HK1204011B - Method for production of furane derivatives from glucose - Google Patents

Description

The present invention relates to a process for the production of furan derivatives from D-glucose.

Due to rising prices of fossil raw materials and an expected decrease in the supply of such raw materials, there is a strong interest in using renewable raw materials. The areas of energy production and the production of basic chemicals should be distinguished. The present invention relates to the latter area and concerns a process for producing furan derivatives from D-glucose.

D-Glucose is present in large quantities in various biopolymers, which are components of renewable raw materials. Examples include starch (e.g., corn starch) or cellulose (e.g., from lignocellulosic biomass). However, for the production of furan derivatives, fructose is a much more suitable starting material.

A common method for converting D-glucose to D-fructose involves the use of a suitable D-glucose isomerase, such as D-xylose isomerase, which accepts D-glucose as a substrate. Such processes have been known for a long time, for example, from US2950228, and are also suitable for industrial applications, as described, for instance, in US3616221 or US3868304.

WO 2011/124639 describes a process for producing the furan derivative hydroxymethylfurfural starting from fructose, glucose, or mannose, wherein the conversion of D-glucose to D-fructose is carried out using D-glucose isomerase.

One problem is that usually at most about 42% of D-glucose can be converted into D-fructose. A further enrichment of D-fructose relative to D-glucose can be achieved by separation processes. One possibility is the use of chromatographic methods, as described, for example, in US5221478. In the food industry, often only a partial enrichment of D-fructose is aimed for. However, chromatographic processes are very demanding, especially for the production of relatively pure or highly pure D-fructose.

In addition to the use of isomerases, enzymatic redox reactions on carbohydrates have also been described in the literature.

For example, DE69839381 describes a sorbitol dehydrogenase that can be used to convert D-sorbitol into L-sorbos, and can be applied in the production of ascorbic acid.

In DE10247147, a process is described in which D-fructose is reduced to D-mannitol using D-mannitol-2-dehydrogenase.

In US4467033, the enzymatic oxidation of L-sorbitol to L-fructose is described.

Examples of the reduction of D-xylose to xylitol are disclosed, for example, in US20060035353 or in Woodyer R. et al., FEBS J., 2005, Volume 272, p. 3816-3827.

It has already been shown that suitable xylose reductases can be used to reduce D-glucose to D-sorbitol (e.g., Wang X. et al., Biotechnol. Lett., 2007, Volume 29, p 1409-1412).

Sugar redox enzymes, such as sorbitol dehydrogenase, are also used for diagnostic purposes (e.g., DE60006330).

These processes involve individual redox reactions, in which either reduction or oxidation occurs for product formation.

Enzymatically catalyzed redox reactions are used in industrial processes, for example, in the production of chiral alcohols, α-amino acids, and α-hydroxy acids. The previously known industrial processes usually employ a redox enzyme for product synthesis, as well as optionally another enzyme for cofactor regeneration. This is different from processes where two or more enzymatic redox reactions involved in product formation, as well as any necessary enzymatic reactions for cofactor regeneration (simultaneously or sequentially), are carried out in a single reaction setup without isolating an intermediate product. In recent years, such enzymatic cascade reactions—here referred to as one-pot reactions—have gained significant attention because they effectively reduce operating costs, operating time, and environmental impact. Additionally, enzymatic redox cascades enable transformations that are not easily achievable by classical chemical methods.

Thus, an attempt was described to achieve the deracemization of racemic secondary alcohols via a prochiral ketone as an intermediate product using a one-pot system (J. Am. Chem. Soc., 2008, Volume 130, p. 13969–13972). The deracemization of secondary alcohols was achieved by two alcohol dehydrogenases (S- and R-specific) with different cofactor specificities. A disadvantage of this method is the very low concentration of the substrate used, 0.2–0.5%, which is not suitable for industrial purposes.

Another one-pot system was described in WO 2009/121785, where a stereoisomer of an optically active secondary alcohol was oxidized to the corresponding ketone and then reduced to the respective optical antipode, using two alcohol dehydrogenases with opposite stereoselectivities and different cofactor specificities. The cofactors were regenerated by means of a so-called "hydride transfer system" using only one additional enzyme. To regenerate the cofactors, various enzymes such as formate dehydrogenase, glucose dehydrogenase, and lactate dehydrogenase were used. A disadvantage of this method is the low concentration of the substrates used.

In contrast, many individual enzymatic redox reactions are already known. An example of application is the production of chiral hydroxy compounds starting from corresponding prochiral ketone compounds. In these processes, the cofactor is regenerated by means of an additional enzyme. What these processes have in common is that they represent isolated reduction reactions and regenerate NAD(P)H (see, for example, EP1152054).

Further examples of the enzymatic production of chiral, enantiomerically enriched organic compounds, such as alcohols or amino acids, have been described (Organic Letters, 2003, Volume 5, p. 3649-3650; US7163815; Biochem. Eng. J., 2008, Volume 39(2) p. 319-327; EP1285962). In these systems, a NAD(P)H-dependent oxidase from Lactobacillus brevis or Lactobacillus sanfranciscensis was used as a cofactor regeneration enzyme. The experiments also involved single reactions for product formation.

The advantages of a one-pot reaction, such as cost-effectiveness through time and material savings, are lost in the mentioned individually proceeding oxidation or reduction reactions.

Isolation of fructose from aqueous solutions is, for example, possible according to a method described in US4895601 or US5047088.

In the literature, various examples of the production of furan derivatives from carbohydrates are known.

In such processes, a wide variety of acidic catalysts have been used: inorganic acids (see, for example, Chheda, J. N.; Roman-Leshkow, Y.; Dumesic, J. A. Green Chem. 2007, 9, 342-350), organic acids (e.g., oxalic acid), zeolites (H-form), transition metal ions (see, for example, Young, G.; Zhang, Y.; Ying, J. Y. Angew. Chem. Int. Ed. 2008, 47, 9345-9348; Tyrlik, S. K.; Szerszen, D.; Olejnik, M.; Danikiewicz, W. Carbohydr. Res. 1999, 315, 268-272), heterogeneous dissolved metal phosphates (see, e.g., Asghari, F. S.; Yoshida, H. Carbohydr. Res. 2006, 341, 2379-2387), or strongly acidic cation exchangers (see, e.g., Villard, R.; Robert, F.; Blank, I.; Bernardinelli, G.; Soldo, T.; Hofmann, T. J. Agric. Food Chem. 2003, 51, 4040-4045).

As a solvent in such processes, water, as a green solvent, was preferentially investigated. Although a system consisting of biomass and water can be considered a "green approach," this cannot be said anymore at temperatures above 300°C and pressures exceeding 20 MPa, which are required to achieve acceptable yields (see, for example, Qi, X.; Watanabe, M.; Aida, T. M.; Smith Jr., R. S. Cat. Commun. 2008, 9, 2244-2249).

A special furan compound that can be produced from carbohydrates in the presence of acidic catalysts is hydroxymethylfurfural (hereinafter referred to as HMF). Processes for the production of HMF are also known from the literature. HMF can be obtained from carbohydrates in aqueous solution in the presence of both homogeneous and heterogeneous acids. The yields achieved vary between 30 and 60% depending on the carbohydrate substrate and reaction conditions. When water is used as a solvent, reaction conditions of 300 °C and 27 MPa are also described. Furthermore, the formation of by-products such as levulinic acid (LA) or insoluble humic acids is described (see, for example, Bicker, M., Kaiser, D., Ott, L., Vogel, H., J. of Supercrit. Fluids 2005, 36, 118-126; Szmant, H. H., Chundury, D. D., J. Chem. Techn. Biotechnol. 1981, 31, 135-145; Srokol, Z., Bouche, A.-G., van Estrik, A., Strik, R. C. J., Maschmeyer, T., Peters, J. A., Carbohydr. Res. 2004, 339, 1717-1726).

A flow process under supercritical conditions starting from D-glucose was described by Aida et al. (Aida, T. A.; Sato, Y.; Watanabe, M.; Tajima, K.; Nonaka, T.; Hattori, H.; Arai, K. J. of Supercrit. Fluids, 2007, 40, 381-388).

Organic solvents can also be suitable for the production of HMF. However, an important limitation is that they are sometimes difficult to separate from the product (see, for example, Bao, Q.; Qiao, K.; Tomido, D.; Yokoyama, C. Catal. Commun. 2008, 9, 1383-1388; Halliday, G. A.; Young Jr., R. J.; Grushin, V. V. Org. Lett. 2003, 5, 2003-2005). Additionally, many solvents used in the past are not suitable for possible subsequent reactions but generate by-products if not separated. Commonly used solvents for the conversion of carbohydrates into HMF are dimethyl sulfoxide (DMSO) and dimethyl formamide (DMF). Compared to water as a solvent, the conversion of carbohydrates to HMF can be carried out at relatively low temperatures of 80–140 °C and yields significantly higher yields (up to 95% in DMF) within shorter reaction times (30 minutes to 2 hours) (see, for example, Halliday, G. A., Young Jr., R. J., Grushin, V. V., Org. Lett. 2003, 5, 2003-2005; WO2009076627). It is assumed that DMSO acts as a catalyst in the dehydration of D-fructose (or other carbohydrates) to HMF (and similar compounds) (see: Amarasekara, A. S.; Williams, L. D.; Ebede, C. C. Carbohydr. Res. 2008, 343, 3021-3024).

Reaction mixtures of water/DMSO or water/toluene were also used in a continuous reaction setup, requiring reaction times of 4–6 hours at 140–180°C to achieve a maximum yield of 80% HMF (see Chheda, J. N., Roman-Leshkov, Y., Dumesic, J. A., Green Chem. 2007, 9, 342–350).

Ionic liquids can act both as neutral solvents and as active Brønsted acids, with the separation of ionic liquids still being a problem. Immobilized ionic liquids have also been used as Brønsted acid catalysts (see Bao, Q.; Qiao, K.; Tomido, D.; Yokoyama, C. Catal. Commun. 2008, 9, 1383-1388).

All known processes up to today have various disadvantages, such as low initial substrate concentration, low overall yields.

It was surprisingly found that a way has been discovered to achieve a better overall yield in the production of furan derivatives from D-glucose, where surprisingly high initial concentrations of D-glucose can be used.

In one aspect, the present invention provides a process for obtaining furan derivatives from D-glucose, characterized in that A) D-glucose is converted into D-fructose by means of an enzymatic process using and regenerating redox co-factors, wherein D-glucose is converted into D-fructose with the participation of two or more oxidoreductases, and during the conversion of D-glucose to D-fructose, first an enzymatically catalyzed reduction to D-sorbitol is carried out, followed by an enzymatically catalyzed oxidation of D-sorbitol to D-fructose, wherein NAD+/NADH and NADP+/NADPH are used as redox co-factors, and wherein as a result of two further enzymatically catalyzed redox reactions occurring simultaneously in the same reaction system, one of the redox co-factors is formed in its reduced form and the other in its oxidized form,and B) D-fructose is converted into furan derivatives and in step A), during the regeneration reaction which converts the reduced cofactor back into its original oxidized form, oxygen or a compound of the general formula in which R1 represents a straight-chain or branched-chain (C1-C4)-alkyl group or a (C1-C4)-carboxyalkyl group, is reduced, and in the regeneration reaction which converts the oxidized cofactor back into its original reduced form, a (C4-C8)-cycloalkanol or a compound of the general formula in which R2 and R3 are independently selected from the group consisting of H, (C1-C6)-alkyl, wherein alkyl is straight-chain or branched, (C1-C6)-alkenyl, wherein alkenyl is straight-chain or branched and contains one to three double bonds, aryl, particularly C6-C12 aryl,Carboxyl, or (C1-C4)carboxyalkyl, in particular also cycloalkyl, for example C3-C8 cycloalkyl, is oxidized.

In another aspect of a method according to the present invention, R1 is a substituted or unsubstituted, for example an unsubstituted C1-C4 alkyl group.

In another aspect, in a process according to the present invention, R2 and R3 are independently selected from the group consisting of H, (C1-C6)-alkyl, wherein alkyl is straight-chain or branched, (C1-C6)-alkenyl, wherein alkenyl is straight-chain or branched and contains one to three double bonds, aryl, particularly C6-C12 aryl, carboxyl, or (C1-C4)-carboxyalkyl.

A method provided by the present invention is also referred to here as a method according to/under the present invention.

"Oxidation reaction(s)" and "reduction reaction(s)" refer here to those enzyme-catalyzed redox reactions that are not part of cofactor regeneration and are involved in the formation of the product in a process according to the present invention. "Oxidation reaction(s)" and "reduction reaction(s)" are summarized under the term "product-forming reactions." The product-forming reactions in a process according to the present invention each include at least one oxidation reaction and at least one reduction reaction. In a preferred embodiment of the present invention, a process according to the present invention is characterized in that the oxidation reaction and the reduction reaction proceed simultaneously in time.

Enzymes and redox enzymes in a method according to the present invention include oxidoreductases. Oxidoreductases are enzymes that catalyze redox reactions. For example, oxidoreductases include dehydrogenases, reductases, oxidases, and peroxidases.

The mention of an acid or the salt of an acid here includes the respective term that is not mentioned. Similarly, the mention of acids, particularly bile acids, here includes all esters derived from them. Furthermore, compounds that are (partially) protected with protective groups are also included when referring to the underlying substances.

In a preferred embodiment of the present invention, the method according to the present invention is characterized in that a compound of formula I (2-oxo acid) is used, namely pyruvate (redox cosubstrate), which is reduced to lactate by means of lactate dehydrogenase, that is, during the regeneration reaction, which converts the reduced cofactor back into its original oxidized form, pyruvate is reduced to lactate by means of lactate dehydrogenase.

In a preferred embodiment of the present invention, the method according to the present invention is characterized in that a compound of formula II (redox cosubstrate) is a secondary alcohol, particularly 2-propanol (isopropyl alcohol, IPA), which is oxidized to acetone by means of an alcohol dehydrogenase, that is, during the regeneration reaction, which converts the oxidized cofactor back into its original reduced form, 2-propanol is oxidized to acetone by means of an alcohol dehydrogenase.

In a preferred embodiment of the present invention, the method according to the present invention is characterized in that oxygen is used as a redox cosubstrate, which is reduced by means of NADH oxidase.

In a preferred embodiment of the present invention, the method according to the present invention is characterized in that a secondary alcohol malate is used as a redox cosubstrate, which is oxidized by an oxaloacetate decarboxylating malate dehydrogenase ("malate enzyme") into pyruvate and CO2, for example, that during the regeneration reaction, which converts the oxidized cofactor back into its original reduced form, malate is oxidized by a malate dehydrogenase into pyruvate and CO2.

The resulting pyruvate is converted in another redox reaction in this embodiment, which does not serve product formation but represents the second cofactor regeneration reaction.

Suitable sources of D-glucose in a method according to the present invention include, for example, enzymatic or non-enzymatic hydrolysates of starch, particularly corn starch, enzymatic or non-enzymatic hydrolysates of sucrose, or enzymatic or non-enzymatic hydrolysates of cellulose. Cellulose used in a method according to the present invention can, for example, be obtained from biomass, preferably from lignocellulosic biomass, such as wood, straw, such as wheat straw, corn straw, bagasse, sisal, energy grasses. For the enzymatic hydrolysis of corn starch, for example, amylases can be used. For the enzymatic cleavage of sucrose, invertases, for example, are suitable. For the enzymatic cleavage of cellulose, cellulases, for example, can be used. For the non-enzymatic cleavage of the aforementioned polysaccharides, for example, an acid-catalyzed cleavage is suitable.

Stage A) of a method according to the present invention is carried out in an aqueous system, which may optionally contain a buffer. Suitable buffers include, for example, acetate, potassium phosphate, Tris-HCl, and glycine buffers, for example with a pH value of 5 to 10.5, preferably from 6 to 10. Furthermore, or alternatively, ions can be added to the system during the conversion of D-glucose to D-fructose for stabilizing the enzymes, such as Mg²⁺ or other additives such as glycerol.

In a method according to the present invention, D-glucose is converted into D-fructose according to Reaction Scheme 1 in Step A).

The present invention is characterized in that, during the conversion of D-glucose to D-fructose, first an enzymatically catalyzed reduction and then an enzymatically catalyzed oxidation are carried out. In a particular aspect, the present invention is characterized in that an isomerization of D-glucose occurs via reduction to D-sorbitol, which is then oxidized to D-fructose, particularly according to the following reaction scheme 2.

The method according to the present invention is characterized in that both reduction and oxidation reactions occur for converting D-glucose into D-fructose in the same reaction setup without isolating intermediate products.

Suitable enzymes for the reduction of D-glucose to D-sorbitol are known and include, for example, xylose reductases, which are available, for instance, from Candida tropicalis or Candida parapsilosis.

Suitable enzymes for the oxidation of D-sorbitol to D-fructose are known and include, for example, sorbitol dehydrogenases, which are available, for instance, from sheep liver, Bacillus subtilis, or Malus domestica.

A particular embodiment of the method according to the present invention is characterized in that at least one dehydrogenase is used in the conversion of D-glucose to D-fructose.

Both the enzymes and the redox cofactors can be used either in soluble form or immobilized on a support (solid phase).

Redox enzymes that are suitable for regenerating NAD+/NADH and/or NADP+/NADPH are known to experts in the field and include, for example, dehydrogenases.

In a method according to the present invention, both individual enzymes and fusion proteins can be used in step A), comprising two redox enzymes.

Another particular embodiment of the method according to the present invention is characterized in that enzymatic redox reactions are catalyzed by such dehydrogenases to convert D-glucose into D-fructose, which use the redox co-factors NAD+/NADH and/or NADP+/NADPH.

Here, NAD+ denotes the oxidized form and NADH the reduced form of nicotinamide adenine dinucleotide, while NADP+ denotes the oxidized form and NADPH the reduced form of nicotinamide adenine dinucleotide phosphate. The addition of redox cofactors may not be necessary if the enzyme solutions already contain them in sufficient concentrations. If the redox cofactors NAD(P)+ and/or NAD(P)H are added during the conversion of D-glucose to D-fructose, the concentration added in a method according to the present invention is usually from 0.001 mM to 10 mM, preferably from 0.01 mM to 1 mM.

Other redox enzymes for the regeneration of redox factors are known to experts and include, for example, alcohol dehydrogenases, NADH oxidases, hydrogenases, lactate dehydrogenases, or formate dehydrogenases.

Another specific embodiment of the method according to the present invention is characterized in that NAD+ is regenerated in the reaction of D-glucose to D-fructose in the same reaction mixture by means of a NADH oxidase.

Another particular embodiment of the method according to the present invention is characterized in that NADPH is regenerated in the conversion of D-glucose to D-fructose within the same reaction mixture by means of an alcohol dehydrogenase.

NADH oxidases and alcohol dehydrogenases are well known to experts. Alcohol dehydrogenases, for example, include those from Lactobacillus kefir. Suitable NADH oxidases are, for example, available from Leuconostoc mesenteroides, Streptococcus mutans, and Clostridium aminovalericum.

Another particular embodiment of the method according to the present invention is characterized in that NADPH is regenerated in the conversion of D-glucose to D-fructose in the same reaction mixture by alcohol dehydrogenase from Lactobacillus kefir.

For the regeneration of redox cofactors, co-substrates must be present and may need to be added.

Co-substrates are substances that are reduced or oxidized during the regeneration of NAD+/NADH and/or NADP+/NADPH (or other redox cofactors). Suitable co-substrates in a method according to the present invention include, for example, alcohols (e.g., 2-propanol), lactic acid and its salts, pyruvic acid and its salts, oxygen, hydrogen and/or formic acid and its salts.

NADPH can be regenerated, for example, by alcohol dehydrogenase from Lactobacillus kefir in the presence of the co-substrate 2-propanol (isopropanol), which is oxidized to acetone.

Possible reaction pathways for the conversion of D-glucose to D-fructose according to a method of the present invention are shown in the following reaction schemes 3 and 4: CtXR = Xylose reductase from Candida tropicalis SISDH = Sorbitol dehydrogenase from sheep liver LkADH = Alcohol dehydrogenase from Lactobacillus kefir, NADP(H)-dependent LacDH = Lactate dehydrogenase, NAD(H)-dependent CtXR = Xylose reductase from Candida tropicalis BsSDH = Sorbitol dehydrogenase from Bacillus subtilis LkADH = Alcohol dehydrogenase from Lactobacillus kefir, NADP(H)-dependent SmOxo = NADH oxidase from Streptococcus mutans

It was found that in a process according to the present invention, a high initial concentration of D-glucose in the aqueous reaction mixture of ≥ 5% (w/v) D-glucose, preferably ≥ 10% (w/v) D-glucose, particularly preferably ≥ 15% (w/v) D-glucose can be used.

In a further, preferred embodiment, D-glucose is used in the process according to the present invention in an aqueous reaction mixture at a concentration of ≥ 5% (w/v) D-glucose, preferably ≥ 10% (w/v) D-glucose, particularly preferably ≥ 15% (w/v) D-glucose, with a concentration of 50% (w/v), preferably 40% (w/v), particularly preferably 35% (w/v) should not be exceeded.

Due to the temperature-dependent solubility of D-glucose, the glucose concentration must be adjusted according to the respective reaction temperature during the procedure.

Enzymes can be used in the process according to the present invention as such, optionally in the form of cell lysates, optionally as recombinantly overexpressed proteins, for example as recombinantly overexpressed proteins in E. coli, with the further preference that the corresponding cell lysates can be used without additional purification. Depending on the enzyme to be produced, other microorganisms can also be used for expression, for example microorganisms known to those skilled in the art. Solid components of the respective microorganisms can either be separated or used together with the reaction in a process according to the present invention (e.g., whole-cell biocatalysts). Culture supernatants or lysates of microorganisms that already exhibit sufficient enzyme activity without the use of recombinant DNA technology can also be employed. In a process according to the present invention, both enzymes and redox cofactors can be used either in soluble form or immobilized on solid supports. One enzyme unit (1 U) corresponds to the amount of enzyme required to convert 1 µmol of substrate per minute.

Surprisingly, it was found that a high conversion can be achieved in a process according to the present invention when converting D-glucose to D-fructose, for example a conversion of ≥70% (w/v), such as ≥90% (w/v), e.g., ≥98% (w/v) and up to 99.9% (w/v), or even complete conversion can be achieved.

Depending on the enzymes used, the method according to the present invention can, for example, in step A), be carried out at temperatures of 10°C to 70°C, preferably at room temperature, e.g., 20°C to 50°C.

The D-fructose that can be obtained according to step A) of the present invention can be isolated, for example by means of crystallization.

For example, the 50% D-glucose content obtained during the hydrolysis of sucrose can be converted into D-fructose using a two-step enzymatic redox process according to the present invention, resulting in an increased proportion of D-fructose in the total sugar content. Thus, a suitable starting material for further conversion into furan derivatives becomes available. Surprisingly, it was found that the intermediate product D-fructose, which is obtained by a method according to the present invention, can be particularly well used for further conversion into furan derivatives.

The conversion of D-fructose to furan derivatives in step B) according to the present invention can be carried out by a suitable method, for example by a conventional method, or as described herein.

According to common methods, the conversion of D-fructose into furan derivatives can be carried out in a process according to the present invention in the presence of a catalyst, for example an acidic catalyst such as an inorganic acid, an organic acid, e.g., oxalic acid, a zeolite (H-form), transition metal ions, a heterogeneous dissolved metal phosphate, or a strongly acidic cation exchanger.

As a solvent in such processes, water or an organic solvent can be used, for example dimethyl sulfoxide (DMSO), dimethyl formamide (DMF), or N-methylpyrrolidone.

Preferably, the conversion of D-fructose to furan derivatives in step B) according to the present invention takes place in the presence of an acidic catalyst and in the presence of N-methylpyrrolidone (N-methyl-2-pyrrolidone, NMP) of the formula

The conversion of D-fructose to furan derivatives in step B) according to the present invention can be carried out either as a batch process or as a continuous process.

In a preferred embodiment, step B) is carried out according to the present invention by microwave heating.

Particular embodiments of the method according to the present invention are characterized in that N-methyl-2-pyrrolidone (NMP) is used either as a reaction solvent or as a co-solvent, namely as an additive to another solvent, during the conversion of D-fructose into furan derivatives.

In a particular embodiment of a method according to the present invention, NMP is used as a (co)solvent, for example, as a reaction solvent or as an additive to another solvent, in step b).

In a method according to the present invention, when NMP is used as a solvent, NMP can be used as the sole solvent, or NMP can be used together with a co-solvent. In the case where a co-solvent is used, an NMP concentration of up to 70% (v/v), for example up to 60% (v/v), based on the total amount of solvent can be employed. As co-solvents, for example, water or organic solvents known from the prior art, such as N,N-dimethyl sulfoxide (DMSO) or N,N-dimethyl formamide (DMF), may be considered.

In a process according to section B) of the present invention, D-fructose can be used in an amount up to 40% (w/v) and is generally used in an amount of 5 to 20%, although the reaction can also proceed at lower concentrations, for example, with a D-fructose concentration of about 1% (w/v). The minimum value is rather defined by economic considerations and not chemically.

Acid catalysts in stage B) in a method according to the present invention include common acid catalysts that can be used in the conversion of fructose into furan derivatives. Preferably, the catalyst is a Brønsted acid. Homogeneous acid catalysts, such as sulfuric acid or hydrochloric acid, or heterogeneous acid catalysts, for example cation exchange resins like montmorillonite, preferably montmorillonite KSF® or amberlite, such as amberlite®, preferably amberlite 15®, can be used. Furthermore, Lewis acid catalysts, such as CrCl2, AlCl3, SiO2-MgCl2 or a SILP (silica supported ionic liquid phase) catalyst, can also be used in a method according to the present invention. However, these generally provide less favorable results than the aforementioned catalysts.

In another aspect, a process according to the present invention is characterized in that, during the conversion of D-fructose to furan derivatives in step B), as an acidic catalyst a homogeneous acid catalyst, preferably sulfuric acid or hydrochloric acid; a heterogeneous acid catalyst, preferably an ion exchange resin, for example montmorillonite, such as Montmorillonite KSF® or an Amberlite, such as Amberlite®, preferably Amberlite 15®; a Lewis acid catalyst, such as CrCl2, AlCl3 or SiO2-MgCl2; a SILP catalyst, preferably a homogeneous or heterogeneous acid catalyst, is used.

The amount of catalyst needed in step B) can be easily determined by an expert through simple preliminary experiments. The amount depends on the type of catalyst used.

As an example, the following catalyst amounts are given relative to the amount of fructose used, particularly in the case where NMP is used as a solvent:

Katalysator	Menge
IN HCl	20 bis 200% (v/w)
HCl (37%)	2 bis 25% (v/w)
20 bis 200% (v/w)
2 bis 25% (v/w)
Montmorillonite KSF®	1 bis 50% (w/w)
Amberlite 15®	1 bis 50% (w/w)
1 bis 20% (w/w)
20 bis 200% (w/w)
SILP	10-200% (w/w)

In this case, the indicated values are not problematic at a concentration of about 10% (w/v) D-fructose; however, at higher fructose concentrations, the amount of catalysts should be limited so that the fructose can still be dissolved in the remaining amount of solvent.

The process in step B) according to the present invention is carried out at suitable temperatures. Suitable temperatures include, particularly when NMP is used as a solvent, temperatures from 100 to 220°C, preferably from 115 to 200°C, and particularly preferably from 135 to 185°C.

The reactions in step B), using NMP as a solvent, were carried out experimentally throughout in closed vessels (batch, microwave) without active pressure control. From the microwave runs, a maximum pressure of about 2–4 bar can be assumed for NMP, strongly depending on the additives. For example, if HCl is used as a catalyst, the resulting pressure can rise up to 15 bar. In continuous operation, a constant back pressure was applied to prevent boiling of the solvent, up to approximately 40 bar. Pressure arises either as the vapor pressure of the solvent(s) or additives, or a system-related (pump) pressure is applied. However, pressure does not seem to be crucial for the reaction mechanism.

It has been found that the mainly formed furan derivative in a process according to the present invention is hydroxymethylfurfural (HMF) of the formula.

In another aspect, a method according to the present invention is characterized in that the furan derivative is hydroxymethylfurfural.

In a method according to the present invention, the term "HMF selectivity" is intended to mean the portion of consumed D-fructose that is converted into HMF.

Furan derivatives produced by the method according to the present invention can be used directly or further converted into subsequent products in other chemical reactions. For example, hydroxymethylfurfural can be oxidized further to 2,5-furandicarboxylic acid (FDCA). FDCA is known as a monomer for producing polymers such as polyethylene furanoate (PEF), which can be used similarly to polyethylene terephthalate (PET), for example, for hollow bodies, especially bottles such as beverage bottles, cosmetic bottles, or bottles for cleaning agents. When ethylene glycol from renewable sources and FDCA, which is accessible from HMF produced in a process according to the present invention, are used simultaneously, PEF can be obtained that consists practically entirely of renewable raw materials.

In another aspect, the present invention is characterized in that produced furan derivatives are further processed, for example, that hydroxymethylfurfural is further oxidized to 2,5-furandicarboxylic acid, which may be subjected to polymerization, for example, to produce polymers such as polyethylene furanoate (PEF).

Description of the Figures

Fig. 1 shows the results of the dehydration of D-fructose in N-methyl-2-pyrrolidone with sulfuric acid as a catalyst according to Example 5. Fig. 2 and Fig. 3 show the results of the dehydration of D-fructose in N-methyl-2-pyrrolidone with sulfuric acid as a catalyst - performed in a microwave reactor according to Example 12. Fig. 4 and Fig. 5 show the results of the dehydration of D-fructose in N-methyl-2-pyrrolidone with hydrochloric acid as a catalyst - performed in a microwave reactor according to Example 13. Fig. 6 shows the results of the dehydration of D-fructose in N-methyl-2-pyrrolidone with Montmorillonite KSF® as a catalyst - performed in a microwave reactor according to Example 14. Fig. 7 shows the results of the dehydration of D-fructose in N-methyl-2-pyrrolidone with hydrochloric acid as a catalyst - reaction in a flow reactor according to Example 15. Fig. 8 shows an overview of the tested conditions for the dehydration of D-fructose. Fig. 9 shows a schematic reaction setup for stopped flow microwave reactions and continuous flow reactions for the production of furan derivatives from D-fructose.

In the following examples, all temperature values are given in degrees Celsius (°C). The following abbreviations are used: EtOAc Ethyl acetate FDCA Furandicarboxylic acid h Hour(s) HMF 5-Hydroxymethylfurfural HPLC High-performance liquid chromatography IPA Isopropyl alcohol (2-propanol) LSL Levulinic acid MeOH Methanol NMP N-Methylpyrrolidone (N-Methyl-2-pyrrolidone) PET Polyethylene terephthalate PEF Polyethylene furanoate RT Room temperature SILP Supported Ionic Liquid Phase TFA Trifluoroacetic acid

Example 1 Conversion of D-glucose to D-fructose using a xylose reductase and a sorbitol dehydrogenase, with the use of an alcohol dehydrogenase for recycling NADPH and a lactate dehydrogenase for recycling NAD+

A 0.5 ml reaction mixture contains 50 mg/ml D-glucose and 6 U/ml of recombinant xylose reductase from Candida tropicalis (overexpressed in E. coli BL21 (DE3)), and 0.1 mM NADP+. For cofactor regeneration, 7% (v/v) isopropanol (IPA) and 6 U/ml of recombinant alcohol dehydrogenase from Lactobacillus kefir (overexpressed in E. coli BL21 (DE3)) are added. The enzymes are used in the form of cell lysate. The reaction takes place for 24 hours at 40°C and pH = 9 (50 mM Tris-HCl buffer) under continuous shaking (900 rpm) in an open system. The open system allows removal of the formed acetone, which drives the reaction towards D-sorbitol formation. In the open system, water and IPA also evaporate.so that they are redosed after 6 hours and 21 hours. In each case, a total volume of 0.5 ml and an IPA concentration of 7% (v/v) are set again. After 24 hours, the reaction vessel is incubated at 60°C under vacuum to inactivate the enzymes and evaporate the organic solvents. After cooling to room temperature, recombinant D-sorbitol dehydrogenase from Bacillus subtilis (overexpressed in E. coli BL21 (DE3)) is added at a final concentration of 5 U/ml, ZnCl2 at a final concentration of 1 mM, and NAD+ at a final concentration of 0.1 mM. For cofactor regeneration, 5 U/ml (final concentration) of lactate dehydrogenase from rabbit muscle (Sigma Aldrich) and 300 mM pyruvate are used.The mixture is adjusted to 0.5 ml with water. The reaction proceeds for another 24 hours at 40°C under continuous shaking (900 rpm) in a closed system. A conversion of D-glucose to D-fructose of more than 90% is achieved.

Example 2 Conversion of D-glucose to D-fructose by a xylose reductase and a sorbitol dehydrogenase, using an alcohol dehydrogenase for the recycling of NADPH and an oxidase for the recycling of NAD+

A 0.5 ml reaction mixture contains 50 mg/ml D-glucose, 6 U/ml of recombinant xylose reductase from Candida tropicalis (overexpressed in E. coli BL21 (DE3)), and 0.1 mM NADP+. For cofactor regeneration, 7% (v/v) IPA and 6 U/ml of recombinant alcohol dehydrogenase from Lactobacillus kefir (overexpressed in E. coli BL21 (DE3)) are added. The enzymes are used in the form of cell lysate. The reaction proceeds for 24 hours at 40°C and pH = 8 (50 mM Tris-HCl buffer) with continuous shaking (900 rpm) in an open system. The open system allows removal of the produced acetone, which drives the reaction towards D-sorbitol formation. In the open system, water and IPA also evaporate.so that they are re-dosed after 6 hours and 21 hours. In each case, a total volume of 0.5 ml as well as an IPA concentration of 7% (v/v) are adjusted again. After 24 hours, the reaction vessel is incubated at 60°C under vacuum to inactivate the enzymes and to evaporate IPA as well as the generated acetone. After cooling to room temperature, recombinant D-sorbitol dehydrogenase from Bacillus subtilis (overexpressed in E. coli BL21 (DE3)) is added in a final concentration of 5 U/ml, CaCl2 in a final concentration of 1 mM, and a mixture (1:1) of NAD+ and NADH in a final concentration of 0.1 mM. For co-factor regeneration, 10 U/ml (final concentration) of NADH oxidase from Leuconostoc mesenteroides (overexpressed in E. coli) is added.The enzymes are used in the form of cell lysate. The reaction mixture is filled up to 0.5 ml with water. The reaction proceeds for another 24 hours at 40°C with continuous shaking (900 rpm) in an open system to ensure sufficient oxygen supply for the NADH oxidase from the air. In the open system at 40°C, water evaporates. Therefore, it is replenished with water to 0.5 ml after 6 hours and 21 hours. A conversion rate of approximately 98% of D-glucose to D-fructose is achieved.

Example 3 Processing and Analytics of Sugar

The sample is incubated for 10 minutes at 65°C to inactivate the enzymes and then centrifuged. The supernatant is subsequently filtered through a 0.2 µm PVDF filter and analyzed by Ligand-Exchange-HPLC (Agilent Technologies, Inc.). Sugars and polyols are separated on a lead column from Showa Denko K.K. (Shodex® Sugar SP0810) with a flow rate of 0.5 ml/min of water (VWR International GmbH, HPLC grade) at 80°C. Detection is performed using a Refractive Index Detector (RID, Agilent 1260 Infinity®, Agilent Technologies, Inc.). An inline filter from Agilent Technologies, Inc. is used, as well as an anion-exchange pre-column (Shodex® Axpak-WAG), a reversed-phase column (Shodex® Asahipak® ODP-50 6E), and a sugar pre-column (SUGAR SP-G), all from Showa Denko K.K.

Example 4 Materials and Methods for the Conversion of D-Fructose to Furan Derivatives

In the context of this invention, dehydration reactions of D-fructose to HMF were carried out under various reaction conditions, either as a standard batch process, under microwave-assisted heating, or using "continuous flow" conditions. Fig. 8 shows an overview of the tested conditions. Surprisingly, it was found that NMP as a solvent provides higher yields in the reaction compared to previously known systems when combined with either homogeneous or heterogeneous catalysts, both in the microwave-assisted process and under "continuous flow" conditions.

Synthesis of SiO₂-MgCl₂

SiO₂-MgCl₂ was prepared according to a procedure described by Yasuda et al. (Yasuda, M.; Nakamura, Y.; Matsumoto, J.; Yokoi, H.; Shiragami, T. Bull. Chem. Soc. Jpn. 2011, 84, 416-418).

Synthesis of SILPs

The SILP catalyst was prepared according to known procedures (Fu, S.-K.; Liu, S.-T. Synth. Commun. 2006, 36, 2059-2067) using N-methylimidazole. For immobilization, the obtained ionic liquid was mixed with 200 wt.% silica gel in dry chloroform (100 mL per 10 g SiO₂) and heated for 24 h at 70 °C. The resulting solid was filtered off, washed with chloroform, and dried under reduced pressure. The obtained silica gel showed a catalyst loading of approximately 16 wt.%.

General Terms Batch Reactions

If not otherwise described, all batch reactions were carried out in 4 mL screw-cap vials. Heating was performed in suitable aluminum blocks up to the desired temperature.

Batch Microwave Reactions

Microwave reactions were carried out in a batch process using a Biotage Initiator Sixty laboratory microwave equipped with an autosampler to enable sequential reaction procedures. The absorption level was set to maximum, automatically regulating the maximum energy input to 400 W.

Stopped-flow microwave reactions and continuous-flow reactions

Stopped-flow reactions for optimizing a semi-continuous process were carried out on a CEM® Discover System with a CEM® Voyager upgrade and using an external pressure sensor. For continuous process reactions, a cartridge-based reactor system X-Cube from ThalesNano® was used, equipped with a Gilson® GX-271 autosampler for automatic product collection. Two quartz sand cartridges (CatCart®, 70 x 4 mm) were installed as the reaction zone.

Alternatively, a perfluorooxyalkane capillary (PFA capillary, 0.8 mm inner diameter, 1.6 mm outer diameter) was wrapped around a heated aluminum cylinder. The substrates were introduced into the system at the desired flow rate using a Shimadzu LC-10AD HPLC pump. Exact volumes (column volume 16.0 mL; dead volume before and after the column each 1.0 mL) were determined by tracking defined flow rates of the pure solvent with a digital stopwatch. The reaction setup is shown in Fig. 9.

Analysis of the Reactions for the Conversion of D-Fructose to Furan Derivatives

For quantitative HPLC analysis, samples of the reaction mixtures (22 µL unless otherwise stated) were diluted with deionized water to a final volume of 1 mL. For reaction samples with different concentrations, the dilution was adjusted so that the maximum concentration did not exceed 2 mg/mL.

To this solution, 100 µL of 3-hydroxybenzyl alcohol was added as an internal standard, after which the sample was thoroughly mixed. Solid residues were separated by centrifugation (5 min, 20000 G) or filtration (Phenex PTFE, 4 mm, 0.2 µm). Quantification was performed based on the peak areas in the RI spectrum compared to the internal standard.

The samples were analyzed using HPLC on a Thermo Scientific® Surveyor Plus system or a Shimadzu® Nexera system, each equipped with a PDA Plus and RI detector. For separation, a ion exclusion column from Phenomenex® (Rezex RHM-Monosaccharides H+ (8%), 150 x 7.8 mm, composed of a cross-linked matrix of sulfonated styrene and divinylbenzene, H+-form) was used as the stationary phase, and a mobile phase consisting of water (HPLC-grade) and 0.1% TFA (HPLC-grade) as the eluent. The column temperature was kept constant at 85 °C, and the run time was optimized to 25 minutes. Product quantification was performed by integrating the RI signal using an internal standard. Additionally, wavelengths of 200 nm, 254 nm, and 280 nm were recorded using the PDA for further reaction analysis.

GP1 - Dehydration of D-Fructose in a Batch Process

In a standard reaction for reaction optimization, 100 mg of D-fructose (0.56 mmol) and the respective catalyst in the desired amount were placed into a glass vial and then mixed with 1 mL of freshly distilled NMP. The resulting solution/suspension was heated to the selected temperature and allowed to react for the desired time.

GP2 - Dehydration of D-Fructose in the Microwave Batch Process

In a standard reaction for reaction optimization, 100 mg of D-fructose (0.56 mmol) and the respective catalyst in the desired amount were placed into a microwave vessel (0.5-2.0 mL). The vessel was equipped with a stirring magnet and filled with 1 mL of NMP. The radiation intensity of the microwave was automatically adjusted by an in-house control algorithm to reach the desired temperature. A rapid cooling of the reaction vessel was achieved using blown-in compressed air at a pressure of at least 6 bar.

GP3 - D-Fructose Dehydration in the Microwave Stopped-Flow Procedure

In a standard reaction for reaction optimization, a D-fructose standard solution (1 mL; c = 100 mg/mL in NMP) and hydrochloric acid (100 µL; c = 1 mol/L) were placed into a microwave vessel and equipped with a magnetic stirrer. After sealing the vial with a Snap-Cap, the solution was heated to the desired temperature for the required time. To achieve the fastest possible heating, the supplied energy was set according to Table 1 below. Tabelle 1

Leistungseinstellung der Mikrowelle und zugehörige Temperaturen
100°C	50 W	180°C	125 W
125°C	65 W	200°C	140 W
150°C	100 W	220°C	160 W

A rapid cooling of the reaction vessel was achieved by blowing in compressed air at least 6 bar.

GP4 - Dehydration of D-Fructose in a Cartridge-Based Reactor System

In a standard reaction for reaction optimization, a D-fructose standard solution (1 mL; c = 100 mg/mL in NMP) was mixed with hydrochloric acid (c = 1 mol/L) and pumped into the reaction system using a reagent pump. During the heating process, several pre-samples were taken to monitor a stable temperature and a stable flow rate. Reaction temperatures of 150°C, 180°C, and 200°C were selected, with the reaction pressure regulated at 40 bar. For this purpose, flow rates between 0.2 and 0.6 mL/min were chosen. Reaction samples were collected in 2.5 mL portions and analyzed.

Example 5 Use of sulfuric acid as a catalyst for the dehydration of D-fructose

Various temperatures, reaction times, and acid concentrations were compared. The reactions were carried out according to "GP1" (Example 4). As a catalyst, either 100 µl of 1 N sulfuric acid or 10 µl of concentrated sulfuric acid was used. The results are summarized in Table 2. Tabelle 2

Schwefelsäure als Katalysator zur Dehydratisierung von D-Fructose
	100°C	3 h	69%	45%	65%	< 1%
	120°C	4 h	95%	77%	81%	< 1%
	150°C	15 min	98%	88%	90%	< 1%
	180°C	10 min	100%	85%	85%	< 1%
	120°C	45 min	98%	85%	90%	< 1%
	150°C	10 min	100%	90%	90%	< 1%
	180°C	5 min	100%	82%	82%	< 1%

The formation of black insoluble polymers and humic substances was not observed under the used optimal conditions. To analyze the reaction progress, a time series for a representative reaction was recorded (concentrated H2SO4, 150°C, see Fig. 1).

Example 6 Use of Chromium(II) Chloride as a Catalyst for the Dehydration of D-Fructose

As described by Zhao, H.; Holladay, J. E.; Brown, H.; Zhang, Z. C. Science 2007, 316, 1597-1600, chromium(II) chloride can be used as an efficient catalyst for the dehydration of D-fructose. In this example, the effect of CrCl2 in N-methyl-2-pyrrolidone is shown. The experiments were carried out according to procedure "GP1" (Example 4). While relatively low yields of HMF were achieved, significant amounts of tar-like compounds were observed (Table 3). Tabelle 3

Chrom-(II)-chlorid als Katalysator zur Dehydratisierung von D-Fructose
	100°C	3 h	86%	51%	59%	< 1%
	150°C	3h	100%	39%	39%	< 1%

Example 7 Use of Montmorillonite KSF® as a Catalyst for the Dehydration of D-Fructose

100 mg of D-fructose was incubated with 1 ml of N-methyl-2-pyrrolidone under stirring (procedure "GP1", example 4). A uniform reaction time of 3 hours was chosen. Different amounts of Montmorillonite KSF® were added as a catalyst. Table 4 summarizes the results. Under the best conditions, a yield of 61% HMF and a selectivity of 63% HMF could be achieved. Tabelle 4

Montmorillonit KSF® als Katalysator zur Dehydratisierung von D-Fructose
1 mg	120°C	37%	11%	31%	< 1%	nein
3 mg	120°C	54%	20%	38%	< 1%	nein
5 mg	120°C	65%	30%	46%	< 1%	nein
7 mg	120°C	73%	32%	44%	< 1%	nein
10 mg	120°C	80%	41%	52%	< 1%	nein
20 mg	120°C	90%	43%	48%	< 1%	nein
40 mg	120°C	94%	43%	46%	< 1%	nein
1 mg	130°C	31%	11%	35%	< 1%	nein
3 mg	130°C	73%	35%	48%	< 1%	nein
5 mg	130°C	87%	46%	53%	< 1%	nein
7 mg	130°C	92%	50%	55%	< 1%	nein
10 mg	130°C	94%	49%	52%	< 1%	nein
20 mg	130°C	96%	54%	57%	< 1%	nein
40 mg	130°C	97%	54%	55%	< 1%	ja
1 mg	140°C	72%	30%	42%	< 1%	nein
3 mg	140°C	91%	46%	51%	< 1%	nein
5 mg	140°C	95%	53%	56%	< 1%	nein
7 mg	140°C	96%	53%	55%	< 1%	nein
10 mg	140°C	98%	55%	56%	< 1%	nein
20 mg	140°C	98%	56%	57%	< 1%	nein
40 mg	140°C	99%	56%	56%	< 1%	ja
1 mg	150°C	94%	44%	46%	< 1%	nein
3 mg	150°C	96%	52%	54%	< 1%	nein
5 mg	150°C	98%	56%	57%	< 1%	nein
7 mg	150°C	98%	57%	59%	< 1%	nein
10 mg	150°C	98%	58%	59%	< 1%	ja
20 mg	150°C	97%	61%	63%	< 1%	ja
40 mg	150°C	97%	61%	63%	< 1%	ja

Example 8 Use of Amberlite 15® as a Catalyst for the Dehydration of D-Fructose

This example demonstrates the use of a strong ion-exchange resin based on sulfonic acid groups, which is cross-linked in a macroporous structure. 100 mg of D-fructose was incubated for 3 hours at 100°C with stirring in the presence of 1 ml of N-methyl-2-pyrrolidone (Procedure "GP1, Example 4). Amberlite 15® was added as a catalyst. The results of this experiment are shown in Table 5. In contrast to Montmorillonite KSF®, a higher yield was achieved at the relatively low temperature. The formation of tar-like compounds was avoided. Tabelle 5

Amberlite 15® als Katalysator zur Dehydratisierung von D-Fructose
10 mg	100°C	3h	70%	50%	71%	< 1%

Example 9 Use of SiO₂-MgCl₂ as a Catalyst for the Dehydration of D-Fructose

Since a silica gel-magnesium chloride complex showed catalytic activity in the dehydration of carbohydrates in acetonitrile (Yasuda, M.; Nakamura, Y.; Matsumoto, J.; Yokoi, H.; Shiragami, T. Bull. Chem. Soc. Jpn. 2011, 84, 416–418), this catalyst was tested for its suitability in N-methyl-2-pyrrolidone. Under reaction conditions similar to those in "GP1" (Example 4), a maximum yield of 26% HMF was achieved (see Table 6). However, when only silica gel was used, the yield dropped below 1%, and the formation of large amounts of tar-like compounds was observed. Tabelle 6

200 mg	150°C	30 min	99%	26%	26%	4%

Example 10 Use of AlCl3 as a catalyst for the dehydration of D-fructose

AlCl3 was tested as an example of a Lewis acid catalyst under reaction conditions "GP1" (Example 4). For this purpose, freshly sublimed AlCl3 was used. Similar results were obtained as with Amberlite 15®. However, the catalyst is sensitive to hydrolysis and therefore cannot be used for repeated applications or in continuous processes. Moreover, larger amounts of tar-like compounds were formed (see Table 7 for the results). Tabelle 7

10 mg	100°C	3h	100%	50%	50%	< 1%

Example 11 Use of SILPs combined with Chromium(II) chloride as a catalyst for the dehydration of D-fructose

Using the reaction conditions "GP1" (Example 4), a combination of CrCl₂ and SILPs (silica-supported ionic liquid phase, see Example 4) was tested. After 20 minutes, a yield of almost 50% HMF could be achieved. However, this yield could not be increased with longer reaction times. Furthermore, a conversion of D-fructose to D-glucose was observed at shorter reaction times (Table 8). Tabelle 8

120°C	5 min	85%	5%	39%	46%	< 1%
120°C	10 min	94%	3%	45%	48%	< 1%
120°C	15 min	99%	1%	44%	45%	< 1%
120°C	20 min	97%	2%	49%	51%	< 1%
120°C	25 min	97%	1%	47%	48%	< 1%
120°C	30 min	98%	< 1%	49%	50%	< 1%
120°C	45 min	99%	< 1%	48%	49%	< 1%
120°C	1 h	99%	< 1%	52%	52%	< 1%

Example 12 Use of Sulfuric Acid as a Catalyst for the Dehydration of D-Fructose (Microwave Heating)

To achieve better control over the heating and cooling phases as well as the reaction temperature, a microwave-based system for temperature adjustment was employed. Samples were prepared as described in specification "GP2" (Example 4), using N-methyl-2-pyrrolidone. Under the reaction conditions used, no formation of tar-like compounds was observed. A maximum complete conversion of D-fructose and a yield of 83% HMF were achieved (Fig. 2 and Fig. 3).

Example 13 Use of Hydrochloric Acid as a Catalyst for the Dehydration of D-Fructose (Microwave Heating)

The dehydration of D-fructose was carried out in a stopped-flow microwave reactor according to procedure "GP3" (Example 4). Higher temperatures were necessary to achieve complete conversion of D-fructose. While longer reaction times at lower temperatures improved the yield of HMF, the yield decreased at higher temperatures with increasing reaction time (Figures 4 and 5). A maximum yield of 89% HMF could be achieved with complete conversion of D-fructose.

Example 14 Use of Montmorillonite KSF® as a Catalyst for the Dehydration of D-Fructose (Microwave Heating)

Since microwave methods allow rapid heating/cooling and excellent temperature control within the reaction vessel, the heterogeneous catalyst Montmorillonite KSF® was also used for the dehydration of D-fructose to N-methyl-2-pyrrolidone. Reaction conditions according to "GP2" (Example 4) were applied. The reaction time was 5 minutes. Although only relatively low conversion rates of D-fructose and yields of HMF were achieved, the formation of tar-like compounds could be avoided (results see Table 9). Tabelle 9

Montmorillonit KSF® als Katalysator zur Dehydratisierung von D-Fructose (Mikrowellen-Erhitzung)
5 mg	150°C	51%	20%	39%	< 1%	nein
7 mg	150°C	61%	26%	43%	< 1%	nein
10 mg	150°C	64%	30%	46%	< 1%	nein
15 mg	150°C	76%	38%	50%	< 1%	nein
20 mg	150°C	82%	43%	52%	< 1%	nein

To find the best reaction conditions, various reaction times were tested using 20 mg of catalyst at 150°C (Figure 6).

Example 15 Use of sulfuric acid for catalyzing the conversion of D-fructose into furan derivatives (continuous process)

D-Fructose (10% w/v) and concentrated sulfuric acid (1% v/v) were dissolved in N-methyl-2-pyrrolidone. The mixture was pumped through the reactor using a PFA capillary under continuous flow (reaction temperature 150°C). After discarding the first 18 ml, an additional 10 ml were collected for analysis. The effect of different residence times in the reactor was tested by varying the flow rates (Table 10). Tabelle 10

Schwefelsäure zur Katalyse der Umsetzung von D-Fructose zu Furanderivaten (kontinuierlicher Prozess)
0.8 ml/min	20 min	100%	74%	74%	< 1%
1.6 ml/min	10 min	100%	75%	75%	< 1%
3.2 ml/min	5 min	100%	76%	76%	< 1%

Under the examined conditions, no formation of black insoluble polymers and humic substances was observed.

Example 16 Use of hydrochloric acid for catalyzing the conversion of D-fructose into furan derivatives (continuous process)

In this example, hydrochloric acid was used as a catalyst for the dehydration of D-fructose in NMP under continuous flow (reaction conditions see specification "GP4", example 4). A maximum yield of 75% HMF was achieved at a reaction temperature of 180°C and a flow rate of 0.6 ml/min. A selectivity of 76% HMF was obtained. Usually, the content of levulinic acid (LA) was below 1% (results see Fig. 7).

Claims (20)

1. A process for the production of furan derivatives from D-glucose, characterized in that

   A) D-glucose is converted into D-fructose in an enzymatic process, wherein redox cofactors are used and regenerated, wherein D-glucose is converted into D-fructose, involving two or more oxidoreductases in product-forming reactions, and, during the conversion of D-glucose into D-fructose, at first an enzymatically catalyzed reduction to D-sorbitol and, subsequently, an enzymatically catalyzed oxidation of D-sorbitol to D-fructose are performed, wherein NAD⁺/NADH and NADP⁺/NADPH are used as redox cofactors and, as a result of at least two further enzymatically catalyzed redox reactions proceeding in the same reaction batch, one of the two redox cofactors accumulates in its reduced form and, respectively, the other one in its oxidized form, and

   B) D-fructose is converted into furan derivatives, and

   in stage A)

   - in the regeneration reaction reconverting the reduced cofactor into its original oxidized form, oxygen or a compound of general formula wherein R₁ represents a linear-chain or branched (C₁-C₄)-alkyl group or a (C₁-C₄)-carboxyalkyl group, is reduced, and

   - in the regeneration reaction reconverting the oxidized cofactor into its original reduced form, a (C₄-C₈)-cycloalkanol or a compound of general formula

   wherein R₂ and R₃ are independently selected from the group consisting of H, (C₁-C₆)-alkyl, wherein alkyl is linear-chain or branched, (C₁-C₆)-alkenyl, wherein alkenyl is linear-chain or branched and contains one to three double bonds, aryl, in particular C₆-C₁₂-aryl, carboxyl, or (C₁-C₄)-carboxyalkyl, in particular also cycloalkyl, in particular C₃-C₈-cycloalkyl, is oxidized.
2. A process according to claim 1, characterized in that at least one dehydrogenase is used for the conversion of D-glucose into D-fructose.
3. A process according to claim 2, characterized in that NAD⁺/NADH and NADP⁺/NADPH, either in a soluble form or immobilized onto solids, are used as redox cofactor(s).
4. A process according to any one of claims 1 to 3, characterized in that at least one redox cofactor is regenerated during the conversion of D-glucose into D-fructose in the same reaction batch by at least one further redox enzyme selected from alcohol dehydrogenases, NADH oxidases, hydrogenases, lactate dehydrogenases or formate dehydrogenases, under consumption of co-substrates.
5. A process according to claim 4, characterized in that co-substrates are selected from alcohols, lactic acid and salts thereof, pyruvic acid and salts thereof, oxygen, hydrogen and/or formic acid and salts thereof.
6. A process according to any one of claims 1 to 5, characterized in that the reaction in stage a) proceeds according to Reaction Scheme 3 wherein

   CtXR = xylose reductase from Candida tropicalis

   SISDH = sorbitol dehydrogenase from sheep liver

   LkADH = alcohol dehydrogenase from Lactobacillus kefir, NADP(H)-dependent

   LacDH = lactate dehydrogenase, NAD(H)-dependent

   or according to Reaction Scheme 4 wherein

   CtXR = xylose reductase from Candida tropicalis

   BsSDH = sorbitol dehydrogenase from Bacillus subtilis

   LkADH = alcohol dehydrogenase from Lactobacillus kefir, NADP(H)-dependent

   SmOxo = NADH oxidase from Streptococcus mutans.
7. A process according to any one of the preceding claims, characterized in that the D-fructose obtained according to stage A) of the present invention is isolated.
8. A process according to claim 7, characterized in that the D-fructose is isolated in a crystalline form.
9. A process according to any one of the preceding claims, characterized in that an acidic catalyst and a solvent are used in stage B).
10. A process according to claim 9, characterized in that N-methyl-2-pyrrolidone of formula is used as a solvent.
11. A process according to claim 10, characterized in that N-methyl-2-pyrrolidone is used either as a reaction solvent or as a co-solvent.
12. A process according to any one of the preceding claims, characterized in that the conversion of D-fructose into furan derivatives in stage B) is performed as a batch process or as a continuous process.
13. A process according to claim 12, characterized in that the conversion of D-fructose into furan derivatives in stage B) is performed under microwave heating.
14. A process according to any one of claims 9 to 13, characterized in that

    - a homogeneous acid catalyst,

    - a heterogeneous acid catalyst,

    - a Lewis acid catalyst,

    - an SILP catalyst,

    is used as the acidic catalyst during the conversion of D-fructose into furan derivatives in stage B).
15. A process according to claim 14, characterized in that sulphuric acid or hydrochloric acid is used as the homogeneous acid catalyst.
16. A process according to claim 14, characterized in that an ion exchanger, in particular a montmorillonite or an Amberlite, is used as the heterogeneous acid catalyst.
17. A process according to claim 14, characterized in that CrCl₂, AlCl₃ or SiO₂-MgCl₂ is used as the Lewis acid catalyst.
18. A process according to any one of the preceding claims, characterized in that the furan derivative is hydroxymethylfurfural of formula
19. A process according to any one of the preceding claims, characterized in that furan derivatives which are produced are converted further.
20. A process according to claim 19, characterized in that hydroxymethylfurfural is oxidized further into 2,5-furan dicarboxylic acid (FDCA) of formula which, optionally, is subjected to polymerization, in particular for the production of polyethylene furanoate.