Abstract
Crisscross cycloaddition of citronellal azine (6) with 2 equiv. of TFA and powdered 3 Å molecular sieves in CH2Cl2 at reflux for 22 h afforded 37% of the desired C2-symmetric hydrazine 7 and 5%–10% of diastereomer 8 in which one of the 6–5 ring fusions is cis. Methylation of the hydrazine of 7 and reduction of the resulting salt (9) with Li in NH3 cleaved the N—N bond to give secondary tertiary amine 10 in 97% yield. Eschweiler–Clarke methylation afforded the C2-symmetric bis tertiary amine 11 in 69% yield. Racemic products were obtained in initial attempts at asymmetric catalysis using 7 or 11 as asymmetric bases, using bistertiary amine 11 as a ligand analogous to sparteine for alkyllithiums, or using the lithium amide from secondary tertiary amine 10 as an asymmetric base. Apparently, the proton is buried in the core of 11, leaving a hydrophobic surface; the free counterion is not an asymmetric catalyst. Diamine 11 may be too hindered to complex to s-BuLi. Tertiary amine 11 (pKa1 = 24.7) is more basic than DBU (pKa = 24.3) in CH3CN, in good agreement with theory.Key words: crisscross cycloaddition, azine, dipolar cycloaddition, calculation of pKa.