We have been studying the intramolecular hydroamination of substituted aminoallenes with chiral titanium amino-alcohol derived catalysts.  The ligands are prepared in two steps from natural or unnatural amino acids and are enantiomerically pure.  We performed the hydroamination reaction catalytically with an enhanced regioselectivity for a-vinylpyrrolidine products over other titanium catalysts, with enantiomeric excesses of up to 16% for our first generation ligands (R' = H).  The aims over the last year were to improve upon our initial results through systematic ligand tuning.

Asymmetric intramolecular hydroamination of aminoallenes with titanium amino alcohol complexes;  R* = CH₂Ph or CHMe₂;  R = ⁱPr, c-C₆H₁₁ or 2-adamantyl;  1^st generation ligands, R' = H, Me, Bu or Ph.

Our first goal was to prepare ligands with increased steric bulk (R' = Me, Bu or Ph) by alkylating with the appropriate Grignard reagent instead of reduction with LiAlH₄ during the ligand synthesis.  Except for one derivative that has resisted multiple attempts at synthesis (R = ⁱPr, R* = CHMe₂, R' = Ph), we have completed the series and have synthesized and characterized the remaining 17 ligands. 

The substrates for the hydroamination reaction must be synthesized in a six-step procedure, and we sought a more rapid screen for catalytic activity by investigating the alkylation of benzaldehyde with diethylzinc.  Our ligands and their titanium complexes catalyze this reaction, and since the substrate is commercially available, it is a more rapid screen for enantioselectivity.  When this reaction is carried out with our first generation ligands (R' = H), we observe enantiomeric excesses below 10% in most cases.  However, increasing the steric bulk (R' = Bu or Ph) generally increases the ee values significantly, with values near 75% for our best ligands. 

We then continued our investigation of the catalytic intramolecular hydroamination with an aminoallene substrate.  We used 6-methyl-hepta-4,5-dienylamine as our substrate as it gives only a single product.  Enantioselectivities were measured for the 17 newly prepared ligands;  however, there is not a substantial increase in enantioselectivity with the second generation ligands in this reaction as was seen for the benzaldehyde screen.  This led us to two conclusions.  First, that the benzaldehyde reaction is not a good catalytic screen for the hydroamination.  The second is that we require more substantial ligand tuning to achieve higher enantioselectivity in this reaction.

Three of the other ligand derivatives from our proposal were also investigated during the past year.  We attempted the synthesis of N-aryl amino alcohols by an aromatic amination of an aryl halide but have not been successful.  We have prepared three sulfonamide ligands, including two with electron withdrawing CF₃ groups and are in the process of determining the enantioselectivities of the hydroamination reaction with these new ligands.  We have also prepared four new tridentate aminophenol ligands that await further investigation.

We have also continued to examine the coordination chemistry of the amino-alcohol ligands (H₂L) with three titanium starting materials in a collaborative computational project.  We have experimental results showing substantial reactivity differences of the ligands with Ti(NMe₂)₄, TiCl(NMe₂)₃, and TiCl₂(NMe₂)₂.  We studied the energetics (equilibrium and transition state geometries) of reasonable reaction coordinate pathways corresponding to the reactions of the three titanium starting materials with the H₂L ligands.  Calculations (B3LYP, 6-311G(d)) have been carried out in collaboration with Robert J. Cave at Harvey Mudd College.  We hope that completion of this theoretical project will shed some insights into catalyst design in the hydroamination reaction.