Ketone reduction with axial hydride attack

In hydrogenolyses with HAICU, the dimethyl acetals of cyclobutanone and cyclohexanone are cleaved more slowly than that of 3-pentanone, while those of cyclopentanone and cycloheptanone are cleaved more rapidly (Table 1), as would be expected for a carbonium ion process. The differences in rate are small, suggesting that carbonium ion character is not strongly developed in the transition state. With the dimethyl acetal of 4-t-butylcyclohexanone, the hydride addition step occurs with strongly predominating axial addition when HAlCh is used Zn(BH4)2 with TMS-Cl, and TMS-H with TMSO-Tf are less selective (Table 2). Equatorial attack predominates, however, in the reduction of the ketone itself with TBDMS-H and TBDMS-OTf. ... 

Substituents which destabilize the chair conformation give stereoisomeric products, as demonstrated by the reduction of substrate 92, which in contrast to the results above, exclusively generates the exo alcohol 93, Eq. 74 [116], Ketone 92 presumably exists in a boat conformation with pseudo-equatorial methyl groups rather than in a pseudo-chair conformation which would place the methyl substituents in pseudo-axial positions. Hydride attacks from the less-hindered endo face to provide the exo and pseudoaxial alcohol. 

Any explanation of facial selectivity must account for the diastereoselection observed in reactions of acyclic aldehydes and ketones and high stereochemical preference for axial attack in the reduction of sterically unhindered cyclohexanones along with observed substituent effects. A consideration of each will follow. Many theories have been proposed [8, 9] to account for experimental observations, but only a few have survived detailed scrutiny. In recent years the application of computational methods has increased our understanding of selectivity and can often allow reasonable predictions to be made even in complex systems. Experimental studies of anionic nucleophilic addition to carbonyl groups in the gas phase [10], however, show that this proceeds without an activation barrier. In fact Dewar [11] suggested that all reactions of anions with neutral species will proceed without activation in the gas phase. The transition states for reactions such as hydride addition to carbonyl compounds cannot therefore be modelled by gas phase procedures. In solution, desolvation of the anion is considered to account for the experimentally observed barrier to reaction. 

Table 10. Polycyclic 2-Cyclohcxcnols by Stereoselective Reduction of the Corresponding Ketones with Lithium Aluminum Hydride or Sodium Borohydride (Selective Axial Attack)...

---