New P,O-donor ligands and their use in hydroformylation and carbonylation reactions
The synthesis of 2 new senes of potential hard - soft donor ligands has been investigated. These ligands have the general formula [ph;J>(CH2)nP(O)PhR] where: n=I [R==CH3 (MPPMO), R=C2Hs (EPPMO), R=C3H7 (PPPMO), R=CJI9 (BPPMO)] n=2 [R=CH3 (MPPEO), R=C~s (EPPEO), R=C3H7 (PPPEO), R=CJI9 (BPPEO)]. The ligands are made by a 2 step process. Firstly, careful alkylation of the appropriate diphosphine leads to monophosphonium salts. Subsequent alkaline hydrolysis of these intermediates produces the corresponding diphosphine monoxides in high yields. The substituent R group bonded to the P=O group, corresponds to the alkyl group on the alkylating agent. The study involved examination of a range of experimental conditions for both steps. This lead to different optimal experimental conditions for each synthesis depending on the nature of R for alkylation, the critical factor was the strict stoichiometric 1: 1 addition of starting material to alkylating agent, as any excess lead to disalt formation. Alkylating agents of high reactivity were pre-dissolved in the solvent and required drop-wise addition in the form of a dilute solution, while those oflower reactivity were added as neat ligands. These lower reactivity alkylating agents also required the application of heat, but only after an initial quantity of mono salt had formed. Hydrolysis of these alkali resistant salts was achieved by substitution of most of the water in the reaction medium by the aprotic solvent, tetrahydrofuran. Union Carbide had previously prepared stable well characterised rhodium(I) complexes containing the related chelated ligand [ph;J>(CH2)nP(O)Ph2] where n=l (DPPMO), n=2 (DPPEO); and had reported their use in catalyst hydroformylation and carbonylation reactions. The proposed mechanism involves chelate ring opening and closing via the p=o to rhodium bond. The aim of the present study was to examine the effect of changing the nature of the R group bonded to the P=O unit. Changing the R group might be expected to affect the co-ordinating power of the P=O group (i.e. varying the hardness of the oxygen donor atom affects its ability to co-ordinate to the soft rhodium cation). Thus, if the above mechanism for catalysis is in operation, changes might be expected to be observed in catalyst efficiency as the ease of ring opening and closing varies. A re-examination of the Union Carbide experiments had been made by workers at B.P. Chemicals' and while this agreed with the trends of the Union Carbide work, there was need for a more detailed study. Part of the study reported here was carried out in the laboratories of B.P. Chemicals'. Despite taking great care in reproducing the experimental conditions of the earlier work, little difference was found between the catalytic efficiency of the diphosphine oxide ligand and PPh3. This threw doubt on the proposed mechanism A separate set of experiments performed at Kingston University substantiated, these findings. These results suggest a mechanism involving active species containing unidentate P¬bonded coordination of ligands. Variation in efficiency was only observed in the case of the bulkiest R group (n-butyl) and this can be ascribed to a steric effect. For the carbonylation of methanol, it had been reported that [RhCI(CO)(DPPEO)] is the precursor to an active catalyst containing P-bonded unidentate ligands. Catalyst stability in this system is attributed to ligand hemilability allowing alteration between the chelated resting state of the catalyst and an active species containing the unidentate P-bonded ligand. This contrasts with other reports in which a wide range of rhodium - phosphine catalyst precursors all generated [Rbl2(CO)2f as the active species. The present work measured average reaction rates for catalysis by homogeneous rhodium systems containing PPh3, DPPMO, MPPMO and PPPMO. - The rates were almost identical, evidence which is consistent with generation of [Rbl2(CO)2f in each case.