EP4665167A1

EP4665167A1 - Activated carbon extraction process for food and beverage ingredients

Info

Publication number: EP4665167A1
Application number: EP24757663.0A
Authority: EP
Inventors: Zhu Gao; Kasi V. SOMAYAJULA; Sangphyo HONG; Robert Kriegel; Nicholas Pierce MCCARTY
Original assignee: Coca Cola Co
Current assignee: Coca Cola Co
Priority date: 2023-02-16
Filing date: 2024-02-15
Publication date: 2025-12-24
Also published as: WO2024173631A1; JP2026505493A; MX2025009572A; CN121240777A

Abstract

This disclosure provides new extraction (filtration) systems, extraction media, and continuous extraction processes for effectively removing low molecular weight compounds produced during the commercial production of edible caramel color, particularly the compound 4-MeI. In an aspect, the process can include a sequence of an ultrafiltration to provide a retentate solution and a permeate solution, and contacting the permeate solution with a granular activated carbon for a contact time sufficient to provide a purified permeate solution with substantially lower concentrations of 4-MeI. If desired, the purified permeate solution can be reused by combining with the retentate solution.

Description

ACTIVATED CARBON EXTRACTION PROCESS FOR FOOD AND BEVERAGE INGREDIENTS

CROSS-REFERENCE TO RELATED APPLICATIONS

[0101] The present application claims priority to U.S. Provisional Application No. 63/485,367, filed February 16. 2023, the contents of which are hereby incorporated by reference herein.

TECHNICAL FIELD

[0102] This disclosure relates to a purification process for reducing the amount of low molecular weight species from food and beverage ingredients.

BACKGROUND

[0103] Caramel color compositions are widely used to impart a desired color to food and beverage products. These edible color compositions are generally produced by the controlled heating of carbohydrates such as sugars, com-syrup, and the like, either alone or in the presence of other ingredients. The caramel colors produced using these processes can have colors ranging from pale yellow to red or red-brown to dark brown, and typically contain complex mixtures of compounds.

[0104] The International Technical Caramel Association (ITCA) classifies caramel colors according to four classes, based on their method of manufacture, composition, and functional properties. One element of the color classification system is whether the color is produced by heating the carbohydrate in the presence or absence of compounds such as ammonia, ammonium compounds, or sulfite compounds. Each method of manufacture produces its own set of complex impurities, and subsequent processing is typically used to remove or reduce the concentrations of certain components.

[0105] Therefore, there remain substantial challenges for developing new and improved processes which can efficiently remove certain low molecular weight species produced during the commercial production of edible caramel colors. SUMMARY OF THE DISCLOSURE

[0106] This disclosure provides new and improved processes and methods for effectively and efficiently removing certain low molecular weight species produced during the commercial production of edible caramel color used in food or beverage applications, including compounds such as 4-methylimidazole (4-MeI or 4-MEI). Although methods of producing caramel colors have been disclosed which generate reduced levels of 4-MeI, for example, in U.S. Patent Appl. Publ. No. 2010/0003383 and U.S. Patent No. 4,416.700, there is a need for improved separation, filtration, and/or extraction methods to remove 4-MeI from such color components.

[0107] Generally, edible caramel color compositions are manufactured by heating carbohydrates including commercially available food-grade nutritive sweeteners either alone or in the presence of food-grade acids, alkalis or salts. The acids, alkali, and/or salts function as catalysts which give the caramel specific color and functional properties. Edible caramel color compositions made without acids, alkali, and/or salts are generally labeled as “burned sugar”, and are used as a flavor with incidental color, rather than as strictly coloring agents.

[0108] Edible caramel compositions are divided into four general classes. Class I compositions, also referred to as plain caramel or caustic caramel, are prepared by heating carbohydrates with acids or alkalis, but without the use of any ammonium or sulfite compounds. Class II caramel, also referred to as caustic sulfite caramel, is prepared by heating carbohydrates with acids or alkalis in the presence of sulfite compounds, but without the use of ammonium compounds. Generally, class II caramels have a reddish hue and are stable in acid above pH values of about 2.5-3. Class III, also referred to as ammonia caramel, ammonia process caramel, closed-pan ammonia process caramel, open-pan ammonia process caramel, bakers’ caramel, confectioners' caramel, or beer caramel, is prepared by heating carbohydrates with acids or alkalis in the presence of ammonium compounds, but without the use of sulfite compounds. Class IV, also referred to as sulfite ammonia caramel, is prepared by heating carbohydrates with acids or alkalis in the presence of both sulfite and ammonium compounds. Generally, class IV caramels have a brown or black hue and are generally stable in acid above pH of about 1.5.

[0109] Compounds such as 4-MeI and other methyl imidazoles can arise as by-products from the production of the caramel color that uses ammonia as a reactant. Therefore, in an aspect, the methods of this disclosure are particularly useful for class III caramel color or class IV caramel color. The structure of 4-MeI is illustrated here. 4- Methylimidazole may be produced, for example, as a by-product from a production method that uses ammonia as a reactant.

4-MeI

[0110] In an aspect of this disclosure, an ultrafiltration is used to remove as much of the low molecular weight compounds as possible from a caramel color composition, including 4-MeI. Ultrafiltration provides a permeate which contains the low molecular weight compounds but also substantial solids most of which could be reused except for the presence of the 4-MeI compound.

[OHl] Another aspect of this disclosure provides a continuous extraction process using activated carbon adsorbent which can remove low molecular weight compounds such as 4-MeI from an aqueous permeate fraction with a high removal efficiency of about 90% to 99%. allow ing the remaining compounds such as glucose and sucrose to be available for recovery and re-use. The ability to add the recovered solids back to the retentate solution from the ultrafiltration process may provide a more micro-stable ingredient as well as closer and more consistent taste profile. Developing such a continuous extraction process using activated carbon adsorbent required identifying adsorbents, discovering the various factors which could affect the separation of low molecular w eight compounds and enable recovery of the solids, and understanding how the adsorbents and factors could be altered to enhance the separation and recovery' process.

[0112] Accordingly, in an aspect, this disclosure provides a process for purifying a caramel color solution, the process comprising:

(a) providing a caramel color solution comprising low molecular weight species and high molecular weight species;

(b) filtering the caramel color solution by a ultrafiltration process to provide a retentate solution comprising the high molecular weight species and a permeate solution comprising the low' molecular weight species in a first concentration; (c) contacting the permeate solution with a granular activated carbon for a contact time sufficient to provide a purified permeate solution comprising the low molecular weight species in a second concentration which is lower than the first concentration.

In an aspect, the granular activated carbon can be contained in a filter bed through which the permeate solution flows at a certain flow rate. This type of filter bed, such as a column of activated carbon, provides a convenient continuous process. According to a further aspect, the purified permeate solution can be combined with the retentate solution to provide a purified caramel color solution, to recover and reuse the remaining compounds such as glucose and sucrose.

[0113] Accordingly, this disclosure also provides a continuous process for purifying a caramel color solution, the process comprising:

(b) filtering the caramel color solution by a ultrafiltration process to provide a retentate solution comprising the high molecular weight species and a permeate solution comprising the low molecular weight species in a first concentration;

(c) flowing the permeate solution through a first granular activated carbon filter bed at a flow rate sufficient to provide a first purified permeate solution comprising the low molecular weight species in a second concentration which is lower than the first concentration.

In another aspect, this process described immediately above can further comprise the step of:

(d) flowing the first purified permeate solution through a second granular activated carbon filter bed in fluid communication with the first granular activated carbon filter bed at the flow rate to provide a second purified permeate solution comprising the low molecular weight species in a third concentration which is lower than the second concentration.

[0114] These and other features, embodiments and aspects of the processes, methods, and compositions are described more fully in the Detailed Description and claims and further disclosure such as the Examples provided herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] FIG. 1 illustrates a plot of 4-MeI extraction efficiency (%) versus the number of bed volumes for the 3.5% Brix solutions, demonstrating the effect of flow rate on extraction efficiency.

[0116] FIG. 2 provides a plot of the residual 4-MeI versus the number of bed volumes for the 3.5% Brix solutions, demonstrating that bed contact times of from 18-36 minutes (flow rate of 0.5-1 L/h) was the minimum contact time to efficiently extract 4-MeI.

[0117] FIG. 3 illustrates the evolution of pH of the eluted purified permeate solution as a function of bed volumes at flow rates of 0.5 L/h and 1 L/h, demonstrating the initial increase in the pH which stabilized to a value close to the starting pH of the permeate of pH 7.

[0118] FIG. 4 illustrates the evolution of Brix of the eluted purified permeate solution as a function of bed volumes at flow rates of 0.5 L/h and 1 L/h. demonstrating the initial decrease in Brix which stabilized to a value close to the starting Brix of the permeate.

[0119] FIG. 5 illustrates the evolution of Absorbance of the eluted purified permeate solution as a function of bed volumes at flow rates of 0.5 L/h and 1 L/h, demonstrating significant adsorption of color bodies from the permeate in the initial fractions, which trend toward the starting Absorbance of the permeate in a linear fashion.

[0120] FIG. 6 shows the effect of permeate pH on 4-MeI extraction efficiency (%) at a permeate Brix of 3.5% as a function of bed volumes, plotted for pH 3 and pH 7, each at a flow rate of 0.5 L/h and 1 L/h.

[0121] FIG. 7 demonstrates the effect of Brix % on the extraction efficiency for 4-MeI, plotting extraction efficiency (%) versus bed volumes for high Brix (about 7%), mid Brix (2.7-3.5%) and low Brix (0.8-1%) permeate.

[0122] FIG. 8 illustrates the effect of activated carbon contact on the evolution of pH, plotting pH of the eluent versus bed volumes for high Brix (about 7%). mid Brix (2.7-3.5%) and low Brix (0.8-1%) permeates.

[0123] FIG. 9 illustrates the effect of activated carbon contact on the evolution of Brix%, plotting Brix% of the eluent versus bed volumes for high Brix (about 7%), mid Brix (2.7-3.5%) and low Brix (0.8-1%) permeates. [0124] FIG. 10 illustrates the effect of activated carbon contact on the evolution of Absorbance, plotting Absorbance of the eluent versus bed volumes for high Brix (about 7%), mid Brix (2.7-3.5%) and low Brix (0.8-1%) permeates.

DETAILED DESCRIPTION

[0125] To define more clearly the terms used herein, the following definitions are provided, and unless otherwise indicated or the context requires otherwise, these definitions are applicable throughout this disclosure. If a term is used in this disclosure but is not specifically defined herein, the definition from the IUPAC Compendium of Chemical Terminology, 2^nd Ed (1997) can be applied, as long as that definition does not conflict with any other disclosure or definition applied herein, or render indefinite or nonenabled any claim to which that definition is applied. To the extent that any definition or usage provided by any document incorporated herein by reference conflicts with the definition or usage provided herein, the definition or usage provided herein controls.

[0126] The terms low molecular weight species and high molecular weight species referring to components of the caramel color solution are relative terms, and may vary according to the molecular weight cut-off of the specific membrane used in any specific filtration process such as ultrafiltration. Thus, for all membranes, filtering the caramel color solution by an ultrafiltration or other size exclusion process provides a retentate solution which is enriched in the high molecular weight and a permeate solution which contains higher amounts or is enriched in the low molecular weight species present in the original solution. In an ultrafiltration process such as used in this disclosure, molecules such as 4- methylimidazole (4-MeI), 2-methylimidazole (2-MeI), 5-hydroxy-2-methylpyridine, 2- hydroxypyridine, 2-hydroxy-6-methylpyridine carboxylic acid, glucose, and sucrose and the like can be concentrated in the permeate solution, while other larger molecules, oligomers, and polymers such as larger oligosaccharides can be concentrated in the retentate solution. In an aspect, the disclosed ultrafiltration process can employ a membrane having a molecular weight cut off (MWCO) of about 800 Da (dalton), such that species having a molecular weight greater than about 800 Da are retained in the retentate.

[0127] When numerical ranges are disclosed, Applicant's intent is to disclose or claim individually each number that such a range could reasonably encompass, including end points of the range as well as any sub-ranges and combinations of sub-ranges encompassed therein, unless otherwise specified. For example, by disclosing a temperature of from 70 °C to 80 °C, Applicant’s intent is to recite individually 70 °C, 71 °C, 72 °C, 73 °C, 74 °C, 75 °C, 76 °C. 77 °C, 78 °C, 79 °C, and 80 °C, including any sub-ranges and combinations of sub-ranges encompassed therein, and these methods of describing such ranges are interchangeable. Applicant reserves the right to proviso out any individual values, ranges, or sub-ranges that can be claimed according to a range, if for any reason Applicant chooses to claim less than the full measure of the disclosure.

[0128] Values or ranges may be expressed using the term “about”, and when such values or ranges are expressed, other embodiments disclosed include the specific value recited, from the one particular value, and/or to the other particular value. In one aspect, use of the term “about” can be replaced with ±15% of the stated value, ±10% of the stated value, or ±5% of the stated value.

[0129] In an aspect, this disclosure provides new and improved processes and methods for effectively removing low molecular weight compounds produced during the commercial production of edible caramel color, particularly the compound 4-MeI. In an aspect, this disclosure provides a process for purifying a caramel color solution, the process comprising:

(b) filtering the caramel color solution by an ultrafiltration process to provide a retentate solution comprising the high molecular weight species and a permeate solution comprising the low molecular weight species in a first concentration;

(c) contacting the permeate solution with a granular activated carbon for a contact time sufficient to provide a purified permeate solution comprising the low molecular weight species in a second concentration which is lower than the first concentration.

The caramel color solution is an aqueous solution or composition. Low molecular weight species described in this process includes the compound 4-methylimidazole (4-MeI), but also may include other compounds such as 2-methylimidazole (2 -Mel), 5-hydroxy-2- methylpyridine, 2-hydroxypyridine, 2-hydroxy-6-methylpyridine carboxylic acid, or combinations thereof. [0130] The low molecular weight species 4-MeI is particularly present when the caramel color is selected from a class III caramel color or a class IV caramel color. Therefore, the process developed in this disclosure and further set out in the examples focused on the 4-MeI compound.

[0131] Accordingly, the ultrafiltration process selectively removes as much of these low molecular weight compounds as possible from the caramel color composition, particularly 4-MeI, but a number of high molecular weight compounds such as glucose and sucrose can also pass through the membrane and occur in the permeate. This carbon filtration process allows selective removal of the low molecular weight compounds from the permeate, thereby allowing the purified permeate to be added back to the retentate to produce a more micro-stable ingredient as well as closer and more consistent taste profile.

[0132] While the activated carbon step can be carried out in a batch mode, this disclosure also provides an efficient continuous extraction process with the activated carbon adsorbent to remove compounds such as 4-MeI from an aqueous permeate fraction with a high removal efficiency of about 90% to 99%, allowing the remaining compounds such as glucose and sucrose to be available for recovery and re-use. The steps of identifying compounds to be removed, identifying activated carbon as a suitable filtration media or adsorbent, discovering the various factors which could affect the separation efficacy and enable recovery' of the desirable solids, and understanding how' the adsorbents and factors could be altered to enhance the separation and recovery process provided unexpected results, such as unexpected combinations of features that could improve the separation efficiency.

[0133] By the ultrafiltration process the low' molecular weight compounds have been removed from the caramel color solution to provide a retentate and a permeate containing these compounds. By the activated carbon purification process, those low molecular weight components have been separated from the high molecular weight components of the permeate such as glucose and sucrose, and the now -purified permeate solution can be recycled and reused. For example, the purified permeate solution can be combined with the retentate solution to provide a purified caramel color solution.

[0134] Accordingly, in an aspect, this disclosure provides a process for purifying a caramel color solution, the process comprising: (a) providing a caramel color solution comprising low molecular weight species and high molecular weight species;

(b) filtering the caramel color solution by an ultrafiltration process to provide a retentate solution compnsing the high molecular weight species and a permeate solution comprising the low molecular weight species in a first concentration;

In this aspect, the granular activated carbon can be contained in a filter bed through which the permeate solution flows at a certain flow rate, which enables this process to be carried out as a continuous process.

[0135] According to a further aspect, once the purified permeate solution is collected with reduced concentrations of the low molecular weight compounds, the purified permeate solution can be combined with the retentate solution to provide a purified caramel color solution. Several low molecular weight species and various types of low molecular weight species could be removed from the permeate in the disclosed process. Examples of low molecular weight species include but are not limited to 4-methylimidazole (4-MeI), 2- methylimidazole (2-MeI), 5-hydroxy-2-methylpyridine, 2-hydroxypyridine, 2-hydroxy-6- methylpyridine carboxylic acid, or a combination thereof. The compounds 4-MeI (4- methylimidazole) is particularly targeted with this methodology and the examples describe quantitative data for their removal under various conditions.

[0136] One aspect of this disclosure provides for sufficient contact time between the permeate solution and the granular activated carbon to provide a purified permeate solution comprising the low molecular species. The contact time will vary according to the flow rate of the permeate through a bed or column of the activated carbon, but in embodiments, the contact time can be from about 15 min (minutes) to about 2 h (hours), from about 20 min to about 1.5 h, or from about 25 min to about 1 h. The flow rates needed to achieve these contact times will depend on the size and shape of the activated carbon bed. In an aspect, using a lab scale column having a volume of the carbon bed (bed volume) of 300 cm³, the permeate solution flows through the filter bed at a flow rate of from 0. 1 L/h to 1.2 L/h, or a flow rate of 0.3 L/h to 1 L/h.

[0137] Using the disclosed process, particularly when operating at low flow rates to provide sufficient contact time as described herein, about 90% or more of the low molecular weight species 4-MeI can be removed upon contacting the permeate solution with granular activated carbon to provide the purified permeate solution. For example, at a permeate Brix of about 3.5%, the 4-MeI concentration of around 500 ppb (parts per billion, by weight) in the permeate can be reduced to a 4-MeI concentration of around 50 ppb in the purified permeate solution. In some aspects, with sufficient contact time (e.g. 18-36 mins) at low flow rates, around 98% of the low molecular weight species 4-MeI can be removed using activated carbon, such that the 4-MeI concentration of around 500 ppb in a Brix 3.5% permeate can be reduced to a 4-MeI concentration of around 10 ppb in the purified permeate solution. Contact time and flow rate examples are provided in Example 3.

[0138] In an aspect, contacting the permeate with the activated carbon resulted in a change or evolution of the permeate eluent in pH, Brix %, and color, as compared with the permeate itself, as detailed in Example 4. It was observed that as the permeate elutes through the column, the early or initial fractions of the eluent showed a higher pH, a lower Brix, and a lower UV-Vis absorbance as compared to the starting permeate solution. The values of pH, Brix, and absorbance of the eluents approached or tended towards the starting values in the permeate solution as more bed volumes were eluted.

[0139] It was also observed that the Brix % of the permeate affected the pH, Brix %, and color of the early or initial fractions of the eluent, and their evolution over time and bed volumes, as examined in Example 7. The initial changes in pH of the eluent from the starting pH of the permeate were significant, and the pH of the eluent approached the starting values in the permeate solution as more bed volumes were eluted; see Example 7.

[0140] The natural pH of the permeate solution from the ultrafiltration step usually has a pH of about pH 3±0.5 or about pH 3±0.25. Removal of the low molecular weight compounds by activated carbon can occur at this pH, but it has been discovered that removal is not as efficient at pH around 3 as it is at higher pH values; see Example 5. The data in Example 5 demonstrate that at pH 3, the activated carbon saturates earlier compared to at pH 7, however increasing the contact time by slowing the flow rate can significantly improve the performance at pH 3, and reduces the residual concentrations of 4-MeI. In an aspect, the retentate at pH of around 3 can be treated with a base, such an alkali metal hydroxide, to raise the pH and improve the efficiency of low molecular weight compound removal by activated carbon. Therefore, in an aspect, the pH of the permeate solution can be adjusted to a pH of about 6.5-7.5 or a pH of about 7 prior to contacting the permeate solution with a granular activated carbon.

[0141] In an aspect, the permeate Brix % was also observed to affect the extraction efficiency of 4-MeI, as demonstrated in Example 6. The 4-MeI content also was found to vary with Brix, and permeate samples having a range of % Brix values from 1% to 7% were examined, using high Brix (about 7%), mid Brix (2.7-3.5%) and low Brix (0.8- 1%) permeate samples. It was found that activated carbon consumption was higher for permeate of Brix >4%, due to higher 4-MeI. At high Brix, increased concentrations of 4- Mel in permeate also leads to a higher uptake of 4-MeI. However at these high Brix values and uptake rates, the carbon bed will saturate earlier leading to an increased carbon consumption.

[0142] The effect of permeate temperature on the extraction process using activated carbon was examined, as described in Example 8. Bacterial growth within the activated carbon bed can be an issue, and microbial build up can lead to problems within activated carbon units over days of use. In these tests, elevated temperatures of above about 60 °C may mitigate bacterial growth, therefore temperatures from about 60 °C to 80 °C may be useful to mitigate microbial growth in the carbon bed during extraction. No adverse impact of elevated temperatures (60-80 °C) on the extraction of 4-MeI was observed. Accordingly, contacting the permeate solution with the granular activated carbon can be carried out at temperatures of from 55°C to 85°C or from 60°C to 80°C, without adverse effects.

[0143] Example 9 described tests of activated carbon that has been regenerated by removal of adsorbed low molecular weight compounds. Regenerated carbon that had been prepared by w ashing spent carbon with organic solvents such as ethyl acetate or acetic acid did show activity' for removing 4-MeI, which can minimize waste through recycling of the activated carbon. Performance for 4-MeI removal was lower compared to virgin carbon. Regeneration with ethyl acetate lowered the extraction efficiency of 4-MeI. Regeneration with acetic acid lowered the extraction efficiency of 4-MeI. [0144] In Example 10, two activated carbon beds in series were tested for their efficacy, with the possibility of using dual columns for applications that require frequent adsorbent replacement or have a high consumption rate of carbon, for example, in high Brix % permeate. When tested with a permeate flow rate of 0.3-0.5 L/h (residence time of 36-40 mins) for a 3-4% Brix permeate at a temperature of 60 °C, these tests did not saturate the carbon beds and extraction efficiency remained high (>99%) for 4-MeI for up to 50 bed volumes tested.

[0145] In an aspect, the pre-conditioning and post-conditioning the activated carbon bed using a so-called “sweetening on’’ and “sweetening off’ sequence of steps may reduce or minimize the initial changes in the early eluted permeate fractions in pH, loss of color bodies, and changes in Brix. This “sweetening on’’ step uses an initial reverse flow of cold water to remove fines and particulates, followed by a reverse flow of a solution similar to the permeate in Brix and pH, but absent of 4-MeI, until the outlet solution Brix and pH are equivalent to the inlet Brix and pH. This pre-conditioned column can be used for continuous extraction of the permeate until low molecular weight compound breakthrough is observed. A post-conditioning “sweetening off’ step then can be conducted using a reverse-flow hot water flush for 2-3 bed volumes to remove sugar which adsorbs to the carbon bed. If the carbon bed cools down prior to completing the sw eetening off process, it may induce solidification which can make it difficult to remove salts and regenerate the carbon.

[0146] Accordingly, in an aspect, combinations of conditions useful in the disclosed process can include: (i) a permeate Brix of 3-4% and the pH of the permeate being adjusted to a pH of 6.5-7.5; (ii) the contact time is from 15 min (minutes) to 3 h (hours), from 30 min to 2 h, or from 40 min to 1.5 h; and (ii) contacting the permeate solution with the granular activated carbon at a temperature of from 55°C to 85°C or from 60°C to 80°C.

[0147] In a further aspect, combinations of conditions useful in the disclosed process can include: (i) a permeate Brix of 5-7% and a pH of 2.5-3.5; (ii) the contact time is from 1 h (hour) to 4 h (hours); and (ii) contacting the permeate solution with the granular activated carbon at a temperature of from 55°C to 85°C or 60°C to 80°C.

[0148] Another aspect of this disclosure provides a continuous process for purifying a caramel color solution, the process comprising: (a) providing a caramel color solution comprising low molecular weight species and high molecular weight species;

This process can further comprise the step of:

[0149] In another aspect, the process disclosed immediately above can further comprise combining the first purified permeate solution, the second purified permeate solution, or both with the retentate solution to provide a purified caramel color solution. In embodiments, the permeate solution flows through the first granular activated carbon filter bed and the first purified permeate solution through the second granular activated carbon filter bed at a flow rate of filter bed at a flow rate of from 0.3 L/h to 1 L/h, a flow rate of from 0.3-0.5 L/h.

EXAMPLES

[0150] The examples of this disclosure illustrate the effectiveness of the filtration or extraction method using activated carbon extraction media for removing low molecular weight compounds in certain ingredients for food or beverages, including the compound 4-MeI (4-methylimidazole).

[0151] The activated carbon used in the examples was obtained from Chemviron, and had an average particle diameter of 1.2-1.4 mm. This activated carbon was manufactured as granular carbon from selected grades of bituminous coal by a previous agglomeration and steam-activation, and was pretreated by acid washing and neutralization.

EXAMPLE 1

Continuous flow extraction system

[0152] A simple continuous flow extraction system was constructed and used in low molecular weight compound extraction experiments to identify and develop aspects of the process. This continuous flow extraction system included a reservoir for the permeate solution (permeate reservoir) which was in fluid communication with a dosing pump, which was used to feed the permeate from the reservoir to a column containing the activated carbon at the set dosage rate. The pump used was a HANNA® Instruments positive displacement solenoid driven pump which has an adjustable flow rate to a maximum of 15.2 L/h (liters/hour) at a pressure of 1 bar.

[0153] The dosing pump collected permeate solution from the permeate reservoir and delivered it at the top of the activated carbon column using a foot valve assembly fitted with a filter. The flow rate at the column outlet was controlled by the tap at the base of the column, and the filtered fractions of the purified permeate were collected for analysis.

EXAMPLE 2

Conditions used in extraction experiments

[0154] The conditions used in extraction experiments were based on results obtained in initial testing. Liquid phase isotherms were used to show the relationship between the loading of the adsorbent (carbon) and the residual concentration of the low- molecular weight compound. The relationship was modelled by an empirical Freundlich isotherm for heterogeneous surfaces. The data was obtained by adding increasing amounts of activated carbons to different flasks containing a fixed volume of liquid permeate. A blank flask containing the liquid with no carbon was used as a control. The flasks were agitated over a period of 24 hours at constant temperature to reach equilibrium. The flask contents were then filtered and the liquid was analyzed for residual low molecular weight compound. The data from the isotherm tests can be used to determine the amount of granular carbon required (carbon consumption) to meet the treatment objective. Isotherm tests generated estimations of parameters for the continuous process based on permeate samples with 5% Brix and at pH 7. The amount of carbon used was 140 g/bed. The volume of liquid treated was about 20-25 L. at a flow rate of 4 L/h for an estimated test time of from about 5 to about 7 hours.

[0155] For a given experiment, total number of bed volumes (BVs) to treat may be about 80-85 BVs for 24-25.5 L. A bed volume is equivalent to the volume of the carbon bed, which was 300 cm3 for the lab scale column set-up implemented in this work.

[0156] Small samples can be collected every n number of BVs for analysis. Because it is expected that the filtered permeate will be free from 4-MeI during the early BVs, the sampling will be spaced at the beginning and collected more frequently once breakthrough is observed at the outlet. Breakthrough was monitored using TLC analysis and confirmed by LCMS-MS analysis. Breakthrough curves were then generated which plot the concentration of 4-MeI at the outlet with respect to that of the inlet (%) against the number of bed volumes treated. These plots helped identify the point where the performance drops significantly and the volume of permeate a carbon bed is able to process.

[0157] These conditions were used in the continuous flow experiments set out in the following examples to examine and determine the effect of contact time (linear velocity), pH, Brix %, temperature, activated carbon regeneration, and activated carbon columns (beds) in series on the extraction process.

EXAMPLE 3

Effect of contact time (linear velocity, flow rate) on extraction efficiency

[0158] The effect of contact time of the permeate within an activated carbon column on extraction efficiency was examined, in order to identify the highest flow rate that can provide sufficient extraction, in order to enable faster processing times. These experiments were conducted with a permeate solutions at Brix 3.5% having a pH of 7.

[0159] The following flow rates were tested, and the contact times of the permeate solution for each flow' rate with the activated carbon bed are shown in the table, along with the starting concentrations of the 4-MeI. Table 1. Activated carbon contact time for the flow rates tested and 4-MeI starting concentrations

[0160] The effect of contact time (linear velocity or flow rate) on the extraction of 4-MeI is illustrated in the plots of the extraction efficiency (%) versus the number of bed volumes is shown at FIG. 1 for 4-MeI extraction, for the 3.5% Brix solutions.

[0161] This study demonstrates that using a flow' rate of <1 L/h, no breakthrough was observed within the first 15 bed volumes. At higher flow rates greater than 1 L/h. the contact time was insufficient to extract 4-MeI efficiently. The earlier breakthrough (loss in extraction efficiency %) for 4-MeI was observed and, while not intending to be bound by theory, was likely due to the lower concentration of 4-MeI (ppb) in the permeate. The following data demonstrate that the residual concentrations of 4-MeI were low when operating at low flow rates, thus indicating contact times of 18-36 minutes offer good extraction performance; see Table 1. Thus, FIG. 2 plots the residual 4-MeI versus the number of bed volumes for the 3.5% Brix solutions, demonstrating that bed contact times of from 18-36 minutes (flow' rate of 0.5-1 L/h) was the minimum contact time to efficiently extract 4-MeI. Thus, reduction efficiencies of >98% w ere observed for both 4- Mel with very low residual concentrations of 4-MeI (~10 ppb) for the first 14 bed volumes at contact times of 18 mins-36 mins. Higher flow' rates (2-4 L/h) provided insufficient contact time between permeate and the carbon, leading to poorer extraction efficiency.

[0162] These experiments demonstrated that a higher contact time improves the extraction efficiency. A contact time of from 18-36 minutes was identified as minimum contact time that was useful to extract 4-MeI efficiently. Therefore, the minimum contact time of about 40 minutes w as used in the experiments.

[0163] Similar analyses of permeate solutions at Brix 1% also show ed that higher contact time (for example, >18 mins) can achieve removal of 4-MeI even at very low starting concentrations. EXAMPLE 4

Effect of activated carbon contact on the evolution of permeate eluent pH. Brix. and color

[0164] It was observed that as the permeate filters through the column, the pH, Brix, and color of the permeate were altered. These initial changes from the starting pH, Brix, and color were observed by the initial fractions of the eluent, which showed a higher pH, a lower Brix, and a lower UV-Vis absorbance as compared to the starting permeate solution. In each case, the values of pH, Brix, and absorbance of the eluents approached or tended towards the starting values in the permeate solution as more bed volumes were eluted.

[0165] The initial increase in the pH was observed, which then stabilized to a value close to the starting pH of the permeate, as shown in FIG. 3. The control value line in FIG. 3 refers to the pH of the permeate that is loaded onto the carbon bed, before it is contacted with the activated carbon. While not intending to be bound by theory', this observation is thought to result from various factors. For example some degree of anion exchange or chemical interactions with anions of weak acids or cations of weak bases may occur at the surface of the carbon with formation of salts, affecting the pH. Further, the soluble ash content of the carbons which are acid-washed such as the one tested here may affect the pH to some degree, although the relatively low ash content suggests this factor may be less important.

[0166] A larger increase in the pH of the initial fractions of the eluent was observed for permeate solutions at pH 3 as compared to pH 7. This change in the pH appears to be largely unaffected by the flow rate used, as there were no significant differences observed between flow rates for permeate at pH 3 and pH 7 as shown in FIG. 3.

[0167] Some solids were retained or adsorbed by the activated carbon bed during elution, which may be recovered upon washing the bed with water. Permeate Brix was also observed to change during the elution process. This evolution of the measured Brix of the filtered permeate suggests some initial adsorption of other soluble solids within the permeate may occur, which then stabilizes once the carbon saturates with these soluble solids. Color bodies were also retained by the carbon bed leading to a loss in color of the initial fractions of eluent. Although color bodies from the permeate are also adsorbed, in comparison to the change in Brix, adsorption of color bodies occurs in a more linear fashion as a function of bed volumes. These effects are demonstrated in the data shown in FIG. 4 and FIG. 5. The control value lines in FIG. 4 and FIG. 5 refer to the pH of the permeate that is loaded onto the carbon bed, before it is contacted with the activated carbon.

[0168] The extent of color extraction and adsorption of solids from permeate by the carbon also appears to be largely unaffected by the pH of the permeate and the flow rate used. While an increase in pH of permeate from pH 3 to 7 leads to a darkening of the permeate, the extent of color body extraction (%) is similar across both pH 7 and pH 3. Further, the evolution of Brix is similar when varying pH of permeate and contact time.

[0169] It has been discovered that pre- and post-conditioning the activated carbon bed using a so-called “sweetening on” and “sweetening off’ sequence of steps may reduce or minimize the changes in pH, loss of color bodies, and changes in Brix.

EXAMPLE 5

Effect of permeate pH on extraction efficiency

[0170] The effect of pH of the permeate solution on the low molecular weight species extraction or removal efficiency was examined. These tests were conducted on a permeate solution at Brix of 3.5%.

[0171] At a permeate pH of 3, which is close to the natural pH of the permeate, removal efficiency was reduced as compared to a permeate which w as pH adjusted to pH 7. This effect was most notable for 4-MeI removal, where removal at pH 7 exhibited a superior performance compared to pH 3. However, extraction of 4-MeI at pH 3 does improve at high contact time (> 36 mins, corresponding to <0.5 L/h). Therefore, removal at pH 3 at high contact time (for example, > 40 mins) is expected to provide a viable removal efficiency. These results are demonstrated in FIG. 6.

[0172] One advantage of performing the extraction at pH 3 is eliminating the need for additional processing steps such as neutralization and reacidification, as pH 3 is approximately the natural pH of the permeate. Therefore, removal at pH 3 w ill undergo a higher carbon consumption to achieve comparable removal efficiencies to pH 7, therefore will incur a higher operational cost.

[0173] These data also demonstrate that the residual concentrations of 4-MeI after passing through the column are lower at pH 7 compared to pH 3. At pH 3, the carbon saturates earlier compared to pH 7, indicating that the adsorption capacity of carbon may be lower at pH 3 for 4-MeI. However increasing the contact time by slowing the flow rate can significantly improve the performance at pH 3, and reduces the residual concentrations of 4-MeI significantly.

[0174] These tests indicated the extraction performance at pH 7 was superior to performance at pH 3. However, increasing the contact time of the permeate solution with the activated carbon improves the performance at pH 3 significantly. To achieve a similar performance at pH 3 as was observed at pH 7, extraction at pH 3 will likely have a higher carbon consumption.

EXAMPLE 6

Effect of Brix % on the extraction efficiency

[0175] The effect of soluble solids in the permeate on extraction efficiency was examined. The 4-MeI content was also found to vary with Brix. Permeate samples received from pilot trials have a range of % Brix values from 1% to 7%. These have been classed into 3 groups according to high Brix (about 7% Brix), mid Brix (2.7-3.5% Brix) and low Brix (0.8-1% Brix) permeate. It was found that activated carbon consumption was higher for permeate of Brix >4%, due to higher 4-MeI content. At high Brix, increased concentrations of 4-MeI in permeate also leads to a higher uptake of 4-MeI, as illustrated in the following table.

Table 2. Concentrations of 4-MeI in permeate solutions as a function of Brix

[0176] Uptake of 4-MeI was found to increase at higher Brix (~7%), likely due to the high concentrations present in permeate. However at these Brix values and uptake rates, the carbon bed will saturate earlier leading to an increased carbon consumption. Mid and low Brix samples exhibited no drop in performance for the bed volumes examined at contact time of 18 mins. However, at high Brix, drop in performance was observed significantly earlier due to high concentrations of 4-MeI and other compounds present in the permeate. These factors are observed in FIG. 7 for 4-MeI. At higher Brix, expanding the mass transfer zone for adsorption, for example by increased contact time using multiple columns in series, will improve the performance and extraction efficiency.

[0177] Therefore in general, under the experimental conditions used, high levels of extraction efficiency were observed for samples with Brix < 4%. Breakthrough of 4-MeI therefore will govern the carbon bed-life.

EXAMPLE 7

Effect of activated carbon contact on the evolution of permeate eluent pH, Brix, and color as a function of Brix

[0178] It was observed that as the permeate elutes through the column, changes from the starting pH, Brix, and color were observed by the initial fractions of the eluent, which showed a higher pH. a lower Brix. and a lower UV-Vis absorbance as compared to the starting permeate solution. In each case, the values of pH, Brix, and absorbance of the eluents approached or tended toward the starting values in the permeate solution as more bed volumes were eluted.

[0179] While not intending to be bound by theory, the changes in the pH and Brix of the permeate are thought to be due to adsorption of sugars and salts from the early bed volumes of the eluent. As described herein, pre-conditioning the carbon bed can reduce these changes in both pH and %Brix. These effects were discovered to vary as a function of the Brix of the starting permeate, as illustrated in FIG. 8. In FIG. 8, the starting pH values of the eluent prior to elution were as follows: Brix 0.8-1%, pH 7.4; Brix 2.7-3.5%, pH 7.07; and Brix about 7%, pH 7.14.

[0180] As show n, an initial change in pH of the eluent from the starting pH of the permeate was significant, and the pH of the eluent approached the starting values in the permeate solution as more bed volumes were eluted. A lower extent of change in the initial pH of the eluent was observed for the high and mid Brix permeates as compared with the low Brix permeate, likely due to the higher concentration of salts and buffers present in high and mid Brix permeate which moderate this change, as compared with the low Brix permeate. All samples exhibited this decrease in pH with increasing bed volumes, and the lower the Brix, the higher the initial pH and the more gradual the decrease in pH, possibly due to lower buffer concentrations present in the lower Brix samples. Pre-treatment steps can be applied to pre-condition the activated carbon bed to reduce or minimize these changes in pH and solid retention as described herein.

[0181] Regarding the initial changes in the Brix of the permeate observed as sugars and salts from early bed volumes are adsorbed by the carbon, it is thought that the activated carbon becomes saturated by adsorption of sugars and salts from the permeate over about 5-10 bed volumes, after which the Brix of the eluent returns to that of the input value of the permeate. FIG. 9 demonstrates the evolution of Brix as a function of the starting Brix of the permeate.

[0182] Each permeate of low, mid, and high Brix were tested at pH 7. The Brix percent for the samples as received and the Brix percent of the control (Brix after neutralization to pH 7), are shown in the following Table.

Table 3. Permeates used to test the evolution of Brix as a function of starting permeate Brix

[0183] Each of the samples exhibited an initial decrease in Brix from the starting permeate followed by an increase in Brix with each bed volume. A sharper increase over a wider range of Brix percent w as observed for permeate with a higher initial Brix % value as the carbon bed saturates earlier.

[0184] In contrast, the extraction or adsorption of color bodies was observed to vary more linearly with bed volumes, as opposed to the observed asymptotic change in pH and Bnx. As illustrated in FIG. all samples exhibited an increase in absorption as a greater volume of the permeate was passed through the column, and the higher concentrations of color bodies in the higher Brix permeate produced a steeper increase in the absorption with the number of bed volumes. For FIG. 10, the control (unfiltered) absorbance level for each ofthe Brix levels (at 500 nm) were as follows: Brix 0.8-1%, 0.327; Bnx 2.7-3.5%, 0.501; and Bnx about 7%, 1.035. EXAMPLE 8

Effect of Temperature on the extraction process

[0185] Bacterial growth within the activated carbon bed can be an issue, and microbial build up can lead to problems within activated carbon units over days of use. Extraction was tested at temperatures of about 20 °C ('‘room temperature’’), 60 °C, and 80 °C. When testing permeate solutions at Brix 3.5% having a pH of 7 at these different temperatures, essentially no losses of extraction efficiency for 4-MeI were observed over about 18 bed volumes, when tested at a flow rate of 1 L/h, for a residence time of 18 min.

[0186] In these tests, elevated temperatures (above about 60 °C) may mitigate bacterial growth, therefore using temperatures from about 60 °C to 80 °C can be useful to mitigate microbial growth in the carbon bed during extraction. No adverse impact of elevated temperatures (60-80 °C) on the extraction of 4-MeI was observed, therefore, continuous extraction can operate at temperatures of from about 60 °C to 80 °C.

EXAMPLE 9

Activated carbon regeneration tests

[0187] An initial evaluation of whether regeneration of spent activated carbon could be used was carried out. Regenerating spent carbon would allow recycling of carbon beds to minimize waste. These tests were conducted as follow s. First, spent carbon was washed with water for tw o bed volumes to release trapped sugars, followed by a dynamic solvent wash for one bed volume using either ethyl acetate or acetic acid, followed by water washing for three bed volumes. The regenerated carbon beds were then tested to compare performance for 4-Mel extraction against virgin carbon.

[0188] The data in the following table compares the pre-regeneration extraction efficiency of a used carbon bed, the post-regeneration extraction efficiency of the bed when washed with either ethyl acetate or acetic acid, and compares these to the efficiency of a virgin bed for 4-MeI extraction.

Table 4. Regenerated carbon efficiency versus virgin carbon efficiency for 4-MeI extraction.

[0189] These initial tests show that regenerated carbon is able to extract 4- Mel, however performance is lower compared to virgin carbon. Therefore, there is potential for regeneration of carbon using solvent extraction, and using regenerated carbon alone or in combination with virgin carbon for extraction.

EXAMPLE 10

Use of two activated carbon columns (beds) in series

[0190] Using two columns in series was investigated, as double columns can be used for applications that require frequent adsorbent replacement or have a high consumption rate of carbon.

[0191] Two columns were connected in series and the breakthrough curve of each column was recorded against the total bed volumes treated. Test conditions included a permeate flow rate of 0.3-0.5 L/h (residence time of 36-40 mins) for a 3-4% Brix permeate at a temperature of 60 °C. These tests did not saturate the carbon beds when operating at the lab scale under these conditions identified. Extraction efficiency remained high (>99%) for 4-MeI for up to 50 bed volumes tested.

[0192] Studies at shorter contact times have shown that 4-MeI extraction performance drops off. Therefore, using columns in series can work well for 4-MeI removal when operating at high Brix (>5%).

EXAMPLE 11

Pre-conditioning and post-conditioning the activated carbon columns

[0193] Based on the experimental data described above which demonstrate that during the early bed volumes of the extraction, loss of solids (change in %Brix) and a change in the pH (increase) occurred, the following pre-conditioning and post-conditioning processes of the carbon bed can be developed as follows. These pre-conditioning and postconditioning processes are referred to as “sweetening on” and “sweetening off’ steps.

[0194] An initial cold water wash is conducted by passing cold water up through the column in a reverse flow mode, so that the cold water enters the column from the bottom and exits the top. This reverse flow cold water flush displaces residual fines and particulates effectively, as cold water is more viscous than hot water. This cold water wash is followed by a wash with hot water in reverse flow, which warms the bed and minimizes crystallization of the sugar.

[0195] The pre-conditioning “sweetening on” step is then conducted by using a solution of similar composition to the permeate in Brix and pH, but without low molecular weight species (e.g., 4-MeI). This sweetening solution is passed through the column from the top of the column at a rate of 2 L/h until the %Brix and the pH at the outlet is the same as that of the inlet. The column is now pre-conditioned, and the continuous extraction of the permeate is then conducted until low molecular weight compound breakthrough is observed.

[0196] The post-conditioning “sweetening off step is then conducted at the end of the carbon bed usage as follows. A reverse-flow hot water flush is conducted for 2- 3 bed volumes to remove sugar which has adsorbed to the carbon bed. If the bed cools down prior to completing the sweetening off process, it can induce solidification w hich may make it difficult to remove salts and regenerate the carbon. It is expected that 90-100% recovery of solids is obtained from the sweetening off solution, thus, the sweetening off fraction can be recycled to condition new carbon beds during a sweetening on process.

EXAMPLE 12

Changes in permeate composition during extraction

[0197] Changes in the composition of the permeate occurred as fractions were collected from the ultrafiltration process, including a gradual decline in the Brix and 4-MeI concentrations with later collected fractions. The changes in the composition of the permeate during extraction have been found to be detrimental to the sustainable operation and longevity of the carbon bed. For example, the larger levels of 4-MeI in the initial fractions of the permeate can lead to earlier breakthrough and diminished extraction performance for mid and low Brix permeates, and desorption of adsorbed compounds (4- Mel) as Brix progresses to mid and low.

[0198] To avoid these issues, a feed tank can be used which can collect multiple fractions of the permeate as it is generated from the ultrafiltration process. This feed tank can then be used to supply the permeate to the carbon bed. This process ensures a relatively constant composition of the permeate supplied to the carbon bed and allows the carbon bed to perform in an efficient and sustainable way. This also maintains the longevity of the carbon bed which is impacted by significant changes in composition (Brix and 4-MeI concentrations etc.). A collection tank may also be used post extraction, which can provide a more consistent quality of permeate.

Claims

CLAIMS What is claimed is:

1. A process for purifying a caramel color solution, the process comprising:

2. The process according to claim 1, wherein the granular activated carbon is contained in a filter bed through which the permeate solution flows at a flow rate.

3. The process according to claim 1 or claim 2, wherein the contact time is from 15 min (minutes) to 2 h (hours), from 20 min to 1.5 h, or from 25 min to 1 h.

4. The process according to claim 2 or claim 3, wherein the permeate solution flow s through the filter bed at a flow rate of from 0. 1 L/h to 1.2 L/h, or a flow rate of 0.3 L/h to 1 L/h.

5. The process according to any preceding claim, further comprising combining the purified permeate solution with the retentate solution to provide a purified caramel color solution.

6. The process according to any preceding claim, wherein the low molecular weight species comprise 4-methylimidazole (4-MeI), 2-methylimidazole (2-MeI). 5- hydroxy-2-methylpyridine, 2-hydroxypyridine, 2-hydroxy-6-methylpyridine carboxylic acid, or a combination thereof.

7. The process according to any preceding claim, wherein the low molecular weight species comprise 4-MeI (4-methylimidazole).

8. The process according to any preceding claim, wherein the caramel color is selected from a class III caramel color or a class IV caramel color.

9. The process according to any preceding claim, wherein a molecular weight cut off of the ultrafiltration process is about 800 Da.

10. The process according to any preceding claim, wherein the permeate solution from the ultrafiltration step has a pH of about pH 3±0.5 or about pH 3±0.25.

11. The process according to any preceding claim, wherein the pH of the permeate solution is adjusted to a pH of about 6.5-7.5 or a pH of about 7 prior to contacting the permeate solution with a granular activated carbon.

12. The process according to any preceding claim, wherein the pH of the permeate has a Brix of 3-4% and the pH of the permeate solution is adjusted to a pH of about 6.5- 7.5 prior to contacting the permeate solution with a granular activated carbon.

13. The process according to any preceding claim, wherein contacting the permeate solution with the granular activated carbon occurs at a temperature of from 55°C to 85°C or from 60°C to 80°C.

14. The process according to any preceding claim, wherein the Brix % of the permeate solution prior to contacting the permeate solution with a granular activated carbon is from about 1% to about 7%.

15. The process according to any preceding claim, wherein the low molecular weight species comprise 4-MeI and 90% or more of the 4-MeI concentration in the permeate solution is removed upon contacting the permeate solution with granular activated carbon to provide the purified permeate solution.

16. The process according to any preceding claim, wherein the purified permeate solution comprises 4-MeI in a concentration of 50 ppb (parts per billion, by weight).

17. The process according to any of claims 1-2, wherein:

(i) the permeate has a Brix of 3-4% and the pH of the permeate is adjusted to a pH of6.5-7.5;

(ii) the contact time is from 15 min (minutes) to 3 h (hours), from 30 min to 2 h, or from 40 min to 1.5 h; and

(ii) contacting the permeate solution with the granular activated carbon occurs at a temperature of from 55°C to 85°C or from 60°C to 80°C.

18. The process according to any of claims 1-2, wherein:

(i) the permeate has a Bnx of 5-7% and a pH of 2.5-3.5;

(ii) the contact time is from 1 h (hour) to 4 h (hours); and

(ii) contacting the permeate solution with the granular activated carbon occurs at a temperature of from 55°C to 85°C or 60°C to 80°C.

19. A continuous process for purifying a caramel color solution, the process comprising:

20. The process according to claim 19, further comprising:

21. The process according to any of claims 19-20, further comprising combining the first purified permeate solution, the second purified permeate solution, or both with the retentate solution to provide a purified caramel color solution.

22. The process according to any of claims 19-21, wherein the low molecular weight species comprise 4-MeI (4-methylimidazole).

23. The process according to any of claims 19-22, the permeate solution flows through the first granular activated carbon filter bed and the first purified permeate solution through the second granular activated carbon filter bed at a flow rate of filter bed at a flow rate of from 0.3 L/h to 1 L/h, a flow rate of from 0.3-0.5 L/h.