Continuous processes for selective conversion of aldohexose-yielding carbohydrate to ethylene glycol using low concentrations of retro-aldol catalyst

Abstract

To provide a continuous catalytic method for producing ethylene glycol from a feed comprising a carbohydrate that yields aldose.SOLUTION: A method comprising the steps (a) continuously or intermittently supplying a feed comprising an aldose-yielding carbohydrate to a reaction zone containing a liquid medium having therein a heterogeneous, nickel-containing hydrogenation catalyst; (b) continuously or intermittently supplying to the reaction zone a homogeneous, tungsten-containing retro-aldol catalyst, wherein the concentration of solubilized tungsten compounds in the liquid medium in the reactor, calculated as tungsten atoms, is 200 to 1500 milligrams per liter; and (c) continuously or intermittently withdrawing from the reaction zone a raw product stream containing ethylene glycol.SELECTED DRAWING: None

Description

[Technical Field]

[0001] This invention relates to a continuous process for the selective conversion of carbohydrates to ethylene glycol using a retro-aldol reaction and hydrogenation of the intermediate to yield aldoses, and more particularly to such a process in which low concentrations of a homogeneous tungsten-containing retro-aldol catalyst are used. [Background technology]

[0002]

[0002] Ethylene glycol is a useful, mass-produced chemical that has a wide range of uses, both as a building block for other materials such as polyethylene terephthalate (PET) and because of its inherent properties, such as antifreeze. Demand for ethylene glycol is substantial, making it one of the largest organic chemicals produced in the world. It is currently made by a multi-step process starting with ethylene from a hydrocarbon feedstock.

[0003]

[0003] Proposals have been made to produce ethylene glycol from renewable resources such as carbohydrates. These alternative processes include catalytic routes such as hydrogenolysis of sugars and a two-catalyst process using a retro-aldol catalyst to produce intermediates from sugars that can be hydrogenated over a hydrogenation catalyst to produce ethylene glycol and propylene glycol. The former process is referred to herein as the hydrogenolysis process or hydrogenolysis pathway, and the latter process is referred to as the hydrogenation or retro-aldol process or hydrogenation or retro-aldol pathway. For ease of reference, the latter is referred to herein as the retro-aldol process or retro-aldol pathway. The terms "catalytic process" or "catalytic pathway" are intended to encompass both the hydrogenolysis pathway and the retro-aldol pathway. The term "Hcat," as used herein, is intended to encompass both the hydrogenolysis catalyst and the hydrogenation catalyst.

[0004] In the catalytic route, a carbohydrate (which may be a single carbohydrate or a mixture of carbohydrates) that produces aldoses or ketoses is fed to a reaction zone containing a catalyst in an aqueous medium. In the presence of elevated temperature and hydrogen, the carbohydrate is converted to ethylene glycol and / or propylene glycol. The hydrocracking process uses a hydrocracking catalyst and is typically at temperatures below about 225°C. In many instances, high conversion of carbohydrate can occur at temperatures below about 220°C. The hydrocracking route often uses low concentrations of carbohydrate fed to the reaction zone to minimize by-product production. The retro-aldol route is fundamentally different in that the carbohydrate is converted to an intermediate over a retro-aldol catalyst, which is then catalytically converted to ethylene glycol and / or propylene glycol over a hydrogenation catalyst. The desired initial retro-aldol reaction is endothermic and requires high temperatures, often above 230°C, to provide a sufficient reaction rate so that conversion of the carbohydrate to the intermediate occurs preferentially over hydrogenation of the carbohydrate to a polyol such as sorbitol.

[0005] The hurdles to achieving a commercially competitive process are considerable. High conversion of carbohydrates to ethylene glycol and saleable by-products alone is not sufficient to achieve competitiveness. The process must be continuous with long run times for stable operation. Because the retro-aldol route involves hydrogenation at high pressure, high throughput is desirable to minimize capital costs for reactors and related equipment. An additional operating cost for the retro-aldol route is the use of catalysts, both homogenous, retro-aldol catalysts and Hcat.

[0006] The retroaldol route to ethylene glycol is fraught with complications. A carbohydrate must undergo a retroaldol conversion to produce glycolaldehyde, which is then hydrogenated to ethylene glycol. The temperatures required for the catalytic retroaldol reaction are sufficient to cause other reactions of the carbohydrate. For example, glucose isomerization to fructose occurs, which predominantly forms propylene glycol in the retroaldol process. Glycolaldehyde is highly reactive and can react noncatalytically to, for example, 1,2-butanediol. For these reasons, the retroaldol reaction and hydrogenation reactions are carried out as close in time and proximity as possible, sometimes in the same reactor. In addition, the hydrogenation conditions can cause further reactions, including hydrogenation of the desired ethylene glycol product. The kinetics of some of the reactions that can occur during the retroaldol pathway were reported by Zhou et al. in "Ethylene Glycol Production from Glucose over W-Ru Catalysts: Maximizing Yield by Kinetic Modeling and Simulation," AIChE Journal, Vol. 63, No. 6, June 2017. Their main focus was on variations in temperature, glucose feed concentration, and feed rate. They stated that lower glucose concentrations were preferable for improving ethylene glycol yield, and therefore chose to conduct simulations and experimental validation using a 10% glucose feed concentration. From their simulations, they concluded that a low glucose feed rate is advantageous for obtaining a high ethylene glycol yield. Their simulations showed that increasing the reaction temperature from 453°K to 473°K or higher reversed the reaction selectivity from hexitol to ethylene glycol. Further conclusions by the authors include that gas production becomes significant when the reaction is carried out at high temperatures with extremely low glucose feed rates.

[0007] For a given production rate of ethylene glycol, a low glucose concentration and a low feed rate will require the use of a larger reactor volume than a higher feed rate and a higher glucose concentration in the feed. Lin et al., in U.S. Patent Application Publication No. 2017 / 0210687 A1, disclose a process for producing diols from sugars using a specific main catalyst and a soluble tungstate or tungsten compound. The main catalyst is said to be an acid-resistant alloy of nickel and various other components. Lin et al. provide various embodiments including glucose concentrations up to 60 wt.%. Embodiment 2, for example, uses a feed containing 50 wt.% glucose at a feed rate of 2 L / hour to a 10-liter reactor containing 6 liters of liquid medium. The main catalyst is present in an amount of 1000 grams, and a soluble tungstate catalyst (sodium tungstate) is provided in an amount of 2 wt.% of the feed. Lin et al. reported achieving a 71% ethylene glycol yield and 7% propylene glycol by weight with 3% butylene glycol by weight. They did not specifically report the yields of sugar alcohols (itols), such as sorbitol and glycerin. The large amounts of main catalyst and sodium tungstate used by Lin et al. appear to be related to the glucose concentration in the feed, as in other embodiments where a lower concentration of glucose is used and a smaller amount of catalyst is used. In Example 4, a 40% glucose feed by weight is used with 100 grams of tungsten trioxide and 500 grams of main catalyst. The ethylene glycol yield was 67% and the propylene glycol yield was 2%. In Example 10, the glucose concentration in the feed was 40% by weight, the main catalyst was used in an amount of 1500 grams, and sodium tungstate was used at 0.5% by weight. The ethylene glycol yield was 80% and the propylene glycol yield was 5%. While the work by Lin et al. demonstrates that it may be possible to obtain higher ethylene glycol yields using higher concentrations of glucose in the feed, it comes at the expense of using large amounts of catalyst, which affects the practical operation and economics of an industrial-scale plant.

[0008]

[0008] Schreck et al., in U.S. Patent No. 10,544,072 B2, disclose a continuous process for the highly selective conversion of aldohexose-bearing carbohydrates to ethylene glycol. Using a predetermined ratio of retro-aldol catalyst to hydrogenation catalyst and temperature, high conversion efficiency to ethylene glycol is obtained with minimal co-production of sorbitol and at least one of an ethylene glycol to propylene glycol weight ratio of at least about 15:1 and a glycerin to propylene glycol weight ratio of less than about 0.5:1. In the examples, the feed is about a 32 weight percent or 50 weight percent glucose solution. The hydrogenation catalyst used in most examples is a combination of nickel, rhenium, and boron supported on an extruded silica-alumina support. This type of hydrogenation catalyst contains 6.8 weight percent nickel. Subsequent studies in the patent application indicate that the silica-alumina support may collapse under the conditions employed in the examples. Two examples use a carbon-supported ruthenium catalyst. The retroaldol catalyst was either ammonium metatungstate, sodium metatungstate, or a combination of sodium metatungstate and sodium tungstate. In all examples using a nickel catalyst, the catalyst was provided in an amount of 6 grams per 100 milliliters (30 grams / liter), and the retroaldol catalyst was provided at 0.1 to 1.5 weight percent. Feed rates of approximately 1 to 2.5 milliliters / minute were disclosed in the examples. Selectivities to ethylene glycol as high as 86 weight percent were reported. The highest selectivities to ethylene glycol were achieved in examples containing 1 weight percent or more of the retroaldol catalyst. Generally, a feed rate of 1 milliliter / minute provided higher selectivities to ethylene glycol.

[0009] Schreck et al. demonstrated the ability to achieve high selectivity to ethylene glycol at glucose feed concentrations of 32 to 50 percent, and the amount of catalyst used was economical from an operational standpoint. However, the economics of a commercial plant would be improved by reducing the amount of catalyst used.

[0010]

[0010] De Vlieger et al., in International Publication No. 2020 / 055831, disclose a start-up process for producing glycols from sugars. They disclose the use of one or more agents to suppress tungsten precipitation during process start-up using a tungsten-containing retroaldol catalyst and a hydrogenation catalyst. The patent applicant states: "Homogeneous tungsten-based catalysts typically used in sugar-to-glycol processes can be prone to conversion to undesirable products, for example, by reduction and precipitation of the metal (tungsten). Precipitated solids in the reactor system can cause undesirable chemical and / or physical reactions between the tungsten metal and other chemical species present (e.g., catalyst poisoning) in addition to line blockage and clogging" (paragraph

[10] ).

[0011]

[0011] As examples of suitable substances, the patentee applicant lists sugar feed or products formed during the process, such as sorbitol, monoethylene glycol, monopropylene glycol, 1,2-butanediol, glycerol, or other sugar alcohols, aldehydes, ketones, and carboxylic acids. [Prior art documents] [Patent documents]

[0012] [Patent Document 1] U.S. Patent Application Publication No. 2017 / 0210687A1 [Patent Document 2] U.S. Patent No. 10,544,072B2 [Patent Document 3] International Publication No. 2020 / 055831 [Non-patent literature]

[0013] [Non-Patent Document 1] Zhou et al., Ethylene Glycol Production from Glucose over W-Ru Ca talysts: Maximizing Yield by Kinetic Modeling and Simulation, AIChE Journal, Vol. 63, No. 6, June 2017 Summary of the Invention [Problem to be solved by the invention]

[0014]

[0012] In accordance with the present invention, it has been discovered that the retro aldol route to ethylene glycol can utilize extremely low concentrations of retro aldol catalyst while still achieving high selectivity to ethylene glycol and high reactor throughput. In the disclosed process, these extremely low concentrations of retro aldol catalyst are used in combination with a hydrogenation catalyst having a specific activity, size, and spatial distribution to obtain high selectivity to ethylene glycol. The spatial distribution of the hydrogenation catalyst reduces the risk of hydrogen starvation. Furthermore, the concentration of catalytic metal for hydrogenation can often be lower than previously preferred for the retro aldol route to ethylene glycol, which reduces the cost and facilitates the practical operation of commercial-scale facilities for the retro aldol route to convert aldoses, particularly glucose, to ethylene glycol. [Means for solving the problem]

[0015] The retroaldol catalyst used in the disclosed process is a homogeneous tungsten-containing catalyst, or a precursor thereof, introduced continuously or intermittently into a reaction zone containing a heterogeneous hydrogenation catalyst comprising nickel, preferably nickel on a low-surface-area, substantially inert support. The disclosed process utilizes a catalyst with relatively low hydrogenation activity, e.g., a nickel-containing catalyst with controlled hydrogenation activity supported on individual particles, in combination with an extremely low retroaldol catalyst concentration, while still providing sufficient activity and selectivity for commercial-scale operation. The retroaldol catalyst can be added together with the aldose-producing feed, separately, or a combination of both. The retroaldol catalyst addition rate is typically such that the concentration of solubilized tungsten compounds, calculated as tungsten atoms, in the liquid medium in the reactor is about 200 to 1500, preferably about 300 to 1200 milligrams per liter. At low concentrations of solubilized tungsten compounds, the amount of solubilized tungsten compounds leaving the reactor is reduced, thereby improving the economics of the process. Without wishing to be limited by theory, it is believed that under reaction conditions, the contamination of the hydrogenation catalyst by the accumulation of tungsten compounds is weakened or stabilized, or even reversed in some cases. Furthermore, the replacement rate of the nickel-containing catalyst can be reduced, allowing for further economic benefits to be realized.

[0016] According to a particular embodiment, a catalytic process for producing ethylene glycol from a feed containing aldose-yielding carbohydrates comprises: (a) continuously or intermittently supplying said feed to a reaction zone containing a liquid medium having a heterogeneous nickel-containing hydrogenation catalyst therein, said feed being supplied at a rate of at least about 50 grams / hour, preferably at least about 100 grams / hour of carbohydrate per liter of liquid medium, said liquid medium being under catalytic conversion conditions comprising the presence of dissolved hydrogen, a temperature of at least about 235°C, a pH greater than 3, and a residence time sufficient to react at least 99 weight percent of the carbohydrate to produce aldoses; (i) the heterogeneous hydrogenation catalyst has a maximum particle dimension of less than about 100 micrometers, preferably less than about 50 micrometers, and is preferably a supported nickel-containing catalyst, more preferably a supported nickel-containing catalyst stabilized with at least one of rhenium and iridium, the support being an inert support, preferably having a surface area of less than about 100 square meters per gram, more preferably less than about 50 square meters per gram, and containing less than about 10 weight percent nickel (calculated as elemental nickel); (ii) dispersing the hydrogenation catalyst in a liquid medium in an amount of less than about 100 grams per liter, thereby providing a spatial relationship between catalytically active hydrogenation sites in said liquid medium; (b) continuously or intermittently feeding a homogeneous tungsten-containing retro aldol catalyst to the reaction zone, wherein the concentration of solubilized tungsten compounds, calculated as tungsten atoms, in the liquid medium in the reactor is about 200 to 1500 milligrams per liter, and the relative amounts of hydrogenation catalyst and retro aldol catalyst are sufficient to provide a cumulative conversion efficiency of the aldose-containing carbohydrate to ethylene glycol of at least 75 percent over a duration of 100 hours under catalytic conversion conditions; and (c) continuously or intermittently withdrawing from the reaction zone a raw product stream containing ethylene glycol.

[0017] Preferably, the hydrogenation catalyst is present in the liquid medium in an amount to provide 0.1 to 5 grams per liter of nickel (calculated as elemental nickel) in the reaction zone. Preferably, only a portion of the nickel in the hydrogenation catalyst is catalytically active to control the hydrogenation activity in the reaction zone as well as the hydrogenation activity of the catalyst particles. In most cases for supported catalysts, less than about 50 percent, and sometimes less than about 35 percent, of the nickel in the hydrogenation catalyst (calculated as elemental nickel) is in the zero-valent state.

[0018] Preferably, the carbohydrate-containing feed is introduced into the reaction zone in a spatially dispersed manner while still providing a commercially attractive loading per unit volume of the reaction zone, e.g., at least 100 or 150 grams, and sometimes at least about 200 or 250 grams, of carbohydrate per liter of liquid medium. Thus, the feed may be introduced at two or more locations in the reaction zone, or the feed may be diluted, for example, with at least one solvent or by recycling the product stream. Spatial dispersion of the carbohydrate helps to avoid localized regions around the hydrogenation catalyst where the hydrogen demand for hydrogenation exceeds the hydrogen supply. Preferably, the carbohydrate is mixed with a liquid, and then this mixture is introduced into the reaction zone. The mass ratio of carbohydrate to liquid is often less than about 0.6:1, e.g., about 0.05:1 to 0.5:1, and sometimes about 0.1:1 to 0.4:1.

[0019]

[0017] While multiple embodiments are disclosed, other embodiments of the present disclosure will be apparent to those skilled in the art from the following detailed description, which shows and describes illustrative aspects of the invention. As will be understood, the present disclosure is capable of modifications in various obvious forms, all without departing from the spirit and scope of the present disclosure. Accordingly, the drawings and detailed description are to be regarded as illustrative in nature and not as restrictive. DETAILED DESCRIPTION OF THE INVENTION

[0020] All patents, published patent applications and articles referenced herein are hereby incorporated by reference in their entirety. definition

[0019] As used herein, the following terms have the meanings set forth below unless otherwise specified or apparent from the context in which they are used.

[0021]

[0020] Where ranges are used herein, only the endpoints of the range are recited to avoid the need to lengthily list each and every value included in the range. Any suitable intermediate values and ranges between the recited endpoints may be selected. As an example, if a range of 0.1 to 1.0 is recited, all intermediate values (e.g., 0.2, 0.3, 6.3, 0.815, etc.) are included, as are all intermediate ranges (e.g., 0.2 to 0.5, 0.54 to 0.913, etc.).

[0022]

[0021] Use of the terms "a" and "an" is intended to include one or more of the described elements.

[0022] Mixing or mixed means the formation of a physical combination of two or more elements, such a combination may be homogeneous throughout or may have a heterogeneous composition, examples of which include, but are not limited to, solid mixtures, solutions, and suspensions.

[0023] Aldoses contain only a single aldehyde group (-CH=O) per molecule and have the general chemical formula C n (H2O) n

[0039] Non-limiting examples of aldoses include aldohexoses (all sugars containing an aldehyde having six carbons, examples of which include glucose, mannose, galactose, allose, altrose, idose, talose, and gulose); aldopentoses (all sugars containing an aldehyde having five carbons, examples of which include xylose, lyxose, ribose, and arabinose); aldotetroses (all sugars containing an aldehyde having four carbons, examples of which include erythrose and threose), and aldotrioses (all sugars containing an aldehyde having three carbons, examples of which include glyceraldehyde).

[0024] Aldose-yielding carbohydrates refer to aldoses or disaccharides or polysaccharides that can yield aldoses upon hydrolysis. For example, sucrose is a carbohydrate that also yields ketoses upon hydrolysis but still yields aldoses.

[0025] Aqueous and aqueous solutions mean that water is present, but need not be the predominant component. For illustrative purposes, but not limited to, a solution of 90 volume percent ethylene glycol and 10 volume percent water would be an aqueous solution. Aqueous solutions include liquid media containing dissolved or dispersed components, such as, but not limited to, colloidal suspensions and slurries.

[0026]

[0026] Biologically derived carbohydrate feedstock means a product containing carbohydrates that are sourced in whole or in substantial part from, derived from, or synthesized from biological products or renewable agricultural materials (including, but not limited to, plants, animals, and marine materials) or forestry materials.

[0027]

[0027] Initiating contact means that a fluid begins to contact a component, e.g., a medium containing a homogeneous or heterogeneous catalyst, but it is not necessary that every molecule of the fluid contact the catalyst.

[0028] The composition of the solution is determined using gas chromatography for the lower boiling components, typically those with three or fewer carbons and normal boiling points below about 300°C, and high performance liquid chromatography for the higher boiling components, typically those with three or more carbons and those that are thermally unstable.

[0029] The efficiency of conversion of aldohexose to ethylene glycol is reported as a percentage by mass and is calculated by dividing the mass of ethylene glycol contained in the product solution by the mass of aldohexose theoretically provided by the carbohydrate feed, i.e., the mass including all aldohexoses themselves contained in the carbohydrate feed and aldohexoses theoretically produced upon hydrolysis of all disaccharides or polysaccharides contained in the carbohydrate feed.

[0030] Cumulative conversion efficiency refers to the combined conversion efficiency over a given period of time in terms of the mass of aldose-producing carbohydrates supplied during that period and the mass of, for example, ethylene glycol produced during that period. Measurements can be made by any suitable means, provided that they take into account all carbohydrates and ethylene glycol produced over such period.

[0031] Dispersed means separated in all directions. A dispersed hydrogenation catalyst is one having a liquid medium surrounding particles containing both dissolved hydrogen and hydrogenatable organic compounds.

[0032] Hexitol is CH 14 It refers to a six-carbon compound with the empirical formula O6, which has one hydroxyl per carbon.

[0033] High shear mixing involves providing fluids moving at different velocities relative to adjacent regions, which can be accomplished via stationary or moving mechanical means to create shear and promote mixing. As used herein, the components subjected to high shear mixing may be immiscible, partially immiscible, or miscible. Hydraulic distribution refers to the distribution of an aqueous solution in a vessel, including contact with any catalyst contained therein.

[0033]

[0034] Hydrogen starvation means that the demand for molecular hydrogen for the desired catalytic hydrogenation exceeds the mass transfer capacity for a timely supply of molecular hydrogen. Hydrogen starvation can lead to the production of partially hydrogenated compounds, the formation of organic acids, and the transfer of hydrogen from one organic compound to another.

[0034]

[0035] Immediately before means that there are no intervening unit operations requiring a residence time greater than 1 minute.

[0036] Intermittent means from time to time, and the time intervals may be regular or irregular.

[0035]

[0037] Sugar alcohols (itols) refer to hydrocarbons substituted only at each carbon atom with hydroxyl moieties, such as sorbitol, mannitol, and glycerin.

[0036]

[0038] Ketose refers to a monosaccharide that contains one ketone group per molecule.Non-limiting examples of ketose include ketohexose (all sugars that contain ketones with six carbons, examples of which include fructose, psicose, sorbose, and tagatose), ketopentose (all sugars that contain ketones with five carbons, examples of which include xylulose and ribulose), ketotetrose (all sugars that contain ketoses with four carbons, examples of which include erythrulose), and ketotriose (all sugars that contain ketoses with three carbons, examples of which include dihydroxyacetone).

[0037]

[0039] Liquid medium refers to the liquid in the reactor. This liquid is a solvent for the carbohydrate, intermediates, and products, as well as a solvent for the homogeneous tungsten-containing retroaldol catalyst. Typically, and preferably, such liquid contains at least some water, and is therefore referred to as an aqueous medium.

[0038]

[0040] Hydrogenatable organics ("HOCs") are oxygen-containing hydrocarbons that can be hydrogenated to one or more products under process conditions. HOCs include, but are not limited to, sugars and other ketones and aldehydes, and hydroxyl-containing hydrocarbons such as alcohols, diols, and sugar alcohols.

[0039]

[0041] The pH of an aqueous solution is determined at ambient pressure and temperature. For example, in determining the pH of an aqueous hydrogenation medium or product solution, the liquid is cooled and allowed to stand at ambient pressure and temperature for 2 hours before determining the pH. If the solution for which a pH measurement is desired contains less than about 50 weight percent water, water is added to the solution to provide more than 50 weight percent water. For consistency, dilutions of the solution are made to provide the same weight percent water.

[0040]

[0042] pH control agent means one or more of a buffer and an acid or a base.

[0043] Sufficient pressure to maintain at least partial hydration of the carbohydrate means sufficient pressure to maintain sufficient moisture of hydration of the carbohydrate to prevent caramelization reactions. At temperatures above the boiling point of water, the pressure is sufficient to allow moisture of hydration to be retained on the carbohydrate.

[0041]

[0044] Rapid diffusion mixing is mixing in which at least one of the two or more fluids to be mixed is finely divided to facilitate mass transfer and form a substantially homogeneous composition.

[0042]

[0045] The reaction medium is a liquid phase under catalytic conversion conditions capable of dissolving the carbohydrate-containing feed.

[0046] The reactor may be one or more vessels in series or parallel, and the vessel may contain one or more zones. The reactor may be of any suitable design for continuous operation, including, but not limited to, tanks and pipe or tubular reactors, optionally with fluid mixing capabilities. Reactor types include, but are not limited to, laminar flow reactors, fixed bed reactors, slurry reactors, fluidized bed reactors, moving bed reactors, simulated moving bed reactors, trickle bed reactors, bubble column reactors, cavitation reactors, and loop reactors.

[0043]

[0047] Soluble means capable of forming a single liquid phase or a colloidal suspension in the reaction medium.

[0048] The solubilized tungsten compound is a tungsten compound that is dissolved in the reaction medium or that is colloidally suspended.

[0044]

[0049] With respect to nickel catalysts, stabilized means that one or more of rhenium and iridium are combined with nickel in an amount sufficient to reduce the solubility of nickel in the reaction medium compared to nickel without rhenium and iridium. Stabilized nickel may or may not be alloyed. Rhenium and / or iridium may, in some cases, affect the catalytic performance of the nickel catalyst.

[0045] Carbohydrate Feed

[0050] The disclosed process uses a carbohydrate feed containing aldohexose-yielding carbohydrates. When a product solution containing a high weight ratio of ethylene glycol to propylene glycol is desired, the carbohydrates in the feed comprise at least about 90 weight percent, preferably at least about 95 or 99 weight percent, of aldohexose-yielding carbohydrates. Often, the carbohydrate feed contains a carbohydrate polymer, such as starch, cellulose, or a partially hydrolyzed fraction of such a polymer, or a mixture of polymers, or a mixture of polymers with a partially hydrolyzed fraction.

[0046]

[0051] Most biologically-sourced carbohydrate feedstocks yield glucose when hydrolyzed. The disclosed process can be effectively used to convert glucose and glucose precursors to ethylene glycol. Glucose precursors include, but are not limited to, maltose, trehalose, cellobiose, kojibiose, nigerose, isomaltose, β,β-trehalose, α,β-trehalose, sophorose, laminaribiose, gentiobiose, and mannobiose. It is also acceptable to have glucose as the majority or only reactive component of the carbohydrate feed. Of course, other aldoses may be used in the disclosed process. Other carbohydrate polymers and oligomers, such as hemicellulose, partially hydrolyzed forms of hemicellulose, disaccharides such as sucrose, lactulose, lactose, turanose, maltulose, palatinose, gentiobiulose, melibiulose, and melibiulose, or combinations thereof, may also be used. However, these properties can result in various mixtures of ethylene glycol and propylene glycol.

[0047]

[0052] The carbohydrate feed may be solid, preferably in the form of a liquid suspension, or dissolved in a solvent such as water. When the carbohydrate feed is in a non-aqueous environment, it is preferred that the carbohydrate be at least partially hydrated. Non-aqueous solvents include alkanols, diols and polyols, ethers, or other suitable carbon compounds having 1 to 6 carbon atoms. Solvents include mixed solvents, particularly mixed solvents containing water and one of the aforementioned non-aqueous solvents. Certain mixed solvents may have a higher concentration of dissolved hydrogen under the conditions of the hydrogenation reaction, thereby reducing the possibility of hydrogen depletion. Preferred non-aqueous solvents are those that can be hydrogen donors, such as isopropanol. Often, these hydrogen donor solvents contain hydroxyl groups that, upon donating a hydrogen atom, are converted to carbonyls, which can be reduced under the conditions in the reaction zone. Most preferably, the carbohydrate feed is provided in the form of an aqueous solution. In either case, the volume of the feed and the volume of the withdrawn raw product must be balanced to produce a continuous process.

[0048]

[0053] A further consideration in providing carbohydrate to the reaction zone is minimizing energy and capital costs. For example, in steady-state operation, the solvent contained in the feed exits the reaction zone with the raw product and must be separated to recover the desired glycol product.

[0049]

[0054] Preferably, the feed is introduced into the reaction zone in a manner that avoids inappropriate concentrations of HOC, which can cause hydrogen starvation. The use of a greater number of carbohydrate supply arrangements per unit volume of the reaction zone allows for a greater concentration of carbohydrate in the feed. Generally, the mass ratio of water to carbohydrate in the carbohydrate feed is preferably in the range of 4:1 to 1:4. Sometimes, aqueous solutions of certain carbohydrates, such as dextrose and sucrose, at concentrations of 600 grams per liter or higher are commercially available.

[0050]

[0055] In some cases, recycled hydrogenation solution substantially free of hydrogenation catalyst, or an aliquot or separated portion thereof, is added as a component to the carbohydrate feed. The recycled hydrogenation solution may be one or more of a portion of the raw product stream or an internal recycle from which the hydrogenation catalyst has been removed. Suitable solids separation techniques include, but are not limited to, filtration and density separation, such as cyclones, vane separators, and centrifugation. This recycle allows the carbohydrate to be fed at a rate sufficient to maintain high conversion per unit volume of the reaction zone while reducing the amount of fresh solvent required for the feed. The use of recycle, especially when the recycle is a fractionated portion of the raw product stream, allows for low carbohydrate concentrations in the reaction zone while maintaining high conversion of carbohydrate to ethylene glycol. Additionally, it is feasible to maintain the recycle stream at or near the temperature of the reaction zone because it contains a tungsten-containing catalyst, allowing retro-aldol conversion to occur before the feed enters the reaction zone. With the use of recycled hydrogenation solution, the mass ratio of carbohydrate to the recycled total product stream and added solvent is often in the range of about 0.05:1 to 0.4:1, and sometimes about 0.1:1 to 0.3:1. The recycled raw product stream is often about 20 to 80 volume percent of the product stream.

[0051]

[0056] In some cases, the introduction rate of the recycled hydrogenation stream is such that such stream can be used as a driving fluid for an injector or eductor to introduce small bubbles or microbubbles of hydrogen into the reaction zone at the point of carbohydrate introduction, thereby reducing the possibility of hydrogen starvation. For example, the recycled hydrogenation stream may be sent to a venturi mixer to which hydrogen is fed. The hydrogen may be supplied at least partially from the headspace of the reaction zone; furthermore, if a venturi mixer is used, a vacuum may be used to draw hydrogen from the headspace into the mixer. The mixture is then sent to one or more injectors in the reaction zone. The injectors may be jet mixers / aerators or slot injectors, such as those disclosed in U.S. Pat. No. 4,162,970. Injectors, particularly slot injectors, can operate over a wide range of liquid and gas flow rates, thereby significantly reducing gas transfer capacity. The energy required to provide microbubbles of a given size is often lower than the energy required to form microbubbles of that size using a microbubble sparger. The bubble size produced by the injector is expected to be affected by, among other factors, the rate of liquid flow through the injector and the ratio of hydrogen to reaction products passing through the injector. Preferably, the hydrogen introduced by the injector is in the form of microbubbles having diameters ranging from 0.01 to 0.5 millimeters, preferably 0.02 to 0.3 millimeters. The microbubbles serve to enhance the mass transfer rate of hydrogen to the liquid medium in the reaction zone.

[0052]

[0057] The carbohydrate contained in the carbohydrate feed is provided at a rate of at least 50 or 100 grams per hour, preferably about 150 to 500 grams per hour per liter of reactor volume. Optionally, a separate reaction zone containing a retro-aldol catalyst but essentially free of hydrogenation catalyst may be used.

[0053] Retro-aldol catalyst

[0058] The disclosed process uses a homogeneous tungsten-containing retroaldol catalyst. The homogeneous catalyst source can be a solid or a dissolved or suspended tungsten compound. When added as a solid, the tungsten compound can be added as the compound itself, in a mixture with other materials, or supported on a support. Tungsten compounds include oxides, sulfates, phosphides, nitrides, carbides, halides, acids, and the like. Examples also include tungsten carbide, soluble phosphotungsten, and tungsten oxide supported on zirconia, alumina, and alumina-silica. Preferred catalysts are provided by soluble tungsten compounds and mixtures of tungsten compounds. Soluble tungstates include, but are not limited to, ammonium and alkali metal paratungstates, such as sodium and potassium paratungstate, partially neutralized tungstic acid, ammonium and alkali metal metatungstates, and ammonium and alkali metal tungstates. The presence of ammonium cations often leads to the formation of undesirable amine by-products in lower glycol products. While not wishing to be bound by theory, the catalytically active species may or may not be the same as the soluble tungsten compound introduced as the catalyst. Rather, the catalytically active species may form as a result of exposure to retro-aldol reaction conditions. Tungsten-containing complexes are typically pH-dependent. For example, a solution containing sodium tungstate at a pH greater than 7 is expected to produce sodium metatungstate when the pH is lowered. The form of the complexed tungstate anion is generally pH-dependent. The rate at which complexed anions formed from the condensation of tungstate anions are formed is affected by the concentration of the tungsten-containing anions.

[0054]

[0059] Preferred retro aldol catalysts include ammonium tungstate or alkali metal tungstates partially neutralized with an acid, preferably an organic acid having 1 to 6 carbon atoms, such as, but not limited to, formic acid, acetic acid, glycolic acid, and lactic acid. Partial neutralization often results in approximately 25 to 75% of the tungstate cations being acid sites, i.e., an average of 25 to 75%. Partial neutralization may be carried out before introducing the tungsten-containing compound into the reactor, or with the acid contained in the reactor. While not wishing to be bound by theory, it is believed that the spatial spacing between tungstate anions in the reaction medium and the time required to achieve equilibrium with polytungsten complexes inhibit the formation of polytungsten oxide complexes, such as metatungstate and paratungstate. However, typically, partially neutralized tungstates are rapidly formed, providing retro aldol activity while diminishing the catalytic activity of, for example, sodium tungstate, which catalyzes the isomerization of aldoses to ketoses.

[0055] hydrogenation catalyst

[0060] The disclosed process employs a heterogeneous hydrogenation catalyst. The hydrogenation catalyst uses nickel and has specific properties. Without wishing to be limited by theory, the hydrogenation catalyst may catalyze the conversion of products such as ethylene glycol into less desirable materials, such as ethanol or ethane, in addition to catalyzing the conversion of intermediates to ethylene glycol. Nickel is a relatively weak hydrogenation catalyst and has therefore been used to convert carbohydrates to ethylene glycol.

[0056]

[0061] Nickel catalysts can be supported or unsupported. Unsupported nickel catalysts include Raney nickel catalysts and nickel alloys, and may contain promoters, modifiers, and other adjuvants. As discussed below, the size and catalytic activity of nickel-containing unsupported catalysts are important. Controlled activation or selective poisoning can be used to provide the desired hydrogenation catalytic activity.

[0057]

[0062] The supported hydrogenation catalysts used in the disclosed processes contain relatively little nickel, calculated as elemental nickel, e.g., less than about 10 weight percent, and in some cases less than about 5 weight percent, e.g., about 0.1 to 2.5 weight percent nickel. Thus, each catalyst particle has a tailored hydrogenation activity that is believed to reduce the catalyzed reaction of products to less desirable materials and further reduce the risk of hydrogen starvation resulting in acid production.

[0058]

[0063] Not all nickel in a catalyst is catalytically active. The activity of nickel-containing catalysts is expected to be influenced in part by the portion of nickel on the catalyst that is in the zero-valent state, as well as by the configuration of nickel particles on the support and by blockage of catalytically active nickel by the support or deposits. Nickel activity and / or stability may also be affected by the inclusion of catalytic promoters, modifiers, and other adjuvants. Catalyst activity also typically decreases with use. Generally, about 1 to 50 percent, preferably about 20 to 50 percent, and in some cases about 10 to 35 percent of the nickel atoms in the catalyst are in the zero-valent state. The size of the nickel deposits on the support can also affect the portion of catalytically active nickel. Larger particles have a smaller surface area per unit mass than smaller particles.

[0059]

[0064] The absolute amount of catalytically active species in the volume of the reaction zone can also be adjusted by treating the catalyst or its environment and / or by ameliorating issues of catalyst maldistribution in the reaction zone. The absolute amount of catalytically active species in the catalyst is based on the effectiveness of the catalyst, not simply the overall mass of the catalyst. Therefore, physical process conditions such as maldistribution in the reaction zone and in situ catalyst conditions that cause catalyst losses, such as loss of catalytic species due to coating, poisoning, particle growth, and solubilization, are reflected in the absolute amount of catalytically active species available for reaction. For example, the formation of deposits on the catalyst reduces the available catalytic sites. Removing all or part of the deposits increases the number of available catalytic sites, thereby increasing the absolute amount of catalytically active species in the catalyst. In some cases, catalytic species in the catalyst become oxidized or reduced. For example, nickel used in hydrogenation catalysts may become oxidized and no longer catalytically active. Reduction of nickel oxide is expected to increase the amount of catalytically available nickel, thus increasing the absolute amount of catalytically active species in the catalyst. Similarly, catalytic species may be oxidized or reduced, or converted to more or less active species.

[0060]

[0065] Hydrogenation catalysts, particularly supported catalysts, preferably contain stabilizers such as rhenium and iridium, and other adjuvants. The stabilizers are often present in an amount of about 1-20 atomic percent of the nickel, sometimes about 3-15 atomic percent (all calculated as elemental metal). Without wishing to be bound by theory, the stabilizers form alloys or other interactions with the nickel to reduce its particle growth rate, oxidation, and solubility in the reaction medium, thereby extending the useful life of the hydrogenation catalyst. The stabilizers may also have other effects that increase the effectiveness of the catalyst. The catalyst may also contain other adjuvants, such as promoters such as boron.

[0061]

[0066] The nickel-containing catalyst has a low surface area, e.g., less than 50 square meters per gram, preferably less than 10 square meters per gram (as measured by BET). Without wishing to be bound by theory, it is believed that the catalytic sites within the catalyst are prone to hydrogen starvation and therefore to the production of acids and other less desirable products that reduce the selectivity of the conversion to ethylene glycol. Generally, the larger the support, the more desirable it is to have a smaller surface area. For supported catalysts, preparation techniques that selectively place nickel at or near the surface would be advantageous, especially for supports with larger surface areas. The maximum dimension of the catalyst is less than 100 micrometers, sometimes less than about 50 micrometers. The aspect ratio (maximum dimension to thickness) is often 1:10 to 1:1. In some cases, the smallest dimension from the surface to the center of the catalyst is less than about 20 micrometers.

[0062]

[0067] When used, supports for hydrogenation catalysts are substantially inert in the reaction medium under process conditions. In many cases, the conditions of the reaction medium are detrimental to the structural stability of many conventional supports, such as silica-containing supports and silica-alumina-containing supports. Supports include alumina, preferably alpha-alumina; zirconia; carbon, including but not limited to, graphene; and refractory metals, including but not limited to, molybdenum, tantalum, and tungsten. Supports include surface coatings on substrates.

[0063]

[0068] Supported hydrogenation catalysts can be prepared by any suitable technique known in the art. A frequently used technique involves depositing a soluble compound of the catalytic metal on the support, drying, and then calcining. Incipient wetness techniques may also be used. In some cases, incipient wetness techniques result in a greater concentration of catalytic metal toward the outside of the support, where a solution containing the dissolved catalytic metal is drawn out of the pores by drying. Optionally, techniques may be used such as filling the interior of the support with a water-insoluble liquid and then depositing the catalytic metal from an aqueous solution at the outer region of the support. The water-insoluble liquid may then be removed, for example, during drying or calcination. In some cases, a chelating agent may be used in the mother liquor used to prepare the catalyst. It is believed that the chelating agent promotes more uniform deposition of nickel across the surface area of the support. Chelating agents include, but are not limited to, oxalic acid, citric acid, ethylenediaminetetraacetic acid, ethylenediamine, and phosphates and phosphonates. It is also contemplated that after deposition of the catalytic metals on the support, the catalyst may be crushed or broken down to smaller particle sizes to reduce the possibility of hydrogen starvation.

[0064]

[0069] Calcination, if used, typically converts the catalytic metal to an oxide, and activation at elevated temperatures in the presence of hydrogen provides nickel in the zero-valent state. In some cases, because the solubility of nickel oxide in aqueous acidic solutions at elevated temperatures is low, it is desirable to carry out activation for a time, temperature, and hydrogen pressure sufficient to ensure that at least about 20 percent, e.g., 25-35 percent or 50 percent, of the nickel is in the zero-valent state.

[0065]

[0070] Although nickel is a relatively mild hydrogenation catalyst, if necessary, the hydrogenation catalyst may be treated before use to reduce the catalyst's hydrogenation activity. This reduction can be achieved by any suitable technique, such as selective sulfiding and coking, which are well known in the art. One technique involves contacting the heterogeneous catalyst with a solution of a water-soluble tungstate-containing compound and reducing the pH to precipitate tungsten compounds on the catalyst. The pH is typically reduced by at least 0.5, and sometimes by at least 2, to provide the desired amount of deactivation. The precipitated tungsten compounds can partially block active nickel sites and reduce the hydrogenation activity by blocking or partially blocking the interior of the catalyst, easily causing hydrogen starvation.

[0066] Process Conditions

[0071] In the disclosed process, the combination of reaction conditions (e.g., temperature, hydrogen partial pressure, catalyst concentration, water flow distribution, and residence time) is sufficient to convert at least about 99 mass percent, and sometimes essentially all, of the carbohydrate to yield aldoses. In summary, the process is complex because the carbohydrate feed must be subjected to retro-aldol conditions while minimizing isomerization to ketoses, and then the intermediates from the retro-aldol conversion, particularly glycolaldehyde, must be subjected to hydrogenation before they undergo other reactions to form less desirable products. If the retro-aldol and hydrogenation conversions are not balanced, ethylene glycol production cannot be optimized.

[0067]

[0072] In this disclosure, a low concentration of homogeneous tungsten-containing catalyst is used. Therefore, the balance between retro-aldol conversion and hydroconversion may be affected. Nevertheless, the disclosed process allows for high selectivity to ethylene glycol; for example, when the carbohydrate yielding the aldose is glucose, at least about 75 percent, preferably at least about 80 percent, of the glucose is converted to ethylene glycol. It is understood that if the carbohydrate feed contains ketoses or carbohydrates that do not have the potential to be completely converted to glycolaldehyde, the selectivity to ethylene glycol will be reduced proportionately.

[0068]

[0073] The rate of addition of the tungsten-containing catalyst is sufficient to provide solubilized tungsten compounds to the liquid medium of the reactor, and is about 200 to 1500 milligrams / liter, preferably about 300 to 1200 milligrams / liter (calculated as elemental tungsten). It is well understood that tungsten can form precipitates that can lead to the accumulation of solids, including hydrogenation catalysts, on surfaces. See, for example, International Publication No. 2020 / 055831. At low concentrations of solubilized tungsten in the reactor used in accordance with this disclosure, the deposition of tungsten precipitates is stabilized, and in some cases, an equilibrium between solubilized and precipitated tungsten compounds is achieved. Varying or cycling the rate of addition of tungsten compounds to the reactor, including stopping the addition of tungsten compounds, is within the scope of this disclosure, thereby varying the concentration of solubilized tungsten in the liquid medium from a higher concentration to a lower concentration. At lower concentrations, at least a portion of the tungsten precipitates on the surfaces are removed. The lower concentration of solubilized tungsten in the liquid medium is often at least about 100 milligrams / liter lower than the higher concentration, and sometimes the lower concentration is in the range of about 10 to 700 milligrams / liter, e.g., 50 to 500 milligrams / liter (calculated as elemental tungsten). Generally, the solubilized tungsten from the deposit provides sufficient solubilized tungsten so that the conversion of the carbohydrate feed to ethylene glycol is relatively unimpeded.

[0069]

[0074] The equilibrium between solubilized and deposited tungsten compounds is often affected by pH, with higher pH values shifting the equilibrium more toward solubilized tungsten. Therefore, pH control agents such as hydroxides and carbonates may be used, if necessary, especially when reducing or halting the tungsten compound addition rate for the purpose of removing precipitated tungsten. In an example where the pH is increased to remove tungsten deposits, the pH is increased by at least about 0.5 pH units. If used, reductions in tungsten compound addition may be intermittent or periodic. If the solubilized tungsten concentration in the reactor is maintained at a higher level, e.g., in the range of 1000 to 1500 milligrams per liter, at least one cycle of lower tungsten compound addition every 250 hours is used. The lower rate of tungsten compound addition is often about 0 to 50 or 75 percent of the higher rate. The duration of the reduced tungsten compound addition rate is often about 0.1 to 24 hours or longer, but preferably not so long as to cause a 5 percent or greater decrease in selectivity to ethylene glycol. In some cases, cycling is performed only when a loss of hydrogenation catalyst activity is observed. One indicator of loss of hydrogenation catalyst activity is a decrease in the pH of the liquid medium in the reactor, e.g., a decrease of at least 0.2 pH units. In other examples, the tungsten compound addition rate can be varied on a pre-established schedule. The cycle to a lower tungsten compound addition rate typically lasts from about 0.5 hours to 250 hours, e.g., from about 1 hour to 100 hours.

[0070]

[0075] The amount of hydrogenation catalyst used in the disclosed process can vary widely. For practical purposes, it is economically preferable to use a smaller amount of hydrogenation catalyst. By using a lower concentration of tungsten-containing retro aldol catalyst and thereby reducing the deposition of tungsten compounds on the hydrogenation catalyst, a smaller amount of hydrogenation catalyst can provide longer periods of activity. Typically, the hydrogenation catalyst is provided in an amount of nickel (calculated as elemental nickel) of less than 10 grams per liter, sometimes less than about 5 grams per liter, e.g., about 0.1 or 0.5 to 3 grams per liter of liquid medium in the reactor. As noted above, not all of the nickel in the catalyst is in the zero-valent state, or not all of the nickel in the zero-valent state is readily accessible to glycolaldehyde or hydrogen. Therefore, the optimal mass of nickel per liter of liquid medium for each specific hydrogenation catalyst is expected to vary. In some cases, a portion of the hydrogenation catalyst is continuously or intermittently withdrawn from the reactor and replaced with fresh or rejuvenated hydrogenation catalyst. This exchange helps to maintain a relatively constant hydrogenation activity within the reactor.

[0071]

[0076] In addition to the spatial relationship of the hydrogenation catalyst particles, the activity of the hydrogenation catalyst, especially in the region where the carbohydrate feed enters the reactor, is important for optimizing ethylene glycol selectivity. If the hydrogenation catalyst particles are too close together, the retro-aldol conversion may not be completed, thereby increasing the production of sorbitol and other sugar alcohols. If the hydrogenation catalyst particles are too far apart, reactions between intermediates may occur, producing undesirable products such as 1,2-butanediol. The activity of the hydrogenation catalyst particles is also a factor in optimizing selectivity to ethylene glycol. Too much activity increases the risk of hydrogen starvation, leading to further hydrogenation of ethylene glycol and propylene glycol to alcohols and hydrocarbons, in addition to the formation of less desirable products such as acids.

[0072]

[0077] In addition to the hydrogenation activity of each catalyst particle, the spatial relationship, size of the catalyst particle, and the concentration of organic matter available for hydrogenation ("HOC") all play a role in the local hydrogen demand in the reactor. Hydrogen starvation can occur when the hydrogenation potential exceeds the ability to supply hydrogen to a localized region of the hydrogenation catalyst. Hydrogen is typically diffusely soluble in the liquid reaction medium at process conditions. Increasing the hydrogen concentration can be achieved by increasing the partial pressure of hydrogen, by using a solvent or cosolvent with greater hydrogen solubility, or by using a hydrogen donor compound such as isopropanol in addition to molecular hydrogen to supply hydrogen. Hydrogen in the liquid medium undergoes mass transfer to the hydrogenation catalyst, but this mass transfer is primarily related to the driving force caused by concentration differences. The duration of the time required for mass transfer, combined with hydrogen depletion via catalytic hydrogenation, can result in localized regions around the catalyst particles that do not have enough hydrogen to supply the hydrogen demand of the hydrogenation catalyst, i.e., localized hydrogen-starved regions. Thus, using a less active hydrogenation catalyst and / or a smaller sized hydrogenation catalyst will result in a smaller demand for hydrogen in the local region around each hydrogenation catalyst particle. Similarly, a reduced concentration of HOCs in the local region around a catalyst particle will result in a smaller demand for hydrogen in that local region.

[0073]

[0078] As noted above, the hydrogenation catalyst used in the disclosed process has a major dimension of less than about 100 micrometers and a nickel concentration (calculated as elemental nickel) of less than about 5 weight percent. With low hydrogenation activity, the hydrogen demand in a given area around the catalyst particle is expected to be less than the demand that would exist for a larger catalyst particle with a higher nickel concentration, all other factors remaining the same. Thus, a low hydrogenation activity catalyst can result in more dispersed catalytic sites, thus reducing the risk of hydrogen starvation. An additional feature of the disclosed process, along with the ability to reduce the risk of hydrogen starvation, is that in some cases the pressure for the process can be lowered without unduly increasing the risk of hydrogen starvation. Lower pressures can reduce compression costs and, in some cases, capital costs. Typically, in a retro-aldol process, the pressure (absolute) typically ranges from about 15 to 200 bar (1500 to 20,000 kPa), e.g., about 25 to 150 bar (2500 to 15,000 kPa). Pressures in the range of 2500 to 12000 kPa may find use where the hydrogenation catalyst and carbohydrate feed are optimally dispersed.

[0074]

[0079] Other factors that affect the risk of hydrogen starvation are the concentration of HOCs and the uniformity of that concentration. For example, if a carbohydrate-containing feed is introduced into a localized region of the reactor, the retro-aldol catalyst, particularly the hydrogenation catalyst, in that region may become overloaded, resulting in a loss of selectivity to ethylene glycol, for example, by allowing glucose to bypass the aldol catalyst and be hydrogenated to sorbitol, or by increasing the demand for hydrogen for hydrogenation in that region. If excellent mixing of the feed in the liquid medium in the reactor is achieved and the feed is added at multiple points to avoid regions with excessive concentrations of HOCs, the low concentrations of retro-aldol catalyst used in the disclosed process can still handle high rates of carbohydrate feed. Generally, with excellent mixing and distribution of the carbohydrate feed, feed rates of up to 1 kilogram / hour of carbohydrate per liter of reactor can be achieved in some instances. The carbohydrate feed is at least 50 grams / hour of carbohydrate per liter, and often in the range of about 100 to 700 or 1000 grams / hour of carbohydrate per liter.

[0075]

[0080] The residence time in the reactor will depend in part on the reactor design and the concentration of catalyst in the reaction medium. Both the retroaldol conversion and the hydrogenation reactions are extremely rapid. As a result, productivity is limited by the ability to supply hydrogen to the hydrogenation catalyst to prevent hydrogen starvation. The residence time is typically about 1 minute to 5 hours, e.g., 5 to 200 minutes. In some cases, the weight hourly space velocity is about 0.01 or 0.05 to 1 hour, based on the total carbohydrates in the feed. -1 The residence time is sufficient for glycolaldehyde and glucose to be less than 0.1 weight percent of the reaction product, and most preferably less than 0.001 weight percent of the reaction product.

[0076]

[0081] Isomerization of aldoses to ketoses reduces the selectivity of conversion to ethylene glycol. Therefore, it is desirable to minimize conditions that promote isomerization, such as low pH. As discussed in U.S. Pat. No. 10,544,072, the retro-aldol reaction has a high activation temperature, and isomerization to ketoses can occur at lower temperatures. Therefore, the carbohydrate feed is preferably rapidly passed through a temperature zone ranging from 170° C. to 230° C., preferably at least about 240° C.

[0077]

[0082] The carbohydrate feed may be in the presence of other chemicals during heating. For example, hydrogen for hydrogenation may be at least partially supplied along with the carbohydrate feed. Optionally, other adjuvants, such as pH control agents, may also be present. In one embodiment, the carbohydrate feed contains a retroaldol catalyst, and in such instances, catalytic conversion of the carbohydrate to produce an aldohexose occurs during heating. The extent of conversion of the carbohydrate to produce an aldohexose during heating is expected to be affected by, among other things, the duration of heating, the relative concentrations of the carbohydrate and the retroaldol catalyst, and the activity of the retroaldol catalyst.

[0078]

[0083] Some tungsten-containing compounds that provide retro-aldol catalysts also act as isomerization catalysts. One such compound is sodium tungstate. When such tungsten-containing compounds are used, at least a portion of the compounds is preferably introduced separately, more preferably at a location distant from the area where the carbohydrate feed is introduced into the reactor. This allows the tungsten-containing compound to be converted into an active retro-aldol catalyst and further diluted. Furthermore, contact with the tungsten-containing compound is expected to occur at a temperature where the retro-aldol reaction occurs, thereby reducing the degree of isomerization.

[0079]

[0084] Because the carbohydrate feed is at least partially converted to intermediates before contact with any potential hydrogenation catalyst, it may be beneficial to add a retro-aldol catalyst to the carbohydrate feed in a pre-mixing zone before the carbohydrate feed is introduced into the reactor. The isomerization of aldoses to ketoses is promoted by a lower or higher pH environment. Therefore, maintaining the carbohydrate at a more or less neutral pH before introduction into the reactor can minimize the degree of isomerization. Therefore, when it is desirable to utilize at least a partial retro-aldol conversion in a pre-mixing zone, the tungsten-containing compound may be at least partially neutralized, such as partially neutralized tungstic acid, an alkali metal salt of partially neutralized tungstic acid, or ammonium tungstate.

[0080]

[0085] Heating of the carbohydrate feed can be achieved in any suitable manner, and one or more types of heating can be used. All, none, or some of the heating of the carbohydrate feed can be performed before the carbohydrate feed is introduced into the liquid medium. In embodiments in which the heated carbohydrate feed is maintained in contact with the retro-aldol catalyst in a premixing zone, the duration of such contact before introduction into the liquid medium is generally less than about 15 seconds, preferably less than about 10 seconds, and in some cases less than about 5 seconds. Typically, any retention time of the heated carbohydrate feed before introduction into the liquid medium is a result of the arrangement of equipment, such as piping distance, and the residence time in auxiliary equipment, such as a fluid distributor, from the heat exchange zone to the hydrogenation zone. As will be understood, turn-up and turn-down operations are expected to affect the inherent retention time.

[0081]

[0086] The rate of heating is expected to be affected by heat and mass transfer parameters. Generally, it is desirable to promote mixing of the carbohydrate feed during heating to facilitate both mass and heat transfer, thereby reducing the time required for the carbohydrate feed to completely pass through this temperature zone. This mixing may be carried out in any suitable manner, including, but not limited to, mechanical mixing, static mixing, and rapid diffusion mixing. The thoroughness of mixing can also affect the mass transfer of reactants, intermediates, catalysts, and products, and therefore the selectivity of the conversion to ethylene glycol and the rate of by-product formation.

[0082]

[0087] A particularly useful stream for direct heat exchange with the carbohydrate feed is the withdrawn product solution (recycle). When a soluble retro-aldol catalyst is used in the liquid medium, recycle results in substantial return of the retro-aldol catalyst to the reaction system. Recycle may be conducted at a temperature of at least about 180°C, e.g., in the range of about 230°C to 300°C. The mass ratio of recycle to carbohydrate feed will depend on the relative temperatures of the two streams and the desired combined temperature. When a recycle is used, the mass ratio of recycle to carbohydrate feed is often in the range of about 0.1:1 to 100:1. The recycle may be a fractionated portion of the withdrawn product solution, or may be subjected to unit operations to separate one or more components from the recycle stream, such as, but not limited to, degassing to remove hydrogen or filtration to remove any entrained heterogeneous catalyst. When the product solution is degassed to recover at least a portion of the hydrogen, the recycle is often a fractionated portion of the degassed product solution. One or more ingredients can be added to the recycle before combining with the carbohydrate feed during direct heat exchange operation. These ingredients include, but are not limited to, a retro-aldol catalyst, a pH control agent, and hydrogen. By using a recycled product solution, the combined carbohydrate feed and recycle can contain unreacted carbohydrates to produce aldoses, intermediates to ethylene glycol, and ethylene glycol. When a carbohydrate feed that is not in the form of an aqueous solution, such as a solid or melt, is used, the recycle provides water to dissolve the carbohydrate and stabilize it from caramelization reactions.

[0083]

[0088] The reactor is provided at a temperature of at least about 235°C, more preferably at least 240°C, and most often at least about 245°C, up to about 280°C or 300°C. The disclosed process involves the retroaldol conversion, which is primarily kinetically limited, and the hydrogenation conversion, which is primarily diffusion-limited. Adjusting the temperature provides an additional tool for balancing the retroaldol conversion and the hydrogenation conversion. Thus, an increase in sorbitol concentration in the product may be evidence of a need to increase the activity of the retroaldol catalyst, which can be achieved by increasing the reactor temperature. Conversely, for example, as glucose aldehyde molecules condense to form 1,2-butanediol, an increase in 1,2-butanediol may indicate that hydrogenation is slowing; in this case, either the temperature may be reduced or additional hydrogenation catalyst may be added. The amount of hydrogenation catalyst used is less than about 100 grams, often about 10 to 75 grams (calculated as grams of dry catalyst) per liter of liquid in the reactor.

[0084]

[0089] The pH of the liquid reaction medium is greater than 3, preferably greater than 3.5, e.g., 3.8 to 8, and in some cases about 4 to 7.5. Lower pH enhances the isomerization of aldoses to ketoses, as discussed above. Furthermore, lower pH also shifts the equilibrium between solubilized and precipitated tungsten compounds toward precipitated tungsten compounds.

[0085]

[0090] The disclosed process can be utilized in a variety of reactors, including, but not limited to, fixed-bed, fluidized-bed, trickle-bed, moving-bed, slurry-bed, and loop-type reactors, such as the Buss Loop® reactor available from BUSS ChemTech AG, and structured-bed reactors. Most commonly, stirred reactors are used to facilitate mass transfer of hydrogen and feed and intermediates to the catalyst. One type of reactor that can provide high hydrogen concentrations and rapid heating is a cavitation reactor, such as that disclosed in U.S. Pat. No. 8,981,135 B2, the entire contents of which are incorporated herein by reference. Cavitation reactors generate heat in localized regions; therefore, the temperatures in these localized regions, rather than the bulk temperature of the liquid medium in the reaction zone, are the temperature process parameter for purposes of this disclosure. Cavitation reactors are of interest for this process because the retro-aldol transformation can be extremely rapid at temperatures achievable in cavitation reactors.

[0086]

[0091] In the typical operation of a fluidized, moving, or slurry bed reactor, a portion of the hydrogenation catalyst is continuously or intermittently withdrawn from the reactor and replaced with regenerated or fresh hydrogenation catalyst. This procedure maintains relatively constant productivity and selectivity to ethylene glycol over long periods of operation.

[0087]

[0092] While the present disclosure has been described with reference to various embodiments, workers skilled in the art will recognize that changes can be made in form and detail without departing from the spirit and scope of the disclosure.

Claims

1. 1. A continuous catalytic process for producing ethylene glycol from a feed containing carbohydrates that yields aldoses, comprising: (a) continuously or intermittently supplying said feed to a reaction zone containing a liquid medium having a heterogeneous nickel-containing hydrogenation catalyst therein, said feed being supplied at a rate of at least about 50 grams / hour of carbohydrate per liter of liquid medium, said liquid medium being under catalytic conversion conditions comprising the presence of dissolved hydrogen, a temperature of at least about 235° C., a pH greater than 3, and a residence time sufficient to react at least 99 weight percent of the carbohydrate to produce said aldose; (i) the heterogeneous hydrogenation catalyst has a maximum particle size of less than about 100 microns; (ii) the hydrogenation catalyst is dispersed in the liquid medium in an amount of less than about 100 grams per liter, thereby providing a spatial relationship between catalytically active hydrogenation sites in the liquid medium; (b) continuously or intermittently feeding a homogeneous tungsten-containing retro aldol catalyst to the reaction zone, wherein the concentration of solubilized tungsten compounds, calculated as tungsten atoms, in the liquid medium in the reactor is from about 200 to 1500 milligrams per liter, and the relative amounts of hydrogenation catalyst and retro aldol catalyst are sufficient to provide a cumulative conversion efficiency of at least 75 percent of the aldose-containing carbohydrate to ethylene glycol over a duration of 100 hours under catalytic conversion conditions; and (c) continuously or intermittently withdrawing from said reaction zone a crude product stream containing ethylene glycol.

2. 2. The process of claim 1, wherein the aldose-yielding carbohydrate comprises glucose.

3. The process of claim 1 , wherein the liquid medium comprises water.

4. 4. The process of claim 3, wherein the liquid medium comprises water and a co-solvent having a solubility of hydrogen greater than the solubility of hydrogen in water.

5. 10. The process of claim 1, wherein the hydrogenation catalyst is a nickel-containing catalyst supported on an inert support having less than 10 weight percent nickel (calculated as elemental nickel).

6. 6. The process of claim 5, wherein the support has a surface area of less than 10 square meters per gram.

7. The process of claim 6, wherein the support is alpha-alumina.

8. 10. The process of claim 1, wherein the hydrogenation catalyst is pretreated to reduce its hydrogenation activity.

9. 9. The process of claim 8, wherein the pretreatment comprises depositing a tungsten-containing compound on the hydrogenation catalyst.

10. 10. The process of claim 1, wherein the solubilized tungsten compound in the liquid medium is at a concentration of about 300 to 1200 milligrams per liter, calculated as elemental tungsten.

11. 10. The process of claim 1, wherein the concentration of the solubilized tungsten compound in the liquid medium is cycled between higher and lower concentrations, either continuously or intermittently.

12. 12. The process of claim 11, wherein the pH is increased by at least 0.5 pH units during the cycle to a lower concentration of solubilized tungsten.

13. 10. The process of claim 1, wherein the hydrogenation catalyst is a supported catalyst and is present in the liquid medium in an amount to provide from 0.1 to 3 grams of nickel per liter (calculated as elemental nickel).

14. 10. The process of claim 1, wherein the catalytic conversion conditions comprise an absolute pressure of 2500 to 20,000 kPa.

15. 10. The process of claim 1 carried out in a loop reactor.

16. 10. The process of claim 1 carried out in a stirred bed reactor.

17. 10. The process of claim 1, wherein a portion of the hydrogenation catalyst is continuously or intermittently withdrawn from the reactor and replaced with rejuvenated or fresh hydrogenation catalyst.

18. 10. The process of claim 1, wherein the aldose-yielding carbohydrate-containing feed is supplied to the reactor in an amount of from 100 to 1000 grams of carbohydrate per liter of liquid medium per hour.

19. 20. The process of claim 18, wherein the aldose-yielding carbohydrate-containing feed is fed to the reactor at multiple points to reduce the risk of hydrogen starvation.

20. 2. The process of claim 1, wherein a portion of the homogeneous tungsten-containing retro-aldol catalyst is added to the aldose-producing carbohydrate-containing feed sent to the reactor.

21. 21. The process of claim 20, wherein the homogeneous tungsten-containing retro aldol catalyst has catalytic activity for isomerizing aldoses.

22. 22. The process of claim 21, wherein the homogeneous tungsten-containing retro aldol catalyst is partially neutralized.

23. 10. The process of claim 1, wherein the catalytic conversion conditions comprise a pH of from 3.8 to 8.

24. 10. The process of claim 1, wherein less than about 35 percent of the nickel on the hydrogenation catalyst is in the zero-valent state.

25. The process of claim 1 , wherein the support has a largest dimension of less than about 50 micrometers.

26. 10. The process of claim 1, wherein the carbohydrate-containing feed is introduced into the reaction zone in a spatially dispersed manner.

27. 27. The process of claim 26, wherein the weight ratio of the carbohydrate to liquid in the feed is about 0.1:1 to 0.4:

1.

28. 27. The process of claim 26, wherein the feed is introduced at two or more locations in the reaction zone.

29. 27. The process of claim 26, wherein the feed comprises a carbohydrate and a portion of the liquid medium from the reaction zone absent a hydrogenation catalyst.

30. 30. The process of claim 29, wherein a portion of the liquid medium from the reaction zone for the feed is used as a motive fluid for introducing hydrogen microbubbles into the reaction zone.