Process for reducing the amount of 2,3-pentanediol from propylene glycol

Abstract

A process for removing 2,3-pentanediol from a mixture rich in propylene glycol, which process comprises extractive distillation using an entrainer in said extractive distillation having a Hansen solubility parameter δH of between 5 and 15 and the entrainer having a boiling point at atmospheric pressure of at least 200°C.

Description

[0001] PROCESS FOR REDUCING THE AMOUNT OF 2,3-PENTANEDIOL FROM PROPYLENE GLYCOL

[0002] Introduction

[0003] The present invention relates to a process for reducing the amount of 2,3-pentanediol from a mixture rich in propylene glycol, and which mixture comprises a small amount of 2,3-pentanediol. The process comprises extractive distillation using an entrainer in said extractive distillation having a Hansen solubility parameter 6H of between 5 and 15 and the entrainer having a boiling point at atmospheric pressure of at least 200°C.

[0004] Background of the invention

[0005] Since about a decade, there is ongoing research in developing processes for manufacturing chemical building blocks from renewable sources. One such process of interest is obtaining monoethylene glycol (in short: ethylene glycol, MEG, or EG) from a (renewable) carbohydrate source such as sugar.

[0006] WO 2016 / 114661 discloses a continuous process for preparation of ethylene glycol from a carbohydrate source. Said process is carried out in a stirred tank reactor (CSTR) in which a catalyst system is present. Said catalyst system comprises a tungsten compound and at least one hydrogenolysis metal. The hydrogenolysis metal is preferably present in the form of a catalyst supported on a carrier. Such heterogeneous catalyst particles can fairly easily be separated from the effluent stream and added back.

[0007] The liquid effluent obtained is usually subjected to a series of separation steps, e.g. to separate the tungsten compound used as co-catalyst (or a tungsten compound that is formed by the process), but also to separate the various (volatile) alcohols, polyols and glycols produced. Generally, the target product in the processes like the reference above and similar processes is ethylene glycol. Whilst such processes may have a selectivity for ethylene glycol of about 40 to 70%, considerable amounts of other components are also produced, mainly monopropylene glycol (MPG), but also polyols like glycerol, sorbitol and erythritol. The mixture produced usually contains also other glycols and diols next to MEG and MPG, such as 1,2- butanediol, 1,4-butanediol, 1,2-pentanediol and 2,3-pentanediol. Although the typical way to separate miscible liquids on an industrial scale would be fractional distillation, such method has its limits in separating some of these glycols, as boiling points are often close together, and other glycols or polyols to be separated tend to form azeotropes with either ethylene glycol or propylene glycol, which makes separation by distillation even more difficult. Whilst the purification of ethylene glycol in the context of the conversion of carbohydrates to ethylene glycol has been widely researched, there are fewer publications dealing with removing minor compounds from propylene glycol.

[0008] As mentioned, mono propylene glycol (for short: MPG, or propylene glycol, or PG) is also produced in substantial quantities when aiming for producing ethylene glycol from carbohydrates with hydrogenolysis. Depending on the source of carbohydrates and processing conditions, the weight ratio of EG : PG produced may range from 10 : 1 to 1 : 1. Propylene glycol is a valuable product as well, although the market is smaller than ethylene glycol. Since it has value, in particular in pure form, and since it is produced in substantial amounts, it is desirable to be able to obtain the propylene glycol in a fairly pure form. However, among the side products produced by the hydrogenolysis of carbohydrates is 2,3- pentanediol. Depending on the source of carbohydrates and processing conditions, the amount of 2,3- pentanediol produced may be between 0.2 and 3% by weight on propylene glycol. In fact, four different stereoisomers of 2,3-pentanediol exist due to the presence of two chiral centres (2 and 3 position). These four stereoisomers are: (2R,3R) 2,3-pentanediol, (2S,3S) 2,3-pentanediol, (2R,3S) 2,3-pentanediol, and (2S,3R) 2,3-pentanediol. As far as is known to the inventors, all four forms are produced by said reaction. Of these four, the (2R,3R) and (2S,3S) enantiomers have a boilingpoint (at atmospheric conditions) of about 175°C, and these can be separated fairly easy from monopropylene glycol, which has a boilingpoint of about 188.2°C (at atmospheric conditions). However, the (2R,3S) and (2S,3R) enantiomers have a boilingpoint of about 187.5-189.5°C, which makes it very difficult if not impossible to separate these components from monopropylene glycol (bp 188.2°C) by straightforward fractional distillation to a reasonable degree. For a monopropylene glycol produced by hydrogenolysis of carbohydrates it is desired that not only the (2R,3R) and (2S,3S) stereoisomers are removed to a substantial degree, but also a reduction in the (2R,3S) and (2S,3R) stereoisomers is achieved.

[0009] WO 2017 / 050847 discloses a process for extractive distillation to remove 1,2-butanediol from ethylene glycol, using as extractant C3-C6 sugar alcohols, e.g. glycerol. In the process of this reference, the extractive distillation is such that the ethylene glycol is extracted by the extractant: 1,2-butanediol and propylene glycol (and side products 2,3 butanediol, 1,2 pentanediol, 2,3-pentanediol, 1,2-hexanol and 1,2- heptanol) are removed at the top of the extraction distillation column, and at the bottom the mixture of ethylene glycol and extractant are obtained. Said bottom mixture is then subjected to a separation to yield ethylene glycol without 1,2-butanediol and extractant. It should be noted that this is exemplified in this reference by modelling only, not by experiments actually carried out. The description and figure 2 suggest that the top stream may be separated by (fractional) distillation in a stream of purified propylene glycol and a stream of the mentioned side products. The examples do not give any evidence, and given the closeness of the boiling points of 2,3-pentanediol (that is, the mixture of (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol) and propylene glycol (187-189.5°C and 188.2°C, respectively) this is not possible in practice. Another disadvantage of the method of WO 2017 / 050847 is that the extractant is to be separated from the main product (ethylene glycol) which means that the stream that is processed in the extractant recovery column has to handle a big volume: all extractant and all ethylene glycol. This has consequences from an investment point of view (larger equipment) and operational-cost point of view (large volumes which have to be heated and cooled) and hence is undesired.

[0010] WO2023 / 041842 discloses a process for reducing the amount of MEG and butanediols (1,2- and 2,3-) from a composition containing MPG. The model example contains about 53% MPG, 21% MEG, 4% butanediols, and about 0.6% of an organic impurity forming an azeotrope with MPG. The method relies on distillation in which a solvent is added. The solvent disclosed in the examples is tri-ethylene glycol (TEG). The ratio solvent : feed is high in the experiments (e.g. > 12 in experiments 9 and 10) and also the reflux ratio is high (between 5 and 50 in the experiments), which leads to high energy costs.

[0011] CN112920021 discloses a process for purification of 1,3-propanediol, by liquid-liquid extraction.

[0012] Hence, there is a desire for a process that allows reducing the amount of 2,3-pentanediol present in a mixture comprising at least 80% (preferably at least 90%) by weight of propylene glycol and from 0.05 to 5% by weight of 2,3-pentanediol (in particular (2R,3S)- and (2S,3R) 2,3-pentanediol). "Reducing" in this connection means: less of these pentanediols than were present in the glycol mixture prior to applying the separation process of the present invention, and preferably such should relate to (2R,3S)- and (2S,3R) 2,3- pentanediol). Preferably the reduction of 2,3-pentanediol should be such that the amount of 2,3- pentanediol in any ethylene glycol produced is 50% or less (preferably 30% or less, more preferably 1% or less, even more preferably 0.2% or less) of the amount of 2,3-pentanediol that was in the feed of the product (when taken on propylene glycol).

[0013] As to the propylene glycol produced, it is desired that a purification process is provided that can achieve a reduction in the amount of (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol (hereinafter called: (2R,3S), (2S,3R) 2,3-pentanediol) in a stream comprising predominantly propylene glycol of at least 30% by weight, preferably a reduction of at least 50% by weight. The percent reduction is then the amount of the enantiomers (2R,3S) and (2S,3R) combined, in the stream subjected to the purification after said purification, compared to the amount in such stream prior to said purification. It is furthermore desired that such purification method can be integrated with a separation process that can be used to separate a product stream from hydrogenolysis of carbohydrates that separates into an ethylene glycol-rich fraction and a propylene glycol rich fraction, wherein also the ethylene glycol rich fraction is subjected to a process that reduces the amount of 1,2-pentanediol, all in a practical manner with only a minimal number of chemical compounds needed to achieve the three separation / purification step (three steps as in: separation EG / PG, removal of 1,2-pentanediol from EG, and removal of (2R,3S), (2S,3R) 2,3-pentanediol from PG).

[0014] It is also preferred that the process will provide propylene glycol comprising less than 1% by weight (on propylene glycol) of the combination of (2R,3S) 2,3-pentanediol and (2S,3R) pentanediol, preferably such is less than 0.5%, by weight on propylene glycol.

[0015] The process should preferably such that the number of auxiliary compounds (e.g. extractants or solvents) is minimized, as this facilitates operation, and preferably the volume of streams to be processed should be minimized. It is furthermore desired that the number of unit operations required is as low as possible and that the energy consumption is preferably minimized and that the process preferably does not have the disadvantages of the prior art.

[0016] Summary of the invention

[0017] It has now been found objective that the above may be achieved, at least in part, by a process for reducing the content of 2,3-pentanediol in a mixture comprising 90-99.9% by weight propylene glycol and 0.05-5% by weight 2,3-pentanediol by extractive distillation, using an entrainer in said extractive distillation having a Hansen solubility parameter 6H of between 5 and 15 and the entrainer having a boiling point at atmospheric pressure of at least 200°C.

[0018] Preferably, the entrainer further has a Hansen solubility parameter 6P of between 2 and 12, preferably of between 3 and 11, more preferably between 3.7 and 10, most preferably between 3.8 and 9. Even more preferably, said entrainer further has a Hansen solubility parameter 6D of between 15 and 20, preferably of between 15 and 18. As a Hansen solubility parameter 6H, it is preferred that said entrainer has a Hansen solubility parameter 6H of between 6 and 14, more preferably of between 7 and 13. Suitable entrainers are preferably selected from a C6-C10 glycol ether, or a C9-C18 primary or secondary aliphatic unbranched alcohol, and mixtures thereof.

[0019] Since 2,3-pentanediols are inevitably formed by the hydrogenolysis of carbohydrates with hydrogen in the presence of a catalyst, and the (2R,3R) 2,3-pentanediol and (2S,3S) 2,3-pentanediol can be easily removed by conventional fractional distillation, yet only now the presently claimed process achieves removal of [(2R,3S) 2,3-pentanediol + (2S,3R) 2,3-pentanediol to a sufficient degree, the invention further relates to a propylene glycol composition comprising propylene glycol in an amount of at least 99.5 weight%, which further comprises 2,3-pentanediol in an amount of less than 0.5 weight %, and which 2,3-pentanediols have a weight ratio [(2R,3R) 2,3-pentanediol + (2S,3S) 2,3-pentanediol] : [(2R,3S) 2,3-pentanediol + (2S,3R)

[0020] 2.3-pentanediol] of between 1 : 1 to 1 : 20, preferably between 1 : 1 and 1 : 10.

[0021] Detailed description of the invention

[0022] The word "entrainer" herein encompasses pure compounds but also mixtures of compounds having the Hansen solubility parameters and boiling point claimed.

[0023] In the process according to the invention, it is referred to the two enantiomers (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol. There are also the two stereoisomers (2R,3R) 2,3-pentanediol and (2S,3S) 2,3- pentanediol. Without wishing to be bound by theory, it is believed that all four stereoisomers are formed in the hydrogenolysis of carbohydrates, although perhaps in unequal amounts. The enantiomers (2R,3R)

[0024] 2.3-pentanediol and (2S,3S) 2,3-pentanediol have a boiling point of about 175°C, and thus can be separated from propylene glycol with fractional distillation to a sufficient degree. However, the (2R,3S)

[0025] 2,3-pentanediol and (2S,3R) 2,3-pentanediol enantiomers have a boiling point that is very close to that of propylene glycol, and hence separation of these from propylene glycol requires a different technique, as now provided. Although the extractive distillation claimed also can remove part or all of the (2R,3R) 2,3- pentanediol and (2S,3S) 2,3-pentanediol, generally such components with a boiling point sufficiently lower than that of propylene glycol will have been removed already prior to the extractive distillation now claimed, e.g. by fractional distillation.

[0026] Inventors of the present case found that entrainers suitable for the purpose have in common that they have a Hansen solubility parameter 6H of between 5 and 15 (and preferably a Hansen solubility parameter amount of (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol) in propylene glycol prior to this invention was problematic, in particular when present in small amounts of e.g. 0.5-2%, due to the closeness of the boiling points of (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol and propylene glycol. With the present method, it was found that the amount of these 2,3-pentanediols present in a composition mainly comprising monopropylene glycol can be reduced to a substantial degree. For practical reasons (in view of the temperatures involved), the entrainer used in the present method has a boiling point of at least 200°C, at atmospheric pressure.

[0027] Surprisingly, it was also found that extractive distillation with the same entrainer can be used to reduce the content of 1,2-pentanediol present in a composition comprising predominantly (e.g. at least 90% by weight, preferably at least 95%) of ethylene glycol.

[0028] Hence, inventors of the present case found that the entrainers specified can be used as an entrainer in extractive distillation of ethylene glycol as well as in extractive distillation of propylene glycol. That the same entrainer can be used for two different separations is very convenient in a process such as the hydrogenolysis of carbohydrates that yields, next to predominantly ethylene glycol also propylene glycol in substantial amounts, and which ethylene glycol and propylene glycol are to be obtained in a relatively pure form. In such hydrogenolysis process side products are obtained, amongst which are 1,2-pentanediol and 2,3-pentanediol. The first generally ends up in the ethylene glycol stream and is difficult to remove to a sufficient degree from the desired ethylene glycol, whereas the latter (notably the enantiomers (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol) end up in the propylene glycol stream and can be difficult to remove from the desired propylene glycol.

[0029] In order to identify suitable entrainers, numerous chemical components were subjected to laboratory screening of the relative volatility of propylene glycol and 2,3-pentanediol, and how such is modified by the presence of an entrainer. This provides a prediction of the suitability of an entrainer for the purpose. The details are set out in example 2. Compounds that came out promising in this test were several glycol ethers, and some aliphatic alcohols, and all were found to fit the Hansen solubility parameters now claimed. Hence, it is preferred that the entrainer for the present invention is selected from a C6-C10 glycol ether, or a C9-C18 (preferably C10-C12) primary or secondary aliphatic unbranched alcohol, and mixtures thereof. A "C6-C10 glycol ether" herein means a glycol ether having from 6 to 10 (including 6 and 10) carbon atoms in its formula. Similarly, a "C9-C18 primary or secondary aliphatic unbranched alcohol" herein means a primary or secondary aliphatic unbranched alcohol containing from 9 to 18 carbon atoms.

[0030] Following this (and the requirement that the boiling point at atmospheric conditions should be at least 200°C), it is preferred that in the process of the present invention, the entrainer, when such is a glycol ether, such is selected from the group consisting of as appears in table 1, and mixtures thereof. Table 1 also gives the Hansen solubility parameters of the individual compounds.

[0031] Table 1

[0032] Hence, these are preferred glycol ethers in the present invention. Depending e.g. on cost, availability and safety, impurities, thermal stability, in case a glycol ether is used as entrainer, said glycol ether is preferably selected from the group consisting of triethylene glycol monoethyl ether, triethyleneglycol monobutyl ether, and mixtures thereof.

[0033] As alternative to glycols, the entrainers may also be a C9-C18 primary or secondary aliphatic unbranched alcohol, preferably C10-C12 primary or secondary aliphatic unbranched alcohol. Following this (and the requirement that the boiling point at atmospheric conditions should be at least 200°C), it is preferred that in the process of the present invention, the entrainer, when such is a alcohol such is selected from the group consisting of as appears in table 2, and mixtures thereof.

[0034] Table 2

[0035] Yet a further suitable entrainer for the purpose of the invention is triethylphosphate. This compound has a 6D of 16.7, a 6P of 11.4, and a 6H of 9.2.

[0036] Potential entrainers that are used in industry for similar purposes have been tested for the purpose of the invention, and these appeared to be unsatisfactory in the context of the invention. These are listed in table 3. Table 3

[0037] For reasons of e.g. cost, availability and safety, impurities, thermal stability, in case the entrainer is a primary or secondary aliphatic unbranched alcohol, preferred alcohols for such are 1-dodecanol and 2- decanol. Preferred glycol ethers in this invention are: triethyleneglycol monoethyl ether, triethyleneglycol monobutyl ether. Hence, it is preferred that in the present invention the entrainer is selected from triethyleneglycol monoethyl ether, triethyleneglycol monobutyl ether, 1-dodecanol and 2-decanol, and mixtures thereof.

[0038] The invention further relates to a process, wherein said process comprises: a) feeding said mixture to a distillation column 1, b) feeding to said distillation column 1 said entrainer, c) removing propylene glycol from the top section of the column, d) removing from the bottom section of the distillation column 1 a mixture comprising entrainer and 2,3-pentanediol (preferably: (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol).

[0039] Clearly, the entrainer will need to be freed of the 2,3-pentanediol in order to be reused again. Such process is a regeneration process of the entrainer. Hence, the invention further relates to a process comprising: feeding the mixture comprising entrainer and 2,3-pentanediol obtained in step d) to a column 2 yielding regenerated entrainer at the bottom section of column 2 and a mixture comprising entrainer and 2,3-pentanediol (preferably: (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol) at the top section of column 2, combining said regenerated entrainer with feeding of the entrainer of b) to column 1.

[0040] The distillation in column 1, to which is fed the (crude) propylene glycol to be purified (which thus contains some 2,3-pentanediol (preferably: (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol), e.g. between 0.05-5%, preferably between 0.1 and 3% by weight on the crude propylene glycol) can be carried out in a distillation column as known in the art.

[0041] Regarding this extractive distillation column 1 it is generally preferred for effective extractive distillation that the feed of the stream to be purified is neither added near the top, nor near the bottom section of column 1. Preferably, the feed is added in the middle half of the column (between the top quarter and the bottom quarter). This is to be understood as: if column 1 has 100 theoretical stages, the feed to be purified is preferably added between stages 25 and 75 of this column. Also, it is generally preferred that the entrainer is added above the feed. Hence, in the process according to the present invention, it is preferred that feeding the mixture comprising propylene glycol and 2,3-pentanediol to column 1 in a) is at a stage N of the column and wherein the feeding of the entrainer to column 1 in b) is at a stage M of the column, wherein stage M is above stage N.

[0042] As the entrainer is preferably to be fed at a point in the column above that of the feed of the crude propylene glycol, said distillation column 1 preferably has two inlets. The distillation column 1 is connected to a reboiler at the bottoms section and a condenser at the top section.

[0043] In the process according to the present invention the weight ratio of entrainer fed to column 1 : mixture fed to column 1 is preferably from 15 : 1 to 1: 1 for a good operation, such e.g. depending on the amount of 1,2-pentanediol to be removed, and the entrainer used. More preferably, such ratio is between 12 : 1 and 2 : 1.

[0044] By employing the process of the invention, the propylene glycol obtained in c) preferably has an amount of 2,3-pentanediol (notably: (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol) which is less than 50% (preferably: less than 30%) by weight of the amount of 2,3-pentanediol (preferably: (2R,3S) 2,3- pentanediol and (2S,3R) 2,3-pentanediol) of the mixture fed to distillation column 1 in a).

[0045] The process of this invention has the advantage, compared to the prior art processes, that for the extractive distillation of two different products (ethylene glycol and propylene glycol) this can be achieved using a single entrainer. As for the ethylene glycol concerned, this is obtained as a relatively pure component (containing less 1,2, -pentanediol that the feed) without being mixed with the entrainer as some prior art methods. Rather, the entrainer contains the component to be removed: 1,2-pentanediol.

[0046] The separation process according to the present invention is preferably carried out downstream of a hydrogenolysis reactor in which carbohydrates are converted with hydrogen, in the presence of catalysts, to ethylene glycol. Propylene glycol is usually a by-product, next to 1,2-pentanediol, but in larger quantities than 1,2-pentanediol. Preferably, in the process according to the present invention, prior to the now claimed separation process there are one or more other separation processes. Downstream of the reactor and prior to the now claimed separation process there is preferably first a removal of volatiles and water. Furthermore, there is preferably also a separation of high boiling side products (higher boiling than ethylene glycol and propylene glycol) such as glycerol and erythritol. Also, it is preferred that there is already a separation of ethylene glycol and propylene glycol preceding the now claimed purification process. All above means that the stream of propylene glycol to be purified is smaller. Hence, it is preferred in the present invention that the mixture fed to column 1 in a) comprises less than 1 % by weight of ethylene glycol.

[0047] To effect the desired extractive distillation, it may be preferred to operate column 1 at such conditions that the reboiler of column 1 operates at a pressure of between 20 and 100 mbara, preferably at a pressure of between 40 and 90 mbara, and wherein the condenser of column 1 operates at a temperature of between 70 and 140°C, preferably between 90 and 130°C, and wherein the reboiler of column 1 operates at a temperature of between 150 and 240°, preferably between 170 and 210°C. The invention provides the benefit that low reflux ratios are possible. Hence, it is preferred in the present invention that column 1 operates with a reflux ratio of between 0.5 and 3, preferably between 0.7 and 2.4.

[0048] As to the column 2 for regenerating the entrainer, it is desired that of column 2 the reboiler is operated at a temperature of between 150 and 220°, preferably between 165 and 210°C, and at a pressure of between 20 and 400 mbara, preferably at a pressure of between 30 and 200 mbara, and wherein the condenser of column 2 in c) is operated at a temperature of between 90 and 180°C, preferably between 100 and 160°C.

[0049] In the process of the invention it is preferred that the propylene glycol mixture comprises propylene glycol and (2R,3S) 2,3-pentanediol and / or (2S,3R) 2,3-pentanediol in a weight ratio propylene glycol : (2R,3S) 2,3- pentanediol and / or (2S,3R) 2,3-pentanediol of between 200 : 1 and 50 : 1.

[0050] In step b), the weight ratio entrainer : propylene glycol of between 15 : 1 and 4 : 1, more preferably between 10 : 1 and 5 : 1. In step c), the weight ratio entrainer : propylene glycol of between 15 : 1 and 4 : 1, more preferably between 10 : 1 and 5 : 1.

[0051] With the process according to the present invention, it is possible to obtain a propylene glycol which has a reduced amount of (2R,3S) 2,3-pentanediol and / or (2S,3R) 2,3-pentanediol when compared with the feed to the process (when considered on the basis of propylene glycol). The process according to the invention thus may yield a top stream of column 1 in b) which comprises at least 98% by weight (based on the weight of the top stream) of propylene glycol and less than 2% by weight of (2R,3S) 2,3-pentanediol and / or (2S,3R) 2,3-pentanediol. Preferably, this purified propylene glycol top stream of column 1 comprises at least 99 % by weight, more preferably at least 99.5% by weight of propylene glycol, and such stream comprises preferably less than 1%, more preferably less than 0.5% by weight of (2R,3S) 2,3- pentanediol and / or (2S,3R) 2,3-pentanediol combined.

[0052] EXAMPLES

[0053] Example 1

[0054] Example 1 is a laboratory screening of the relative volatility of ethylene glycol and 1,2-pentanediol, and how such is modified by the presence of an entrainer. This provides a prediction of the suitability of an entrainer for the purpose. The relative volatility is determined by gas chromatography of the headspace.

[0055] Mixtures of components (ethylene glycol, 1,2-pentanediol and entrainer) were prepared in ratios presented in table 4 below. A 5 mL volume of the mixture was inserted into a 10 mL vial. The vial was purposefully left half empty to ensure proper volume of headspace available for achieving liquid-vapor equilibrium. The entrainers tested were: Triethylene glycol monobutyl ether, Propylene glycol monophenyl ether (l-Phenoxy-2-propanol), Dodecanol, Triethylphosphate, Diethylene Glycol Monoethyl Ether Acetate, 2-Decanol, diethylene glycol monobutyl ether, Di(ethylene glycol) hexyl ether, Tripropylene glycol methyl ether, Dipropylene glycol methyl ether, Di(ethylene glycol) divinyl ether, Ethylene glycol monobenzyl ether.

[0056] As comparative entrainers: Dipropylene glycol, Glycerol, Triethylene glycol, Diethylene glycol, Octadecane, Decane, Dodecane.

[0057] Table 4

[0058] After preparing the vial, the content of the vial was thoroughly mixed to ensure homogeneous distribution of content in the liquid phase. Subsequently, the vials were placed in a Headspace-Gas chromatography (HS-GC) autosampler where the vials were heated to 130°C. After 60 minutes, the vial's headspace which contains the gas content was sampled and analyzed by the machine. This injection resulted in a chromatogram where the separated peaks each represent one of the mixtures' components, and the area under the peak indicates their respective concentration in the gas phase.

[0059] The obtained areas are proportional to the components partial pressure and the relative volatility (a) was calculated as: wherein:

[0060] Anis HS-GCMS area of component n with entrainer addition An,o is HS-GCMS area of component n in reference sample (no entrainer addition)

[0061] Pn,sat is saturation pressure of component n xnis molar fraction of component n in liquid phase

[0062] 1 is index for component 1 (here: ethylene glycol)

[0063] 2 is index for component 2 (here: 1,2-pentanediol)

[0064] The relative volatility in distillation is an indication of feasible separation of the components in a mixture. Higher relative volatility shows that the relative concentration of component 1 to 2 is higher in the gas phase than in liquid phase. Hence, in a single stage separation such as in this test for triethylene glycol monobutyl ether, the ethylene glycol is shown to be more volatile than the 1,2-pentanediol in the presence of triethylene glycol monobutyl ether. And so on for other entrainers.

[0065] As a point of reference, the relative volatility of the mixture without the entrainer present shows that almost no separation of the two components is possible (relative volatility =~1). The results of the relative volatility a are set out in figure 1 (EG means ethylene glycol) for the compounds according to the invention and figure 2 for comparatives.

[0066] The higher the relative volatility of a given compound in this test the better the suitability of this compound as entrainer in extractive distillation. There is no hard cut-off, but the lower the relative volatility in this test the more entrainer (in relation to the ethylene glycol) would be needed to remove 1,2-pentanediol from ethylene glycol, or the 1,2-pentanediol is not removed to the same extent. A lower amount of entrainer is desired for operational and economic reasons.

[0067] Example 2

[0068] In the same way as in example 1, now the suitability of the same glycol ethers for removing 2,3- pentanediol from mono propylene glycol was tested. This included screening the relative volatility of propylene glycol and 2,3-pentanediol, and how such is modified by the presence of an entrainer. This provides a prediction of the suitability of an entrainer for the purpose. A 10 mL volume of the mixture was inserted into a 20 mL vial. The vial was purposefully left half empty to ensure proper volume of headspace available for achieving liquid-vapor equilibrium. The relative volatility is determined by gas chromatography of the headspace as in example 1.

[0069] There was a difference in which mixtures were tested, as 2,3-pentanediol was not available in a pure enough form (only up to 95%) for results to be reliable. Hence, an experiment like that of experiment 2 of example 1 could not be performed for 2,3-pentanediol. For this reason, a reference sample was taken from a hydrogenolysis experiment from which the propylene glycol was obtained ("crude MPG"), which contained (as appeared from analysis) about 0.72 wt% of 2,3-pentanediol (and from this the amount of 2,3-pentanediol was calculated, which is assumed to be a mixture of all four stereoisomers). This combination is denoted "crude MPG" in the table 5 below.

[0070] Mixtures of components were prepared in rations presented in table 5 below. Thereafter, the same methodology as in example 1 was followed.

[0071] Table 5

[0072] The relative volatility a can be calculated in a similar way as for example 1, and the results of this is set out in figure 3. It shows that of the compounds tested, the glycol ethers are the most suitable compounds (of the compounds tested) as entrainer for extractive distillation to remove 2,3-pentanediol from propylene glycol.

[0073] Example 3

[0074] For this experiment, a trial was conducted in a distillation column having maximum 100 theoretical stages, with triethylene glycol monobutyl ether (BTEG) as entrainer, on a model feed.

[0075] The composition of the model feed was: a mixture of 97.72 wt% ethylene glycol, 1.68 wt% 1,2- pentanediol and 0.54 wt% 1,2-hexanediol and some minor components (other impurities).

[0076] In continuous operation mode, the ethylene glycol-containing feed mixture (at 78-85°C and ambient pressure) and triethylene glycol monobutyl ether entrainer (at 111-112°C and ambient pressure) were introduced (the entrainer higher than the ethylene glycol feed mixture) into a maximum 100 theoretical stage column where the purified ethylene glycol is separated in the distillate and the impurities of the feed leave the column with the entrainer as bottom stream.

[0077] The column was operated at 116°C and 50 mbara condenser conditions, with a column pressure drop of 13.5-15 mbara. A reflux ratio of 2 was used throughout the experiment.

[0078] Three experiments were performed where the entrainer to feed ratio was varied, according to the data in table 6 below. The table also lists the results. The amount of 1,2-pentanediol in the top stream was below the detection limit of 0.001 wt%.

[0079] Table 6 Note: "sep mass % in bottom stream" refers the percentage of the stated component that is present in the bottom stream, based on the amount of said component in the EG feed stream.

[0080] Example 4

[0081] In an experiment similar as in the previous example, a model propylene glycol was made containing a commercial 2,3-pentanediol. The model feed contained: 98.684 wt% propylene glycol, 0.758 wt% mixture of (2R,3R)- and (2S,3S) 2,3-pentanediol, and 0.558 wt% mixture of (2R,3S)- and (2S,3R) 2,3-pentanediol. In continuous operation mode, the propylene glycol-containing feed mixture (at 78-85°C and ambient pressure) and triethylene glycol monobutyl ether (BTEG) entrainer (at 111-113°C and ambient pressure) was introduced (the entrainer higher than the propylene glycol feed mixture) into a maximum 100 theoretical stage column where the purified propylene glycol was separated in the distillate and the impurities of the feed leave the column with the entrainer as bottom stream.

[0082] The column was operated at 108-109°C and 50 mbara condenser conditions, with a column pressure drop of 12.8-18.1 mbara. A reflux ratio of 2 was used throughout the experiment.

[0083] The results are set out in table 7.

[0084] Table 7

[0085] Note: the purity for the MPG in the top stream would be higher than the percentage indicated if no (2R,3R)- and (2S,3S) 2,3-pentanediol had been present in the model MPG composition. The (2R,3R)- and (2S,3S) 2,3-pentanediol were present in the model composition as no commercial composition of 2,3- pentanediol was easily available at the time of the experiments that was free of these stereoisomers (and would only contain (2R,3S)- and (2S,3R) 2,3-pentanediol), and the process of this invention is better suited to remove the (2R,3S)- and (2S,3R) 2,3-pentanediol (which is the aim) than to remove the (2R,3R)- and (2S,3S) 2,3-pentanediol.

[0086] In an actual commercial process of manufacturing MPG from carbohydrates most of the (2R,3R)- and (2S,3S) 2,3-pentanediol would have been removed before performing the extractive distillation of this invention, as these stereoisomers have a boiling point which is sufficiently lower than that for MPG to allow fractional distillation. Hence, in such set up the purity of MPG in the top stream of this extractive distillation would be higher than the numbers here suggest.

Claims

CLAIMS1. Process for reducing the content of 2,3-pentanediol in a mixture comprising 90-99.9% by weight propylene glycol and 0.05-5% by weight 2,3-pentanediol by extractive distillation, using an entrainer in said extractive distillation having a Hansen solubility parameter 6H of between 5 and 15 and the entrainer having a boiling point at atmospheric pressure of at least 200°C.

2. Process according to claim 1, wherein said entrainer further has a Hansen solubility parameter 6P of between 2 and 12, preferably of between 3 and 11, more preferably between 3.7 and 10, most preferably between 3.8 and 9.

3. Process according to any of the preceding claims, wherein said entrainer further has a Hansen solubility parameter 6D of between 15 and 20, preferably of between 15 and 18.

4. Process according to any of the preceding claims, and wherein said entrainer has a Hansen solubility parameter 6H of between 6 and 14, more preferably of between 7 and 13.

5. Process according to any of the preceding claims, wherein the entrainer is selected from a C6-C10 glycol ether, or a C9-C18 primary or secondary aliphatic unbranched alcohol, and mixtures thereof.

6. Process according to claim 5, wherein the entrainer is selected from triethyleneglycol monoethyl ether, triethyleneglycol monobutyl ether, 1-dodecanol and 2-decanol, and mixtures thereof.

7. Process according to any of the preceding claims, wherein the 2,3-pentanediol comprises at least one of (2R,3S) 2,3-pentanediol and (2S,3R) 2,3-pentanediol.

8. Process according to any of the preceding claims, wherein said process comprises: a) feeding said mixture to a distillation column 1, b) feeding to said distillation column 1 said entrainer, c) removing propylene glycol from the top section of the column, d) removing from the bottom section of the distillation column 1 a mixture comprising entrainer and 2,3-pentanediol.

9. Process according to claim 8, feeding the mixture comprising entrainer and 2,3-pentanediol obtained in step d) to a column 2 yielding regenerated entrainer at the bottom section of column 2 and a mixture comprising entrainer and 2,3-pentanediol at the top section of column 2, combining said regenerated entrainer with feeding of the entrainer of b) to column 1.

10. Process according to claim 8 or 9, wherein the propylene glycol obtained in c) has an amount of 2,3-pentanediol which is 5 to 60% by weight of the amount of 2,3-pentanediol of the mixture fed to distillation column 1 in a).

11. Process according to any of claims 8 to 10, wherein feeding the mixture comprising propylene glycol and 2,3-pentanediol to column 1 in a) is at a stage N of the column and wherein the feeding of the entrainer to column 1 in b) is at a stage M of the column, wherein stage M is above stage N.

12. Process according to any of claims 8 to 11, wherein the weight ratio of entrainer in b) to the mixture in a) is from 15 : 1 to 1: 1, preferably between 12 : 1 and 2 : 1.

13. Process according to any of claims 8 to 12, wherein the reboiler of column 1 operates at a pressure of between 20 and 100 mbara, preferably at a pressure of between 40 and 90 mbara, and wherein the condenser of column 1 operates at a temperature of between 70 and 140°C, preferably between 90 and 130°C, and wherein the reboiler of column 1 operates at a temperature of between 150 and 240°, preferably between 170 and 210°C.

14. Process according to any of claims 8 to 15, wherein column 1 operates with a reflux ratio of between 0.5 and 3, preferably between 0.7 and 2.4.

15. A propylene glycol composition comprising propylene glycol in an amount of at least 99.5 weight%, which further comprises 2,3-pentanediol in an amount of less than 0.5 weight %, and which 2,3-pentanediols have a weight ratio[(2R,3R) 2,3-pentanediol + (2S,3S) 2,3-pentanediol] : [(2R,3S) 2,3-pentanediol + (2S,3R) 2,3- pentanediol] of between 1 : 1 to 1 : 20, preferably between 1 : 1 and 1 : 10.

Images