WO2020016118A1

WO2020016118A1 - Method for producing oligomeric polyisocyanates with subsequent fractional liquid-liquid extraction

Info

Publication number: WO2020016118A1
Application number: PCT/EP2019/068782
Authority: WO
Inventors: Thomas Ernst MÜLLER; Christoph Gürtler; Reinhard Halpaap; Frank Richter; Martina Peters; Petro SA GOMES; Burkhard KÖHLER; Walter Leitner; Kai LEONHARD
Original assignee: Covestro Deutschland Ag
Priority date: 2018-07-16
Filing date: 2019-07-12
Publication date: 2020-01-23

Abstract

The present invention relates to a method for producing oligomeric polyisocyanates with subsequent fractional liquid-liquid extraction, wherein a solvent pair which forms two phases in the extraction and comprises a first solvent system and a second solvent system is used in the extraction, and the polyisocyanate accumulates in the first solvent system and the oligomeric polyisocyanate accumulates in the second solvent system. The first solvent system comprises a linear, branched or cyclic alkane, chloroaromate and/or bromoaromate, and the second solvent system comprises a solvent selected from the group comprising nitriles, cyclic carbonates, cyclic ethers, sulfones, sulphoxides, halogenated (wherein halogen does not represent fluorine) alkanes and/or dialkyl ethers of oligoethylene oxides.

Description

Process for the preparation of oligomeric polyisocvanates with subsequent fractional liquid-liquid extraction

The present invention relates to a process for the oligomerization of isocyanates, comprising the step of reacting, in particular, a low molecular weight mono-, di- or triisocyanate (mixture) with subsequent fractional liquid-liquid extraction, the extraction comprising a two-phase solvent pair comprising first solvent system and a second solvent system is used and the starting material accumulates in the first solvent system and the oligomeric polyisocyanate in the second solvent system.

The oligomerization of diisocyanates such as especially hexamethylene diisocyanate (HDI) to low molecular weight oligomers is characterized by a classic consecutive reaction sequence in which a diisocyanate monomer molecule (A) with another monomer molecule (with uretdione formation, ideal structure: A2) or two other monomer molecules (under isocyanurate) - and / or iminooxadiazinedione formation, ideal structures A ₃ ) reacts. These ideal structures react further with all compounds of the system containing isocyanate groups (A, A2, A ₃ , and higher oligomers with more than one uretdione, isocyanurate and / or hninooxadiazinedione group). The target product for technical use is an oligomer mixture with a high ideal structure content. The higher molecular weight oligomers are less desirable because they bring about a higher viscosity of the overall system with a reduced isocyanate group content, based on the weight of the oligomer mixture. A mechanistic study of cyclooligomerization can be found, for example, in IT Horvath, LU Richter et al., Angew. Chem. Int. Ed. 2005, 45, pages 107-110. An overview of isocyanate oligomerizations can be found in J. Prakt. Chem./Chem. Ztg. 1994, 336, pages 185-200.

No. 5,298,431 describes a multistage process for isolating a cyclotrimerized isocyanate from a mixture which contains this cyclotrimerized isocyanate and higher isocyanate oligomers. The method comprises the steps of: (a) contacting the mixture with a liquid solvent to obtain a solvent-containing mixture; (b) extracting the solvent-containing mixture by means of liquid-liquid extraction in order to separate the mixture into an extract and a residue, the extract being a cyclotrimerized isocyanate which is essentially free of oligomers and which further (after separation of the solvent) is one has reduced viscosity compared to the original polyisocyanate mixture; and (c) separating the uncyclotrimerized monomer from the extract to obtain a reduced viscosity trimer product which is substantially monomer free. The product with reduced viscosity, which is characterized by this procedure was obtained. Although this patent speaks of a liquid-liquid extraction, no solvent pairs are used here. Rather, the reaction mixture is extracted with a single solvent.

No. 4,871,460 discloses how isocyanate prepolymers with free NCO groups are separated from reaction mixtures which contain unreacted excess di- or polyisocyanate monomer. This is done by extracting the reaction mixture with an inert gas, especially carbon dioxide, in either the liquid or supercritical state. The highly purified polyisocyanates are well suited for the production of foams, elastomers, adhesives, lacquers and paints.

EP 0 550 908 A2 relates to a process for the isolation of polyisocyanate isomers in pure or enriched form from at least two polyisocyanate mixtures containing polyisocyanate isomers by liquid-liquid extraction using a two-phase system consisting of a non-polar and a polar phase containing hydrogen halide. This patent application relates to the separation of isomers of polyisocyanates. The isomer separation of the polyisocyanate MDI with its isomers such as 4,4'- and 2,4'-MDI is specifically described.

DE 10 2004 06 0131 A1 relates to a process for the preparation of polyisocyanates from diisocyanates and their use. The process is a continuous process for the partial trimerization of (cyclo) aliphatic isocyanates in the presence of at least one catalyst and is distinguished in that the process is carried out at least partially in at least two backmixed reaction zones.

Examples of further processes for the preparation of polyisocyanate prepolymers are in US Pat. No. 5,258,482, EP 2,042,485 Al, US Pat. No. 5,202,001, US Pat. No. 6,133,3415, US Pat. No. 3,883,577, US Pat 4,888,442.

In order to maximize the proportion of ideal structures in the oligomer mixture, the reaction has so far mostly been terminated after partial conversion by deactivating the catalyst and the monomer, e.g. by means of vacuum rectification, separated from the product mixture and returned to the reactor. The higher the conversion in the oligomerization reaction, the greater the proportion of higher oligomers. Due to the distillative step, such a process is associated with a high energy input and with a thermal load on the process products and recycled process starting materials.

The object of the present invention was therefore to provide a process by which the desired polyisocyanate has a higher molecular uniformity, ie a higher one Ideal structure share can be obtained, the recycling of the catalyst in a recycle stream enables and / or in which less energy is required compared to conventional processes.

This object is achieved according to the invention by a process for the preparation of oligomeric polyisocyanates, comprising the steps:

(a) reaction of a starting material to an oligomeric polyisocyanate, the reaction taking place in a reaction vessel;

(b) performing a fractional liquid-liquid extraction on the reaction mixture obtained from step (a), the extraction taking place in an extraction container different from the reaction container, wherein in the liquid-liquid extraction a two-phase solvent pair comprising a first solvent system and a second solvent system is used, the first solvent system being a linear, branched or cyclic, optionally partially or perfluorinated alkane (preferably all isomeric pentanes, hexanes, heptanes, octanes, decanes, undecanes, dodecanes, tridecanes, tetradecanes, pentadecanes, hexadecanes, cyclohexane, Tetralin and any mixtures with one another), chloroaromatics and / or bromoaromatics (such as p-dibromobenzene) and the second solvent system comprises a solvent which is selected from the group comprising nitriles (preferably acetonitrile, propionitrile, phthalonitrile, dimethylmalonitrile, M alononitrile, 3-butenitrile, particularly preferably acetonitrile), cyclic carbonates (preferably propylene and ethylene carbonate), cyclic ethers (preferably trioxane), sulfones (preferably diethylsulfone), sulfoxides, nitroaromatics (preferably 2-nitrofuran, 1,3,5- Trinitrobenzene), lactones (preferably 3H-furan-2-one, angelicalactone, gamma-butyrolactone), esters (preferably dimethyl oxalate), inorganic esters (preferably dimethyl sulfate), aldehydes (preferably 2-furfural, 3-furfural), formamides ( preferably N-formylmorpholine), halogenated (where halogen is not fluorine) alkanes and / or dialkyl ethers of oligoethylene oxides.

In the process according to the invention, an oligomerization of the polyisocyanates takes place in step (a). This is preferably a trimerization and particularly preferably an isocyanurate and / or iminooxadiazinedione formation. Suitable starting materials for the oligomerization are, for example, hexamethylene diisocyanate (1,6-HDI), 1,8-octane diisocyanate, 4-isocyanatomethyl-1, 8-octane diisocyanate, 1,10-decane diisocyanate, 1,11-undecane diisocyanate, 1,12-dodecane diisocyanate, 2-methyl-1,5-diisocyanatopentane, 1,3- and 1,4-cyclohexane diisocyanate, norbornane diisocyanate, 1,3- and 1,4-bisisocyanatomethylcyclohexane, 1,3- and 1,4-bis (isocyanatoethyl) cyclohexane, 2 , 4- and 4,4-diisocyanatodicyclohexylmethane, 1-isocyanato-l-methyl-4 (3) -isocyanatomethylcyclohexane (IMCI), isophorone diisocyanate (IPDI), 1,3- and 1,4-bis (isocyanatomethyl) benzene, 2, 2,4-trimethyl-l, 6-hexamethylene diisocyanate, 2,4-tolylene diisocyanate (TDI), 2,6-TDI, 4,4-diphenylmethane diisocyanate (MDI), 4,2-MDI, 1,3- and 1 , 4-diisocyanatobenzene, 3,3-dimethyl-4,4-diisocyanatodiphenyl, 3,4- and 4,4-diisocyanatodiphenyl ether and / or 1,5-diisocyanatonaphthalene. In addition, monoisocyanates such as, for example, methyl, ethyl, 1- and / or 2-propyl, butyl (all isomers), pentyl (all isomers) and / or hexyl isocyanates (all isomers) and the higher thereof can also be used in the educt mixture Homologues may be present individually or in any mixture with one another and with the di- and triisocyanates mentioned above.

In step (b) of the process, a liquid-liquid extraction takes place in an extraction container different from the reaction container. In this respect, the method according to the invention can also be regarded as a "retrofit" method.

By definition, the solvent system pair forms two phases if a phase boundary between the two liquid phases can be seen with the naked eye. The first and the second solvent system each contain a solvent such that the two pairs of solvent systems show a miscibility gap within a certain temperature range. Likewise, the first and second solvent systems can each be present as mixtures of solvents, so that the two pairs of solvent systems show a miscibility gap within a certain temperature range.

Also included is the fact that the first and second solvents originally present in pure Lorm partially mix with each other, whereby a first solvent system with the first solvent as the main component and the second solvent as a secondary component and a second solvent system with the second solvent as the main component. and the first solvent is obtained. Here too there is a phase boundary between the first and second solvent systems.

By choosing the solvent system pair, the educt accumulates in the first solvent system and the more polar oligomeric polyisocyanate accumulates in the second solvent system. As a result of this separation, there is little starting material in the second solvent system and this is therefore missing as a reaction partner for the further reaction of the oligomeric polyisocyanates to undesirable higher oligomers. In the example of the trimerization of diisocyanates, after the reaction of one diisocyanate monomer molecule (A) with two other monomer molecules to form an isocyanurate (A3), further isocyanate (A) for the formation of higher molecular weight oligomers (A5 ₊ ) is missing.

A technical advantage of the procedure according to the invention is the energy saving in comparison with the known processes and the lower temperature load on the oligomers and, if appropriate, unreacted monomers, since in the event of a subsequent separation only solvents need to be converted into the gas phase and not the high-boiling monomers such as, for example, diisocyanates. The process also enables the catalyst to be recycled in a recycle stream.

A further energy saving results from the fact that the conversion of the monomer can be driven to higher conversions in comparison with the known methods without the gel point being exceeded. This is made possible by the separation of the oligomers, in particular the ideal structures, during the reaction and the resulting slowdown or at least partial prevention of their further reaction. Turnovers, based on the amount of substance of the monomer used, are possible between 30% and 99.9% or between> 30% and <99%. As a result, the amount of recirculated monomer is significantly reduced compared to classic single-stage oligomerization in a homogeneous phase.

Embodiments of the method according to the invention are described below, the embodiments being able to be combined with one another as desired, unless the context clearly indicates the opposite.

In one embodiment of the process according to the invention, the volume ratio of first to second solvent system in step (b) is> 1: 1 to <50: 1. The volume ratio is preferably> 5: 1 to <10: 1.

In a further embodiment of the process according to the invention, the reaction in step (a) takes place in the absence of a solvent. After a reaction of the educt in bulk, the extraction then follows with the solvent system pair.

In a further embodiment of the process according to the invention, the reaction in step (a) takes place in the presence of the first and / or the second solvent system. This variant is also possible and can bring advantages, for example, if the reaction mixture has a high viscosity. In a further embodiment of the process according to the invention, the starting material comprises hexamethylene diisocyanate. The 1,6-isomer is preferred. Its isocyanurate trimer can be used for a wide range of applications, particularly in the manufacture of polyurethane coatings. With the exception of technically unavoidable impurities, the starting material used is preferably exclusively hexamethylene diisocyanate.

In a further embodiment of the method according to the invention: in step (b), the extraction takes place in a multi-stage mixer-separator system operated in countercurrent with the first and second solvent systems within the extraction container, the extraction container having a first end, an opposite end to the first has a second end and an intermediate region arranged therebetween and the reaction mixture from step (a) is introduced into a mixer-separator stage located in the central region of the extraction container; wherein a stream comprising the first solvent system is withdrawn from the first end of the extraction container, introduced into a first solvent recovery unit, the first solvent system is separated, the product obtained by this separation is returned to the reaction container and the first solvent system obtained by this separation in the second end of the extraction container is returned, and wherein a stream comprising the second solvent system is withdrawn from the second end of the extraction container, is introduced into a second solvent recovery system, the second solvent system is separated off, the product obtained by this separation is taken off and by this separation obtained second solvent system is returned to the first end of the extraction container.

The extraction container with its various mixer-separator stages is usually upright. Then the first end of the container is the upper end, the second end is the lower end and the central area to which the reaction mixture is applied is preferably the geometric center of the entirety of the mixer-separator stages or is based on half of the theoretical Floor number of the system defined.

In an alternative embodiment of the method according to the invention, the extraction container is designed as a packed column or pulsed or non-pulsed sieve plate column. The separation within the first and second solvent recovery units is preferably done thermally. However, it is also possible that other methods, such as separation via a semipermeable membrane or a temperature reduction, are used.

With the procedure described here, a cycle with recovery of the catalyst and unreacted starting materials is possible.

It is furthermore possible that, at a point further to the first end of the extraction container than the removal of the stream comprising the first solvent system, another stream comprising the first solvent system is withdrawn and introduced into a third solvent recovery unit, a separation of the first solvent system takes place, the product obtained by this separation is returned to the reaction container and the first solvent system obtained by this separation is returned to the second end of the extraction container.

The product obtained by the separation can in particular be a catalyst which is preferably in the first solvent system. An energy-efficient internal recycling of the first solvent system is achieved by this procedure, especially if the other components of the reaction system with the exception of the catalyst have already been removed beforehand. The separation within the third solvent recovery unit is preferably done thermally. However, it is also possible that other methods, such as separation via a semipermeable membrane or a temperature reduction, are used.

It is furthermore possible that, in addition to a point further to the second end of the extraction container than the removal of the location of the stream comprising the second solvent system, another stream comprising the second solvent system is removed and this is returned to the first end of the extraction container. In this way, internal recycling without energy-intensive separation steps for the second solvent system is made possible, especially if the remaining components of the reaction system have been removed from the second solvent system beforehand.

It is further preferred that the first solvent system comprises linear, branched and / or cyclic alkanes, such as, for example, “-hexane, nI ieptane, ao-octane, cyclohexane and / or hexadecane, and the second solvent system comprises acetonitrile. With the exception of technically unavoidable traces, there are preferably no further solvents other than the solvents mentioned in the first and second solvent systems. With this selection of solvent combinations, gaps in the mixture were verified experimentally.

In a further embodiment of the process according to the invention, step (b) is carried out after a conversion of> 1% by weight to <35% by weight, based on the amount of the isocyanate groups originally present in step (a). The degree of conversion can be determined, for example, by titration of the NCO groups of a reaction sample, by means of in situ IR spectroscopy or by means of HPLC chromatography. This conversion is preferably> 15% by weight to <32% by weight and particularly preferably> 20% by weight to <28% by weight.

In a preferred embodiment of the process according to the invention, step (a) is carried out in the presence of a catalyst. The catalyst is used, for example, in a proportion of> 0.1% by weight to <35% by weight, based on the amount of starting material used. A proportion of 1% by weight to <30% by weight is particularly favorable. The fact that the catalyst can be separated off and recycled in the process according to the invention means that high catalyst contents can also be used expediently.

In a preferred embodiment of the process, no deactivation of the catalyst is carried out after step (a). Deactivation here means in particular the addition of a catalyst poison or a substance which reacts chemically with the catalyst.

Preferred catalysts for the isocyanate trimerization with high solubility in the non-polar solvent system are representatives of the proazaphosphatranes such as 2,8,9-trialkyl-2,5,8,9-tetraaza-l-phosphabicyclo [3.3.3] undekan, 2.8 , 9-Tricycloalkyl-2,5,8,9-tetraaza-l-phosphabicyclo [3.3.3] undecane or 2,8,9-triaralkyl-2,5,8,9-tetraaza-l-phosphabicyclo [3.3.3 ] undecane, where the substituents in the 2-, 8- and 9-position may be the same or different. CI-CIO radicals which can be linear or branched, preferably branched, come as alkyl substituents, C4-C12 radicals which can in turn be substituted by alkyl radicals and cycloalkyl radicals and benzyl radicals which in turn can be substituted by alkyl or alkoxy radicals, in question. The catalyst preferably comprises 2,8,9-tri-o-methoxybenzyl-2,5,8,9-tetraaza-1-phosphabicyclo [3.3.3] undecane, 2,8,9-tricyclopentyl-2,5,8, 9-tetraaza-l-phosphabicyclo [3.3.3] undecane, 2,8,9-triisopropyl-2,5,8,9-tetraaza-l-phosphabicyclo [3.3.3] undecane and / or 2.8, 9-tri-iec-butyl-2,5,8,9-tetraaza-l-phosphabicyclo [3.3.3] undecane. The latter compound is shown as an example in the following formula:

In the solvent system _pair n-heptane / acetonitrile, for example, it has a distribution coefficient K = x «- _HePtan / x _Acetomt rii of 3.9.

The catalyst preferably comprises polyhedral oligomeric silsesquioxanes of the general formula (I):

wherein RI, R2, R3, R4, R5, R6, R7 and R8 are independently selected from groups (Ia) and / or (Ib) with the proviso that at least one of the radicals RI, R2, R3, R4, R5 , R6, R7 and R8 is selected from group (Ib): (Ia): alkyl, cycloalkyl, aryl, heteroaryl, heterocyclyl, alkenyl, alkynyl, perfluoroaryl,

Perfluoroalkenyl, perfluoroalkynyl, alkoxy, perfluoroalkyl, perfluoroalkoxy, polyoxyalkylene,

(Ib): -A-R9, where:

A = (CR ₂ ) _n with n = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 or 15, and R for identical or different substituents from the group H and / or Ia; or

A = ortho-, meta- or para-C _ö R _t , where R has the meaning given above; or

A = (CR ₂ ) ₂ -Si (R ₂ ) - (CR ₂ ) ", with n = 1, 2, 3, 4, where R has the meaning given above; and R9 = N (R10) (R11), P (R10) (R11), N (R10) (R11) (R12) X, P (R10) (R11) (R12) X,

COOM, OM, or a further substituent according to formula (I), wherein

RIO, RI 1 and RI 2 are independently alkyl or aryl,

X is carboxylate, alcoholate, hydrogen carbonate, or carbene and M is ammonium, phosphonium, an alkali or alkaline earth metal cation.

Such catalysts are particularly suitable for the trimerization of monomeric polyisocyanates such as 1,6-HDI to isocyanurates.

This choice of the catalyst is based on the knowledge that its solubility or its distribution coefficient can be influenced in a two-phase system by a modular structure consisting of a non-polar partial structure and a catalytically active partial structure.

For example, the non-polar and the catalytically active partial structure can carry opposite ionic charges and be held together by means of electrostatic interactions. It is also possible for the two substructures to be connected to one another by a covalent bond.

The non-polar partial structure of the catalyst is formed in the present invention by the substituted silsesquioxane structure. The catalytically active substructure is bound via one or more substituents as described above.

The catalytic partial structure comprises an ionic or non-ionic nucleophile that catalyzes the starting oligomerization. An example of an ionic catalytically active partial structure is a silsesquioxane in which one or more of the substituents RI to R8 are - (CH2) _n -P (alkyl) 3X and X = R-COO with R = CH3-, CH3CH2-, « -Propyl, ao-propyl, ieri.-butyl, tert.-amyl, CH3 (CH2) _n - with n = 4, 5, 6, 7, 8, CH3CH (OH), phenyl, 4-methoxyphenyl and / or 3 , 4,5-trimethoxyphenyl or RO with R = phenyl or 2,6-bis (l, l-dimethylethyl) -4-methylphenyl. An example of a non-ionic catalytically active partial structure is given if one or more of the substituents RI to R8 are - (CH2) _n -P (alkyl) 2. The exocyclic Si-substituted substituents R2 to R8 which are not substituted by catalytically active substructures are preferably ao-butyl, cyclohexyl and / or ao-octyl groups, particularly preferably ao-butyl groups, both in the ionic and in the nonionic case ,

According to the invention it can be provided that the substituent R9 is a further substituent according to formula (I). This includes oligomeric catalysts with more than one silsesquioxane unit, for example dimers or trimers. The fact that the distribution of the catalyst in a two-phase system can be influenced means that a process can be carried out which, instead of exclusively separating the starting material by distillation, permits extraction which takes place at the same time as the catalyzed reaction or is carried out downstream of the catalyzed reaction. In particular, a process can be implemented in which the catalyst and the starting material are preferably in the non-polar phase and the oligomeric polyisocyanate is preferably in the polar phase. To obtain the desired oligomer, only the polar solvent has to be distilled off after phase separation. Reactive extraction processes can also be implemented. According to the present invention:

Alkyl: acyclic aliphatic hydrocarbon residues that do not contain any C-C multiple bonds. Alkyl is preferably selected from the group consisting of methyl, ethyl, propyl, 2-propyl, n-butyl, isobutyl, sec-butyl, isobutyl, n-pentyl, ao-pentyl, eo- Pentyl, "-I Iexyl," -I Ieptyl, n-octyl, n-nonyl and / or n-decyl.

Cycloalkyl: cyclic aliphatic (cycloaliphatic) hydrocarbons with in particular 3, 4, 5, 6, 7, 8, 9 or 10 carbon atoms, where the hydrocarbons can be saturated or unsaturated (but not aromatic), unsubstituted or mono- or polysubstituted. The cycloalkyl radicals can furthermore be bridged once or several times, for example in the case of adamantyl, bicyclo [2.2.1] heptyl or bicyclo [2.2.2] octyl. Cycloalkyl is preferably selected from the group comprising cyclopropyl, cyclobutyl, cyclopentyl, cyclohexyl, cycloheptyl, cyclooctyl, cyclononyl, cyclodecyl, adamantyl, cyclopentenyl, cyclohexenyl, cycloheptenyl and / or cyclooctenyl.

Alkenyl: acyclic aliphatic hydrocarbon radicals which have at least one C-C double bond. Alkenyl is preferably selected from the group comprising ethenyl (vinyl), propenyl, butenyl, pentenyl, hexenyl, heptenyl, octenyl, nonenyl and / or decenyl.

Alkynyl: acyclic aliphatic hydrocarbon radicals which have at least one C-C triple bond. Alkynyl is preferably selected from the group comprising ethynyl, propynyl, butynyl, pentynyl, hexynyl, heptynyl, octynyl, nonynyl and / or decynyl.

Aryl: aromatic hydrocarbons with up to 14 ring members, especially phenyls and naphthyls. Each aryl radical can be unsubstituted or mono- or polysubstituted, and the aryl substituents can be the same or different and can be in any desired position of the aryl. Aryl is preferably selected from the group containing phenyl, 1-naphthyl and 2-naphthyl, each of which is unsubstituted or one or more times can be substituted. A particularly preferred aryl is phenyl, unsubstituted or mono- or polysubstituted.

Heteroaryl: a 5- or 6-membered cyclic aromatic radical which contains at least 1, optionally also 2, 3, 4 or 5 heteroatoms, where the heteroatoms are each independently selected from the group S, N and O and the heteroaryl Radical may be unsubstituted or mono- or polysubstituted; in the case of substitution on the heteroaryl, the substituents can be identical or different and can be in any desired position of the heteroaryl. It is preferred that the heteroaryl residue is selected from the group consisting of benzofuranyl, benzoimidazolyl, benzothienyl, benzothiadiazolyl, benzothiazolyl, benzotriazolyl, benzooxazolyl, benzooxadiazolyl, quinazolinyl, quinoxalinyl, carbazolyl, quinolothienyl, dibenzylyl, dibenzylyl, dibenzylyl , Imidazolyl, imidazothiazolyl, indazolyl, indolizinyl, indolyl, isoquinolinyl, isoxazoyl, isothiazolyl, indolyl, naphthyridinyl, oxazolyl, oxadiazolyl, phenazinyl, phenothiazinyl, phthalazinyl, pyrazolyl, pyridyl, 3-pyridyl, 2-pyridyl, 2-pyridyl, 2-pyridyl, 2-pyridyl, 2-pyridyl, Pyridazinyl, pyrimidinyl, pyrazinyl, purinyl, phenazinyl, tetrazole, thienyl (thiophenyl), triazolyl, tetrazolyl, thiazolyl, thiadiazolyl and / or triazinyl.

Heterocyclyl: aliphatic saturated or unsaturated (but not aromatic) cycloalkyls with in particular three to ten, ie 3, 4, 5, 6, 7, 8, 9 or 10 ring members, in which at least one, optionally also two or three carbon atoms by a hetero atom or a heteroatom group each independently selected from the group consisting of O, S, S (= 0) ₂ , N, NH and N (Cl-8-alkyl), preferably N (CH3), are replaced, the ring members being unsubstituted or can be substituted one or more times. Preferred are heterocyclyl radicals from the group comprising azetidinyl, aziridinyl, azepanyl, diazepanyl, dithiolanyl, dihydroquinolinyl, dihydropyrrolyl, dioxanyl, dihydroindolol, dihydrofurylol, oxetanyl, pyrrolidinyl, piperazinyl, 4-methylpiperazinyl, piperidinyl, pyrazolidinyl, pyranyl, tetrahydropyrrolyl, tetrahydropyranyl, tetrahydroquinolinyl, tetrahydroisoquinolinyl, Tetrahydroindolinyl, tetrahydrofuranyl, tetra- hydropyridinyl, tetrahydro thiophenyl, tetrahydro-pyridoindolyl, tetrahydronaphthyl, hydrocarbolinyl tetra-, tetrahydro-isoxazolo-pyridinyl, thiazolidinyl and / or thiomorpholinyl.

Perfluoroaryl: aryl radicals as defined above, where all aromatically bound hydrogen atoms have been replaced by fluorine atoms.

Perfluoroalkenyl: alkenyl radicals as defined above, where all hydrogen atoms have been replaced by fluorine atoms. Perfluoroalkynyl: alkynyl radicals as defined above, where all hydrogen atoms have been replaced by fluorine atoms.

Alkoxy: alkyl group attached via an oxygen atom as defined above.

Perfluoroalkyl: alkyl radicals as defined above, where all hydrogen atoms have been replaced by fluorine atoms.

Perfluoroalkoxy: alkoxy radicals as defined above, where all hydrogen atoms have been replaced by fluorine atoms.

Polyoxyalkylene: polyether groups obtained from the polymerization of alkylene oxide units, in particular polymers, copolymers and block copolymers of ethylene oxide and propylene oxide.

Carboxylate: salts of carboxylic acids, especially of alkyl and aryl carboxylic acids.

Preferably:

R2, R3, R4, R5, R6, R7 and R8 ao-butyl,

RI is - (CFO _ni -PAlkyF or - (CFO _ni -PAlkyBX with nl = 2, 3, 4, 5, 6, 7 or 8 and

X is selected from the group comprising

R13-COO with R13 = CH3-, CH3CH2-, n-propyl, ao-propyl, ieri.-butyl, tert.-amyl, CH3 (CH2) _n - with n = 4, 5, 6, 7, 8, phenyl , 4-methoxyphenyl and / or 3,4,5-trimethoxyphenyl,

R14-0 with R14 = phenyl or 2,6-bis (l, l-dimethylethyl) -4-methylphenyl.

For example, the partition coefficient K = x _«-HePtan / x _Acetomt rii, defined as the amount of a compound in the non-polar phase divided by the amount of the compound in the polar phase, of a catalyst of the general formula (I) with RI = - ( CH2) 3-P (Ao-Butyl) 3 ⁺ CH3COO and R2 to R8 = Ao-Butyl in the n-heptane / acetonitrile system determined to be 0.74.

The synthesis of the ionic silsesquioxane catalysts SQ ⁺ R-COO and SQ ⁺ RO (SQ + stands for the [3- (3,5,7,9, ll, 13,15-heptaisobutylpentacyclo [9.5.1.1 ³ ' ⁹ .l ⁵ ^'15 . L ^{7 13} ] - octasilox-l-yl) propyl] (triisobutyl) phosphonium substituents) can be achieved as follows:

A mixture of l- (3-iodopropyl) -3,5,7,9, ll, 13,15-heptaisobutylpentacyclo [9.5.1.1 ³ ' ⁹ .l ^{5 15} . l ⁷ ^'13 ] - octasiloxane and triisobutylphosphine in toluene is heated under reflux and stirred. The quaternary salt obtained is isolated by removing the solvent in vacuo. Then it is dissolved in dichloromethane and mixed with a solution of the corresponding silver carboxylate Ag ⁺ R-COO or silver alcoholate Ag ⁺ RO in water. The two-phase mixture is stirred at room temperature. The organic phase is then separated off and dried. The end product SQ ⁺ R-COO or SQ ⁺ RO is obtained after removing the solvent in vacuo.

Examples of catalysts that can be obtained in this way are:

Catalyst 1 (SQ ⁺ CH ₃ -COO)

Catalyst 2 (SQ ⁺ tert-butyl-COO)

Catalyst 3 (SQ ⁺ CH ₃ (CH ₂ ) ₈ -COO)

Catalyst 4 (SQ ⁺ phenyl-COO)

Catalyst 5 (SQ ⁺ 4-methoxyphenyl-COO)

Catalyst 6 (SQ ⁺ 3,4,5-trimethoxyphenyl-COO)

Catalyst 7 (SQ ⁺ phenyl-O)

Catalyst 8 (SQ ⁺ 2,6-bis (l, l-dimethylethyl) -4-methylphenyl-0)

Catalyst 9 (SQ ⁺ CH ₃ -CH (OH) -COO)

Examples

The present invention is explained in more detail with reference to the following examples and the figures, but without being restricted thereto. Show it:

FIG. 1 shows the arrangement of the glass vessels and schematically the test sequence in Example 1

FIG. 2a the concentration profile of the non-polar phases from Example 1

FIG. 2b the concentration profile of the polar phases from example 1

FIG. 3a a simulation of a concentration profile in the non-polar phases

FIG. 3b a simulation of a concentration profile in the polar phases

FIG. 4a the concentration profile of the non-polar phases from example 3

FIG. 4b the concentration profile of the polar phases from example 3

FIG. 5 shows a reaction method with liquid-liquid extraction

FIG. 6 another reaction method with liquid-liquid extraction

In the following examples, the following were used as reagents: 1,6-diisocyanatohexane (1,6-HDI), the corresponding isocyanurate of 1,6-HDI (1,3,5-tris (6-isocyanatohexyl) -1,3,5 -triazinane- 2,4,6-trione, "symmetrical trimer"). A mixture of the isocyanurate of 1,6-HDI

("Symmetrical trimer") and the iminooxadiazinedione of 1,6-HDI ("asymmetrical trimer") in a ratio of> 95: 1 and traces of higher oligomers is referred to below as 1,6-HDI trimer. Furthermore, a commercially available 1.6-HDI oligomer mixture containing 1.6-HDI trimer and higher oligomers in a ratio of 63.05% by weight 1.6-HDI trimer ("A3"), 21.00% by weight 1,6-HDI pentamer ("A5"), 8.76% by weight of 1,6-HDI heptamer ("A7") and

7.19% by weight 1,6-HDI nonamer ("A9"). 1- (2-Methoxyphenyl) piperazine (MPP) was used as the internal standard for HPFC chromatography. 2,8,9-Tri-.vec-butyl-2,5,8,9-tetraaza-l-phosphabicyclo [3.3.3] undecane (TTPU) was also used.

The concentrations of 1,6-HDI, trimer and higher oligomers were determined by HPFC chromatography. For the measurement, 0.1 g of the sample to be measured was diluted to a volume of 5 ml with acetonitrile. From this solution, 0.5 ml was diluted to a volume of 5 ml with acetonitrile. 0.1 ml of this solution was mixed with 0.4 ml of a solution of 250 mg MPP in 50 ml of acetonitrile and 0.5 ml of a solution of 4.07 g of dichlorobenzene in 391.53 g of acetonitrile. The solution was shaken for 30 seconds. Then 0.01 ml of Solution was injected into the sample loop of an HPIL device from Agilent (1200 Series), which was equipped with a Lichrospher 60RP Select B column (5 pm, 125 x 4 mm). The column was thermostatted to 35 ° C. A mixture of (A) 550 ml water, 360 ml acetonitrile and 90 ml phosphate buffer pH = 7 and (B) acetonitrile was used as eluent (1.5 ml / min). The gradient was 97% A (2 min) - 15 min - 37% A (2 min). The substances were detected with a UV detector at 254 nm (reference 360 nm). The HPLC method was calibrated with solutions of pure 1,6-HDI and the isocyanurate of 1,6-HDI. In the following, the symmetrical and the asymmetrical 1,6-HDI trimer were integrated together. The same response factor as for 1.6-HDI trimer was assumed for the higher oligomers.

In the context of the present invention, the description "SQ ⁺ " is used to describe the catalyst for the silsesquioxane cation, in which R2, R3, R4, R5, R6, R7 and R8 are ao-butyl and RI for - (C 11: ) ₃ - P (Ae. »- B uty I) ₃ stands. An example of this name is the acetate SQ ⁺ CH ₃ COO-.

The synthesis of the ionic silsesquioxane catalyst SQ ⁺ CI LCOO was carried out according to the following procedure: a mixture of l- (3-iodopropyl) -3,5,7,9, ll, 13,15-heptaisobutyl-pentacyclo [9.5.1.1 ^{3 , 9} .l ^5.15 .l ^7.13 ] octasiloxane (9.86 g, 10 mmol) and triisobutylphosphine (2.02 g, 10 mmol) in 30 mL toluene was heated under reflux for 6 hours and stirred. The quaternary salt obtained was isolated by removing the solvent in a partial vacuum. The residue was then dissolved in 50 ml of dichloromethane and a suspension of the corresponding silver carboxylate Ag ⁺ CI LCOO (1.67 g, 10 mmol) in 100 ml of water was added. The mixture was stirred at room temperature for 8 hours. The mixture was then filtered, the organic phase separated and dried over sodium sulfate. The final product SQ ⁺ CI FCOC V was obtained after removal of the solvent in a partial vacuum and characterized by means of high-resolution mass spectroscopy, infrared spectroscopy, X-ray fluorescence analysis (RFA) and multi-core NMR spectroscopy, the proton and carbon spectra with those in the figure Letter-marked groups were clearly assigned by comparison with literature data based on displacement and integral.

The mass spectra were measured by electrospray ionization on a Thermo Fisher Scientific Orbitrap XL. For this purpose, the samples were dissolved in CHCI3, the solution was diluted with MeOH (containing 0.1% acetic acid) and injected into the ESI source system via the direct inlet using a syringe pump (electrospray voltage 4 kV, current 0.7 mA, volume flow Sheath Gas 5 arb, volume flow of surge gas 5 arb, capillary temperature 300 ° C, capillary voltage 0.05 V, tube lens voltage 150 V). The detection was carried out with an Orbitrap ion trap.

The infrared spectra were measured on a Bruker alpha FT-fR spectrometer, the samples in each case being applied in bulk to the ATR crystal. The spectra were measured in the range from 4000 to 400 cm ¹ with a resolution of 4 cm ¹ by averaging 32 individual spectra against air as the background spectrum (averaging 100 individual spectra).

The ICP analysis was carried out on an ICP-MS-HR Element2 equipped with an ECM detector using argon as cooling, auxiliary, carrier and additional gas, generator power 1200 - 1600 watts, masses: ³¹ P, ²⁹ Si, quantification via external Calibration with rhodium as internal

Default.

The ³¹ P { ¹ H} NMR spectra were measured at 161.9 MHz in CDCL on a Bruker AV400 Ultrashield. The 11 NMR spectra were measured at 400 MHz in CDCL with a Bruker AV400. The chemical shifts were calibrated relative to the solvent signal (CDCL, d = 7.26 ppm). ¹³ C { ¹ 11} NMR spectra were measured at 100.6 MHz in CDCL with a Bruker AV400 Ultrashield. The chemical shifts were calibrated relative to the solvent signal (CDCL, d = 77.16 ppm). The signals were assigned by comparison with the ¹³ C APT NMR spectra of the products (APT = Attached Proton Test). Corresponding to the solvent signal from CDCL (positive polarity), signals with positive polarity were assigned to quaternary C atoms or CH2 groups, signals with negative polarity to CH or CH3 groups. The type of splitting is specified as a singlet (s), two singlets (2xs), three singlets (3xs), doublet (d), triplet (t), multiplet (m) or broad signal (b).

HR-MS (ESI): calculated 1059.4794 (C43I Le.OizPSiC), found 1059.4753; calculated 59.0100 (C2H3O2), found 59.0131.

FT-fr 2913 (s), 1464 (s), 1366 (s), 1229 (m), 1084 (b), 837 (m), 744 (m), 472 (m) cm ¹ . The silicon and phosphorus content was determined by ICP analysis to be 17.0% by weight Si (theoretically: 20.1%) and 2.1% by weight P (theoretically: 2.8%).

³¹ R {Ή} NMR (CDC1 ₃ , 161.9 MHz): d = 30.24 (s) ppm.

Table 1-la: Chemical shift and assignment of the characteristic signals for catalyst 1 (SQ ⁺ CH ₃ -COO) in the II NMR spectra.

Table 1-lb: Chemical shift and assignment of the characteristic signals for catalyst 1 (SQ ⁺ CH ₃ -COO) in the ³ C f ¹ 11} NMR spectra.

example 1

The fractional extraction was demonstrated using the following procedure:

First, 15 mL n-I Ieptane and 2.5 mL anhydrous acetonitrile were equilibrated in 15 glass vessels at 60 ° C. The glass vessels were numbered 1 to 15. A cyclic process was then repeated 40 times:

0.1 mL of a mixture of approx. 50% (mass fraction in each case) acetonitrile and 25% 1,6-HDI and 25% 1,6-HDI trimer were added to the central vessel (No. 8).

All glass jars were stirred for 3 minutes. After settling (approx. 5 seconds), 5 mL were removed from the upper, n-Icptan-rcichen phase. 0.5 mL (from tubes 1 to 8) and 0.4 mL (from tubes 9 to 15) of the lower,

Acetonitrile-rich phase removed.

From vials 1 to 14 (number j), the upper phase sample was added to the next vial (number _j +1). The upper phase sample from vessel 15 was collected as raffinate and 5 mL of monomer and trimer-free equilibrated nI ieptane was added to vessel # 1.

The samples from the lower phases were distributed as described in the previous step, but distributed in the opposite direction to the vessels. The extract was removed from vessel No. 1 and monomer-free and trimer-free acetonitrile was added to vessel 15.

At the end of the experiment, a concentration profile was created and analyzed by HPLC. The procedure is schematic in LIG. 1 reproduced.

To simulate the procedure, a discontinuous model based on batch contacts and a linear driving force approximation was used for the mass transfer between the phases. The liquid-liquid equilibrium parameters in such a model (non-linear equilibrium isotherm) were measured for the system acetonitrile / nI ieptane / I IDI / I IDI-Trimcr at 60 ° C. Mass transition parameters for monomer, trimer, pentamer, heptamer and nonamer were estimated to be 0.75, 0.39, 0.29, 0.23 and 0.20 min ¹ , respectively. The partition coefficient for the catalyst was the same as that for the nonamer.

The experimentally determined concentration profiles after 40 runs and the mathematical modeling in equilibrium are in LIG. 2a (upper, nI Icptau-rciche phase) and LIG. 2b (lower, acetonitrile-rich phase). With the counter j, the x-axes indicate the number of the respective vessel and the y-axes the mass fraction M of the respective component in the non-polar phase (M „) and polar phase (M ,,). Rectangles () for the measured values and the upper line for the result of the simulation with regard to vessel number 1 are assigned to the HDI trimer ("A3"). Rectangular rectangles (¨) and the lower line with regard to vessel number 1 are assigned according to the monomeric HDI ("A").

From FIG. 2a and 2b it can be seen that it is possible to separate monomeric and trimeric HDI from one another by means of fractional liquid-liquid extraction. Mathematical modeling correlates best with the experimental data on the trimer. From this it is concluded that its profile is already close to the state of equilibrium.

Example 2

In this example, the simulations from Example 1 were additionally carried out with 1% by weight (based on the amount of HDI originally used) of a catalyst which is distributed predominantly in the n-heptane-rich phase. The mass fraction of the catalyst is shown by the dashed line. Results are analogous to FIG. 2a and 2b in FIG. 3a and 3b. The curves for the HDI monomer and trimer correspond to those from FIG. 2a and 2b. It can be seen that the catalyst is distributed towards vessel number 15 and can thus be separated from the product, the HDI trimer.

Example 3

In this example, the procedure was analogous to example 1. However, a reaction mixture was used, as was also obtained in an actual experiment. The composition of the mixture was 0.5% by weight of the catalyst SQ ⁺ CI LCCXL ("Cat."), 24.88% by weight of 1,6-HDI monomer ("A"), 15.68% by weight of 1 , 6-HDI trimer ("A3"), 5.22% by weight 1.6-HDI pentamer ("A5"), 2.18% by weight 1.6-HDI heptamer ("A7") and 1.79% by weight 1,6-HDI nonamer ("A9") in 49.75% by weight acetonitrile. Such a composition can be obtained in a simple manner from 50% by weight of acetonitrile, 25% by weight of 1,6-HDI and 25% by weight of a commercially available 1,6-HDI oligomer mixture and the corresponding addition of catalyst.

The experimentally determined mass fractions of A, A3, A5, A7 and A9 are shown in FIG. 4a and FIG. 4b reproduced as corresponding data points. FIG. 4a shows the mass fractions in the non-polar phase and FIG. 4b in the polar phase. FIG. 4a and 4b as solid or different dashed lines, results of simulations of the concentration profile of the individual components. It can be seen that the catalyst follows the nI icptan and is driven completely towards the raffinate (vessel number 15). In this way, the catalyst can be separated from the product. It can also be seen how unreacted monomer can be separated from the reaction mixture and is thus available for a continuous reaction process.

It should be generally noted that in the figures described above, the deviations of the simulated concentration profiles from the experimentally determined ones stem from the fact that a steady state was taken as a basis in the simulations. In the experiments, this has in some cases not yet been reached after 40 cycles.

Example 4

In this example the distribution of the catalyst SQ ⁺ CI ECOCT was investigated experimentally. The extract and the raffinate from Example 4 after the 40th cycle were analyzed for the presence of silicon and phosphorus using ICP analysis. 0.0084% by weight of Si and 0.0009% by weight of P were found in the raffinate. In contrast, the extract had 0.0380% by weight of Si and 0.0038% by weight of P. In the extract there is a 4.4 times higher catalyst concentration than in the raffinate. Evidence of the enrichment of the catalyst in the extract was thus provided.

FIG. 5 schematically shows a reaction process with downstream extraction. The trimerization of HDI is explained using this figure as an example.

The trimerization reaction takes place in reaction vessel 1. The reaction mixture comprising HDI, HDI trimer and catalyst obtained is then introduced into the central region of the extraction column 2. The extraction column 2 is designed as a multi-stage mixer-separator system operated in countercurrent for a liquid-liquid extraction. Acetonitrile is passed through the column from top to bottom, following gravity. N-I Icptan is led from bottom to top. In column 2 there is now an enrichment of HDI and the catalyst in the n-I ieptane-rich phase analogously to Example 3. The catalyst is selected so that it partitions more strongly than HDI into the n-I ieptane-rich phase.

The extract thus obtained containing HDI is removed as stream 3 from the upper region of column 2 and introduced into the solvent recovery unit 6. The nI-ieptane-rich phase is distilled off in the solvent recovery unit 6. The recovered HDI is fed back into reaction vessel 1 as stream 4. Furthermore, the distilled-off non-polar solvent is fed back to the lower end of the extraction column 2 as stream 5. Due to its higher polarity, the HDI trimer is enriched in the acetonitrile-rich phase. This is withdrawn at the bottom of the column 2 as stream 10 and introduced into the solvent recovery unit 7. The polar solvent is distilled off and fed as stream 8 to the upper end of column 2. The HDI trimer obtained after the distillation is removed as product stream 9.

Because the catalyst is partitioned even more strongly than HDI into the n-Icptane phase, the stream 11 containing catalyst can be removed at the top of column 2. The non-polar solvent is distilled off in a further solvent recovery unit 12. The recovered catalyst is fed back into reaction vessel 1 as stream 13. The solvent obtained from the distillation is fed as stream 14 to the lower end of column 2.

As can be seen from FIG. 5 can be seen, the process can be operated as a continuous, circular process, HDI trimer being continuously separated and unreacted HDI and the catalyst being made available to the reaction again. It is therefore practical to carry out a separation of the HDI trimmer even at low sales. In this way, the formation of by-products, especially higher HDI oligomers, is suppressed.

Further advantages are that there is no deactivation of the catalyst and the catalyst is no longer contained in the HDI trimer. Furthermore, the HDI trimer can be obtained monomer-free. The energy balance of this process is very favorable. Only solvents have to be distilled off. Vacuum rectification of HDI and the oligomers is not necessary. At the same time, the thermal load on the HDI and trimmer is reduced.

FIG. 6 shows schematically one of the previously in connection with FIG. 5 discussed reaction processes analogous process with subsequent extraction. The trimerization of HDI is also explained here as an example.

The trimerization reaction takes place in reaction vessel 1. The reaction mixture comprising HDI, HDI trimer and catalyst obtained is then introduced into the central region of the extraction column 2. The extraction column 2 is designed as a multi-stage mixer-separator system operated in countercurrent for a liquid-liquid extraction. Acetonitrile is passed through the column from top to bottom, following gravity. IItata is not led from bottom to top. In column 2, analogously to example 3, HDI and the catalyst are now enriched in the nIcptan-rcichen phase. The catalyst is selected so that it partitions more strongly than HDI into the nIptan-rciche phase. The extract thus obtained containing HDI is removed as stream 3 from the upper region of column 2 and introduced into the solvent recovery unit 6. In the solvent recovery unit 6, the nIcptan-rciche phase is distilled off. The recovered HDI is fed back into reaction vessel 1 as stream 4. Furthermore, the distilled-off non-polar solvent is fed back to the lower end of the extraction column 2 as stream 5.

Due to its higher polarity, the HDI trimer is enriched in the acetonitrile-rich phase. This is drawn off in the lower section of column 2 as material stream 10 and introduced into the solvent recovery unit 7. The polar solvent is distilled off and added as stream 8 to the upper end of column 2. The HDI trimer obtained after the distillation is removed as product stream 9.

Because the catalyst is partitioned even more strongly than HDI into the n-Icptane phase, the stream 11 containing catalyst can be removed at the top of column 2. The non-polar solvent is distilled off in a further solvent recovery unit 12. The recovered catalyst is re-entered as stream 13 in the reaction vessel 1. The solvent obtained from the distillation is fed as stream 14 to the lower end of column 2.

In the in LIG. 6 variant, the fact that an acetonitrile phase is present at the lower end of column 2, which is essentially free of further constituents, is also used. Then, bypassing the solvent recovery unit 7 in an internal recycling, the material stream 14 can be fed directly onto the upper end of the column 2. This further improves the energy balance of the process.

Of course, this variant also has the one previously described in connection with the method from FIG. 5 described advantages.

Claims

claims

1. A process for the preparation of oligomeric polyisocyanates, comprising the steps:

(b) performing a fractional liquid-liquid extraction on the reaction mixture obtained from step (a), the extraction taking place in an extraction container different from the reaction container, wherein in the liquid-liquid extraction a two-phase solvent pair comprising a first solvent system and a second solvent system is used, the first solvent system comprising a linear, branched or cyclic alkane, chloroaromatic and / or bromoaromatic and the second solvent system comprising a solvent which is selected from the group comprising nitriles, cyclic carbonates, cyclic ethers, sulfones, sulfoxides , Nitroaromatics, lactones, esters, inorganic esters, aldehydes, formamides, halogenated (where halogen does not represent fluorine) alkanes and / or dialkyl ethers of oligoethylene oxides.

2. The method according to claim 1, wherein in step (b) the volume ratio of first to second solvent system is> 1: 1 to <50: 1.

3. The method according to claim 1, wherein the reaction in step (a) takes place in the absence of a solvent.

4. The method according to claim 1, wherein the reaction in step (a) takes place in the presence of the first and / or the second solvent system.

5. The method according to claim 1, wherein the starting material comprises hexamethylene diisocyanate.

6. The method according to claim 1, wherein in step (b) the extraction takes place in a multi-stage mixer-separator system operated in countercurrent with the first and second solvent systems within the extraction container (2), the extraction container (2) has a first end, a the first end has opposite second end and a middle region arranged between them and the reaction mixture from step (a) to a in the middle region of the extraction container

(2) located mixer-separator stage is entered; wherein at the first end of the extraction container (2) a stream (3) comprising the first solvent system is withdrawn, introduced into a first solvent recovery unit (6), the first solvent system is separated off, the product (4) obtained by this separation into the Reaction container (1) is returned and the first solvent system (5) obtained by this separation is returned to the second end of the extraction container (2); and wherein at the second end of the extraction container (2) a stream (10) comprising the second solvent system is removed, is introduced into a second solvent recovery system (7), the second solvent system is separated off and the product (9) obtained by this separation is removed and the second solvent system (8) obtained by this separation is returned to the first end of the extraction container (2).

7. The method according to claim 6, wherein additionally at a further to the first end of the extraction container (2) than the withdrawal of the stream comprising the first solvent system

(3) a further stream (11) comprising the first solvent system is removed, this is introduced into a third solvent recovery unit (12), the first solvent system is separated off, and the product (13) obtained by this separation is transferred to the reaction container (1 ) is returned and the first solvent system (14) obtained by this separation is returned to the second end of the extraction container (2).

8. The method according to claim 6, wherein in addition at a further to the second end of the extraction container (2) than the removal of the stream (10) comprising the second solvent system, another stream comprising the second solvent system (15) is withdrawn and into the first end of the extraction container (2) is returned.

9. The method of claim 1, wherein the first solvent system comprises linear, branched and / or cyclic alkanes and the second solvent system comprises acetonitrile.

10. The method according to claim 1, wherein step (b) is carried out after a conversion of> 1% by weight to <35% by weight, based on the amount of the isocyanate groups originally present in step (a).

11. The method according to claim 1, wherein step (a) is carried out in the presence of a catalyst.

12. The method according to claim 11, wherein the catalyst is used in a proportion of> 1% by weight to <35% by weight, based on the amount of polyisocyanate used. 13. The method according to claim 11, wherein the catalyst comprises polyhedral oligomeric silsesquioxanes of the general formula (I):

wherein RI, R2, R3, R4, R5, R6, R7 and R8 are independently selected from groups (Ia) and / or (Ib) with the proviso that at least one of the radicals RI, R2, R3, R4, R5 , R6, R7 and R8 is selected from group (Ib):

(Ia): alkyl, cycloalkyl, aryl, heteroaryl, heterocyclyl, alkenyl, alkynyl, perfluoroaryl,

(Ib): -A-R9, where: A = (CR2) _n with n = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,

13, 14 or 15, and R for identical or different substituents from the group H and / or Ia; or

A = ortho-, meta- or para-C _ö R _t , where R has the meaning given above; or

A = (CR ₂ ) ₂ -Si (R ₂ ) - (CR ₂ ) _n , with n = 1, 2, 3, 4, where R has the meaning given above; and

R9 = N (R10) (R11), P (R10) (R11), N (R10) (R11) (R12) X, P (R10) (R11) (R12) X,

COOM, OM, or a further substituent according to formula (I), wherein

RIO, RI 1 and RI 2 are independently alkyl or aryl, X is carboxylate, alcoholate, hydrogen carbonate, or carbene and

M is ammonium, phosphonium, an alkali or alkaline earth metal cation.

14. The method of claim 13, wherein:

R2, R3, R4, R5, R6, R7 and R8 are a-butyl, RI is - (CH2) _ni -PAlkyl2 or - (CH2) _ni -PAlkyl3X with nl = 2, 3, 4, 5, 6, 7 or 8 and

X is selected from the group comprising

R13-COO with R13 = CH3-, CH3CH2-, n-ropyl, ao-propyl, ieri.-butyl, tert.-amyl, CH3 (CH2) _n - with n = 4, 5, 6, 7, 8, phenyl , 4-methoxyphenyl and / or 3,4,5-trimethoxyphenyl, R14-0 with R14 = phenyl or 2,6-bis (l, l-dimethylethyl) -4-methylphenyl.

15. The method according to claim 11, wherein after step (a) no deactivation of the catalyst is carried out.