Combinatorial and supramolecular saproaches to monodentate phosphorus-ligands for transition metal catalyzed asymmetric reactions.

In recent years, monodentate phosphorus ligands (e.g. phosphites, phosphonites, phosphoramidites and phosphinamines) have held the stage in asymmetric catalysis.1 In addition to their outstanding activity and selectivity, comparable or even superior to those of bidentate ligands, the convenient, fast and practical preparation from commercially available materials underlines their potential for industrial applications. Furthermore, the modular nature of all these ligands allows the synthesis of a variety of representatives, thereby making a combinatorial approach possible. My PhD research project deals with the synthesis of libraries of new chiral monodentate P-ligands and their screening in asymmetric transition metal-catalyzed reactions through two innovative approaches: combination of binaphthol-derived phosphites and C1-symmetric phosphinamines for the selective generation of heteroleptic catalysts in Rh- and Pd-mediated reactions (Ligand Combination Approach, Research line 1); supramolecular ligand-ligand interactions for highly selective transition metal catalysis (Supramolecular Approach, Research line 2). Research line 1 In 2002-3, the groups of Reetz and Feringa independently used a binary mixture of chiral monodentate Pligands in several asymmetric rhodium-catalyzed reactions. By mixing two different ligands (La and Lb) in the presence of a transition metal [M], three species can be formed: [M]LaLa, [M]LbLb (homocomplexes) and [M]LaLb (heterocomplex). The heterocomplex is often more reactive and more (regio-, diastereo-, enantio-) selective than either of the two homocomplexes. Moreover, under thermodynamic control the heterocomplex : homocomplexes ratios usually exceed the statistical value (2 :1 :1). The ideal case would constitute an equilibrium completely in favor of the heterocomplex [M]LaLb, because a single well-defined catalyst would then exist in the reaction mixture and the undesired competition of the less selective homocomplexes would be avoided. We were intrigued by the remarkable selectivities reported for the mixtures of chiral ligands (binolderived phosphites, phosphonites and phosphoramidites) with achiral phosphines.4 In particular, the 1:1 mixture of a chiral phosphite with an achiral phosphine was reported to induce reversal of the enantioselectivity in the Rh-catalysed hydrogenation of N-acetamido acrylate (compared to the chiral phosphite alone). The only possible explanation for this peculiar behaviour is the selective formation of the phosphite-phosphine Rh-heterocomplex, favoured by electronically matching one -donor ligand (phosphine) and one -acceptor ligand (phosphite). As enantiomerically pure chiral phosphines are not easy to synthesize, we were wondering whether the phosphine ligands could be substituted by other - donor phosphorus ligands still retaining the thermodynamic preference for the formation of the heterocomplex. For this reason, we turned our attention to chiral phosphinamines, which are easy to prepare enantiomerically pure, and have electronic properties similar to those of phosphines.5 DFT calculations showed that the phosphite-phosphinamine rhodium heterocomplex is more stable than the two homocomplexes by 11.29 kcal/mol We synthesized a library of monodentate chiral phosphites 1a-d by reacting enantiopure BINOL-PCl with different chiral and achiral alcohols. Monodentate chiral phosphinamines 2a-e were prepared by reaction of Ph2PCl with a number of C2-symmetric secondary amines and C1-symmetric secondary and primary amines. Complexation studies were performed by means of 31P-NMR, using Rh(acac)(C2H4)2 as the rhodium source. When C1-symmetric phosphinamine ligands were employed, the cis-heterocomplexes were formed with selectivity ranging from moderate (70%) to excellent (_ 99%). The homo- and heterocombinations of phosphites and C1-symmetric phosphinamines were then screened in the rhodium-catalyzed hydrogenation of methyl 2-acetamidoacrylate. Remarkably, the 1:1 combination of a BINOL-derived phosphite and a phosphinamine induced reversal of the enantioselectivity, compared both to the phosphite and the phosphinamine alone. This heterocombination induced a peculiar stereochemical outcome also in the palladium-catalyzed asymmetric allylic substitution of rac-1,3- The heterocomplexes formed in this way are expected to have reduced degrees of freedom7 compared to the complexes of normal monodentate ligands, and thus supramolecular ligands somehow resemble traditional bidentate ligands. According to this analogy, they are often referred to as supramolecular bidentate ligands or self-assembled ligands, thanks to their ability to spontaneously form bidentate systems in solution. These terms also apply to those supramolecular ligands that are only capable of noncomplementary interactions: indeed these systems are still capable of forming rigid and conformationally restricted complexes, although they cannot selectively form heterocomplexes when used in a mixture, which quite reduces their "combinatorial appeal". Prompted by the studies of Prof. Reek and co-workers on the synthesis of supramolecular bidentate heterocomplexes through complementary interactions (e.g. coordinative bondings)8 and by the efficient resolution strategy of racemic N-benzyl _-amino acids (N-Bn-AA) by liquid-liquid extraction using a chiral salen–cobalt(III) complex as enantioselective receptor, accomplished in our laboratories, we decided to use the salen-cobalt(III)-N-benzyl-L-serine complex as a chiral platform for the preparation of new families of supramolecular mono- and bi-dentate P-ligands (Scheme 4 A and B). The salen-cobalt(III)-N-Bn-AA complexes, in fact, possess a rigid framework with the salen ligand in a cis-_-folded arrangement around the octahedral cobalt ion. The remaining two cis coordination sites are occupied by the N-benzyl _-amino acid, which is thus accommodated in the “binding pocket” of the chiral cobalt complex. 9b Ligands 3 were synthesized in good yields starting from the symmetrical tetra-t-butyl-salen which was transformed into the corresponding cobalt(III) acetate complex, and the acetate ion was exchanged with Nbenzyl- L-serine. In the case of symmetric salen backbone, the complex was obtained pure in high yield and the phosphite moieties were then introduced by reaction with different diol-derived chlorophosphite. In the case of the supramolecular bidentate ligands 4, the unsymmetrical hydroxymethyl-containing salen afforded the corresponding cobalt(III) acetate complex as a mixture of two inseparable diastereoisomers. Complexation of monodentate ligands 3 to rhodium(I) was studied by 31P-NMR spectroscopy, which showed the formation of the desired Rh-complexes. The reactivity of supramolecular monodentate Pligands was investigated in two Pd-catalyzed-allylic alkylation on (E)-1,3-diphenylallyl acetate and Pdcatalized desymmetrization of meso-cyclopenten-2-ene-1,4-diol biscarbamate. We are now exploring new synthetic pathways to obtain the unsymmetrical (S,S)-salen-ligand and the corresponding unsymmetrical (S,S)-salen-cobalt(III)-N-Bn-L-serine bidentate P-ligand in a pure diastereoisomeric fashion. We have also investigated the design and synthesis of a novel class of chiral monodentate phosphite ligands, named PhthalaPhos, which contain a phthalic acid primary diamide moiety. Such phthalamidic group displays both donor and acceptor hydrogen bonding properties that, in principle, can give rise to supramolecular interactions both between the ligands and with the reaction substrate The pre-catalytic Rh complex of one of these ligands was fully characterized and studied by NMR, IR and mass spectroscopy, which confirmed the presence of hydrogen bonds between the coordinated ligands, and the formation of a supramolecular bidentate ligand. The catalytic properties of these new ligands were assessed in the rhodium-catalyzed hydrogenation of benchmark olefins (e.g. methyl 2-acetoamidoacrylate and N-(1-phenylvinyl)acetamide), taking the known phosphite 9 as a touchstone. Four ligands gave e.e. values higher than 97% with methyl 2-acetoamidoacrylate as substrate, and six reached the same level of performance with N-(1-phenylvinyl)acetamide. Remarkably, the reference ligand 9, featuring the same BINOL phosphite moiety, gave only 84% and 90% e.e. respectively for the same substrates, thus suggesting that the phthalimide residue significantly influences the catalytic properties of these ligands. Evaluation of the catalytic properties of the Phthalaphos ligands were also extended to the Rh-catalyzed hydrogenation of more challenging substrates of potential industrial interest, such as N-(3,4-dihydronaphthalen-1-yl)acetamide and (E)-methyl 2-(acetamidomethyl)-3-phenylacrylate. The results of this screening were quite variegated both in terms of activity and enantioselectivity, but the employment of a Phthalaphos ligand gave the highest e.e. value ever obtained on N-(3,4-dihydronaphthalen-1-yl)acetamide with phosphite ligands and the highest e.e. value ever obtained with (E)-methyl 2-(acetamidomethyl)-3-phenylacrylate.