FEEDSTOCK

FIELD OF THE INVENTION

[0001] This invention relates to processes for the production of chemicals, including glycols, from renewable feedstocks, such as sugar-based materials, and, in particular, to the production of ethylene glycol and propylene glycol from biomass derived saccharide- containing feedstocks.

BACKGROUND OF THE INVENTION

[0002] Biomass is considered as an ideal feedstock to partially or completely replace fossil derived resources for the sustainable production of specialty chemicals due to its availability from various resources. Waste or surplus biomass from agricultural residues, forestry wastes and energy crops can ensure renewability as well as non-competition with food supply.

[0003] Glycols such as mono-ethylene glycol (MEG) and monopropylene glycol (MPG) are valuable materials with a multitude of commercial applications, e.g., as heat transfer media, antifreeze, and precursors to polymers, such as polyesters. Ethylene and propylene glycols are typically made on an industrial scale by hydrolysis of the corresponding alkylene oxides, which are the oxidation products of ethylene and propylene, produced from fossil fuels. Production of glycols, from non-petrochemical renewable feedstocks, such as biomass, is highly desirable w ith the conversion of sugars such as glucose to glycols representing an efficient use of the starting materials since the oxygen atoms remain intact in the desired end product.

[0004] Conventional methods for the conversion of saccharides to glycols revolve around a nickel-promoted tungsten carbide catalytic hydrogenation/retro-aldol process that was initially described in Ji et al. (“Direct Catalytic Conversion of Cellulose into Ethylene Glycol using Nickel-Promoted Tungsten Carbide Catalysts” Angew. Chem, Int. Ed. 47:8510-8513; 2008). Typically, the hydrogenation catalyst compositions tend to be heterogeneous, which means that they are not in the same phase as the predominantly liquid feedstock and remain physically separable from it. However, the retro-aldol catalysts are generally homogeneous and are dissolved into the reaction mixture. Such catalysts are inherently limited due to solubility constraints and can precipitate out of solution under certain conditions. However, efficient recovery of the homogenous catalyst can be a challenge and its presence in reactor effluent can affect a range of dow nstream process reactions. [0005] Continuous flow processes for the production of glycols from saccharide feedstock have been described in US8410319, CN102675045, CN102643165, US8222462 and CN103731258. A process for the co-production of bio-fuels and glycols is described in US8877985.

[0006] CN 102643164 A describes a method for producing ethylene glycol and 1,2- propylene glycol by continuously hydrocracking cellulose. In the described continuous process, the reactor effluent contains a diverse mixture of reaction products as well as homogeneous and heterogenous catalyst. A rectification system downstream of the reactor is used to separate polyol products from heavy components. A major proportion of the rectified heavy components is returned to the reactor for further hydrocracking. In this way, soluble homogenous catalyst present in the heavy component products is returned directly to the reactor for recycling. According to CN 102643164 A, by recovering and partially refluxing the soluble catalyst the consumption of the catalyst is reduced, the cost of the catalyst is saved, and the problem of the loss of catalyst is addressed. A minor portion of the rectified heavy component tungstate ions are recovered by a combination of precipitation and ion exchange.

[0007] A disadvantage of this process is that the presence of homogeneous retro-aldol catalyst in rectification system promotes degradation of sugar alcohols, resulting in an organic acid by-product. When recycled to the hydrogenation reactor, there is a resulting build-up of organic acids that, in turn, destabilizes the heterogenous catalyst. Furthermore, the reduction in heavies fraction by conversion to organic acids and other components reduces the overall glycol yield, as the resulting degradation products cannot be effectively converted to glycols. Moreover, pH control in the main reactor with caustic would result in a build-up of sodium in the recycled retro-aldol catalyst stream, causing a high sodium/retro-aldol metal ratio (e.g., a Na/W molar ratio greater than 3), which adversely impacts the retro-aldol catalytic activity and/or the hydrogenation reaction. In addition, the presence of retro-aldol catalyst in the heavies stream and related waste streams presents an environmental problem in that the waste cannot be safely/responsibly incinerated and/or requires costly and energy -intensive treatment processes, such as submerged flame incineration.

[0008] There remains a need to further improve glycol production from renewable feedstocks to facilitate efficient recovery and recycling of homogenous retro-aldol catalyst, as well as reducing the effect of such catalyst on downstream aspects of the process.

[0009] These and other uses, features and advantages of the invention should be apparent to those skilled in the art from the teachings provided herein. SUMMARY OF THE INVENTION

[0010] According to one aspect of the present invention, there is provided a process for the production of glycol from a saccharide-containing feedstock, the process comprising the steps of: subjecting the saccharide-containing feedstock to a catalytic conversion reaction, in a conversion zone, in the presence of a heterogenous hydrogenation catalyst and a homogeneous retro-aldol catalyst resulting in a glycol product; removing an effluent from the conversion zone, wherein the effluent comprises a first portion of the homogeneous retro-aldol catalyst and the glycol product; contacting at least a portion of the effluent with an ion exchange material to adsorb at least a portion of anions from the homogeneous retro-aldol catalyst present in the effluent; desorbing the adsorbed anions from the ion exchange material and recycling at least a portion of the desorbed anions to the conversion zone; after the contacting step, separating a product stream and a heavies fraction from the effluent; and passing the product stream to a glycol recovery zone for recovering a purified glycol product.

BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The process of the present invention will be better understood by referring to the following detailed description of preferred embodiments and the drawings referenced therein, in which:

[0012] Fig. l is a schematic representation of one embodiment of the process of the present invention;

[0013] Fig. 2 is a schematic representation of a preferred embodiment of the separator zone of Fig. 1;

[0014] Fig. 3 is a schematic representation of another embodiment of the process of the present invention;

[0015] Fig. 4 is a schematic representation of a further embodiment of the process of the present invention;

[0016] Fig. 5 is a schematic representation of yet another embodiment of the process of the present invention; and

[0017] Fig. 6 is a graphical representation of results of Example 3 herein.

DETAILED DESCRIPTION OF THE INVENTION

[0018] The present invention provides a process for the production of glycol from a saccharide-containing feedstock in the presence of a catalyst system having a homogeneous retro-aldol catalyst and a heterogeneous hydrogenation catalyst. Effluent removed from the conversion zone comprises a portion of the homogeneous retro-aldol catalyst and the glycol product. The homogeneous retro-aldol catalyst has a transition metal oxide anion. At least a portion of the effluent is contacted with an ion exchange material to adsorb at least a portion of the transition metal oxide anions from the retro-aldol catalyst present in the effluent. The transition metal oxide anions originating from the homogeneous catalyst are desorbed from the ion exchange material and recycled to the conversion zone, without degrading the heavies byproducts that would result in formation of carboxylic acids and reduced valorisation of the heavies by-product stream as seen in conventional processes. A glycol-containing product stream and a heavies fraction are separated from the effluent that has been contacted with the ion exchange material. The glycol-containing product stream is passed to a glycol recovery zone for recovering a purified glycol product. Because the heavies fraction is not degraded by heating in the presence of retro-aldol catalyst during the separation step, the heavies fraction may be valorised, for example, by burning for heat generation, co-feeding to another process such as thermal cracking, energy recovery, anaerobic conversion to methane, catalytic cracking, and combinations thereof. In one preferred embodiment, the heavies fraction is subjected to catalytic cracking to produce a cracked product, which preferably comprises glycols. Preferably, the cracked product is passed to the glycol recovery zone for recovering a purified glycol product, thereby increasing overall product yield.

[0019] As discussed above, some conventional processes separate the product glycol from the effluent and return the heavies fraction and homogeneous catalyst to the conversion zone. However, such separation processes typically require a heat source (for example, a reboiler). The heavies fraction typically comprises oxygenated hydrocarbons, including sugar alcohols, polyalcohols, and combinations thereof. When the retro-aldol catalyst is heated in the presence of the oxygenated hydrocarbons, they are partially converted to organic acids. Under water- free conditions, the acids might undergo esterification reaction, which, in turn, might hydrolyse again to carboxylic acids once mixed in the reactor with the water-containing liquid. Thus, when the fractionated stream is recycled back to the conversion reactor there is a build-up of organic acids and other compounds that, in turn, deleteriously affect the stability of the hydrogenation catalyst. Furthermore, the value of the heavy fraction is reduced by conversion to organic acids and other components that cannot be effectively converted to glycols.

[0020] In general terms, the present invention provides a process for the efficient recovery of the homogenous retro-aldol catalyst from reactor effluent at a stage in the process before separation of glycol products from the effluent occurs. The effluent comprises a diverse mixture of glycol molecules, such as ethylene glycol and propylene glycol (the main products of the process), as well as sugar alcohols, organic acids, and trace components. Recovery of the homogenous retro-aldol catalyst is carried out by contacting this effluent with an ion exchange material. The inventors have surprisingly found that the removal of homogeneous catalyst at this stage in the process contributes to greater stability of the product stream during subsequent refining stages. In addition, removal of the retro-aldol catalyst from the stream at a location close to the reactor outflow provides greater energy efficiency and process simplification for the overall biomass to glycols process.

[0021] Advantages of the present invention include: recovery of the retro-aldol catalyst and being able to re-use it again as catalyst for the main hydrogenolysis process without degrading the hydrogenation catalyst; decoupling recycle of the retro-aldol catalyst from downstream processing so that, even if acids are formed in distillation, there is little to no effect on the conversion zone; reducing conversion of sugar alcohols to components, including carboxylic acids, that cannot be effectively cracked to glycols, thereby limiting the maximum overall glycol yields that can be obtained. Accordingly, the heavies fraction can be treated more cost effectively because it does not contain the retro-aldol catalyst, and/or the heavies fraction may be otherwise valorised, e.g., enabling energy' recovery. Further, the process of the present invention allows for the use of caustic for pH control in the conversion zone.

[0022] Referring now to Fig. 1, the process of the present invention 10 is a process for the production of glycol from a saccharide-containing feedstock 12.

[0023] Preferably, the glycol is selected from the group consisting of ethylene and propylene glycols. More preferably, the glycol is selected from monoethylene glycol (MEG), monopropylene glycol (MPG), and combinations thereof.

[0024] The saccharide-containing feedstock 12 for the process of the present invention 10 is selected from the group consisting of monosaccharides, disaccharides, oligosaccharides and polysaccharides. Preferably, the saccharide-containing feedstock 12 comprises saccharides selected from glucose, sucrose, starch, maltose, cellobiose, com syrup, cellulose, hemicellulose, glycogen, chitin, and combinations thereof. More preferably, the saccharide- containing feedstock 12 comprises saccharides selected from glucose, sucrose, starch, and combinations thereof. Most preferably, the saccharide-containing feedstock 12 comprises glucose.

[0025] If the saccharide-containing feedstock 12 includes, or is derived from, oligosaccharides or polysaccharides, the oligosaccharides and poly saccharides are preferably subjected to pre-treatment before being used in the process of the present invention 10. Suitable pre-treatment methods are known in the art and one or more may be selected from the group including, but not limited to, sizing, drying, milling, hot water treatment, steam treatment, hydrolysis, pyrolysis, thermal treatment, chemical treatment, biological treatment, saccharification, fermentation, and solids removal. However, after pre-treatment, the starting material still comprises mainly monomeric and/or oligomeric saccharides. The saccharides are, preferably, soluble in the reaction solvent.

[0026] Preferably, the saccharide-containing feedstock 12, after any pre-treatment, comprises saccharides selected from glucose, starch and/or hydrolysed starch. Hydrolysed starch comprises glucose, sucrose, maltose, and oligomeric forms of glucose. The saccharides are suitably present as a solution, a suspension, or a slurry in a solvent.

[0027] After pre-treatment, the treated feedstock stream is suitably converted into a solution, a suspension, or a slurry in a solvent. The solvent may be water, a Ci to Ce alcohol, a polyalcohol, or a mixture thereof. For example, the solvent may be a mixture of water and Ci to Ce alcohol or polyalcohol. An example is a 50:50 mixture of water and Ci to Ce alcohol or poly alcohol. Suitable Ci to Ce alcohols include methanol, ethanol, 1 -propanol and isopropanol. Suitable polyalcohols include glycols, particularly products of the hydrogenation reaction, glycerol, erythritol, threitol, sorbitol, 1,2-hexanediol, and mixtures thereof. More suitably, the polyalcohol may be glycerol or 1,2-hexanediol. Further solvent may also be added to a reactor vessel or reactor vessels in a separate feed stream or may be added to the treated feedstock 12 stream before it enters the conversion zone 14.

[0028] The concentration of the saccharide-containing feedstock 12 as a solution in the solvent supplied to the conversion zone 14 is at most at 80 wt.%, preferably at most 60 wt.%, more preferably, at most 45 wt.%.

[0029] In the conversion zone 14, the saccharide-containing feedstock 12 is subjected to a catalytic conversion reaction in the presence of a catalyst system comprising a heterogeneous hydrogenation catalyst and a homogeneous retro-aldol catalyst. The catalytic conversion reaction results in a glycol product.

[0030] The conversion zone 14 may include one or more reactors. Suitable reactors include stirred tank reactors, slurry reactors, ebullated bed reactors, jet flow reactors, mechanically agitated reactors, bubble columns, such as slurry bubble columns, and external recycle loop reactors. The use of these reactors allows dilution of the feedstock 12 and reaction intermediates to an extent that provides high degrees of selectivity to the desired glycol product (mainly ethylene and propylene glycols), such as by effective back-mixing.

[0031] In one embodiment, the conversion zone 14 includes a main hydrogenolysis zone and a catalytic finishing hydrogenation zone (not shown). The main hydrogenolysis reactor in the conversion zone 14 is preferably a back-mixed system. The nature of a back-mixed system is that the product stream from the reactor contains intermediate carbonyl-containing components (e.g., glycolaldehyde, hydroxyacetone). The concentration of intermediates m the product stream is typically <1 wt.%. In a preferred embodiment, the catalytic finishing hydrogenation zone, for example a fixed bed reactor, is provided to fully hydrogenate intermediates to optimize glycol yields and to remove reactive carbonyl components to suppress reactions dow nstream (e.g., in reboilers of a distillation section).

[0032] Where a catalytic finishing hydrogenation zone is employed, it is preferably located directly after the main hydrogenolysis zone, and before the ion exchange material 18 (discussed more fully below). Suitable catalysts for the catalytic finishing hydrogenation zone include any suitable heterogeneous hydrogenation catalyst. One example is a Ni/ZrO2 catalyst. The catalytic finishing hydrogenation reaction can be successfully operated at low temperatures (e.g., 50°C), but the reaction is preferably operated at the same temperature as the main hydrogenolysis zone for process economics (by avoiding the need to cool down the stream first). However, a temperature lower than in the main hydrogenolysis reactor can be considered to suppress glycols cracking (yield loss) if this is observed at the elevated temperatures. In a preferred embodiment, the temperature of the catalytic finishing hydrogenation zone is in a range of from 180°C to 190°C. In a more preferred embodiment, the temperature in the catalytic finishing hydrogenation zone is close to the operating temperature of the main hydrogenolysis zone. The finishing zone is preferably operated at similar pressure as the main hydrogenolysis zone

[0033] The catalyst system in the conversion zone 14 catalyses a conversion reaction, in the presence of hydrogen, selected from hydrolysis, retro-aldol condensation, and/or hydrogenation reactions. For example, when the saccharide-containing feedstock 12 comprises cellulose, reactions include hydrolysis of cellulose to glucose, C-C cleavage of glucose by retro-aldol condensation to form glycolaldehyde, and hydrogenation of glycolaldehyde to ethylene glycol. In addition, glucose may be isomerized to fructose, and fructose is converted to propylene glycol. By-products of hydrogenation include sorbitol and/or mannitol.

[0034] The hydrogenation catalyst comprises a transition metal having catalytic hydrogenation capabilities selected from Groups 8, 9 and 10 of the periodic table. Preferably, the hydrogenation catalyst comprises a metal selected from the group consisting of iron, cobalt, nickel, ruthenium, rhodium, palladium, iridium, platinum, and combinations thereof. The metal or metals may be present in elemental form or as compounds. It is also suitable that this component is present in chemical combination with one or more other ingredients in the hydrogenation catalytic composition.

[0035] In some embodiments, the hydrogenation catalytic composition comprises metals supported on a solid support. In some embodiments, the solid supports may be in the form of a powder, or in the form of regular or irregular shapes, such as spheres, extrudates, pills, pellets, tablets, and/or monolithic structures. Alternatively, the solid supports may be present as surface coatings, for example on the surfaces of tubes or heat exchangers. Suitable supports for the hydrogenation catalyst are those known to the skilled person and include, without limitation, alumina, silica, zirconium oxide, magnesium oxide, zinc oxide, titanium oxide, carbon, activated carbon, zeolite, clay, silica alumina, and combinations thereof.

[0036] In one embodiment, the hydrogenation catalyst is a skeletal catalyst, preferably a skeletal nickel catalyst. Those skilled in the art typically refer to a Raney ©-metal type catalyst, such as a Raney®-nickel catalyst, when describing a skeletal catalyst. Commercially available equivalents or substitutes for a Raney®-Ni catalyst will be recognized and understood Preferably, the skeletal metal catalyst is provided in a pelletised form as a slurry. In another embodiment, the hydrogenation catalyst is a supported catalyst, such as ruthenium supported on activated carbon.

[0037] The homogeneous retro-aldol catalyst has a transition metal oxide anion. The retro- aldol catalyst preferably comprises a transition metal compound, complex or elemental material comprising tungsten, molybdenum, lanthanum, tin, vanadium, niobium, chromium, titanium, zirconium, and combinations thereof. More preferably, the retro-aldol catalyst composition comprises one or more material selected from the group consisting of tungstic acid, molybdic acid, ammonium tungstate, ammonium metatungstate, ammonium paratungstate, tungstate compounds comprising at least one Group 1 or 2 element, such as sodium tungstate, metatungstate compounds comprising at least one Group 1 or 2 element, such as sodium metatungstate, paratungstate compounds comprising at least one Group 1 or 2 element, such as sodium paratungstate, heteropoly compounds of tungsten, heteropoly compounds of molybdenum, tungsten oxides, molybdenum oxides, vanadium oxides, metavanadates, chromium oxides, chromium sulfate, titanium ethoxide, zirconium acetate, zirconium carbonate, zirconium hydroxide, niobium oxides, niobium ethoxide, and combinations thereof. Preferably, the retro-aldol catalyst composition comprises one or more compound, complex or elemental material selected from those containing tungsten, or alternatively, those containing molybdenum. [0038] The weight ratio of the retro-aldol catalyst to saccharide-containing feedstock 12 in the main hydrogenolysis zone is suitably in the range of from 1 : 1 to 1 : 1000, based on the weight of metal in the catalyst.

[0039] A buffer is preferably supplied to the conversion zone 14. The purpose of the buffer is to maintain the pH in a preferred range. Suitable buffers will be known to the skilled person but include, for example, without limitation, caustic (NaOH) and sodium bicarbonate. The amount of buffer supplied to the reactor is suitably in a range of from 0.05 to 5 wt.% buffer based on the total weight of feedstock 12 supplied to the reactor, preferably in a range of from 0. 1 to 1 wt.% buffer. As explained more fully below, in the process of the present invention, transition metal oxide anions originating from the homogeneous catalyst are recycled to the conversion zone 12. The recycle stream includes the desorbing solvent, which is preferably a caustic solvent. Accordingly, the amount of buffer added to the conversion zone 14 is preferably adjusted to provide an alkali-metal: metal molar ratio is less than 3, where the metal originates from the homogeneous retro-aldol transition metal oxide anions.

[0040] Glycol is produced by hydrogenolysis of the saccharide-containing feedstock 12. The temperature in a reactor of the conversion zone 14 is suitably at least 80°C, preferably at least 130°C, more preferably at least 160°C, most preferably at least 190°C. The temperature in the reactor is suitably at most 300°C, preferably at most 280°C, more preferably at most 250°C, most preferably at most 230°C. Operating at higher temperatures has the potential disadvantage of increased amounts of side-reactions, leading to lower yield, while operating at a low temperature might result in suppression or inactivation of the retro-aldol catalyst. Preferably, the glycol production is conducted at a temperature in a range of from 180°C to 250°C, more preferably in a range of from 210°C to 250°C.

[0041] The pressure in the reactor is suitably at least 1 MPa, preferably at least 2 MPa, more preferably at least 3 MPa. The pressure in the reactor is suitably at most 25 MPa, preferably at most 20 MPa, more preferably at most 18 MPa. Preferably, glycol production is conducted at a pressure in a range of from 3 MPa to 14 MPa. Preferably, the reactor is pressurised to a pressure within these limits by addition of hydrogen before introducing the feedstock 12 and is maintained at such a pressure as the reaction proceeds through continuous addition of hydrogen.

[0042] The residence time in the reactor during the glycol production step is suitably at least 1 minute, preferably at least 2 minutes, more preferably at least 5 minutes. Suitably the residence time in the reactor is no more than 5 hours, preferably no more than 2 hours, more preferably no more than 1 hour. [0043] It will be understood by those skilled in the art that reaction parameters may be adjusted as needed over time to achieve a steady state concentration of product in the conversion zone 14. For example, the parameters that may be adjusted include, without limitation, feed rate, residence time, saccharide concentration in the feed, temperature, and/or pressure. A corresponding amount of effluent 16 is removed continuously from the conversion zone 14.

[0044] Effluent 16 is drawn from the conversion zone 14 in a manner such that the heterogeneous catalyst is substantially retained in the conversion zone 14. For example, the effluent 16 may be draw n through a filter or screen (not shown) to retain the heterogeneous hydrogenation catalyst inside a reactor in the conversion zone 14, while allowing a flow of effluent 16 to leave the reactor. The effluent 16 comprises the product glycol and a portion of the homogenous retro-aldol catalyst. The pH of the effluent 16 from a main hydrogenolysis reactor in the conversion zone 14 is acidic, typically in a range of from 3 to 6, practically in a range of from 3.5 to 5.0. The pH of the effluent from an optional catalytic finishing step may be higher than the pH from the main hydrogenolysis reactor.

[0045] Effluent 16 is contacted with ion exchange material 18 to separate the transition metal oxide anions of the homogeneous catalyst from the remainder of the effluent 16. After the effluent 16 is contacted with the ion exchange material 18, a gly col-containing product stream 22 is separated from heavies fraction 24 in separator zone 20. The present inventors have surprisingly discovered that, by separating the transition metal oxide anions of the homogenous catalyst with the ion exchange material 18 before the separation step, degradation of the heavies fraction 24 in the separator zone 20 is reduced or avoided altogether. The gly col- containing product stream 22 is passed to a glycol recovery zone 50. The heavies fraction 24 may be recycled to the conversion zone 14, further processed in the process of the present invention (see, for example, the embodiment of Fig. 3), thermally cracked to produce product and/or feedstock for another process, anaerobically converted to methane, and/or burned as a fuel. These options for valorising the heavies fraction 24 are typically not an option when they have been degraded to acids in the presence of retro-aldol catalyst, and/or are contaminated with retro-aldol catalyst. The retro-aldol catalyst can be recycled in recycle stream 28 to the conversion zone 14 without adversely impacting the hydrogenation catalyst.

[0046] In conventional processes, acidic conditions destabilize the hydrogenation catalyst, and a more stable catalyst is needed. A highly active retro-aldol catalyst is needed for the reaction to work well under the conditions of high-concentration glucose in the reaction mixture, but that activity cannot be provided by simply raising the concentration of the homogeneous transition metal because of its tendency to destabilize by precipitation, leading to clogging of the reactor and/or deleterious deposits on process equipment.

[0047] The ion exchange material 18 may be provided as ion exchange resin and used in the form of a plug-flow operated column or provided in an annular space between a pair of concentric cylinders having openings through which fluid may pass to/from the ion exchange material. Ion-exchange resins are known for removal of tungsten from aqueous streams. See, for example, Di Natale et al. (“Recovery of Tungstate from Aqueous Solutions by Ion Exchange” Ind. Eng. Chem. Res. 46:21:6777-6782; 2007). Ion exchange resins are also used for separation of tungsten from molybdenum (for example, see CN102162030A).

[0048] Synthetic ion exchange resins typically comprise high molecular weight polymeric materials containing a plurality of ionic functional groups per molecule and associated counter ions bonded to an inside surface layer of their pores. Exemplary anion exchange resins suitable for removal of metal anions from a process stream can comprise a strongly basic QAE (quaternary amino ethyl) group, or alternatively a weakly basic DEAE (diethylamino ethyl) group. In a preferred embodiment, the resin is used in the form of a packed bed of microporous resin beads, in a range of from 0. 1 to 1 mm in diameter, through which the effluent 16 can be passed. Preferably, the resin is packed in a manner known by those skilled in the art to avoid channelling of the effluent through the resin. Surprisingly it has been found that adsorption of transition metal oxide anions, such as tungstate, on the ion-exchange resin may be performed selectively, thereby avoiding removal of valuable product (e g., MEG or MPG), from the product stream along with the anions from the retro-aldol catalyst.

[0049] Preferred ion exchange resins include ammonium and/or quaternary ammonium cations suitable for adsorbing transition metal oxide anions originating from the homogeneous catalyst from the effluent 16. Preferably, the ion exchange resin is a strong base anion (Type II) exchange resin. Examples of suitable ion exchange resin that may be used include, without limitation, AmberLite™ IRA-410, AmberLite™ IRA-900, Amberlyst™ A21, Amberlyst™ A26, and combinations thereof.

[0050] The ion exchange resins are often supplied in chloride form. Preferably, prior to being placed into service, the ion exchange resin is conditioned to replace chlorine ions with hydroxyl ions, for example, by washing with and/or soaking in a sodium hydroxide solution.

[0051] The ion exchange material 18 is preferably regenerated, once saturated, to remove transition metal oxide anions originating from the retro-aldol catalyst from the material 18 and allow them to be recycled as catalyst to the conversion zone 14. Regeneration of the ion exchange material 18 may be performed, for example, without limitation, by washing with a concentrated brine solution or sodium hydroxide. Preferably, the regeneration is conducted to achieve a molar ratio of sodium to transition metal originating from the retro-aldol catalyst (for example, tungsten) that is less than or equal to 3, preferably the molar ratio is in a range of from 2 to 4, more preferably in a range of from 2 to 3.

[0052] At least a portion of the retro-aldol catalyst is recycled to the conversion zone 14 in stream 28. It may be advantageous to, periodically or continuously, remove a bleed stream 32 of retro-aldol catalyst from recycle stream 28 to remove a possible build-up of acids, chlorine, and the like, which could otherwise adversely impact reactions in the main conversion zone 14. In this case, it may not be required to remove a bleed stream 32 earlier in process, but possibly, after a number of adsorption/desorption cycles, it may be desirable to start periodically or continuously removing a bleed stream 32 prior to recycling stream 28 to the conversion zone 14.

[0053] In conventional processes, where retro-aldol catalyst has been recycled with the heavies fraction after separation of the glycol-containing product stream, there has been a problem of using a caustic solution as a buffer due to the build-up of sodium in the recycle stream, which impacts the alkali -metal metal ratio due to the presence of organic acids formed as by-products in the main reactor. Accordingly, based on an understanding of acid-make in the main reactor, the present inventors expected that a correction to the sodium concentration in the recycle stream 28 would be required before adding the recycle stream 28 to the feedstock 12 or the conversion zone 14. In particular, the inventors expected that the molar ratio of sodium to retro-aldol transition metal would be greater than 3. The inventors surprisingly found that the desorbed anions in caustic/brine was closer to the theoretical molar ratio of 2, and at least <3, and that correction was therefore not required.

[0054] The ion exchange material 18 may be contained within one or more beds across or through which the effluent 16 is passed. Preferably, the ion exchange material 18 is provided in a lead-lag configuration, where a majority of the anions are adsorbed in the lead ion exchange material and a remainder of the transition metal oxide anions are adsorbed in a lag ion exchange material. The ion exchange material in the lead and lag may be the same or different. The inventors have found that the lead ion exchange material reduces the transition metal oxide anion concentration by more than 95 wt.%, even 97 wt.% (e g., from 3000 ppm to 100 ppm). The lag ion exchange material improved the recovery to more than 99 wt.%, even 99.9 wt.% (e.g., less than 100 ppm, preferably less than 10 ppm, even more preferably to about 1 ppm). The lag ion exchange material can also mitigate the risk associated with early or unexpected slip of transition metal oxide anions from the lead ion exchange material (for example, by misoperation or channelling in an ion exchange bed).

[0055] To ensure process continuity during a desorption cycle, the ion exchange material 18 has two or more parallel units so that effluent 16 may be routed to one ion exchange material 18, while the other is undergoing a desorption cycle. More preferably, the ion exchange material 18 is provided in two or more parallel lead-lag configurations, where each parallel stream has a lead and a lag ion exchange material 18.

[0056] In accordance with the present invention, the concentration of transition metal oxide anions present in the effluent 16 is depleted by at least 50 wt.%, relative to the amount of active metal in the retro-aldol catalyst present prior to contacting with the ion exchange material 18. Preferably, at least 60 wt.%, at least 70 wt.%, at least 80 wt.%, at least 90 wt.%, or even greater than 95 wt.% of transition metal oxide anions are depleted from the effluent 16. The amount of retro-aldol catalyst metal may be determined by a method known to those skilled in the art, including, without limitation, ICP (Inductively Coupled Plasma) and XRF (X-Ray Fluorescence) analysis. The inventors have found a reduction in transition metal oxide anions of at least 97 wt.% with one ion exchange bed and up to 99.9 wt.% or more when ion exchange beds were operated in a lead-lag configuration.

[0057] Separator zone 20 is provided to separate heavies fraction 24 from the glycol- containing product stream 22. In the embodiment illustrated in Fig. 1, the separator zone 20 schematically illustrates outlet streams comprising gly col-containing product stream 22, lights stream 26, and heavies fraction 24.

[0058] Preferably, the separator zone 20 includes a gas/liquid separator, a lights stream recovery column, water removal, and/or one or more distillation columns for separating the gly col-containing product stream 22 from the heavies fraction 24.

[0059] A preferred embodiment of the separator zone 20 is illustrated in Fig. 2. In this embodiment, the effluent 16 from the conversion zone 14 is directed to a hot high-pressure separator 34 for separating a vapour phase from a liquid phase. The liquid phase is directed to a cold low-pressure separator 36 for further gas/liquid separation. The liquid phase from the cold low-pressure separator 36 is passed to a lights recovery column 38, from which the lights stream 26 is recovered. The lights stream 26 comprises C1-C4 alcohols produced in the conversion zone 14. The liquid stream from the lights recovery column 38 is directed to the ion exchange material 18.

[0060] It will be understood by those skilled in the art that the hot high-pressure separator 34 operates at a pressure that is close to the conversion zone 14 pressure, suitably in a range of from 0 to 10 bar (0 to 1 MPa) below the preceding reactor outlet pressure, while the cold low- pressure separator 36 is operated at a pressure that is lower than a preceding reactor in the conversion zone 14 pressure or a preceding high-pressure separator 34, suitably in a range of from 0 to 15 barg (0 to 1.5 MPaG). Similarly, it will be understood by those skilled in the art that hot means that the hot high-pressure separator 34 is operated at a temperature that is close to a preceding reactor in the conversion zone 14 temperature, suitably not lower than the operating temperature of the ion exchange material 18, while the cold low-pressure separator 36 is at a reduced temperature relative to the preceding reactor in the conversion zone 14, optionally recovering the heat (not shown). For example, a cold low-pressure separator 36 is suitably operated at a temperature that can be achieved via an air cooler.

[0061] In the embodiment depicted in Fig. 2, the stream from the ion exchange material 18 is directed to water removal step 42. This embodiment is also depicted more generally in Fig. 4. Water removal step 42 may be effected using equipment such as, for example, without limitation, a concentrator, a condenser, a water evaporation column, a dehydrator, and the like. The selection of equipment type, and operation can be determined by a person skilled in the art in light of process and/or product targets, downstream equipment, and the like. For example, the equipment/ operation may be selected to achieve a target of 0.05 wt.% water to meet a target specification for MEG. But, for example, where the downstream equipment includes a multieffect distillation in the glycol recovery zone 50, it may not be necessary to achieve the water content target at this point of the process.

[0062] For example, the concentrator may be a series of three concentrators, the first operating at high temperature and high pressure, while the second and third concentrators operate at progressively lower temperatures and pressures. The product of the concentrator stream typically has a water content of about 10 wt.%. Conventionally, the product is then routed to a high temperature dehydrator to further remove water before separation and purification of products.

[0063] In the embodiment of Fig. 2, the dewatered stream from the water removal step 42 is then directed to a glycol/heavies separator 44. The glycol/heavies separator 44 is preferably one or more distillation columns. The distillation column may be any suitable distillation column known to those skilled in the art and may be equipped with trays, structured packing, and/or unstructured packing. The column preferably has a number of theoretical trays in a range of from 3 to 140. The actual number of theoretical trays is readily determined by a skilled person. [0064] In a preferred embodiment, the distillation column is operated under vacuum, for example, at a pressure in a range of from 1 to 200 mbar (0.1 to 20 kPa), preferably at a pressure in a range of from 70 to 90 mbar ( 7 to 9 kPa). A reboiler for the distillation column may be operated at a temperature in a range of from 160°C to 220°C, preferably in a range of from 195°C to 210°C. Water is preferably dosed to the reboiler to suppress dehydration reactions.

[0065] The specific equipment and configuration of equipment in the separator zone 20 may be selected by those skilled in the art. As shown in Fig. 2, the light stream 26 is removed prior to contacting the effluent with the ion exchange material 18. This embodiment may be particularly advantageous to remove C1-C4 compounds that may interact with the ion exchange material 18.

[0066] The composition of the glycol-containing product stream 22 drawn from the separator zone 20 is dependent on the feedstock and the reaction conditions. As an example, the glycol-containing product stream 22 contains MEG, MPG, 1,2-butanediol (1,2-BDO), water, and possibly 1 ,2-propanediol (1,2-PDO), 1,2-hexanediol (1,2-HDO) and/or trace components. Likewise, the composition of the heavies fraction 24 is dependent on the feedstock and the reaction conditions. As an example, the heavies fraction 24 contains sorbitol, erythritol, threitol, glycerol, carboxylic acids, carboxylic acid esters, 1,2-PDO, 1,2-HDO, and combinations thereof. The heavies fraction 24 may also contain a residual amount of MEG (for example, <10 wt.%, preferably much lower).

[0067] In one embodiment of Fig. 2, the liquid stream from the lights recovery column 38 may be passed through a heat exchanger 56 cool the stream to a temperature suitable for ion exchange material 18 and re-heat prior to forwarding the liquid stream to the water removal step 42. A single counter-current heat exchanger is preferred, although a sequence of separate heat exchangers is also possible.

[0068] In the embodiment depicted in Fig. 3, the heavies fraction 24 from separator zone 20 is subjected to a cracking reaction in a cracking zone 60. Preferably, the cracked product 62 is passed to the glycol recovery zone 50 for independent glycol recovery and/or for processing together with the glycol-containing product stream 22. For example, the heavies fraction 24 may result in a cracked product 62 that has a higher concentration of MPG, whilst the glycol-containing product stream 22 has a higher concentration of MEG. In this case, it is preferred to independently recover the respective target glycols in the glycol recovery zone 50. The independent recovery of MEG and MPG may be further improved by integration of the independent glycol recovery schemes, for example by directing streams of certain fractions between the schemes. [0069] The cracking zone 60 is provided to crack the heavies fraction 24 to smaller molecules that can themselves be valorised, separated in the glycol recover}' zone 50, and/or used as a feedstock in another process. The cracking zone 60 is preferably a catalytic cracking zone. In a preferred embodiment, the cracking zone 60 comprises a catalytic trickle bed reactor. There are many different types of catalysts that may be used to crack the heavies fraction 24 comprising Cs-Ce sugar alcohol by-products (e.g., glycerol, erythritol, threitol, and/or sorbitol), carboxylic acids, and their respective esters. One example of a suitable cracking catalyst is a Ni/ZrO2 catalyst. Suitable catalysts and process conditions are also described in Edulji et al. (US10,414,708, 2019 Sep 17 “Method for the production of glycols from sorbitol”).

[0070] In the embodiment illustrated in Fig. 4, the process of the present invention 10 further comprises a step of dewatering the effluent 16 after passing through the ion exchange material 18. Effluent from the ion exchange material 18 is passed to a water removal step 42 for separating water from the effluent 16. Optionally, the embodiment of Fig. 4 is combined with the embodiment of Fig. 3 to add a cracking zone 60 for cracking the heavies fraction 24 from separator zone 20. An advantage of this embodiment is that no solvent change is required between the resin loading and un-loading steps.

[0071] In the embodiment illustrated in Fig. 5, the process of the present invention 10 further comprises a step of dewatering the effluent 16 before passing through the ion exchange material 18. Effluent from conversion zone 14 is passed to a water removal step 42 for separating water from the effluent 16. Optionally, the embodiment of Fig. 5 is combined with the embodiment of Fig. 3 to add a cracking zone 60 for cracking the heavies fraction 24 from separator zone 20. An advantage of this embodiment is energy savings in that the liquid stream temperature has a reduced temperature after water separation and closer to the temperature applied during exchange resin treatment, reducing energy requirements.

[0072] The glycol recovery zone 50 may have a variety of configurations and process equipment known to those skilled in the art. Examples of suitable processes and apparatus include, without limitation, extractive distillation (such as described by van der Heide et al. (WO2021/122853A1, 2021 Jun 24), Perez Golf et al. (US, 4 Jun 2019), Fischer et al. (EP3464226B1, 2020 Aug 12) and Huizenga et al. (US10246390, 2019 Apr 2)), azeotropic distillation (such as disclosed by Huizenga et al. (US10266470, 2019 Apr 23), and Fischer et al. (US10081584, 2018 Sept 25)) and multi-effect distillation (such as described by van der Heide et al. (co-pending EP application number 21207919 filed 2021 Nov 12)). These disclosures are incorporated herein by reference. [0073] The glycol recovery zone 50 preferably results in a glycol product 52 that meets or exceeds specifications for the target glycol and its desired end-use. For example, where the target glycol is MEG, it is preferred that it meets or exceeds the specifications for fibre-grade MEG. The specifications for a fibre-grade MEG from a petroleum source are presented in Table 1.

TABLE 1

[0074] The specifications provided in Table 1 provide a maximum of 0.05 wt.% DEG in the purified MEG. This would apply to a petroleum feedstock. For a feedstock 12 derived from a renewable feedstock, the specifications would likely be adjusted to recite a maximum of 0.05 wt.% 1,2-BDO and/or 1,2-HDO.

[0075] The bottoms stream 54 may be partly or fully recycled, further processed for recovery of residual glycol or additional products, and/or disposed of. Alternatively, or in addition, the bottoms stream 54 may be burned as a fuel source for the process.

EXAMPLES

[0076] The following non-limiting examples of embodiments of the method of the present invention as claimed herein are provided for illustrative purposes only. While the examples were conducted using a retro-aldol catalyst containing tungsten, similar results are expected for retro-aldol catalysts containing other active metals. [0077] Experimental tests were conducted with liquid samples of effluent from a conventional reaction comprising catalytic hydrogenolysis of a saccharide-containing feedstock. The effluent was produced according to Example 1 A or IB.

EXAMPLE 1A

[0078] The catalytic hydrogenolysis reaction was conducted in a 1-L autoclave operated in continuous mode. The reactor was filled with 15.0 gram of WR Grace Raney-Ni 2800 (supplied by Aldrich-Sigma chemical company), which was washed prior to loading it into the reactor to remove any residual alumina. Water was added to the reactor to obtain a 60% liquid level. The reactor was closed and flushed with N2 to remove any oxygen prior pressurizing it under H2 to 104 barg (10.4 MPa(g)). A H2 flowrate of 75 NL/hr was applied.

[0079] A stirrer speed of 1250 RPM was applied to obtain good distribution of Raney-Ni in the reactor. A hot water wash was conducted at a temperature of 230°C by feeding water at 420 ml/hr, followed by a validation of the Raney-Ni hydrogenation activity.

[0080] Validation of hydrogenation activity was done by considering the glucose to sorbitol conversion for feeding a 10 wt.% glucose solution to the reactor at a feed rate of 300 g/hr at a reaction temperature of 70°C. The glucose feed was stopped after completing the validation, which typically takes 18 hrs, and the temperature was increased to the reaction temperature of 230°C. The feedstock was fed to the reactor with the retro-aldol catalyst, in the amounts shown in Table 2 once the target reaction temperature was achieved in the reactor.

TABLE 2

[0081] Total reactor pressure remained at 104 barg (10.4 MPa(g)) and H2 feed rate at 75 NL/hr throughout the experiment. Effluent samples from this part of the process are referred to herein as Example 1A effluent.

EXAMPLE IB

[0082] The remainder of the effluent from the hydrogenolysis reactor was subjected to a catalytic finishing hydrogenation step in a fixed bed reactor. The finishing reactor was filled with a catalyst bed with a total bed height of 24.5 inches (0.6 m) comprising a homogeneous mixture of 218 grams of 3 mm particles Ni-ZrCh catalyst diluted with 241.6 grams of SiC 20 grid particles. The Ni/ZrCh catalyst was reduced prior use in hydrogen (10 NL/hr) at 250°C (ramp 1.2 C°/hr) for 16 hrs. Effluent samples from this part of the process are referred to herein as Example IB effluent.

EXAMPLE 2

[0083] Two samples of Example 1A effluent were labelled as Sample A and Sample B, respectively. The tungstate concentration in the effluent was 3000 mg/kg. Sample A was untreated, while Sample B was contacted with AmberLite®IRA-900 ion exchange resin to reduce the tungstate concentration to 70 mg/kg.

[0084] Samples A and B were then subjected to rotary evaporation to remove water, resulting in a bench-top approximation of the embodiment of Fig. 4 for Sample B.

[0085] Samples A and B were heat treated (250°C, 240 minutes) in Hastelloy pressure cells, thereby approximating the effect of the heavies distillation column on the effluent.

[0086] Visually, the samples were very distinct, with Sample A showing a much darker colour after heat treatment, indicative of side reactivity having occurred within the sample, whilst Sample B was actually slightly lighter in colour. This suggested the depletion of tungstate catalyst in Sample B may have reduced the propensity for side reactions to occur.

[0087] The samples were analysed by HPLC, as well as GC-MS, with and without contacting with ion exchange resin, before and after heat-treating. GC-MS analysis of the samples showed the formation of furan to be significant only in the case of Sample A (with high levels of tungstate present). The results are presented in Tables 3 and 4.

[0088] Table 3 presents results of HPLC analysis of by-products of hydrogenolysis. The results show that all sugar alcohols are degraded to a certain extent in presence of W. Sorbitol degradation does occur in absence of tungstate but to a lesser extent than in the presence of tungstate (e.g., 41.6% versus 57.5%). The main advantage is that the resulting degradation products are other sugar alcohols that can still effectively being converted to glycols. The latter is based on the fact that all other sugar alcohol concentrations increased, while they decreased in presence of W. TABLE 3

[0089] Table 4 shows the recovery of glycol products. The results show that the glycols in absence of W are stable when undergoing the heat treatment (no significant loss of MEG, and perhaps even formation). In addition, potentially the glycols are less stable in presence of W.

TABLE 4

[0090] The results demonstrate that the retro-aldol catalyst can be removed from the effluent by contacting with ion exchange resin before heating to separate glycols from the heavies fraction, thereby decoupling any effects of downstream processing of the tungstate- free effluent from recycling the tungstate back to the conversion zone. In addition, the results demonstrate that there is less conversion of sugar alcohols to components that cannot be effectively converted to glycols when the effluent is first contacted with ion exchange resin before heating to separate glycols from the heavies fraction. Accordingly, more heavies fraction is available for valorisation when retro-aldol catalyst is depleted or removed before separation of gly col product from the heavies fraction. A further advantage of removing or reducing the amount of tungstate catalyst from the effluent at an early stage is that it has a tendency to precipitate particularly at more acidic pH causing clogging of lines and reducing overall process efficiency. A further advantage is that the absence of retro-aldol catalysts opens-up alternative ways to treat or dispose of the heavies fraction (e.g., by normal incineration)

EXAMPLE 3

[0091] Samples of Example 1A effluent were contacted with different ion exchange resin materials to test adsorption of tungstate from the effluent and desorption of the tungstate for recycling to the conversion zone.

[0092] The ion exchange materials tested include AmberLite™ IRA-410, Amberlyst™ A21, AmberLite™ IRA-900 and Amberlyst™ A26.

[0093] Batch isotherm tests were carried out to determine the tungsten loading capacity and recovery of the selected resins. The tested dosage range for the resins for adsorption was 12.5- and 15-mL resin per liter of feed solution. The volume of the resin is based on the conditioned resin. Conditioning was performed prior to each experiment, by soaking the resins in 10% (w/w) NaOH solution to convert the original exchangeable site into OH" form to ensure no chloride was present in the effluent. Afterwards, the conditioned resins were preserved in demineralized water (DW).

[0094] A feed solution volume of 200 mL was applied for each adsorption test condition. The different resin dosages were added to the feed solution in bottles, which were shaken on an orbital shaker for 6 hours to ensure sorption equilibrium.

[0095] Samples were taken of the supernatant for measurement of the tungsten equilibrium concentration ce (mg-W/L), and accordingly, the anionic tungstate species removal and the resin loading capacity qe (mg/mL) were determined based on the equation qe = ((c0 - ce ) * V feedyVresin where: cO (mg-W-L -1 ) is the initial tungstate concentration;

Vfeed (L) is the feed volume, (i.e., 0.2 L); and Vresin (mL) is the applied resin volume.

[0096] Prior to the desorption tests, the resin dosages 12.5 and 15 mL/L were rinsed with deionized water. The regeneration was performed with a 4% (w/w) NaOH solution, considering a regenerant volume four times the resin volume. The bottles of the desorption tests were also shaken for 6 hours on an orbital shaker. Like the calculation of the adsorbed tungstate, the desorbed tungstate is expressed as desorbed mg-metal per mL of resin. Furthermore, the desorption efficiency of the resin was calculated, which is based on the eluted tungsten in relation to the loaded tungstate of the corresponding resin dosage. Based on the results, a comparative assessment was made between the resins. The results are presented in Fig. 6.

[0097] W isolated from product is defined as the percentage of tungsten that was transferred from the feed solution to the brine. This was calculated by multiplying the anionic tungstate species adsorption (%) with the anionic tungstate species desorption (%).

EXAMPLE 4

[0098] A 2-meter column with internal diameter of 1” (2.5 cm) was filled to about 60 vol.% with AmberLite™ IRA-410 resin. This resin comes in Cl form, and it was therefore first treated with a 4 wt.% NaOH solution to exchange the Cl ions with OH ions. Example IB effluent was contacted with this resin to remove the W. It was found that about 40 BVs (bed volumes) of effluent could be treated before breakthrough of the anionic tungstate species was observed (e.g., tungstate > 100 ppm via XRF analysis, 96% adsorption). 100 Liters of anionic tungstate species-free (tungstate<100 ppm) effluent was produced this way.

[0099] Anionic tungstate species-saturated resin was used in a dedicated bench scale column to study the desorption characteristics of tungstate. 342 ml of saturated resin was charged into the column. Removal of air pockets was ensured via water backwashing. Resin regeneration (W desorption) was conducted by treating the bed with a 4 wt.% NaOH solution. [0100] The regeneration was conducted at a flowrate of 0.75 BV/hr (257 ml/hr) to ensure plug flow' conditions and a contact time of 32 minutes. Regeneration was considered complete when 10 BV (3.4 L) were processed.

[0101] The results showed that 98+ wt.% of anionic tungstate species could be recovered by desorption after processing 10 BVs of 4 wt.% NaOH solution. Anionic tungstate species is preferably recovered at aNa/W molar ratio <3, preferably as close to 2 (as Na2WO4 has aNa/W molar ratio of 2). The Na/W ratio increased over time due to increased slip of NaOH solution when the resin becomes less saturated with anionic tungstate species. The majority of the tungsten (~60 wt.%) was recovered in the first BVs (0.5 - 2 BV), with maximum anionic tungstate species concentrations in the order of 70-85 g/kg, and with the Na/W molar ratio during was in the desired order of 2.0-2.5, reflecting the recovery' of pure Na2WOr especially at the beginning of the experiment. Acid analysis indicated a maximum contribution of Na- carboxylates to the Na/W ratio of 0.2.

[0102] By combining the brines of the first two BV’s, it was possible to obtain a brine with an anionic tungstate species concentration of 61.8 g/kg and with a Na/W molar ratio of 2.5. This represents an anionic tungstate species recovery of 60 wt.%, and thereby reflects the w orking capacity of the resin. [0103] This blend was used for studying the effect of recycling the recovered retro-aldol catalyst directly to the main reactor. The results in Table 5 show that the retro-aldol catalyst is still very active, giving total glycol (MEG + MPG + 12BD0) yields of 65+ wt.%.

TABLE 5

[0104] Further advantages of using ion exchange material to recover/reuse retro-aldol catalysts before separation of a glycol product stream from the heavies fraction unlocks the possibility to use NaOH to control pH, while not affecting the Na/W molar ratio of the recovered retro-aldol catalyst in the brine (target Na/W molar ratio is 2, or at least < 3). Further, the advantage of the increased sugar alcohol stability in downstream processing (distillation) allows for higher yields in the heavies cracking reactor.