US2380196A - Process for the preparation of tartaric acid values - Google Patents

Description

" Patented July 10, 1945 PROCESS FOR THE PREPARATION OF TARTARIC ACID VALUES Solomon Soltlberg, Tamaqua, Pa.,assignor to Atlaa Powder Company, Wilmington, Del., a corporation of Delaware No Drawing. Application May 20, 1942, Serial No. 443,694

4 Claims. (01. 260 -528) The present invention relates to the nitric acid oxidation of carbohydrates to dibasic acids, and more particularly to the oxidation of "glucosecontaining carbohydrate material to produce tar-- taric acid.

An object of my invention is to provide a more efficient and economical process for oxidizing carbohydrates, particularly glucose, or a polysaccharide capable of yielding glucose-on mild acid hydrolysis, to theindustrially valuable tartaric acid.

A further object of the invention is to produce a relatively high tartaric acid yield from carbohydrates.

Another object is to provide a process easy to control, economical of equipment and safe to operate.

Other objects will become apparent from the following description.

The controlled oxidation of glucose or its polymers to give high yields of tartaric acid is inherently a difllcult problem. To get d-tartaric acid,

a gluconic chain must be cut selectively between I the 4 and carbon atoms, which scission. also yields oxalic acid, a valuable by-product. Forthe highest yield of oxalic acid compatible with production of theoretical tartaric acid, it should be the only point of attack.- That is, the theoretical'yields of tartaric and oxalic acids from the oxidation of glucose are 83.3% and 50% respectively based on the glucose. Thus:

.CHO coon Hi'JOH noon M. wt. Yield (onglucose) no H 0, no 11 150 83.3% n on 00H ist r. a

m. door! 50% HIGH M.W.=l80

My investigations show that the oxidative decomposition or glucose by nitric acid can simultaneously follow several paths. The primary oxidatlonproducts seem to be gluconic and saccharic acids, but 2- and fi-ketogluconic acids can occur. These intermediates in turn canbe broken down to give S-carbon and B-carbon acids in addition to mesoand d-tartaric and oxalic acids. The tartaric acids and all materials containing 2 or more carbon atoms can be broken down further to oxalic acid and carbon dioxide.

Tartaric acid is ordinarily one of the more easily attacked oxidation products, and, consequently, oxidation normally proceeds beyond it to oxalic acid. In the presence ofdirective catalysts, however, oxidation has been carried out so as to produce tartaric acid. So far as is known, however,

hydrates to tartaric acid have not heretofore proved commercially operative. As a result of my study of the reaction it appears to me that this is largely .due to inherent process inefliciency and resultant lack of reproducibility of results.

Since tartaric acid is an intermediate product of reaction and the reaction proceeds so complexly and by different mechanisms, a reacted solution contains a number of different products of differing stages'of oxidation. Those present to the greatest degree are the ones most resistant to nitric acid under the conditions employed. Usually if catalysts directive for tartaric acid are employed, the oxidation yields a mixture of tartaric acid, oxalic acid, and 'a residue of acids of lower stages of oxidation which appears largely to be made up of saccharic acid. Tartaric acid, of course, is present only to the extent that it has not been further oxidized to oxalic acid and carbon dioxide. .In any given oxidation to produce tartaric acid, a considerable quantity of the tartaric acid formed during the reaction is decomposed. This decomposition could be eliminated and yield improved if tartaric acid could be removed from the solution as it is formed and before any opportunity for attack upon it by nitric acid. In practice, however, this is impossible since there are no available separation procedures for the cheap removal of small increments of tartaric acid from the nitric acid solution.

Another loss of potential tartaric acid has been in the, residue of intermediate oxidation products. It has been proposed to oxidize the residue further after removal of tartaric and oxalic acids, and some improvement in overall yield has been effected by this means. However, such procedure has not adequately remedied inefilcient perform: ance in the first oxidation step, and the second oxidation has been difficult to perform. It has necessitated a second and entirely different set of operating conditions and resulted in further losses of tartaric acid by destructive oxidation and a further residue of partially oxidized material.

As an oxidizing agent nitric acid ossesses the advantages of cheapness and re'generability, but

its action depends upon a complex decomposition to nitrous acid and oxides Of nitrogen so that the oxidizing power varies continually and in a manner very difiicult of exact control throughout the period of use. Its reaction with carbohydrate materials igaurocatalytic which makes control of temperature and concentration diflicult to achieve. The reaction proceeds at first with too great violence and then tapers oi! as nitric acid becomes used and oxidation products formed. The violent phase of the oxidation must be kept in hand if the reaction is to proceed to ood yields of the desired products. Processes hereprocesses for the nitric acid oxidation of carbo- '60 more po h v n wmp adequate which readily hydrolyze tohexoses.

dioxide which hydrolyzes readily to erythritol 60' 2 I 2,seo,19e

control'of the oxidation and as a result have been inefficient.

I have invented a process which is more efficient, subject to better control, more economical and safer to operate than the previously known 5 processes for making tartaric acid from carbohydrate material. First, my process is carried out in cyclic manner. In each cycle I oxidize both fresh carbohydrate and residue from a previous oxidation with nitric acid. The reaction is preferably performed under carefully controlled and stage-wise conditions. Residue from each cycle is oxidized with fresh carbohydrate in each next succeeding cycle, and so on.

By operating in my manner a number of advanl5 tages are achieved. The oxidation of the mixture of partially oxidized material and fresh carbohydrate proceeds more evenly and completely to produce the desired product. All residue is subjected to further oxidation so that no material capable of forming tartaric acid or oxalic acid, the recoverable by-product, is wasted, and the reaction proceeds with but one set of constantly repeated conditions. I have found that by operating so as to obtain a smaller conversion per pass I am able to obtain a greater overall yield over a succession of cycles. This represents a removal of tartaric and oxalic acids at an earlier stage in the oxidation, and, hence, approaches the.

ideal of removing these acids as they are formed.

The conversion per cycle should be suflicient, however, to permit adequate removal of product. I have found a number of features of procedueand reaction control are important to the obtention of optimum yields. terial I may use water-soluble carbohydrate materials oxidizable in solution by nitricacid to carbon dioxide, tartaric acid and other soluble compounds, and a residue containing intermediates further oxidizable to tartaric acid. By the 40 term "carbohydrate material, I mean to include not only compounds. containing hydrogen and oxygen in the proportions of water but also other polyhydroxylic materials, such as hexitols, penti- .tols, erythritol, sugar acids, including aldonic and ketonic acids, other similar materials and'materials readily hydrolyzable to these. Such materials are now commonly classed with carbohydrates. Most hexoses, such as fructose or mannose, and particularly d-glucose which is available in pure form free from ash, of constant composition, and at a low price, are readily usable in my process. Carbohydrate materials readily hydrolyzable by acid to other starting materials are equivalent to those materials. Among such ma- '55 terials, I include oligoor poly-saccharides, such as starch, dextrine, corn syrup, sucrose and hightest molasses (partially inverted raw can sugar) Butadiene products, for insolubles are not Dracticallly further oxidizable in my reaction. With the exception of d-glucose and sucrose, the' at present commercially available saccharides (economically useful) have relatively high ash contents. The

ash content causes difliculty in acyclic process 76 For starting ma- 35 due to build-up in the residue and usually not as many cycles can be performed. Sucrose while obtainable ash-free gives lower yields of d-tartaric than does glucose. It is understood that various less desirable 'saccharide materials can be rendered suitable as starting materials by proper deashing or demineralizing as, for example, by ion exchange methods.

I may introduce nitric acid oxidizing agent into the reaction mixture as such, in which case I prefer that it contain a little lower oxide of nitrogen which appears to act catalytically toward the reaction; or I may form it in situ by passing into the reaction mixture a mixture of oxidizable nitrogen oxides and air or other oxygen containing gas or by using higher oxides of nitrogen which have been formed outside the reaction mixture.

' The reaction requires a catalyst to-give a practical yield of tartaric acid. Any oxidation catalyst specific for tartaric acid product, a number of which are known in the art, can be used. Various polyvalent metal compounds are operative for this purpose, including compounds of vanadium, manganese, iron and molybdenum. Vanadium in the form of its soluble pentavalent compounds,

such as sodium orthovanadate hexadecahydrate nonrecoverable nitrogen as N2 or N20, can be converted to nitric acid, without further purification, either at atmospheric pressure, or, after compression, as for example, to the customary 6-7 atmospheres; or they may be oxidized and absorbed in the reaction mixture.

The oxidation process is carried out on a commercial scale suitably in stirred, stainless steel reactors fitted with internal cooling coils, and a means for recovering the' nitric acid fumes evolved.

Control of the various operating variables is necessary for uniform operation and optimum yields. Uniform operation, in turn, calls for approximately constant residue composition and an approximately constant total feed composition, hence, an approximately constant carbohydrate-residue ratio. The amount and composition ofthe residue are determined by the independent operating variables, such as acid-carbohydrate ratio, catalyst nature and amount, in-

itial acid strength, heating schedule, and, to a minor extent, equipment design; or these, the

,, acid-carbohydrate ratio is preferably used for control.

As the cyclic process is initiated with fresh carbohydrate and built up to equilibrium, the composition of the residue, and, hence, its behavoir on oxidation, changes somewhat until after several cycles, the equilibrium point is reached where substantially constant tartaric acid yields are obtained.

When this equilibrium is reached, the size of the residue, and hence the point to which the oxidation has been carried, is controlled, of course, to give the optimum yield of tartaric acid. As has been noted above, the best possible yield of tartaric acid would require removal of the acid asformed, and one of the important advantages of material which can be conveniently recycled and the amount of product necessaryfor a'prac- '5 tical separation do limit .the extent to which the q ideal *is practically obtainable. I have found that with glucose good yields "may be obtained by controlling oxidizing conditions so that a weight ratio of residue to fresh glucoseof from about 0.5 to about 1.0 is maintained. This residueglucose ratio for equilibrium recycling conditions, is achieved by using an acid-glucose mol ratio of from about to 1.5 and preferably about 6 to about 7.5 mols of 100% nitric acid per mol of glucose, and holding the other operating conditions within the limits hereinafter disclosed.

' When starting up the cycle, and glucose alone is being oxidized, the ratio should be about 5, and this should be gradually increased for successive cycles, to a point between about 6 and 7.5 (determined bybatch size and the characteristics of the reaction vessel), until the cycle is in equilibrium. The initial aqueous nitric acid concentration in the reaction mixture should prefer-v ably be from 45-70% HNO; (calculated as aqueous HNO; without reference to amount of sugar and/or residue). The effect of changing this concentration is reflected principally in the violence of the reaction. The oxidation of carbohydrate to tartaric acid will proceed with nitric acid as weak as but the yields are inferior action in a series of temperature stages.

' duction or heating-up stage.

a short time. The blow stage is maintained to those obtainable with higher acid strengths, and the reaction times much longer. Similarly, nitric acid stronger than 62-63%, up to 10%, can be used,'but theyields are not as good, and the reaction becomes much more diflicult to control. The control of temperature in the oxidation is highly desirable to insure uniformity of product and yield and also to make the reaction largely independent of external heating by conserving the heat generated'by the reaction. I have found it of considerable advantage to conduct the re- During the mixing of ingredients I prefer to maintain a low temperature of about C. or lower. Following mixing, the temperature should be raised gradually or allowed to rise spontaneously to about to C. This period is called the in- Thelength of the heating up stage is subject to considerable variation and may be shortened by the addition of a nitrite radical or by heating. By prolonging it the violence of the blow stage described below may be moderated, but the time necessary for i5 the oxidation is thereby increased. During the heating-up period intermediate oxides of nitrogen are formed from the nitric acid. The presence of intermediate oxides or acids derived therefrom is necessary for the reaction. A period of 20 to minutes represents a desirable compromise between these condition under the reaction conditions I employ. when the temperature reaches 30 to. 35 C. an autocatalytic and, at least at first, strongly exothermic reaction called the -blow" sets in. The temperature rapidly rises and under most operating conditions it is necessary to cool the reactants strongly so that the temperature does not exceed 75 C. for more than etc. temperature of about fill-75. 0., and preferably" arcs-'10? (2., for aperiodof time the length of which is not critical (butshould be constant 1 for uniform results) and may be determined empirica'llygfor operating convenience. Generally this stage of the reaction will be allowed to proceed from 45 to minutes. The final temperature stage of the oxidation is the fume-oi!" stage in which the last of the nitric acid is reacted and passed, off as lower nitrogen oxides. nitric acid and nitrogen oxides is necessary for eincient recovery of products from the reaction mixture. During the fume-oil!" stage the reac- Absence of tion mixture is maintained at a high tempera-' mits temperature conditions more nearly appro-- priate to the different oxidation stages.

Optionally my process, at least at certain stages, may also be operated either under reduced or superatmospheric pressure.

' The oxalic and tartarlc'acids can be recovered from the oxidized reaction mixture partly, in the case of oxalic acid, by direct precipitation, by ester separation as described below, and by a number of known methods such as by crystallizing in the form of acid potassium salt, zinc salts, calcium salts or lead salts, although for practical reasons one type of recovery may be preferred over another. My cyclic .process is of definite benefit regardless of the recoverystep used.

Methods of operation are illustrated by the following examples:

EXAMP LE 1 RUN 1 An initial, non-cyclic oxidation, to prepare an oxidizable residue for use in a cyclic series, was 7 performed as follows:

500grams crystalline, anhydrous d-glucose were dissolved in 875 cc. chilled 70% nitric acid. To

this solution were added 19 cc. aqueous solution containing 0.075 gram sodium orthovanadate hexadecahydrate as catalyst. 'The entire solution was then transferred to a water jacketed.

stainless steel reactor and cc. water were added. At this time the acid had a strength of 62% and was present in the ratio of 5 mols acid to 1 mol glucose. Heating was started and-the temperature of the solution rose to 32 C. in the course of 78 minutes. Then the blow" cbmmenced. The temperature of the solution was controlled by passing .cooling water through the Jacket of the reactor so that it did not exceed 76 C. (except for a short time at the start when the reaction was most violent). Thereafter the I solution was kept at 65-70 C. for 90 minutes and. then heated at 94-96 C. on a water bath, until nitrogen oxides were no longer detectable by test with starch-iodide paper. It required over 6 hours to drive of! the HNO: completely.

The above constituted the preparation of an 1 oxidized mixture. The oxidized product was then cooled for 24 hours at 5 0., after which a crop of 97.3 gr crystalline oxalic acid was,removed by filtering. 41.5 grams of basic zinc carbonate were added to the nitrate (equivalent to 15% of the total titratable acidity) to precipitate residual dissolved oxalicacid as thezinc salt.. After iiilation of averages.

tering to remove the zinc oxalate, a further 138.5 grams of basic zinc carbonate were added to the filtrate with stirring at 45-55 C. to precipitate the tartaric acids as zinc salts. After filtering off the zinc tartrates the filtrate was concentrated by evaporation and a residue syrup weighing 375 grams recovered which contained about 90% solids.

Recovery data for this run are listed in the table below as Run'l.

RUN 2 vide 0.015% catalyst (calculated as hexadecahydrate) based on sugar. The entiresolution was transferred to the reactor and the acid concentration adjusted to 62% by adding. water as in Run 1. Reaction was induced b heating the solution as in Run 1. In this case, however, the "blow" started in 55 minutes at 31.0 C. at which time cooling water was run through the Jacket and the temperature controlled to-a maximum temperature of 52.5 C. The rate of heat evolution in this blow was substantially milder than the corresponding stage of Run 1. The HNO: fumeof! was conducted as before and the oxidized product worked up to remove oxalic and tartaric acids as in Run 1; Recovery of the residue in this case yielded a syrup (about 90% solids) weighing 505 grams, for recycling.

Subsequent runs were made according to the procedure of Run 2, nine runs constituting the series. Nitric acid initial strength was adjusted as necessary in these runs by dilution with water to give 6263% solutions. The ratio of acid to sugar and residue was adjusted from run to run to keep the weight of residue fairly constant. If the residue appeared to be increasing the next oxidation was given a higher acid ratio, and vice versa.

ing with d-glucose'the tartaric acid yields (calculated from yields of separated salts) have been found to average more than 42% under preferred Y conditions such as those illustrated in Example 1. The oxalic acid yields average about 44% (as oxalic acid dihydrate). Furthermore the process The following table gives the results for the Y nine run series just described. Run 1 was not cyclic and is therefore omitted from the calcu- Table I Yield of oxalic acid cac to a 0 rec Run (calculated acid and g fgg" fjg ggy,

from Zn amount of Wei ht u at mm salt) per acid calcug s g 100 g. sugar lated from Zn salt) per 100 g. sugar Grams Grams Grams "C.

pure glucose, the amount and nature of the residue reaching a steady state. Hence, while the acid strength or ratio to sugar and residue may .have to be'varied slightly during the first few cycles, while the system is approaching the steady is readily controllable, much less violent. and considerably safer to operatethan the prior oxidation processes. I

By means of my process cheap sources of gludose can be used to substantial advantage. Yields are not as good as with pure d-glucose and the number of cyclic runs is limited by the extent ditions it will be preferable to use this type of raw material instead of pure d-glucose. The following example illustrates one series of five runs using a crude glucose source.

' EXAMPIE 2 A typical residual corn syrup obtained as a byproduct in the process of preparing d-glucose from cornstarch and analyzing about 71% reducing sugar'as dextrose, 4.5% ash (including sodium chloride), and 79% total solids, was used. This material is known to be a mixture of various di-, tri-, and tetraglucosides, with some glucose, and some of the acid reversion products of glucose.

The reactor was a 3-liter flask cooled byimmersion in a large water bath with initial temperature of 24 C. v

The initial run was performed by mixing together 746 grams of the corn syrup described, containing 600 grams solids, 1050 cc. concentrated nitric acid, 22.2 cc. water, and 0.09 grams sodium orthovanadate hexadecahydrate in 22.5 cc. water. After mixing the ingredients, the temperature of the solution rose to 30 C. The blow started 25 minutes after mixing, rose to 87 C. in 3 minutes, and was then kept at 65-70 C. for 45 minutes by running cooling water into the water bath. The flask was then removed from the water bath and the temperature permitted to rise spontaneously (reaching about 90 0.). The flask was thereupon packed in sawdust (for heat insulation) until the nitric acid had been driven off by the heat generated by the reaction.

Oxalic and tartaric acids were thereafter removed from the product of this oxidation by the method of Example 1. The residue obtained after concentrating the remaining liquid consisted of 450 grams solids.

For the second run (first cyclic run) the resiacid,.and 012- gram sodium orthovanadate hexadecahyd'rate in 30, cc. water. The blow started in 30 minutes and the solution reached a maxi mum temperature of 825 C. The reaction was controlled and nitric acid removed as in Run 1';

Oxalic and tartaric acids were removed as before and the remaining liquid concentrated to yield a-rcsidue syrup (about 90% solids) for recycling.

Three further cyclic ms were made following the same procedure. The-details and recovery figures for these five runs are-tabulated below,

Run 1' being the non-cyclic oxidation first described:

Table II Yield of oxalic acid t ii a (3th??? or ncaci oao ee (calculated acid and Resume Runt fro-m Zn amount of syrup HNOs/mol. blow salt) per acid calcu- Welght sugar temp 100g. sugar lated from The accumulation of ash in the corn syrup series limits the number 015 cycles to about 5. The residue of the fifth run can be used in another cyclic series by removing the ash by suitable procedures. Alternatively, the residue of the fifth run can be oxidized without adding more sugar to it and a further amount of product, principally oxalic acid, obtained.

The average yields of the cyclic process in Example 2 are 29.6%- tartaric acids and 44.02% oxalic acid. For comparison, the residue from a single oxidation (not cyclic) was reoxidized and a total yield of 27.7% tartaric acids and 41.6%

oxalic acid obtained from both. This stepwise oxidation is not only less efficient with respect to yield but is a longer process involving more steps and also a more expensive process to practice.

The cyclic process of this invention can be carried out with lower concentrations of nitric acid than are shown in the preceding examples.

EXAMPLE3 The process of Example 1 was carried out with 50% initial concentration of nitric acid instead of 62-63% but with other conditions of temperature and catalyst the same. The ratio of acid to sugar was increased progressively in each ofthe four runs to keep the weight of residue fed back approximately constant. The oxalic and tartaric acids were separated as the zinc salts. 400 g. batches of glucose were run using a stainless steel reactor. The following table lists data for the initial oxidation (Run 1) and two cyclic oxidations according to this process:

In the runs of Example 3, owing to the dilution, the blow was much less vigorous, the temperature did not rise as high, and a much longer period was required for the boil-off. Nevertheless, the characteristic improvement in tartaric acid yield from the recycling was evinced.

The zinc salts of oxalic acid can be worked up for saleable oxalic acid values by treatment with soda ash to produce sodium oxalate and regenerate the basic zinc carbonate, or by other known methods. The zinc tartrates can be converted to free tartaric acids, (9; mixture of dand meso in which the .dpredominates) by treatment with oxalic acid, and the free acids resolved, either by fractional crystallization of the free acids or by means of the acid potassium salt.

The recovery of oxalic and tartaric acids may likewise be effected in a similar manner by means of their calcium salts.

I have discovered a third method of product removal which I call ester-separation. This method is particularly usefulwhere the products of my reaction are desired in the form of esters such as dibutyl tartrate or where it is desired to avoid contamination with inorganic compounds.

In this ester separation the mixture of acids re-. suiting from the oxidation are esterified by ordinary esterification procedures to produce diesters of lower alcohols, such as those having less than 6 carbon atoms and particularly butyl alcohol. For example, the butyl ester may be prepared by refiuxing the reacted solution with butyl alcohol and permitting water to distill oif. The esters of the tartaric and oxalic acids are then.

separated from the other esters as by solvent extion carried out in batch operation. However,

my process can also be'conifniently operated in a continuous manner. :For example, nitric acid, carbohydrate material and residue may be continuously fed into a reaction system made up of successive zones in which are maintained temperature conditions corresponding to the desired temperature stages such as the blow and fume-ofi periods, and solution from which recovery of product is to be made may be continuously drawn oif. Under continuous operation I have found the heating-up stage may usually be merged with the blow.

Optimum conditions vary for difierent types of the specific examples shown,,but is of general 5 applicability to various types of carbohydrate oxidation to tartaric acid, and is to be held as limited only by the scope of the appended claims.

Having described my invention, what I claim is:

1. A cyclic process for the production of tartaric acid values which comprises oxidizing in aqueous solution a mixture of a carbohydrate material selected from the group consisting of glucose,

fructose, pentoses, gluconic acid, 5 ketogluconic acid erythritol, and materials rapidly hydrolyzable to these by dilute nitric acid; and an oxidation residue from a previous oxidation, with nitric acid, in thepresence of a soluble vanadium compound catalyst specific for tartaric acid production, conducting the reaction in a series of temperature stages comprising a prior stage conducted substantially entirely at a temperature of not more than approximately 75 C. and a, later stage conducted at a temperature of at least approximately 90 C. and below the boiling point of the reaction mixture until evolution of nitrogen oxides ceases, removing tartaric and oxalic acids from the reaction, thereby leaving said oxidation residue, adding additional carbohydrate material to said residue and repeating the cycle.

2. A process according to claim 1 wherein the carbohydrate material is d-glucose and the ratio of residue to d-glucose is within the range of from approximately 0.5 to approximately 1.0.

3. A process according to claim 1 wherein the carbohydrate material is d-glucose and the ratio of residue to d-glucose is within the range of from approximately 0.5 to approximately 1.0, and within the range of from approximately 5 to approximately 7.5 mols of nitric acid per mol of glucose are employed for the reaction.

4. A process according to claim 1 wherein the tartaric and oxalic acids are removed from the reacted mixture by esterification of the acids in the reacted mixture with a lower alcohol, separating the tartaric and oxalic esters from the esters of other acids in the reacted mixture, hydrolyzing the said esters of other acids to reform those acids thereby forming the residue.

SOLOMON SOLTZBERG.