FI93857C - A method for separating citric acid from a fermentation broth using a polymeric adsorbent - Google Patents

Description

93857

A method for separating citric acid from a fermentation broth using a polymeric adsorbent

The invention relates to the adsorbent separation of citric acid in a solid layer from fermentation broths containing citric acid, carbohydrates, amino acids, proteins and salts. More particularly, the invention relates to a process for separating citric acid from fermentation broths containing it using a non-zeolite polymeric adsorbent that selectively adsorbs citric acid and is selected from the group consisting of a neutral, cross-linked polystyrene ionic polymer, a non-zeolite a ryl ester polymer, a weakly basic anionic ion exchange resin having tertiary amine or pyridine functional groups, and a strongly basic anion exchange resin having quaternary amine functional groups, and mixtures thereof.

Citric acid is used as a food acidifier and in pharmaceutical, industrial and detergent compositions. The growing popularity of liquid detergents formed using citric acid has mainly caused the global growth of citric acid production to about 320 million kilos per year, which is expected to continue in the future.

• ·

Citric acid is produced by an internal fermentation process using molasses and the microorganism strain Aspergillus niger. The fermentation product contains carbohydrates, amino acids, proteins and salts as well as citric acid, which must be separated from the fermentation broth.

»

Currently, two different methods are used to separate citric acid. The second involves the precipitation of citric acid as the calcium salt. The calcium citrate obtained is acidified with sulfuric acid. In another method, citric acid is extracted from the fermentation broth with a mixture of trilaurylamine, n-octanol and C10 or C '- paraffin. The citric acid is re-extracted from the solvent phase into water using 2 93857 heat. However, both of these methods are complex and expensive and generate significant amounts of waste.

The patent literature has suggested the use of a possible third method for separating citric acid from fermentation broth, which method comprises membrane filtration to remove raw materials or high molecular weight impurities and subsequent adsorption of the impurities as after concentration or precipitation of citric acid as a calcium salt and subsequent acidification with H 2 SO 4, separation of CaSO 4 and treatment with cationic or anionic ion exchange resins. This method, disclosed in EP Patent Application No. 151,470, August 14, 1985, is also a rather complicated and time-consuming method for the separation of citric acid. The process according to the invention, on the other hand, makes it possible to separate the citric acid in one adsorption step and to recover the citric acid from the adsorbent, whereby purified citric acid is obtained using an easily separable desorbent.

In the process of the invention, the essential features of which are set out in the appended claims, citric acid is separated from the fermentation broth into a polymeric adsorbent selected from the group consisting of neutral, crosslinked polystyrene polymer, nonionic hydrophobic polyacrylic ester , and a strongly basic anionic ion exchange resin having quaternary amine functional groups, and * ·. mixtures thereof, and then recovered by desorption with a suitable desorbent under suitable conditions. According to the invention, it has been found that a highly selective separation of citric acid from salts and carbohydrates is achieved by adjusting and maintaining the pH of the feed solution below the first ionization constant of citric acid (pKai = 3.13). The value to which the pH must be lowered to achieve sufficient selectivity appears to be independent of the concentration of citric acid in the feed mixture; The pH is inversely proportional to the concentration. At concentrations of less than 13% to very low values, the pH may be close to the pKaj of citric acid 3.13; at a concentration of 13%, the pH may be in the range 0.9 to 1.7, but with 40% citric acid fed, the pH must be reduced to at least 1.2 or below. At higher concentrations, the pH should be even lower, e.g., at a citric acid concentration of 50%, the pH should be 1.0 or below. Thus, it is preferred to keep the pH of the feed mixture in the range of 0.5-2.5, with 0.5-2.2 being most preferred. According to the invention, it has also been found that the adsorption temperature of the polymeric adsorbent used can be reduced by adding acetone or some other low molecular weight ketone to the desorbent, thus avoiding the decomposition of the adsorbent by high temperatures.

The process of the invention for separating citric acid from a feed mixture of citric acid-containing fermentation broth using a polymeric adsorbent as described above can be applied to a process comprising the steps of: (a) maintaining a net flow of liquid through said column of adsorbent; at least three zones in which separate operations take place and which are connected to each other in series, the end zones of the statue being connected to provide a continuous connection between these zones; (b) maintaining an adsorption zone in this column, wherein this zone is defined by an adsorbent located between the feed inlet flow at the upstream boundary of the zone and the raffinate outlet flow in this zone: ·; (c) maintaining the purge zone immediately upstream of said adsorption zone, said purge zone being defined by an adsorbent located between the extract effluent at the upstream boundary of the purge zone and the feed inlet stream at the downstream boundary of said zone; 4 93857 (d) maintaining a desorption zone upstream of said purification zone, wherein said desorption zone is defined by an adsorbent located between the inlet stream of the desorbent at the upstream boundary of this zone and the extract effluent at the downstream boundary of this zone; (e) passing the feed mixture under adsorption conditions to the adsorption zone to effect selective adsorption of citric acid by the adsorbent in the adsorption zone, and removing the raffinate inlet stream containing the non-adsorbed portions of the fermentation broth leaving the adsorption zone; (f) introducing a desorbent into the desorption zone under desorption conditions to displace citric acid from the adsorbent in the desorption zone; (g) removing the extract effluent containing citric acid and desorbent from the desorption zone; (h) directing at least a portion of the effluent extract stream to the separation devices and removing at least a portion of the desorbent under the separation conditions; and (i) periodically passing the zone boundaries formed by the feed inlet stream, the raffinate effluent stream, the desorbent effluent stream, and the extract effluent stream through the adsorbent column downstream of the adsorption zone to provide flow zones in the adsorption zone and the effluent.

The process can produce a raffinate product that contains an even smaller amount of desorbent. In addition, in the process, a buffer zone can be maintained immediately upstream of the desorption zone, thereby providing this buffer zone. defines an adsorbent located between the inlet stream of the desorbent at the upstream boundary of the buffer zone and the effluent effluent at the upstream boundary of this buffer zone.

5,93857

Details of embodiments of the invention regarding feed mixtures, adsorbents, desorbents and processing conditions are set forth below.

Figure 1 is the concentration of different citric acid species as a function of citric acid dissociation pH, showing the change in the equilibrium point of citric acid dissociation with varying concentrations of citric acid, citrate anions, and hydrogen ion.

Figure 2 shows the effect of pH on the amount of citric acid that the adsorbent is capable of adsorbing.

Figures 3A, 3B and 3C graphically show the pulse experiments of Example I using XAD-4 resin to separate citric acid from a feed containing 13% citric acid at pH 2.4, 1.7 and 0.9, respectively.

Figures 4A, 4B, 4C, 4D and 4E show the pulse experiments of Example II at pH 2.4, 1.7, 0.9, 2.8 and 1.4, respectively, using different adsorbent samples.

Figures 5A and 5B show the pulse experiments of Example III at pH 2.8 and 1.4 at 93 ° C, respectively.

Figures 6A, 6B and 6C show the pulse experiments of Example IV at pH 1.94, 1.13 and 0.5, respectively.

Figures 7A, 7B and 7C show the pulse experiments of Example V at pH 1.82, 0.5 and 0.3, respectively.

Figures 8A and 8B show the pulse experiments of Example VI at pH 1.5 and 1.0, respectively.

Figure 9 shows the pulse test according to Example VII and the ad sorption obtained at lower temperatures (93 ° C vs. 45 ° C) when 10% acetone is added to the desorption water.

Figure 10 shows a pulse experiment according to Example VIII using a weakly basic anion exchange resin having tertiary amine functionality in crosslinked acrylic; in a resin matrix for separating citric acid from a feed containing 40% citric acid at pH 1.6 by desorption with water; Figures 11A, HB and 11C show the pulse experiments of Example IX at pH 7.0, 3.5 and 2.4, respectively.

6 93857

Figures 12, 13A and 13B show pulse experiments according to Example X at pH 1.6 using various samples of a weak basic anion exchange resin having pyridine functionality in a crosslinked polystyrene resin matrix. Citric acid is desorbed with 0.05 N sulfuric acid or water.

Figure 14 shows a pulse test according to Example XIII at pH 1.6.

Figure 15 shows the pulse test of Example XIV at pH 2.2 using a different adsorbent sample of a less basic anion exchange resin having quaternary ammonium functionality in a crosslinked polystyrene resin matrix and desorbed with dilute sulfuric acid.

The following definitions of the various terms used in this specification are useful in explaining the treatments, objects, and advantages of the present invention.

A "feed mixture" is a mixture containing one or more extract components and one or more raffinate components which are separated by the process of the invention. The term "feed stream" means the feed mixture stream that is passed to the adsorbent used in the treatment.

An "extract component" is a compound or type of compound that is more selectively adsorbed by an adsorbent, while a "raffinate component" is a compound or type of compound that is less selectively adsorbed. In this method, the citric acid is an extract component and the carbohydrate salts are raffinate components. The term "desorbent" generally refers to a material capable of desorbing an extract component. The term "desorption stream" or "desorption input stream" means the flow through which the desorbent passes into the adsorbent. The term "raffinate stream" or "raffinate effluent" means the stream through which the raffinate component is removed from the adsorbent. The composition of the raffinate stream can range from substantially 100% desorbent to substantially 100% raffinate components. The term "extract stream" or "extract effluent stream" means a stream by which the extract material desorbed by the desorbent 93857 is removed from the adsorbent. The composition of the extract stream may also range from substantially 100% desorbent to substantially 100% extract components. At least a portion of the extract stream, and preferably at least a portion of the raffinate stream from the separation treatment, is passed to separation equipment, usually fractionation equipment, in which at least a portion of the desorbent is separated to produce an extract product and a raffinate product. The terms "extract product" and "raffinate product" refer to products which are prepared by a process comprising the extract component and the raffinate component at higher concentrations, respectively, than those present in the extract stream and the raffinate stream, although the process of the invention makes it possible to prepare very pure It should be noted that the adsorbent never completely adsorbs the extract component Similarly, the adsorbent does not completely adsorb the raffinate component. on the basis of the concentration ratios of the extract component and the raffinate component present in the stream in question, in particular the ratio of citric acid to with respect to the electrically adsorbed constituents is the lowest in the raffinate stream, followed by the highest in the feed mixture and the highest in the extract stream. Similarly, the ratio of the concentration of the least selectively adsorbed components to the more selectively adsorbed citric acid is highest in the raffinate stream, followed by the highest in the feed stream, and lowest in the extract stream.

The term "selective pore volume" of an adsorbent is the volume of adsorbent that selectively adsorbs the extract component from the feed mixture. The term "non-selective pore volume" of the adsorbent is that volume of the adsorbent which does not retain the extract component from the feed mixture. This volume comprises those cavities of the adsorbent which do not contain adsorbent bonds and the spaces between the particles of the adsorbent. Selective pore volume and non-selective pore volume are generally defined as 93857 units of volume and are important in determining appropriate flow rates for the liquid introduced into the treatment zone to provide efficient treatments using a given amount of adsorption. When the adsorbent "enters" the treatment zone (described in more detail below) used in accordance with one embodiment of the invention, its non-selective pore volume, together with the selective pore volume, transports the liquid to this zone. The non-selective pore volume is used to determine the amount of liquid that flows into the zone upstream of the adsorbent, displacing the liquid in the non-selective pore volume. If the flow rate of the liquid entering the zone is less than the non-selective pore volume of the adsorbent entering this zone, a net flow of liquid into this zone occurs. Because this net flow is its liquid in the non-selective pore volume of the adsorbent, it contains in most cases less selectively attached feed stream components. The selective pore volume of the adsorbent may in certain cases adsorb portions of the raffinate from the liquid surrounding the adsorbent, because in certain cases there is competition between the extract material and the raffinate material at the adsorbent sites in the selective pore volume. If a large amount of raffinate material surrounds the adsorbent relative to the extract material, the raffinate material may be competitive enough to adsorb the adsorbent.

The feed material according to the invention is a fermentation product obtained by means of the internal fermentation of molasses liquid and the microorganism Aspergillus Niger. This fermentation product may have, for example, the following composition: citric acid 12.9% + 3% salts 6000 ppm carbohydrates (sugars) 1% others (proteins and amino acids) 5%

The salts are K, Na, Ca, Mg and Fe. Carbohydrates include sugars such as glucose, xylose, mannose and DP2 and DP3 oligosaccharides plus up to 12 or more unidentified saccharides. The composition of the substance to be treated may vary from the above without still falling outside the scope of the invention. However, juices such as citrus juices are not suitable or appropriate because the other substances they contain are adsorbed simultaneously with citric acid. See. Johnson, J .:

Sci. Food Agric, Vol. 33 (3) s, 287-293.

It has now been found that the separation of citric acid can be greatly improved by adjusting the pH of the feed to a level below the first ionization constant of citric acid. The first ionization constant (pKa 2) of citric acid is 3.13; Handbook of Chemistry & Physics, 53rd Edition, 1972-1973, CRC Press, and consequently the pH of the citric acid feed stream should be less than 3.13. If the pH of a 13% solution of citric acid is e.g. 2.4 or higher, as shown in Figure 3A (Example I), "citric acid breaks through" (desorbs) along with salts and carbohydrates at the beginning of the cycle, indicating that citric acid is not adsorbed. On the other hand, a lower "breakthrough" of citric acid is observed at a pH of 1.7, and no "breakthrough" is observed at a pH of 0.9 at a concentration of 13%, as shown, for example, in Figures 3B and 3C.

The aqueous solution contains combined citric acid in equilibrium with several citrate ions and hydrogen ions. This is evident from the following equations, where the acid dissociation constants pKa 2, pKa 2 and pKa 2 of citric acid are 3.13, 4.74 and 5.40, respectively, at 25 ° C:

Equation 1 = 3.13 pKa2 = 4.74 pKa3 = 5.40 H ^ CA, H2CA-1 ______ ~ HCA-2 —_______— 1 ~ CA-3 + H + + H + + H +

The equilibrium point for the dissociation of citric acid can be changed by varying the citric acid, citrate anion or hydrogen ions. concentration. This is shown in Figure 1 when comparing the concentrations of several species of citric acid in solution to pH at 90 ° C. The results show that higher levels of 10 93857 percent of non-ionized citric acid (H 2 CA) are obtained at higher hydrogen ion concentrations (lower pH). Lowering the pH (increasing the H + ion concentration) results in more non-ionized citric acids while decreasing the amount of -1 -2 -3 citrate anion species (^ CA, HCA and CA) in solution.

Based on the balance of citric acid and the above-mentioned properties of the resin, non-ionized citric acid is separated from other ionic species in the fermentation broth (including citrate ions) using the adsorbent resins described above. However, solutions with lower pH are required to achieve better recovery of citric acid. To determine the static adsorption iso-term of citric acid of a resin of the invention, Amberlite XAD-4, treatment was performed at room temperature, about 25 ° C, as a function of feed pH. Figure 2 shows the results of this study. The results show the adsorption of non-ionized citric acid as the pH is lowered. Without being limited to this definition, it is apparent that the non-ionized citric acid species in solution are adsorbed on the adsorbents of the invention either by an acid-base mechanism or a hydrogen binding mechanism or a mechanism based on a strong relative affinity of these hydrophobic species.

The desorbents used in the various prior art adsorbent separation methods vary depending on such factors as the nature of the treatment used. In a pendulum bed system in which a selectively adsorbed feed component is removed from the adsorbent by a cleaning stream, the choice of desorbent is less critical, and desorbents comprising gaseous hydrocarbons such as methane, ethane, etc., or other types of gases such as nitrogen and hydrogen can be used or under reduced pressure, or both, to effectively remove the adsorbed feed component from the adsorbent. However, in such adsorbent separation treatments, which are generally applied continuously, mainly at constant pressures and temperatures, to obtain a liquid phase, the desorbent must be chosen correctly to ensure a number of factors. First, the desorbent should displace the extract component from the adsorbent at reasonable flow rates without significantly adsorbing itself, which would adversely prevent displacement of the extract component from the desorbent in the subsequent adsorption step. Expressed as selectivity (discussed in more detail below), it is preferred that the adsorbent be more selective for all other extract components relative to the raffinate component than for its selectivity over the desorbent material over the raffinate component. Second, the desorbents must be compatible with the particular adsorbent and feed mixture. More specifically, they must not reduce or eliminate the critical selectivity of the adsorbent over the extract component compared to the raffinate component. In addition, the desorbents must be such that they can be easily separated from the feed mixture to be treated. Both the raffinate stream and the extract stream are removed from the adsorbent as a mixture with the desorbent, and without a method to separate at least a portion of the desorbent, the purity of the extract product and the raffinate product is not very high and the desorbent agent cannot be reused in the process. As a result, it has been found that each desorbent used in this process should have a significantly different average boiling point compared to the feed mixture so that at least a portion of the desorbent can be separated from the feed components in the extract and raffinate streams by simple fractional distillation. As used herein, the term "substantially different" means that the average boiling point between the different desorbent and the feed mixture is at least about 5 ° C. The boiling range of the desorbent may be higher or lower than the boiling range of the feed mixture. Finally, desorbents should be readily available and therefore reasonably cost effective. In the preferred isothermal, isobaric liquid phase treatment according to the invention, water has been found to be a very effective desorbent. It has also been found that acetone and other low molecular weight ketones, such as methyl ethyl ketone and diethyl glycone, are effective in admixture with water in small amounts up to 15% by weight. Their usefulness is based on their water solubility. However, their advantages are based on their ability to reduce the temperature at which desorption occurs. When using some adsorbents and water as the desorbent, the temperature must be raised to enhance the desorption step. High temperatures can cause premature deactivation of the adsorbent. The solution to this problem during such separation is to add 1 to 15% by weight, preferably 1 to 10% by weight and most preferably 5 to 10% by weight of acetone to the desorbent. A low molecular weight ketone can also affect the stability of the adsorbent in potentially two ways, namely by removing soluble constituents that cause deactivation and effecting regeneration, i.e., removing deactivating agents or altering their effect. A reduction in the desorption temperature of such a separation of about 50 ° C is achieved, for example, by adding 10% by weight of acetone to the desorbent. A reduction of about 5-70 ° C can be achieved by adding 1-15% by weight of acetone to the water used as desorbent. Dilute inorganic acids have also been found to give good results when used as desorbents. Aqueous solutions of sulfuric acid, nitric acid, hydrochloric acid and phosphoric acid and mixtures thereof can be used in amounts corresponding to concentrations of 0.01 to 1.0 n (normality), with the best results being obtained with dilute sulfuric acid at a concentration of 0.01 to 1.0 n.

It is known that certain properties of adsorbents are highly desirable, if not absolutely necessary, for the successful implementation of selective adsorption treatment. These properties are also important in this method. Such properties are, for example, the following: 1) adsorption capacity relative to a certain volume of extract component per volume of adsorbent, 2) selective adsorption of extract component with respect to raffinate component and desorbent, and 3) sufficiently rapid adsorption and desorption of extract component from and adsorbent. The ability of the adsorbent to adsorb a given volume of extract component is, of course, essential. Without this ability, the adsorbent is unusable in adsorbent separation. In addition, the greater the ability of the adsorbent to extract the extract component, the better the adsorbent. The higher efficiency of an adsorbent makes it possible to reduce the amount of adsorbent required to separate the extract component of known concentration when using a given feed mixture addition rate. The reduction in the amount of adsorbent required in any adsorption treatment reduces the cost of the separation treatment. It is important that the good initial efficiency of the adsorbent is maintained economically long enough during the separation treatment. Another necessary ability of the adsorbent is to separate the components of the feed stream, i.e. that the adsorbent has a certain adsorption selectivity (B) over one component over another. Relative selectivity can not only be shown for one component over another, but can also be shown between any component of the feed mixture and the adsorbent. Selectivity (B), a term used in this specification, is the ratio of the two components of the adsorbed phase to the ratio of the same two components in the non-adsorbed phase at equilibrium conditions. Relative selectivity is shown in Equation 2 below:

Equation 2

Selectivity = (B) = (volume percentage C / volume percentage P) A

(volume percentage C / volume percentage D) U

wherein in Equation C and D are the two components of the feed in% by volume, and A and U are the adsorbed and non-adsorbed phases, respectively. Equilibrium conditions prevail when the feed passing through the adsorption layer does not change its composition upon contact with the adsorption layer. In other words, there is no net transfer of the substance between the non-adsorbed and the adsorbed phase. As the selectivity of the two components approaches 1.0, the adsorbent does not adsorb one component to the other, but both are adsorbed (or not adsorbed) in approximately the same manner relative to each other.

14 93857

When the value (B) becomes less than or greater than 1.0, the adsorbent mainly adsorbs one component relative to the other. When comparing the selectivity of the adsorbent with respect to the second component C over the second component D, a value (B) greater than 1.0 indicates that the adsorbent adsorbs predominantly component C. When (B) is less than 1.0, this indicates that component D adsorbs predominantly and more adsorb remains in the non-adsorbed phase and component D in the adsorbed phase. The ideal desorbents are such that their selectivity is about 1 or slightly less than 1 for the extractant components so that all extractant components can be desorbed as one group at reasonable flow rates by the desorbent, that the extract components can replace the desorbent in the next adsorption step. Although separation of the extract component from the raffinate component is theoretically possible when the selectivity of the adsorbent with respect to the extract component relative to the raffinate component is greater than 1, it is preferred that this selectivity approaches 2. As well as relative volatility, the higher the selectivity . Greater selectivity allows a smaller amount of adsorbent to be used. A third important feature is the exchange rate of the extract component of the feed mixture, i. the relative desorption rate of the extract component. This property is directly dependent on the amount of desorbent that must be used in the treatment to recover the extract component from the adsorbent. Higher exchange rates reduce the amount of desorbent required to remove the extract component, thus allowing lower processing costs. At higher exchange rates, less desorbent must be pumped through the treatment and separated from the extract stream for reuse in the treatment.

Decomposition is a measure of the degree of separation of a two-component system and can be used to determine the efficiency of a quantified combination of adsorbent, desorbent, and treatment conditions during separation treatment. In this specification, scattering means the distance between two centers of peaks divided by the average width of these peaks at 9/2-height of the peaks as determined by the pulse test described below. The equation for determining dispersion is therefore as follows:

Equation 3 *. L * ~ L1_ 1/2 (Wx + W 2) wherein in formula and L 2 are respectively the distance from the reference point in ml, sa with respect to the peak center O, and and W2 are the peak widths at half-height of the case.

In the experiment of different adsorbents, dynamic test equipment, certain feed mixtures and desorbents were used to determine the adsorption capacity, selectivity and exchange rate of the adsorbent. The apparatus consisted of an adsorption chamber consisting of a helical statue of about 70 ml with inlet and outlet points at opposite ends of the chamber. The chamber was in contact with the temperature control devices and also with the pressure control devices used to make the chamber operate at a predetermined constant pressure. Quantitative and qualitative analytical equipment, such as a refractometer, a Polarimeter, and chromatographic equipment, can be connected to the chamber outlet line and used to quantitatively or qualitatively determine one or more components of the flow out of the chamber. A pulse test performed using this equipment and the following method was used to determine selective and other values in different adsorption systems. The adsorbent was filled to equilibrium with the desorbent used by passing it through the adsorption chamber. At the appropriate time, a feed pulse containing a tracer of known extract and a specific extract component or raffinate component, or both, diluted with a desorbent, was injected for several minutes. The desorbent flow is then restarted and the tracer and extract component or raffinate component (or both) were eluted using liquid-solid chromatography. The effluent can be analyzed continuously or the effluent samples can be collected periodically and subsequently analyzed separately with analytical equipment and the peak curves of the corresponding component can be developed.

Based on the results obtained from the adsorption experiments, its efficiency can be expressed as the pore volume, the retention volume of the extract or raffinate component, the selectivity of one component over another, and the desorption rate of the desorbent over the extract component. The retention volume of the extract or raffinate component can be expressed as the distance between the center of the peak curve of the extract or raffinate component and the peak curve of the tracer or some other known reference point. It is expressed as the volume of desorbent in cubic centimeters pumped during this period, represented by the distance between the peak curves. The selectivity (B) of the extract component relative to the raffinate component can be expressed as the ratio of the distances between the centers of the peak curve of the extract component and the peak curve of the tracer (or some other reference) to the corresponding distance between the peak curve of the raffinate component and the peak of the tracer component. Exchange of an extract component with the desorbent speed can generally be defined as the peak of the curve half width. The narrower the width of the peak, the higher the desorption rate. The desorption rate can also be determined from the distance between the center of the tracer peak curve and the disappearance of the freshly desorbed extract component. This distance is also the volume of desorbent pumped during this time.

Further evaluation of the preferred adsorption systems and the conversion of these values into a practical separation method requires the best system to be tested in a practically continuous, countercurrent liquid-solid contact system. The general principles of operation of such a device have been described previously in U.S. Patent 2,985,589. A laboratory-scale apparatus that implements these principles is described in U.S. Patent 3,706,812. This device has several: adsorption layers and a plurality of supply lines attached to layered distribution devices , and terminate in a rotating it 93857 distribution valve. At a given valve position, the feed and desorbent are passed through two lines and the raffinate and extract streams are removed through the other two lines. All other supply lines are inactive, and when the distribution valve position is shifted by one cycle, all active functions are shifted by one layer. This simulates a state in which the adsorbent moves countercurrent to the liquid flow. Further details of the above-mentioned nonionic adsorbent test apparatus and adsorbent assay technique are set forth in "Separation of Cg Aromatics by Adsorption" by A.J. deRosset, R.W. Neuzil, D.J. Korous and D.H. Rosback, presented at the Society's American Chemical Society, Los Angeles, California meeting on March 28th. - 2.4. 1971.

One group of adsorbents that can be used in the process of the invention are nonionic, hydrophobic, water-insoluble styrene-poly (vinyl) benzene copolymers and copolymers thereof with monoethylenically unsaturated compounds or polyethylenically unsaturated poly (ethylenically unsaturated) monomers. which include acrylic esters, e.g., those described in U.S. Patent Nos. 3,531,463 and 3,663,467, which are incorporated herein by reference. As disclosed in U.S. Patent No. 3,531,463, these polymers can be prepared as disclosed in U.S. Patent Application No. 749,526, which has resulted in U.S. Patent Nos. 4,221,871, 4,224,415, 4,256. 840, 4,297,220, 4,382,124 and 4,501,826, all of which are incorporated herein by reference. The above adsorbents are manufactured by Rohm and Haas Company and sold under the trademark "Amberlite". Amberlite polymers of the type known to be effective in the practice of the invention are mentioned in the brochures of the above-mentioned company as Amberlite adsorbents XAD-1, XAD-2, XAD-4, XAD-7 and XAD-8 and are described in these brochures as "hard, high into a porous polymer formed by insoluble spheres having a surface area ". Different types of adsorbent Amberlite polymers. differ somewhat in physical properties, such as porosity, volume percentage, density and screen size, ie 93857, but even more in their surface area, average pore diameter and 2-pole torque. The most preferred adsorbents have a surface area of 10-2000 m / g and most preferably 100-1000 m / g. These properties are shown in the following table.

table 1

Properties of Amberlite Polymeric Adsorbents XAD-1 XAD-2 XAD-4 XAD-7 XAD-8

Chemical poly-poly-poly-acrylic-acrylic nature styrene styrene styrene ester ester

Porosity% by weight 37 42 51 55 52

Actual wet density g / cm 1.02 1.02 1.02 1.05 1.09

Area m2 / g 100 300 780 450 160

Avg. diameter,

Angström 200 90 50 90 225

Structure density g / cm3 1.07 1.07 1.08 1.24 1.23

Nominal part size 20-50 20-50 20-50 20-50 25-50

2-pole torque of functional groups 0.3 0.3 0.3 1.8 1.8

The uses of the adsorbent Amberlite polymers disclosed in the Rohm and Haas Comapny brochures include decolorization of the pulp mill bleach waste solution, decolorization of paint waste, and removal of insecticides from the waste solutions. This, of course, does not suggest a surprising finding according to the invention that the adsorbent Amberlite polymers are effective in separating citric acid from Aspergillus Niger fermentation broth.

93857 19

Another group of adsorbents that can be used in the process of the invention are weak basic anion exchange resins having tertiary amine or pyridine functionality in a crosslinked polymer matrix, e.g. acrylic or styrene. They are very suitable in the form of spheres, have a very uniform polymer porosity, good chemical and physical stability and good abrasion resistance (which is not common in macroreticular resins).

Adsorbents, such as those described above, are manufactured by Rohm and Haas Company and sold under the trademark "Amberlite". Such Amberlite polymers are known to be effective for the purposes of the invention and are referred to in the Rohm and Haas Company brochures as Amberlite adsorbents XE-275 (IRA-35), IRA-68, and are described in these brochures as "insoluble in conventional solvents, and having an open structure for efficient adsorption and de-sorption of large molecules without losing their efficiency due to fouling by organic matter ". Also suitable are resins AG-3X4A and AG4-X4, manufactured by Bio Rad, and corresponding resins sold by Dow Chemical, such as Dowex 66 and Dow test resins, which can be prepared by the methods of U.S. Patents 4,031,038 and 4,098,867.

The various polymeric adsorbents in this class differ to some extent in physical properties such as porosity, volume percent, structural strength, and nominal particle size, and even more in terms of surface area, average pore diameter, and 2-pole moment. The most preferred adsorbents have a surface area of 10-2000 m 2 / g and preferably 100-1000 m 2 / g. The special features of the above products are described in the literature and technical brochures published by the manufacturers, such as the products listed in Table 2, which are referred to in this context. Other types of products can also be used.

20 93857

Table 2

Adsorbent_ Matrix species Literature references AG3-4A Polystyrene Chromatography Electrophoresis (Bio Rad) Immunochemistry Molecular

Biology - HPLC - Price List M April 1987 (Bio Rad) AG4-X4 Acrylic Chromatography Electrophoresis

Immunochemistry Molecular Biology - HPLC - Price List M April 1987 (Bio Rad)

Dow Polystyrene U.S. Patent Nos. 4,031,038 and Experimental 4,098,867 Resins

Dowex 66 Polystyrene Material Safety Data Sheet printed 17.2. 1987 (Dow Chemical USA) IRA-35 Acrylic Amberlite Ion Exchange Resins (XE-275) (XE-275) Rohm & Haas Co. 1975 IRA-68 Acrylic Amberlite Ion Exchange Resins-

Amberlite IRA-68

Rohm & Haas Co. April 1977

Applications for polymeric Amberlite adsorbents, as described in the literature published by Rohm and Haas Comany, include decolorization of the pulp mill bleach waste solution. decolorization of paint waste and removal of insecticides from waste solutions. The present invention has not been disclosed in the literature for the efficient separation of chiric acid from polymeric Amberlite adsorbents from Aspergillus Niger fermentation broth.

A third group of adsorbents that can be used in the process of the invention are strongly basic anion exchange resins having quaternary ammonium functionality in a crosslinked polymer matrix, i. di-vinylbenzene crosslinked acrylic or styrene resins.

They are very suitable in the form of spheres and have a uniform • polymer porosity and good chemical and physical stability. In the present case, these resins may be gel-like (or "gel-type") or "macroreticular", a term used in the current literature, namely Kun and Hetherington: "A Progress Report on the Removal of Colloids From Water by Macroreticular Ion Exchange Resins ", an article presented at the International Water Conference (Pittsburgh, PA. October 1969), reproduced by Rohm and Haas Comapny. In current adsorption technology, "the term microreticular refers to a gel structure per se, pore size that is atomic in size and depends on the swelling properties of the gel," while "macroreticular pores and true porosity means structures with pores larger than pores. Their size and shape are not significantly affected by changes in the environment, for example those due to variations in osmotic pressure ", while the dimensions of the gel structure" depend significantly on environmental conditions ". In "classical adsorption", the terms microporous and macroporous normally refer to pores longer than 20 A and larger and 200 A. Pores with a diameter between 20 A and 200 A are transfer pores. "The authors use the term" macroreticular "in its sense. instead of defining the novel ion exchange resins of the invention "having both a microreticular and a macroreticular pore structure. The former means the distance between the chains and the cross-linking points of the swollen gel structure, and the latter pores which do not form part of the actual chemical structure. The macroreticular part of the structure may in fact comprise micro-, macro- and transitional pores depending on the pore size distribution. "(The above quotations are from page 1 of Kun et al.) Adsorbents with a macroreticular structure also have good abrasion resistance (which is not common with conventional macroreticles). References in this specification to "macroreticular" thus refer to adsorbents of the type described which have two porosities as defined by Kun and Hethesing, and the terms "gel" and "gel-type" are used herein in their conventional sense.

22 93857

When considering ion exchange resins containing a quaternary ammonium function according to the invention, the quaternary amine has a positive charge and can form an ionic bond with a sulfate ion. The ion exchange resin containing a quaternary ammonium anion has a weak basic property in the sulfate form, which in turn can adsorb citric acid due to the acid-base reaction.

P - N + (R) 3 or P - N + - (R) 2 (C2H4OH) O- O- .. I ·· ·· I ..

: 0 = S = O:: 0 = S = O: I - + - I _ + - O - H - C.A. O - H -C.A.

wherein P = resin group R = lower alkyl, C1-3 C.A. = citrate ion

Adsorbents, e.g., those described above, are manufactured by Rohm and Haas Company and sold under the trademark "Amberlite". Amberlite polymers known to be effective for the purposes of the invention are disclosed in Rohm and Haas Company's brochures in Amberlite IRA 400 and 900 series and are described in these brochures as "insoluble in all conventional solvents, open in structure to provide efficient adsorption and desorption to large molecules. * that they lose their effectiveness due to contamination by organic matter ". Also suitable are resins AG1, AG2 and AGMP-1 manufactured under the tradename Bio Rad, and corresponding resins manufactured by Dow Chemical Co., such as Dowex 1, 2, 11, MSA-1 and MSA-2, etc. Useful in the practice of the invention are: also the so-called. moderately basic ion exchange resins which are mixtures of strong and weak base resins. Examples include the following commercially available resins: Bio-Rex 5 (Bio-Rad 1); Amberlite IRA-47 and Duolite A-340 (both manufactured by Rohm & Haas). They are useful, for example, if necessary, a basic ion exchange resin which is not as basic as strongly basic resins, or one which is more basic than weakly basic resins.

23 93857 The different types of polymeric adsorbents in this group differ to some extent in their physical properties such as porosity, volume percentage, structural strength and nominal particle size, and even more in terms of surface area, average pore diameter and 2-pole moment. More preferred adsorbents have a surface area of 10-2000 m / g and preferably 100-1000 m / g. The special properties of these materials are described in the literature and technical brochure published by the manufacturer, e.g. those listed in Table 3 and referred to herein.

Table 3

Properties of Adsorbents Adsorption- Matrix Substance_ Resin Type Literature References IRA 458 Acrylic Gee- Amberlite Ion Exchange Resins (Rohm & Haas) Lite Type 1986 & Technical Bulletin IE-207-74 84 IRA 958 Acrylic Mak- Technical Bulletin and Material Grassy Safety Data Sheet Available IRA 900 Technical Bulletin available in Macroporous 3a Amberlite Ion Exchange

Resins, IE-100-66.

IRA 904 Polystyrene Technical Bulletin, 1979 and Macroporous IE-208/74, Jan. 1974 • IRA 910 Polystyrene Technical Bulletin, 1979 and Macroporous IE-101-66, May. 1972 IRA 400, 402 Polystyrene Amberlite Ion Exchange Resins, Macroporous Oct., Sep. 1976, April 1972 and IE-69-62, Oct. 1976 IRA 410 Polystyrene Amberlite Ion Exchange Resins Gel Type IE-72-63, El. 1970 AG 1 Polystyrene Chromatography Electrophoresis (Bio Rad) Gel Type Immunochemistry Molecular

Biology HPLC, price list M aprik. 1987 AG 2 Polystyrene Chromatography Electrophoresis

gel type Immunochemistry Molecular: Biology HPLC, price list M

April 1987 24 93857

Table 3 (continued)

Properties of Adsorption Agents Adsorption Matrix Substance_ Resin Type References AG-MP-1 Polystyrene Chromatography Electrophoresis Macroporous Immunochemistry Molecular

Biology HPLC, price list M April 1987

Bio Rex 5 Strong Chromatography Electrophoresis (Bio Rad) and Weakly Immunochemistry Molecular

basic hart- Biology HPLC, price list M

mixture (eg April 1987 AG-2 and AG-3 or AG-4

The adsorbent can be used in the form of a dense solid layer which is contacted alternately with the feed mixture and the desorbent. In the simplest embodiment of the invention, the adsorbent is used in the form of a single static layer, in which case the treatment is only semi-continuous. In another embodiment, two or more stationary layers may be used in the fixed bed together with suitable valves so that the feed mixture is passed through one or more layers of adsorbent, while the desorbent may be passed through one or more other layers of series. The flow of the feed mixture and the desorbent can take place either upwards or downwards through the desorbent. Any conventional apparatus for contacting a liquid and a solid in a static layer can be used.

However, in a countercurrent system, the countercurrent layer or the simulated countercurrent layer are such that they have a much higher separation efficiency than fixed adsorption layer systems, and are therefore most preferred. In processes using a moving bed or a simulated moving bed, the adsorption and desorption treatments take place continuously, which allows the continuous preparation of the extract and raffinate stream and the continuous use of the feed and desorption streams. A preferred embodiment of this method is one that uses a simulated, backflow system based on a moving 25 9 3 8 5 7 bed. The processing principles and order of processing of such a flow system are described in U.S. Patent 2,985,589. Only four inlet lines operate at each time point, namely feed inlet flow, desorbent inlet flow, raffinate outlet flow, and extract outlet flow. Simultaneously with this simulated, upward flow of solid adsorbent, the flow of liquid that fills the pores of the dense layer of adsorbent occurs. Thus, a contact in the opposite direction is maintained, whereby a downward flow of liquid in the adsorption chamber can be provided by means of a pump. As the active fluid inlet moves through the cycle, i. from the top of the chamber to its bottom, circulates the chamber circulation pump through different zones of fluid that require different flow rates. A programmed flow control device may be used to provide and control such flow rates.

The active liquid supply points divide the adsorption chamber into separate zones, each with a different function. In this embodiment of the method according to the invention, it is generally necessary to use three separate treatment zones to carry out the treatment, although in a few cases a fourth zone can also be used.

The adsorption zone, zone 1, is defined by the adsorbent located between the feed inlet stream and the raffinate outlet stream. In this zone, the feed is contacted with the adsorbent, the extract component is adsorbed and the raffinate stream is removed. Since the general flow through zone 1 occurs from the feed stream entering this zone to the raffinate flow leaving the zone, the flow is considered to occur downstream in this zone as it moves from the feed inlet to the raffinate outlet.

„93857 26 Immediately upstream of the liquid flow to zone 1 there is a purification zone, zone 2. The purification zone is defined by the adsorbent between the extract discharge stream and the feed inlet stream. The basic treatments that take place in zone 2 are the displacement of the raffinate material from the non-selective pore volume of the adsorbent that has migrated to zone 2 by changing the adsorbent to this zone and the desorption of the raffinate material to the adsorbent that has adsorbed the adsorbent. Purification is accomplished by directing a portion of the extract stream exiting zone 3 to zone 2 at the countercurrent boundary of that zone, i.e., the extract effluent so as to provide displacement of the raffinate material. The material flow to zone 2 takes place downstream from the extract discharge stream to the feed inlet stream.

Immediately upstream of zone 2 and of the liquid flowing to zone 2 is the desorption zone, i.e. zone 3. This desorption zone is defined by the adsorbent between the desorbent inlet stream and the extract outlet stream. The purpose of the desorption zone is to allow the desorbent that enters this zone to displace the extract component that the adsorbent has adsorbed due to previous contact with the feed in zone 1 during the initial treatment period. The flow of liquid to zone 3 takes place in substantially the same direction as to zones 1 and 2.

In a few cases, a buffer zone, zone 4, may also be used. a portion of the raffinate stream removed from zone 1 may be directed to zone 4 to displace the desorbent in that zone therefrom 2 'to the 93857 desorption zone. Zone 4 should contain sufficient adsorbent so that the raffinate material in the raffinate stream passing from zone 1 to zone 4 can be prevented from entering zone 3 and contaminating the extract stream from zone 3. In cases where the fourth treatment zone is not used, the raffinate flow from zone 1 to zone 4 must be carefully adjusted so that the flow directly from zone 1 to zone 3 can be interrupted when a sufficient amount of raffinate material is not present in the zone.

Circulation of the feed and discharge streams through the solid bed of adsorbent can be accomplished by using piping in which the valves are operated intermittently to vary the supply and discharge streams and to allow fluid flow in countercurrent to the solid adsorbent. Another treatment that can provide a flow of solid adsorbent countercurrent to the liquid involves the use of a rotary poppet valve, the supply and discharge flows being connected to this valve, and the lines through which the feed is introduced, the extract is removed, the desorbent is introduced and the raffinate is removed. transferred in the same direction through the adsorbent layer. Both said pipe arrangement and the poppet valve are known per se.

Various rotary poppet valves which can be used in this treatment are disclosed in U.S. Patents 3,040,777 and 3,422,848. Both of the above patents describe a rotary valve in which a suitable provision of different supply and discharge flows from fixed sources can be achieved without difficulty.

In many cases, the treatment zone contains a much higher amount of adsorbent compared to another treatment zone. For example, in a few treatments, the buffer zone may contain a smaller amount of adsorbent; compared to the amount of material required in the adsorption and purification zones. It can also be seen that in cases where a desorbent is used which readily desorbs the extract from the adsorbent, a relatively small amount of adsorbent is required in the desorption zone compared to the amount of adsorbent required in the buffer zone, adsorption zone and purification zone. Since it is not necessary for the adsorbent to be present in only one column, the invention also includes the use of several chambers or sets of statues.

It is not necessary that all feed or discharge streams be used simultaneously, and in many cases one of the streams may be shut off while the other feeds and discharges the feed. The apparatus which can be used to carry out the method according to the invention may also comprise a series of separate layers connected to interconnectors and fitted with supply and discharge connections to which different supply and discharge streams can be connected and which can be alternately and periodically varied to provide continuous processing. In a few cases, connecting lines can be connected to transfer points that, during normal processing, do not act as lines through which material is led to and from the processing.

It has been found that at least a portion of the extract effluent passes to separation devices, whereby at least a portion of the desorbent can be separated to produce an extract product containing the desorbent at a reduced concentration. Preferably, but this is not necessary to carry out the process, part of the raffinate effluent is also passed to separation devices in which at least part of the desorbent can be separated to form a desorbent stream which can be reused and to form a raffinate product containing a desorbent at reduced concentration. The separation device usually consists of a fractionation column, the structure and operation of which are well known in the art.

Reference may be made in this connection to U.S. Patent 2,985,589 and to the article "Continuous Adsorptive Processing — A New Separation Technique" (D.B. Broughton), issued to the Society of 29 93857

At the 34th Annual Meeting of Chemical Engineers in Tokyo, Japan on April 2, 1969, referred to herein to clarify the flow diagram of a countercurrent simulated moving bed method.

Although many such adsorbent separation methods can use treatments in both the liquid and vapor phases, the liquid phase treatment is more advantageous because it requires lower temperatures and because an extract product can be obtained in higher yield compared to the treatment in the vapor phase. Under adsorption conditions, the temperature range is about 20-200 ° C, with the range of about 65-100 ° C being most preferred, and the pressure range is from about atmospheric pressure to about 3450 kPa, which values are preferred to ensure a liquid phase. The same temperature and pressure range is used under the desorption conditions as under the adsorption conditions.

The size of the treatment units that can be used in the practice of the invention can range from experimental scale (e.g., disclosed in U.S. Patent 3,706,812) to factory-scale devices, and flow rates can range from a few milliliters per hour to several thousand kilos per hour.

The following examples illustrate the selectivity that allows the method of the invention to be implemented. However, these examples do not limit the invention.

Example I

In this example, three pulse experiments were performed using a neutral siren-divinylbenzene polymer as an adsorbent (XAD-4, manufactured by Rohm & Haas Company) to determine the citric acid resolution of this adsorbent at different pH values from its fermentation mixture containing carbohydrates (DP3, DP xylose, arabinose and raffinose), and salt ions, including Na +, K +, Mg ++, Ca ++, Fe +++, Cl ", SO 4", PO 4 "and NO 2 -, amino acids and proteins. The first experiment was performed at pH 2.4 and 45 ° C.

: The other two experiments were performed at pH 1.7 and 0.9. Citric diacid was desorbed with water. The composition of the feed mixture obtained from fermentation was as follows: 30 93857

Mixture fed Quantity

Citric acid 12.9%

Salts (K +, Na +, Ca ++, Mg ++, FE +++) 0.60% (6000 ppm)

Carbohydrates (sugars) 1%

Other (S04 =, Cl “, P04“, NO3 ~, 5% proteins and amino acids)

Water 81.5%

Retention rates and redissolution values were obtained using a pulse pig apparatus and the treatment method described above. The Ad sorbent was tested in a vertical 70 ml column using the following set of treatments in pulse experiments. The desorbent was continuously passed up through the column containing the adsorbent at a nominal liquid flow rate (LHSV) of about 1.0 per hour. The pore volume was determined from the volume of desorbent required to fill the dry filled statue. The flow of desorbent was stopped at the appropriate time and a 10 ml sample of the feed mixture was injected into the column using a sample ring and the flow of desorbent was restarted. Samples of the effluent were automatically collected in a sample collector and their hydrochloric and citric acid content was then analyzed chromatographically. A few subsequent samples were also analyzed for carbohydrates, but because they were eluted at approximately the same rate as the carbohydrates, they were not analyzed in these examples, nor were some of the other minor constituents, namely amino acids and proteins. From the analyzes of these samples, the concentrations of the peak curves of the components of the feed mixture were developed. The retention volume of citric acid was calculated by measuring the distance from the center of the salt retention volume curve to the center of the citric acid curve. The scattering value R is calculated from the previously presented Equation 3. The separation factor B was calculated as previously described.

The results of these pulse experiments are shown in Table 4 below.

* 31 93857

Table 4

Test A - pH - 2.4

Retention Peak width net 0.5 height- Scattering

Ingredient volume della_ (0.5 in height)

Salts 0 14.4 1.39

Citric acid 44.4 49.5 comparison

Test B - pH - 1.7

Retention Peak width net 0.5 height- Scattering

Ingredient volume della_ (0.5 in height)

Salts 0 11.6 1.49

Citric acid 42.2 45.1 comparison

Experiment C - pH - 0.9

Retention Peak width net 0.5 height- Scattering

Ingredient volume della_ (0.5 in height)

Salts 0 13.3 1.4

Citric acid 40.9 45.1 comparison

The results are shown in Figure 3A, which shows that although citric acid is adsorbed more strongly compared to the other ingredients, there is a significant loss of citric acid which remains unadsorbed and is eliminated with salts and carbohydrates (not shown). Citric acid differs satisfactorily in the process of Figure 3B, where the results are good, and in the process of Figure 3C, where the results are excellent. The method is obviously preferred at pH 1.7 and below. However, at pH 2.4 (Figure 3A), it has been found that a substantial amount of citric acid is recovered in the form of citrate H2CA-4 in the raffinate along with salts and carbohydrates.

It is clear from this that ionized solutes must be reduced, as described above, by maintaining a lower pH of the feed, whereby in Equation 1 the equilibrium shifts to the left.

Example II

This example shows the results obtained using 32,939,557 neutral crosslinked styrene-divinylbenzene copolymer (XAD-4) and neutral crosslinked polyacrylic ester copolymer (XAD-8) and the same feed mixture as in Example I at different pH values, due to its poor pH difference. to be obtained at a pH of 2.4 or higher or above the citric acid ionization constant pKa ^ = 3.13. The separation was performed using the same method and apparatus described in Example I above, except that the temperature was 60 ° C and 5 ml of feed mixture was used.

Figures 4A, 4B and 4C are graphical representations of wood test results using copolymer XAD-4 at pH values of 2.4, 1.7 and 0.9, respectively. Figure 4A shows that citric acid "breaks through" with salts (and carbohydrates). This problem can be partially avoided by lowering the pH to 1.7 in Figure 4B. Excellent separation is achieved by further lowering the pH to 0.9, as shown in Figure 4C. This difference, when adjusting the pH, clearly has good industrial applicability.

Figures 4D and 4E graphically show the results of wood experiments under the above conditions using XAD-8 copolymer at pH 2.8 and 1.4 and temperatures of 65 ° C, respectively. In Figure 4D, which was performed at pH 2.8, there is no difference. salts, carbohydrates and citric acid elute together. After about 67 ml of citric acid has been recovered after a small fraction of carbohydrates, salts and citric acid, relatively pure citric acid can be obtained, but the recovery rate is low. In Figure 4E, which was performed at pH 1.4, there is a selectivity between citric acid and carbohydrates and salts, resulting in satisfactory separation and recovery of citric acid.

Example III

This example illustrates the results obtained using a neutral crosslinked styrene-divinylbenzene copolymer (XAD-4) and the same feed mixture as in Example I at different pH values to detect poor separation at pH 2.4 or higher. The same method and apparatus were used as in Example I, except that the temperature was 93 ° C in quarters 5A and 5B and the volume of the feed mixture was 10 ml.

Figures 5A and 5B show the results of pulse experiments using copolymer XAD-4 at pH 2.8 and 1.4, respectively. Figure 5A shows that citric acid "breaks through" with salts and carbohydrates. This problem can be avoided by lowering the pH to 1.4 as shown in Figure 5B. Such separation and pH adjustment clearly have industrial applicability.

Example IV

In this example, the method and apparatus of Example I were used. The temperature was 60 ° C and 5 ml of feed mixture was used. The feed mixture was similar to that previously used except that the citric acid concentration was 40%. The effect of concentration on pH was studied. In Figure 6A, also at 60 ° C, the pH of 1.9 is too high to separate citric acid at 40% concentration. By adjusting the pH downwards according to Figures 6B and 6C, mainly citric acid is adsorbed, and excellent separation is obtained at pH 1.13 and 0.5. In none of these examples were carbohydrates analyzed, but it can be assumed that the carbohydrates separated together with the salts.

Example V

Using the three samples of this example, the method and apparatus of Example I were applied. The temperature was 93 ° C and the volume of the feed mixture was 5 ml. The feed mixture was similar to that previously used except that citric acid was concentrated to 40% to determine the effect of pH concentration. In the process of Figure 7A, the pH of 1.8 is still too high at 93 ° C to separate citric acid at 40% concentration. By adjusting the pH downwards, as in Figures 7B and 7C, mainly sulfuric acid is adsorbed and excellent separation is obtained. Again, 34,938,57 carbohydrates were not analyzed, but it can be assumed that the carbohydrates separated together with the salts.

Example VI

The pulse experiment of Example I was repeated using two 50% samples of citric acid and copolymer XAD-4 as the adsorbent. In both cases, the desorbent was water. The composition of the feed used was the same as in Example I except that the citric acid was concentrated to 50%. The temperature was 93 ° C. The pH of the first sample was 1.5. As shown in Figure 8A, no citric acid separated. After lowering the pH to 1.0 in the second sample, citric acid separated easily, as shown in Figure 8B. Again, carbohydrates were not analyzed, but were assumed to be eliminated with the salts. The separation according to Figure 8B was defined as good.

Example VII

The separation example shown in Figures 7B and 7C requires high temperatures, e.g., 93 ° C, to achieve 40% citric acid separation due to difficulties in desorbing citric acid from the XAD-4 adsorbent. In this example, those high temperatures that adversely affect the service life and handling costs of the adsorbent are eliminated, and the separation can be easily performed at 45 ° C using a desorption mixture containing 10% by weight of acetone and 90% by weight of water. In the method of Figure 9, a feed mixture containing 40% by weight citric acid, 4% carbohydrates and 2% salts of the following elements: K +, Na +, Mg ++, Fe +++, Ca ++, and proteins and amino acids was fed to the pulse test apparatus previously described and the experiment was performed except that the temperature was 45 ° C. In this experiment, the pH was maintained at 0.5 and the desorbent contained acetone as described above. The retention volume of citric acid was 10.7 ml and the decomposition was 0.61, and as a result, the separation could be easily performed.

35 93857

Example VIII

In this example, four pulse experiments were performed using a weakly basic resin with a tertiary, amine-functional hydrogen bound to a sulfate ion in a cross-linked gel-type acrylic resin matrix (AG4-X4, manufactured by Bio Rad Laboratories, Richmond, CA) with a tertiary hydrogen bound to a sulfate ion in a crosslinked acrylic resin matrix to determine the ability of this adsorbent to separate citric acid from its fermentation broth containing carbohydrates (DPI, DP2, DP3 and Caucose, xylose, arabinose and raffinose), and salt, Fe +++,

Cl, SO 4 -, PO 4 "and NO 2, amino acids and proteins, at pH 1.6. The first experiment was performed at 75 ° C. The other experiments were performed at 60 ° C. Citric acid was desorbed with water in pulse experiment No. 1 (Figure 10). and sulfuric acid at two concentrations: 0.05 (Pulse Test No. 2) and 0.25 N (Pulse Test No. 3) Pulse Test No. 4 was similar to Pulse Test No. 2 except that the adsorbent was used thereafter. after passing through it a volume of 24 layers of feed mixture, the composition of the fermentation feed mixture being as follows:

Mixture fed Quantity

Citric acid 40%

Salts (K +, Na +, Ca ++, Mg ++, Fe +++) 1.5%

Carbohydrates (sugars) 4%

Other (so4_, Cl, po4 ", no3, proteins and amino acids) 5%

Water 49.5%

Retention volumes and separation factor (B) were obtained using the pulse test apparatus and method described in Example I above, except that 5 ml of sample was used. The net retention volume (NRV) of citric acid was calculated by measuring the distance from the center of the hydrochloride curve to the center of the citric acid curve. The difference value B was calculated from the ratio of the retention volumes of the components to be separated to the retention volume of the first salt component (i.e. salts 1).

3 * 93857 The results of these pulse tests are shown in Table 5 below.

Table 5

Resin / de- Feed-

Pulse sorption

test substance NRV B

1 AG4-X4 / water salts 1 1.6 34.25 citric acid 54.8 comparison unknown A 0 offspring unknown B 6.6 8.30 salts 2 54.6 1.00 2 AG4-X4 / 0.05 n H2SO4 salts 3.2 11.87 citric acid 38.0 comparison unknown A 0 residue unknown B 2.7 14.7 3 AG4-X4 / 0.25 n H2SO4 unknown A 0 offspring citric acid 26.9 comparison salts 2.3 11.70 unknown B 7.6 3.54 4 AG4-X4 / 0.05 n H2SO4 unknown A 0 prodrug citric acid 38.0 comparison salts 2.4 15.8 unknown B 7.2 5.28

The results of pulse experiments 2-4 are similar to Figure 10. It can be seen from Table 5 that although citric acid is satisfactorily separated in this process in very pure form using water, desorption by water is slower than with dilute sulfuric acid, resulting in a higher net retention volume. When the adsorbent is aged by passing 24 bed volumes of the feed mixture therethrough, there are no signs of deactivation, as can be seen in Figure 13, which is essentially identical to Figure 11 (performed under identical conditions using fresh adsorbent).

Example IX

The first pulse test according to Example 8 was repeated using the same method and equipment, except that the temperature was 65 ° C.

37 93857

The desorbent was water. This example describes the results obtained using a macroporous, weakly basic anion exchange resin with a cross-linked polystyrene matrix (Dowex 66) and the same feed mixture as in Example VIII (40% citric acid) in the first two pulse experiments, pH 7.0 and 3.5 (resp. Figs. 11A and 11B) to illustrate the failure of the desired separation when the pH is above the first ionization constant of citric acid pKa ^ = 3.13, and more specifically in these two samples at a pH above 1.7 with a concentration of 40% when citric acid is used. In the third part of the example (Figure 11C), the feed was diluted to 13% citric acid and the pH was lowered to 2.4. Although there is a clear improvement, it is obvious that the pH and / or concentration must be further reduced to "prevent breakthrough" of citric acid. For example, at a concentration of 13%, it is estimated that the pH must be lowered to about 1.6-2.2.

Figures 11A and 11B graphically show the results of pulse experiments using Dowex 66 resin at pH 7.0 and 3.5, respectively. Figures 11A and 11B show that citric acid "breaks through" together with salts (and carbohydrates) at higher pH values. This problem can be partially avoided by reducing the concentration to 13% and lowering the pH to 2.4, as shown in Figure 11C, which shows, ‘! that only a small amount of citric acid is not adsorbed and "breaks through" the raffinate, while most of it is adsorbed on the adsorption resin (but not desorbed in this figure). This difference, when adjusting the concentration and pH to the optimum level, apparently has good industrial applicability.

Example X

Three more pulse experiments were performed under the conditions of Example 8, unless otherwise noted, using citric acid samples from the same feed mixture, but with two different adsorbents. The desorbent was 0.05 n in the first two samples: H2SO4 (Figures 12 and 13A), while water was used in the third sample (Figure 13B). The composition of the feed mixture used was the same as in Example VIII. The temperature was 60 ° C and the pH was 1.6. The adsorbent No. 1 used in the first experiment was a macroporous, pyridine-function crosslinked divinylbenzene resin having the following formula: CH3

P - CH2 - N - CH2 -O

·· H + so4 = where P is a resin-forming polystyrene group.

The second adsorbent (No. 2) used in the second and third samples was a tertiary amine which also had a pyridine functional group having the following formula:

P - CH2 - N - CH2 -O

CH ~ I 2 + = CHOH H SO.

I 4 ch3 where P is as defined above. Both resins are crosslinked with divinylbenzene. In a few cases, although water is an effective desorbent that provides excellent separation, it is not strong enough to recover adsorbed citric acid fast enough to make the process industrially useful. See Fig. 13B, where the conditions are the same as above when using adsorbent No. 2 and water as desorbent. In this case, the citric acid does not elute until about 95 ml of desorbent has passed through the adsorbent. Dilute sulfuric acid is therefore the most preferred desorbent, as shown by the results in Figures 12 and 13A. It can also be seen from Figures 12, 13A and 13B that excellent citric acid separation is achieved.

Example XI

The procedure of Example VIII, its conditions and equipment were used to separate citric acid from four samples of the same feed mixture using two different resins belonging to the group of adsorbents in Example VIII (except that in the first and fourth samples the column temperature was 50 ° C). and the desorbent was 0.05 N I 2 SO 2, and the second and third samples had a pH of 2.2 and the desorbent was dilute sulfuric acid at a concentration of 0.15 N). Both IRA-68 and IRA-35 resins, manufactured by Rohm and Haas, have an amine function and the following structural formula: R 'I ♦ = P - CH, - N: H SO.

2 I 4 R "wherein P is a polyacrylic matrix and R * and R" = CH2.

Amberlite IRA-68 (Samples 1, 2 and 3) is a gel type resin. IRA-35 (Sample 4) is a macroreticular resin. Sample 3 was identical to Sample 2 except that the adsorbent had previously been used to separate the bed volume of the feed mixture 69. Samples 1 and 2 are both excellent adsorbents for separating citric acid from its fermentation broth in the pH range 1.6-2.2. Sample 3 shows the stability of the resin (little or no deactivation) after the adsorbent has been used in an amount of 69 bed volumes of feed! in this distinction. The net retention volume (NRV) and selectivity (B) are shown in Table 6 below.

4o 93857

Table 6 Sample

n; o Resin Ingredient NRV B

1 Amberlite salts 5.5 8.24 IRA-68 citric acid 45.3 comparison unknown A 0 progeny unknown B 9.3 4.87 2 Amberlite salts 2.3 12.61 IRA-68 citric acid 29.0 comparison unknown A 0 prodrug unknown B 6.5 4.46 3 Amberlite salts 2.85 10.32 IRA-68 citric acid 29.4 comparison unknown A 0 progeny unknown B 7.0 4.2 4 Amberlite salts 1.3 27.38 IRA-35 citric acid 36.9 comparison unknown A 0 offspring unknown B 5.9 6.25

To compare the adsorbents of Examples I to VII with the adsorbents of Examples VIII to XI, several extract samples were analyzed for the determination of readily carbonized contaminants (RCS) (Food & Chemical Codex (FCC) Monograph No. 3) and potassium content. The amount of RCS was determined as follows: 1 g of the extract sample (the actual concentration of citric acid was determined) was carbonized at 90 ° C using 10 ml of 95% SO 2 solution. The carbonated material was examined spectrophotometrically at 500 nm using a 2 cm cell and a 1.27 cm diameter tube, and the amount of RCS was calculated for 50% citric acid solution. The value obtained can be compared to that obtained using this test method in the FCC test mentioned above and using a standard solution of cobalt. Potassium was determined spectroscopically by atomic adsorption. For comparison, the same analytical determinations were performed using the same feed sample, and the RCS value was calculated for 50% citric acid using resins XAD-4, AG4-X4 and adsorbents 1 and 2 according to Example III. The results are shown in Table 7.

4i 93857

Table 7

Extract quality (RCS / potassium) in the pulse test RCS C.A. net

Adsorption Desorption calculated (50% retention aid_ substance_ citric acid) ppmK volume XAD-4 H 2 O 6.86 f 8.98 59, 137 13.0 AG4-X4 0.05 n.H 2 SO 4 1.77, 1.42 24 .81 34.8 No. 2 (e.g. X) 0.05 n * H 2 SO 4> 3.17, 3.33 24, 54 30.8 No. 1 (e.g. X) 0.05 n «H 2 SO 4 2.17 62 31.0 These values indicate the advantage of reducing the RCS and K values of weakly basic resins over neutral resins. In all samples the RCS value had decreased by at least 50% and in two samples the K value had decreased by more than 50%. From the net retention volume, it can be stated that both groups of adsorbents have a good dispersion value, but adsorbents based on strong bases have the disadvantage of somewhat increased cycle lengths. The lengths of the period can be reduced by using sulfuric acid at higher concentrations, e.g. always about 0.2 n and preferably 0.1-0.2 n.

In another embodiment, the citric acid adsorbed on the adsorbent can be turned in situ to citrate prior to desorption, e.g., by reaction with an alkaline earth metal or alkali metal hydroxide or ammonium hydroxide, and then immediately eluted using metal hydroxide, ammonium hydroxide or water as the desorbent. Over time, the adsorbent may be deactivated by unknown impurities, but the adsorbent may be regenerated by rinsing it with a stronger desorbent, e.g. sulfuric acid stronger than the desorbent, time metal hydroxide or NH 4 OH, or as an organic solvent, or an organic solvent.

42 93857

Example XII

In this example, two pulse experiments were performed using a gel-type, strongly basic anion exchange resin (IRA 458, manufactured by Rohm & Haas Co.) having the structural formula (1) shown above on page 22 and substituted with three methyl groups to determine the citric acid resolution of the adsorbent. carbohydrate-containing fermentation broth (DPI, DP2, DP3, as well as glucose, xylose, arabinose and raffinose), and salt ions such as Na +, K +, Mg ++, Ca ++, Fe +++, Cl ”, SO 2, PO 0 and NO 2, amino acids and proteins , at pH 2.2.

P is acrylic crosslinked with divinylbenzene. Pulse test sample 1 was treated at 50 ° C. Pulse test sample 2 was treated at 60 ° C, but after the bed was aged using 33 bed volumes of feed mixture.

Treatments were then performed using 62 bed volumes without substantial deactivation in the adsorbent. Citric acid was desorbed with a 0.1 N solution of sulfuric acid in each sample. The composition of the fermentation broth was as follows:

Feed mix%

Citric acid 40

Salts (K +, Na +, Ca ++, Mg ++ Fe +++) 1.5

Carbohydrates (sugars) 4

Other (SO 4 =, Cl ", PO 4 =, NO ~ 5; proteins and amino acids)

Water 49.5

Retention volumes and separation factor were obtained using the pulse test apparatus and method described in Example I above.

The results of these pulse tests are shown in Table 8 below.

Table 8 43 9 3 8 5 7 Sample

No. Resin Ingredient NRV B

1 IRA-458 salts 1.0 38.9 citric acid 38.9 comparison unknown S 0 progeny unknown B 6.6 5.89 2 IRA-458 salts 0.9 43.3 citric acid 39.0 comparison unknown A 0 progeny unknown B 7.1 5.49

It is evident that the citric acid separates satisfactorily in this treatment, and when the adsorbent was aged using 33 bed volumes of the feed mixture, there were no signs of deactivation, giving substantially similar results as under the same conditions with fresh adsorbent.

Example XIII

The pulse experiment of Example XII was repeated using additional samples of citric acid and the same feed mixture, but different macroporous, strongly basic anion exchange resin IRA-958 with quaternary ammonium function and acrylic resin matrix cross-linked with divinylbenzene. The desorbent was 0.05 n 2 SO 2. Used feed mixture; the composition was the same as in Example XII. The temperature was 60 ° C and the pH was 1.6. An adsorbent resin manufactured by Rohm & Haas having the structure shown on page 22 (1) in which R is methyl was used in this experiment.

As shown in Figure 14, the citric acid begins to elute after 45 ml of desorbent has passed through the adsorbent and is very easily separated from the fermentation broth in very pure recovery with excellent recovery.

Example XIV

The pulse experiment of Example XII was repeated using additional citric acid samples and the same feed mixture but a different strongly basic anion exchange resin adsorbent AG2-X8 (manufactured by Bio Rad Company) having the structural formula (2) above on page 22 where R is methyl, together with a crosslinked polystyrene-gel type resin matrix having quaternary ammonium functional groups. The desorbent was 0.15 N H2SO4. The composition of the feed mixture used was the same as in Example XI. The temperature was 50 ° C and the pH was 2.2.

As shown in Figure 15, citric acid begins to elute after 43 ml of desorbent has flowed through the adsorbent and separates very efficiently from the fermentation broth in very pure recovery with excellent recovery.

Further comparing the adsorbents of Examples I-VII with the adsorbents of Examples XII-XIV, several carbon samples were analyzed for readily carbonized impurities (RCS) (Food & Chemical Codex (FCC) Monograph No. 3) and potassium as described above. The results obtained for each of the adsorbents XAD-4, IRA 458, IRA 959 and AG2-X4, as well as the quality of the adsorbent, are shown in Table 9 below.

Table 9

Extract quality (RCS / potassium) in pulse test ·, RCS unit ppmK CA net

Adsorbents (calculated (calculated retention aid Desorbent tu 50% citric acid) 50% citric volume _ _ _ acid) _ XAD-4 H 2 O 8'98 137 13 '° IRA 458 0.1 n H 2 SO 4 1, 5 80 37.9 IRA 958 0.05 n H2SC> 4 2.73 82 32 AG2-X4 0.15 n H2SO4 5.3 131 43 These values show that the improvement brought about by the decrease in RCS and K values is clear in strongly alkaline resins compared to neutral resins. In all samples, the RCS value was reduced by 40-85% and the K value by 0-20%. Examples 45 9 3 8 5 7 XII, XIII and XIV (Figure 4) show that each class of adsorbents has good resolution, but the adsorbents used have the disadvantage of somewhat longer circulation times. These cycle times can be reduced by using higher concentrations of sulfuric acid up to about 0.2 n, with the most preferred range being 0.1-0.2 n.

According to one embodiment, the citric acid adsorbed by the adsorbent can be converted to citrate in situ before desorption, e.g. by reaction with alkaline earth metal hydroxide, alkali metal hydroxide or ammonium hydroxide, and then immediately eluted using metal hydroxide, ammonium hydroxide or water as desorbent. Over time, unknown impurities can deactivate the adsorbent, but the adsorbent can be regenerated by rinsing it with a stronger desorbent, e.g. stronger sulfuric acid flowing into the desorbent, alkali metal hydroxide or NH 4 OH, or an organic solvent, e.g.

Claims (8)

93857 A process for separating citric acid from a fermentation broth consisting of a mixture containing citric acid, characterized in that a mixture having a pH lower than the pKal of citric acid is contacted with a polymeric adsorbent selected from the group consisting of neutral, cross-linked polystyrene polystyrene , a hydrophobic polyacrylic teripolymer, a weakly basic anion exchange resin having tertiary amine or pyridine functional groups, and a strongly basic anion exchange resin having an adsorbent adsorbent adsorbent, in which the adsorbent, and mixtures thereof, and that the citric acid is then recovered from said adsorbent with a desorbent consisting of water or a mixture of water and an inorganic acid under desorption conditions. A process according to claim 1, characterized in that the adsorption and desorption conditions consist of a temperature of about 20 to 200 ° C and a pressure of about atmospheric pressure to about 3450 kPa. Process according to Claim 1, characterized in that the adsorbent has a tertiary amine-functional group in the matrix formed by the crosslinked acrylic resin. Process according to Claim 1, characterized in that the adsorbent has a pyridine-functional group in a matrix formed by a crosslinked polystyrene resin. Process according to Claim 1, characterized in that the adsorbent has a quaternary amine-functional group in the matrix formed by the crosslinked acrylic resin. Process according to Claim 1, characterized in that the adsorbent contains a quaternary ammonium salt of pyridine in a matrix formed by a crosslinked polystyrene resin. Method according to any one of claims 1 to 6, characterized in that it comprises the steps of: (a) maintaining a net flow of liquid through said adsorbent column in parallel, said column comprising at least three zones of separate operations connected in series with the column terminal -the zones are connected to provide a continuous connection between the zones; (b) maintaining an adsorption zone in this column, the zone being defined by an adsorbent located between the feed inlet flow at the upstream boundary of the zone and the raffinate effluent flowing in the downstream boundary of the zone; (c) maintaining the purge zone immediately upstream of said adsorption zone, said purge zone being defined by an adsorbent located between the extract effluent at the upstream boundary of the purge zone and the feed inlet stream at the downstream boundary of said zone; (d) maintaining a desorption zone upstream of said purification zone, said desorption zone being defined by an adsorbent located between the inlet stream of the desorbent at the upstream boundary of this zone and the extract effluent at the downstream boundary of this zone; (e) introducing the feed mixture into the adsorption zone under adsorption conditions to effect selective adsorption of citric acid by the adsorbent in the adsorption zone, and removing the raffinate inlet stream containing the fermentation broth components leaving the adsorption zone; (f) introducing a desorbent into the desorption zone under desorption conditions to displace citric acid from the adsorbent in the desorption zone; 93857 (g) removing an extract effluent containing citric acid and a desorbent from the desorption zone; (h) directing at least a portion of the effluent extract stream to the separation devices and removing at least a portion of the desorbent under the separation conditions; and (i) periodically passing the zone boundaries formed by the feed inlet stream, the raffinate effluent stream, the desorbent effluent stream, and the extract effluent stream through the adsorbent column downstream of the adsorption zone to provide flow zones in the adsorption zone and the effluent. A method according to claim 7, characterized in that it comprises the step of maintaining a buffer zone immediately upstream of the desorption zone, said buffer zone being defined by the adsorbent located between the inlet of the desorbent at the downstream boundary of the buffer zone and the upstream boundary of the buffer zone.