Reaction scheme

My book writes that:

Recall that epoxides rearrange with Lewis acids in a pinacol-like fashion, and that Grignard reagents in THF exist in the following equilibrium:

Schlenk equilibrium for generic Grignard

I approached this question by reasoning that the Grignard reagent would lead to opening of the epoxide:

Proposed reaction outcome

but the answer given in the book is instead

(Incorrect) book solution

How is the five-membered ring formed?

  • 1
    $\begingroup$ I think in this case the question is asking you to think of the Grignard rearranging to give a Lewis acid that induces a rearrangement of the epoxide $\endgroup$
    – Waylander
    Jan 22, 2019 at 14:16
  • 4
    $\begingroup$ I'm calling BS on the provided answer. The semi-pinacol rearrangement seems reasonable, but that means you're going to end up with an aldehyde in the presence of methyl Grignard. I have trouble believing that it will stay an aldehyde. In some sense, the problem suggests that as well. Why else would one need to oxidize with PCC at all? $\endgroup$
    – Zhe
    Jan 22, 2019 at 14:47
  • $\begingroup$ Perhaps the carbonyl remains bound to the LA-like Mg species which deactivates it for methyl Grignard attack - but I do rather agree with you. Odd choice of reduction conditions too. Has someone database access to find this reaction? $\endgroup$
    – Waylander
    Jan 22, 2019 at 15:02
  • 1
    $\begingroup$ @Waylander I don't like the binding argument because such a bound species would putatively be present during the additional of methyl Grignard to an aldehyde. On the other hand, if the Grignard were used in catalytic amounts (not explicitly specified in reaction description), then it would be possible not to obtain the product with methyl addition, except sacrificially to generate the true active Mg catalyst. $\endgroup$
    – Zhe
    Jan 22, 2019 at 15:59
  • 1
    $\begingroup$ @Chakravarthy Kalyan: Acyclic epoxides rearrange with Lewis acids [DOI:10.1021/jacs.5b10419] Cuprates, with or without a Lewis acid, react in the "normal" SN2 fashion. That is, without rearrangement. J. F. Normant, et al., Tetrahedron, (1986), 42, 5607. $\endgroup$
    – user55119
    Jan 24, 2019 at 17:23

1 Answer 1


Sara Jane: Lest you miss the point of the stimulating discussion between Waylander and Zhe, I would like to sum up their thoughts with a diagram. Your first clue should have been the opening line: "Recall that epoxides rearrange with Lewis acid...". It is common that catalytic Lewis acids rearrange 2,3-dialkyl oxiranes [epoxides] to ketones. In the case of cyclohexane epoxide (1), rearrangement to ring-contracted cyclopentane carboxaldehyde (4) occurs. In the illustration I have used CH3MgBr as the Lewis acid. Alternatively, one might use the other components of the Schlenk equilibrium, i.e., (CH3)2Mg or MgBr2. The rearrangement is shown as two separate steps (2 and 3) although it is likely to be a concerted process.

The aldehyde 4 in the presence of the Grignard reagent will undergo addition. Given sufficient Grignard reagent, addition to the carbonyl group will occur to afford the secondary alcohol 6. The only ways that the aldehyde 4 can be isolated is if there is insufficient methyl magnesium bromide or that the Grignard reagent acts as a base and forms the unreactive enolate of the aldehyde. The aldehyde 4 would be produced upon aqueous workup. While Grignard reagents with β-hydrogens, e.g., isopropyl and ethyl Grignard reagents, are capable of reduction of aldehydes and ketones, methyl Grignard reagents are incapable of this reaction. Therefore, the primary alcohol 9 may be assumed to be a misprint in place of secondary alcohol 6.

The oxidation of alcohol 6 to ketone 7 is straightforward. The final step is the formation of the alkene 8 (most likely of the (E)-configuration) via the McMurry coupling.

enter image description here


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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