Diastereoselection in Aldol Reactions

DIASTEREOSELECTION IN ALDOL REACTIONS

The reaction of enolates with aldehydes or ketones to produce β-hydroxy car-bonyl derivatives is a very common and a very useful way to make carbon–carbon bonds. A fundamental stereochemical feature of the reaction is that two new chiral centers are produced from achiral starting materials. Hence syn and anti diastereomers will be produced, each as a pair of enantiomers. This is shown schematically for the reaction of a propionate enolate with isobutyraldehyde. Because they have different energies, the syn and anti diastereomers will be produced in unequal amounts, but each will be produced in racemic form because both starting materials are achiral.

The diastereoselectivity of the reaction results from a combination of three factors. First the carbonyl electrophile can undergo addition on either its Re or Si face. Second, the enolate nucleophile is planar and can attack the carbonyl group from either of its faces. Third, the enolate geometry can be either Z or E. To control the diastereoselectivity, it is first necessary to use a single isomer of the enolate. In general, the E enolate is the kinetic enolate and the Z enolate is thermodynamically favored. Methods are available to produce either as the major isomer by α-proton removal from carbonyl compounds with strong bases. This is particularly true of esters and amides. Pure Z and E enolates can also be prepared by first converting the carbonyl compound to a Z and E mixture of silyl enol ethers, separating these isomers, and regenerating the Z and E enolates with methyl lithium. Suffice it to say that there are known ways to produce either Z or E enolates in pure form.

The stereoelectronic requirements for carbonyl addition are that electron dona-tion occurs by interaction of the donor with the π ^∗ orbital of the carbonyl group. To meet the stereoelectronic requirements and explain the diastereoselectivity, the Zimmerman–Traxler model is used. Interaction of the lithium cation with the oxygen of the enolate and of the carbonyl electrophile leads to a six-membered chairlike transition state. If the geometry of the enolate is fixed, the only variable is the orientation of the electrophile. The preferred orientation has the larger sub-stituent in a pseudoequatorial position. This preferred orientation produces the major diastereomer. An example is shown for the Z enolate of ethyl propionate reacting with isobutyraldehyde, which predicts that the anti diastereomer should be favored (and it is!). A similar analysis predicts that the E enolate should give the syn diastereomer as the major product (and it does!).

This model is extremely useful in understanding the stereochemical outcomes of aldol processes. It also provides a framework for influencing the diastereose-lectivity in a rational way. For instance, if the ethoxy group in the above example is changed to a much bulkier group, increased transannular interactions in the pseudoaxial transition state would make it even higher in energy and result in increased selectivity for the anti isomer (and it does!)

Even greater diastereoselectivity in the aldol reaction can be achieved using boron enolates as the carbon nucleophile. Boron enolates are easily prepared from aldehydes and ketones, and the syn and the anti isomers can be separated as pure compounds. They react with aldehydes and ketones to give aldol products by a similar transition state. The difference is that boron oxygen bonds are shorter than lithium oxygen bonds, and thus steric interactions in the transition state are magnified and result in greater diastereoselectivity.