= Solution
\b[B is the major product: the newly formed OH is the dashed bond in the printed zigzag structure.] Retain the existing solid-wedge OH and the configuration at the carbamate-bearing <carbon>.
<zinc> coordinates the <ketone> <oxygen> and the adjacent hydroxyl <oxygen>, forming a five-membered O–Zn–O chelate. This constrains rotation around the <carbonyl>–alpha-carbon bond. In the favored chelated <conformation> the bulky $\mathrm{CH(NHCO_2R')CH_2OR}$ group shields one face of the <ketone>, while the alpha <hydrogen> leaves the other more accessible. <hydride> attacks from that less crowded face, the <carbonyl> $\pi$ bond opens onto <oxygen>, and aqueous work-up gives the second <alcohol>.
Mapping this \b[<Cram chelation model>] back onto the printed zigzag gives B, not A. Merely placing a new OH “on the same side” of an unrotated page drawing is not a valid <chelation> argument: the reactive chelate and that drawing use different <conformations>. Zinc-borohydride reductions of alpha-hydroxy <ketones> are experimentally known to favor the corresponding anti-1,2-diol family under <chelation> control.<Stereoselective Reduction of 2-Hydroxy Ketones towards syn- and anti-1,2-Diols by S. M. Husain and others>[Abstract, chelation-controlled zinc-borohydride reductions]{ref}
The reagent combination can generate <zinc borohydride> and related zinc-associated reducing species. The stereochemical justification is coordination-controlled facial selectivity; it does not require identifying a unique solution aggregate from the reagent list. The quoted yield is not itself a measurement of a 95:5 diastereomeric ratio.
Back to article page
