PROCESS FOR PRODUCING 2,5-FURANDICARBOXYLIC ACID FROM 5-HYDROXYMETHYLFURFURAL ETHERS
Patent Information
- Authority / Receiving Office
- MX · MX
- Patent Type
- Patents
- Current Assignee / Owner
- FURANIX TECH BV
- Filing Date
- 2022-06-17
- Publication Date
- 2026-05-19
AI Technical Summary
Existing oxidation processes for producing 2,5-furandicarboxylic acid from 5-alkoxymethylfurfural face challenges such as difficulty in maintaining process stability, high metal incorporation into the product cake, and sensitivity to process parameters, leading to unwanted by-products and reduced yields, especially on an industrial scale.
A process involving the use of a catalyst system comprising cobalt, manganese, and bromine in a saturated organic acid solvent, with controlled addition of hydrobromic acid or dicarboxylic acids to manage metal incorporation and maintain process stability, by monitoring and adjusting the ratio of manganese to cobalt in the cake and optimizing reaction conditions.
The process ensures reliable operation with reduced metal incorporation into the product cake, maintaining high yields and purity, even on an industrial scale, by implementing precise control measures to counteract metal enrichment and stabilize the reaction.
Abstract
Description
PROCESS FOR PRODUCING 2,5-FURANDICARBOXYLIC ACID FROM 5-HYDROXYMETHYLFURFURAL ETHERS TECHNICAL FIELD The present invention relates to a process for producing a carboxylic acid composition comprising 2,5-furandicarboxylic acid, specifically a process for producing a carboxylic acid composition comprising 2,5-furandicarboxylic acid using 5-alkoxymethylfurfural as a starting material. BACKGROUND OF THE INVENTION 2,5-Furandicarboxylic acid (FDCA) is known in the industry as a very promising building block for replacing petroleum-based monomers in the production of high-performance polymers. In recent years, FDCA and the novel plant-based polyester polyethylene furanoate (PEF), a fully recyclable plastic with superior performance properties compared to currently widely used petroleum-based plastics, have attracted considerable attention. These materials could make a significant contribution to reducing dependence on petroleum-based polymers and plastics, while also enabling more sustainable management of global resources.Consequently, comprehensive field research was conducted to arrive at a technology to produce FDCA and PEF in a commercially viable manner, in order to enable the successful commercialization of these promising materials. FDCA is typically obtained as a crude carboxylic acid composition by the oxidation of molecules containing furan fractions, for example, 5-hydroxymethylfurfural (5-HMF), as well as the corresponding 5-HMF esters or 5-HMF ethers, for example, 5-alkoxymethylfurfural, and similar starting materials, which are usually obtained from plant-based sugars, for example, by sugar dehydration. A wide variety of prior art oxidation processes are known, including, for example, enzymatic or metal-catalyzed processes. One of the most established techniques in the field uses a catalyst system comprising cobalt, manganese, and bromine to oxidize compounds containing a furan moiety to FDCA using oxygen or air as an oxidizing agent. The respective processes, which are applicable to a wide variety of starting materials, are disclosed, for example, in WO 2014 / 014981 Al or WO 2011 / 043660 Al. MA / a / ZUZZ / UU 1040 Since the purity achievable for crude carboxylic acid compositions in the processes mentioned above is often insufficient to reach the level of purity required for the polymerization of FDCA to PEF or other high-performance polymers, purification processes have been developed to further purify crude carboxylic acid compositions to produce a purified carboxylic acid composition. These processes include, for example, hydrogenation steps, post-oxidation steps, distillation steps, recrystallization steps, or similar methods, often combined with comprehensive purification schemes involving several washing and isolation steps of the resulting carboxylic acid composition. Examples of purification processes are disclosed, for example, in WO 2014 / 014981 Al or WO 2016 / 195499 Al. In recent years, it has been discovered that one of the most promising approaches to obtaining FDCA in an economically viable manner employs a significant amount of 5-HMF ethers, such as 5-alkoxymethylfurfural, as the starting material for oxidation. As a result, the crude carboxylic acid composition obtained in such processes not only comprises the free diacid, i.e., FDCA, but also includes a significant amount of the monoalkyl ester of FDCA. Currently, processes employing 5-methoxymethylfurfural as the starting material for oxidation, which yield significant amounts of the monomethyl ester of FDCA (FDCA-Me), appear to be the most established. While some prior art documents are eager to report high yields and good purities for their stated oxidation processes, less attention is often paid to the fact that the underlying reactions are, in most cases, very difficult to carry out in practice and / or are quite sensitive to external influences. This is particularly true for batch experiments with long residence times or (semi-)continuous processes that need to be operated (preferably at steady state) for an extended period. In particular, these difficulties are very serious if several subsequent process steps need to be linked together to reach the desired product, since a small deviation in one process step can potentially multiply its negative effect on subsequent reactions. Furthermore, most prior art documents describe only laboratory-scale experiments. However, producing a new compound in a commercially viable manner requires large-scale reactors, which makes maintaining reactions even more challenging. In a real industrial-scale plant, gradients in process parameters such as temperature and concentration, variations in compound mass flow rates, or other influences can result in a process that stops completely or produces an unwanted product. For example, using recycle streams to increase process efficiency or economics can lead to the accumulation of materials, whether desirable or undesirable. Unfortunately, while the oxidation process that begins with 5-alkoxymethylfurfural as the starting material has several advantages—for example, efficient sugar dehydration and product recovery to produce 5-alkoxymethylfurfural, high yields, and good product purity—compared to comparable prior art processes that do not produce FDCA monoalkyl esters, such processes have sometimes proven particularly difficult to control. In establishing the technology, it was discovered that it can be challenging to keep the processes running (often referred to as "live" or "alive" by those skilled in the art) for extended periods.For various sets of process parameters, several of which are reported to be preferred in the prior art based on laboratory-scale experiments, the process tends to stall after a while (the practitioner often refers to these processes as dead or moribund), and sometimes it is not even possible to start the process in the first place. While a dye-oxidation process to produce FDCA can manifest itself in various ways, it is often observed that the color of the resulting product, in most cases obtained as a solid cake, changes from white to yellow and then to brown, and the FDCA yields decrease significantly as more and more unwanted byproducts are produced. Consequently, the cake color on the white-to-brown scale is a good qualitative indicator of whether the process has gone out of the desired regime and is running in an undesirable state or is even in the process of dying out completely.Additionally, the stopping of the reaction can usually be evidenced by a rapid increase in the oxygen content of the reactor outlet gas stream, and a reduction in the production of CO2 and CO. In addition to the problems described above for oxidation processes starting with 5-alkoxymethylfurfural, it is also unfortunately observed that these processes are more likely than prior art processes to suffer from the incorporation of metal catalysts into the product cake. This not only contaminates the product but also removes valuable catalysts from the system that could otherwise be reused or recycled. The incorporation of catalyst metals into the product cake appears to be particularly problematic with respect to the oxidation of furan-containing portions to form 2,5-furandicarboxylic acid. Although this problem has not been previously reported, to the inventors' knowledge, it is particularly pronounced with respect to the incorporation of manganese into the cake. In such cases, the product cake will be particularly enriched in manganese, relative to cobalt, compared to the catalyst feed. The inventors believe that this effect is different and distinct from the over-oxidation of manganese(II) to manganese(IV) sometimes reported in the Co / Mn / Br literature for the production of organic acids via oxidation, in which manganese is oxidized to form Mn(IV)Oz, which subsequently precipitates from the solution as black specks in the product cake.The observations made and reported herein show a phenomenon in which a pink color appears in the cake, associated with an excess of manganese in the cake. If we wish to be limited by any theory, it is believed that a relatively insoluble complex forms between 2,5-furandicarboxylic acid and manganese, possibly involving the doubly ionized form of 2,5-FDCA and Mn(II). The two effects described above—the challenge of maintaining the process and the unfavorable tendency for metal to be incorporated from the catalyst system into the cake—appear to be two separate effects observed in oxidation processes that begin with 5-alkoxymethylfurfural. For example, metal incorporation from the catalyst system is also observed in active processes and is qualitatively manifested by the pink color of the cake. However, the two effects could potentially have a similar or at least related origin. In any case, it is believed that the removal of the catalyst from the process in the cake—that is, the precipitate—will likely contribute to the difficulty of maintaining the process, since the process requires the metal catalyst. DETAILED DESCRIPTION OF THE INVENTION In view of the problems described above, a primary objective was to overcome the disadvantages of prior art oxidation processes using 5-alkoxymethylfurfural as starting materials, while maintaining the general benefits of the basic process. In particular, there was a need for a process to produce 2,5-furandicarboxylic acid from 5-alkoxymethylfurfural that could be reliably initiated and operated for extended periods without going out of the acceptable range, or at least with a lower probability of doing so, even at an industrial scale. Furthermore, there was a need for a process to produce 2,5-furandicarboxylic acid from 5-alkoxymethylfurfural that would reduce the problem of metal incorporation into the product cake.In particular, an objective was to provide a process for producing 2,5-furandicarboxylic acid that allows for adjustments to the process in the event of high metal incorporation in the cake, where it would be desirable to require only minor adjustments, preferably to parameters that can be precisely controlled and adjusted quickly. It would be particularly desirable if the process for producing 2,5-furandicarboxylic acid could achieve the respective benefits through sophisticated process control without the need for additional substances or devices. Another objective was to provide a process for producing 2,5-furandicarboxylic acid from 5-alkoxymethylfurfural that employs an optimized catalyst system that increases the robustness of the respective processes and reduces the tendency for metal incorporation into the cake, while maintaining the beneficial properties reported for such processes. Another objective was to provide a process for producing 2,5 furandicarboxylic acid that employs acetic acid, or acetic acid with smaller amounts of water, as the main washing fluid for the raw cake. Without wishing to be bound to any particular theory, the presence of 2,5-furandicarboxylic acid monoalkyl ester in the oxidation reactor appears to be responsible for some of the beneficial effects typically associated with the respective prior art technology. Therefore, a minimum amount of 2,5-furandicarboxylic acid monoalkyl ester in the oxidation reactor was found to be desirable. In particular, a certain amount of 2,5-furandicarboxylic acid monoalkyl ester in the feed to the oxidation reactor appears to reduce the tendency for manganese to appear in the product cake. However, the monoalkyl ester of 2,5-furandicarboxylic acid was also found to be the cause of some of the problems associated with the respective process described above. In particular, it appears that both the robustness of the process and the ability to wash metals from the cake are negatively affected if the concentration of the monoalkyl ester of 2,5-furandicarboxylic acid exceeds a maximum value. The solubility of the monoalkyl ester of 2,5-furandicarboxylic acid, for example, monomethyl-2,5-furandicabroxylate, in acetic acid-rich systems is much higher than that of 2,5-furandicarboxylic acid. While FDCA will crystallize extensively, leaving only a small residue in the solution, FDCA-Me will tend to remain in solution with only a portion co-crystallizing in the product cake. As a result, FDCA-Me will be retained in the mother liquor and will tend to accumulate in the system.If the level rises sufficiently, it will exceed the solubility limit at the temperature used for product isolation, and a second phase, primarily FDCA-Me, will crystallize out of solution. This precipitate has been found to be particularly difficult to filter, consisting of a waxy, spongy particle, which increases filtration time and also hinders cake washing and can lead to increased metal retention (in a proportion similar to the catalyst metal feed). While not wishing to be subject to any theory, it is believed that the amount of monoalkyl ester of 2,5-furandicarboxylic acid must be kept within a specific range in which the most convenient reference system for concentration was found to be the mother liquor, i.e., the liquid that is obtained from the reaction mixture and the composition of crude carboxylic acid after the separation of the FDCA in a solid-liquid separation zone, since the mother liquor allows information to be obtained about the reaction medium in the oxidation reactor. Accordingly, in the process for producing a carboxylic acid composition comprising 2,5-furandicarboxylic acid of the present process, the mother liquor comprises ester ML / a / ZUZZ / UU 1040 monoalkyl of 2,5-furandicarboxylic acid in the range of 0.5 to 7% by weight with respect to the weight of the mother liquor. However, we found that this limitation alone was not enough to completely eliminate the problem of metal incorporation from the catalyst into the cake. Fortunately, the inventors found a solution to control the process so that metal incorporation into the cake could be counteracted at an early stage without needing to stop the ongoing process. In the present process, the amount of manganese and / or cobalt in the cake is determined, and only if the determined amount exceeds a predefined threshold value is an additional process step performed to adjust the process. The tendency for manganese to be incorporated into the cake, in a higher proportion than that of cobalt, is particularly pronounced. A useful measure of this tendency is the ratio of manganese to cobalt in the cake, divided by the ratio of manganese to cobalt in the catalyst feed. When this ratio is approximately 1.0, the metals in the dry cake exactly reflect those in the catalyst system. When this ratio is significantly higher than 1, say 2 or greater, then manganese is preferentially sequestered in the cake, and corrective control action, as described herein, must be undertaken. Surprisingly, we discovered that the incorporation of metals into the cake can be counteracted by increasing the amount of one or more control acids in the oxidation reactor, where the complete experiments revealed that one or more control acids must be selected from the group consisting of hydrobromic acid and mono- or dicarboxylic acids having 2 to 5 carbon atoms and a pKa less than 3.2 for the process to operate correctly. Additionally, we discovered that for an oxidation process that produces monoalkyl ester of 2,5-furandicarboxylic acid, specific temperatures are required to allow reasonable operation, where a specific catalyst system was identified that has proven to be particularly robust and in itself reduces the tendency to incorporate metals into the cake even in the presence of larger amounts of monoalkyl ester of 2,5-furandicarboxylic acid. The invention relates to a process for producing a carboxylic acid composition comprising 2,5-furandicarboxylic acid, comprising the steps of: a) oxidizing an oxidizable compound comprising 5-alkoxymethylfurfural in an oxidation reactor in the presence of a saturated organic acid solvent having 2 to 6 carbon atoms and a catalyst system comprising cobalt, manganese and bromine using an oxidizing gas at a temperature in the range of 160 to 210°C to obtain a crude carboxylic acid composition comprising 2,5-furandicarboxylic acid monoalkyl ester and solid 2,5-furandicarboxylic acid, b) isolating at least a portion of the solid 2,5-furandicarboxylic acid from the crude carboxylic acid composition in a solid-liquid separation zone to generate a solid cake and a mother liquor, c) determine the amount of manganese and / or cobalt in the cake, and d) increasing the amount of one or more control acids in the oxidation reactor, if the determined amount of manganese and / or cobalt in the cake exceeds a predefined threshold value, wherein the one or more control acids are selected from the group consisting of hydrobromic acid and mono- or dicarboxylic acids having 2 to 5 carbon atoms and a pKa of less than 3.2, and wherein the mother liquor comprises monoalkyl ester of 2,5-furandicarboxylic acid in the range of 0.5 to 7% by weight relative to the weight of the mother liquor. The present process overcomes the disadvantages of prior art oxidation processes while using 5-HMF ethers as starting materials and maintaining the general benefits associated with this technology, such as high yields, good product purity, and the availability of inexpensive raw materials. The process of the present invention can be reliably started and operated for extended periods without compromising acceptable product quality, even at an industrial scale. The process of the present invention addresses the problem of metal incorporation into the product cake. As soon as the amount of metal in the cake exceeds a predefined threshold, a suitable countermeasure is defined that allows the process to be directly adjusted to produce a product cake with a reduced amount of metal.In the process according to the invention, only comparatively minor intervention is required for process control, where the addition of control acid can be precisely controlled and the intensity of the intervention, if required, can be quickly adjusted. Methods for carrying out the invention Step a) of the present process corresponds to a typical oxidation reaction for obtaining FDCA, where the temperature was determined to be within the range that proved particularly beneficial for the production of FDCA from the starting material and particularly suitable for enabling process control by regulating the acid. Furthermore, it was found that the temperature ensures the formation of sufficient monoalkyl ester of 2,5-furandicarboxylic acid. Additionally, the oxidizable compound oxidized in step a) as a starting material is defined as 5-alkoxymethylfurfural, i.e., an ether of 5-hydroxymethylfurfural. Pre-existing art also has many instances where temperatures of ML / a / ZUZZ / UU 1040 lower oxidation. However, the current process should be preferred, as the higher temperatures allow the oxidation reactor to operate at elevated pressure while simultaneously allowing the large amount of heat generated by the oxidation reaction to be removed by vaporization. This is known to those skilled in the art as adiabatic operation, which is a way of saying that the heat of reaction is not being removed by external sources such as coolers, loss through the walls, and the like. In general, higher temperature requires higher operating pressure for adiabatic operation. Higher pressure, in turn, allows for a higher partial pressure of oxygen (at a prescribed oxygen volume percentage) in the reactor and reduces the risk of oxygen deficiency.The volume percentage of oxygen in the exhaust gas is generally limited for safety reasons to be below the lower explosive limit, for example, to a level of 10% by volume, or more preferably below approximately 6% by volume, to allow for a safety margin. This results in the formation of a crude carboxylic acid composition comprising a monoalkyl ester of 2,5-furandicarboxylic acid and 2,5-furandicarboxylic acid. In the process according to the invention, the catalyst system for the oxidation comprises cobalt, manganese, and bromine, wherein these compounds are preferably provided as cobalt acetate, manganese acetate, and hydrobromic acid, with hydrobromic acid being particularly preferred. The oxidation reactor can be any typical oxidation reactor known in the art. The saturated organic acid solvent used in the reaction has 2 to 6 carbon atoms, with acetic acid being particularly preferred. In step b) at least a portion of the solid 2,5-furanedicarboxylic acid is isolated, i.e., separated from the crude carboxylic acid composition, wherein the isolation is carried out in a solid-liquid separation zone. Within the framework of the present invention, the term at least a portion preferably means at least 10% by weight with respect to the weight of the crude carboxylic acid composition, more preferably at least 50% by weight, more preferably at least 80% by weight. In the solid-liquid separation zone, a solid cake and a mother liquor are generated. In step c) of the process of the present invention, the amount of manganese and / or cobalt in the cake is determined. Preferably, the amount of manganese is determined. It is understood that the cake comprising solid FDCA may be moistened due to the presence of residual mother liquor. However, a person skilled in the art intending to determine the amount of a compound in the cake shall sufficiently dry the respective cake to ensure confidence in the measurement result, or shall adjust or correct the measurement and its result for the residual amount of mother liquor, respectively. Preferably, the cake in step c) comprises more than 90% solids, more preferably more than 95% solids, more preferably more than 99% solids, by weight of the cake, the latter also being referred to as dry cake. The moisture content of the cake can be determined using any of several techniques known in the art, for example, by weight loss under controlled heating conditions, and the results of the metal determination are reported on a moisture-free basis. The amount of manganese and / or cobalt in the cake can be determined using any suitable measurement technique, provided the respective techniques are well known to the expert, for example, properly calibrated X-ray fluorescence (XRF) or inductively coupled plasma (ICP). Aside from chemical analysis, spectroscopic and optical measurement methods are particularly preferred. In the most basic case, the amount of manganese in the cake is determined by optical inspection of the cake by the process operator, where the color of the resulting cake is evaluated against the intensity of the pink color, which is now known to be typically associated with manganese in the cake during the oxidation of furfural-related compounds to form FDCA. If the determined amount of manganese and / or cobalt, preferably manganese, exceeds a predefined threshold value, for example, because the cake was found to be too pink, the amount of one or more control acids in the oxidation reactor tends to increase. The term control acids is arbitrarily chosen to clearly denote the group of specific acids that were found to be suitable in the present process. The appropriate threshold values are defined by the expert in the technique based on the individual characteristics of the process and the amount of metals that are considered acceptable in the cake for subsequent processing steps and / or for other applications. We have found that, in addition to the total concentration of catalyst metals in the cake, which can also be influenced by the retention of the catalyst-rich mother liquor, the manganese enrichment, relative to cobalt, in the cake relative to the feed catalyst is also a suitable indicator. This ratio is particularly useful since the applicants have discovered that in the oxidation of furfural-related compounds to form FDCA with a catalyst system comprising cobalt, manganese, and bromine, manganese is the most sensitive indicator of a problem and is observed at unusually high levels in the cake. A manganese enrichment factor has been developed, which is defined according to the following equation: (Mn / Co)cake / (Mn / Co)catalyst This equation has the advantage of not being affected by absolute catalyst levels or the presence of, for example, washing liquids, while still reflecting undesired manganese enrichment in the cake. A value of 1 or less does not indicate preferential manganese enrichment in the cake. For practical reasons, when absolute manganese contents are low, for example, less than approximately 10 ppm in the cake, the value can fluctuate as high as approximately 1.5. Values above this level, or for example, above 2.5, indicate undesired manganese incorporation into the cake and require corrective action. Consequently, a process is preferred where the weight ratio of manganese to cobalt in the cake to the weight ratio of manganese to cobalt in the catalyst system is less than 2.5, preferably less than 2, and more preferably less than 1.5. As defined above, one or more control acids are selected from the group consisting of hydrobromic acid and mono- or dicarboxylic acids having 2 to 5 carbon atoms and a pKa less than 3.2. Hydrobromic acid is specifically preferred and was among the first control acids identified by the inventors, as it is often available at the plant as part of the catalyst system typically employed in step a). However, additional organic acids were tested, and the inventors found that only those acids with a certain acidity can function as control acids. It was found that the pKa of suitable control acids should be less than 3.2, where pKa is measured in water. Furthermore, screening experiments confirmed that only monocarboxylic or dicarboxylic acids can serve as control acids, as higher polycarboxylic acids can cause a loss of activity, possibly due to complex formation. For example, trimellitic acid and pyromellitic acid are relatively strong aromatic polycarboxylic acids with initial pKa values of 2.52 and 1.92, respectively. However, these acids can cause a loss of activity in certain oxidations and are not suitable as control acids. In the present work, FDCA, a relatively strong aromatic dicarboxylic acid, was found to also cause a loss of activity or prevent the system from starting. This effect is not observed, to our knowledge, in the oxidation of p-xylene to terephthalic acid.If they wish to be bound by theory, the applicants speculate that FDCA's greater solubility and higher acidity are at least partially responsible for this effect. Oxidation at higher temperatures amplifies this problem due to the significantly greater solubility at higher temperatures and the resulting availability of the diacid to form complexes with the catalyst components. FDCA monomethyl ester, a monocarboxylic acid with the required pKa but with 7 carbon atoms, can provide the desired effect when added as a control acid. However, FDCA monoester tends to accumulate in a recycling operation and has been found to have a detrimental effect on oxidation at high levels, leading to slower and more intermediate reaction rates, browning, or even a failed reaction. Additionally, the FDCA monoester can cause filtration difficulties and make it difficult to remove the mother liquor from the cake, leading to higher overall metal levels. However, surprisingly, the monocarboxylic acid FCA, i.e., 2-furancarboxylic acid, was found to be a very suitable control acid for the process of the present invention. However, when using FCA, the purification system used after oxidation must be capable of removing it from the FDCA composition if the FDCA is to be used in the polymerization. Consequently, the inventors deduced that control acids can be defined as hydrobromic acid and mono- or dicarboxylic acids having 2 to 5 carbon atoms and a pKa less than 3.2.Within the scope of the present invention, the FDCA monoester, which also functions as a controlling acid, is considered separately due to the special problems associated with its accumulation and is therefore limited to a range of use. In step d), the amount of one or more control acids is increased by the deliberate and intentional addition of one or more control acids to the oxidation reactor. Any in situ formation of a control acid that could potentially occur in the oxidation reactor is not considered a deliberate and intentional addition of one or more control acids and does not correspond to increasing the amount of one or more control acids in the oxidation reactor. Since hydrobromic acid is often used to provide bromine ions to the catalyst of the oxidation reaction, it is convenient to discuss what constitutes an increase in hydrobromic acid within the scope of the present invention.Most (semi-)continuous processes add additional catalyst, often including additional hydrobromic acid, to compensate for bromine loss during the reaction—for example, loss in the overhead or mother liquor—in order to maintain desired catalyst concentrations. Adding hydrobromic acid to the oxidation reactor to maintain the desired concentration does not constitute an increase in the amount of hydrobromic acid within the meaning of the present invention. In other words, step d) would require changing the catalyst composition in the reactor by increasing the bromine-to-metal ratio. Accordingly, step d) of the process of the present invention could be: d) increasing the amount of one or more control acids in the oxidation reactor or increasing the bromine-to-metal ratio in the catalyst system, preferably to weight ratios greater than 2, even more preferably to weight ratios greater than 2.5, if the determined amount of manganese and / or cobalt in the cake exceeds a predefined threshold value, wherein one or more control acids are selected from the group consisting of mono- or dicarboxylic acids having 2 to 5 carbon atoms and a pKa less than 3.2. In other words, if the cake using the existing catalyst system is found to have a high metal content, especially manganese, and the catalyst composition is then changed to increase the bromine level by adding HBr, such a change would constitute an increase in the amount of a control acid within the meaning of the present invention. In stark contrast, increasing bromine through, for example, NaBr or NhLBr would neither constitute nor increase the amount of a control acid, as these are not control acids within the meaning of this invention. The person skilled in the art is familiar with several suitable methods for adjusting the concentration of 2,5-furandicarboxylic acid monoalkyl ester in the oxidation reactor and, correspondingly, in the mother liquor to the level defined above. From among the options known in the art, the person skilled in the art selects an appropriate option based on their general knowledge. For example, the person skilled in the art may increase the amount of solvent or other starting materials to dilute the solution in the oxidation reactor, or, if mother liquor recycling is used, the person skilled in the art may deliberately remove the 2,5-furandicarboxylic acid monoalkyl ester from the mother liquor to alter the ester concentration in the oxidation reactor and in the fresh mother liquor. A portion of the mother liquor may also be removed, or purged, from the system to reduce the concentration of monoester and / or control acids.If a portion of the mother liquor is purged, the acetic acid can still be recovered, for example, by distillation, and the residue can be discarded or treated to recover the catalyst for reuse. However, we found other convenient options for decreasing the amount of 2,5-furandicarboxylic acid monoalkyl ester in the mother liquor, which otherwise tends to accumulate under certain conditions. It was found that a convenient way to decrease the amount of this substance is by increasing the temperature in the oxidation reactor of step a), and / or increasing the residence time of the crude carboxylic acid in the oxidation reactor of step a), and / or applying a post-oxidation step a1a) after step a), and / or decreasing the temperature in the solid-liquid separation zone. A post-oxidation step has been found to be particularly effective when employed at high temperatures. In view of this observation, a preferred embodiment of the present process comprises controlling, preferably decreasing, the amount of monoalkyl ester of 2,5-furandicarboxylic acid in the mother liquor by increasing the temperature in the oxidation reactor of step a) and / or decreasing the temperature in the solid-liquid separation zone and / or increasing the residence time of the crude carboxylic acid in the oxidation reactor of step a) and / or applying a post-oxidation step a1) after step a), wherein the post-oxidation is carried out in a post-oxidation reactor under the conditions described for step a). The most preferred process is one in which the temperature in step a) is 170°C or higher and in which a post-oxidation step a) is applied after step a), wherein the post-oxidation is carried out in a post-oxidation reactor under the conditions described for step a). This process is especially preferred because the amount of ester was found to be ML / a / ZUZZ / UU 1040 FDCA monoalkyl in this case tends to plateau, meaning it increases only up to a certain level, where this level was found to be within the previously defined desired range. In fact, this observation has proven very useful for operating various processes starting from 5-HMF ethers, regardless of the metal incorporation problem. Accordingly, a process for producing a carboxylic acid composition comprising 2,5-furandicarboxylic acid is disclosed herein, comprising the following steps: a) oxidizing an oxidizable compound comprising 5-alkoxymethylfurfural in an oxidation reactor in the presence of a saturated organic acid solvent having 2 to 6 carbon atoms and a catalyst system comprising cobalt, manganese, and bromine using an oxidizing gas at a temperature in the range of 170 to 210°C to obtain a crude carboxylic acid composition comprising a monoalkyl ester of 2,5-furandicarboxylic acid and solid 2,5-furandicarboxylic acid; 12) oxidizing the crude carboxylic acid composition of step a) in a post-oxidation reactor in the presence of a saturated organic acid solvent having 2 to 6 carbon atoms and a catalyst system comprising cobalt, manganese, and bromine using an oxidizing gas at a temperature in the range of 170 to 210°C to obtain a crude carboxylic acid composition comprising a monoalkyl ester of 2,5-furandicarboxylic acid 2,5-furandicarboxylic acid and solid 2,5-furandicarboxylic acid, b) isolating at least a portion of the solid 2,5-furanedicarboxylic acid from the crude carboxylic acid composition in a solid-liquid separation zone to generate a solid cake and a mother liquor, c) determine the amount of manganese and / or cobalt in the cake, and d) increase the amount of one or more control acids in the oxidation reactor, if the determined amount of manganese and / or cobalt in the cake exceeds a predefined threshold value, wherein the mother liquor comprises 2,5-furandicarboxylic acid monoalkyl ester in the range of 0.5 to 7% by weight with respect to the weight of the mother liquor. A process is preferred in which one or more control acids are selected from the group consisting of hydrobromic acid, bromoacetic acid, dibromoacetic acid, 5-bromo-2-furoic acid, fumaric acid, acetoxyacetic acid, maleic acid, and furoic acid. More preferably, the control acid is selected from the group consisting of hydrobromic acid, bromoacetic acid, dibromoacetic acid, acetoxyacetic acid, and 5-bromo-2-furoic acid. More preferably, the control acid is selected from the group consisting of hydrobromic acid, bromoacetic acid, dibromoacetic acid, acetoxyacetic acid, and 5-bromo-2-furoic acid. The previous process is preferred because the respective control acids were found to provide a particularly good and pronounced effect, while at the same time being comparatively inexpensive and easy to handle, or are available as waste and / or byproducts of the process according to step a) and can be obtained in the mother liquor of step b). More preferably, a mixture of control acids comprising bromoacetic acid, dibromoacetic acid, acetoxyacetic acid, and 5-bromo-2-furoic acid is added to the oxidation reactor. It is also preferred that the control acids be relatively stable, i.e., resistant to chemical decomposition, so that they do not require frequent replenishment. A process according to the invention is preferred, wherein the process for producing a carboxylic acid composition is a continuous or semi-continuous process, preferably a continuous process, wherein at least a portion, preferably at least 60% by weight, more preferably at least 80% by weight, of the mother liquor is directed from the solid-liquid separation zone to the oxidation reactor as a recycled mother liquor stream, wherein preferably the portion of mother liquor that is not directed to the oxidation reactor as a recycled mother liquor stream is treated in an evaporation step to recover the organic acid solvent as a condensed vapor stream and / or wherein preferably one or more bases are added to the mother liquor treated in the evaporation, preferably in an amount that is equal to or greater than the amount of free bromine ions in the mother liquor, on a molar basis. The process of the present invention provides acceptable results for batch processes, where, for example, a sample of the crude carboxylic acid composition comprising the solid precipitate from the batch reactor is taken and processed in a solid-liquid separation zone according to step b). If required, control acids can be added to the oxidation reactor of the ongoing batch process. Furthermore, it is possible to complete a first batch process in order to analyze the resulting product cake and provide additional control acids to the second batch if the amount of metal in the cake of the first exceeds the predefined threshold value. However, the process defined above is clearly preferred, as the process of the present invention shows its full potential in continuous or semi-continuous processes. These processes require suitable control mechanisms that allow for minimally invasive adjustment of the operating system to counteract the problem of metal incorporation into the cake. Such processes generally involve the continuous or intermittent addition of an oxidizable compound and the removal of the carboxylic acid composition comprising 2,5-furandicarboxylic acid. Advantageously, the mother liquor obtained in step b) can be reused in a subsequent series of batch experiments to increase the amount of control acids in the oxidation reactor.However, the design of the process of the present invention as a continuous or semi-continuous process allows the mother liquor to be redirected from the solid-liquid separation zone back to the oxidation reactor as a recycled mother liquor stream. This allows the technical expert to increase the amount of control acids in the oxidation reactor, if the mother liquor stream comprises control acids. A process according to the invention is preferred, wherein the oxidizable compound comprises 5-methoxymethylfurfural, and wherein the crude carboxylic acid composition comprises 2,5-furandicarboxylic acid monomethyl ester. It is believed that the present process can be employed for 5-alkoxymethylfurfural regardless of the alkoxy chain length, and especially for 5-alkoxymethylfurfural where the alkoxy group comprises 1 to 6 carbon atoms. The best results were found to be obtained when 5-methoxymethylfurfural is used as the oxidizable compound. This is particularly advantageous because 5-methoxymethylfurfural has proven to be one of the most cost-effective starting materials for producing FDCA. A process according to the invention is preferred, wherein the mother liquor comprises monoalkyl ester of 2,5-furandicarboxylic acid, preferably monomethyl ester of 2,5-furandicarboxylic acid, in the range of 1.0 to 4% by weight with respect to the weight of the mother liquor. The above process is preferred because it was found that the above range for the amount of 2,5-furandicarboxylic acid alkyl ester in the mother liquor ensures that the beneficial effect of the compound is strong enough to be felt while at the same time a sufficient buffer is established towards the upper limit of FDCA monoalkyl ester that was identified, so the process has high flexibility with respect to variations and peaks in the concentration of 2,5-furandicarboxylic acid alkyl ester. A process is preferred in which the mother liquor comprises bromoacetic acid, preferably in an amount of 0.5% or more by weight relative to the weight of the mother liquor, and / or dibromoacetic acid, preferably in an amount of 0.1% or more by weight relative to the weight of the mother liquor, and / or 5-bromo-2-furoic acid preferably in an amount of 0.02% or more by weight relative to the weight of the mother liquor. Depending on the chosen parameters, the mother liquor was found to comprise bromoacetic acid and / or dibromoacetic acid and / or 5-bromo-2-furoic acid. The formation of these compounds, which can act beneficially as a control acid, has not been previously reported for other FDCA production processes, for example, processes starting from 5-HMF, and could potentially be a characteristic feature of processes employing 5-alkoxymethylfurfural as the oxidizable compound, at least if the specific reaction conditions are established as defined above in step a). A process is preferred in which the predefined threshold value for cobalt in the cake is 200 ppm by weight, preferably 50 ppm by weight, and more preferably 30 ppm by weight, relative to the weight of 2,5-furandicarboxylic acid, and / or in which the predefined threshold value for manganese in the cake is 100 ppm by weight, preferably 25 ppm by weight, and more preferably 15 ppm by weight, relative to the weight of 2,5-furandicarboxylic acid. This process is preferred because the respective threshold values ensure that a solid cake is obtained, sufficiently free of metal to allow efficient downstream processing. Additionally, a process according to the invention is preferred in which the predefined threshold value also includes the weight ratio of manganese to cobalt in the cake to the weight ratio of manganese to cobalt in the catalyst system. A process is preferred in which the amount of one or more control acids in the oxidation reactor is increased by adding one or more control acids to the oxidation reactor by increasing the portion of the mother liquor that is directed to the oxidation reactor as recycled mother liquor. This process is preferred because it eliminates the need to handle and / or store additional control acids on the production side, reducing costs and eliminating the need for additional equipment. The amount of control acids in the oxidation reactor can be increased, especially for continuous or semi-continuous processes, by directing the mother liquor to the oxidation reactor as a recycled mother liquor stream. If the mother liquor stream contains one or more control acids and replaces the fresh solvent in the oxidation reactor, the concentration of the one or more control acids contained in the mother liquor will increase in the oxidation reactor. This configuration allows for sophisticated process control, where the amount of control acids in the oxidation reactor can be increased by increasing the portion of the mother liquor directed to the oxidation reactor as a recycled mother liquor stream. A process according to the invention is preferred, wherein the weight ratio of cobalt to manganese in the catalyst system is 10 or more, preferably 15 or more, and / or wherein the weight ratio of bromine to the combined weight of cobalt and manganese in the catalyst system is 1 or greater, preferably 1.5 or greater, more preferably 2 or greater, wherein the value is preferably less than 4.0, more preferably less than 3.5. If the catalyst system comprises other metals besides cobalt and manganese in an amount of 5% by weight or more, it is preferred that the above ratios be achieved for the weight ratio of bromine to the combined weight of all metals in the catalyst system. This process is especially preferred because we found that the above catalyst system significantly outperforms other catalyst systems under the conditions defined above for step a).In particular, the inventors surprisingly discovered that the respective catalyst system reduces the tendency of manganese and cobalt from the catalyst system to be incorporated into the product cake, even when increased amounts of 2,5-furandicarboxylic acid monoalkyl ester are present, and reduces the manganese enrichment relative to cobalt in the cake. Furthermore, the use of the aforementioned catalyst system opens up a much wider range of process parameters, such as pressure or residence time, allowing the process according to the invention to be run very reliably, significantly reducing the likelihood of unintended process shutdowns and / or minimizing the formation of unwanted byproducts. A process according to the invention is preferred, wherein isolating at least a portion of the solid 2,5-furandicarboxylic acid in a solid-liquid separation zone comprises washing the solid 2,5-furandicarboxylic acid with a washing solution comprising a saturated organic acid solvent having 2 to 6 carbon atoms, preferably acetic acid, and less than 15%, preferably less than 10%, by weight of water. The above process is preferred because the amount of metals in the cake can be further reduced by employing a washing step in the solid-liquid separation zone. It was quite surprising that washing solutions comprising predominantly a saturated organic acid solvent, preferably acetic acid, produce reasonable success with respect to metal removal from the cake.In the prior art, it was often considered necessary to use larger quantities of water to ensure sufficient metal removal from the cake. However, with the process of the present invention, washing with organic acid was found to be sufficient to obtain cakes that are adequately metal-free. This is considered particularly advantageous because large quantities of water, which could be introduced into the system with the washing solution, are undesirable for any system using mother liquor recycle in an oxidation reactor, as the oxidation reaction is typically sensitive to higher water concentrations. A process is preferred where the weight ratio of manganese to cobalt in the cake, divided by the weight ratio of manganese to cobalt in the catalyst system, is less than 1.5, preferably less than 1.3. This process is preferred because it introduces a well-defined criterion for the process operator to judge whether the process is running within the desired regime, allowing for a very easy and reliable way to identify errors in the system. The process of the present invention requires the addition of potentially large quantities of control acid, where several of these control acids comprise ionic bromine or organic bromine compounds. These strong acids are corrosive and can potentially form highly oxidizing gaseous compounds. Consequently, the mother liquor stream comprises highly acetic and / or corrosive compounds and / or additional metal ions originating from the reactor equipment, for example, iron, nickel, or chromium. Furthermore, the oxidizing gaseous substances can potentially damage the head equipment of the oxidation reactor. Therefore, it is a further objective of the present invention to provide countermeasures to protect the equipment, particularly the overhead equipment and tubes in contact with the mother liquor stream, and / or to remove unwanted metals from the mother liquor.The inventors found that the following processes are preferred to address these challenges. ML / a / ZUZZ / UU 1040 A process according to the invention is preferred, wherein the process further comprises the step of: h) contacting at least a portion of the mother liquor with a composition comprising a base selected from the group consisting of Na2CO3 and NaOH to increase the pH to more than 7, wherein one or more metal hydroxides or carbonates are precipitated from the mother liquor, wherein the metal is selected from the group consisting of cobalt, manganese, iron, nickel or chromium. A process is preferred in which the mother liquor comprises cobalt in an amount greater than 2000 ppm by weight and manganese in an amount greater than 130 ppm by weight relative to the mother liquor weight. This process is preferred because it allows for efficient recycling of the mother liquor, where a sufficiently high quantity of catalyst metals can be retained in the oxidation reactor based on the cobalt and manganese supplied with the recycled mother liquor stream. A process is preferred in which the solid 2,5-furandicarboxylic acid isolated in step b) is further washed with a second wash solution comprising water in an amount greater than 95%, preferably greater than 99%, by weight relative to the wash solution. This process is advantageous because it can yield an FDCA cake containing minimal amounts of manganese and cobalt. However, as noted above, this process is often found to be more suitable for batch processes that do not employ mother liquor recycling. If the second wash solution is mixed with the mother liquor in the solid-liquid separation zone, reusing the mother liquor stream becomes more difficult, as the presence of large amounts of water is often undesirable in the oxidation reactions defined in step a). A process in which the organic acid solvent is acetic acid is preferred. This process is preferred because acetic acid has repeatedly proven to be the most suitable solvent used in most prior art processes. Acetic acid is inexpensive, readily available, and comparatively acceptable from an environmental perspective. A process is preferred in which the oxidizing gas comprises molecular oxygen and is preferably air. This process is preferred because the use of air is, in most cases, the most economically viable way to oxidize 5-alkoxymethylfurfural to FDCA. A process is preferred where the temperature in step a) is in the range of 170 to 190°C. This process is preferred because the inventors found that within this specific temperature range, the amount of monoalkyl ester of 2,5-furandicarboxylic acids tends toward a plateau value that is well within the desired range defined above. Furthermore, this temperature was found to produce FDCA with high yields and good purity, as qualitatively indicated by the white cake observed in several of the experiments. ML / a / ZUZZ / UU 1040 used the respective temperatures. A process according to the invention is preferred, wherein the pressure in step a) is in the range of 700 to 2000 kPa, and / or wherein the oxidation reactor comprises one or more continuous stirred-tank reactors, preferably two or more continuous stirred-tank reactors in series. Among the possible sets of parameters and equipment tested by the inventors, the above parameters were found to be ideal for obtaining high-purity FDCA with good yields while minimizing the energy costs required to pressurize the reactors. A process according to the invention is preferred, wherein the solid-liquid separation zone comprises a filter or centrifuge, preferably a filter, and more preferably a rotary pressure filter. The above process is advantageous because filters and centrifuges have been found to be particularly suitable means for isolating solid FDCA from the mother liquor comprising monoalkyl ester of 2,5-furandicarboxylic acid, although these compounds are often difficult to separate from solid FDCA. A process according to the invention is preferred, wherein the cake comprises 2,5-furandicarboxylic acid in an amount greater than 95%, preferably greater than 98%, by weight relative to the weight of the dry cake and preferably monoalkyl ester of 2,5-furandicarboxylic acid in an amount in the range of 0.1 to 3%, preferably 0.15 to 2.3% by weight relative to the weight of the dry cake, and / or wherein the cake comprises a combined amount of cobalt and manganese of less than 300 ppm, preferably less than 75 ppm, by weight relative to the weight of 2,5-furandicarboxylic acid in the cake. It is thought that the amount of FDCA monoalkyl ester should be less than 3% by weight relative to the weight of the dry cake to avoid a detrimental effect of this compound on subsequent purification methods. In view of the foregoing description of the present invention, those skilled in the art understand that the results obtained by the inventors of the present invention also allow for the definition of an optimized oxidation process for the production of FDCA. Such a process utilizes all the information discussed above to provide a process that can operate reliably and provides low initial metal incorporation into the cake, thus requiring less effort for process control. Consequently, an associated aspect of the present process may be: Process for producing a carboxylic acid composition comprising 2,5-furandicarboxylic acid, comprising the step of: a) oxidizing an oxidizable compound comprising 5-alkoxymethylfurfural in an oxidation reactor in the presence of a saturated organic acid solvent having 2 to 6 carbon atoms and a catalyst system comprising cobalt, manganese and bromine using a gas MA / a / ZUZZ / UU 1040 oxidant at a temperature in the range of 160 to 210°C to obtain a crude carboxylic acid composition comprising monoalkyl ester of 2,5-furandicarboxylic acid and solid 2,5-furandicarboxylic acid, wherein the liquid phase in the reactor comprises monoalkyl ester of 2,5-furandicarboxylic acid in the range of 0.5 to 7% by weight with respect to the weight of the liquid phase, wherein bromine is provided as hydrobromic acid, wherein the weight ratio of cobalt to manganese in the catalyst system is 10 or higher, preferably 15 or higher, and wherein the weight ratio of bromine to the combined weight of cobalt and manganese in the catalyst system is 1 or higher, preferably 1.5 or higher, more preferably 2 or higher. It is clear to the person skilled in the art that the preferred embodiments of this process correspond to the preferred embodiments of the previously disclosed process, for example, with respect to the solvent, starting materials, catalyst, temperature, and 2,5-furandicarboxylic acid alkyl ester ranges. From here on, the invention is described in more detail using experiments. EXAMPLES Unless otherwise described, the oxidation reactor is a 600 mL stirred pressure vessel with two impellers. The reactor is precharged with a solvent comprising acetic acid and water in a 95:5 weight ratio and catalyst components to obtain the specified composition of the catalyst system comprising cobalt, manganese, and bromine. The catalyst components are provided as cobalt(II) acetate tetrahydrate, manganese(II) acetate tetrahydrate, and HBr at 48% by weight in water. The typical precharge amount is 310 grams. The oxidation reactor is purged, pressurized, and heated to the desired operating temperature with stirring at 2000 rpm. The oxidizable compound provided as feed is either 5-methoxymethylfurfural (MMF) or a mixture of MMF with 5-hydroxymethylfurfural and small amounts of levulinic acid. The process is started with a typical feed rate of 8.3 mmol / min for 60 minutes (total feed of 500 mmol). A lean air flow rate (8% oxygen) is started at a typical flow rate of 10 L / min. The typical reaction begins within 3 minutes, noted by a sharp decrease in oxygen at the outlet and an increase in CO and CO2. Heat is generated during the reaction, and a steam stream is taken from above and condensed. This steam stream comprises mainly acetic acid and water.The amount of solvent captured at the top is continuously monitored, and is replenished in the oxidation reactor with a fresh flow of solvent into the reactor. The typical operating pressure is 12 to 14 barg at 160°C and 17.5 barg at ML / a / ZUZZ / UU 1040 175°C. At the end of the desired feeding period, the feeding of the oxidizable compound is stopped and the contents of the oxidation reaction are either rapidly cooled to room temperature (or the desired filtration temperature) or subjected to a prolonged post-oxidation period at the same reaction temperature and oxygen flow rate indicated above. EXAMPLE A Co / Mn ratio The experiments in Example A used a MMF feed at 160°C with a feed time of 1 hour and 15 minutes of post-oxidation. The cake was isolated by filtration and washed with 1 part solvent (95 parts acetic acid to 5 parts water, by weight) to 1 part of the estimated dry cake weight. The results are shown in Table 1. TABLE 1 Co / Mn ratios # Co in cat. (ppm) Mn in cat. (PPm) Br in cat. (PPm) Co in cake (PPm) Mn in cake (PPm) Al 2920 2820 6990 3500 9850 A2 2110 120 2740 185 31 A3 2290 130 2200 76 400 A4 2310 130 2020 116 7 The table shows that the present process can use different catalyst systems. High cobalt-to-manganese ratios appear to reduce the overall tendency of the system to incorporate metals into the cake. EXAMPLE B Proportion of metals in the cake versus the catalyst The experiments in Example B used a MMF feed at 160°C for 1 hour and 15 minutes of post-oxidation. The catalyst was 3300 ppm Co, 185 ppm Mn, and 7000 ppm Br of aqueous HBr for all experiments. After each run, the reaction slurry was cooled to 80°C and filtered. A minimal acetic acid / water wash (95 / 5 by weight) was used to displace the remaining mother liquor from the cake. The combined mother liquor and wash liquid were analyzed and corrected for cobalt, manganese, ionic bromine, and water (5%). This material was then used as the pre-charge for the next run to simulate a recycling operation. Consequently, subsequent experiments were performed with a larger amount of control acid in the oxidation reactor. The cobalt recovery in each case was 90–95%, allowing for a high recycling rate.The recycling process in each case used the mother liquor from the previous run as a pre-charge, after adjusting to make the catalyst at the desired level. A total of 8 runs were performed, with a total of 7 recycles. Table 2 shows both the cake quality and the mother liquor composition. The yield is the combined yield of FDCA and FDCA monoester recovered in the cake. The results are summarized in Table 2. ML / a / ZUZZ / UU 1040 TABLE 2 Mη / Co ratio in the cake versus the catalyst # Cake properties Mother liquor properties Yield (mol%) Co (PPm) Mn (PPm) Mn / Co cake / cat. FDCAMe (% by weight) Acetic acid (% by weight) Dibromoacetic acid (% by weight) Acetoxyacetic acid (% by weight) Fumaric acid (% by weight) B1 81.3 209 62 5.3 1.24 38 0.02 0.09 0.16 B2 79.8 154 9 1.0 2.50 0.64 0.05 0.16 0.26 B3 79.2 67 400 1.1 3.34 0.77 0.07 0.16 0.33 B4 78.7 88 5 1.0 4.10 0.85 0.09 0.18 38 B5 79.4 302 17 1.0 4.72 0.95 0.11 0.24 0.44 B6 79.6 354 18 0.9 4.96 0.92 0.08 0.26 0.40 B7 79.4 301 16 0.9 5.40 0.93 0.10 0.29 0.51 B8 80.0 639 31 0.9 5.76 0.87 0.07 0.25 0.49 The results show that in run B1, i.e., the run with fresh feed (without recycling), the Me / Co ratio in the cake divided by the same ratio in the catalyst is very high, indicating excessive Mn precipitation in the cake (often associated with a pink cake). Furthermore, relatively high amounts of Co and Mn were detected in the cake. The remaining runs, all of which included some FDCA-Me and control acids due to recycling (bromoacetic acid, dibromoacetic acid, acetoxyacetic acid, 5-bromo-2-furoic acid, and fumaric acid), had yields that remained constant and showed no excess manganese precipitation over cobalt. FDCA-Me was found to accumulate steadily with each recycle. The experimental setup did not allow for a significant further increase in the control acid, as only the mother liquor was used in each case, and the accumulation of acids at any point is governed by the steady state reached between reproduction and removal. It can be seen that the amount of metals in the cake decreases to highly desirable values as the concentration of FDCA-Me and the control acid initially increased in experiments B2 through B4. Furthermore, the Mn / Co ratio in the cake relative to that in the catalyst feed was consistently close to unity, indicating that no excess metal (Mn) was incorporated into the cake. However, while the amount of control acids plateaued, the accumulation of FDCA-Me continued, and its detrimental effect on metal incorporation apparently began to dominate the system. Increased filtration difficulty was observed when FDCA-Me levels were very high. In this case, the metals remained in the same ratio as in the catalyst feed, but were comparatively high, suggesting a problem with achieving thorough washing and subsequent removal of the mother liquor. EXAMPLE C Different diets The experiments in Example C were performed in the same manner as in Example B, but the feed used consisted of a mixture of 5-HMF (6.4 wt%), MMF (86.4 wt%), and a small amount of levulinates (2.3 wt%) with small amounts of other compounds. The recycling rate was over 90% for each of runs C2 to C4. The results are summarized in Table 3. TABLE 3 Different diets # Cake properties Properties of the mother cake Yield (mol%) Co (ppm) Mn (PPm) Mn / Co / cat cake. FDCAMe (% by weight) Acetic acid (% by weight) Dibromoacetic acid (% by weight) Acetoxyacetic acid (% by weight) Fumaric acid (% by weight) Cl 87.2 163 54 5.9 1.30 0.31 0.08 0.02 0.23 C2 86.4 29 1 0.6 2.43 0.53 0.09 0.03 0.42 C3 86.9 23 1 0.8 3.83 0.65 0.13 0.05 0.52 C4 81.7 45 1 0.4 4.82 0.53 0.19 nd 0.59 As in example B, the first run, i.e., the run performed without the addition of highly acidic components, resulted in a high level of manganese in the cake. In runs C2 to C4, which included controlling the acids in the oxidation reactor, the performance was good, and the total amount of metals in the cake was low. The experiments exemplify the process of the present invention for different feed compositions. In particular, the 5-HMF and MMF mixtures gave good results. EXAMPLE D Temperature The experiments in Example D were performed using the same feed as Experiment C, but at a temperature of 175°C (except for the first two runs at 160°C). The recycle rate was set at 80% of the mother liquor, using cobalt as the standard. A 1-hour post-oxidation was used, also at 175°C. The pressure was increased to 17.5 barg to accommodate the higher temperature and vapor pressure of the solvent. The first two runs were performed at 160°C to establish a mother liquor composition for the high-temperature experiments. The results are summarized in Table 4. TABLE 4 Temperature # Cake properties Mother liquor properties Yield 0 (mol %) Co (PPm) Mn (PPm) Μη / Co cake at cat. FDCAMe (% by weight) Br-acetic acid (% by weight) Dibromoacetic acid (% by weight) Acetoxyacetic acid (% by weight) Fumaric acid (% by weight) 5-Br-2 FCA (% by weight) DI 81.9 306 70 4.1 1.42 0.25 0.10 0.10 0.24 0.02 D2 84.2 16 1 1.1 2.60 0.39 0.18 0.15 0.37 0.03 D3 79.3 301 27.2 3.6 2.26 0.47 0.12 0.20 38 0.04 D4 81.7 14 1 1.3 2.45 0.49 0.14 0.21 0.44 0.04 D5 81.2 12.2 2 1.3 2.57 0.51 0.15 0.23 0.47 0.05 D6 80.6 24 2 1.5 2.50 0.45 0.13 0.22 0.44 0.05 D7 80.4 12 1 1.5 2.10 0.45 0.15 0.20 0.45 na Similar to examples B and C, the first run without added control acids resulted in a very high level of manganese in the cake. Again, the first run with a larger amount of control acid in the oxidation reactor (still at 160°C) yielded excellent metal content in the cake, even with comparatively low amounts of control acids. The third run, the first at 175°C, had a somewhat elevated level of manganese in the cake, which could be attributed to the new steady state of the process with respect to its mother liquor composition being established at the increased temperature. In other words, it is suspected that D3 would exhibit even more metal in the cake if run without control acids. In fact, the process was expected to terminate when runs were attempted directly at 170°C without adding control acids. The first two runs were performed at 160°C to establish a viable system.All subsequent runs show the desired low level of metals in the cake and provide excellent yields and a surprisingly low amount of metal in the cake, even for the final run. The monoester content stabilized at a lower level when run at 175°C with post-oxidation and 80% recycling compared to full recycling cycles at 160°C without post-oxidation. Along with the other experiments, this shows that excellent results can be achieved over several runs if the monoester content is managed. Additionally, the beneficial effect of the control acids is again confirmed by Example D. EXAMPLE E Effect of bromine content. The experiments in Example E were performed using the same setup as described above. The feed used was purified 5-methoxymethylfurfural (MMF), using a total of 500 mmol of MMF, which was fed at a constant rate for a total of 1 hour. The reaction temperature was 160°C, and the pressure was 12 barg. No post-oxidation was used. The reactor pre-charge was 310 grams, with catalyst and FDCA alkyl monoester (FDCAMe) added, as indicated below. The total reported yield is the sum of FDCA and FDCA-Me, after subtracting the initial FDCA-Me, on a molar basis relative to the MMF feed. In addition to the exact measurements, the cake color was used as a qualitative indicator for the concentration of manganese in the cake (ranging from white to pink) and the concentration of unwanted byproducts and colored bodies (ranging from white to yellow to brown).This qualitative analysis is quick and provides a fairly reliable impression of cake quality, since even small amounts of impurities result in a noticeable discoloration of the cake. The results are summarized in Table 5. TABLE 5 Effect of bromine content. # Co / Mn / Br ppm Initial FDCAMe, wt% Total yield, mol% Co cake, ppm Mn cake, ppm Mn / Co / Br cake Cake color E2 2200 / 125 / 2000 0 82.6 1771 466 4.6 Light pink E2 2200 / 125 / 2000 3 78.7 848 41 0.9 Yellow E3 2200 / 125 / 2000 6 73.3 2554 128 0.9 Yellow-brown E4 2200 / 125 / 2000 9 62.4 2558 124 0.9 Brown E5 3300 / 185 / 3000 0 82.3 9057 1067 2.1 Pink E6 3300 / 185 / 7000 0 90.2 74 13 3.1 White E7 3300 / 185 / 7000 6 90.1 294 14 0.8 White-yellow E8 3300 / 185 / 7000 6 83.5 2774 147 0.9 Yellow E9 3300 / 185 / 7000 6 86.4 1420 73 0.9 White Comparing E1 and E2, or E6 and E7-E9, shows that FDCA-Me appears to reduce the problem of manganese enrichment in the cake, as demonstrated by both the reduced absolute values and the drop in the Mi / Co cake / catalyst ratio, from numbers well above 1 without added FDCA-Me to numbers close to or below 1 with added FDCA-Me. However, E1 to E4 show that excess FDCA-Me negatively affects the process, with the E4 process, which yields a brown product, being considered a failure. This example also suggests that thorough cake washing is more difficult with excess FDCA-Me present, as indicated by elevated levels of both metals despite a Mi / Co cake / catalyst ratio close to unity. Comparing E5 and E6 demonstrates that adding additional Br as hydrobromic acid significantly reduces the amount of manganese in the cake.Additionally, E5 failed prematurely, with only 362 mmol of feed (out of the planned 500 mmol) when the reaction suddenly stopped, while E6 proceeded without issue. Based on these results, it can be deduced that HBr can function as a control acid within the meaning of the present invention. Indeed, it can be seen that the catalyst system employed in E6 to E9 is an optimized catalyst system for the oxidation of 5-alkoxymethylfurfural. While E3 yields a yellow-brown cake at 6 wt% FDCA-Me, the cake color in E7 to E9 does not reach the brown level. In contrast, even white cakes can be obtained in E9. Finally, the yields in E6 compared to E1, or in E7-E9 compared to E3, also increase substantially. EXAMPLE F Control acids The experiments in Example E were performed using the same setup as described above. The feed used was purified 5-methoxymethylfurfural (MMF), using a total of 500 mmol of MMF, which was fed at a constant rate for a total of 1 hour. The reaction temperature was 160°C, and the pressure was 12 barg. No post-oxidation was used. The reactor pre-charge was 310 grams, with catalyst and components added, as indicated below. The total reported yield is the sum of FDCA and FDCA-ME, after subtracting any initial FDCA-Me, on a molar basis relative to the MMF feed. The catalyst in all cases was 3300 ppm cobalt, 185 ppm manganese, and 7000 ppm bromine. The results are summarized in Table 6, where a pink cake was observed for experiments F4. IVIA / a / ZUZZ / UU 1040 TABLE 6 Control Acids # Added Component, Added Component, wt. % pKa Total Yield, mol% Co Cake, ppm Mn Cake, ppm Mn / Co / Catalyst Cake F1 none - - 90.2 74 13 3.1 F2 Fumaric Acid 1.1 3.03 85.2 30 <0.5 0.3 F3 Bromoacetic Acid 1.3 2.86 91.6 76 2 0.5 F4 Formic Acid 1.5 3.77 88.7 1704 494 5.2 F5 Maleic Acid 1.1 1.9 88.1 99 9 1.6 F6 2Furoic Acid 1.1 3.16 85.0 144 6 0.7 F7 FFCA 1.4 2.57 (est.) 75.4 711 259 6.5 Not all the acids added in the experiments resulted in the desired white cake that was also low in precipitated metals and free of manganese enrichment, as evidenced by the Miη / Co cake / catalyst ratio. The reference case (without added control acids) used the robust catalyst system identified earlier and produced a relatively white cake but still exhibited undesirable manganese enrichment. The addition of formic acid and 2-carboxy-5-(formyl)furan (FFCA) did not produce the desired white cake or the desired low Miη / Co cake / catalyst ratio. Considering that a white cake can be obtained with this catalyst in some cases, even without the addition of a control acid, it was concluded that the respective acids could even have a detrimental effect.Each of the other acids in the table had a positive impact on reducing the Min / Co cake / catalyst ratio, evidence that they are suitable for mitigating the problem of excess manganese in the cake. In most cases, the total metal content was also good, demonstrating good cake washing capabilities. Considering all the experimental evidence, it was deduced that suitable control acids are selected from the group consisting of hydrobromic acid and mono- or dicarboxylic acids with 2 to 5 carbon atoms and a pKa less than 3.2. EXAMPLE G Addition of HBr to bromoacetic acid The experiments in Example G were performed using the same setup as described above. The feed was a mixture of 5-HMF (6.4 wt%), MMF (86.4 wt%), and a small amount of levulinates (2.3 wt%), with minor amounts of other compounds. In each case, the reactor was pre-charged with 310 grams of acetic acid / water 95 / 5 wt and catalyst compositions as indicated. The reactor temperature was 170°C. The catalyst in all cases was Co / Mn / Br at 3300 / 185 / 7000 ppm wt, with bromine supplied using an aqueous HBr solution. It was observed that the catalyst composition that functioned well at 160°C, allowing a full 1-hour feed for a total of 500 mmol, did not function for the entire hour at 170°C and 17–18 barg.At some point during the execution, it would be observed that the reaction stopped abruptly, as evidenced by a rapid increase in the oxygen content of the outlet gas stream and a reduction in the production of CO2 and CO. TABLE 7 Addition of HBr or bromoacetic acid # Compound Added Lived / Died Cake Color Yield, FDCA+FDCA-ME, mol% G1 None Died NDND G2 Mother liquor from run at 160°C Lived Pink 89% G3 1.6 wt bromoacetic acid Lived White 92% G4 1.9 wt FDCA No start NANA G5 HBr, to total 8500 ppm Lived White approx. 87% In experiment G1 (no added acid), the process failed, preventing meaningful analysis of yields and cake color. In experiment G4 (no added FDCA), the process could not be started. In G2, mother liquor from a previous operation was added, increasing the amount of FDCA-Me and control acids in the oxidation reactor (by a relatively small amount). While G2 showed significant manganese incorporation into the cake, it resulted in a live process with acceptable yields. In experiment G3 (addition of bromoacetic acid) and experiment G5 (addition of HBr), a live process produced a desirable white cake, indicating low manganese content. EXAMPLE H Addition of HBr This example was performed using a single continuous stirred-tank oxidation reactor (CSTR). Cobalt and manganese levels were kept constant throughout the process, and feeds with different HBr levels were processed to observe the effect of the added HBr on the system. The reactor is equipped with a reflux condenser and pump to allow heat removal by solvent evaporation while the reflux is pumped back into the reactor. The reactor was precharged with approximately 100 grams of the specified catalyst pack in acetic acid. The reactor was heated to 160°C under nitrogen pressure. After reaching this temperature, the gas was changed to an air-nitrogen mixture, comprising 8% oxygen, at a flow rate of 3.3 Nl / min. The feed is 20 wt% RMF (a mixture of 5-HMF, MMF, and levulinics as previously used) in acetic acid with the desired cobalt and manganese (3000 ppm and 300 ppm, respectively). The feed contains approximately 1 wt% water, establishing a steady-state concentration in the reactor of approximately 6%, due to the water formed during oxidation.A valve at the bottom of the reactor was opened approximately every 30 seconds, removing a small amount of material to maintain a constant level and establish CSTR conditions. The temperature was maintained at 160°C, with a pressure of 13 barg and a residence time of 60 minutes. After at least 3 hours of operation, the reactor was considered to be at steady state and sampling could begin. At the end of each run, the feed was shut off and post-oxidation was performed. After post-oxidation, the reactor was cooled, and the contents were filtered, washed with acetic acid / water, and dried before analysis. The table below shows the metal content of the cakes. TABLE 8 Addition of HBr Bromine run in feed, ppm Bromine / (Co+Mn) wt / wt Co cake, ppm Mn cake, ppm Mn / Co / catalyst cake H1 514 0.16 2705 1100 4.1 H2 2023 0.61 1246 593 4.8 H3 4475 1.36 98 14 1.4 H4 5998 1.82 35 400 1.2 H5 6996 2.12 29 400 1.6 H6 8879 2.69 16 2 1.2 Analyses H1 and H2 show both a high overall level of metals in the cake and a high value for the cake / catalyst ratio of Mn / Co. All the remaining runs have a low total metal incorporation, decreasing as the HBr increased, and all show a very good, close to unity, cake / catalyst ratio for Mn / Co.
Claims
1. A process for producing a carboxylic acid composition comprising 2,5-furandicarboxylic acid, comprising the steps of: a) oxidizing an oxidizable compound comprising 5-alkoxymethylfurfural in an oxidation reactor in the presence of a saturated organic acid solvent having 2 to 6 carbon atoms and a catalyst system comprising cobalt, manganese, and bromine using an oxidizing gas at a temperature in the range of 160 to 210°C to obtain a crude carboxylic acid composition comprising a monoalkyl ester of 2,5-furandicarboxylic acid and solid 2,5-furandicarboxylic acid, b) isolating at least a portion of the solid 2,5-furandicarboxylic acid from the crude carboxylic acid composition in a solid-liquid separation zone to generate a solid cake and a mother liquor, c) determining the amount of manganese and / or cobalt in the cake, and d) increasing the amount of one or more control acids in the oxidation reactor,If the determined amount of manganese and / or cobalt in the cake exceeds a predefined threshold value, wherein one or more control acids are selected from the group consisting of hydrobromic acid and mono- or dicarboxylic acids having 2 to 5 carbon atoms and a pKa less than 3.2, and wherein the mother liquor comprises monoalkyl ester of 2,5-furandicarboxylic acid in the range of 0.5 to 7% by weight relative to the weight of the mother liquor.
2. The process according to claim 1, further characterized in that the one or more control acids are selected from the group consisting of hydrobromic acid, bromoacetic acid, dibromoacetic acid, 5-bromo-2-furoic acid, fumaric acid, acetoxyacetic acid, maleic acid, and furoic acid.
3. The process according to claim 1 or 2, further characterized in that the process for producing a carboxylic acid composition is a continuous process wherein at least 60% by weight, preferably at least 80% by weight, of the mother liquor is directed from the solid-liquid separation zone to the oxidation reactor as a recycled mother liquor stream.
4. The process according to any of claims 1 to 3, further characterized in that the oxidizable compound comprises 5-methoxymethylfurfural, and wherein the crude carboxylic acid composition comprises 2,5-furandicarboxylic acid monomethyl ester.
5. The process according to any of claims 1 to 4, further characterized in that the mother liquor comprises monoalkyl ester of 2,5-furandicarboxylic acid, preferably monomethyl ester of 2,5-furandicarboxylic acid, in the range of 1.0 to 4% by weight relative to the weight of the mother liquor.
6. The process according to any of claims 1 to 5, ML / a / ZUZZ / UU 1040, further characterized in that the mother liquor comprises bromoacetic acid, preferably in an amount of 0.5% or more by weight relative to the weight of the mother liquor, and / or dibromoacetic acid, preferably in an amount of 0.1% or more by weight relative to the weight of the mother liquor, and / or 5-bromo-2-furoic acid, preferably in an amount of 0.02% or more by weight relative to the weight of the mother liquor.
7. The process according to any of claims 1 to 6, further characterized in that the predefined threshold value for cobalt in the cake is 200 ppm by weight, preferably 50 ppm by weight, more preferably 30 ppm by weight, with respect to the weight of 2,5-furandicarboxylic acid and / or wherein the predefined threshold value for manganese in the cake is 100 ppm by weight, preferably 25 ppm by weight, more preferably 15 ppm by weight with respect to the weight of 2,5-furandicarboxylic acid.
8. The process according to any of claims 3 to 7, further characterized in that the amount of one or more control acids in the oxidation reactor increases by adding one or more control acids to the oxidation reactor by increasing the portion of the mother liquor that is directed to the oxidation reactor as recycled mother liquor.
9. The process according to any of claims 1 to 8, further characterized in that the weight ratio of cobalt to manganese in the catalyst system is 10 or higher, preferably 15 or higher, and / or wherein the weight ratio of bromine to the combined weight of cobalt and manganese in the catalyst system is 1 or higher, preferably 1.5 or higher, more preferably 2 or higher.
10. The process according to any of claims 1 to 9, further characterized in that isolating at least a portion of the solid 2,5-furandicarboxylic acid in a solid-liquid separation zone comprises washing the solid 2,5-furandicarboxylic acid with a washing solution comprising a saturated organic acid solvent having 2 to 6 carbon atoms, preferably acetic acid, and less than 15%, preferably less than 10%, by weight of water.
11. The process according to any of claims 1 to 10, further characterized in that the solid 2,5-furandicarboxylic acid isolated in step b) is further washed with a second washing solution comprising water in an amount of more than 95%, preferably more than 99%, by weight relative to the weight of the washing solution.
12. The process in accordance with any of claims 1 to 11, further characterized in that the temperature in step a) is in the range of 170°C to 190°C.
13. The process according to any of claims 1 to 12, further characterized in that the pressure in step a) is in the range of 700 to 2000 kPa, and / or wherein the oxidation reactor comprises one or more continuous stirred-tank reactors, preferably two or more continuous stirred-tank reactors in series.
14. The process according to any of claims 1 to 13, further characterized in that the weight ratio of manganese to cobalt in the cake to the weight ratio of manganese to cobalt in the catalyst system is less than 2.5, preferably less than 2, more preferably less than 1.
5.
15. The process according to any of claims 1 to 14, further characterized in that the cake comprises 2,5-furandicarboxylic acid in an amount greater than 95%, preferably greater than 98%, by weight relative to the weight of the dry cake and preferably monoalkyl ester of 2,5-furandicarboxylic acid in an amount in the range of 0.1 to 3%, preferably 0.15 to 2.3% by weight relative to the weight of the dry cake.