WO2003006416A1

WO2003006416A1 - Process for the carbonylation of a conjugated diene and use of such process in the preparation of caprolactam or adipic acid

Info

Publication number: WO2003006416A1
Application number: PCT/NL2002/000461
Authority: WO
Inventors: Eit Drent; Johannes Adrianus Maria Van Broekhoven; Anthonius Johannes Maria Breed
Original assignee: Dsm Ip Assets B.V.
Priority date: 2001-07-13
Filing date: 2002-07-11
Publication date: 2003-01-23

Abstract

Process for the carbonylation of a conjugated diene, comprising reacting the conjugated diene with carbon monoxide and an alkanol in the presence of a metal-based catalyst to form an ester, whereby further a polymeric by-product is formed, wherein the polymeric by-product is separated from the metal-based catalyst with help of a solvent. The process can be used in the preparation of caprolactam or adipic acid.

Description

PROCESS FOR THE CARBONYLATION OF A CONJUGATED DIENE AND USE OF SUCH PROCESS IN THE PREPARATION OF CAPROLACTAM OR

ADIPIC ACID

The present invention relates to a process for the carbonylation of a conjugated diene to an ester product by reaction with carbon monoxide and an alkanol in the presence of a metal-based catalyst. In such a process for the carbonylation of a conjugated diene, mono and di-esters are formed. However, the reaction conditions applied in such a process for the carbonylation of a conjugated diene also cause a certain amount of polymerization. Thus, a main by-product formed in the carbonylation process is the polymer of the conjugated diene used. Depending on the concentration of conjugated diene, the reaction conditions applied and the activity of the catalyst used, this polymeric by-product can further react with carbon monoxide and alkanol to the ester of the formed polymer. The presence of such a polymer of conjugated diene, or the ester product thereof, leads to an increased viscosity of the reaction mixture and is disadvantageous for the process. US-A-5495041 describes a continuous process for the preparation of a pentenoate ester by carbonylation of butadiene. In the described process a part of the effluent of the reactor is led to a drain (purge) in order to prevent a build up of high boiling by-products in a circulating reaction mixture.

A drain to purge reactor effluent containing the formed polymeric by- product, however, is considered to be a very disadvantageous feature. Such a drain will contain not only polymeric by-product but also a valuable amount of catalyst. The costs involved to make-up for the loss of catalyst, and especially the loss of precious metals, are high. There is thus an economic desire to separate used catalyst components in a reactor effluent, and especially the metal, from the formed polymeric by-product in a reactor effluent. Such a separation would enable the economically advantageous re-use of the retrieved catalyst components.

It has now been found that catalyst components in a reactor effluent, and especially the metal, can be separated from the formed polymeric by-product in a reactor effluent in an economically advantageous way with help of a solvent. Accordingly this invention provides a process for the carbonylation of a conjugated diene, comprising reacting the conjugated diene with carbon monoxide and an alkanol in the presence of a metal-based catalyst to form an ester, whereby further a polymeric by-product is formed, wherein the polymeric by-product is separated from the metal-based catalyst with help of a solvent.

Thus, the process provides an economically attractive method to separate formed polymeric by-product, whilst the loss of valuable catalyst components, such as the metal, is reduced considerably or even avoided completely.

The carbonylation reaction can be a mono- or dicarbonylation. If the carbonylation reaction is a monocarbonylation, the carbonylation product is a mono- ester. When the carbonylation reaction is a dicarbonylation the reaction product is a di- ester. The conjugated diene preferably is a conjugated diene having from 4 to 20, more preferably from 4 to 8 carbon atoms per molecule. By "conjugated diene" is understood dienes having at least two double bonds which alternate with single bonds such as, for example 1 ,3-butadiene, 1 ,3-pentadiene, 1 ,3-hexadiene, 1 ,3- cyclohexadiene, 2,4-heptadiene and 2-methyl-1 ,3-butadiene. Aromatically delocalised double bonds are excluded from the scope of the present invention. However, the aliphatic conjugated dienes can have non-aliphatic groups, such as phenyl groups, substituted onto the -C=C-C=C- backbone. Preferably, the conjugated diene is 1 ,3- butadiene.

The carbon monoxide can be used in its pure form or diluted with an inert gas such as nitrogen, carbon dioxide or noble gases such as argon. Small amounts of hydrogen can also be present. In general, the presence of more than 5% hydrogen is undesirable, since this can cause hydrogenation of the conjugated diene. Preferred alkanols in the process of the invention are alkanols with 1 to 20, more preferably with 1 to 6 carbon atoms per molecule and alkanediols with 2- 20, more preferably 2 to 6 carbon atoms per molecule. The alkanols can be aliphatic, cycloaliphatic or aromatic. Suitable alkanols in the process of the invention include methanol, ethanol, ethanediol, n-propanol, 1 ,3-propanediol, iso-propanol, butanol, iso- butanol, tert.butanol, pentanol, hexanol, cyclohexanol, dodecanol, hexadecanol and phenol. Preferably methanol or ethanol is used. The use of these alkanols or alkanediols as a coreactant in the carboxylation process of 1 ,3-butadiene enables the production of alkyl pentenoates and alkyl adipates of which the alkyl group contains 1 to 20, more preferably 1 to 6, carbon atoms.

Methanol is an especially preferred alkanol. The use of methanol in the carbonylation process of 1 ,3-butadiene enables the production of methyl- pentenoate (for example methyl-2-pentenoate, methyl-3-pentenoate and/or methyl-4- pentenoate) and/or dimethyl adipate. Dimethyl adipate is an intermediate compound in the preparation of adipic acid, an intermediate compound to prepare Nylon 6,6. Methyl- pentenoate is an important intermediate in the production process of ε-caprolactam. For the production of ε-caprolactam, methyl-3-pentenoate is preferred. For the preparation of ε-caprolactam, methyl-pentenoate is hydroformylated to methyl formylvalerate, which after reductive amination and cyclisation forms ε-caprolactam. ε- Caprolactam is a starting material in the manufacture of Nylon 6 fibres or engineering plastics. Hydroformylation, reductive amination and cyclisation can be performed in any manner known to a person skilled in the art. Suitable processes are described in WO- A-9733854, WO-A-9835938 and WO-A-9837063.

The molar ratio of conjugated dienes, in particular 1 ,3-butadiene, and alkanol in the liquid phase of the reactor can vary between wide limits and suitably lies in the range of 1 :0.1 to 1 :10, more suitably from 1 :1 to 1 :6. Generally a molar ratio near the lower limit of this range favours the preparation of diesters and a molar ratio near the higher limit of this range favours the preparation of mono-esters.

Preferably the metal in the metal-based catalyst is a group VIII metal, such as for example Pd, Pt, Ni, Rh, Ru, Ir, Co or Fe. Preferred metals are Pd, Pt, Rh and Co. Most preferably the metal-based catalyst is a Pd-based catalyst.

Preferably the metal-based catalyst comprises a source of metal and a ligand, preferably a bidentate diphosphine ligand. More preferably the catalyst comprises in addition a source of anions.

A particularly preferred metal-based catalyst is a catalyst including (i) a source of palladium cations;

(ii) a bidentate diphospine ligand having the general formula I X¹-R-X² (I) wherein X^ and χ2 represent a cyclic group with at least 5 ring atoms, of which one is a phosphorus atom, and R represents a bivalent organic bridging group, connecting both phosphorus atoms; (iii) a source of anions. Suitable sources for palladium cations include its salts, such as for example the salts of palladium and sulphuric acid or sulphonic acids; palladium complexes, e.g. with carbon monoxide or acetylacetonate, or palladium combined with a solid material such as an ion exchanger. Preferably, a salt of palladium and a carboxylic acid is used, suitably a carboxylic acid with up to 12 carbon atoms, such as salts of acetic acid, propionic acid and butanoic acid, or salts of substituted carboxylic acids such as trichloroacetic acid and trifluoroacetic acid. A very suitable source is palladium(ll) acetate.

In the general formula I for component ii) of the catalyst system, R preferably represents a bivalent organic bridging group containing from 1 to 6, preferably containing from 2 to 6, more preferably from 2 to 4, and most preferably 2 to 3 atoms in the bridge. By "a bridge" is understood the shortest connection between both phosphorus atoms. Preferably, the organic bridging group R represents an alkylene group, but it can also comprise a carbon chain, interrupted by one or more hetero atoms, such as nitrogen, sulphur, silicon or oxygen atom. Most preferably the bridging group R contains 2 or 3 carbon atoms in the bridge. The bridge can be substituted or non-substituted or can form part of an aliphatic or aromatic ring structure.

Preferably the bridging group is a substituted alkylene group with at least one substituent and more preferably at least two substituents. Preferably the alkylene group is substituted with two to four substituents and more preferably with two to three substituents. Most preferably the alkylene group is substituted with two substituents.

The substituents can be substituted to any part of the bivalent bridging group R. In an advantageous embodiment, the substituents are substituted at carbon atoms connected to the phosphorus atoms. In case the substituents are substituted at the carbon atoms connected to the phosphorus atoms, the bidentate ligand can have chiral C-atoms and have a R,R, S,S or R,S (meso) form.

The substituents can contain carbon atoms and/or hetero atoms, such as halides, sulphur, phosphor, oxygen and nitrogen. Preferably the substituents are hydrocarbyl groups. The hydrocarbyl groups itself can be aromatic, aliphatic or cycloaliphatic and can contain carbon atoms and hetero atoms. The hydrocarbyl groups include straight-chain or branched saturated or non-saturated carbon containing groups.

Preferred hydrocarbyl groups are alkyl groups, preferably having from 1 to 10 carbon atoms, more preferably from 1 to 4 carbon atoms. Linear, branched or cyclic alkyl groups can be used. Suitable alkyl groups include, methyl, ethyl, propyl, iso- propyl, butyl and iso-butyl. Most suitably methyl groups are used.

Most preferably the bivalent bridging group R is an alkylene group which is di-substituted, preferably with two alkyl groups, most preferably with two methyl groups. χ1 and X² represent a substituted or non-substituted cyclic group with at least 5 ring atoms, of which one is a phosphorus atom, and preferably with from

6 to 10 ring atoms. X^ and X² can represent both the same or each a different cyclic group. Preferably X¹ and X² both represent the same cyclic group. More preferably X¹ and X² represent a substituted or non-substituted cyclic group with 9 ring atoms, of which one is a phosphorus atom. By "a cyclic group" is understood a monocyclic or polycyclic group. Preferred cyclic groups are bicyclic groups. Particularly preferred is a substituted or non-substituted bicyclononyl group wherein one carbon atom is replaced by the phosphorus atom, i.e. a 9-phosphabicyclononyl group. The 9-phosphabicyclononyl group can have several isomeric structures. For the purpose of the invention the [3,3,1] and [4,2,1] isomers are preferred. Most suitably X¹ and X² are substituted or non-substituted [3,3,1] or [4,2,1] 9-phosphabicyclononyl groups. The two

9-phosphabicyclononyl groups can have both the same or each a different isomeric structure.

One or both of the phosphabicyclononyl rings can be substituted with one or more suitable hydrocarbyl groups containing carbon atoms and/or hetero atoms.

If a phosphabicyclononyl ring is substituted it is preferably substituted with one or more alkyl groups, preferably having from 1 to 10 carbon atoms, more preferably from 1 to 4 carbon atoms.

Suitable bidentate ligands of formula (I) thus include isomers of for example 1 ,2-P,P'bis(9-phosphabicyclononyl)ethane;

1 ,3-P,P'bis(9-phosphabicyclononyl)propane; 1 ,3-P,P'bis(2,6- dimethyl,9-phosphabicyclononyl)propane; 2,3-P,P'bis(2,6- dimethyl,9-phosphabicyclononyl)butane; 2,3-P,P'bis(9-phosphabicyclononyl)butane;

2,3-P,P'bis(9-phosphabicyclononyl)pentane; 1 ,2-P,P'bis(9- phosphabicyclononyl)propane; 1,2-P,P'bis(9-phosphabicyclononyl)butane; 3,4-

P,P'bis(9-phosphabicyclononyl)hexane; and mixtures thereof. Isomers of 2,3-P,P'bis

(9-phosphabicyclononyl)butane are preferred.

The ligands can for example suitably be prepared by reacting

P-cyclo-octylene hydride (e.g. phosphabicyclononane hydride) and butyllithium to generate a lithium cyclo-octylene phosphide (lithiated phosphabicylononane). The latter phosphide is reacted with an aliphatic group substituted with suitable leaving groups, preferably tosylates or cyclic sulfates, in an appropriate manner. The P-cyclo-octylene hydride can conveniently be prepared as described by Eisner et al. (Chem. Abstr.

1978, vol. 89, 180154x). The ratio of moles of bidentate ligand per mole atom of palladium preferably ranges from 0.5 to 10, more preferably from 1 to 2 and most preferably the ratio is equimolar.

The source of anions of component (iii) is preferably an acid. A wide range of acids can be used, including mineral acids, such as sulphuric acid, nitric acid and phosphoric acid, and organic acids, such as acetylacetonic acids, sulphonic acids, carboxylic acids and halogenated carboxylic acids such as trifluoroacetic acid. Preferably, a carboxylic acid is used. When a carboxylic acid, is used, preferably an acid with a pK_a value > 1 and more preferably an acid with a pK_a in the range from 1 to

6, in aqueous solution at a temperature of 25 C is used. Exemplary carboxylic acids are benzoic acid, acetic acid, valeric acid, butanoic acid, or nonanoic acid. Also acids corresponding with the ester (by-)products can be advantageously used in the process of the invention. The use of these acids is advantageous because they are readily obtainable by hydrolysis of these ester (by-)products. Examples of these acids are dicarboxylic acids like for example adipic acid, glutaric acid and fumaric acid; monoesters of dicarboxylic acids like for example monoalkyladipate and monoalkylmethylglutarate.

In another preferred embodiment the source of anions is a tertiary carboxylic acid, i.e. an acid with the formula (I)

R¹ - C(R³)- C(O)-OH (I)

I

R² wherein R^ , R² and R³ independently represent alkyl or aryl groups. Suitably the tertiary carboxylic acid used contains a total of from 5 to 20 carbon atoms, more preferably from 5 to 15 and most preferably from 8 to 10 carbon atoms.

Preferably R¹ , R² and R are alkyl groups, preferably having from 1 to 16 carbon atoms, more preferably from 1 to 10 carbon atoms, such as for example methyl, ethyl, n-propyl, iso-propyl, n-butyl, iso-propyl, tert-butyl, n-pentyl, n-decyl, 2- methyl-octyl, n-dodecyl and n-tetradecyl.

Suitable tertiary carboxylic acids include for example 2,2-dimethyl propanoic acid (pivalic acid), 2,2-dimethyl butanoic acid, 2-methyl-2-ethyl-butanoic acid, 2,2-dimethyl-pentanoic acid, 2-methyl-2-ethyl-pentanoic acid, isomers of tertiary C-9 acid (i.e. tertiary acids containing a total of 9 carbon atoms), isomers of tertiary C- 10 acid, and isomers of tertiary C-11 acid. The acids can be gradually esterified during the carbonylation process and can be regenerated, if desired, by hydrolysis.

In another preferred embodiment, the carboxylic acid is the hydrocarboxylation product of the conjugated diene with carbon monoxide and water. Such a carboxylic acid is obtainable by carbonylation of the conjugated diene with carbon monoxide and water or by hydrolysis of the ester product. For example, if the conjugated diene is 1 ,3-butadiene, component (iii) can conveniently be pentenoic acid. Pentenoic acid could be prepared in-situ by carbonylation of 1 ,3-butadiene with carbon monoxide and water or by hydrolysis of a pentenoate product.

The quantity of, for instance in-situ prepared, carboxylic acid used in the carbonylation reaction can vary within wide ranges. Preferably, the amount of acid ranges from 0.1 mole to 1000 mole of acid per mole of palladium cation. In a preferred embodiment the catalyst system further comprises a substoichiometric amount of halide anions, based on the amount of metal cations. The presence of such an amount of halide anions is thought to have a significantly favourable effect in that the conversion reaction proceeds at high rate, even at moderate temperatures. By "substoichiometric" is understood that less halide anions are present than required to neutralise the metal cations, e.g. that the molar ratio of dissociated halide anions versus palladium cations is less than 2:1. Preferably the source of halide anions is a source of chloride, bromide or iodide anions, and more preferably a source of iodide anions is used. In a special preferred embodiment hydrogen iodide is used as a source of anions. The amount of catalyst can vary between wide limits. Preferably, the amount of catalyst system is small. For example, the ratio of mole conjugated diene per mole of metal suitably lies in the range from 1x10^ to 1 x10 mole conjugated diene per mole of metal.

Preferably the reaction takes place in a solvent in which the metal- based catalyst is soluble and which enables separation of the polymeric by-product as described herein. Although liquid carbonylation product and optionally surplus of conjugated diene or alkanol can serve as solvent during the reaction, the reaction is preferably performed in the presence of an additional, inert solvent. Suitable inert solvents are for example aprotic compounds. Examples include ketones, for example methylbutylketone, methylisobutylketone, acetone, methylethylketone; ethers, for example anisole, 2,5,8-trioxanonane (diglyme), diethylether, tetrahydrofuran, diphenylether, diisopropylether and the dimethylether of diethyleneglycol; esters, for example methylacetate, dimethyladipate and butyrolactone; amides, for example dimethylacetamide and N-methylpyrrolidone; and sulphones, for example diisopropylsulphone, sulfolane, 2-methylsulfolane and 2-methyl-4-ethylsulfolane. Preferred solvents are mono- and di-esters. Of these, diesters are more preferred. More preferred such diesters are derived from diacids containing 2 to 12, more preferably 3 to 7 carbon atoms. The diester can be a diaryl ester, a dialkyl ester or an arylalkylester. Preferred diesters include dimethyladipate, diphenyladipate, dibutyladipate, diethyloxalate, diisopropylsuccinate, diphenyl succinate and dimethylsuccinate. When the conjugated diene is 1 ,3-butadiene, more preferred solvents include dimethyladipate (the dimethyl ester of hexanedioic acid), dimethyl α- methyl glutarate (the dimethyl ester of 2-methylpentanedioic acid), dimethyl α-ethyl succinate (the dimethyl ester of 2-ethyl butanedioic acid) and dimethyl α-propyl malonate (the dimethyl ester of 2-propyl propanedioic acid) and mixtures thereof. Of these dimethyladipate is most preferred.

In a preferred embodiment, the solvent is the dicarbonylation product of the conjugated diene with the alkanol. For example, if the conjugated diene is 1 ,3-butadiene and the alkanol is methanol, the solvent can conveniently be dimethyladipate. In this preferred embodiment the solvent can thus be different for each conjugated diene used.

If di-carbonylation is aimed for and the desired product is a di-ester, preferably no solvent other than this di-carbonylation product is used.

If mono-cabonylation is aimed for and the desired product is a mono- ester, preferably an additional solvent, preferably a di-ester as described herein above, is used in addition to the mono-carbonylation product.

For the carbonylation reaction temperatures in the range from 20 to 200 °C, more preferably in the range from 60 to 160°C and carbon monoxide pressures in the range of 0.1-8.5 MPa are preferred. The carbonylation reaction yields a composition containing amongst others metal-based catalyst and a polymeric by-product. More in particular the yielded composition can contain metal-based catalyst, reaction solvent, an ester product and a polymeric by-product.

The amount and type of polymeric by-product can vary widely depending on the conjugated diene, the specific catalyst and the reaction conditions used. It is believed that the polymeric by-product can comprise non-carbonylated polymer and/or carbonylated polymer. Non-carbonylated polymer can be formed by a polymerisation reaction of the conjugated diene. Carbonylated polymer can be formed by a carbonylation reaction of the non-carbonylated polymer with carbon monoxide and alkanol. The polymer is mainly carbonylated at those units where the polymer contains a branch having an unsaturated bond at the end. For example, a polybutadiene is mainly carbonylated at the 1 ,2 vinyl units. Without wishing to be bound to any theory it is believed that a more active catalyst, such as a catalyst system comprising a substoichiometric amount of halide anions as described herein, leads to a higher degree of carbonylation of the polymer. Furthermore it is believed that higher concentrations of conjugated diene result in a higher rate of polymerisation.

The formation of polymeric by-product, especially the polymeric byproduct with an average number molecular weight in the range from 1000 to 50000, more especially in the range from 3000 to 30000, as measured by polystyrene calibrated Gel Permeation Chromatographgy (GPC), gradually increases the viscosity of the yielded composition, which is disadvantageous for the process.

The process according to this invention is especially suitable to separate such polymers, i.e. those polymers with an average number molecular weight in the range from 1000 to 50000, more especially in the range from 3000 to 30000, as measured by polystyrene calibrated Gel Permeation Chromatographgy (GPC), from the process, and in particular from the metal-based catalyst.

The amount and type of ester product depends to a large extent on the reactants and the reaction conditions used. The ester product formed can for example, be a mono- or di-ester as described herein before. The desired ester end- product is preferably removed from the process in an amount about equal to the amount in which it is produced in the process. If the desired ester end-product is a di- ester, such an amount to be removed preferably constitutes only part of the di-ester present. Since di-esters are a preferred reaction solvent, preferably a certain amount of di-ester is added upfront as a solvent and present during the whole process. Thus for practical reasons the amount of diester removed preferably lies in the range of 5 to 95 %v/v, more preferably in the range of 20-70%, on the total amount of di-ester present.

If the desired ester end-product is a mono-ester, the amount of desired ester end-product removed is preferably about the same as the amount in which this product is present. If the mono-ester is not used as a reaction solvent, the mono-ester is for practical reasons preferably removed for at least 30 %v/v, more preferably for at least 50-100 %v/v on the total amount of mono-ester present.

The ester product can be separated from the yielded composition before or after separation of the polymeric by-product. Preferably the ester product is separated from the yielded composition before the polymeric by-product is separated. The ester product can be separated by any method known to one skilled in the art to be suitable therefore. Preferably the ester product is removed from the composition by distillation, more preferably by wiped film evaporation. Together with the ester product other parts of the composition, for example unconverted reactants, such as the conjugated diene or alkanol, and/or light boiling by-products, can be distilled off. In such a case the ester product is preferably further purified in one or more additional, preferably distillation, steps.

The polymeric by-product in the yielded composition, before or after removal of the ester product, is separated from the catalyst components with help of a solvent.

Preferably the yielded composition is first cooled, preferably to a temperature below 100°C, more preferably to a temperature in the range from 10°C to 50°C, before separation of the polymeric by-product. Suitably the solvent used differs in the solubility of the metal-based catalyst and the polymeric by-product. Preferably at least one solvent is used in which the metal-based catalyst essentially dissolves and in which the polymeric by-product does not or essentially not dissolve. The solvent can be used simultaneously as a reaction solvent or can be added after reaction has taken place. For the separation the help of at least one solvent is needed. The help of more than one solvent can be advantageous, depending on the specific polymeric by-product and/or metal-based catalyst. Preferably at least two solvents are used. Preferably the polymeric by-product is separated from the metal-based catalyst by means of phase separation and/or extraction. The later measures both make use of differences in solubility.

If the polymeric by-product is a non-carbonylated polymer or a polymer only carbonylated to a small extent separation can be accomplished with the help of just one solvent. In this case the polymeric by-product is preferably separated from the metal-based catalyst by using a solvent, in which the polymeric by-product is not or only sparingly soluble, as a reaction solvent. Preferred solvents for this purpose are the diesters described herein as possible reaction solvents.

If such a solvent is used, formed polymeric by-product will not or only sparingly dissolve in the reaction solvent. At an appropriate temperature, optionally after cooling, a separate polymeric phase can spontaneously form. Catalyst components, and especially the source of metal, will remain dissolved in a reaction solvent phase. The formed polymeric phase can conveniently be separated from the reaction solvent phase by phase separation.

Carbonylated polymer, however, has been observed to dissolve to a considerable extent in many of the possible reaction solvents. It was therefore found difficult to separate a carbonylated polymeric by-product from the metal-based catalyst with the help of just one solvent. If the polymeric by-product is mainly carbonylated polymer, the polymeric by-product is preferably separated from the metal-based catalyst with the help of two or more solvents. Preferably the relative solubility of one or more solvents for the polymeric by-product and/or the metal-based catalyst differs. Preferably one or more of the solvents is/are simultaneously used as a reaction solvent. After the reaction this reaction solvent can contain dissolved catalyst and dissolved polymeric by-product. Separation can be accomplished by extracting either polymeric by-product with a suitable extraction solvent or extracting catalyst components with a suitable extraction solvent. Such an extraction of either polymeric by-product or catalyst components can also be applied where non-carbonylated polymer or polymer only carbonylated to a small extent is formed.

Preferably separation is accomplished by extracting polymeric byproduct from the reaction solvent with a suitable extraction solvent.

In a preferred embodiment of the process of this invention separation comprises:

I) intimately contacting a composition containing a reaction solvent, a metal-based catalyst and a polymeric by-product with at least one extraction solvent, yielding a mixture containing metal-based catalyst, polymeric by-product, reaction solvent and extraction solvent; II) allowing the mixture of I) to settle into at least a phase containing polymeric byproduct and extraction solvent and a phase containing metal-based catalyst and reaction solvent;

III) separating the phase containing polymeric by-product from the phase containing metal-based catalyst. In a preferred embodiment the phase containing metal-based catalyst is re-used in a carbonylation reaction.

Preferably at least a hydrocarbon solvent is present as an extraction solvent for extracting polymeric by-product. More preferably two or more extraction solvents are used, of which one extraction solvent is a hydrocarbon solvent. In a preferred embodiment the polymeric by-product is separated from the metal-based catalyst with the help of a composition of extraction solvents containing at least a hydrocarbon solvent and an alkanol.

The hydrocarbon extraction solvent can comprise aromatics, alicyclics and aliphatics such as alkanes, alkenes, alkadienes, alkynes. Of these solvents, aliphatics, especially linear or branched alkanes or mixtures thereof, are preferred. Preferably the hydrocarbon solvent comprises saturated alkanes having at least 3 carbon atoms, more preferably at least 6 carbon atoms and most preferably at least 8 carbon atoms. The hydrocarbon solvent can advantageously be a paraffin or mixture of paraffins, such as for example SHELLSOL D40, D60, D70 and D100 (SHELLSOL is a trademark). There is no upper limit for the molecular weight of the alkanes, however, for practical reasons, alkanes and mixtures of alkanes that are liquid at ambient temperature, i.e. 25 °C, are preferred. Hydrocarbons, which can be used include for example toluene, xylene, cumene, butadiene, pentane, hexane, cyclohexane, hexene, heptane, octane, octene, nonane, decane, undecane, dodecane and tetradecane. Preferred hydrocarbons include n-decane and higher homologues, such as for example n-undecane, n-dodecane and n-tetradecane. Preferably the hydrocarbon solvent has a boiling point well above or well below the boiling point of the ester product, because small amounts of hydrocarbon solvent can, to a certain extent, dissolve in the phase containing metal-based catalyst. If the boiling point of the hydrocarbon solvent lies in the same range as that of the ester product, the hydrocarbon solvent might interfere with ester product recovery.

The alkanol extraction solvent is preferably an alkanol, having from 1 to 20, more preferably from 1 to 6 carbon atoms per molecule or an alkanediol, having from 2-20, more preferably 2 to 6 carbon atoms per molecule, or mixtures of alkanols and/or alkanediols. Most preferred alkanol extraction solvents are methanol, ethanol, n- propanol or iso-propanol. In a preferred embodiment the alkanol used in the carbonylation reaction is used as the alkanol extraction solvent.

In addition the composition of extraction solvents can contain a further, third, solvent such as the inert solvent used in the reaction, for example a diester. The composition of extraction solvents can consist of one or more phases, depending on the hydrocarbon solvent and alkanol used. For example, the composition of extraction solvents can comprise a single-phase mixture of alkanol and hydrocarbon or two phases, an alkanol-rich phase and a hydrocarbon-rich phase.

The composition of extraction solvents is brought into contact with the composition yielded by the reaction, that is a composition containing a reaction solvent, a metal-based catalyst and a polymeric by-product. Temperature and pressure during this contacting step can vary between wide ranges and are dependent on the hydrocarbon solvent and alkanol used. For example, if hydrocarbon solvents having less than 6 carbon atoms are used, temperature and pressure are adjusted such that the hydrocarbon solvent is liquid under the conditions applied. The extraction solvents can be added one by one or can be added simultaneously. For example, the composition containing a reaction solvent, a metal- based catalyst and a polymeric by-product can be brought first into contact with an alkanol solvent, which can result in a certain degree of precipitation of the polymeric by-product. Subsequently the precipitated polymeric by-product can be re-dissolved by means of a hydrocarbon solvent. This will result in a hydrocarbon-hch/polymer phase and a metal-containing/alkanol-rich phase, which can be separated.

Preferably, however, the composition containing a reaction solvent, a metal-based catalyst and a polymeric by-product is simultaneously brought into contact with all extraction solvents. Extraction can be carried out co-currently or counter-currently, but is preferably carried out counter-currently. That is, the composition containing a reaction solvent, a metal-based catalyst and a polymeric by-product is brought into contact with one or more extraction solvents, which flow in a direction counter currently to the yielded composition. Extraction/phase separation systems that can be used include for example mixer-settler combinations, packed columns and/or rotating disks contactors. In a preferred embodiment the yielded composition is contacted with at least one extraction solvent in a mixer, where after the resultant mixture is allowed to settle into at least two phases in a settler. Preferably a sequence of mixer-settler combinations is used, wherein the yielded composition flows in a direction counter- currently to the direction of the extraction solvent or solvents.

In another preferred embodiment a rotating disk contactor is used wherein the yielded composition is added at one side of the contactor and flows counter-currently to at least one extraction solvent, which is added at the other side of the contactor. In such a rotating disk contactor contacting and settling occurs more or less simultaneously and separated phases can be obtained at the top and bottom of the rotating disk contactor.

Also a combination of extraction/phase separation systems can be used, for example a combination of mixer-settlers and rotating disk contactors. The amount of extraction solvent or solvents to be used depends on the amount of polymeric by-product formed during the reaction and can be optimised by one skilled in the art. The volumetric ratio of solvent or solvents to yielded composition can thus vary widely. For practical purposes a volumetric ratio of extraction solvent or solvents to the composition to be extracted in the range from 0.01 :100 to 100:0.01 , more preferably in the range from 0.1 :10 to 10:0.1 is preferred.

The ratio of hydrocarbon solvent to alkanol to be used in a composition of extraction solvents comprising a hydrocarbon solvent and an alkanol can also vary widely. Again the optimum ratio depends on the amount of polymeric byproduct formed during the reaction. For practical purposes a volumetric ratio of hydrocarbon solvent to alkanol in the range from 0.01 : 100 to 100:0.01 , more preferably in the range from 0.1 : 10 to 10:0.1 is preferred.

Both phases generated in a phase separation can be extracted further. Preferably a phase comprising the polymeric by-product is extracted once or more with alkanol or a mixture of alkanol and reaction solvent to retrieve residual metal. The phase comprising the metal-based catalyst, is preferably extracted once or more with hydrocarbon solvent to remove residual polymeric by-product from this phase.

The composition yielded by the reaction from which both ester- product and polymeric by-product are removed as described above can be recycled to the reaction zone.

The process according to the invention can be a batch, continuous or semi-continuous process. In a preferred embodiment the process is a continuous process. This invention therefore also provides a continuous process for the carbonylation of a conjugated diene, comprising: a) reacting the conjugated diene with carbon monoxide and an alkanol in the presence of a metal-based catalyst system, yielding a composition containing metal-based catalyst, an ester product and a polymeric by-product; b) separating ester product from the composition obtained in step a), yielding an ester product-depleted composition; c) separating polymeric by-product from the composition obtained in step b) with help of a solvent; yielding an ester product-depleted and polymer-depleted composition; d) recycling the ester product-depleted and polymer-depleted composition obtained in step c) to reaction step a).

Preferences are as previously described hereinabove for the process in general. In the preferred process as outlined above, ester product is separated in step b) from the composition obtained in step a) before separating the polymeric by- product in step c). However, it will be understood by one skilled in the art that this invention also provides a process wherein the order of steps b) and c) is reversed, i.e. wherein first polymeric by-product is separated yielding a polymer-depleted composition whereafter ester-product is separated yielding an ester product-depleted and polymer-depleted composition. Such a process comprising a) followed by c) followed by b) followed by d) is also part of this invention. ln a continuous process separation of the polymeric by-product can occur batch-wise or continuous. By a batch-wise phase separation is understood that step c) is carried out intermittently whenever the concentration of polymer by-product exceeds a certain threshold value. By a continuous phase separation is understood that step c) is carried out continuously, regardless the concentration of polymer byproduct. The separation can be carried out on the whole of the composition obtained in the previous step or on part of that composition. Preferably step c) is carried out on part, preferably on a bleed or purge stream, of the composition obtained in the previous step. Polymeric by-product is then removed from such a part or purge stream, whereafter the obtained polymer-depleted stream can, for example, be recycled to the reaction in step a).

The process as described herein can advantageously be used in a process for the preparation of caprolactam or adipic acid.

An illustration of a continuous process according to the invention is now described by reference to figure 1.

A continuous stirred tank reactor (102) is charged (for start-up) with a metal-based catalyst (Pd acetate, bidentate diphosphine ligand and a source of anions) and a diester (e.g. dimethyladipate) as a reaction solvent.

A stream of appropriate amounts of carbon monoxide, 1,3-butadiene and methanol (101) is continuously introduced to the continuous stirred tank reactor (102) at reactor conditions suitable for carbonylation, for example a pressure of 50 bar and a temperature of 135 °C. A stream comprising unreacted 1 ,3-butadiene, methanol and carbon monoxide; metal-based catalyst; diesters (e.g dimethyladipate and possibly other diesters formed during the reaction); methyl pentenoate; polymeric by-product and some other compounds (e.g. methoxy butenes) (103) is withdrawn from the reactor (102) and led into a gas-liquid separator (104). A stream of gas (105), containing mainly carbon monoxide and 1,3-butadiene is withdrawn at the top of the gas-liquid separator (104), and recycled into the reactor (102). A stream of liquid (106) containing dissolved metal-based catalyst; possibly some 1 ,3-butadiene; methanol; diesters; methyl pentenoate; polymeric by-product and some other compounds (e.g. methoxy butenes) is withdrawn at the bottom of the gas-liquid separator (104), and led into a flash vessel (107), preferably a wiped film evaporator. A stream of light components (108), such as unconverted 1 ,3-butadiene, methanol, methyl pentenoate, methoxybutenes and some di-ester is withdrawn at the top of the flash vessel (107). A stream containing metal-based catalyst; diesters; and polymeric by-product (109) is withdrawn at the bottom of the flash vessel (107). From this stream (109) a purge stream (110) is withdrawn, and a main stream (111 ) is recycled to the reactor (102). The purge stream (110) is led into a first mixer (112), where it is mixed with a stream of fresh hydrocarbon (113) and a stream of used methanol (114) obtained from a second settler (122). From the first mixer (112) a mixture of metal-based catalyst; diesters; and polymeric by-product; hydrocarbon solvent and methanol (115) is withdrawn and led into a first settler (116). In the first settler (116) this mixture is phase separated into a metal-based catalyst-depleted/polymer-rich hydrocarbon phase (116a) and a polymer- depleted/metal-based catalyst rich methanol phase (116b). A stream of polymer- depleted/metal-based catalyst-rich methanol phase (117) is recycled to the reactor (102). A stream of metal-based catalyst-depleted hydrocarbon phase (118) is withdrawn from the first settler (116) and led to a second mixer (119) where it is mixed with a stream of fresh methanol (120) to extract any residual metal-based catalyst. From the second mixer (119) the obtained mixture (121) is led to a second settler (122), where it is phase separated into a metal-based catalyst-depleted/polymer-rich hydrocarbon phase (122a) and a polymer-depleted/metal-based catalyst-rich methanol phase (122b). A stream (114) of polymer-depleted/metal-based catalyst-rich methanol is recycled to the first mixer (112). A stream of metal-based catalyst-depleted/polymer- rich hydrocarbon (123) is discarded.

The invention is further illustrated by the following non-limiting examples.

For comparative experiment A and examples 1 to 5, a bench-scale unit as illustrated in figure 2 was used. In this unit a continuous stirred tank reactor (201) is charged with a metal-based catalyst and a reaction solvent. A feed of appropriate amounts of carbon monoxide, 1 ,3-butadiene and methanol (203) is added to the reactor (201) at reactor conditions suitable for carbonylation. The mixture in the reactor is stirred. A product stream (205) is withdrawn from the reactor (201) and led into a stripper (207). Light products and unreacted reactants (209) are stripped off by a stream of nitrogen gas (211) at the top of the stripper (207). A stream containing dissolved metal-based catalyst; reaction solvent; and polymeric by-product (213) is withdrawn at the bottom of the stripper (207), and led into a buffer vessel (215). A stream containing in each case metal-based catalyst is hereafter recycled from the buffer vessel (215) to the reactor (201).

Comparative experiment A A bench-scale unit as illustrated in figure 2 was operated during 150 hours for the carbonylation of 1 ,3-butadiene to methylpentenoate using a homogeneous Pd-based catalyst.

The unit was charged with 425 ml of a catalyst solution, 200 ml in the reactor (201) and 225 ml in the buffer vessel (215). The catalyst solution was prepared by stirring a mixture of 3 mmole Pd acetate, 4.5 mmole 1 ,2-bis(9- phosphabicyclononyl)ethane as a ligand, 30 mmole pivalic acid and 400 ml anisole under nitrogen. The reaction was started by pressurising the reactor (201) with carbon monoxide, starting the methanol, butadiene, anisole and recycle pumps and heating the reactor to 135°C, while keeping the pressure at 50 bar. The feed rate of butadiene was varied between 6 and 16 grams butadiene/hour. The feed rate of methanol was varied between 4 to 10 grams methanol/hour. The anisole feed rate of 10 g/hour was sufficient to compensate distillation losses. The recycle stream from the buffer vessel (215) to the reactor (201) was pumped at 60 ml/hour. The liquid level in the reactor (201) during operation was kept at 250 ml. Catalyst stability was maintained by continuously feeding a small flow of 0.2 g/hour of pivalic acid and 3.6 mg/hour of 1 ,2- bis(9-phosphabicyclononyl)ethane (as a salt of pivalic acid dissolved in anisole). Light products, anisole and reactants were removed in a stripper (207) at about 100 °C using pure nitrogen as stripping gas. The nitrogen flow was adjusted for maintaining a constant liquid level in the buffer vessel (215). The stripper (207) bottoms containing the catalyst, anisole, some methylpentenoate, some other high boiling esters, and dissolved polybutadiene were continuously recycled to the reactor. After 150 hours the methylpentenoate product yield was 750 g. A sample from the stream (217) recycled from the buffer vessel (215) to the reactor (201) was analysed and had the following composition: 71 %w anisole 21 %w methylpentenoate

8 %w polybutadiene (by I^C NMR) with an average number molecular weight of 20.000 (by styrene calibrated GPC) 320 ppmw Pd (by ICP AES).

Polybutadiene could only be removed from the process by discarding part of the homogeneous liquid phase also comprising 320 ppmw Pd.

Example 1

Comparative experiment A was repeated with the following differences: anisole was replaced by dimethyladipate (DMA) as a reaction solvent. The feed rate of butadiene was varied between 8 and 15 grams butadiene/hour. The feed rate of methanol was varied between 4 to 6 grams methanol/hour. The recycle stream from the buffer vessel (215) to the reactor (201) was pumped around at 60 ml/hour. The methylpentenoate product yield after 280 hours was 1100 g. A small flow of acid was fed to the reactor (201) to make up for acid losses. During the test the buffer vessel (215) contained two liquid phases, an upper polybutadiene-hch phase which increased in volume from zero to 310 ml over a period of 280 hours, and a lower dimethyladipate phase with little dissolved polybutadiene. The dimethyladipate phase was recycled to the reaction zone. Both liquid phases were analysed for polybutadiene by ¹³C NMR and for Pd by ICP AES after 280 hours.

Upper phase Lower phase

310 ml 250 ml

50 %w polybutadiene 4 %w polybutadiene

45 %w dimethyladipate 85 %w dimethyladipate

5 %w methylpentenoate 1 10 %w methylpentenoate

75 ppmw Pd 345 ppmw Pd The polybutadiene structure determined by ^³C NMR consisted of

80% 1 ,4 (linear) units and 20% 1 ,2 vinyl units and was not carbonylated. The formed polybutadiene had an average number molecular weight of 20000 (by styrene calibrated GPC).

The example shows that by using a di-ester such as dimethyladipate as a reaction solvent, spontaneous phase separation occurs, generating a polybutadiene-hch upper phase and a polybutadiene-depleted lower phase. The lower phase can be recycled to the reaction zone. Polybutadiene can be removed from the process by discarding part of the upper phase, containing only 75 ppmw Pd. Thus loss of valuable catalyst components is considerably decreased compared with comparable experiment A. Example 2

Example 1 was repeated with the following difference: The catalyst consisted of 6 mmole Pd acetate, 13 mmole meso-1 ,2 dimethyl-1 ,2-bis (9- phosphabicyclononyl) ethane, 80 mmole pivalic acid (PVA) and 0.6 mmole hydrogen iodide. Pivalic acid was continuously dosed at a rate of 0.2 g/hour and additional ligand at 3.6 mg/hour (as pivalic acid salt in dimethyladipate). The feed rate of butadiene was varied between 8 and 12 grams butadiene/hour. The feed rate of methanol was varied between 6 and 12 grams methanol/hour. The recycle stream from the buffer vessel (215) to the reactor (201) was varied between 30 and 60 ml/hour. After 140 hours the methylpentenoate product yield was 840 g. The buffer vessel (215) contained a homogeneous liquid phase containing 7 %w dissolved (carbonylated) polybutadiene and 715 ppmw Pd. A quantity of 250 ml of this liquid was removed from the buffer vessel (215) and was mixed with a 2-phase mixture of 250 ml methanol and 250 ml tetradecane at ambient temperature under nitrogen blanketing. After settling two liquid phases were collected with the following composition: Upper phase Lower phase

270 ml 500 ml

82 %w tetradecane 8 %w tetradecane

4 %w methanol 48 %w methanol

6 %w dimethyladipate 40 %w dimethyladipate 1 %w methylpentenoate 3 %w methylpentenoate

7 %w polybutadiene (15 g) <1 %w polybutadiene (<5 g) 9 ppmw Pd (2 mg Pd) 405 ppmw Pd (178 mg Pd)

The lower phase, containing 99 %w of the Pd catalyst, was introduced again in the buffer vessel (215) and gave unaltered carbonylation activity. The polybutadiene-rich upper phase, containing more than 75 %w of the polybutadiene and 1 %w of the Pd catalyst was extracted a second time with methanol/dimethyladipate. This resulted in a polybutadiene-rich liquid phase containing less than 1 ppmw Pd (less than 0.2 mg Pd) and a polybutadiene depleted methanol/dimethyladipate phase containing the remaining Pd. Thus 99.9 %w of the Pd catalyst inventory of the purge stream is recovered. The polybutadiene-rich phase was analysed by ^³C NMR and consisted of partly carbonylated polybutadiene, having an average number molecular weight of 5000 (determined by polystyrene calibrated GPC), with the following microstructure: 80% 1 ,4 (linear) units 10% 1 ,2 vinyl units 10% carbonylated 1 ,2 structure.

The higher degree of polybutadiene carbonylation was responsible for the enhanced polybutadiene solubility in dimethyladipate. Carbonylation of polybutadiene vinyl structure is believed to be caused by the use of a more active metal-based catalyst and by operating the reactor (201) at a higher butadiene conversion. Example 3

Example 2 was repeated with the difference that the catalyst inventory was only 50% and that no extra ligand was dosed during the run. The recycle flow varied between 60 and 120 ml per hour. After 330 hours the methylpentenoate product yield was 2600 g. The buffer vessel (215) contained a homogeneous liquid with 14 %w dissolved polybutadiene and 495 ppmw Pd. The polybutadiene vinyl 1 ,2 structure was largely carbonylated. 400 ml liquid of the buffer vessel (215) was contacted with a 2-phase mixture of 230 ml methanol and 230 ml tetradecane. Phase separation gave the following result:

Upper phase Lower phase

370 ml 500 ml

59 %w tetradecane 13 %w tetradecane

13 %w methanol 36 %w methanol

14 %w dimethyladipate 34 %w dimethyladipate 6 %w methylpentenoate 8 %w methylpentenoate

9 %w polybutadiene (27 g) 9 %w polybutadiene (40 g)

8 ppmw Pd (2 mg Pd) 444 ppmw Pd (198 mg Pd)

The polybutadiene-rich upper phase contains 40 %w of the polybutadiene and 1 %w of the Pd. Purging the upper phase, to discard 40 %w of the polybutadiene thus results in a Pd loss of only 1 %w. The polybutadiene structure was analysed by ^C NMR and contained:

80% 1 ,4 (linear) units

5% 1,2 vinyl units

15% carbonylated 1 ,2 structure. The polybutadiene had an average number molecular weight of 5000

(determined by polystyrene calibrated GPC).

Example 4

Example 2 was repeated with the difference that the polybutadiene was extracted from the catalyst recycle with a single-phase mixture of toluene/methanol after 360 hours of operation. At that moment methylpentenoate yield was 4600 g. 196 g of liquid from the buffer vessel (215), containing 17 %w dissolved polybutadiene and 636 ppmw Pd was contacted with a single phase mixture of 170 g methanol and 100 g toluene. Phase separation gave the following result: Upper phase Lower phase

450 ml 35 ml

25 %w toluene 29 %w toluene

44 %w methanol 9 %w methanol

3 %w methylpentenoate 1 %w methylpentenoate

28 %w dimethyladipate 12 %w dimethyladipate

1.3 %w polybutadiene (5 g) 50 %w polybutadiene (17 g)

285 ppmw Pd (105 mg Pd) 45 ppmw Pd (1.6 mg Pd)

The polybutadiene structure was analysed by 1³c NMR and contained:

80% 1 ,4 (linear) units 10% 1,2 vinyl units 10% carbonylated 1 ,2 structure.

Contrary to previous examples the hydrocarbon/polybutadiene liquid was now the lower phase. This is caused by the very high polybutadiene content of this phase (density of the polybutadiene is about 1 kg/litre). The polybutadiene-rich lower phase contains 77 %w of the polybutadiene and 1.5 %w of the Pd. Purging the upper phase, to discard 77 %w of the polybutadiene thus results in a Pd loss of only 1.5 %w. Example 5

Example 4 was repeated with the difference that after 360 hours of operation the product stream from the reactor (201) was extracted before stripping of the methylpentenoate product in the stripper (207). The polybutadiene was extracted with pentane/methanol by mixing 190 g reactor product containing 10 %w dissolved polybutadiene and 440 ppmw Pd with 100 g methanol and 100 g n-pentane. Phase separation gave the following result: Upper phase Lower phase

370 ml 35 ml

25 %w pentane 37 %w pentane

42 %w methanol 14 %w methanol

9 %w methylpentenoate 5 %w methylpentenoate

23 %w dimethyladipate 10 %w dimethyladipate

2 %w polybutadiene (6 g) 34 %w polybutadiene (8 g)

288 ppmw Pd (85 mg Pd) 41 ppmw Pd (1.4 mg Pd)

The polybutadiene structure was analysed by ^ ³C NMR and contained: 80% 1 ,4 (linear) units

10% 1,2 vinyl units

10% carbonylated 1 ,2 structure.

Pd loss by a purge of the lower phase was 1.6 %w, whilst the efficiency of polybutadiene removal by this purge was about 57 %w. Also in this example the polybutadiene/hydrocarbon liquid was the higher density phase, because of the high polybutadiene content. Example 6

Example 2 was repeated with the following differences (see figure 3): A feed (202) of appropriate amounts of carbon monoxide and 1 ,3-butadiene was added to the continuous stirred tank reactor (201). The buffer vessel (215) was filled with 1 litre of 5 mm glass beads (216). The buffer vessel (215) contained about 200 ml of tetradecane phase (215a) and 200 ml dimethyladipate/methanol (215b) phase. The dimethyladipate containing stream (213) exiting the bottom of the stripping vessel (207) with a flow rate of about 120 ml per hour was admixed with a flow (214) of 20 ml methanol per hour prior to entering the buffer vessel (215) to maintain a two liquid phase system in this vessel (215). The Pd and polymer containing dimethyl adipate- methanol mixture was continuously percolated through the tetradecane phase (215a) for 260 hours. A small flow (218) of 3 ml dimethyladipate/tetradecane mixture (volume ratio 2:1) was continuously fed to the bench-scale unit to make up for distillation losses. After 260 hours operation the composition of both phases was analysed by GLC, ICP-

AES (for Pd) and ¹³C NMR (for polybutadiene) with the following result:

Upper phase Lower phase

86 %w tetradecane 15 %w tetradecane

1.5 %w methanol 24 %w methanol

5.4 %w dimethyladipate 53 %w dimethyladipate

1.6 %w methylpentenoate 9 %w methylpentenoate 4 %w polybutadiene 0.7 %w polybutadiene 17 ppmw Pd 1025 ppmw Pd

The polybutadiene structure was analysed by 1³C NMR and contained: 80% 1 ,4 (linear) units 15% 1 ,2 vinyl units 5% carbonylated 1 ,2 structure. The results show that the polybutadiene by-product can be very effectively removed when it is extracted from the catalyst recycle liquid before its vinyl structure is carbonylated to a large extent by subsequent reactor passages. 85% of the polybutadiene is present in the upper (tetradecane) phase (215a) in the buffer vessel (215), whereas more than 98% of the catalyst is retained by the lower (dimethyladipate/methanol) phase (215b) that is being recycled to the reactor.

Claims

1. Process for the carbonylation of a conjugated diene, comprising reacting the conjugated diene with carbon monoxide and an alkanol in the presence of a metal-based catalyst to form an ester, whereby further a polymeric by-product is formed, wherein the polymeric by-product is separated from the metal- based catalyst with help of a solvent.

2. A process as claimed in claim 1 , wherein the conjugated diene is 1 ,3- butadiene.

3. A process as claimed in claim 1 or 2, wherein the alkanol is an alkanol with 1 to 6 carbon atoms per molecule or an alkanediol with 2 to 6 carbon atoms per molecule.

4. Process as claimed in any one of claims 1-3, wherein the metal-based catalyst is a catalyst including

(i) a source of palladium cations;

(ii) a bidentate diphospine ligand having the general formula I X¹-R-X² (I) wherein X^ and X² represent a cyclic group with at least 5 ring atoms, of which one is a phosphorus atom, and R represents a bivalent organic bridging group, connecting both phosphorus atoms; (iii) a source of anions.

5. Process as claimed in any one of claims 1-4, wherein the polymeric byproduct is separated from the metal-based catalyst with the help of two or more solvents.

6. Process as claimed in any one of claims 1-5, wherein the separation comprises: I) intimately contacting a composition containing a reaction solvent, a metal- based catalyst and a polymeric by-product with at least one extraction solvent, yielding a mixture containing metal-based catalyst, polymeric by-product, reaction solvent and extraction solvent;

II) allowing the mixture of I) to settle into at least a phase containing polymeric by-product and extraction solvent and a phase containing metal-based catalyst and reaction solvent;

III) separating the phase containing polymeric by-product from the phase containing metal-based catalyst.

7. Process as claimed in anyone of claims 1-6, wherein the polymeric by-product is separated from the metal-based catalyst with the help of a composition of extraction solvents containing at least a hydrocarbon solvent and an alkanol.

8. Process as claimed in claim 7, wherein the hydrocarbon solvent comprises alkanes having at least 3 carbon atoms.

9. Continuous process for the carbonylation of a conjugated diene, comprising a) reacting the conjugated diene with carbon monoxide and an alkanol in the presence of a metal-based catalyst system and a solvent, yielding a composition containing metal-based catalyst, solvent, an ester product and a polymeric by-product; b) separating ester product from the composition obtained in step a), yielding an ester product-depleted composition; c) separating polymeric by-product from the composition obtained in step b) with help of a solvent, yielding an ester product-depleted and polymer- depleted composition; d) recycling the ester product-depleted and polymer-depleted composition obtained in step c) to reaction step a).

10. Use of a process claimed in any one of claims 1-9, in a process for the preparation of caprolactam or adipic acid.