Isolation and X‐Ray Structures of Reactive Intermediates of Organocatalysis with Diphenylprolinol Ethers and with Imidazolidinones

Abstract

Abstract Reaction of 2‐phenylacetaldehyde with the Me 3 Si ether of diphenyl‐prolinol, with removal of H 2 O, gives a crystalline enamine ( 1 ). The HBF 4 salts of the MePh 2 Si ether of diphenyl‐prolinol and of 2‐( tert ‐butyl)‐3‐methyl‐ and 5‐benzyl‐2,2,3‐trimethyl‐1,3‐imidazolidin‐4‐one react with cinnamaldehyde to give crystalline iminium salts 2, 3 , and 4 . Single crystals of the enamine and of two iminium salts, 2 and 3 , were subjected to X‐ray structure analysis ( Figs. 1, 2 , and 6 ), and a 2D‐NMR spectrum of the third iminium salt was recorded ( Fig. 7 ). The crystal and NMR structures confirm the commonly accepted, general structures of the two types of reactive intermediates in organocatalysis with the five‐membered heterocycles, i.e. , D, E ( Scheme 2 ). Fine details of the crystal structures are discussed in view of the observed stereoselectivities of the corresponding reactions with electrophiles and nucleophiles. The structures 1 and 2 are compared with those of other diphenyl‐prolinol derivatives (from the Cambridge File CSD ; Table 1 ) and discussed in connection with other reagents and ligands, containing geminal diaryl groups and being used in enantioselective synthesis ( Fig. 4 ). The iminium ions 3 and 4 are compared with N ‐acylated imidazolidinones F and G ( Figs. 9, 12 , and 13 , and Table 3 ), and common structural aspects such as minimalization of 1,5‐repulsion (the ‘A 1,3 ‐effect’), are discussed. The crystal structures of the simple diphenyl‐prolinol⋅HBF 4 salt ( Fig. 3 ) and of Boc‐ and benzoyl‐( tert ‐butyl)methyl‐imidazolidinone (Boc‐BMI and Bz‐BMI, resp.; Figs. 10 and 11 ) are also reported. Finally, the crystal structures are compared with previously published theoretical structures, which were obtained from high‐level‐of‐theory DFT calculations ( Figs. 5 and 8 , and Table 2 ). Delicate details including pyramidalization of trigonal N‐atoms, distortions around iminium CN bonds, shielding of diastereotopic faces, and the π ‐interaction between a benzene ring and a Me group match so well with, and were actually predicting the experimental results that the question may seem appropriate, whether one will soon start considering to carry out such calculations before going to the laboratory for experimental optimizations.