Karabatsos model

A further improvement of the theory of 1,2-asymmetric induction was introduced by Felkin15. Neither Cram s open-chain model nor the Karabatsos model is able to explain why the stereoselectivity increases when either the incoming nucleophile R2e or the substituent at the carbonyl group (R1) increases in bulk. To explain these experimental observations the following assumptions are made for the Felkin model ... 

Figure 6-4. The Karabatsos model for nucleophilic attack at acyclic carbonyls.

In 1968, Felkin noted that neither the Cram nor the Karabatsos models predict the outcome of nucleophilic addition to cyclohexanones [15], and fail to account for the effect of the size of R on the selectivity [16]. The point about cyclohexanones is particularly well-taken, since it is unlikely that the mechanisms of Grignard and hydride additions to cyclic and acyclic ketones differ significantly. The data in Table 4.1 indicate that as the size of the substituent on the other side increases, so does the selectivity, except for the single example where the large substituent is cyclohexyl and the carbonyl is flanked by a ter/-butyl. 

In the Karabatsos model, the medium-sized group is placed eclipsed with the carbonyl oxygen. Nucleophilic attack is again predicted to preferentially occur from the diastereotopic face that has the smaller group (Figure 10.8 B). Note that the Cram and Karabatsos methods lead to the same predictions. However, the Karabatsos model has been found to more accurately predict product ratios. 

A. Cram s model, B. Karabatsos model, and C. the Felkin-Ahn model for nucleophilic addition to a carbonyl. 

Cram model Karabatsos model Felkin model... 

In contrast to the open-chain and dipolar models, which are based on conformations of the carbonyl compound not representing energy minima, Karabatsos proposed a different model assuming an early, reactant-like transition state in which the most stable conformation of the free carbonyl compound is preserved1314. Thus, the C-M bond eclipses the carbonyl double bond and, in order to minimize the energy of the transition state, the nucleophile approaches close to the small substituent on the stereogenic center (Figure 5). 

Solvolytic experiments specifically designed to test Bartell s theory were carried out by Karabatsos et al. (1967), who were primarily interested in an assessment of the relative contributions of hyperconjugation and non-bonded interactions to secondary kinetic isotope effects. Model calculations of the (steric) isotope effect in the reaction 2- 3 were performed, as well as that in the solvolyses of acetyl chloride... 

If it is assumed that the Curtin-Hammett principle applies, one need only to compare the energies of the minima on the solid and dashed curves to be able to predict the structure of the major product. These curves also allow a direct comparison of Cram s, Cornforth s, Karabatsos s and Felkin s model for 1,2 asymmetric induction. Both Figures show the Felkin transition states lying close to the minima. The Corn-forth transition states (Fig. 3) are more than 4 kcal/mol higher and should contribute little to the formation of the final products assuming a Boltzmann distribution for the transition states, less than one molecule, out of a thousand, goes through them. Similarly, Fig. 4 shows the Cram and Karabatsos transition states to lie more than 2.7 kcal/mol above the Felkin transition states, which means that they account for less than 1% of the total yield. 

Karabatsos87 introduced a variant of Cram s model based on the following assumptions (1) the transition states for addition to carbonyls are reactant-like (2) the reactive conformations are then the most stable ones, which have a neighboring o bond eclipsing the carbonyl group (cf. p. 188) and (3) the nucleophile approaches from the less hindered side. Assumption (2) is questionable even in the reactant-like transition state, the stable and reactive conformations may be completely different. Karabatsos s model was the first to draw attention on the importance of conformational factors in asymmetric induction. 

Karabatsos, G. J. Asymmetric induction. A model for additions to carbonyls directly bonded to asymmetric carbons, J. Am. Chem. Soc. 1967, 89, 1367-1371. 

Prediction for the reduction of cyclic ketones on the basis of the Cram, Karabatsos and Felkin-Ahn models is usually unreliable and a simple model has yet to emerge. 

The stereochemistry and mechanism of reduction of cyclic ketones by metal hydride reagents provided a unique Of rtunity for comparison of experimental results with theoretical expectation. The models proposed by Cram, Comforth and Karabatsos described above were inadequate to explain the stereochemical outcome, and so a wide range of models was developed to explain the dichotomy between cyclic and acyclic results. The theoretical basis, applications and limitations of these models have been critically reviewed. The effect of steric influences, torsional and electronic factors, and the nature of the cation on the rate of reduction, stereochemical outcome and position of the transition state have also been surveyed. ... 

The predictive value of Cram s rule notwithstanding, the rationale was speculative, and as spectroscopic methods developed, it was called into question. For example, Karabatsos studied the conformations of substituted aldehydes [8] and dimethylhydrazones [9] by NMR, and concluded that one of the ligands at the a position eclipses the carbonyl. It was felt that in the addition reaction, the organometallic probably did coordinate to the carbonyl oxygen as Cram had suggested, and Karabatsos used the conformations of the dimethylhydrazone as a model for the metal-coordinated carbonyl. He concluded that since the aldehyde and the hydrazone have similar conformations, so should the metal-complexed carbonyl... 

Thus Karabatsos concluded that the rationale for Cram s rule was incorrect [10]. In 1967, he published a new model, which took into account the approach of the nucleophile from either side of all three eclipsed conformers [10]. He noted that the enthalpy and entropy of activation for Grignard or hydride additions to carbonyls are 8 to 15 kcal/mole and -20 to -40 eu, respectively. Since the barrier to rotation around the sp -sp carbon-carbon bond is much lower [12], the selectivity must arise from Curtin-Hammett kinetics [13,14]. Of the six possible conformers (Figure... 

In 1977, Anh [23] used ab initio methods to evaluate the energies of all the postulated transition structures (Figures 4.2 - 4.4) for the reaction of 2-methyl-butanal and 2-chloropropanal (the former to test the Cram, Karabatsos, and Felkin models, and the latter to test the Felkin and Comforth models). The nucleophile was H , located 1.5A from the carbonyl carbon, at a 90° angle, on each face of the carbonyl. Rotation of the C1-C2 carbon-carbon bond then provided an energy trace which included structures close to all of the previously proposed conformational models. The results for both compounds clearly showed the Felkin transition states to be the lowest energy conformers for attack on either face of the carbonyl. Inclusion of a proton or lithium ion, coordinated to the oxygen, produced similar results. It therefore appeared that Felkin s notion of attack antiperiplanar to the large substituent was correct. 

In each case, draw the transition state for the Cram, Karabatsos and Felkin-Ahn models of the reaction of each molecule with (l)MeMgBr (2) PhMgCl (3) CH3C C -Na+ (4) EtLi (5) 2-lithio-l,3-dithiane... 

A modification of the Cram model, in which the medium sized group, M, eclipsed the carbonyl oxygen, was developed by Karabatsos however, it generally predicted the same product as the Cram model. In this model, which assumes two major conformations, the major product is that which is derived from attack at the less hindered side of the more stable conformer. 

---