PL166018B1

A method for recovery of transition metal and an ion-exchange resin

Abstract

A method for recovering a transition metal such as rhodium from a polar or non-polar liquid by contacting said transition metal-containing liquid with an ion-exchange resin having bonded ionically thereto an organophosphorus ligand; the transition metal can be eluted from the bed using a liquid containing a sufficient concentration of organophosphorus ligand.

Term

No projected expiry on record.

10 claims: 1 independent, 9 dependent

1
Patent claims Zastrzeżenia patentowe 1. A method of recovering rhodium from a rhodium-containing liquid in an amount below 20 ppm, characterized in that the rhodium-containing liquid is contacted with an ion exchange resin in which said resin has an ion-associated first organophosphorus ligand and then the said resin is contacted with a liquid containing a second organophosphorus ligand in sufficient amount to remove rhodium from the resin, preferably the same compound being used as the first and second ligand. 1. Sposób odzyskiwania rodu z cieczy zawierającej rod w ilości poniżej 20 ppm, znamienny tym, że kontaktuje się ciecz zawierającą rod z żywicą jonowymienną, w której wymieniona żywica ma związany z nią jonowo pierwszy ligand fosforoorganiczny i następnie kontaktuje się wymienioną żywicą z cieczą zawierającą drugi ligand fosforoorganiczny w ilości wystarczającej dla usunięcia rodu z żywicy, przy czym korzystnie jako pierwszy i drugi ligand stosuje się ten sam związek.

206 paragraphs in 5 sections, as filed

The present invention relates to a method for recovering rhodium from liquids, in particular from liquids containing low concentration transition metal in solution.

The invention particularly relates to a method of recovering rhodium from its liquid solutions, such as an aldehyde product from an olefin hydroformylation process.

Processes using homogeneous transition metal catalysts such as rhodium are well known. For example, transition metal catalysts are used in the processes of hydrogenation of unsaturated compounds such as copolymers of conjugated dienes with copolymerizable monomers as described in U.S. Patent Nos. 4,464,515 and 4,503,196, in carbonylation of methanol to acetic acid, in oligomerization of olefins, in hydrocyanation of butadiene to adipic acid dinitrile, in decarbonylation of aldehydes and in hydrosilylation of olefins. A particularly spectacular example of such homogeneous catalyst systems is the catalytic hydroformylation of olefinic compounds with carbon monoxide and hydrogen to produce aldehydes. In such a system, the rhodium catalyst is generally stabilized with a complexing agent, referred to as a ligand, although it is known to use a simple rhodium catalyst without such additional stabilizing agents.

In one known embodiment, the rhodium complex-catalyzed hydroformylation process is carried out in a non-aqueous hydroformylation reaction medium; the reaction medium contains an organic solvent and both a catalytic complex solubilized with an organic solvent and a solubilized free ligand, i.e. a ligand not bound or not attached to the rhodium catalytic complex. Solvents that do not interfere with the hydroformylation process are used. Suitable organic solvents include those used in known transition metal catalyzed Group VIII processes of the Periodic Table of Elements, such as alkanes, ethers, aldehydes, ketones, esters, amides, aromatic hydrocarbons and mixtures of various organic solvents. Examples of such non-water processes

166 OS8 hydroformylation is described in U.S. Patent Nos. 4,148,830, 3 527 809, 4247486, 4260828, 4283 562, 4306087, 4400548, 4429161, 4482749 4491675, 4528403, 4593011, 4593127, 4599206, 463302, 466868 4 717 775, 4731486, 4737 588, 4748261, in European Patent Application Publication Nos. 96986, 96987 and 96988 (all published on December 28, 1983), in published PCT applications No. WO 80/01690 (publication of August 21, 1980), and WC 87/07600 (publication of December 17, 1987), to name a few. In these systems, the product can be recovered by selective evaporation of the aldehyde under reduced pressure and at temperatures below about 150 ° C, preferably below about 130 ° C.

It is also known to use water or a similar polar solvent such as methanol as a catalytic reaction medium during olefin hydroformylation. One such process is described in U.S. Patent No. 4,248,802, which uses water-soluble rhodium complexes and certain salts of sulfonated triarylphosphines as hydroformylation catalysts. The separation of the aldehyde product from the catalyst in this process is facilitated by the product and the catalyst solution not being mixed together. However, this approach usually requires stricter reaction conditions than in non-aqueous systems.

A new method of recovering, by simple phase separation, aldehyde products produced during hydroformylation of olefinic compounds in a non-aqueous reaction medium containing an ionic phosphorus ligand, separated from rhodium catalyst components, was also developed. This process is described in U.S. Patent Application No. 218911, relating to a method for separating a catalyst and an aldehyde product, of July 14, 1988, to AG Abatjoglou, K. R. Peterson and DR Bryant.

For both the non-aqueous (non-polar) and aqueous (polar) hydroformylation processes, various streams of rhodium-containing liquid can be generated that are not recycled to the hydroformylation reactor. For example, in hydroformylation processes involving phase separation to collect the aldehyde product, it is not unusual that some of the rhodium is removed with the aldehyde product. Due to the high costs and limited supply of rhodium, it is necessary to reduce rhodium losses in the aldehyde product to the lowest possible level. In fact, for economic reasons, the rhodium concentration in the aldehyde product should be reduced to less than about 50 parts per billion (ppb) and more preferably to less than 20 ppb.

Thus, the object of the invention was to develop a method for recovering rhodium from liquids containing rhodium at very low concentrations, and especially to recover rhodium from a liquid such as an aldehyde product from hydroformylation.

The method for recovering rhodium from a rhodium-containing liquid in an amount below 20 ppm, according to the invention is that (i) contacting the rhodium-containing liquid with an ion exchange resin in which said resin has the first organophosphorus ligand associated with it, causing rhodium to be removed from the liquid and binding to the resin and (ii) then contacting said resin with a liquid containing a second organophosphorus ligand in an amount sufficient to remove rhodium from the resin, preferably the same compound is used as the first and second ligand.

The attached figure is a schematic diagram of the process according to the invention for the recovery of transition metal e.g. rhodium.

The invention can be widely used to recover rhodium from any polar or non-polar solution that may be derived from any rhodium process, for example from the hydrogenation of unsaturated compounds, such as copolymers of conjugated dienes with copolymerizable monomers, as described, for example in U.S. Patent Nos. 4,465,315 and 450,396, in the carbonylation of methanol to acetic acid, in olefinization of olefins, in the hydrocyanation of butadiene to adipic acid dinitrile, in decarbonylation of aldehydes, in olefin hydrosilylation, and in other known processes. Thus, the present invention is widely used to recover rhodium from both polar and nonpolar solutions. However, for convenience, the invention will be further described with particular reference to the recovery of rhodium from liquid solutions

166 018 olefin derived from hydroformylation, in particular from water-immiscible (non-polar) organic solutions derived from this process. As noted above, many known hydroformylation processes potentially face the problem of rhodium loss, for example when receiving an aldehyde product. Those skilled in the art, however, envisage the widespread use of the present invention to recover other transition metals from other liquid streams generated by or undergoing processes other than hydroformylation, in the light of the following description and specific examples.

The invention is illustrated in the attached drawing with reference to the only figure. As shown, stream 1 of a polar or non-polar liquid containing rhodium in solution flows through a column 10 containing an ion exchange resin having an organophosphorus ligand attached to it, hereinafter simply referred to as "ion ligand functionalized resin". As the solution flows through column 10 in contact with the ligand functionalized ionic resin, rhodium is removed from the solution. The liquid with reduced rhodium concentration is withdrawn via line 2. Without being bound to any particular explanation, rhodium is believed to form a coordination complex with the ion-bound ligand with the resin.

Any polar or non-polar stream containing rhodium can be treated by the method of the invention and the invention is not limited to any particular liquid source. Both aqueous liquids as well as other polar liquids such as methanol and polar and non-polar organic liquids such as aldehydes, alkanes, ethers, ketones, esters, amides and aromatic compounds containing rhodium in solution can be treated with rhodium recovery treatment according to the invention. Thus, for example, the invention is useful for recovering the rhodium present in the aldehyde product of any hydroformylation process or to this extent from any other liquid rhodium-containing byproduct stream that may be bound or may be derived from a hydroformylation process. The invention is particularly useful for the recovery of rhodium from the aldehyde product of the hydroformylation process described in U.S. Patent Application No. 218911 regarding the method of separating the catalyst from the aldehyde product of July 14, 1988 to AG Abatjoglou, RR Peterson and DR Bryant. According to expert knowledge, the treatment of some rhodium-containing liquids may be better carried out using certain resins due to the resin composition associated with physical constraints (e.g. limited solubility, limited swelling and the like).

Of course, it should be borne in mind that the starting material for the method of the invention, i.e. the rhodium-containing liquid, can also be a liquid residue from any other process intended to extract rhodium from the liquid. For example, U.S. Patent Application No. 231,508 entitled "Recovery of catalytic metal from non-polar organic solutions, filed August 12, 1988 on behalf of

DJ Miller and DR Bryant concerns the recovery of rhodium from a non-polar liquid e.g. an aldehyde product with an aqueous solution of an ionic phosphine ligand. The present invention can be used to recover trace amounts of rhodium left in the aldehyde starting material from such a process.

The rhodium concentration in the liquid to be treated according to the invention is not a critical parameter. However, rhodium concentration has an effect on resin load and other related parameters. Although the invention can be used to remove rhodium from liquids having up to about 400 parts per million (ppm) rhodium, calculated as metallic rhodium or more, it is particularly useful for recovering smaller amounts of rhodium, e.g. less than about 20 ppm rhodium. More preferably, the rhodium-containing liquid treated by the process of the invention contains no more than about 2 ppm rhodium, most preferably less than about 1 ppm, and may contain as little as 20 parts per billion (ppb), i.e. just measurable amount. Thus, if a liquid containing dissolved rhodium, such as an aldehyde product stream from a hydroformylation process, contains more than about 20 ppm of rhodium, it may be more economical to extract the main portion of the rhodium from the liquid by contacting the liquid with a ligand functionalized ion exchange resin in accordance with the present invention.

166 018

An ion ligand functionalized resin suitable for recovering rhodium from a polar or non-polar liquid solution as the starting material, can simply be prepared by contacting the ambient anion exchange resin with the acid derivative or salt of the acid derivative of the organophosphorus ligand at ambient conditions or by contacting the cation exchange resin with the basic derivative or base salt organophosphorus ligand derivative. For convenience, such derivatives will simply be referred to as ionic organophosphorus ligands. For example, the ionic organophosphorus ligand can generally be dissolved in water or a polar solvent such as methanol and the solution is then contacted with an ion exchange resin. Sufficient contact is achieved by simply mixing the aqueous resin suspension and ligand solution, or by passing the ligand solution in contact with the resin bed. It is necessary that the contacting is only sufficient for the ion exchange resin to take up the ionic ligand from the solution, i.e. to exchange the original portion of the ionic resin for ionic Ugand. Appropriate conditions for carrying out this contact can be determined by routine experimentation. Such contact may only be sufficient to carry out the ion exchange or reaction between the active sites of the resin and the ionic organophosphorus ligand. The resin is then separated from the aqueous solution and washed accordingly and then dried before use as needed.

In the broad practice of the present invention, the ionic organophosphorus ligands that can be used to prepare the ligand functionalized ionic resin for use in the method of the invention can be any ionic organic phosphine or ionic organic phosphite that are capable of forming a coordination complex with rhodium in solution. Of course, the ionic portion of the ligand cannot be located on the organophosphorus ligand in such a way that the ligand linkage via ionic attachment to the resin interferes with the ligand's ability to form a coordination complex with a transition metal such as rhodium. The particular circumstances that can be encountered in any intended application may affect which of the ionic organophosphorus ligands is most appropriate. Ionic organophosphorus ligands suitable for use in the method of the invention are well known as well as methods for their preparation and are further described in U.S. Patent Application No. 231,508 regarding the recovery of catalytic metal from non-polar organic solutions, of August 12, 1988 in on behalf of DJ Miller and DR Bryant, the disclosure of which is given by reference. Ionic phosphite-organic ligands suitable for use in the method of the invention are described in U.S. Patent Application No. 228,507 regarding ionic phosphites and their use in homogeneous transition metal catalyzed processes of August 5, 1988, on behalf of AG Abatjoglou and DR Bryant, the disclosure of which is given as a reference.

Ionic organophosphorus ligands as well as methods for their preparation are well known and there is no need to describe them in detail. In general, many different ionic organophosphorus ligands described in the literature can be used for hydroformylation processes of rhodium-catalysed olefins. Some suitable ionic organic phosphines are compounds of formulas 5 and 6 in which R1, R2 and R 3in formula 5 and R4, R5, R6 and R7 in formula 6 independently of each other represent a hydrocarbon radical of 1-30 carbon atoms selected from the group consisting of alkyl, aryl, alkylaryl, arylalkyl and cycloalkyl radicals, Q in formula 6 denotes a divalent organic bridging group, and Y, Y and Y inside 5iY, Y, and iY in formula 6 are substituted on a hydrocarbon radical and each individually represents an ionic radical with a general neutral charge selected from the group consisting of -SO3M, where M is inorganic or organic cationic atoms or radicals, -PO3M, where M is inorganic or organic cationic atoms or radicals, -NR3X ', where R is a hydrocarbon radical having 1-30 carbon atoms selected from the group consisting of alkyl, aryl, alkylaryl radicals , arylalkyl and cycloalkyl, and X 'is inorganic or organic anionic atoms or radicals, -CO2M, where M is inorganic or organic cationic atoms or radicals, m, m m in formula 5 and m, m, m m in formula 6 are numbers that can be the same or different in the range 0-5. At least one m2 / m3 and at least one m2, m25,

166 018 7 m6 them7 cannot be zero, i.e. must be equal to 1 or greater than 1. Numbers to m7 indicate the number of substituents at each hydrocarbon radical.

2 3 4 5 6 7

The R, R and R hydrocarbon radicals in the formula 5 and R4 R3, RD and R 'in formula 6 preferably contain 1-18 carbon atoms. Hydrocarbon radicals having 1-12 carbon atoms are more preferred. Such hydrocarbon radicals include, for example, alkyl, aryl, alkylaryl, arylalkyl and cycloalkyl radicals. Examples of hydrocarbon radicals are e.g. methyl, ethyl, propyl, butyl, hexyl, cyclohexyl, phenyl and the like. Most preferably, at least one of R1, R2 and R3 in formula 5 and at least one of R4, R5, r6 and R7 in formula 6 is a phenyl radical. Such hydrocarbon radicals may contain one or more substituents, provided that they do not adversely affect the use of the ligand in the method of the invention. Suitable substituents, in addition to the necessary ionic substituents e.g. sulfonate, carboxylate and the like include straight and branched chain alkyl groups, preferably 1-4 carbon atoms, alkoxy groups, halogen atoms, hydroxy, cyano, nitro and amino groups and the like. More preferably, at least two and most preferably three of R1, R2 and R3 in formula 5 are phenyl groups and at least three, and most preferably four of R4, R5, R6 and R 'in formula 6 are phenyl groups.

The organic divalent bridging group represented by Q in the above formulas is a divalent radical with 1-30 carbon atoms selected from hydrocarbon radicals, oxygen-containing hydrocarbon radicals (i.e. hydrocarbon radicals interrupted by an oxygen atom), sulfur-containing hydrocarbon radicals (i.e. hydrocarbon radicals interrupted by a sulfur atom) and nitrogen-containing hydrocarbon radicals (i.e. hydrocarbon radicals interrupted by a nitrogen atom). Preferably such radicals contain 1-16, and more preferably 1-12 carbon atoms. Examples of divalent hydrocarbon radicals include alkylene radicals (e.g., methylene (CH2-), ethylene, propylene, isopropylene, butylene,

1,2-dimethylethylene, 1,1-dimethyl enough noethylene, neopentylene, 2-methylpropylene, hexylene, 2-ethylhexylene, dodecylene, eicosylene and the like), arylene radicals (e.g. phenylene, substituted phenylene, diphenylene, substituted diphenylene and the like ) as well as arylene-containing alkylene radicals (e.g. methylene phenylene (-CH2C6H4-), ethylene phenylene ethylene (-C2H4-C6H4C2H4), phenylene propylphenylene (-C6H4C / CH3 / 2C6H4-), methylene diphenylmethylene (-CH2H4C6H4C6); alkylidene radicals (e.g. ethylidene (-CH = CH-) and the like) and the like. Examples of oxygen-containing hydrocarbon radicals are alkyleneoxyalkylene radicals (e.g., ethylene oxymethylene (-C2H4OCH2), propylene oxymethylene (-C3H6OCH2-), ethyleneoxyethylene (-C2H4OC2H4-), 1,2-bis (ethyleneoxy) ethane (-C2H4C4H4OC2 -) and similar); aryloxyalkylene radicals (e.g., phenyleneoxymethylene (-C6H4OCH2-) and the like) and the like. Exemplary sulfur or thio-containing hydrocarbon radicals include alkylenethioalkylene radicals (e.g., ethylenethioethylene (-C2H4SC2H4-), 1,2-bis (ethylenethio) ethane (-C2H4SC2H4SC2H4-), propylenethiomethylene (-C3H6SCH2H6), and -C3H6S6-6 ); arylthioalkylene radicals (e.g., phenylene thiomethylene (-C3H6S-CH2) and the like. Examples of amino-containing hydrocarbon radicals include alkyleneaminoalkylene radicals (e.g. methyleneaminomethylethylene (-CH2N / CH3 / C2H4-), ethyleneaminomethylethylene (-C2H4N / CH3 / C2H4-) bis (ethyleneaminomethyl) ethane (-C2H4N / CH3 / C2H4N / CH3 / C2H4-), propyleneaminomethylpropylene (-C3H6H) and similar. Most preferably Q is a divalent hydrocarbon radical, especially a divalent alkylene radical having 2-8 carbon atoms.

Particularly suitable ionic organophosphine ligands are ionic triarylphosphines and, in particular, sulfonated and carboxylated triarylphosphines as described, for example, in U.S. Patent Nos. 4,248, 802, 4,399 312, 4,686,824, 4,762,550 and 4,731,486 and published European Patent Application No. 216250 ( published in April 1987). In this group, monosulfonated and trisulfonated triphenylphosphines and their salts, and monocarboxylated and tricarboxylated triphenylphosphines and their salts are preferred. Another suitable class of ionic organic phosphines are ionic bis-diaryl phosphine such

166 018 as bisdwiphenylphosphinoethane monosulfonates. Mixtures of the corresponding ionic phosphine ligands can also be used.

Such ionic organophosphine ligands capable of forming rhodium coordination complexes are encompassed by the above formulas and, as well as the methods for their preparation, are well known and need not be discussed in detail. See for example J. Chem. Soc. (1958), p. 276-288 and U.S. Patent Nos. 4,248,802, 4,399,312, 4,483,802, 4,633,021, 4,668 824.4 716 250 and 4,731,486, and European Patent Application No. 216315 published in April 1987, all references herein. For example, sulfonated ligands can be prepared by sulfonating the corresponding phosphines, e.g. triphenylphosphine, with fuming sulfuric acid (oleum) under controlled temperature conditions.

As suitable M ions with the opposite sign to the anionic ions of phosphines, hydrogen (i.e. proton), alkali metal and alkaline earth metal cations e.g. lithium, sodium, potassium, cesium, rubidium, mold, barium, magnesium and strontium, ammonium cation and quaternary ammonium cations. Suitable anionic atoms or radicals are, for example, sulfate, carbonate, phosphate, chloride, acetate, oxalate and the like. Of course, it should be borne in mind that the number of anionic and cationic parts in the ligand molecule also depends on the valence of the ions (ionic radical) and opposite ions (M and X ') of each individual ligand.

Suitable ionic organophosphine ligands for use in the method of the invention are selected from the group consisting of (i) polyphosphites of formula 1, in which each Ar group is identical or a different aryl radical, X is a m-valent hydrocarbon radical selected from the group consisting of alkylene, alkylene- oxy-alkylene, aryl, and aryl- (CH2) y- (Q) n-aryl, each y is individually 0 or 1, each Q is independently a divalent bridge group selected from -CR1R2-, -O-, -S-, -NR3-, -SiR4R5- and -CO-, in which R1 and R2 are independently of each other a radical selected from hydrogen, alkyl with 1-12 carbon, phenyl, tolyl and methoxybenzyl, R3, R4 and R5 independently of each other are -H or -CH3, each n is individually equal to 0 or 1, m is equal to 2-6, wherein the polyphosphites of formula 1 contain at least one ionic part selected from the same group of ionic radicals with a general neutral charge, as defined above in connection with ionic organic phosphines substituted in the aryl part of Ar or X; (ii) diorganic phosphites of formula 2 in which T is a monovalent hydrocarbon radical, Ar, Q, than are as defined above, wherein the diorganic phosphites of formula 2 contain at least one part selected from the same group of ionic radicals with a general neutral charge identified above in connection with ionic phosphines substituted in the aryl part of Ar or T, and (iii) open-end bisphosphite of formula 3, wherein D is a divalent bridge group selected from alkylene, alkylene-oxy-alkylene, arylene and aryl- (CH2) yQn- (CH2)s-aryl, Ar, Q, n, y and T have the meanings given above and each T can be the same or different, with bis-phosphites of formula 3 containing at least one ionic part selected from the same group of ionic radicals with a general neutral charge, identified above with ionic organic phosphines substituted in the aryl part of Ar, D or T.

Exemplary aryl radicals from the above-defined Ar, X, D and T groups in the above formulas include aryl moieties which may contain 6-18 carbon atoms such as phenylene, naphthylene, anthracylene and the like.

In addition, as noted above, while each given ionic phosphite in the above formulas must contain at least one ionic portion selected from the same group of ionic radicals with a general neutral charge as identified above, in connection with the ionic organic phosphines substituted in the aryl portion as defined above groups Ar, X, D and T should be understood, that each given phosphite may contain more than one such ionic part and the same may apply to any aryl part in each ionic phosphite, provided that the total number of such ionic parts in the given phosphite is not high enough to adversely affect the usefulness of the ionic phosphite ligand for binding to the resin and forming a coordination compound with rhodium. Thus, each ionic phosphite ligand generally contains 1-3 ionic parts. Preferably, only one such ionic part is

166 018 9 substituted in any given aryl portion in an ionic phosphite ligand when the ligand contains more than one such ionic portion.

In the above organic phosphite formulas, preferably m is from 2 to 4, and each y and n are equal to 0. However, when n is 1, Q is preferably -CR1R2 - a bridging group as defined above, and more preferably a methylene (-CH2-) or alkylidene (CHR2-), wherein R2 is an alkyl radical with 1-12 carbon atoms (e.g. methyl, ethyl, propyl, isopropyl, butyl, dodecyl etc.), especially methyl.

M-valent hydrocarbon radicals represented by Y in ionic polyphosphite ligands of the formula I are hydrocarbons having 2-30 carbon atoms selected from alkylene, alkylene-oxy-alkylene, aryl and aryl- (CHz) yXQ) n - CH2) y-aryl , Q, and n have the meaning given above. Preferably, the alkylene parts of said radicals contain 2-18 carbon atoms, and more preferably 2-12 carbon atoms, while the aryl parts of these radicals preferably contain 6-18 carbon atoms.

The divalent bridge group represented by D in the bis-phosphite ligands with open ends of the above formula 3 are divalent hydrocarbons containing

2-30 carbon atoms selected from the group consisting of alkylene, alkylene-oxy-alkylene, aryl and aryl- (CH2) y- (Q) n- (CH2) y-aryl, Q, than are as defined above. Preferably, the alkylene parts of said radicals contain 2-12 carbon atoms, while the aryl parts of these radicals preferably contain 6-18 carbon atoms.

The hydrocarbon radicals represented by T in the above ionic phosphite ligand formulas include monovalent hydrocarbon radicals having 1-30 carbon atoms selected from the group consisting of straight or branched alkyl radicals, primary, secondary or tertiary alkyl radicals such as methyl, ethyl, n-propyl, isopropyl, amyl, tertiary-amyl, 2-ethylhexyl and the like, aryl radicals such as phenyl, naphthyl and the like, arylalkyl radicals such as benzyl, phenylethyl, triphenylmethylethane and the like, alkylaryl radicals such as tolyl, xylyl and the like; and cycloalkyl radicals such as cyclopentyl, cyclohexyl, cyclohexylethyl and the like.

Preferably, T is selected from the group consisting of alkyl and aryl radicals that contain from about 1 to 30 carbon atoms. Preferably, the alkyl radicals contain 1-18 carbon atoms, most preferably 1-10 carbon atoms, while the aryl, arylalkyl, alkylaryl and cycloalkyl radicals preferably contain 6-18 carbon atoms. Further, although each T group in the ionic phosphite molecule of formula 3 may differ from the other, they are preferably uniform.

Of course, it should further be borne in mind that, in addition to the ionic substitutions described above, the aryl parts from the definitions of Ar, X, D and T in the above formulas may also be substituted with any other substituent that does not adversely affect the method of the invention. Exemplary substituents include radicals having 1-18 carbon atoms such as alkyl, aryl, arylalkyl, alkylaryl and cycloalkyl, alkoxy radicals, silyl radicals such as -Si (R9) 3 and -Si (OR9) 3, amino radicals such as -N (R9) 2, acyl radicals such as -C (O) R9, acyloxy radicals such as -OC (O) R9, carbonyloxy radicals such as -COOR9, amide radicals such as -C (O) N (R9) 2 and -N (R9) COR9, sulfonyl radicals such as -S O2R9, sulfinyl radicals such as -SO (R9) 2, thionyl radicals such as -SR9, phosphonyl radicals such as -P (O) (R9) 2 as well as halogen, nitro, cyano, trifluoromethyl and hydroxyl groups and the like, wherein each R9 may be a monovalent hydrocarbon radical such as alkyl, aryl, alkylaryl, arylalkyl and cycloalkyl, provided that in amino substituents such as -N (R9) 2, each R9 taken together may also be a divalent bridge group, which forms a heterocyclic radical with a nitrogen atom in amide substituents such as C (O) N (R9) 2 and -N (R9) COR9, each R9 attached to N can also be water and in phosphonyl substituents such as -P (OXR9 ) 2, one R9 may be hydrogen. Of course, it should be borne in mind that each R9 group in a given substituent may be the same or different. Of course, such a hydrocarbon substituent may optionally be substituted with a substituent as already mentioned above, provided that it does not adversely affect the process of the invention.

166 018

Among the more preferred ionic phosphite ligands are ligands in which two Ar groups linked by a bridging group represented by - (CH2) y- (Q)n- (CH2) y- in the above formulas, are connected through their ortho positions in relation to oxygen atoms that connect Ar groups with a phosphorus atom. It is also preferred that any substituent, if present on Ar groups, be attached in the para and / or ortho position on the aryl relative to the oxygen atom that connects the substituted Ar group to its phosphorus atom.

Accordingly, a preferred class of ionic phosphite ligands suitable for use in the method of the invention are ligands of formulas 1A and 1B in which each Y1, γ2, Z2, Z3, Z and Z5 independently of each other, a group selected from a hydrogen atom, a monovalent hydrocarbon group of 1-18 carbon atoms (e.g. alkyl, aryl, alkylaryl, arylalkyl and cycloalkyl), hydroxy, alkoxy with 1-10 carbon atoms, and parts of sulfonic acid and carboxylic acid and their salts, X is a m-valuable bridging group containing 6-30 carbon atoms selected from the group consisting of radicals aryl and aryl-Qn-aryl, m is 2-4, the radical Q is individually -CR1R2-, in which each R1 and R2 independently of each other is a hydrogen atom and an alkyl group with 1-12 carbon atoms and in which n is 0 or 1, provided that in each phosphite ligand of formulas 1A and 1B at least one of any of the groups γ1, γ2, Z2, z3 , FROM4 and Z5 is an ionic portion selected from the group consisting of a portion of sulfonic acid and a carboxylic acid and salts thereof, or each named ligand contains at least one such ionic portion substituted at the aryl portion of X.

Another preferred class of ionic phosphite ligands used in the method of the invention are compounds of formulas 2A and 2B in which γ1, γ2, Z2, Z3, Z4, FROM5, Q and n have the meanings given above for formulas 1A and 1B, T is a monovalent hydrocarbon radical having 1-30 carbon atoms selected from the group from the part of sulfonic acid and carboxylic acid and their salts, or each ligand mentioned contains at least one such ionic part substituted on the aryl part T.

Still another preferred class of ionic phosphite ligands used in the method of the invention are compounds of formulas 3A and 3B, wherein D is a divalent bridging group containing 6-30 carbon atoms selected from the group consisting of aryl and aryl-Qn-aryl radicals, γ1, γ2, Z2, Z3, Z4, z5, Q, n and each T are the same as defined above for increases 1A, 1B, 2A and 2B, and in which each T can be the same or different, provided that in each phosphite ligand of formulas 3A and 3B at least one of any of the groups γ1, γ2, Z2, Z3, FROM4 and Z5 are an ionic portion selected from the group consisting of a portion of sulfonic acid and a carboxylic acid and salts thereof, or each of said ligands contains at least one such ionic portion substituted at the D or T aryl portion.

Many preferred versions of the ionic phosphite ligand patterns have been found, e.g. most preferred m is 2 and each y is 0, while Q is preferably -CH2- or -CHCH3-. Further, each ionic phosphite generally contains 1-3 such ionic parts as defined above.

The ionic parts of the preferred ionic phosphorus ligand formulas above are the sulfonic acid and carboxylic acid parts and their salts. Such salts contain many organic and inorganic cations needed to balance the charges of anions of acids substituted on the phosphite ligand. Some counter-ions disclosed in connection with ionic phosphines are suitable. Thus, suitable inorganic cations may be selected from the group consisting of alkali metals, alkaline earth metals and ammonium. Examples illustrating alkali metal cations are lithium (Li+), sodium (Na+), potassium (K +), cesium (Cs +) and rubidium (Rb +), while examples illustrating alkaline earth metal cations are calcium (Ca++), bar (Ba++), magnesium (Mg++) and strontium (Sr++). Suitable organic cations may be selected from the group consisting of quaternary ammonium cations such as those of formula [N (R21) (r22) (r2) (r2 ')] +, where each R21, RM, R23 and R24 are hydrogen or a radical having 1-30 carbon atoms selected from the group consisting of alkyl, aryl, alkylaryl, arylalkyl and cycloalkyl radicals, and wherein each two or three of said R2 groups1, R22, r43 and R24 may be joined together to form a single, two or poly-cyclic ring together with the nitrogen atom of said cation.

166 018

Examples of m-valuable hydrocarbon radicals represented by X in the above formulas include substituted and unsubstituted hydrocarbon radicals having 2-30 carbon atoms selected from the group consisting of alkylene, alkylene-oxyalkylene, phenylene, naphthylene, phenylene- (CH2) y- (Q) n - CH2) phenylene and naphthylene (CH2) y- (Q) y-CH2) y-naphthylene, in which Q, n and n are as defined above. More specific examples of m-valuable hydrocarbon radicals represented by X include, e.g., straight-chain and branched alkylene radicals such as - (CH2) x, where x is 2-18 (preferably 2-12), pentaerythritol which gives the m-valent radical hydrocarbon of formula C (CH2OH) 4-m, 1,2,6-xylene and the like; radicals -CH2CH2OCH2CH2S 1,4-phenylene, 2,3-phenylene, 1,3,5-phenylene, 1,3-phenylene, 1,4-naphthylene, 1,5-naphthylene, 1,8-naphthylene, 2,3-naphthylene , 1,1'-diphenyl-2,2'-diyl, 2,2'-diphenyl (^ 1,1'-diyl, 1,1 '-diphenyl-4-4' (diyl, 1,1'-dunaftyl) ^^ - diyl,

2,2-dinaphthyl-1,1'-diyl, phenylene-C ^ -phenylene, phenylene-S-phenylene, CH2-phenylene-CH2, phenyleneCH (CH3) -phenylene and the like.

Preferably the ionic poly-phosphite ligands of formula 1 include closed-ended phosphites in which X in the formulas of the above ionic phosphite is a divalent radical selected from the group consisting of phenylene, naphthylene, naphthylene (Q) n-naphthylene and phenylene (Q) naphthylene. wherein Q and n are as essentially and preferably given above. Of course, the aryl portions of such X radicals may contain substituent radicals as disclosed and discussed above.

The divalent radical D of the above ionic phosphite of formula 3 can be the same as any m-value radical described above as X, in which m = 2. Further preferred open-end bisphosphite ligands are compounds in which D is a divalent radical selected from of the group consisting of phenylene, naphthylene, naphthylene (Q) naphthylene and phenylene (Q) n-phenylene radicals in which Q and n are as essentially and preferably given above. Of course, the aryl portions of such D radicals may contain substituent radicals as disclosed and discussed above.

Among the more preferred bis-phosphite ligands of formula I and bis-phosphite ligands with open ends of formula 3 are compounds in which the naphthylene radical represented by X or D is selected from the group consisting of, 1,2-naphthylene, 2,3- naphthylene, especially 1,8-naphthylene and compounds, in which two phenylene radicals or two naphthylene radicals X or D connected by a bridging group represented by - (Q) n- are attached through their ortho position relative to oxygen atoms, which connects two phenylene radicals or two naphthylene radicals to their phosphorus atom. It is also preferred that any substituent radical present on the phenylene or naphthylene radical is attached at the para and / or ortho position of the phenylene or naphthylene radical relative to the oxygen atom that connects the given substituted phenylene or naphthylene radical to its phosphorus atom.

The hydrocarbon radicals represented by T in the above ionic phosphite ligand formulas include monovalent hydrocarbon radicals having 1-30 carbon atoms selected from the group consisting of alkyl radicals including straight or branched primary, secondary or tertiary alkyl radicals such as methyl, ethyl, n-propyl , isopropyl, amyl, tert-amyl, t-amyl, 2-ethylhexyl, 1-decyl and the like; aryl radicals such as phenyl, naphthyl and the like; arylalkyl radicals such as benzyl, phenylethyl, triphenylmethyleneethane and the like; alkylaryl radicals such as tolyl, xylyl and the like and alicyclic radicals such as cyclopentyl, cyclohexylethyl and the like.

Preferably, T is selected from the group consisting of alkyl and aryl radicals that contain about 1-30 carbon atoms. Preferably, the alkyl radicals contain 1-18 carbon atoms, most preferably 1-10 carbon atoms, while aryl, arylalkyl, alkylaryl and alicyclic radicals preferably contain 6-18 carbon atoms. Further, although each T group in the ionic phosphite molecule of formula 3 may differ from the other, they are preferably identical.

Further preferred aryl radicals represented by T include groups of formula 4 in which X1, X2 and Z1 are independently of each other a radical such as Y1, Y2, FROM5 and Z3 with the meaning given above. More preferably X1 and X2 the same or different are a hydrogen atom or a radical

166 018 with steric hindrance from isopropanol or greater, and Z1 means the ionic part as defined above.

In addition, as noted, the above-described radicals represented by Ar,, X, D and T in the above formulas can be further substituted with any substituent that does not adversely affect the desired effect of the invention. Examples of substituents are, for example, monovalent hydrocarbon radicals having 1-18 carbon atoms such as alkyl, aryl, alkylaryl, arylalkyl, cycloalkyl and other radicals as defined above. In addition, the various non-hydrocarbon substituents that may be present include, e.g., halogen, preferably chlorine or fluorine, NO2, -CN, -CF3, -OH, -SifCHsja, -SifOCHsjs, -SifCsHyjs, -C (O) CHa, -C ( O) C2H5, 0C (O) CeH5, -C (O) 0CH3, -N (CH3) 2, -NH2, -NHCH3, -NHCH3, -NH (C2Ha), -conh2, -CON (CH3)2, -S (O) 2C2Hs, -OCH3, -OC2H5, -OCeHs, -C (O) CeHa, -O (t-C4H9), -SC2H5, -OCH2CH2OCH3, - (OCH2CH2) 2OCH3, - (OCH2CH2) sOCH3j -SCH3, -S (O) CH3, -SCeHa, P (O) (CeHs) 2, -P (O) (CH3) 2, -P (O) (C2Hs) 2, -P (O) (CsH7) 2, -P (O) (C4H9) 2, -P (OXCsH13) 2, -P (O) CH3 (CsHs), -P (O) (H) (C6H5), -NHC (O) CHe and the like. In addition, each Ar, X, D and T group may contain one or more substituent groups, which may also be the same or different in each given ligand molecule. Preferably, the substituent radicals include alkyl and alkoxy radicals having 1-18 carbon atoms, and more preferably 1-10 carbon atoms, especially t-butyl and methoxy.

Substitution (excluding the bridging group - (CH2) y- (Q) n- (CH2) y- when present) in the ortho positions of the aryl groups Ar, X, D and T of the above formulas in relation to the oxygen atom that binds each aryl group to the phosphorus ion phosphite ligand atom, may affect ligand stability due to steric obstacles around the phosphorus atom of the ionic phosphite ligand caused by substitution at such ortho positions. For example, too many steric obstacles can affect the ability of the ionic phosphite ligand to bind the transition metal of Group VIII of the Periodic Table of Elements (e.g., rhodium), while an insufficient steric obstacle can cause too much ionic phosphite binding. The obstacle may also affect ligand binding to the ion exchange resin.

One class of preferred ligands of the above formulas are ligands of the formulas 1A, 2A and 3A, wherein both γ1 and y2 are radicals having a steric hindrance of at least isopropyl or greater, such as branched chain alkyl radicals having three to five carbon atoms, especially t-butyl, while more preferably Z2 and Z3 both are alkoxy, especially methoxy.

The ionic phosphite ligands used in the process of the invention can easily be prepared by a series of traditional condensation reactions of phosphorus halide with alcohol. Such types of condensation reactions and the manner in which they can be carried out are known, for example, from U.S. Patent Nos. 4,599,206 and 4,717,775 to non-ionic organophosphite dihydrogen ligands of formula 2 above; No. 4,748,261 regarding non-ionic bisphosphite ligands of formula 3 above and No. 4668651 regarding non-ionic polyphosphite ligands of formula 1 above. For example, the process for producing iodine phosphite dihydrogen ligands of formula 2 above may involve reacting the corresponding organic diol - a dihydroxy compound - with phosphorus trichloride to form an intermediate organic phosphorochloride compound, which in turn is reacted with the corresponding organic monool - the monohydroxy compound - to form the desired ionic ligand fosforynodwuorganiczny. Such ligands can be prepared in reverse order, for example, from the corresponding organic phosphorochloride compound and the corresponding diol compound.

Similarly, ionic bis-phosphite ligands with open ends of the above formula 3 can be prepared by (a) reacting the corresponding organic dihydroxysyl compound with three phosphorous phosphorus to form the corresponding organic phosphorous chlorine intermediate,

166 018 (b) the reaction of said derivative with an organic diol (corresponding to D in the above formula 3) to form the corresponding intermediate hydroxy substituted diorganic phosphite, (c) the reaction of said intermediate hydroxy substituted diorganic phosphite with phosphorus trichloride to form the corresponding intermediate organic phosphorodichloride and (d) reacting said dichloride compound with two moles of the corresponding organic mono-ol (or one mole of each of the two different mono-ols) reaching the corresponding desired open-end bisphosphite ligand. Such condensation reactions can also be carried out optionally as one-pot synthesis. Further ionic poly phosphite ligands of the above formula 1 can be prepared by the same type of condensation reaction of phosphorus halide with alcohol as noted above in the preparation of ionic ligands of the above formula 3, for example, using e.g. diol in the above step (b) corresponding to X in formula 1 and the reaction of the dichloride derivative of step (d) with the corresponding diol instead of two moles of mono-ol to produce the desired ionic polyphosphite ligand. Alternatively, such ionic poly-phosphites can be prepared in a one-pot synthesis, e.g. by reacting the corresponding organic phosphorochlorine intermediate from step (a) or a mixture of various corresponding chlorine intermediates with a polyol corresponding to X, where the molar amount of chlorine derivative used is equal to hydroxyl groups in the polyol used. For example, two molar equivalents of the same phosphorochlorine intermediate from step (a) (or one molar equivalent of each of two different such intermediates) can be reacted with one molar equivalent of diol corresponding to X to form a closed-end bisphosphite-type ligand.

In addition, because the ionic phosphite ligands used in the method of the invention must contain at least one ionic part selected from the same group of ionic radicals with a general neutral charge, as defined above in connection with the ionic organic phosphines substituted on the aryl radical, it is preferred that that at least one of the organic reagents or intermediates of the halide-alcohol condensation reaction used to form such ionic ligands contains at least one such ionic portion substituted at the aryl portion, although it would be possible to protect such ionic parts after formation of the nonionic phosphite ligand by conventional methods, e.g. known sulfonation techniques, carboxylation techniques and the like. Salts of sulfonated acids, carboxylic acids and similar simple hydroxyl compounds such as phenolic and naphtholic mono-ols and diols and / or methods for their preparation are known.

In addition, despite the fact that such phosphorus-halide condensation reactions can be carried out in the presence of a solvent e.g. toluene and an HCl acceptor e.g. amines, preferably double salts of ionic monool reagents and triple salts of ionic diol reagents in the presence of a bipolar nonprotonic solvent such as N-methylpyrrolidone (NMP), dimethyl sulfoxide (DMSO), dimethylformamide (DMF), sulfolane and the like ionic reagents are condensed with the desired phosphorus halide containing parts e.g. PCI3, a phosphorochloride or phosphorodichloro compound to avoid the need for any HCl acceptor. Such double or triple salts and / or methods for their preparation are well known.

For example, one starting material could be a double salt of phenol para-sulfonic acid or a double salt of hydroxy para-benzoic acid, e.g. depending on whether the sulfonate or carboxylate form of the ionic phosphite is desired. Other suitable starting materials will be materials obvious to those skilled in the art and these materials may be substituted or unsubstituted elsewhere on the benzene ring. Double salt (e.g. phenolate sulfonate or phenolate carboxylate) can be prepared by adding the appropriate basic material, such as sodium hydroxide or potassium hydroxide, to the corresponding starting mono-ol. The formation of a double salt can conveniently be carried out

166 018 knit in a solution of the starting material and bases in a non-protic solvent and water. Water may be provided in the form of an aqueous base solution.

Water is then removed from the double salt solution, e.g. by azeotropic distillation with toluene. For example, a sufficient amount of toluene is thoroughly mixed with a solution containing the double salt to form a toluene / water azeotrope with a boiling point of about 85 ° C at atmospheric pressure. The azeotropic composition contains about 80% toluene and 20% water. Under these conditions, the non-protic solvent and the double salt have low volatility and are therefore easily separated from the toluene / water azeotrope.

Since the double salt is essentially insoluble in a non-protic solvent, the suspension of the double salt in the non-protic solvent formed as an azeotropic composition is removed. This suspension is then reacted with the desired intermediate phosphor halide, i.e. an organic phosphorochloride or phosphorodichloride compound. The anion of the double salt hydroxyl group selectively reacts with the phosphorus portion of the intermediate phosphide halide to give the desired ionic phosphite, which easily dissolves in a non-protic solvent. The cation / halide side salt is not soluble in the non-protic solvent and can be removed by filtration or by any technique for separating solid and liquid, e.g. by centrifugation.

The ionic phosphite composition can be easily separated from the non-protic solvent, e.g., by removing the non-protic solvent from the mixture under vacuum distillation. The ionic phosphite recovered in this way can be further purified by recrystallization in a manner known to those skilled in the art, and the purified ionic phosphite used in the process of the invention. Other ionic phosphites that can be used in the process of the invention can easily be prepared by similar techniques.

The ionic phosphine and ionic phosphite ligands used in the present invention are preferably used in the form of the free acid, i.e. when M is a cationic hydrogen atom (proton). Salts of ionic organic phosphines and ionic organic phosphites can be converted to the free acid ligand by known methods, for example using ion exchange as illustrated in the following examples.

Anion exchange and cation exchange resins suitable for the preparation of ionic functionalized ligand resins used in the process of the invention include a wide variety of insoluble organic polymers obtained by addition polymerization or polycondensation of the corresponding monomers previously used to prepare ion exchange resins. The organic polymers obtained are then modified by known methods to obtain the desired ion exchange capacity. Insolubility of suitable polymers is achieved by chemical cross-linking, by radiation or by thermosetting. Examples of suitable polymers for ion exchange resins are polystyrene, polyethylene, polyvinyl chloride, polyvinyl acetate, polyethyleneimine, and other polyalkylene imines, polyvinylpyridine, polyacrylonitrile, polyacrylates, Saran®, Teflon® and the like. Suitable crosslinkers to ensure their insolubility, especially polyolefins, are divinylbenzene, butadiene, diallyl maleate, diallyl phthalate, glycol dimethacrylate and other di- or tri-olefins.

Condensation polymers suitable for the preparation of ion exchange resins include phenol-formaldehyde resins, urea-formaldehyde resins, alkyl resins (reaction products of polyhydric alcohols and polybasic acids, polyesters such as Dacron®, and polyamides. Polyamines, polyethers such as phenylene oxide, polyoxide are also suitable styrene or polypropylene oxide, polysulfides such as phenyl polysulfide and polysulfones such as phenyl polysulfone. Copolymers are also suitable. Celluloses may also be used, although they are not normally considered as resin. These resins are modified by known methods to give resins the ability to ion exchange.

A particularly useful resin is the copolymer of styrene and divinylbenzene available in the brake. Such resins are characterized by the fact that they have long polystyrene chains joined together by divinylbenzene crosslinking, forming a three-dimensional insoluble polyomeric phase. However, as indicated above, in the broad practice of the invention, it is not important to use specific resins.

Anion exchange and cation exchange resins are available in both gel and macrogrid form. Although both these gel and macro-mesh forms can be used in the method of the invention, it is preferred to use macro-mesh resin. Generally, macrogrid resins have a substantially uniform macroporous structure with average pore diameters of about 50 / m. Gel resins will generally only be used when the liquid containing the removed metal, according to the invention, will swell the gel resin, which increases the available surface area of the resin.

Anion exchange resins are characterized as strongly basic or weakly basic anion exchange resins. Both strongly basic and weakly basic anion exchange resins can be used to prepare the ligand functionalized ionic resins used in the process of the invention.

Strongly basic anion exchange resins consist of polymers having movable monovalent anions such as hydroxides (OH-) and the like, associated, for example, with covalently attached functional groups such as quaternary ammonium, phosphonium or arsonium, a tertiary sulfonium functional group. Said functional groups are known as active sites and are distributed over the entire surface of the resin molecule. Strongly basic ion exchange resins are able to undergo ion exchange regardless of the pH of the environment due to their intrinsic ionic character. In the practice of the present invention macrogrid highly basic hydroxide ion exchange resins are particularly preferred. Such resins are commercially available or can be easily prepared from resins sold by Rohm and Haas Company under registered names Amberlyst® e.g. Amberlyst® A-26 and Amberlyst® A-27. Other suitable strongly basic resins are commercially available from other companies such as Dow Chemical Company under the registered name Dowex® 21 K, 11 and MWA-1.

The resin's basic substance in weakly basic ion exchange resins contains chemically attached basic, nonionic functional groups. Functional groups include primary, secondary and tertiary amine groups. Among them, tertiary amine groups are preferred. They can be aliphatic, aromatic, heterocyclic or cycloalkane amino groups. They can also be diamino, triamino or alkanolamine groups. For example, the amines may include groups such as alpha, alpha'-dipyridyl, guanidine and dicyandiamine. Other nitrogen-containing basic nonionic functional groups include nitrile, cyanate, isocyanate, thiocyanate, isothiocyanate and isocyanate groups. Pyridine groups may also be used. The present invention is however not limited to any class of weak base anion exchange resins.

Aminated, styrene-divinylbenzene copolymers crosslinked to a variable degree in the molar range of 1-40% of the reacted monomer, which are also available from Rohm and Haas Company under the registered trademark Amberlyst®, are particularly useful as weak base anion exchange resins. For example, these resins can be prepared by methods described in U.S. Patent No. 2591574 (McBurney, 1952) assigned to Rohm and Haas Company. Amberlyst® A-21 cross-linked using divinylbenzene is a resin useful for the process of the invention because it is porous and has an insoluble pearl structure. Amberlyst® A-21 beads contain nitrogen in an amount of between about 4.2 and about 4.8 milliequivalents / g of resin in the form of tertiary NN-dimethylbenzylamine.

Weakly basic anion exchange resins are characterized in that they basically do not have ion exchange properties at pH values above about 7, because above this pH they do not contain any ionic group. As indicated above, they consist of polymers containing primary, secondary or tertiary amines and the like. A further definition of strongly and weakly basic ion exchange resins, together with a discussion of their preparation and properties, is described in "F. Helfferich - Ion Exchange ", McGraw Hill Book Co., New York, NY, 1962 pp. 16, 4758,78,138-40 and in" Dowex-Ion Exchange ", the Dow Chemical Co., Midland, Michigan, 1958.

166 018

Strongly acid cation exchange resins can also be used to prepare ligand functionalized ionic resins for use in the process of the invention. Strongly acid cation exchange resins consist of polymers having ion exchange active sulfonic acid sites. Suitable strongly acidic cation exchange resins can be obtained from Rohn and Haas Company under the registered name Amberlyst® 15. The use of other suitable cation exchange resins will be apparent to those skilled in the art.

As noted above, the resins used should be insoluble in the rhodium-containing liquid. In the broad practice of the present invention, "insoluble means insoluble at temperatures below the decomposition temperature of the resin in the rhodium-containing liquid, e.g. polar and non-polar solvents such as water, alcohols, ketones, aldehydes, ethers, esters, acetic acid, in saturated, unsaturated or cyclic aliphatic and aromatic hydrocarbons and in the same hydrocarbons having substituents consisting of or containing oxygen, sulfur, nitrogen or halides.

The number of active ion exchange sites per unit mass or unit volume of ion exchange resin suitable for preparing the ligand functionalized ion resin used in the process of the invention can vary widely and is not critical in itself. The number of active sites available on a given resin is quantified as "weight capacity and expressed in milliequivalents per gram (mEq / g). Generally, suitable resins have more than 0.5 mEq / g, preferably more than about 1.0 mEq / g. For best results, essentially all active ion exchange sites should be functionalized (i.e. substituted) with an ion ligand.

When a strongly basic anion exchange resin is contacted with an ionic organophosphorus ligand (e.g. acid or salt of an acid derivative of the organophosphorus ligand), the actual exchange of the anionic part of the ligand for the anionic part associated with the active site of the resin, e.g. a hydroxyl radical, occurs. In the case of a weakly basic anion exchange resin, there is no exchange of anionic parts and rather an acid reaction occurs - the principle between the acid form of the ionic ligand and the active site, e.g. tertiary amine on a weakly basic anion exchange resin. In both cases, the resulting resin may be considered a "salt" with the anionic portion of the ligand attached to the resin by ionic forces rather than covalent or coordinating forces. When contacting a strongly acid cation exchange resin with a basic derivative of an organophosphorus ligand, such as dialkylamino functionalized triphenylphosphine, an acid-base reaction occurs between the acid part of the resin and the basic dialkylamino group of the ligand. Proceedings for the preparation of a dialkylamino-substituted organophosphorus ligand derivative are known, see Zhmurova et al. Zh, Obshch. Khim., 36 (7), pp. 1248-54 (1966). The resulting resin can also be considered as a salt with a cationic ligand portion (quaternary ammonium group) attached to the anionic portion (e.g. sulfonate) on the resin by means of ionic forces. It should be understood that all of the above mechanisms and the like can be used in the preparation of the ligand functionalized ionic resin used in the method of the invention.

It should be noted that a commercially available grade of ion-exchange resin in beads, such as Amberlyst® resins, may be available in the form of a halide, e.g. chloride, and may contain halide impurities, e.g. chloride impurities, which are known to poison (harm the rhodium) hydroformylation catalyst complex . It is therefore advantageous for hydroformylation processes where the recovered rhodium is recycled to the hydroformylation process so that the ion exchange resin used here is at least substantially free of halide impurities and more preferably substantially and completely free of such halide impurities. Removal of such halide impurities as well as any other undesirable impurities from the ion exchange resins prior to their use can easily be carried out by conventional ion exchange and washing techniques.

As noted above, the recovery of rhodium from polar and non-polar liquid solutions by the method of the invention can be surprisingly accomplished by simply contacting the liquid with a ligand functionalized ionic resin. The amount of ligand functionalized ion exchange resin relative to the rhodium-containing liquid will depend on the amount of rhodium in the liquid and the binding strength of the ionically bound ligand to the dissolved rhodium. It is sufficient that the amount of ligand functionalized ion exchange resin is only sufficient to reduce the rhodium concentration to the desired value. Calculated as a standard liquid with a rhodium concentration of about 10 ppm, the volume of ligand functionalized ionic resin about 10 ml of resin per liter of treated liquid should be sufficient to remove rhodium from the liquid. Although rhodium can be removed from the liquid even in the current large amount of free ligand as shown in the following examples, it is preferred that the amount of free ligand in the treated liquid is below about 10 molar equivalents per gram of rhodium atom, more preferably below about 5 equivalents molar ligand per gram rhodium atom, and most preferably below about 2 molar equivalents of ligand per gram rhodium atom. The best results are obtained when only trace amounts of free ligand, i.e. less than 1 mole of ligand per gram of rhodium are present in the liquid. Methods for reducing ligand concentration in liquids are known.

It is sufficient that the contact time between the resin and the rhodium-containing liquid is only sufficient to remove some of the rhodium from the liquid. The rhodium-containing liquid is contacted with the ligand functionalized ionic resin both in a continuous (semi-continuous) and batch fashion. When the process is carried out in a batch manner, contacting preferably involves mixing the mixture of rhodium-containing liquid and resin for about 0.01 to 10 hours, typically 0.1 to 5 hours, followed by separation by any known technique, e.g. settling, centrifugation, filtration and similar. Preferably, the invention is carried out continuously. The rhodium-containing liquid flows through one or more resin deposits, e.g. through fixed, moving or fluidized beds, at a liquid flow rate in the range of from about 0.1 to 100 volumes equal to the volume of the bed per hour, and usually from 1 to 20 bed volumes per hour. The invention can be carried out in any conventional ion exchange device and no special devices are required. Of course, for best results, proper contact between the resin and liquid is important. According to experts in this technology, the resin bed is used to remove rhodium from the liquid until the level (concentration) of rhodium in the treated liquid leaving the resin bed rises to above the desired value.

The rhodium can then be removed from the loaded resin by contacting or eluting the resin with a polar or non-polar liquid containing a solubilized organophosphorus ligand. Again, a polar or non-polar liquid can be used, subject only to the composition restrictions referred to above. In a similar manner, the many different ionic and non-ionic organophosphorus ligands referred to above can be used to elute rhodium from loaded resin.

Examples of non-ionic organophosphorus ligands that can be used to elute loaded resins include, e.g. trialkylphosphines and phosphites, dwualkiloarylofosfiny and phosphites, alkyldiarylphosphine and phosphites, trójaryloalkilofosfiny and phosphites, dwucykloalkiloarylofosfiny and phosphites, cykloalkilodwuarylofosfiny and phosphites, trójcykloalkilofosfiny and phosphites, triarylphosphines and phosphites, alkyl and / or aryl bisphosphines and bisfosfino mono-oxides, diorganophosphites, bisphosphites organic and polyphosphites and the like. Optionally, mixtures of such ligands as well as tertiary organophosphine organo ligands can be used.

Preferred organic phosphines include tertiary organic phosphines mentioned above, especially triphenylphosphine, propyldiphenylphosphine, n-butylodwufenylofosfiny, t-butylodwufenylofosfiny, t-butylodwufenylofosfiny, n-heksylodwufenylofosfiny, cykloheksylodwufenylofosfiny, dwucykloheksylofenylofosfiny, trójcykloheksylofosfiny, trójbenzylofosfiny and the like. Preferably organic phosphites include triaryl phosphites e.g. triphenylphosphites as well as diorganic phosphites such as those disclosed e.g. in U.S. Patent No. 4,717,775, organic bisphosphites such as those disclosed e.g. in U.S. Patent No. 4,749,261, and organic bis- and polyphosphites such as those disclosed in Of United States of America No. 4668 651.

The appropriate concentration of ligand for eluting rhodium from the loaded resin can be determined by routine testing. In general, the concentration of dissolved ligand in the eluent of about 0.1-2.0 moles per liter should be sufficient to remove rhodium from the bed, although higher ligand concentrations e.g. to

166 018 moles per liter and more can be used and may be more beneficial in some circumstances. For example, rhodium can be released from the resin by elution with a solution of a non-polar organic organophosphine organic ligand (e.g. triphenylphosphine or cyclohexydiphenylphosphine) or a non-polar organic non-ionic organophosphite solution to obtain a non-polar organic solvent soluble rhodium complex suitable for use in known non-aqueous hydroformylation processes.

A particularly unexpected feature of the present invention is that a solution of the same ionic organophosphorus ligand that is ionically bound to the resin can be used to remove (elute) the complexed rhodium from the previously used resin to remove rhodium from the starting material. For example, rhodium-loaded Amberlyst® A-27 anion exchange resin functionalized with 3- (diphenylphosphino) -benzenic acid e.g. triphenylphosphine monosulfonic acid (TPPMS) can be removed by eluting the resin with a 10% c solution of the sodium salt of this ligand, i.e. TPPMS-Na in methanol (about 0.27 moles per liter). Tests showed significant rhodium removal from the resin bed with just 4 bed volumes of ligand solution. That the same ligand for removing rhodium from the resin bed that was used to bind (complex) rhodium on the resin bed could be used was completely unexpected. This process also regenerates the resin for direct reuse while releasing the rhodium in a form that can be directly returned to the hydroformylation system if desired.

Operating pressures and temperatures to implement various aspects of the present invention are not critical in the narrow range and standard (ambient) conditions can be used. The pressure is limited only by the physical strength of the resin. The temperature and each individual pressure should also be that the liquids remain in the liquid state. The temperature is preferably kept at a relatively low level to minimize resin degradation, for example from about 0 ° C to about 120 ° C.

The following examples illustrate the invention without limiting its scope. It should be understood that all parts and proportions given herein and in the appended claims are by weight unless otherwise stated. Rhodium analyzes are also cited as the result of atomic absorption spectroscopy (AAS), unless otherwise indicated.

Example A. Preparation of triphenylphosphine.

Monosulfonic acid (TPPMS) bound to an anion exchange resin TPPMS (3 (diphenylphosphine / benzenesulfonic acid) was attached to the Amberlyst® A-27 ion exchange resin for use in recovering rhodium traces from a hydroformylation reaction product using a rhodium containing catalyst.

Amberlyst® A-27 was first converted from its chloride form to its bicarbonate form, and then to its hydroxyl form. Sodium bicarbonate (NaHCO3) and sodium hydroxide (caustic soda) solutions were prepared by adding 180 g each time to the appropriate compound to 1800 ml deionized water. 1800 ml of a 10% (w / v) sodium bicarbonate solution was passed through a column containing 90 ml of A-27 resin at a rate of 4 bed volumes per hour. The column was then washed with 450 ml deionized water. A 10% (w / v) caustic sodium solution (1800 ml) was then passed through the sodium bicarbonate treated resin, also at a rate of about 4 bed volumes per hour. The resin thus transformed in its hydroxyl form was washed with deionized water until the effluent had a neutral pH.

An aqueous solution of TPPMS (free acid form) was prepared by dissolving 35 g sodium triphenylphosphine monosulfonate (TPPMS-Na) in 665 g deionized water. To facilitate dissolution, this solution was warmed, then cooled and filtered. The filtered solution was passed through 100 g of Amberlite IRN-77 cation exchange resin to convert the salt (TPPMS-Na) to its free acid form (TPPMS). The resin bed was washed with a volume of deionized water equal to the volume of the bed. The effluent, combined with the humans, was collected to obtain 1000 ml of TPPMS acid solution. The concentration of this TPPMS acid solution was determined by titration with 0.05 n aqueous sodium hydroxide solution and was 0.099 n.

This TPPMS solution was eluted under ambient conditions through A-27 resin (hydroxyl form) loaded into a glass column 2 cm in diameter and 50 cm long. The concentration of eluent recovered from the column was determined by titration and was 0.075 n. To ensure that the resin was completely filled with TPPMS once treated with the resin, it was again contacted with the TPPMS solution. The concentration of this second eluent - TPPMS solution was also determined and after the second elution it was 0.075 n. The unchanged concentration of the solution indicated that the resin was completely filled. The resin with ionically attached TPPMS was then washed with isopropanol. This resin was then used to remove rhodium from various solutions and was identified in the following examples as resin A.

Example B. Preparation of ligand functionalized ion resin by direct ion exchange of TPPMS-Na with anion exchange resin.

100 ml of Amberlyst® A-27 resin (chloride form) was loaded into a glass column 2 cm in diameter and 50 ml high. The resin was washed with 2 liters of a 10% sodium bicarbonate solution to convert the resin to its bicarbonate form. The column was then washed with 200 ml deionized water.

An aqueous solution of TPPMS-Na at a concentration of about 6.16% by weight (determined by HPLC) was eluted under ambient conditions over the bicarbonate form of the resin A-27 on the column. 600 ml of solution was passed through the resin at a rate of about 4 bed volumes per hour. After the first elution, the TPPMS-Na concentration was reduced to about 4.04% by weight. The TPPMS loaded resin thus prepared was collected for next use and identified in the following examples as resin B.

Example C. Preparation of an ionic ligand.

Functionalization of DIPHOS-MS ion bound resin.

A method similar to that described in Example A was used to prepare a resin to which DIPHOS-MS (bis-diphenylphosphine ethane monosulfonic acid) was ionically attached. In a stainless steel column 0.95 cm in diameter and 50 cm in length, Amberlyst A-27 was modified to the hydroxyl form by the method of example A. Then, the aqueous solution of DIPHOS-MS was passed through the hydroxyl form of the resin in the column. Resin saturation with DIPHOS-MS was checked by titration. The resin thus prepared was identified in the following examples as resin C.

Example D. Preparation of an ionic ligand.

Functionalizing the resin with an ion-bound ligand.

The cation exchange resin (about 109 g) was contacted with 27.3 g of triphenylphosphine monosulfonic acid sodium salt (TPPMS-Na) supplied as a 5% aqueous solution to convert the salt to the free acid form (TPPMS). The column was loaded with about 20 g of Amberlyst A-27 resin (chloride form), washed with 300 ml methanol and 300 ml distilled water. The TPPMS solution was then passed under ambient conditions through an A-27 resin column to prepare a ligand functionalized ion resin. Unbound ligand was removed from the resin by washing with isopropanol. This resin was identified in the following examples as resin D.

Example 1. Resins A and C were used to remove rhodium from the aldehyde product stream (tridecanol) withdrawn from the hydroformylation process described in Company Patent Application No. 218 911 of July 14, 1988 in which a rhodium-containing catalyst was used. The tridecanal aldehyde product also contained dodecene, water, a small amount of N-methylpyrrolidone and various hydroformylation by-products at very low concentrations. Table 1 summarizes below the results obtained at different tridecanal flow rates, expressed in bed volumes per hour. The rhodium concentration was expressed in parts per billion (ppb). The amount of resin indicated in Table 1 was loaded into a stainless steel column 0.95 cm in diameter and 50 cm in length, through which the tridecanal stream flowed. The percentage rhodium recovery on the resin from the solution was calculated based on the detection limit of the analytical equipment, thus

166 018 minimum values rather than actual values are presented. The actual recovery was probably much higher.

Table 1

Removing traces of family from tndzganal

<td rowspan="2">Example</td><td rowspan="2">Resin</td><td colspan="2">Quantity of resin Flow rate</td><td rowspan="2">House entrance concentration (PPb)</td><td rowspan="2">Rhodium output concentration (PPb)</td><td rowspan="2">% recovery (minimum)</td>

<td>(G)</td><td>aldehyde</td>

LL = actual mew detectable concentration, but was lower than the lower limit of sensitivity of AAS analyzes, i.e. lower than about 20 ppb.

Example II This example illustrates the use of functionalalizanzanz Ugandem ion resin to remove rhodium from a hydroformylation catalyst solution and then the use of a solution of another ligand to remove rhodium from a rhodium loaded resin bed. Resin B was washed with deionized water, then contacted with 500 mL 10% aqueous TPPMS-Na and washed again with water and methanol. A 15.34 g portion of functional anodized resin Ugandem resin was loaded into a stainless steel column 0.95 cm in diameter and 50 cm in length. A methanol solution containing 300 ppm rhodium (introduced as tetranium dodecarbonyl (Rh4 (CO) 12)) and 2 molar equivalents of TPPMS-Na was prepared. 100 g (126 ml) of this solution was recirculated through the resin bed B in the column at a rate of about 1 ml / minute. After recirculation, the rhodium concentration in the solution was reduced to 4.4 ppm indicating significant rhodium removal from the solution (about 98% recovery).

The treated resin column containing bound rhodium was washed with 100g of a 10 wt% solution of TPPMS-Na in methanol to remove the resin bound rhodium. The rhodium concentration in ethane was monitored by analyzing subsequent eluent fractions to determine rhodium removal from the resin with a TPPMS-Na solution. The results are summarized below in Table 2. The rhodium concentration in the collected washes (eluent) was also analyzed and determined to be 208 ppm.

Table 2 Resin rhodium recovery

<td>A sample</td><td>Bed volume</td><td>Rh, ppm</td>

To illustrate the reusability of the resin, 126 ml (100 g) of an additional rhodium-containing solution (prepared in the same manner as above and containing 300 ppm rhodium) was again passed through the same resin. At the end of rhodium loading, the print had a concentration of 10 ppm rhodium, showing significant rhodium removal from the solution by ligand functionalized resin. Then the rhodium-containing resin was washed again with an additional 100 g of a 10% aqueous methanolic solution TPPMS-Na. The accumulated washes (eluent) had a rhodium concentration of 280 ppm, showing significant rhodium removal from the resin bed with a ligand solution.

Examples III and IV. The following examples also illustrate the ability of the ligand functionalized resins prepared as described herein to remove rhodium from solutions that contain free triphenylphosphine ligand (TPP) by the method of the invention. After prior use as described in Example 1, a column containing resin C was used to remove rhodium from the depleted hydroformylation catalyst. The exhausted hydroformylation catalyst was obtained from the hydroformylation process of propylene to butyric aldehyde using a rhodium-TPP catalytic complex, before stripping the light fractions so that the catalyst solution received contained about 14% by weight of butyraldehyde, 18% by weight of triphenylphosphine (TPP), and the balance difference was fractions heavy hydroformylation process. As the catalyst solution eluted through the resin, samples were periodically taken to determine the rhodium concentration in the eluent. Table 3 below lists rhodium concentrations in subsequent eluent samples as a function of total volume of solution passing through the resin. The solution passed through the resin bed at a rate of 4 bed volumes per hour (Example III).

Table 3 also shows the results of rhodium recovery from fresh catalyst solution. The catalyst solution consisted of a 10% solution of triphenylphosphine in TexanolR (Eastman trademark for 2,2,4-trimethyl-1,3-pentanediol mono-isobutyrate) from 200 ppm rhodium (i.e. 5 g triphenylphosphine in 45 g Texanol with 0.0253 g rhodium carbonylacetylacetonate). Before passing the catalyst solution through the resin, the catalyst was activated under hydroformylation using equimolar parts of hydrogen, CO and propylene, a pressure of 413.7 kPa and a temperature of 100 ° C for 30 minutes. 16 g of resin A was placed on the column and fresh catalyst solution was passed through the bed at a rate of about 4 bed volumes per hour (example IV).

Table 3

Removal of rhodium from solutions containing ligands

<td>A sample</td><td>Total accumulated ml of sample passed</td><td>Rh in the eluent, ppm</td>

<td>Example III</td><td> 0</td><td> 621</td>

<td>Example IV</td><td> 0</td><td> 187</td>

These data illustrate the removal of rhodium from a catalyst solution containing a significant concentration of free ligand.

EXAMPLE. 0.935 liters of tridecanaldehyde solution originally containing 55 ppm of rhodium and approximately 0.2% TPPMS-Na ligand was continuously recirculated at a rate of 1.7 ml / min through a 25.4 cm long column filled with 22.4 cm diameter g of Amberlyst® A-27 ligand functionalized with TPPMS ligand (resin D). Recirculating liquid samples were periodically collected at the site just before entering the bed and analyzed for rhodium content (ppm), wt% TPPMS-Na ligand and wt% TPPMS-Na oxide. The results are shown in Table 4. After taking sample No. 7, the resin bed was replaced with a new resin. The final rhodium concentration was 0.094 ppm which corresponds to a total rhodium recovery of 99.8%.

Table 4

<td>Sample no</td><td>Rod (ppm)</td><td>TPPMS-Na (ppm)</td><td>TPPMS-Na oxide (ppm)</td>

166 018

Example VI. 200 ml of a solution containing 50 ppm of rhodium (supplied as rhodium carbonylacetylacetonate) and 1% by weight of triphenylphosphine (TPP) in Texanol® as solvent was recirculated through a 254 cm long column with a diameter of 1.27 cm containing the same ligand functionalized resin as used in Example V (resin D). After 17 hours of recirculation, the rhodium concentration in the solution decreased to 3 ppm, which corresponded to about 94% of rhodium recovery.

Example VII. This example illustrates the removal of rhodium from the loaded bed obtained in Example 6 using a triphenylphosphine (TPP) solution. After increasing the TPP ligand concentration of the recirculating Texanol® solution from Example 6 containing 3 ppm rhodium to about 10% by weight and after 26 hours of further circulation through the same rhodium loaded resin previously in Example 6, rhodium was eluted from the resin so that the rhodium concentration in the recirculating solution increased from 3 ppm to 10.2 ppm.

Example VIII. A 400 ml tridecanal solution was recirculated through a column of length

25.4 cm and a diameter of 1.27 cm loaded with 50 ml of ligand functionalized ionic resin (resin D). The recirculating liquid flow was set to approximately 1.5 ml / min. A continuous rhodium removal process was simulated, and for this purpose a N-methylpyrrolidone solution containing 175 ppm rhodium and 2 molar equivalents TPPMS-Na per gram rhodium was periodically poured into the recirculating aldehyde stream. The rhodium concentration in the stream entering and leaving the resin bed was periodically analyzed and the results are given in Table 5.

Table 5

<td>Day</td><td>Cumulative volume of rhodium solution introduced (ml)</td><td>Input rod (ppm)</td><td>Output rod (ppm)</td><td>Delete family per transition (%)</td><td>Total rhodium removal (%)</td>

Thus, some embodiments of the invention are described in more detail here, but it should be noted that various modifications are possible and that such modifications and changes that do not depart from the idea and scope of the invention described herein and form the part thereof.

PATTERN 2

16^018

PATTERN 2-A

MODEL 3

MODEL 3-A

χΐ ir '

X2

MODEL 4

<γ2'Π »2

R3-IY3lm3

MODEL 5 R\ ^ R6- (Y6lm6 ^ P —01 — px ly5) M5-R5 R ^7- (Y7) m7

Walkthrough level 6 with?

UP Department of Publications. Circulation of 90 copies Price PLN 1.00

Contents5

16 members in 10 offices

This recordMember, by publication datePriority claim

Priority claims3

Priority claims
Document	Office	Kind	Date
23622188	United States of America	A	1988-08-25
88236221	–	–	–
US19880236221	–	–	–

Members16

Family members, oldest first
Document	Office	Kind
EP0355837A2	European Patent Office (EPO)	A2	1990-02-28
AU4019889A	Australia	A	1990-03-01
KR900003388A	Republic of Korea	A	1990-03-26
CN1040826A	China	A	1990-03-28
BR8904259A	Brazil	A	1990-04-10
BR8904259A	Brazil	A	1990-04-10
EP0355837A3	European Patent Office (EPO)	A3	1990-05-16
JPH02160627A	Japan	A	1990-06-20
AU621396B2	Australia	B2	1992-03-12
US5114473A	United States of America	A	1992-05-19
MX166329B	Mexico	B	1992-12-29
CN1021657C	China	C	1993-07-21
RU2004499C1	Russian Federation	C1	1993-12-15
KR940004628B1	Republic of Korea	B1	1994-05-27
PL166018B1This record	Poland	B1	1995-03-31
JPH0829946B2	Japan	B2	1996-03-27

A method for recovery of transition metal and an ion-exchange resin

Abstract

Term

PATTERN 2

PATTERN 2-A

MODEL 3

MODEL 3-A

MODEL 4

Priority claims3

Members16

Numbers

Titles

Classification