Zimmerman-Traxler model steric interactions

A.iv. The Evans Model. It is known that (Z) enolates are more stereoselective than ( ) enolates even when r1 is not large. The Zimmerman-Traxler model transition states 352-355 do not account for this observation. It has been suggested that the transition states are not chair-like, but skewed, as in 381-384.221 In this representation (Z) enolate 381 leads to the syn aldol. Similarly, (Z) enolate 382 gives the anti aldol, ( ) enolate 383 give the anti aldol and E) enolate 384 is the precursor to the syn aldol. The major steric interactions in this model are those for r1 r3 and r2 - r3. For both (Z) and E) enolates, the r1 r3 interaction favors 381 and 383, respectively. The r2 r3 interaction is more important for the E) enolate and... 

In most cases, Crams rule (sec. 4.7.B) predicts the major isomer when the reaction partner (or partners) contain a chiral center. To understand how this rule applies to orientational and facial selectivity, we must understand the transition state of the reaction (invoke the Zimmerman-Traxler model or one of the other models for predicting diastereoselectivity). The Zimmerman-Traxler model is used most often, and if it is applied to 423 and 424, the syn selectivity can be predicted. The facial selectivity shown in 427 and 428 arises from the methyl group. In 428, the enolate approaches from the face opposite the methyl, leading to diminished steric interactions and syn product (429). If the enolate approaches via 427, the steric impedance of the methyl group destabilizes that transition state relative to 428. In both 427 and 428, a Cram orientation is assumed (see above) although other rotamers are possible. The appropriate rotamer for reaction therefore is that where Rl is anti to the carbonyl oxygen. Since the phenyl group is Rl, 427 and 428 are assumed to be the appropriate orientation for the aldehyde. If an aldehyde or ketone follows anti-Cram selectivity, this aldehyde orientation must be adjusted. 

The models become more complex when they take the structure of the base into account. A simple and very popular hypothesis was proposed for esters by Ireland and coworkers in pioneering work23. This model supposes that a monomeric LDA is the active species and that the lithium-carbonyl interaction leads to a six-membered cyclic Zimmerman-Traxler chair-like transition state24, at which a more-or-less concerted proton transfer occurs. The resulting preference for the E enolate observed in THF and the Z preference in THF-HMPA mixtures, an issue discussed in more detail below, could even be accounted through steric considerations (Scheme 4). 

The observed stereoselectivity in the Evans aldol reaction can be explained by the ZImmerman-Traxler transition state model. There are eight possible transition states, four of which would lead to the anti aldol product. These, however, are disfavored due to the presence of unfavorable 1,3-diaxial interactions (not depicted below). The possible transition states leading to the syn aldol product are shown below. The preferred transition state leading to the product is transition state A, where the dipoles of the enolate oxygen and the carbonyl group are opposed, and there is the least number of unfavored steric interactions. 

A cyclic transition state model, that differs from the Zimmerman-Traxler and the related cyclic models inasmuch as it does not incorporate the metal in a chelate, has been proposed by Mulzer and coworkers [78] It has been developed as a rationale for the observation that, in the aldol addition of the dianion of phenylacetic acid 152, the high ti-selectivity is reached with naked enolate anions (e.g., with the additive 18-crown-6). Thus, it was postulated that the approach of the enolate to the aldehyde is dominated by an interaction of the enolate HOMO and the n orbital of the aldehyde that functions as the LUMO (Scheme 4.31), the phenyl substituents of the enolate (phenyl) and the residue R of the aldehyde being oriented in anti position at the forming carbon bond, so that the steric repulsion in the transition state 153 is minimized. Mulzer s frontier molecular orbital-inspired approach reminds of a 1,3-dipolar cycloaddition. However, the corresponding cycloadduct 154 does not form, because of the weakness of the oxygen-oxygen bond. Instead, the doubly metallated aldol adduct 155 results. Anh and coworkers also emphasized the frontier orbital interactions as being essential for the stereochemical outcome of the aldol reaction [79]. 

---