A superior P-H phosphonite: Asymmetric allylic substitutions with fenchol-based palladium catalysts

The fenchol-based P-H phosphonite BIFOP-H exeeds with 65% ee other monodentate ligands in the Pd-catalyzed substitution of 1-phenyl-2-propenyl acetate with dimethylmalonate.


Introduction
Palladium catalyzed allylic substitutions provide valuable tools for stereoselective C-C-and C-heteroatom connections. [1,2] The control of regio-and enantioselectivity is challenging, especially with unsymmetrical substrates, e.g. with monoaryl allyl units. According to computational analyses of electronic effects, [3,4] regioselectivity in favor of the branched product is supported at strong donor-substituted (e.g. alkyl, O-alkyl) allylic positions. Frequently employed Pd-catalysts most often favor linear, nonchiral products (Scheme 1).

Scheme 1: Pd-catalyzed allylic substitution with unsymmetrical substrates (Nu = dimethylmalonate, Nf = OAc).
Pfaltz et al. improved the yield of the chiral, branched product by employing electron withdrawing substituents on the P-donor atoms in P, N-oxazoline ligands [5] (Scheme 2) [6]. Such phosphites were thought to favor a more S N 1-like addition at the substituted, allylic C-atom.
X-ray crystal structure of the cationic complex Pd-FENOP (CCDC 299944), the perchlorate counterion and hydrogen atoms are omitted   (i.e. malonate) at the weakest (longest) C3-Pd bond, yielding preferably the chiral, branched product.
Monodentate BIFOP ligands yield more of the linear alkylation product (Table 1), despite their huge steric demand. Surprisingly, the chloro-and bromophosphites, BIFOP-Cl and BIFOP-Br, are stable ligands under these reaction conditions: no conversion with nucleophiles (e.g. malonate), as was observed previously with diethylzinc, [26] was found. The ligands were recovered after catalysis. Apparently, the absence of strongly Lewis-acidic electrophiles (Na+ vs. Zn2+) and the huge steric shielding prevents halide substitutions and BIFOP-Cl(Br) decompositions.
With regard to enantioselectivities, some monodentate BIFOPs are even superior to the pyridine-phosphinites (FENOPs). While FENOPs favor the R-enantiomeric product, the S-enantiomer is preferred by all BIFOP ligands. Enantioselectivities increase from FENOP with 19% ee to FENOP-Me with 31% ee and to FENOP-NMe 2 with 37% ee, reflecting the effect of steric demanding and electron donating pyridine groups on enantioselectivity.
The surprisingly stable halogen phosphites BIFOP-Cl and BIFOP-Br yield even higher enantioselectivities (39% and 37% ee) than the corresponding phosphite BIFOP-OPh or the phosphoramidite BIFOP-NEt 2 (10% and 29% ee, Table 1). To our knowledge, this is the first successful application of halogen phosphites as ligands in enantioselective catalysis [26]. The highest enantioselectivity however is achieved with the P-H phosphonite BIFOP-H (65% ee, Table 1). As in copper-catalyzed 1,4-additions of diethylzinc to cyclohexenone, [26] the small steric hindrance of the hydrido-substituent and the shielding by the chiral bis-fenchane cavity provide the best combination among the tested BIFOPs for the P-H phosphonite BIFOP-H.
Computational transition structure analyses of allylic substitutions with ammonia mimicking the malonate nucleophile help to understand origins of enantioselectivities, [30][31][32][33] as we have shown recently for Pd-FENOP catalysts with the diphenyl allyl substrate [25]. For the P, N-bidentate pyridyl FENOP system, an exo allyl arrangement and a trans to phosphorus addition of the nukleophile is slightly preferred (cf. the two most stable transition state in Figure 4). This favored Si-addition of the nucleophile explains the experimentally observed formation of the Ralkylation product (Table 1). Systematic conformational analyses of transition structures with BIFOP-H in allylic substitutions yields BIFOP-H-Re as the most stable transition structure. Its Re-addition of the NH 3 -nucleophile is slightly more favored than the Si-addition in the competing transition structure BIFOP-H-Si ( Figure 5). This agrees with the experimentally observed formation of the Salkylation product with BIFOP-ligands (Table 1).
X-ray crystal structure of the cationic complex Pd-FENOP-NMe 2 (CCDC 600370), the perchlorate counterion and hydrogen atoms are omitted