CA1249303A - Ethylene oligomerization process - Google Patents

Abstract

ABSTRACT

ETHYLENE OLIGOMERIZATION PROCESS

Process for the conversion of ethylene to n-alpha-olefin oligomers by contact with a catalytic nickel complex dissolved in a polar diol-based solvent, containing between 0.5 and 4.0 per cent by weight of water, calculated on diol.

Description

3~3, ETXYLENE OLIGOMERIZATION PROCESS

This invention relates to an improvement in a process for the production of the linear alpha-olefin oligomers of ethylene. More particularly, this invention is directed to an improvement in a process for the conversion of ethylene to oligomers by contact with a catalytic nickel complex dissolved in a polar organic solvent.
Linear mono-olefins are compounds having established utility in a variety of applications. Linear mono-olefins having carbon numbers in the detergent range (e.g., about C8 to C20) are known to be particularly useful as intermediates in the production of various surfactants, including, for example, olefin sulphonates, and alcohol ethoxylates.
It is known in the prior art to prepare linear mono-olefins by oligomerizing ethylene at elevated temperature and pressure in a reaction solvent containing a catalytic nickel complex. Particularly useful as catalysts for this service are the complexes prepared as the reaction product of an olefinic nickel compound and a bidentate ligand. The olefinic nickel compound is suitably a reduced nickel compound or ~-allyl nickel compound. The reaction solvent in such a process is preferably a diol-based solvent, particularly an alipha-tic diol of 2 to 7 carbon atoms. Illustrative ethylene oligomeriza-tion processes employing a nickel complex catalyst and a diol-based solvent are described in: U.S. Patent No. 3,676,523, U.S. Patent No.
3,686,351 and U.S. Patent No. 3,737,475, all to R.F. Mason; U.S.
Patent No. 3,644564 to Van Zwet et al.; U.S. Patent No. 3,647,914 to Glockner et al.; U.S. Patent No. 3,647,915 to Bauer et al.; and U.S. Patent No. 3,825,615 to E.F. Lutz.
The present invention centers on the use for one such oligomeri-zation process of a diol-based solvent which comprises a limited quantity of water as a reaction promotor. Heretofore, processes of this type have been practiced under essentially ~6 3()3 anhydrous conditions. Although small amounts of water are commonly introduced into the reaction system together with the catalyst or the solvent make-up, efforts have consistently been made to hold the water content of the diol-based solvent to a level of no more than about 0,2 %w, calculated on diol.
Summary of the Invention The present invention provides an improvement in the conventi-onal process for the oligomerization of ethylene in the presence of a nickel chelate catalyst and in a diol solvent. According to the invention, the ethylene is contacted with the catalyst in a diol-solvent which comprises a specified minor amount of water. It has been found that the presence of water in a specified proportion relative to the solvent substantially enhances the rate of the ethylene oligomerization reaction. For example, reaction rates are increased by as much as 30 to 50 per cent, under preferred opera-tion.
Only small amounts of water, i.e., between about 0.5 and 4.0 per cent by weight (%w) relative to diol, are sufficient and necessary to achieve the desired rate enhancement. Neither lesser nor greater proportions of water are suitable for this purpose. In the amount specified, water is readily soluble in the diol-based solvent and is compatible with other components of the oligomeriza-tion system.
Thus, in the broad sense, the invention is an improvement in a process for the preparation of oligomers of ethylene which comprises reacting ethylene at elevated temperature and pressure in a diol-based solvent and in the presence of a catalyst which is a chelate of nickel with a bidentate ligand. The particular improvement is directed to the presence in the diol-based solvent of water in a quantity between about 0.5 and 4.0 per cent by weight (calculated on weight of diol).
In a more narrow aspect, the invention further relates to the discovery that such an improved process can be practiced to achieve the desired rate enhancement without a concomitant drawback in the distribution in the product of the various ethylene oligomers. In ~2~LC~ 3 ~ 3 ~ 32a3-2575 general practice the presence of the specified small amount of water is found to have a potential drawback from the s~andpoint of the desirability of the production of oli~omers having a certain carbon number distribution. It has, however, been found that in certain preferred embodiments the invention can be practiced, under quanea-titive limitations on catalyst nickel and ligand ccmponents, wiehout significant sacrifice in product oligomer distribution.
Description of the Preferred Embodiments The improved process of this invention is broadly applicable to any ethylene oligomerization process employing a diol reaction solvent and a nickel complex catalyst.
Oligomers are addition products which contain two or more of the monomer units (in this case, two or more ethylene unies?~ bue not as many units as the relatively high molecular weight addition products which are referred to as polymers. The present invention is particularly adapted for the production of linear mono-olefinic oligomers of ethylene containing from 2 to about 20 monomer units (i.e., from 4 to about 40 carbon atoms).
Catalyses suitable for use in the invention are complexes of nickel comprising an atom of nickel chelated with a bidentate chelating ligand. Such catalysts are typically prepared by reacting a suitable bidentate ligand either with an olefinic nickel compound such as bis (cyclooctadiene)nickel(0) or a ~-allyl nickel compound, or, more preferably, with a simple divalent nickel salt and reducing agent, e.g. boron hydride, in the presence of ethylene and in a suitable polar organic solvent. Preparation and use of catalysts of the former type are described in U.S. Patent No. 3,644,564 to Van Zwet et al., U.S. Patent No. 3,647,914 to Glockner et al., and U.S.
Patent No. 3,647,415 to Bauer et al. Preparation and use of catal-ysts of the latter type are described in U.S. PatentsNo. 3,676,523, No. 3,686,351 and No. 3,737,475, all to R.F. Mason, as well as U.S. Patent No. 3,825,615 to E.F. Lutz.

Preferred bidentate chelating ligands for such catalyst are known to include those having a tertiary organophosphorous moiety with a suitable functional group substituted on a carbon atom attached directly to, or separated by no more than two carbon atoms from, the phosphorous atom of the organophosphorous moiety.
Representative ligands of this type are the compounds ~L PCH2CH2COOM, R

CO~M

CHz R- p _ (OR~ , and R ~1 R _ p _ (CH2~y - C- - N- A2 wherein R (independantly in each occurrence) represents a monovalent organo (preferably aromatic) group suitably of up to 10 carbon atoms, R' (independantly in each occurrence) represents a monovalent hydrocarbyl group, X is carboxymethyl or carboxyethyl, A is hydrogen or an aromatic group of up to ten carbon atoms, M is hydrogen or an alkali metal (preferably sodium or potassium), and x and y (independently) are each either zero, one or two and the sum of x and y is two, with the proviso ehat when x is two, the R groups may, together with the phosphorous atom form a mono- or bycyclic heterocyclic phosphine having from 5 to 7 carbon atoms in each ring thereof. Particularly preferred complexes are those described in ~.S. Patent No. 3,676,523 in which the ligand is an o-dihydrocarbylphosphinobenzoic acid or an alkali metal salt thereof and most preferably o-diphenylphosphinobenzoic acid; in another preferred complex described in U.S. Patent No. 3,825J615, the ligand is dicyclohexylphosphinopropionic acid or an alkali metal salt thereof.
Although it is not desired to be bound by any particular theory, it has been suggested that the catalyst molecule undergoes chemical transfor~nation during the course of the oligomerization reaction, possibly involving coordination and/or bonding of ethylene to the nickel moiety. However, the bidentate chelating ligand apparently remains complexed and/or chemirally bonded to the nickel moiety during the course of the oligomerization reaction and this complex of the nickel and the chelating ligand is then the effective catalytic species of the oligomerization process. In any event, the bidentate ligand, such as the phosphorous-containing chelating ligand, is considered as essential component of the catalyst and, provided the nickel catalyst contains the required bidentate ligand, the nickel catalyst may be complexed with a variety of additional organic complexing ligands.
As is the case in prior art practice with such nickel chelate catalysts, the molar ratio of nickel to bidentate ligand used in catalyst preparation is preferably at least 1 19 i.e., the nickel is present in equimolar amount or in molar excess. In the p.eparation of catalyst complexes from a nickel salt, a ligand and a reducing agent, the molar ratio of nickel salt to ligand is suitably in the range from 1:1 to 5:1 with molar ratios of about 1.5:1 to 3:1 preferred and ratios of about 2:1 especially suitably. In these preparations, the reducing agent such as boron hydride is suitably present in equimolar amount or molar excess relative to the nickel salt. For economic reasons, it is preferred that the reducing agent/nickel ratio not exceed about 15:1. More preferably, this ratio is between about 1:1 and about 10:1, while a ratio of about 1.5:1 is considered particularly preferred. Ratios somewhat below 1:1 are also suitable.
The nickel complex catalysts are suitably preformed by contac-ting the catalyst precursors in the presence of ethylene in a suitable polar organic diluent or solvent, which is relatively unreactive toward the (boron hydride) reducing agent. In a preferred modification of producing the preferred catalyst complexes as detailed in the patents to Mason and Lutz, supra, the solvent, nickel salt and ligand are contacted in the presence of ethylene before the addition of reducing agent. It is essential that such catalyst compositions be prepared in the presence of ethylene. The catalysts are suitably prepared at temperatures of about 0 to 50 C, with substantially ambient temperatures e.g., 10 - 30 C
preferred. The ethylene pressure and contacting conditions should be sufficient to substantially saturate the catalyst solution. For example, ethylene pressures may be in the range from 1.7 to 104.4 bar or higher. Substantially elevated ethylene pressures, e.g., in the range from 28.6 to 104.4 bar are preferred.
The ethylene oligomerization process of the invention is necessarily carried out in a solvent which, for purposes of the particular improvement of the invention, is necessarily a glycol-based solvent. Preferably, this solvent is based on glycol(s) selected from the group consisting of C2 to C7 aliphatic diols and mixtures thereof. These diols include, for example, both the vicinal alkanediols such as ethylene glycol, propylene glycol, 2-methyl-1,2-propanediol, 1,2-butanediol and 2,3-butanediol, and the alpha-omega alkanediols such as 1,4-butanediol, 1,5~pentanediol, 1,6-hexa-nediol, aDd 1,7-heptanediol. Alpha-omega alkanediols of 4 to 3~

6 carbon atoms per molecule are preferred solvents with 1,4-butane-diol being particularly preferred. In some cases, it may be desira-ble to employ mixtures of the above-mentioned alkanediols for the glycol-based solvent in the process of the invention.
While the solvent is glycol-based, it may also suitably contain lesser amounts of other substances which react in some respects as solvents for the ethylene reactant and the catalyst. Such other solvent substances include, for eY~ample, impurities commonly present in (essentially anhydrous) commercial diol solvents or in the nickel chelate catalyst preparation and reaction modifiers other than water. As the oligomerization proceeds, the solvent will, of course, contain substantial amounts of ethylene and of oligomerization products and by-products.
In its broader aspects, the present invention centers upon an improvement in process performance which is provided by the presence of water in the oligomerization solvent, and the amount of water in the diol-based solvent is a critical aspect of the invention. For good results from the standpoint of reaction rate enhancement, water content of the solvent is suitably in the range from about 0.5 to 4.0 per cent by weight, calculated on the weight of the diol. No meaningful rate enhancement is observed for either lesser or greater water content. Water is readily soluble in the diol solvent phase in quantities within the specified range.
It has been the practice in the prior art relative to oligomeri-zation processes of the sort now claimed to utilize an essentially anhydrous diol solvent containing no more than about 0.2 %w water and typically no more than about 0.1 %w water. Water is present up to a level of about 0.1 %w in the commercially available anhydrous grade of diols heretofore used in oligomer production.
Preferably, the diol-based solvent in the process of the invention contains between about 0.7 and 3.5 %w water, while a water content between about 1.0 and 3.0 %w is more preferred. Considered most preferred is a quantity of water in the range from about 1.0 to

2.5 %w.
Without intending to be bound to one theory or mechanism of operation for the invention, it can be said that the presence of water in the solvent appears to essentially eliminate an induction period which has been characteristic of such oligomerization proces-ses in the prior art. In the absence of water, or in the presence of only small amounts (e.g., 0.1 %w to 0.2 ~w) of water, the maximum rate of oligomerization is not attained until near the end of a batch reaction. Under practice of the invention, on the other hand, the maximum rate occurs early in the course of a batch reaction.
An important aspect of ethylene oligomerization processes in general is the distribution of the various ethylene oligomers in the process product. The aforementioned patents rel~ting to ethylene oligomerization catalyzed by nickel chelate complexes describe a geometric distribution pattern for any given product which can be defined by a single constant, referred to as the "product distribut-ion constant" or "K factor", according to the mathematical expres-sion: .

mols of Cn+2 olefin K = ; (for n = 4,6,8...).
mols of C olefin It is known that this product distribution constant is affected by a number of parameters, including the type of bidentate ligand employed in the catalyst complex, the type of reaction solvent or diluent, the reaction conditions of temperature and pressure, the catalyst concentration in the solvent and the degree of ethylene saturation of the reaction solution. Furthermore, it has been recognized that the use of a large proportion of water in the reaction solvent, in particular, the use of a water-based solvent rather than a glycol-based solvent, leads to a product in which ethylene polymers rather than oligomers predominate. It has now also been found out that the K factor is potentially affected to a significant degree by the presence of small quantities of water in the reaction solvent under practice in accordance with the invent-ion. As a rule, it is considered preferable that the presence of water in the diol-based solvent not induce significant change in the process K factor, when compared to the results of a process carried out under substantially anhydrous conditions. For instance, a relatively high K factor9 e,g., on the order of about 0.75 to 0.80, is often desirable to maximize the production of olefins in the intermediate carbon number range of about C8 to C20 and to minimize the production of higher carbon number polymeric molecules. The presence of water in the solvent phase of the process of the inven~-ion is, however, in many cases responsible for a significant decline in K factor, e.g., a decrease of approximately 5 per cent.
A further important aspect of the invention is the discovery that the reaction rate enhancement associated with the presence of water in the diol-based solvent can be achieved without significant change in the product distribution constant. Specifically it has been found that rate enhancement benefits can be realized without - significant modification of the process K factor if the process is practiced under particular restrictions upon the concentration of nickel and ligand in the diol-based solvent. In this regard, is has been found to be very desirable to limit the ligand concentration to a value up to about 700 ppm and in particular 600 ppm (parts per million by weight, calculated on the weight of solvent) and to limit the molar ratio of nickel to ligand in the reaction solvent to a value greater than about 1.8. More preferably, the ligand concentration is less than about 500 ppm and the molar ratio of nickel to ligand in the solvent is at least about 2.0, while carry-ing out the process of the invention with concentration of ligand less than about 400 ppm and with molar ratio of nickel to ligand of at least about 2.3 is considered most preferred.
Under conventional practice, K factor is initially relatively high in a batch reaction, but declines as the reaction progresses.
In the practice of the invention, K factor is initially lower than with dry solvent (e.g., one with no more than about 0.2 %w water) ~L ~2 L?~ 3 t} ~

but remains essentially constant throughout a batch process.
With the exception of the presence of water in the diol-based solvent 9 the process of this invention may be practiced according to methods and under conditions well known in the prior art. Very suitably, a mixture of catalyst precursors and diol-based solvent is prepared and charged to an autoclave or similar pressure reactor.
The water may suitably be introduced into the diol-based solvent either before, during of after catalyst preparation and/or addition to the solvent. Ethylene and then sodium borohydride are introduced to complete catalyst formation and the reaction mixtu{e is maintai-ned with agitation at the desired reaction temperature and pressure for oligomerization. Reaction temperature may vary over a wide range, e.g., ~5 C to 150 C, but is preferably at least about 70 C, particularly between about 80 C and 110 C. The pressure must be at least sufficient to maintain the reaction mixture substantially in the liquid phase, although excess ethylene will also be present in a vapor phase. As a rule, the total pressure is less significant to the performance of the process than is the partial pressure of ethylene. Under preferred practice, ethylene partial pressure is in the range from about 49.3 to 173 bar, particul-arly in the range from aboue 70 to 139 bar. For achieving the greatest benefit from reaction rate enhancement in accordance with the invention, the ethylene partial pressure is most preferably in the range from about 77 to 104 bar. Contact between ethylene and the catalyst in the solvent is continued until the desired degree of oligomerization has occurred. The liquid product mixture is then suitably treated according to conventional procedures, typically including a separation of an oligomer-rich liquid phase from the diol-based solvent phase, scrubbing of the oligomer-rich phase to remove residual catalyst, dP-ethanization of the scrubbed liquid, and further work-up of the de-ethanized product to separate it into various product fractions. The above-referenced patents describe suitable procedures for carrying out the several steps of the overall oligomerization process for purposes of the invention in either a batch or a continuous manner.

l~Z~3~)3 The invention is further illustrated by reference to the following examples.

A series of examples (according to the invention) and compar-ative experiments (not according to the invention) were conducted ina baech mode.
In each case, a solution of nickel complex catalyst in diol-based solvent (95 %w 1,4-butanediol, 5 %w benzene) was first prepared at ambient temperature by introducing into a stirred stainless steel autoclave reactor: solvent (1,000 g), NiCl26H2O
(500 ml of a 0.3 molar solution in the solvent), KOH (499 g of a 0.1 molar solution in the solvent), and ligand (502 g of a 0.1 molar solution of o-diphenylphosphinobenzoic acid in the solvent). Ethy-lene was then introduced to a pressure of 35.5 bar, followed by NaBH4 reducing agent (200 g of 0.8 molar solution in diglyme contain-ing 0.8 g of NaOH), and finally by a further addition of ethylene to bring the pressure to 49.3 bar.
For initiation of thè oligomerization reaction~ the catalyst solution was heated to the desired temperature (65 ~C) and pressure in the reactor was adjusted to and maintained at 76.9 bar by further addition of ethylene. After 250 g of ethylene had been taken up by the reaction, a sample of the mixture was withdrawn to monitor K-factor of the intermediate oligomer product. After 500 g of ethylene had been taken up, the liquid mixture in the reactor was drained and allowed to separate into a solvent phase and a final oligomer product phase. K factors of the intermediate and final products were determined by gas-liquid chromatographic analyses.
Two comparative experiments, designated A and B, were carried out to establish the rate of the oligomerization reaction in the diol-based solvent, which contained only about 0.1 %w water. (This 0.1 %w water was attributed to water impurity in the butanediol, to water in the nickel chloride hydrate and to water resulting from neutralization of KOH). Examples 1, 2, and 3, all in accordance with the invention, were carried out to illustrate the rate enhancement associated with the presence of greater quantities of water (0.6 %w,

3~

1.1 %w, and 2.1 %wl respectively, calculated on cliol-based solvent).
The results (summarized in Table 1) indicate rate enhancement of approximately 30 to 50 per cent for the oligomerization reaction examples in which 0.6 to 2.1 %w was present in the diol-based solvent, relative to those comparative examples in which minimal water was present in the solvent.

Ethylene Uptake Solvent, 8 per g of Comparative Example water content ligand g per l solvent Experiment No. %w per h per h K Factor A 0.1 162 130 0.74 (1.7 h) 0.72 (2.8 h~
B 0.1 151 122 0.77 (1.7 h) 0.76 (2.8 h) 1 0.6 214 170 0.76 (1.5 h) 0.74 (2.2 h) 2a 1.1 233 188 0.71 (1 h) 3 2.1 217 169 0.71 (1.2 h) 0.70 (2.2 h) )In example 2, water was introduced during catalyst preparation through use of an aqueous NaBH4 solution (3.18 g NaBH4 and 0.40 g NaOH/100 g solution).

A further series of examples and comparative experiments were carried out in the following manner to illustrate preferred practices of the invention in which reaction rate enhancement is attained without significant change in product K factor.
For each of examples 4-7 and comparative experiments C-F, a 3~)~

solution of catalyst in 1,4-butanediol sc,lvent was prepared by mixing solvent, NiCl26H20, o-diphenylphosphinobenzoic acid, KOH, and sodium borohydride in the proportions inclicated in Table II, with the reducing agent added in the presence of ethylene. In each case, addition was made to a 500 ml stirred round bottom flask under nitrogen atmosphere of: about 301 g of anhydrous 1,4-butanediol (1,4-BD), about 1.32 g of a solution of NiCl26H20 in 1,4-BD (0.017 g of Ni per gram of solution); about 8.11 g of a solution of o-diphenyl-phosphinobenzoic acid in 1,4-BD (0.0128 g of ligand per gram of solution, and about 2.24 g of a solution of KOH in 1,4-BD
(0.00905 g of KOH per g of solution). The resulting clear yellow mixture was pressured into a one-litre autoclave, together with 35 g of ethylene, and stirred for 30 minutes. To the autoclave was then added a solution containing about 0.0227 g of sodium borohydride in 0.400 g of 1,4-BD and 0.400 g of water, followed by a further 2.47 g of butanediol and 87.5 g of ethylene. In examples 8-10 and comparative experiments I and J, the same procedures were followed - except that the catalyst was formed by adding the NiCl2 and KOH to the solvent, followed by ethylene, NaBH4 and ligand.
Different amounts of water were added to the solvent/catalyst mixture for the several examples. In the comparative experiments, the mixture contained about 0.2 %w water which was introduced with the various components during catalyst preparation.
The autoclave was rapidly heated to 95 C to initiate the oligomerization reaction, and maintained at this temperature for the full reaction period. Ethylene was introduced to maintain a pressure of about 92.7 bar. Samples of the reaction mixture were taken after 60 grams of ethylene uptake and after 136 g of ethylene uptake for determination of K factors.
The results of examples 4-10 and comparative experiments C-J
are presented in Table 2, as a function of water content of the solvene. The reaction rate is expressed in terms of the maximum rate of ethylene uptake in g per litre of catalyst/diol solution per h.

Solvent, K Factor Comperative Example water con.ent maximum Experiment No. %w reaction rate intermediate final C 0.2 354 0.79 0.76 D 0.2 413 0.80 0.77 E 0.2 413 ---- 0.76 F 0.2 435 ---- 0.76

4 1.0 5h5 0.73 0.78 2.0 566 0.77 0.77 6 2.0 558 0.77 0.77 7 4.0 441 0.78 0.775 G 8.0 348 0.765 0.76 H 8.0 396 0.765 0.76 I 0.2 368 0.71 -----J 0.2 364 0.75 0.70 8 2.1 581 0.73 0.70 9 2.1 576 0.71 0.69 2.0 559 0.70 0.67

Claims (22)

1. In a process for the preparation of oligomers of ethylene which comprises reacting at elevated temperature and pressure in a diol-based solvent and in the presence of a catalyst which is a chelate of nickel with bidentate ligand, the improvement which comprises reacting the ethylene in a diol-based solvent which contains between 0.5 and 4 per cent by weight of water, calculated on diol.

2. The process of claim 1, wherein the diol-based solvent contains between 0.7 and 3.5 per cent by weight of water.

3. The process of claim 2, wherein the diol-based solvent contains between 1.0 and 3.0 per cent by weight of water.

4. The process of claim 3, wherein the diol-based solvent contains between 1.0 and 2.5 per cent by weight of water.

5. In a process for the preparation of oligomers of ethylene which comprises reacting ethylene at elevated temperature and pressure in a diol-based solvent and in the presence of a catalyst which is a chelate of nickel with a bidentate ligand, the improvement which comprises reacting the ethylene in a diol-based solvent which contains (a) between 0.5 and 4 per cent by weight of water, calcul-ated on diol, (b) less than 700 parts per million by weight of ligand (calculated on total solvent), and (c) nickel and ligand in a molar ratio of nickel to ligand which is at least 1.8.

6. The process of claim 5, wherein the diol-based solvent contains between 0.7 and 3.5 per cent by weight of water.

7. The process of claim 6, wherein the ligand is present in the solvent in a concentration which is less than 600 parts per million by weight.

8. The process of claim 7, wherein the reaction is carried out under a partial pressure of ethylene in the range from 49.3 to 173 bar and at a temperature of at least 70 °C.

9. The process of claim 8, wherein the diol-based solvent contains between 1.0 and 3.0 per cent by weight of water.

10. The process of claim 9, wherein the molar ratio of nickel to ligand in the solvent is at least 2.0

11. The process of claim 10, wherein the reaction is carried out under a partial pressure of ethylene in the range from 70 to 139 bar.

12. The process of claim 11, wherein the diol-based solvent cont-ains between 1.0 and 2.5 per cent by weight of water.

13. The process of claim 12, wherein the ligand is present in the solvent in a concentration which is less than 500 parts per million by weight.

14. The process of claim 7, wherein the reaction is carried out under a partial pressure of ethylene in the range from 77 to 104 bar and at a temperature in the range from 80 °C to 110 °C.

15. The process of claim 1, wherein the ligand is an o-dihydrocarbyl-phosphinobenzoic acid or an alkali metal salt thereof.

16. The process of claim 15, wherein the ligand is o-diphenylphos-phinobenzoic acid or an alkali metal salt thereof.

17. The process of claim 4, wherein the ligand is an o-dihydrocarbyl-phosphinobenzoic acid or an alkali metal salt thereof.

18. The process of claim 17, wherein the ligand is o-diphenylphos-phinobenzoic acid or an alkali metal salt thereof.

19. The process of claim 5, wherein the ligand is an o-dihydrocarbyl-phosphinobenzoic acid or an alkali metal salt thereof.

20. The process of claim 19, wherein the ligand is o-diphenylphos-phinobenzoic acid or an akali metal salt thereof.

21. The process of claim 11, wherein the ligand is an o-dihydrocarbyl-phosphinobenzoic acid or an alkali metal salt thereof.

22. The process of claim 13, wherein the ligand is o-diphenylphos-phinobenzoic acid or an akali metal salt thereof.