Asymmetric catalysis by chiral rhodium complexes in hydrogenation and hydrosilylation reactions

The preparation of optically active molecules needs a chiral auxiliary. It is important to use the minimum amount of this auxiliary, and from that point of view asymmetric catalysis is much more advantageous than stoichiometric asymmetric synthesis. Some homogeneous catalysts prepared from chiral complexes have become during the past few years a useful tool in asymmetric synthesis. The complexes L2RhQ where L2 is a family of chiral diphosphines were prepared and used in asymmetric reduction. DIOP is a readily available ligand prepared from tartaric acid. Many of its derivatives were obtained as weil as other types of chiral phosphines. General syntheses of optically active cx-amino acids, amines or acids are described. Optical yields as high as 90 per cent could be attained. The same complexes can catalyse the hydrosilylation of ketones and imines, giving rise after hydrolysis to optically active alcohols and amines. To improve the usefulness of asymmetric catalysis a supported chiral catalyst was prepared starting from a Merrifield resin. It was used both in reduction and in hydrosilylation. The mechanism of the reactions and the origin of the asymmetric induction will be discussed. The creation of optical activity is an important problern from both theoretical and practical points ofview. For many useful chiral compounds it is only one enantiomer which is desired, as is the case with pharmaceuticals, food additives, perfumes .... Total synthesis is one of the processes which is becoming important on the route to such compounds. lt is then an interesting goal for organic chemists to devise efficient methods of generating optical activity. Almost a11 the methods take advantage of using a chiral auxiliary which can be recovered at the end of the process. In a resolution only 0.5 mole of the desired enantiomer is at best obtained (for each one mole ofthe auxiliary chiral material). In an asymmetric synthesis one mole can be prepared, and thanks to an asymmetric catalysis (where the catalyst is chiral) an unlimited amount of optically pure compound should be, in principle, synthesized. Asymmetrie catalysis will become operative if the two following objectives are to be reached: high stereospecificity and good catalytic activity, i.e. high substratejcatalyst ratio. In recent years, the utilization of chiral soluble complexes as catalysts has given a new life to the field of asymmetric catalysis (for 401 H.B.KAGAN some recent reviews see refs 1, 2). We will describe the main results that we obtained in this area using optically active rhodium complexes and we deal with several types of reaction. When we started to investigate asymmetric catalysis with soluble complexes there were only a few reports which were not very encouraging since only small asymmetric induction could be observed. In order to take a rational approach in constructing a chiral catalyst we looked for a reaction where the mechanism was reasonably weil known and where there is some sensitivity to steric hindrance. We selected the efficient catalyst system described by Wilkinson3 where the catalyst is a rhodium complex. A feature of rhodium complexes is ability to pass easily from oxidation state I to oxidation state 111 (and the reverse). A covalent molecule X2 for example, can be cleaved and added to a rhodium (I) complex by an oxidative addition(Figure 1). The Oxidative addition L3Rh111X3 -->- L3Rh1X2 Reductive elimination Figure 1 reverse process is called a reductive elimination. IfL is a phosphine and X2 = H2 we have the basis of the catalyst system studied in great detail by Wilkinson3 in 1966 (Figure 2). The complex RhClL2S (where S is a presumed coordinated solvent molecule) can be generated from the precursors RhClL3 or (RhClL2h. The oxidative addition of hydrogen followed by displacement of S by an olefm gives the crucial complex [RhClL2HHolefm] wherein both 18 el. 16 el. Wilkinson Tolman Halpern 1966 1972 1973 Figure 2 402 ASYMMETRIC CATALYSIS BY CHIRAL RHODIUM COMPLEXES hydrogen atoms and substrate are present. Dihydrido complexes were studied and it seems that the two hydrogens are cis to each other. The reduction occurs then by two consecutive internal hydrogen transfers. Foracis addition, the second step must occur with retention of configuration at the carbon atom ofthe alkylrhodium intermediate. The kinetic details ofthe system have given rise to many studies and, especially during the last two years several publications have appeared (Tolman, Halpern, De Aguirre). The main pathway takes place via the hydride route, with apart of the reaction occurring through the dimer (RhClL2h which can fix one hydrogen molecule. The unsaturate route, oxidative addition of hydrogen on [Rh, CL L2, olefin] is not very effective. The catalyst is very sensitive to steric hindrance, tetrasubstituted double bonds are not reduced. Many types of carbon--carbon double bonds are reducible, in contrast to carbonyl or imine groups. How should one use the Wilkinson catalyst in order to reduce prochiral double bonds asymmetrically? lt is necessary to introduce chiral ligands L. We prepared the chiral catalyst by a general method (Figure 3) which starts EtOH + 4L + 28 Benzene 2 Figure 3 from (RhC10lefm2h, a stable complex easily obtained from RhC13 • The addition of ligands such as phosphines gives a displacement of olefms. The new species is usually not isolated and is the catalyst. Phosphines Authors Advantages Disadvantages g * Horner et alii Inducing chiral Difficult y in the T 0 0 Knowles et a/ii centre close to synthesis and R; Rz 'R3 the complex resolution * /C6Hs Morrison et a/ii Facility in the Inducing chiral R-P, synthesis, no centre far from C6Hs resolution the complex ...... c6Hs (p"'- Dang and Kagan Fa<;ility in the Inducing chiral * C6Hs R synthesis, no centre far from \_p/C6Hs resolution, good the complex 'C6Hs chelating properties Figure 40 Chiral phosphines as Iigand 403 H.B.KAGAN To prepare chiral phosphines many approaches can be envisaged (Figure 4). Several teams have worked on the problem~. We used phosphines R-P(C6H 5h where R is chiral. In order to improve the steric control introduced by R we chose 7 to use diphosphines with the idea of decreasing the flexibility around the R7-P bond. To give an equivalent character to the diphenyl-phosphino groups we frrst selected molecules with a twofold axis of symmetry. It is essential to start from a cheap material, with both the enantiomers available. We chose tartaric acid and prepared several phosphinated derivatives. One possibility is to. use the carbon atoms of the carboxylic groups after protecticin ofthe diol system. We prepared a dimethoxydiphosphine (Figure 5) which was not very efficient in asymmetric catalysis. To reduce Figure5 the flexibility of the chelate seven-membered ring we fused it with a dioxolan ring. The trans configuration between them distorts the heptagonal ring thanks to a large dihedral angle. This acetonid diphosphine is a crystallized compound easy to purify and to store, and was called DIOP7 • (-)DIOP is derived from ( +) tartaric acid. It is with DIOP and its derivatives that we obtained results which are among the best known in asymmetric catalysis. In our standard conditions we generally use a substrate to catalyst ratio of 100, and all the experiments were performed at room temperature under atmospheric pressure. The more convenient solvent is a mixture benzeneethanol (1/2), with a substrate concentration of 0.3 M. REDUCTION OF CONJUGATED ACIDS Our frrst experiments were conducted on atropic acid and its derivatives (Figure 6). DIOP induced asymmetric reduction with a moderate enantiomeric excess (% e.e.), but an optical yield as high as 63 per cent could be attained in the specific case of the free acid with addition of a small amount of triethylamine. It is interesting to notice that the absolute configuration is reversed with respect to the methyl ester. It is possible that the carboxylate function is coordinated to the rhodium atom during the reduction (Figure 7), giving a new steric situation (with respect to the ester reduction). Many unsaturated acids were reduced by the system RhjDIOP, the optical yields are often high and strongly related to the E-Z isomerism around the double bond8• Among the conjugated acids which can give rise to an asymmetric reduction, the ot-N -acylaminoacrylic acids will be specially considered since they are precursors of the important class of ot-amino acids. 404 Catalyst: 1A0 eq. Rh/(-)DIOP Figure 6 H CP,,,,,,I ,,,,,,H ."..Rh~ll P I ~ ,,,,,,<1> 0-C~ Figure 7 ASYMMETRie SYNTHESIS OF cz-AMINO ACIDS To test the possibilities of the system Rh/DIOP we frrst investigated7 several potential precursors of phenylalanine. The corresponding azlactone and hydantoin are not reduced, but all the other compounds of Figure 8 react easily. The optical yields are quite high, (R)-amino acids are obtained if (-)DIOP is used. Of course we prepared too natural (S)-amino acids by taking (+) DIOP. Many amino acids were synthesized in this way (Figure 9), the chemical yields being almost quantitative. We were surprised to see that substituents on the phenylring of the N-acetylphenylalanine precursor enhance significantly the optical yield (72 per cent for phenylalanine, 80-83 per cent for DOPA or tyrosine). We recently found that this is not a substituent effect. Indeed the specific rotation described in the Iiterature for N-acetylphenylalanine (used also by us 7) is in error. The value is too high. We could demonstrate this fact by v.p.c. analysis on t-butyl-N-lauroyl valine as the chirat phase, accord