A PROCESS FDR THE RECOVERY OF ASCORBIC ACID
Background of the Invention
The present invention relates to a process for the production of ascorbic acid. More particularly, the present invention relates to the recovery of ascorbic acid from aqueous solutions containing the same in dilute concentrations.
As described, e.g., in Kirk-Othmer 's Encyclopedia of Chemical Technology, Third Edition, ascorbic acid (L-ascorbic acid, L-xylo-ascorbic acid, L-threo-hex-2-enonic acid g-lactone) is the name recognized by the IUPAC-IUB Commission on Biochemical Nomenclature for vitamin C. The name implies the vitamin's antiscorbutic properties, namely, the prevention and treatment of scurvy. L-ascorbic acid is widely distributed in plants and animals. The pure vitamin (CόHsOό, mol. wt. 176.13) is a white crystalline substance derived from L-gulonic acid, a sugar acid, and synthesized both biologically and chemically from D-glucose.
Although natural and synthetic vitamin C are chemically and biologically identical, in recent years a limited amount of commercial isolation from vegetable sources, e.g., rose hips, persimmon, citrus fruit, etc., has been carried out to meet the preference of some persons for vitamin C from natural sources. L-ascorbic acid was the first vitamin to be produced in commercial quantities, and manufacture is based on the well-known Reichstein and Grussner synthesis, which involves the steps of hydrogenation of D-glucose to D-sorbitol; fermentation (oxidation) to L-sorbose; acetonation to bis-
isopropylidene-a-L-sorbofuranose; oxidation to bis-isopropylidene-2-oxo-L- gulonic acid, and hydrolysis, rearrangement and purification to L-ascorbic acid.
A direct fermentation of glucose to ascorbic acid would be very attractive, saving on operations and on expensive reagents, in addition to its being derived from a natural fermentation process, as opposed to a synthesis involving chemical steps. There are indications that such direct fermentation to ascorbic acid is feasible. Yet industrial production of ascorbic acid dirough direct fermentation seems impractical, in view of the low product concentration in the fermentation liquor, which normally is in the range of less than 0.7 mol/kg. Purifying the ascorbic acid by conventional methods would result in a purified product of concentrations similar to those in the fermentation liquor. Due to its high solubility in water, me cost of ascorbic acid crystallization by water evaporation would be prohibitive.
Several methods were proposed for combining purification of carboxylic acids with their concentration. In the case of citric acid, it is achieved by die addition of lime to crystallize calcium citrate, which has very low solubility in water. This salt is separated, washed and acidulated with sulfuric acid. Purified and concentrated citric acid is obtained. This method is not applicable for ascorbic acid, as its alkali and alkali earth salts are highly soluble.
A process was proposed in which carboxylic acids were extracted and then displaced from me extractant by a solution of concentrated mineral acids. Both liquid (long chain amines) and solid (resins carrying amine groups) anion exchangers could be considered for this purpose. The purity of the displaced carboxylic acid depends on the preference of the extractant to the mineral acid.
Such a process might be applicable for ascorbic acid separation and concentration, provided that the extractant is strong enough to reach high extraction yield, d at it shows high preference to die displacing acid, and diat the ascorbic acid is stable at die high acidity of d e displacing solution.
The regeneration of die anion exchanger would require neutralization by a base. Using HCl as the displacing acid and distilling it of d e extractant was proposed, but die high temperatures required and die extractant' s decomposition at diese conditions are prohibitive. If die anion exchanger is represented by B, the ascorbic acid in die fermentation liquor and in d e pure form are AAF and AAP, respectively, the displacing acid is HCl, and me neutralizing base is NaOH, die equations of die process stages and of die overall reaction are as follows:
B + AAF → B-AA
B AA + HCl → B-HC1 + AAP
B-HC1 + NaOH → B + NaCl +H2O
AAF + HCl + NaOH → AAP + NaCl + H2O
Reagents are consumed, and a by-product salt of no (or negative) value is produced.
Thus, despite die widely felt need for a more attractive process to meet die exceedingly high demand for ascorbic acid, to date no such process has been commercialized.
Summary of die Invention
With die above-described state of the art in mind, according to die present invention diere is now provided a process for die recovery of ascorbic acid from an aqueous feed solution containing said acid at a concentration of less dian 0.7 mol/kg, comprising adsorbing a major portion of said ascorbic acid widi a solid phase adsorbent resin selected from resins carrying a pyridine function and resins of similar or weaker basicity; separating said ascorbic acid- containing resin from residual aqueous solution, and subjecting said ascorbic acid-containing resin to a desorbing operation wim a neutral solvent at a temperature of at least 20°C higher dian die temperature at which said adsorbtion is carried out, whereby diere is obtained a solution of ascorbic acid in solvent in which die concentration of ascorbic acid is at least equal to its concenttation in said aqueous feed solution.
The basicity of water-soluble bases is determined by die pH of dieir solutions. That of water-immiscible bases (fatty amines, basic resins) cannot be measured direcdy. Their apparent basicity is determined by various methods having one element in common: die water-immiscible base is contacted widi an acid-containing aqueous solution. The degree of acid transfer from d e aqueous solution into die water-immiscible base, or more particularly, die pH of die aqueous solution in equilibrium with d e base, shows die apparent basicity. A dieoretical treatment is given in several articles, including "Basicities of Weak Base Ion Exchange Resins," by Gustafson, et ai, Ind. Eng. Chem. Fundam., Vol. 9, p. 221 (1970). The authors of this article have studied the basicity of several resins by equilibrating diem with aqueous amine-amine hydrochloride buffer solutions, followed by determination of the degree of neutralization α of
die resins as a function of d e pH of die solution. As explained in die article, me pK of die resin is calculated from pK = pH - log oJ(l-d)
Using diis mediod for poly(2-medιyl-5-vinylpyridine) cross-linked with 5, 7 and 10% divinylbenzene gave a pK of about 4.
Using o er basicity measurement metiiods (for example, equilibration wim HCl + NaCl solutions, as proposed by Nagasawa, et al., Mem. Fac. Eng. Nagoya Univ., Vol. 10, p. 105 (1958) could result in different pK values for a particular resin. Yet, die relative basicities of resins can be determined by comparing their apparent basicity by one of die known methods. The resins suitable for die process of die present invention are ose carrying a pyridine function and resins of similar, or weaker, basicity.
In preferred embodiments of me present invention, mere is obtained a solution of ascorbic acid in solvent in which die concenuation of ascorbic acid is higher dian its concentration in said aqueous feed solution.
In die above process, at least 90% of said ascorbic acid is adsorbed by said solid phase adsorbent resin from said aqueous feed solution.
In preferred embodiments of die present invention, said solid phase resins are polyvinylpyridine polymers such as poly 2- and poly 4-vinylpyridine free base gel or macroreticular resins exhibiting a bead form. These resins are preferably at least about 2% cross-linked, and more preferably, at least about 8% cross-linked, widi a suitable cross-linking agent, desirably divinylbenzene.
More preferred resins to date have been 2% to 25% cross-linked, bead form poly 2- and poly 4-vinylpyridine polymers. For example, preferred polymers in work to date have been poly 2- and poly 4-vinylpyridine resins available from Reilly Industries, Inc., Indianapolis, Indiana, in the REILLEX™ polymer series. These REILLEX™ polymers are 2% to 25% cross-linked, and exhibit good thermal stability and adsorptive and desorptive capacities and other preferred features as described herein.
The preferred resin beads can be of any suitable mesh size, for instance, from about 20 to about 60 mesh. Further, me resins can include a minor amount of functionalization of ti eir pyridine groups, which minor amount can include, for example, functionalization to pyridine N-oxide or quaternary salt species. This functionalization has been incorporated to modify the relative basicity of die non-funcuonalized pyridine groups and diereby to modify tiieir adsorptive and desorptive properties.
U.S. Patent 2,443,583 describes ascorbic acid separation dirough adsorption on an anion exchanger, followed by elution by a strong mineral acid, preferably H2SO4. The excess of sulfuric acid is separated from die ascorbic acid dirough precipitation as gypsum. The anion exchanger is regenerated by a base, rinsed and acidified by a weak acid such as carbonic acid.
The process of said patent is complicated, consumes a strong acid and two bases, and forms two by-product salts of low or negative value. An objective of die process of die present invention is to avoid or reduce such chemical consumption and by-product formation.
In the process described in U.S. Patent 5,391,770, ascorbic acid is displaced from its alkali metal salt by a strong acid to its aqueous metiianol solution, wherein die alkali salt of die mineral acid is only sparingly soluble. After separating die salt, die ascorbic acid-containing solution is passed dirough a cation exchange and anion exchange resins in order to remove the residual alkali metal salt of strong acid wid out adsoφtion of die ascorbic acid. The ascorbic acid is men isolated from d e solution.
The process of U.S. Patent 5,391,770 is designed for separating ascorbic acid produced from 2-keto-L-gulonic acid and not ascorbic acid produced by fermentation, as in die present invention. The product concenuation and die compositions and contents of die impurities accompanying it are also different. Therefore, U.S. Patent 5,391,770 does not teach purification and concentration of a fermentation product.
Kulprathipanja, in U.S. Patent 4,720,579, proposes a process for separating citric acid from a fermentation brotii by contacting widi a polymeric adsorbent selected from d e group consisting of an insoluble crosslinked polystyrene polymer and a non-ionic hydrophobic insoluble polyacrylic ester polymer at adsorption conditions selected to selectively adsorb said citric acid. In another patent, U.S. Patent 4,851,573, Kulpratiiipanja proposes an adsorption process for separating citric acid from a fermentation broth by contacting widi a water-insoluble, weakly basic, anionic exchange resin possessing tertiary amine or pyridine functional groups, at adsorption conduons selected to selectively adsorb said citric acid, desorbing said citric acid widi a desorbent comprising water or a dilute inorganic acid at desorption conditions,
said adsoφtion conditions including pH lower dian die first ionization constant of citric acid. This patent directs a strong preference for desoφtion by a dilute sulfonic acid, because in some cases water is not strong enough to recover the adsorbed citric acid. Desoφtion with a neutral solvent at a temperature of at least 20°C higher man die temperature at which die adsoφtion is carried out, is not claimed or exemplified. In fact, the second patent states, "Desoφtion conditions will include die same range of temperature and pressures as used for adsoφtion condtions. "
U.S. Patent 4,323,702 claims a process for recovering a carboxylic acid from an aqueous solution by adsoφtion on a polymeric material having a pyridine skeletal structure and a cross-linked structure, followed by desoφtion dirough die use of a desoφtion agent selected from the group consisting of an aliphatic alcohol, an aliphatic ketone, and a carboxylic ester. The list of suitable carboxylic acids (column 3, lines 24-39) does not include ascorbic acid, which is not a carboxylic acid. The examples use propionic acid, benzoic acid, phdialic acid, malonic acid, tartaric acid, adipic acid, citric acid, metiiacrylic acid and acetic acid, all of which are carboxylic acids and not lactones.
Said patent claims that die resin is effective, even if die temperature of adsoφtion is high (column 3, lines 53-55), teaching away from elution at elevated temperatures. Elution at a temperature higher d an die adsoφtion temperature was not shown in the examples of said patent. Furthermore, medianol and acetone were used as die desorbing agents in die examples. A nearly complete recovery of d e acid from its aqueous solution was not shown or claimed, particularly not with a resin after being used and eluted.
The invention of PCT Application No. WO 93/00226 is directed to an extractive fermentation of lactic acid, whereby brotii is continuously removed from d e fermentor, separated from die cells and passed dirough a polymer phase-containing pyridine group. The main goal is to maintain die pH and the lactate concentration in d e fermentor at levels mat reduce die product inhibition in the fermentor. Elution (desoφtion) of die adsorbed acid is very briefly referred to: "The adsorbed lactic acid can be recovered using a suitable desorbing agent. Suitable desorbing agents will include, for example, polar organic solvents such as alcohols (e.g., methanol) as well as hot water" (page 10, lines 19-22).
Example 6 of said PCT application uses 5% solutions of NH3, H2SO4 or HCl for lactic acid desoφtion. Examples 2, 4 and 5 use medianol. No examples are given for e use of water for lactic acid desoφtion. No claim is made in said application to desoφtion at a temperature higher man that of die adsoφtion, or to obtaining the desorbed product at a temperature higher man that of die feed solution.
PCT Application WO 92/16490 relates to a process for recovering citric acid from a medium comprising it. In one preferred embodiment, me medium is contacted with a solid-phase, free base polymer having tertiary amine functions to adsorb citric acid, which is dien desorbed by displacement widi a stong acid, e.g., H2SO4 or HCl. In anodier preferred embodiment, die medium is contacted widi a solid phase, free base polymer having pyridine functions at a temperature below about 40°C to adsorb citric acid, which is dien desorbed with hot water at a temperature of at least about 75 °C. No claim is made to achieving a product at a concentration higher d an diat of die feed.
In Example 1 of said application, a 10% citric acid solution was passed dirough a polyvinylpyridine polymer resin until die resin was saturated. The resin was dien rinsed with CO2 saturated water, and men was washed wid water at 85 °C. The citric acid concentration in the aqueous solution obtained (desorbate) was not given in the example.
In Example 3, a polyvinylpyridine resin was used in processes as described in Example 1 , and the collected desorbed fluids were put back into die column after anodier saturation and rinse cycle, instead of water. The internal column temperature was brought to at least 85°C. According to WO 92/16490, "Using that technique, a concentration of up to about 10% citric acid is achieved in two cycles. Additional cycles can be performed to further increase citric acid concentration, but in Applicant's work dius far, due to decreasing usable capacity of die resin wi each cycle, die best efficiency has been achieved after two cycles." Thus, Example 3 teaches mat in order to desorb citric acid at concentrations similar to those of die feed, desorbate should be recycled to desoφtion. As a result, die desoφtion is not completed and die resin loses capacity in die next cycle.
Another aspect diat was not referred to in PCT WO 92/16490 is that of the completion of citric acid recovery from the brotii. Any acid left diere forms a product loss. No data is given in die examples on how complete die recovery is. Yet, it is clear diat desoφtion wid citric acid-containing solution not only decreases d e resin's capacity, but also decreases d e efficiency of citric acid recovery from the medium.
In fact, die Applicant of PCT Application WO 92/16490, Reilly Industries, Inc., togetiier with Advanced Separation Technologies, Inc., optimized and piloted tiieir process. The pilot program results, as published in October, 1994, show a product concentration of 10% citric acid, compared to a feed concenuation of 15%. The recovery was 95% or higher. Therefore, diese results teach diat operating die process at conditions allowing high recovery of die acid, results in a product that is more dilute dian the feed.
Ascorbic acid is not a carboxylic acid, and one could not draw analogies from other acids as to its behaviour in adsoφtion on pyridine-based resins and in desoφtion. Yet, if such analogies could have been drawn, d ey would have indicated diat product concentration on adsoφtion, followed by desoφtion at elevated temperamre, is not attainable. An earlier publication by Reilly Industries, Inc. [Ernst and McQuigg, Paper No. 5AE, AICAE National Meeting (1992)] states: "The shape of die 25° equilibrium curve is quite favorable for adsoφtion... The 90° curve has die same shape, which is not favorable for stripping... The design, developed by Advanced Separations Technologies, Inc., indicates a product stream of 9% citric acid from a feed of 16% citric acid in brotii."
The above statement is made for adsoφtion at 25°C and desoφtion at 90°C. The upper limit of the temperamre range is determined in die case of citric acid by die various partial vapor pressures, by die overall pressure in die system and by die tiiermal stability of the resin. One should keep in mind diat in the case of ascorbic acid, diere is an additional limitation. Ascorbic acid tends to oxidize to dehydroascorbic acid, which decomposes rapidly to odier
by-products. This oxidation could be enhanced by elevated temperamres and by die contact widi the resin.
Thus, as seen from the above discussion, die state of die art does not teach whedier binding to d e pyridine based resin and desoφtion at elevated temperamre is attainable witiiout degradation of die ascorbic acid, and in fact, none of d e above-mentioned publications teaches or suggests die process of the present invention.
As is well-known, a strong adsorbent is needed for high yield recovery from the feed solution. On die odier hand, desoφtion is hindered by strong adsorbents, resulting in dilute desorbate solutions. The state of die art does not teach whedier a pyridine based resin is strong enough to show high yields in adsorbing ascorbic acid from die dilute solutions and still weak enough to allow desoφtion at a concentration higher dian its concentration in die feed.
While d e invention will now be described in connection widi certain preferred embodiments in die following example so diat aspects diereof may be more fully understood and appreciated, it is not intended to limit die invention to diese paπicular embodiments. On me contrary, it is intended to cover all alternatives, modifications and equivalents as may be included witiiin die scope of die invention as defmed by d e appended claims. Thus, die following example which includes preferred embodiments will serve to illustrate die practice of diis invention, it being understood that die particulars shown are by way of example and for purposes of illustrative discussion of preferred embodiments of die present invention only and are presented in d e cause of providing what is believed to be die most useful and readily understood
description of formulation procedures as well as of the principles and conceptual aspects of die invention.
EXAMPLE 1
An aqueous solution comprising 7 g/1 of ascorbic acid was contacted at
25 °C countercurrendy with a series of columns, comprising Rillex™ 425. The flow rate was 7 aqueous solution volumes per volume of resin, and the number of contacts was 7. More man 90% of die acid was adsorbed on die resin.
The resin was then washed at 80°C countercurrendy wid water. Here again, 7 stages were used. Practically all die adsorbed ascorbic acid was recovered, at a concentration of 10 g/1.
It will be evident to those skilled in the art diat die invention is not limited to die details of die foregoing illustrative examples and d at the present invention may be embodied in other specific forms without departing from d e essential attributes thereof, and it is dierefore desired diat die present embodiments and examples be considered in all respects as illustrative and not restrictive, reference being made to d e appended claims, rather dian to die foregoing description, and all changes which come within die meaning and range of equivalency of die claims are dierefore intended to be embraced therein.