I recently heard about Cram's rule for predicting the stereoselectivity in nucleophilic additions to carbonyl groups. Can anybody explain what this is and whether it is an accurate rule?


Cram's rule, proposed by Donald Cram in 1952,[1] was the earliest model proposed for rationalising the stereoselectivity of nucleophilic additions to carbonyl groups with α-stereocentres. The model involved assigning each α-substituent a relative "size" (small, medium, and large), then placing the carbonyl oxygen antiperiplanar to the largest of these three groups. The nucleophile then attacks the carbonyl group opposite the larger of the two remaining groups (i.e., the medium group). This is best explained with a diagram:

Reduction of (S)-3-phenylbutan-2-one with L-Selectride

In the reduction of (S)-3-phenylbutan-2-one with L-Selectride (a bulky source of hydride ion), the anti alcohol is formed as the major diastereomer. Cram's rule rationalises this by placing the phenyl group (RL) anti to the carbonyl oxygen. The hydride nucleophile, drawn here as H for simplicity, then preferentially attacks over the proton (RS) instead of the methyl group (RM).

Although Cram was able to demonstrate that his rule worked for a number of nucleophilic additions, there are some conceptual issues with it. Most obviously, in the reactive conformation, the phenyl group is eclipsing the methyl group on the ketone; this torsional strain is worsened when the nucleophile adds into the carbonyl, as the resultant product is initially in a staggered conformation.

In 1968, Hugh Felkin pointed out two observations which the Cram model could not explain.[2] One is to do with the reduction of cyclohexanones, which we will not go into here, as nowadays a slightly different model is used to explain reduction of cyclic ketones. The other was about how the stereoselectivity of C=O reduction varied when the ketone was made progressively more bulky.

Stereoselectivity in a series of ketone reductions

According to Cram's model, as the size of R increased, the steric repulsions between R and the phenyl group (RL) should increase, leading to a decrease in stereoselectivity of reduction. However, Felkin showed that the opposite was true: the anti product was formed with higher stereoselectivity when R increased in size from Me to t-Bu.

To cut a long story short, Cram's rule was eventually replaced by the Felkin–Anh model. Here, the large α-substituent RL was instead placed perpendicular to the C=O bond. This leads to a choice of two possible conformations. The nucleophile attacks the carbonyl group along the Bürgi–Dunitz angle, and the attack over RS instead of over RM is favoured:

Felkin–Anh model

Like Cram's rule, the Felkin–Anh model correctly predicts the diastereoselectivity observed in this reduction. However, it also allows us to rationalise why the stereoselectivity increases as the size of R increases. A larger R makes the gauche interactions between R and Ph more destabilising, which biases the conformational equilibrium towards the reactive conformer (on the right).

For more detail about the Felkin–Anh rule and other modifications which are necessary (depending on the exact substrate and conditions), see: Effect of protecting group on diastereoselectivity of LiAlH4 reduction of α-hydroxyketones and the answers therein.


  1. Cram, D. J.; Elhafez, F. A. A. Studies in Stereochemistry. X. The Rule of “Steric Control of Asymmetric Induction” in the Syntheses of Acyclic Systems. J. Am. Chem. Soc. 1952, 74 (23), 5828–5835. DOI: 10.1021/ja01143a007.
  2. Chérest, M.; Felkin, H.; Prudent, N. Torsional strain involving partial bonds. The stereochemistry of the lithium aluminium hydride reduction of some simple open-chain ketones. Tetrahedron Lett. 1968, 9 (18), 2199–2204. DOI: 10.1016/S0040-4039(00)89719-1.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.