Field of the invention

The present invention relates to a process for the epoxidation of propene with hydrogen peroxide in the presence of a titanium silicalite catalyst. Background of the invention

The epoxidation of propene with hydrogen peroxide in the presence of a titanium silicalite catalyst is known from EP 0 100 1 19 A1. The reaction of propene with hydrogen peroxide in the presence of a titanium zeolite catalyst is usually carried out in a methanol solvent to achieve high reaction rates and product selectivity. In addition to propene oxide, the epoxidation reaction produces byproducts, such as formaldehyde, acetaldehyde, and hydroperoxides formed by ring opening reactions of hydrogen peroxide with propene oxide.

By-products acetaldehyde and propionaldehyde are difficult to separate from the propene oxide product. WO 2004/048355 discloses a method for removing methanol and acetaldehyde from a crude propene oxide in a single distillation column by an extractive distillation where a compound containing an unsubstituted NH2 group and capable of reacting with acetaldehyde at the conditions of distillation is additionally fed at or above the feeding point of the crude propene oxide. An aqueous hydrazine solution is preferably used as the additionally fed compound. Water is particularly preferred as the extraction solvent. The method provides propene oxide of high purity suitable for making polyether polyols.

WO 03/093255 teaches to hydrogenate a solvent stream, recovered from epoxidation of an olefin with hydrogen peroxide, with a heterogeneous catalyst at conditions, where non-reacted hydrogen peroxide, formaldehyde, acetaldehyde, and hydroperoxides, such as 1-hydroperoxy-2-propanol and 2-hydroperoxy-1-propanol formed in the epoxidation reaction, are hydrogenated before recycling the solvent to the epoxidation reaction. WO 03/093255 teaches in this context that the impurities methyl formate, formaldehyde, acetaldehyde, dimethoxymethane and

1 , 1-dimethoxyethane lead to deactivation of the catalyst.

WO 2004/029032 teaches to carry out epoxidation of olefins with hydrogen peroxide in the presence of a titanium-containing zeolite catalyst in an aqueous reaction mixture comprising less than 100 wppm of strong bases or cations of such bases having a ρΚβ of less than 4.5 and at least 100 wppm weak bases or cations of such bases having a ρΚβ of at least 4.5. Limiting the amount of strong bases reduces or prevents long term deactivation of the catalyst, whereas the presence of a weak base improves the selectivity for the epoxide without affecting the long-term activity of the catalyst. Organic amines are strong bases which usually have a ρΚβ of less than 4.5 and therefore introducing such amines into the epoxidation step with a recycle solvent stream has to be avoided in order to maintain long term activity and selectivity of the epoxidation catalyst. WO 2004/048354 teaches recovery of a solvent stream from the reaction mixture of an epoxidation of an olefin with hydrogen peroxide in the presence of a titanium-containing zeolite catalyst, where the recovered solvent stream is treated to contain less than 50 wppm nitrogen in the form of organic nitrogen compounds before it is recycled to the epoxidation step, in order to reduce deactivation of the catalyst upon recycling of the solvent. The solvent treatment is preferably an acid treatment. WO 2004/048354 teaches that the acid treatment can be carried out by adding a carboxylic acid or a mineral acid to the solvent stream before or during a distillation recovering the solvent as an overhead product or by treating an overhead product obtained by distillation with an acidic ion exchanger. Summary of the invention

The inventors of the present invention have now found that subjecting a methanol solvent stream, recovered from epoxidation of propene, to a hydrogenation as described in WO 03/093255 followed by removal of organic nitrogen compounds from the hydrogenated solvent stream by adding an acid to the hydrogenated solvent stream before or during distillation for recovery of methanol as an overhead product will provide a recovered methanol that may contain more acetaldehyde than the hydrogenated solvent stream contained, if the hydrogenation does not convert all acetaldehyde and acetaldehyde acetals. Recycle of methanol, recovered this way, to the epoxidation reaction leads to deactivation of the catalyst. The inventors of the present invention have further found that treating the methanol recovered this way by additionally passing it through a bed of an acidic ion exchanger will convert most of the acetaldehyde to 1 , 1-dimethoxyethane and that recycling methanol treated this way prevents deactivation of the epoxidation catalyst, contrary to what one would expect from the teachings of WO 03/093255 on 1 , 1-dimethoxyethane causing catalyst deactivation.

Subject of the invention is therefore a process for the epoxidation of propene comprising the steps a) reacting propene with hydrogen peroxide in the presence of a methanol solvent and a titanium zeolite epoxidation catalyst to provide a reaction mixture,

b) separating from the reaction mixture of step a) a crude propene oxide and a solvent

mixture comprising methanol, water and peroxides,

c) subjecting the solvent mixture separated in step b) to a catalytic hydrogenation for

hydrogenating said peroxides, providing a hydrogenated solvent mixture comprising from

1 to 1000 mg/kg of acetaldehyde,

d) separating the hydrogenated solvent mixture of step c) in at least one distillation stage, adding an acid to the hydrogenated solvent mixture of step c) or to at least one distillation stage, providing a recovered methanol as an overhead product,

e) passing the recovered methanol of step d) through a bed of an acidic ion exchange resin, providing a treated methanol, and

f) recycling the treated methanol of step e) to step a). Brief description of drawings

Figure 1 shows the concentrations of acetaldehyde in the hydrogenated solvent mixture (A) and the methanol recovered by distillation before (B) and after (C) treatment with an ion exchange resin determined in example 2. Detailed description of the invention

In step a) of the process of the invention, propene is reacted with hydrogen peroxide in the presence of a methanol solvent and a titanium zeolite epoxidation catalyst to provide a reaction mixture.

Propene is preferably used in a molar excess to hydrogen peroxide, preferably at a molar ratio of propene to hydrogen peroxide of from 1.1 : 1 to 30: 1 , more preferably 2: 1 to 10: 1 and most preferably 3: 1 to 5: 1. In a preferred embodiment, propene is used in an excess sufficient to maintain an additional liquid phase rich in propene throughout step a). The propene may contain propane, preferably with a molar ratio of propane to propene of from 0.001 to 0.15 and more preferably of from 0.08 to 0.12.

Hydrogen peroxide can be used as an aqueous solution, preferably containing from 30 to 75 % by weight hydrogen peroxide and most preferably from 40 to 70 % by weight. The aqueous hydrogen peroxide solution is preferably made by an anthraquinone process.

The methanol solvent can be a technical grade methanol, a solvent stream recovered in the workup of the epoxidation reaction mixture or a mixture of both. The methanol solvent may comprise other solvents in minor amounts, such as ethanol, with the amount of such other solvents preferably being less than 2 % by weight. The methanol solvent may also comprise water, preferably from 2 to 8 % by weight water. The methanol solvent is preferably used in the epoxidation in a weight ratio of 0.5 to 20 relative to the combined weight of water and hydrogen peroxide.

The epoxidation catalyst used in step a) preferably comprises a titanium zeolite containing titanium atoms on silicon lattice positions. Preferably, a titanium silicalite catalyst is used, preferably with an MFI or MEL crystal structure. Most preferably a titanium silicalite-1 catalyst with MFI structure as known from EP 0 100 1 19 A1 , is used. The titanium silicalite catalyst is preferably employed as a shaped catalyst in the form of granules, extrudates or shaped bodies. For the shaping process the catalyst may contain 1 to 99% of a binder or carrier material, all binders and carrier materials being suitable that do not react with hydrogen peroxide or with propene oxide under the reaction conditions employed for the epoxidation, silica being preferred as binder. Extrudates with a diameter of 1 to 5 mm are preferably used as shaped catalysts. The amount of catalyst employed may be varied within wide limits and is preferably chosen so that a hydrogen peroxide consumption of more than 90%, preferably more than 95%, is achieved within 1 minute to 5 hours under the employed epoxidation reaction conditions. The epoxidation reaction of step a) is preferably carried out at a temperature of 20 to 80°C, more preferably at 25 to 60°C. The epoxidation reaction is preferably carried out at a pressure that is higher than the vapor pressure of propene at the reaction temperature in order to maintain the propene dissolved in the solvent or present as a separate liquid phase. The pressure in step a) is preferably from 1.9 to 5.0 MPa, more preferably 2.1 to 3.6 MPa and most preferably 2.4 to

2.8 MPa. Using an excess of propene at a high pressure provides high reaction rate and hydrogen peroxide conversion and at the same time high selectivity for propene oxide.

The epoxidation reaction is preferably carried out with addition of ammonia to improve epoxide selectivity as described in EP 0 230 949 A2. Ammonia is preferably added with a weight ratio of ammonia to the initial amount of hydrogen peroxide of from 0.0001 to 0.003.

The epoxidation reaction of step a) is preferably carried out in a fixed bed reactor by passing a mixture comprising propene, hydrogen peroxide and methanol solvent over a fixed bed comprising a shaped titanium zeolite catalyst. The fixed bed reactor is preferably a tube bundle reactor and the catalyst fixed bed is arranged inside the reactor tubes. The fixed bed reactor is preferably equipped with cooling means and cooled with a liquid cooling medium. The temperature profile along the length of the catalyst fixed bed is preferably adjusted to keep the reaction temperature along 70 to 98 %, preferably along 80 to 95 %, of the length of the catalyst fixed bed within a range of less than 5 °C, preferably within a range of from 0.5 to 3 °C. The temperature of the cooling medium fed to the cooling means is preferably adjusted to a value 3 to 13 °C lower than the maximum

temperature in the catalyst fixed bed. The epoxidation reaction mixture is preferably passed through the catalyst bed in down flow mode, preferably with a superficial velocity from 1 to 100 m/h, more preferably 5 to 50 m/h, most preferred 5 to 30 m/h. The superficial velocity is defined as the ratio of volume flow rate/cross section of the catalyst bed. Additionally it is preferred to pass the reaction mixture through the catalyst bed with a liquid hourly space velocity (LHSV) from 1 to 20 h \ preferably 1.3 to 15 h ^~ . It is particularly preferred to maintain the catalyst bed in a trickle bed state during the epoxidation reaction. Suitable conditions for maintaining the trickle bed state during the epoxidation reaction are disclosed in WO 02/085873 on page 8 line 23 to page 9 line 15. The epoxidation reaction is most preferably carried out with a catalyst fixed bed maintained in a trickle bed state at a pressure close to the vapor pressure of propene at the reaction temperature, using an excess of propene that provides a reaction mixture comprising two liquid phases, a solvent rich phase and a propene rich liquid phase. Two or more fixed bed reactors may be operated in parallel or in series in order to be able to operate the epoxidation process continuously when regenerating the epoxidation catalyst. Regeneration of the epoxidation catalyst can be carried out by calcination, by treatment with a heated gas, preferably an oxygen containing gas or by a solvent wash, preferably by the periodic regeneration described in WO 2005/000827. Regeneration of the epoxidation catalyst is preferably carried out without removing it from the fixed bed reactor.

Different methods of regeneration may be combined.

In step b) of the process of the invention, a crude propene oxide is separated from the reaction mixture of step a) and a solvent mixture comprising methanol, water and peroxides is separated from the reaction mixture of step a). The separation of the crude propene oxide and the solvent mixture from the reaction mixture can be carried out by methods known from the prior art. The separation of the solvent mixture from the reaction mixture is preferably carried out to provide a solvent mixture which comprises less than 5 % by weight propene and less than 2 % by weight propene oxide.

Preferably, the reaction mixture is subjected to a pressure reduction and propene vapor formed by the pressure reduction is recompressed and cooled to recover propene by condensation. The compressed propene vapor is preferably fed to a propene distillation column and separated into an overhead product comprising non-reacted propene and a bottoms product containing compounds having a boiling point higher than propene, such as propene oxide and methanol solvent. The overhead product comprising non-reacted propene can be recycled to the epoxidation reaction. The bottoms product can be combined with the liquid mixture remaining after the pressure reduction. The liquid mixture remaining after the pressure reduction is preferably separated by distillation in a pre-separation column to provide a crude propene oxide comprising propene oxide, methanol and residual propene as an overhead product and a solvent mixture comprising methanol, water and peroxides as a bottoms product. The pre-separation column is preferably operated to provide an overhead product comprising from 20 to 60 % of the methanol contained in the liquid phase of the last pressure reduction step. The pre-separation column preferably has from 5 to 20 theoretical separation stages in the stripping section and less than 3 theoretical separation stages in a rectifying section and is most preferably operated without reflux and without a rectifying section to minimize the residence time of propene oxide in the pre-separation column. The pre- separation column is preferably operated at a pressure of from 0.16 to 0.3 MPa. Propene oxide and methanol are condensed from the overhead product of the pre-separation column and propene is preferably stripped from the resulting condensate in a propene stripping column which provides a bottom stream comprising propene oxide and methanol which is essentially free of propene.

A purified propene oxide is preferably separated from the bottoms stream of the propene stripping column in an extractive distillation using water as the extraction solvent. The extractive distillation is preferably operated with additional feeding of a reactive compound containing an unsubstituted NH2 group and capable of reacting with acetaldehyde during the extractive distillation, as described in WO 2004/048335. Extractive distillation with a reactive compound provides a high purity propene oxide containing less than 50 ppm of carbonyl compounds.

In step c) of the process of the invention, the solvent mixture separated in step b) is subjected to a catalytic hydrogenation for hydrogenating peroxides contained in the solvent mixture. The reaction conditions of this catalytic hydrogenation are selected to provide a hydrogenated solvent mixture comprising from 1 to 1000 mg/kg of acetaldehyde.

The catalytic hydrogenation is preferably carried out at a hydrogen partial pressure of from 0.5 to 30 MPa, more preferably of from 1 to 25 MPa and most preferably of from 1 to 5 MPa. The temperature is preferably in the range of from 80 to 180 °C, more preferably from 90 to 150 °C. The catalytic hydrogenation is carried out in the presence of a hydrogenation catalyst, preferably a heterogeneous hydrogenation catalyst. Raney nickel and Raney cobalt may be used as hydrogenation catalyst. Preferably, a supported metal catalyst comprising one or more of metals selected from the group consisting of Ru, Rh, Pd, Pt, Ag, Ir, Fe, Cu, Ni and Co on a catalyst support is used. The metal is preferably platinum, palladium, iridium, ruthenium or nickel and most preferably ruthenium or nickel. The catalyst support can be any solid which is inert and does not deteriorate under the hydrogenation conditions. Suitable as catalyst support are activated carbon, the oxides S1O2, T1O2, ZrCh and AI2O3, and mixed oxides comprising at least two of silicon, aluminum, titanium and zirconium. Activated carbon is preferably used as the catalyst support for the supported metal catalyst. The catalyst support is preferably shaped as spheres, pellets, tablets, granules or extrudates. Preferred are extrudates with a diameter of from 0.5 to 5mm, especially from 1 to 3 mm, and a length of from 1 to 10 mm. The supported metal catalyst preferably comprises from 0.01 to 60 wt.% metal. Supported noble metal catalysts preferably comprise from 0.1 to 5 % metal. Supported nickel and cobalt catalysts preferably comprise from 10 to 60 % metal. The supported metal catalyst may be prepared by methods known in the art, preferably by impregnating the catalyst support with a metal salt followed by reducing the metal salt to the catalytically active metal. Suitable supported metal catalyst are commercially available, for example from Clariant under the NISAT ^® trade name and from Evonik Industries under the Noblyst ^® trade name.

The catalytic hydrogenation converts unreacted hydrogen peroxide to water and the by-product peroxides 1-hydroperoxy-2-propanol and 2-hydroperoxy-1-propanol formed in step a) to

1 ,2-propanediol and prevents by-product formation by peroxide decomposition in subsequent workup stages. The catalytic hydrogenation is preferably carried out to a conversion of hydrogen peroxide that provides a hydrogenated solvent mixture containing less than 0.1 % by weight hydrogen peroxide.

The hydrogenation also converts aldehyde and ketone by-products to the corresponding alcohols, with the degree of conversion depending on the catalyst and the reaction conditions used. The conversion of the hydrogenation of acetaldehyde to ethanol can be adjusted by varying the reaction time and the hydrogen partial pressure and the temperature used in the catalytic hydrogenation in order to provide a hydrogenated solvent mixture comprising from 1 to 1000 mg/kg of acetaldehyde. Depending on the reaction conditions, a part of the acetals of aldehydes with methanol and 1 ,2-propanediol may be hydrogenated as well. However, the hydrogenated solvent mixture comprising from 1 to 1000 mg/kg of acetaldehyde will in general comprise significant amounts of 1 , 1-dimethoxyethane and 2,4-dimethyl-1 ,3-dioxolane, which are the acetals of acetaldehyde with methanol and 1 ,2-propanediol.

In step d) of the process of the invention, the hydrogenated solvent mixture of step c) is separated in at least one distillation stage to providing a recovered methanol as an overhead product. An acid is added to the hydrogenated solvent mixture of step c) or to at least one of the distillation stages. When the acid is added to a distillation stage, it is preferably added at a feed point above the feed point for the hydrogenated solvent mixture and below the column top. The acid may also be added to the reflux stream of the distillation column. Most preferably, the hydrogenated solvent mixture is separated in two subsequent distillation stages providing recovered methanol as an overhead product from both stages, feeding the acid to the hydrogenated solvent mixture before it is fed to the first distillation stage. The two distillation stages are preferably operated with a higher pressure in the second stage and overhead product vapor from the second stage is used for heating the bottoms evaporator of the first stage in order to save energy. Adding an acid in step d) reduces the content of volatile organic amines in the recovered methanol and prevents deactivation of the epoxidation catalyst by organic amines when the recovered methanol is recycled to step a).

The acid is preferably added in an amount providing a content of less than 250 ppm by weight nitrogen in the form of organic nitrogen compounds in the recovered methanol, more preferably in an amount providing a content of less than 50 ppm by weight nitrogen in the form of organic nitrogen compounds. The acid may be a mineral acid, such as nitric acid, sulfuric acid, hydrochloric acid, phosphoric acid or perchloric acid; a sulfonic acids, such as methane sulfonic acid; or a carboxylic acid, such as formic acid, acetic acid, propionic acid, butyric acid, pentanoic acid, hexanoic acid, heptanoic acid, octanoic acid, nonanoic acid, decanoic acid, undecanoic acid, dodecanoic acid, oxalic acid, malonic acid, succinic acid, glutaric acid, adipic acid, maleic acid or fumaric acid. Preferred are sulfuric acid and phosphoric acid, most preferred is sulfuric acid. The amount of nitrogen in the form of organic nitrogen compounds can be determined as the difference between the total amount of nitrogen and the amount of nitrogen in the form of inorganic nitrogen compounds. The total amount of nitrogen can be determined by the Kjeldahl method as described in DIN 53625. The recovered methanol will usually contain no inorganic compounds other than ammonia and the amount of nitrogen in the form of inorganic nitrogen compounds may therefore be determined by ion chromatography of an acidified sample detecting ammonium ions.

The acid is preferably added in an amount providing an apparent pH of from 1.7 to 5.0, more preferably from 1.8 to 4.0, in the bottoms product remaining after recovery of methanol. The term apparent pH here refers to the value measured at 20 °C with a pH meter equipped with a pH sensitive glass electrode calibrated with aqueous buffer solutions. Maintaining the apparent pH within these ranges provides a recovered methanol with a low content of alkyl amines and at the same time reduces or prevents acid corrosion of the distillation equipment.

The recovered methanol provided in step d) may contain acetaldehyde at a higher concentration than the hydrogenated solvent mixture of step c) due to acid catalyzed hydrolysis of

1 , 1-dimethoxyethane and 2,4-dimethyl-1 ,3-dioxolane during distillation.

In step e) of the process of the invention, the recovered methanol of step d) is passed through a bed of an acidic ion exchange resin to provide a treated methanol. Both strongly acidic ion exchange resins and weakly acidic ion exchange resins may be used. Preferred are strongly acidic ion exchange resins containing SO3H groups and weakly acidic ion exchange resins containing COOH groups. Most preferred are strongly acidic ion exchange resins containing sulfonic acid groups. The acidic ion exchange resin is preferably based on an organic polymer, such as crosslinked polystyrene, or an organic inorganic hybrid polymer, such as a polysiloxane. The acidic ion exchange resin may be a gel type solid or a macroporous solid. Preferably, two ion exchanger beds are arranged in parallel to allow regeneration of the ion exchange resin without interrupting the methanol treatment. Preferably, the apparent pH of the treated methanol is monitored and the acidic ion exchange resin is replaced or regenerated when the apparent pH of the treated methanol exceeds a threshold value. This threshold value is preferably selected from 2 to 4 pH units higher than the apparent pH of treated methanol obtained with a fresh or a regenerated acidic ion exchange resin.

The treated methanol provided in step e) will in general contain acetaldehyde at a lower concentration than the recovered methanol of step d) and 1 ,1-dimethoxyethane at a higher concentration due to acetalisation of acetaldehyde with methanol catalyzed by the acidic ion exchange resin.

In step f) of the process of the invention, the treated methanol of step e) is recycled to step a). The inventors of the present invention have found that, contrary to what one would expect from the teachings of WO 03/093255 on 1 , 1-dimethoxyethane causing catalyst deactivation, the recycling of the treated methanol provided in step e) to epoxidation step a) does not lead to significant deactivation of the titanium zeolite epoxidation catalyst, whereas recycling of the recovered methanol of step d) without treatment with an acidic ion exchange resin does lead to deactivation of the titanium zeolite epoxidation catalyst.

Steps a) to f) of the process of the invention are preferably carried out continuously, preferably using continuously operated reactors in steps a) and c) and rectification columns in separation steps b) and d).

In a preferred embodiment of the process of the invention, a crude propene oxide comprising from 15 to 97 % by weight propene oxide, from 2 to 84 % by weight methanol, and acetaldehyde is separated in step b) and this crude propene oxide is subjected to an extractive distillation in an extractive distillation column. An aqueous extraction solvent is used and a reactive compound containing an NH2 group and capable of reacting with acetaldehyde at the conditions of the extractive distillation is fed to the extractive distillation column with a feed stream or separately at a feed point above the feed point for the crude propene oxide. The extractive distillation provides a purified propene oxide as an overhead product and a bottoms product comprising water and methanol, and this bottoms product comprising water and methanol is subjected to the catalytic hydrogenation of step c) for hydrogenating reaction products resulting from the reaction of acetaldehyde with the reactive compound containing an NH2 group.

A crude propene oxide comprising from 15 to 97 % by weight propene oxide, from 2 to 84 % by weight methanol, and acetaldehyde can be separated in step b) by the sequence of pressure reduction, distillation in a pre-separation column and propene stripping in a propene stripping column described further above.

The extractive distillation of this crude propene oxide is carried out in an extractive distillation column. The extractive distillation column may be a tray column containing discrete trays such as sieve trays or bubble cap trays. The extractive distillation column may also be a packed column and both random packings as well as structured packings, such as metal gauze packings may be used. The extractive distillation column may also combine sections with discrete trays and sections with packings. The extractive distillation column will in general also comprise at least one overhead condenser and at least one column reboiler. The extractive distillation column preferably has at least two feed points, a feed point A for feeding the crude propene oxide in the middle section of the extractive distillation column and a feed point B for feeding aqueous extraction solvent located above feed point A. The feed points define three sections of the extractive distillation column, a stripping section between the column bottoms and feed point A, an extraction section between feed point A and feed point B and a rectifying section between feed point B and the top of the extractive distillation column. Preferably a distillation column is used that has a separation efficiency of 10 to 30 theoretical stages in the stripping section, a separation efficiency of 15 to 40 theoretical stages in the extraction section and a separation efficiency of 20 to 60 theoretical stages in the rectifying section, i.e. feed point B is preferably located from 15 to 40 theoretical separation stages above feed point A and from 20 to 60 theoretical separation stages below the top of the extractive distillation column.

The aqueous extraction solvent preferably comprises more than 80 % by weight water, more preferably more than 90 % by weight water. Preferably, the aqueous extraction solvent comprises no further solvent in addition to water. The extraction solvent is preferably fed in an amount providing a mass ratio of the extraction solvent relative to the amount of methanol contained in the crude propene oxide feed of from 0.01 to 1 , more preferably from 0.03 to 0.2. The use of such an amount of aqueous extraction solvent provides effective extraction of methanol and a propene oxide product with a low content of methanol and at the same time avoids hydrolysis of propene oxide in the extractive distillation column.

In addition to the aqueous extraction solvent, a reactive compound containing an NH2 group and capable of reacting with acetaldehyde at the conditions of the extractive distillation is fed to the extractive distillation column, either with a feed stream to the extraction column or separately at a feed point above the feed point for the crude propene oxide. The reactive compound is preferably fed to the extractive distillation column admixed with the extraction solvent. The amount of reactive compound fed to the distillation column is preferably chosen so that the molar ratio of the reactive compound relative to acetaldehyde is in the range of from 0.5 to 10. The use of such an amount of a reactive compound provides effective conversion of carbonyl compounds to high boiling compounds and provides a propene oxide product with a low content of acetaldehyde and other carbonyl compounds. At the same time, by-product formation by reactions of the reactive compound with propene oxide can be kept at a low level.

In a preferred embodiment, the reactive compound has a structure R -Y-NH2, where Y is oxygen or NR ² and R and R ² independently of one another are hydrogen, an alkyl group or an aryl group. A preferred compound of structure R -Y-NH2 is hydrazine. Hydrazine hydrate and hydrazinium salts may be used instead of hydrazine. The amount of the reactive compound fed to the distillation column is then preferably chosen so that the molar ratio of the reactive compound relative to acetaldehyde is in the range of from 0.5 to 2.

In another preferred embodiment, the reactive compound is a diaminoalkane having from 2 to 6 carbon atoms, preferably 1 ,2-diaminoethane, 1 ,2-diaminopropane or 1 ,3-diaminopropane and most preferably 1 ,2-diaminoethane. The amount of reactive compound fed to the distillation column is then preferably chosen so that the molar ratio of the reactive compound relative to acetaldehyde is in the range of from 0.5 to 10, more preferably from 3 to 8. Compared to a reactive compound of structure R -Y-NH2, the use of a diaminoalkane as reactive compound reduces the formation of volatile amines when reaction products resulting from the reaction of acetaldehyde with the reactive compound containing an NH2 group are hydrogenated in the subsequent step of hydrogenating the bottoms product of the extractive distillation.

The bottoms product provided by the extractive distillation comprises water, methanol and reaction products formed by reaction of acetaldehyde with the reactive compound containing an NH2 group. This bottoms product is subjected to the catalytic hydrogenation of step c) for hydrogenating the reaction products resulting from the reaction of acetaldehyde with the reactive compound containing an NH2 group. Oximes and hydrazones formed with reactive compound of structure R -Y-NH2 will be hydrogenated with hydrogenolysis of the oxygen-nitrogen bond or the nitrogen- nitrogen bond. Imines formed from acetaldehyde and diaminoalkanes will be hydrogenated to the corresponding amines. The bottoms product provided by the extractive distillation is preferably combined with the solvent mixture separated in step b) before subjecting it to the catalytic hydrogenation of step c).

Examples

Example 1

Acetalisation of acetaldehyde with methanol catalyzed by an acidic ion exchange resin

1500 ppm by weight acetaldehyde and 1490 ppm by weight propionaldehyde were added to methanol and 300 ml of the resulting solution were stirred with 160 g of of DOWEX™ Marathon™ C ion exchange resin for 4 h at about 20 °C. The resulting solution was analyzed by GC and contained 200 ppm acetaldehyde, 230 ppm propionaldehyde, 1930 ppm 1 ,1-dimethoxyethane and 1810 ppm 1 , 1-dimethoxyproane. Example 2

Epoxidation of propene with recovery of solvent mixture, hydrogenation of solvent mixture, methanol, recovery by distillation with acid addition and treatment of recovered methanol with ion exchange resin before recycle to epoxidation.

Propene was epoxidized in a cooled tube bundle reactor with a catalyst fixed bed on an extruded titanium silicalite catalyst arranged in the reactor tubes. A mixture comprising 40 % by weight of propene, 7.7 % by weight hydrogen peroxide, 3.3 % by weight water, 49 % by weight of methanol and 80 ppm ammonia was fed to the top of the reactor and passed through the catalyst fixed bed in trickle mode. The pressure in the reactor was kept at 2.6 MPa by introducing nitrogen. The temperature in the reactor was kept essentially constant at a temperature in the range of from 25 to 60 ° C, adjusting the temperature during the epoxidation reaction to maintain an essentially constant conversion of hydrogen peroxide of 97 %.

The reaction mixture exiting the reactor was depressurized to a pressure of 0.25 MPa and the depressurized liquid was fed to a pre-separation column to provide an overhead product comprising propene oxide, methanol, residual propene and acetaldehyde and a bottoms product comprising methanol, water and non-reacted hydrogen peroxide. Propene oxide and methanol were condensed from the overhead product of the pre-separation column and propene was stripped from the resulting condensate in a propene stripping column to provide a crude propene oxide as bottom stream comprising 23 % by weight propene oxide, 70 % by weight methanol and 380 ppm acetaldehyde. The crude propene oxide was purified by extractive distillation using 55 g of an 0.8 % by weight aqueous solution of hydrazine hydrate per kg crude propene oxide as extraction solvent. A purified propene oxide, containing less than 5 ppm of methanol and acetaldehyde was obtained as overhead product of the column.

The bottom product of the extractive distillation column was combined with the bottom product obtained from the pre-separation column and subjected to continuous hydrogenation with a nickel hydrogenation catalyst in a trickle bed reactor. The hydrogenation was performed at 100 °C and 1.5 MPa at a WHSV of 4 h-1. The resulting hydrogenated solvent mixture comprised on average 78 % by weight methanol, 17 % on weight water and 80 ppm acetaldehyde, as determined by GC analysis.

The hydrogenated solvent mixture was depressurized and fed to stage 14 (counted from top) of a first methanol distillation column having 20 theoretical stages operated continuously at 0.5 MPa. A first recovered methanol stream containing 96 % by weight methanol and 4 % by weight water was obtained as overhead product. The bottoms product was fed to a second methanol distillation column having 30 theoretical stages operated continuously at 1 .0 MPa. A second recovered methanol stream containing 96 % by weight methanol and 4 % by weight water was obtained as overhead product. Concentrated sulfuric acid was added to the hydrogenated solvent mixture before it was fed to the first methanol distillation column at a rate providing an apparent pH of 2.2 in the bottoms product of the second methanol distillation. The overhead products of the first and the second methanol distillation column were combined. The combined overhead products contained on average recovered methanol streams upstream of the ion exchanger had an acetaldehyde content of 10Oppm acetaldehyde.

The combined overhead products were passed through one of two parallel ion exchange columns containing DOWEX™ Marathon™ C ion exchange resin with an average residence time of 5 min. After treatment with the ion exchange resin, the recovered methanol contained on average 15 ppm acetaldehyde. Figure 1 shows the concentrations of acetaldehyde in the hydrogenated solvent mixture (A) and in the methanol recovered by distillation before (B) and after (C) treatment with the ion exchange resin over a time span of 140 h.