneously, which grow for a time, and then, sooner or later are engulfed in a solid, virtually unfilterable mass of very small crystals, surrounded by gelatinized syrup.

The unsatisfactory crystallizing behavior of impure sorbitol is reflected in the various methods hitherto proposed for purifying it, such as preparation of the benzal derivative and regeneration therefrom, regeneration from the readily crystallized hexa-acetate by saponification, or from the well-crystallized pyridine-sorbitol complex. All these methods are slow, require expensive reagents, give low yields and are unsuited for the large scale economical purification and preparation of crystalline sorbitol.

The crystallization behavior of sorbitol 'in impure solution is due, as I have found, to the natural tendency of sorbitol to crystallize in relatively long needles. This characteristic crystal habit is, in turn, due to the fact that the ends of the crystals grow much faster than the sides. As a result, supersaturated aqueous solutions of sorbitol somewhat resemble solutions of starch or gelatin, in their consistency when gelled. The gelatinization is in all probability due to the spontaneous formation of numerous nuclei, which then proceed to grow almost entirely along one dimension, forming, in effect, tiny fibers. These fibrous crystals grow very rapidly through the solution, twisting and intertwining, thus forming a network of interlaced fibers which soon immobilize the interstitial liquor. In impure solutions, growth slows down due to adsorption of impurities on the crystal faces, so that crystallization soon comes to a halt, due to coating of all faces of the crystals with impurities, with only a few percent of crystals separated. As a matter of experimental fact, the separation of less than 3% of fibrous crystals can immobilize an impure aqueous sorbitol syrup of 75% concentration at 25 C.

If the spontaneously crystallizing magma is stirred, the effect is to accelerate very much the rate of gelatinization without changing its character. This is due to the breakage of the very slender and fragile sorbitol crystal by the internal friction of the solution, thus increasing the number of active end surfaces, and hence the number of crystals produced.

The present invention in its general form, comprises the crystallization of sorbitol, in motion, from solutions in volatile solvents, with careful control of overload, and in the presence of sufficient seed crystals, the amount of seed crystals present being sufiicient to prevent gelatinization, but not enough to increase the viscosity of the solution unduly, or to yield excessive false grain. The invention produces crystals which are well formed, relatively strong straight needles, not interlaced, and readily purgeable of the mother liquor. The magma thus obtained may be freed of most of the mother liquor by filtration, and the product converted to dry, substantially moisture-free solid crystalline sorbitol by drying in thin layers or in a finely divided condition.

I. CRYSTALLIZATION FROM AQUEOUS SOLUTION It has been found that sorbitol crystals can be satisfactorily grown from water solutions. As a starting material any suitable source of sorbitol may be used, such as the syrup obtained by the electrolysis or hydrogenation of glucose, fructose. invert sugar, inverted lactose, or by the extraction of toyon or mountain ash berries with alcohol. Preferably the starting material should have a P. N. of at least 65-70 with an ash content preferably not exceeding 2%, although it is possible to conduct the process with syrups of diiTerent purities. Thus a syrup of P. N. 50 can be used but the rate of crystal growth is slow and the product still relatively impure.

To get proper growth from a syrup of low purity. the overload must be increased and this increases the viscosity. A syrup of medium purity, such as P. N. 77, crystallizes much faster at a given overload than a material of P. N. 65. The growth rate may be adjusted by changing the overload as described below. Since the process accomplishes only a relative purification, the ash and reducing sugar of the final product are best controlled by keeping the ash and sugar of the starting syrup below the necessary limits. The reducing sugar in the crystals after the crystallization is generally about 0.3 of the reducing sugar concentration of the starting syrup (dry basis) and the ash content of the crystals is about 0.4 that of the starting syrup (dry basis). Also, if the starting material is low in ash and reducing sugar then, providing the original P. N. is high enough (preferably above '79), two crops of sorbitol can be taken out both of satisfactory ash and sugar content.

The preferred syrup for isothermal crystallization at 25 0. should have a concentration of about 75-82% of total solids depending on purity. Such a syrup can be added to a heel of 6090% of the original or a previous days batch, which heel should contain for seed a high percentage, say 10 to 13 separated sorbitol crystals based on the total sorbitol in the heel. In the course of about 24 hours with constant temperature conditions and stirring, the separated sorbitol crystal content should be about 10 to 13% and a withdrawal of 10-40% may be made if the semi-continuous process is followed. While such a syrup can be made to grow 10 to 13% of crystals at 25 C. the viscosity of the solution. is so high that the crystals cannot be purged satisfactorily and it is necessary to dilute the magma before filtering it, as described hereinafter. This dilution is preferably made with methanol, although ethanol may also be used. The alcohol reduces the viscosity and hence increases the filterability of the sorbitol. Without the use of an alcohol diluent the filter cake from a straight aqueous magma contains a very high percentage of mother liquor. e. syrup of P. N. 65-70 purged with an alcohol, a yield averaging 25 of a product which. averages close to P. N. is obtained. The syrup remaining after the separation of crystals from the magma has a P. N. of about 60.

In preparing a new crystallizin run an initial body of seed crystals is required and this may be prepared as follows:

A small quantity of pure dry sorbitol, prepared for instance by the method of Strain, J. A. C. S. 56, 3.757. is first taken and ground in a mortar. or stirred with an egg beater or chopping type of stirrer, with about ten times its weight of slightly supersaturated sorbitol solution which may be made from the pure crystals or may be a less pure technical syrup. Working at 25" 0., at which temperature pure sorbitol is saturated at concentration, a feed of 5% concentration is used corresponding to an overload of about 12?; as determined by the formula forth. above. Material of Sfi-SGit purity such may be obtained by the hydrogenaticn of glucose or invert sugar, would require a 76-78% solids con centration, while a still less pure material would need a total solids concentration up to to 82 7 The strength of feed to be used be calculated from the overload formula after the saturation concentration of the feed with respect to pure sorbitol has been determined at the oper g temperature. The overload at any ccncentration will vary with the purity of the sy up so that the saturation or equilibrium of the particular syrup to be used as is d mus? be determined with respect to pure sorbitcl. This saturation concentration may be determined by finding the concentration of the syrup in which pure sorbitol crystals will neither dissolve grow larger.

After grinding'or chopping minutes,

vigorously for the seed takes hold, and the magma be:

comes appreciably thicker than the original syrup. It is then diluted with one to three times its total weight of fresh syrup and stirred vigorously, preferably with a chopping stirrer, for 15-20 minutes further, when the magma is again diluted with a further one to three times its weight of additional syrup, and the stirring continued. In this way, starting with a few grains of pure sorbitol, a batch of any desired size can be built up by successive additions of supersaturated syrup to the magma of chopped-up sorbitol crystals in the proportion of one to three times the weight of syrup to the weight of magma. With each increase in size of the batch, unless the eiliciency of stirring is increased proportionally, the time required to thicken sufficiently before the next addition becomes successively longer and longer.

During this procedure tiny fragile crystals have been produced which tend to grow into the undesirable interlacing fibrous type. The vigorous chopping has broken up the free crystals and clots or" intertwined crystals into numerous short fragments to give a large amount of fast-growing, end-face crystal area. Agitation disperses these fragments throughout the solution thus adequately seeding it and preventing spontaneous nucleation. This solution tends to become immobile, however. Due to the elongated shape and very small size of the crystals, a 76% sorbitol solution can tolerate only about 3% or" such solid material without becoming stiff and buttery. The fine needle crystals at this point have a length of about 0.03 mm. and a transverse dimension of about 0.0001 mm. The solution is therefore thinned out from time to time when necessary to prevent immobilization. The crystals are by this means allowed to grow separately without appreciable entanglement until appreciable lateral growth is attained.

Instead of dilution, the volume of the batch can be kept constant and a fraction of the batch withdrawn, heated sufficiently to dissolve the crystals and, after cooling to the temperature of the batch, returned to the main portion. This in effect reduces the number of crystals and accomplishes the same eiiect as dilution, namely, the production of a new mixture containing chopped crystals from the preceding solution and overloaded sorbitol solution, the proportion of crystals in said new mixture being less than the I proportion of crystals in the preceding chopped mixture.

A point is ultimately reached where the magma can be allowed to crystallize with only gentle agitation, just enough to bring the crystals into contact with fresh liquor. The reason for this is that the crystals have now suificient thickness and breadth to grow as separate, freefiowing needles instead of fibers which bend and interlace. This point has been reached when the magma will tolerate about 8 to 10% separated solids and the preparation of the seed crystals is now completed.

This magma has present therein a sorbitol solution having an overload sumcient to permit crystallization and, in addition, crystals of surficient size which are present in sufiicient quantity so that additional crystallization may be limited to the formation of well-formed, individual crystals without further formation of fibrous crystals. In accordance with an embodiment of the present invention, this magma is slowly moved in contact with the crystals, care being taken to maintain the relationship of the overload of the sorbitol solution to the quantity of well-formed crystals, thereby to limit further separation of solids substantially to the formation of additional well-formed crystals and to the building up of the size of the crystals.

In this way, the crystals can be brought up to an average maximum length of about 0.15 mm, and a thickness of .002 mm. Any further attempt to increase the size of the crystal results in breakage, with formation of shorter fragments which grow out again rapidly to the limiting size.

In operating a batch method the magma produced by the above steps is separated all at once according to the procedure described hereinafter. The separation should be made when the percentage separated solids has preferably reached a value of 10-13% for an aqueous magma at 25 C.

With further increase of separated solids, the crystals tend to break up into smaller crystals which impair purgeability. The viscosity of the magma also becomes higher and at say 30% separated solids the magma is very difficult to stir and purgeability of the crystals is so poor that operation at such conditions is impracticable.

At temperatures above 25 C. the viscosities of the magmas will be higher than at 25 because, although increasing the temperature lowers the viscosity of a solution of a given concentration, the increase in solubility at the higher temperatures demands a higher concentration (which increases the viscosity) to produce an equivalent overload. Similarly at temperatures below 25 C. viscosities will be lower. Therefore, when operating at temperatures other than 25 C. the preferred values for percent separated solids will be somewhat greater or less depending on the temperature employed.

Obviously it is preferable not to have to repeat this procedure in building up the seed crystals very frequently, and hence a semiontinuous or continuous process is preferable. These latter processes involve the separation of only a portion of the magma at a time so that a substantial part of the magma remains to furnish seed crystals for the crystallization of further syrup.

The growing of the crystals can take place in any suitable vessel at constant temperature and with constant slow stirring. In order to increase the size of the seed crystals, these crystals must be contacted with large quantities of the mother liquor and therefore constant stirring is needed. At the same time the stirring produces breakage of some of the crystals thereby furnishing fresh growing surfaces. The rate of stirring is important in that the breakage of large crystals must not be excessive-a. large number of chips and small crystals leading to excessive false grain, raised viscosities, and dificult purgeability. In place of creating motion of the magma by internal stirring, other means may be employed such as rocking or tumbling the whole vessel in which crystallization is taking place. It has been found that such other ways of moving the liquid produce breakage of crystals also, so that presumably the internal friction of the magma in motion is a large factor in crystal breakage.

In general, the overload of a solution increases as the temperature is lowered and this fact is frequently employed in crystallizing where suiii cient supersaturation is not available at given temperature. However, it is diiiicult to control supersaturation induced by cooling solution and since sorbitol solutions in either water or alcohols may be readily obtained in suihciently supersaturated conditions even at room temperature to give satisfactory overloads, it is preferable and much simpler to conduct the crystallization at one temperature. In general it has been found convenient to conduct the crystallization at a temperature of 1540 C. and preferably 25 is selected as a mean of indoor temperature which can be maintained with aminimum of heating or cooling of the solution at any time. With relatively impure syrups the viscosityoverload relationship favors operations at lower temperatures.

According to the preferred method, a portion of the magma prepared as previously described, is withdrawn for sparation and this portion is replaced with an equal quantity of fresh supersaturated syrup. In practice the process is carried on semi-continuously by the periodic removal of a portion for separation and its replacement by an equal quantity of fresh syrup. The withdrawal of a portion of the magma at a time when the growth of well-formed, readily purgeable crystals has reached a maximum, and the replacement with sorbitol syrup results in a decrease in the proportion of crystals in the magma and an increase in the percentage of the sorbitol in solution. The minor amount of breakage caused by the slow agitation is sufficient to provide fresh crystal surface for the promotion of crystal growth. This crystal growth is thereby controlled so that it tends to take place on the small broken crystals thereby preventing the accumulation of a suificiently large amount of small crystals which would render the magma diflicult to purge, and further preventing spontaneous crystallization in amounts sufficient to immobilize the magma.

Since crystallization is preferably isothermal the variable elements are only concentration of replacement syrup, agitation, and amount and form of seed used. The amount of seed can be controlled by the frequency and depth (quantity) of withdrawal, the amount of seed being what is left in the crystallizer after a withdrawal is made. Amount of seed and also the form thereof can be varied by recycling as before explained.

The replacement syrup has preferably 75-82% total solids depending partly upon purity and partly on the operating conditions in the crystallizer. If a test of the total solids of the magma in the crystallizer shows that it is not holding its own against periodic withdrawal and replacement, the concentration of the replacement is increased until the magma reaches equilibrium. Conversely, too high an actual overload leading to undue false grain and lessened purgeability is corrected by lowering the concentration of the replacement. Another factor to be considered in the choice of concentration is the viscosity of the magma, since too high an overload makes the magma difiicultly stirrable and. also leads to false grain formation.

The frequency and depth of withdrawal are varied in accordance with the purity of the starting material since rate of crystal growth is dependent on purity, all other conditions being equal. With a syrup of P. N. '70 crystallized at 78% concentration at 25 C. a satisfactory growth is maintained when is removed every 12 hours. On processing with methanol as described below, a 22% yield is had. Less pure syrups crystallize more sluggishly so that longer growth periods are necessary. Purer materials make it possible to withdraw larger amounts for separation, If withdrawals are made more frequently, a high overload can be used and a greater capacity achieved Without incurring the risk of undue thickening and false grain formation.

The semi-continuous withdrawal and replacement process is much better adapted to commercial operation than the batch process. The equipment required for semi-continuous crystallization is a slowly stirred crystallizer which can be maintained at a constant temperature, a mixing tank, a centrifuge or filter press, and a dryer. In this form of the invention there is no need for continued use of small scale equipment as in the case of the batch process where the crystals must be built up from small quantities of material. In the semi-continuous process the small depth of withdrawal and replacement make for uniformity and homogeneity of the product despite fairly wide fluctuations in the purity of the feed. The control of the process is very simple, requiring only slight adjustment from time to time in the concentration of the feed syrup to raise or lower the overload to compensate for changes in purity. These advantages more than outweigh the disadvantages of getting a daily yield of only a small fraction of the solids in process.

The semi-continuous, and also the continuous, processes of this invention provide means for crystallizing sorbitol at the maximum obtainable rate of growth compatible with production of purgeable crystals. The attenuated growing habit of the sorbitol crystals is such that in a moving viscous solution breakage occurs at an average length of 0.15 mm. The distribution of crystal sizes, therefore, varies, the majority of crystals being between 0.05 and .15 mm. in length and about 0.00'10.002 mm. in thickness and width. There is a slight amount of false grain and there are also a very few crystals substantially thicker and a little longer than the others. At this particle size and distribution and under the preferred operating conditions, the crystals can increase in total amount from an initial 9% (obtained from the seed) to 13% within 24 hours. If the number of 0.15 mm. sorbitol crystals is increased much above 13% there is a tendency to increase the rate of breakage and false grain formation with attendant impairment of purgeability. Hence, with aqueous magmas at 25 C. it is preferred to operate with not more than about 13% separated solids present in the magma.

Instead of growing the crystals in a semi-continuous manner and periodically removing portions of the magma, the method can be operated continuously with high concentration syrup, say total solids, feeding and withdrawing at constant rates.

Although the entire contents of the crystallizer in a batch process, or the withdrawal in the semi-continuous or continuous process, may be filtered directly in a press if so desired, the removal of the viscous aqueous mother liquor from the crystals is not complete and the purification achieved is not as good as may be desired for some purposes. Accordingly, it has been found advantageous in most cases to thin the viscous aqueous mother liquor with a quantity of alcohol, either ethyl or methyl. The quantity and temperature of alcohol used is such as to reduce the viscosity of the magma substantially but not enough to diminish the overload or to cause the formation of two liquid phases. The ethyl alcohol maybe used in the form of a mixture with methyl alcohol such as the denatured alcohol formula No. l. Ethanol is much less satisfactory than methanol due to formation of two liquid phases at certain concentrations, that is to say, ethanol and aqueous sorbitol solutions are not miscible in all concentrations, particularly those obtained momentarily when stirring the ethanol into the magma. There appear two solutions of different densities, the heavier rich in sorbitol containing a small amount of alcohol and water, the other consisting mainly of alcohol containing sorbitol and water. Droplets of this heavier solution in contact with the sorbitol crystals agglomerate them into a sticky mass which is very difficult to filter. For this reason, ethanol gives a product of lower purity which is much harder to dry than the methanol purged crystals. However, it is to be understood that ethanol can be used to dilute an aqueous magma prior to purging if so desired.

The dilution is preferably made with the alcohol, and water if necessary, to give a mixture whose composition is 35% total solids in 52% alcohol and 13% water. It is essential to mix the alcohol and the magma thoroughly so that the dilution is complete and the resulting mother liquor is a homogeneous alcohol solution. Vigorous stirring for several minutes is therefore used to insure this result.

The diluted magma is thereupon filtered in a centrifuge or other suitable filtering device. In the interest of uniformity it is desirable to maintain constant filtering termperatures. The filter cake thus obtained may be dried readily either in a vacuum dryer or in air of humidity less than 50%. Vifhere atmospheric humidities above 50% relative humidity are encountered, an enclosed solvent centrifuge or a filter press are preferably used. Where heat is employed in the drying, care must be taken that the temperature is not high enough to melt the crystals in the filter cake. The dried product is a soft, readily comminuted mass of very fine crystals having an impurities content of less than 35% of that of the starting syrup.

The dilution and filtration are preferably conducted at the same temperature as the crystallization,

Dilution with alcohol is desirable to get good purification and rapid drying of the filter cake. if after diluting the magma with alcohol any considerable period of time is allowed to elapse before filtering, considerable further crystallization takes place in the alcoholic magma. Hence, the alcoholic magma can be used also as a growth medium for a crystallizing process. As will be seen below, even better crystallization results are obtained where the crystals are grown in an alcoholic, preferably methanolic, magma.

Where it is desired to obtain a product of greater purity than the starting solution it is necessary to remove the mother liquor as a liquid to carry off the dissolved impurities. However, when it is desired only to convert sorbitol syrup to crystalline form without purification, then centrifuging and the alcohol dilution required for the latter can both be dispensed with and the liquor from a water crystallization dried down to solid crystalline sorbitol. It has been found that unfiltered aqueous magmas of high purity will dry if they contain originally even as low as separated crystals, volatiles as water, and 70% uncrystallized or dissolved solids, providing that the magmas are given a large surface area-tofacilitate evaporation and permit diffusion of water vapor from the interior of the crystallizing mass.

The method of getting free running crystals from magmas of relatively high purity has been denominated magmatic drying and is similar to the drying of damp filter cakes, the principal difference being in the relative proportions of crystallized and uncrystallized solids at the start of the drying process because no filtration step is interposed where the magma is dried directly. For example, a soggy filter cake, which may contain crystalline or separated solids, 48% dissolved solids and 12% volatiles can be dried readily under vacuum or in dry air if the overall purity is 80% or better. Likewise, an unfiltered magma of like purity, but having a greater percentage of dissolved solids can be dried if spread out into a thin layer or otherwise given a large surface area and subjected to drying temperatures or placed in dry air.

The temperature used at the start in the magmatic drying should not exceed that used for growing crystals in the magma, otherwise seed crystals will be dissolved and gelled portions produced in the cake, which lose water only very slowly and are very difficult to comminute. As drying proceeds, the temperature may be raised. Accordingly, the temperature used for growing magmas for magmatic drying should be as high as possible so that the Volatile solvent can be removed as rapidly as possible without dissolving any crystals in the mother liquor.

The apparatus and procedure used for growing the crystals in a magma to be separated by magmatic drying are similar to those used in the preparation of the magma for separation by filtration. In preparing magma for magmatic drying, however, the process should be started with a cold, aqueous magma at 25 C. and the temperature gradually raised to LO- C., the temperature and concentration being controlled throughout so that the magma is at all times supersaturated with respect to sorbitol. The preparation of a magma for magmatic drying can be advantageously conducted in a vacuum pan which can be operated as a continuous crystallizer, using periodic withdrawal and replacement. When using a vacuum pan, agitation can be provided by bubbles, so that mechanical stirring is not necessary, but where this is done, the feed should not be supersaturated, but should be saturated or even less, i. e., it should contain excess water, whose removal under vacuum supplies the bubbles necessary for agitation.

This magmatic process is particularly adapted to the preparation of crystals from aqueous magmas due to the high concentration of sorbitol in saturated aqueous solutions. At 35 C. an 87.5% solution of sorbitol is somewhat supersaturated and has only 12.5% of volatile substances to be removed by drying or evaporation.

he rate at which drying should be done depends primarily on the crystallizing power and hence the purity of the syrup. If the evaporation of water is pushed too rapidly, the supersaturation will increase beyond the point Where the seed crystals can control, so that gelatini'zation will take place due to the formation of the fibrous type of crystais.

The rate of agitation used in preparing a magma for magmatic drying may conveniently be greater than that for crystallizing a magma for subsequent filtration, in order to break up crystals. Since purgeability is no object in magmatic drying, andsince. it. is desirable to have crystallization proceed as rapidly as possible to facilitate quick drying, the maximum number of well formed crystals should be produced in the magma, regardless of size and purgeability. This is best accomplished by vigorous agitation of a rather thick magma, which produces numerous short fragments of well formed crystals. These provide sufficient surface for crystallization during the magmatic drying and prevent the spontaneous nucleation which is associated with gelatinization caused by fibrous crystals.

By whichever steps the sorbitol is reduced to crystalline form the dried material may be comminuted or otherwise prepared in usable form.

EXAMPLES or (lRrs'rsLLIzArios l uoir. Aqueous SOLUTION Example 1 50 g. of crystalline sorbitol, either pure or obtained from a previous crystallization from impure syrup, and 300 g. of an 80% aqueous solution of sorbitol syrup, P. N. 65, are stirred at 25 C. with an egg beating type chopping stirrer until the viscosity begins to increase, which requires 28-30 minutes. 1000 g. of fresh solution are then added and the chopping process continued. Thereafter a further quantity of 3000 g. esh solution is added and the chopping and ng again repeated. After this dilution the batch is transferred to a five-gallon vertically s and crystallizer, thermostated at 25 C. and surreal at about 5 R. P. M. After 12 hours the resulting magma is diluted with a further quantity of 8000 g. of fresh solution, allowed to crysr tallize for 12 hours and then the crystallizer is filled up to a total of 30 kg.

After 24 hours the magma contains about 13% separated solids as determined by a filtration test at 25 C. and a 3 kg. withdrawal is taken and beaten at 25 C. with 3000 cc. of absolute methanol using a high speed turbine stirrer and adding the alcohol slowly over a period of five minutes to insure uniformity. The thin aqueous alcoholic magma is then centrifuged and dried partially by spinning for an hour, although the bulk of the filtration is completed in minutes. The centrifuge cake is porous and dries readily to a light friable microcrystalline mass. The yield is 27% of original solids, the purity is about P. N.

85. Since the yield of dry product usually exceeds and the separated solids before stirring up are only 1013%, a substantial part of the crystallizing process takes place during the stirring up, centrifuging and drying, the additional solids coming from the mother liquor.

Example 3 A 3 kg. withdrawal of magma from the batch of Example 1 is replaced with 3 kg. of fresh solution and crystallization allowed to proceed 24 hours with constant stirring and at constant temperature. At the end of this period the separated solids have returned to their original level of 13%. A 3 kg. or 10% withdrawal is taken, stirred up with 3 liters of absolute methanol, centrifuged and dried, giving a 28% yield of dried product, P. N. 85.

' Example 3 A magma is built up at 25 C. in a 30 kg. crystallizer as described in Example 1, using a sorbitol syrupof P. N. 68 at a concentration of total solids of 79.5%. When 13% separated solids are reached, as withdrawals are taken, proceeding as in Example 1, replacements are made with a purer syrup, P. N. 76, and having 77.5% total solids concentration. After 27 daily withdrawals and replacements with the syrup of the above composition, the magma is of 77.5% total solids concentration and substantially P. N. '76, the yield from the 28th withdrawal is 25% and the ash and reducing sugar impurities of the product 0.3 those of the original, while the purity of the product is increased to P. N. 86.

Example 4 150 lbs. of sorbitol syrup of P. N. 73, having a concentration of 78.0% was introduced into a crystallizing vessel capable of providing slow and continuous agitation, which contained 350 lbs. of crystalline sorbitol magma of the same percentage of total solids as the fresh syrup. The amount of crystals in the heel was 13.6% of the sorbitol in the crystallizer before the addition of the fresh syrup and was 9.5% after the addition. The amount of the heel or seed magma was of the total batch after the addition of the fresh syrup.

After stirring at 24 C. for 2-: hours the amount of crystals had built up to 14.9% of the total sorbitol. lbs. of magma or 30% of the contents of the crystallizer was thereupon withdrawn and the crystallizer again filled up with fresh syrup.

The portion of the magma withdrawn was stirred up with methanol and water to give a mixture containing 42.8% sorbitol, the solvent being aqueous methanol 77% strength by weight. This diluted mixture was then filtered on a centrifuge and spun for one hour. The wet cake contained 73.3% solids and after drying in a vacuum drier yielded 32.8 lbs. of a dry, white, fluffy, friable powder representing a yield of 27.4% dry basis. The dry product contained 27% of the ash in the original syrup and 31% of the original reducing sugar. The dry product had a P. N. of 87 .8.

II. CRYSTALLIZATION FROIVI ALCGHOLIC SCLUTIONS In place of growing the crystals in water solutions and thereafter diluting with an alcohol to permit separation of the crystals from the mother liquor, it has been found advantageous to conduct both the crystal growth and the separation in alcoholic solution. As solvent, either methanol or ethanol may be used, but methanol is preferred because of the fact that ethanol produces double layer or two liquid phase formation in many cases. Purification is better with methanol than with ethanol. On the other hand, ethanol is somewhat cheaper than methanol.

Compared to the water solutions, the alcoholic solutions have much lower viscosities for the same overload and therefore the alcoholic solutions are more readily handled. High overloads are possible in alcoholic solutions at lower total solids concentrations than are possible in aqueous solutions giving the same overloads. Hence, more clilute sources of sorbitol may be employed where the crystal growth is to take place in alcoholic solution. The crystals may be grown in a wateralcohol magma by starting with a supersaturated alcoholic solution of sorbitol, withdrawing a portion of the magma and replacing it with solution. The composition of the magma may be adjusted to give supersat rations and growth rates compatible with a suitable withdrawal schedule.

The use of aqueous alcohol as a solvent carries with it several advantages. The viscosity of the tions of impure sorbitol in strong alcohol is much solutions or sorbitol can be used advantageously so that a higher yield in proportion to total dissolved solids can be obtained from an alcoholic magma. Alcohol solutions of low purity can also be worked economically for purified sorbitol. Because of low viscosity the replacement can be mixed into the batch much more quickly without excessive power requirements and Without undue crystal breakage and false grain formation.

Also, much higher overloads can be used in alcohol than in water without is r of gelatinization, enabling faster crystallization and higher capacity for crystallizing equipment without sacrificing purgeability and product quality unduly. By comparison, the aqueous solutions must be kept at relatively low overloads in order to retain the purgeablity of the crystals and therefore crystallization is slower.

In the operation of an alcoholic crystallizing method the equipment and procedure is modifled in accordance with known practice in other chemical arts dealing with volatile and relatively expensive solvents so that the solvent content of the magma is kept constant despite evaporation and so that substantially all or" the solvent is kept in process and not lost.

Where the solvent is dilute aqueous alcohol, high sorbitol concentrations are necessary to produce the necessary supersaturation. On the other hand, where strong alcohol is used, the sorbitol concentration for the necessary supersaturation is considerably less. In this latter case the volume of solution per unit quantity of product obtained is larger than such volume per unit of product when dilute aqueous solutions are employed. The optimum composition of solution is determined by balancing the cost of large equipment with the cost of recovering the sorbitol from the filtrate since in the former case where dilute aqueous alcohol is used a considerable quantity of sorbitol passes througl'i the process unused and must be recovered from the filtrate. It has been found preferable, in general, to employ aqueous alcohol of 93% strength by weight,

To prepare the alcoholic for crystal growth, a sorbitol syrup or the wet cake from a previous filtration, or dry sorbitol, is mixed with alcohol and water to prepare a syrup of the ole-- sired composition. This sorbltol concentration is adjusted to produce supersaturation or overload in amount depending on the purity (P. N.) of the corbitol used so that the magma at the time of filtration will not be too thick. to handle in pumps, pipes, filtering equipment, etc. in the alcoholic crystallization, in contrast to the aqueous process where resistance to flow is due to viscosity, the alcoholic magmas are slightly thixotropic in that they set slightly when not stirred and do not flow easily under low h ad. It has been found experimental y that alca holic sorbitol magma containing not more than 8% solid phase (wet basis) is as hick: can be conveniently handled. lviagmas containing more than this amount of solid phase tend to clog in constrictions, pipe bends, etc. This syrup is seeded with a suitable seed such as the heel from a previous batch so that the actual crystal content is equal to at least 1% or the total sorbitol in the batch. If the heel from a previous batch is not available, a seed magma may be used which is built up as described above with reference to crystallization from aqueous solution and the magma thus grown may be'diluted with alcohol before addition to the syrup to bring it to equal alcoholic content. The seeded magma is maintained in a crystallizer, generally at constant temperature. and stirred continuously for the period necessary to produce the maximum crystal growth. As a rule the minimum period for this growth is 2 hours and at the end of this time a withdrawal can be made and the crystals seperaied from the mother liquor. As in the case of the aqueous crystallization care is taken to maintain the relationship of the overload of the sorbitol solution to the quantity of well-formed crystals thereby to limit further separation of solids substantially to the formation of additional well-formed crystals and the building up of the size of the crystals to an average length of 0.05 to 0.15 mm. and an average thickness of 0.001 to 0.002 mm. In alcoholic solution the withdrawal can amount to about 50% of the contents or" the crystalllzer,

- where the process is operated semi-continuously. The depth and frequency of withdrawals and replacements depend on the rate at which crystal growth proceeds and the form of crystals produced. By observing the progress of the crystal content of the magma the operator will know whether to increase the depth of frequency of withdrawal or to decrease one or both. Thus. if the magma is losing in crystal content or if a large amount of false grain and undersize crystals appear, the depth of withdrawal can be decreased so as to leave a larger heel to seed the replacement. The frequency of withdrawal may be cut down to give the crystals more growing time to reach normal size, Excessive false grain can be remedied by lowering the actual number of crystals in the magma by either diluting the magma with a fresh quantity of feed solution or by withdrawing a portion of th magma, heating the withdrawal suiliciently to dissolve the crystals therein and returning the withdrawal to the magma from which it was taken. This latter alternative in effect makes a new mixture of the in which the proportion of cry tals to dissolved solids is lowered. Loss of crystal content may also be overcome in many cases by increasing the overload of the replacement syrup. Conversely, if the magma appears to show an increase in normal crystal content, larger withdrawals can be made and/or frequency of withdrawal increased until the maximum yield is being taken, that is until further increases produce injury either to the quantity of crystals or to the form thereof.

The magma being already diluted with alcohol, it is only necessary to withdraw the desired portion and filter the same. The filtration can therefore be carried out without the intermediate steps of diluting and mixing, which steps are usually desirable Where the crystal growth took place in aqueous solution. The filtering can be performed in the same manner as described in connection with aqueous magmas.

It has been found advantageous to the purity of the product, although not to the yield, to wash the filter cake with a small quantity or alcohol. This is equally true whether the crystals were grown in aqueous or alcoholic solution. Substantial increases in the P. N. of the product follow an alcohol wash of the filter cake. The wash is used, as, say, 4 to 5% of the quantity of magma filtered EXAMPLES OF CRYSTALLIZATION FROM ALcoHoLIo SOLUTIONS Example 7000 g. of a solution by weight of P. N. 60 sorbitol in 90% by weight methanol, is seeded with a portion of aqueous magma containing suflicient crystals to give a total crystal con-tent of 1% in the methanol solution. This seeded solution is maintained at 25 in a crystallizer and is stirred continuously. The solution is allowed to remain in the crystallizer for 3 days, after which time the filtration and the recovery of a 10% withdrawal show a 22% yield, a valu which is not sensibly increased after several more days crystallization without withdrawal. The withdrawal is replaced in each instance by an equal quantity of solution of the same strength and purity. Thereafter five daily withdrawals are made averaging 22% in yield. .The percentage of residual sugar and ash in the dried product are about those of the starting material, the P. N. of the product being about 83.

Example 6 "7000 g. of an alcoholic magma are made up from an aqueous magma obtained as described in Example 1. The alcohol used is specially de-. natured formula No. 1 ethyl alcohol which 'is' substantially 4.5% methanol and 95% ethanol. The alcoholic magma as made up has a 51 con-. c'e'ntration of sorbitol of P. N. 65 in 66% alcohol and the separated solids are 6-7 After stirring 24 hours at 25 C. in a crystallizer a 44.5% yield is obtained on withdrawal. Two recyclings are then made at 24 hour intervals, the first 90%, the second 50% of the magma. 24 hours after the last recycling, a 10% withdrawal is made. A 33% yield is obtained. The P. N. is 718 and the ash and reducing sugar are half those of the original. The filtered cake can be dried to a friable, semicrystalline solid.

- Example 7 A semi-continuous process was operated in which sorbitol wascrystallized from an 18% solution in 90 methyl alcohol in a gallon horizontal aluminum crystallizer. The P. N. of the original sorbitol was 80.5. Every 24 hours 45 to of the contents of the crystallizer (135-170 lbs.) was removed and filtered. Fresh sorbitol solution was added to the large heel in the crystallizer and after 24 hours another 50% withdrawal and replacement were made. Twelve withdrawals were made in all.

The magma withdrawal was centrifuged and washed with 90% methyl alcohol by weight in an amount equal to 4.6% of the weight of the magma filtered. The wet cake contained an average of 57% solids. After drying in a vacuum dryer, a white friable product was obtained corresponding to an average yield of 43.5%. This material was easily pulverized and screened. It had an average P. N. of 90 indicating a purity of about 94% and contained .32 and .14 respectively of the percentage of ash and reducing sugar present in the original sorbitol.

Example 8 266 lbs. of crystalline sorbitol as a methanol wet cake from a previous crystallization were mixed with 91.5% methyl alcohol and heated to dissolve the sorbitol. After cooling to 66 F. the batch was seeded with 13.5 lbs. of wet press cake, stirred up to a cream with five gallons of the alcoholic sorbitol solution. The mixture now had the composition 15% sorbitol, 8.3% water and 76.7% methanol. Of the sorbitol, about 2% of it was in the form of individual seed crystals dispersed throughout the mixture.

After stirring for 39 hours, during which time the temperature was gradually raised to 75 F., the mixture was ready to filter. Filtration on a frame press at 40 lbs. per sq. inch pressure yielded a press cake containing 35.9% solids which on drying gave a white, friable product weighing 153 lbs, corresponding to a 56.5% yield. The recovered crystals had a P. N. of 90.3.

The method of the present invention produces from a sorbitol syrup of intermediate purity and relatively low value, two products each of which is more valuble than the starting syrup. A crystalline product of improved purity and in dry form is one of the products. The other product is a syrup of lower sorbitol content which is of value in humectant and conditioning uses. Particularly the syrups of low sorbitol content are of value because they are non-crystallizing. The purification of the syrup is a relative one, as seen from the examples, and the quality of the crystalline product obtained by one crystallization depends upon the quality of the starting syrup. The crystalline product will contain a definite fraction of the ash and reducing sugar content found in the starting syrup. Likewise, the P. N. of the product will be dependent upon the P. N. of the-starting syrup. Where the starting syrup is of intermediate or low purity, it is necessary to recrystallize the product to obtain high purity crystals.

For many purposes the magmatic drying of an aqueous magma produces satisfactory products. This particular type of drying can be accomplishedwhere the starting syrup is of relatively high purity. Where this condition is met, a prodnot of the magmatic drying is a crystalline mass capable of reduction to a powder, as distinguished from the glasses normally produced by the evaporation of sorbitol solutions.

Finally,it is to be understood that the description and instructions set forth above are for the purpose of guiding those skilled in this art in the crystallization of sorbitol. The adjustment of concentrations, temperatures, rates of stirring and the like will be determined by the conditions of operation under which the crystallization must be conducted. The examples are illustrative of some of the possible ways of controlling the crystallization and are meant to be illustrative only and not limiting.

I claim:

1. The method of preparing crystalline sorbitol from an overloaded impure feed solution of sorbitol in water, said sorbitol being produced by the hydrogenation of a sugar and having a pyridine number not less than 50, which comprises slowly mixing said solution at approximately constant temperature with a seed magma consisting of a water solution saturated with respect to sorbitol and containing about 10 to 13% separated solids as well-formed sorbitol seed crystals, said seed crystals having substantial thickness and strength such that a gently stirred overloaded sorbitol solution in water at 25 C. and containing at least 8% of such crystals as separated solids remains mobile, said seed magma being present in an amount of at least 60% of the total mixture; continuing said mixing until the amount of well-formed sorbitol crystals in the total mixture increases to to 13%; withdrawing a portion not exceeding 40% of the said total mixture; separating the crystals in the withdrawn portion from the mother liquor; replacing the withdrawn portion with an equal quantity of fresh overloaded impure feed solution of sorbitol in water; and repeating the steps of mixing, withdrawing, separating and replacing to continue the operation of the method.

2. The method of preparing crystalline sorbitol from an overloaded impure aqueous feed solution containing about 75-82% sorbitol, said sorbitol being produced by the hydrogenation of a sugar and having a pyridine number not less than 65, which comprises slowly mixing said solution at approximately C. with a seed magma consisting of a water solution saturated with respect to sorbitol and containing about 10 to 13% separated solids as well-formed sorbitol seed crystals, said seed crystals having substantial thickness and strength such that a gently stirred overloaded sorbitol solution in water at 25 C. and containing at least 8% of such crystals as separated solids remains mobile, said seed magma being present in an amount of from 60 to 90% of the'total mixture; continuing said mixing until the amount of well-formed sorbitol crystalsin the total mixture increases to 10 to 13%; withdrawing a portion of from 10 to 40% of the said total mixture; separating the crystals in the withdrawn portion from the mother liquor; replacing the withdrawn portion with an equal quantity of fresh overloaded impure aqueous feed solution of sorbitol; and repeating the steps of mixing, withdrawing, separating, and replacing to operate themethod semicontinuously.

3. The method of preparing crystalline sorbitol from an overloaded impure feed solution of sorbitol in water, said sorbitol being produced by the hydrogenation of a sugar and having a pyridine number not less than 50, which comprises slowly mixing said solution at approximately constant temperature with a seed magma consisting of a water solution saturated with respect to sorbitol and containing about 10 to 13% separated solids as well-formed sorbitol seed crystals, said seed crystals having substantial thickness and strength such that a gently stirred overloaded sorbitol solution in water at 25 C. and cone taining at least 8% of such crystals as: separated solids remains mobile, said seed magma being present in an amount of at least of the total mixture; continuing said mixing until the amount of well-formed sorbitol crystals in the total mixture increases to 10 to 13% Withdrawing a portion not exceeding 40% of the said total mixture; diluting the withdrawn portion with a monohydric alcohol having not more than 2 carbon atoms in a quantity to reduce substantially the viscosity of the said withdrawn portion; filtering the withdrawn portion to separate the sorbitol crystals from the mother liquor and diluting alcohol; replacing the withdrawn portion with an equal quantity of fresh overloaded impure feed solution of sorbitol in water; and repeating the steps of mixing, withdrawing, separating, and replacing to continue the operation of the method.

a. The method of preparing crystalline sorbitol as in claim 5 wherein the monohydric alcohol used to dilute the withdrawn portion of the said total mixture is methanol.

5. The method of producing sorbitol seed crystals which comprises mixing pure dry sorbitol with an overloaded water solution of sorbitol with a pyridine number at least 65, the sorbitol content of said solution being at least and sufiicient to constitute a substantial overload with respect to sorbitol; violently agitating said mixture to chop and break up the sorbitol crystals to give a large amount of fast-growing end-face crystal area, until the mixture becomes substantially thickened; thinning the thickened mixture by dilution with several times its weight of fresh overloaded water solution of sorbitol to prevent immobilization of the said thickened mixture; violently agitating the thinned mixture to chop and break up the sorbitol crystals and to disperse them throughout the mixture, until the diluted mixture becomes substantially thickened; and thereafter alternately thinning the mixture to prevent immobilization thereof and violently agitating the thinned mixture until the chopped sorbitol crystals grow into strong well-formed seed crystals and the final mixture of seed crystals and overloaded water solution of sorbitol contains at least 8% of seed crystals as separated solids and is of mobile consistency.

RUDOLPH MAX GOEPP, JR.

CERTIFICATE OF CORRECTION. Patent No. 2,51 ,699. April 6, 19 5.

RUDOLPH MAX GOEPP, JR.

It is hereby certified that error appears in the printed specification of the above numbered patent requiring correction as follows: Page 1, second colunm, line 9-10, for "refuse" read -reduce-; page 5, first column,

line 15, for "sparation" read -separation--; page 8, first column line "58, for "corbitol" read sorbitol--; page 10, second column, line 21, for

the claim reference numeral "5" read 5"; and that the said Letters Patent should be read with this correction therein that the same may conform to the record of the case in the Patent Office.

Signed and sealed this 6th day of July, A. D. 19%.

Henry Van Arsdale, (Seal) Acting Commissioner of Patents.