Allylation in its simplest form is the addition of the allyl group, -CH2-CH=CH2, to an organic molecule. The originally sp3 carbon is labeled the alpha carbon, and the far end of the double bond is thus the gamma carbon. That adjacent double bond is a nice charge stabilizer, so it’s possible to envision both electrophilic and nucleophilic allylation reactions. Things get hairy very quickly, however, when you start considering all the various substitution patterns of the allyl group. The perfect allylation sequence combines flawless addition stereochemistry with the ideal regiochemistry of the “allyl-delivering” reagent. It all adds up to one hell of a synthetic and mechanistic challenge.
Allylmetal reagents are the quintessential allylic nucleophiles, and represent the “allyl-delivering” component of a huge percentage of allylation reactions. What metal should be used depends on the electrophilic partner, functional group tolerance, desired reactivity, etc. Ionic allylmetals such as allyllithiums often blur the distinction between the alpha and gamma carbons, as metal coordination is usually a three-atom, pi-cloud affair. Methods were developed in the mid-80’s for regioselective allyllithium additions to aldehydes; these started with allyl phenylsulfides and used some very interesting lithium aryl reagents (Tetrahedron 42, 2803).
Allyl Grignards, zincs, and calciums are mostly covalent and the metal-carbon bond ends up in the alpha position, but new bonds are usually formed at the gamma carbon (electrons in the C-M bond move into a new pi bond, and the old pi bond is “pushed” into acting as the true nucleophile). If alpha-reactivity is what you want, it turns out allylbariums are the way to go. Who saw that coming? Barium can also be used for the coupling of two allylic halides with great regio- and stereoselectivity. Unsung heroes of the periodic table…
One of the neatest aspects of allylation reactions is that they can often frame some interesting sigmatropic rearrangements. The above homoallyl-allyl alcohol was produced by treating 1-methylene-2-phenylthiocyclohexane with lithium dimethylaminonaphthalide (LDMAN, one of those weird lithium reagents from earlier) and adding crotonaldehyde. A little base gets a very nice oxy-Cope rearrangement going.
The mechanism of addition of allylsilanes is worth mentioning. These additions are almost always gamma, because the allylic double bond acts as the nucleophile. It doesn’t need to be “pushed,” but additions to carbonyls require the use of a Lewis superacid like protonated triflic acid to get things going. Guess you could say it needs to be “pulled.” A carbocation is produced, silicon does its weird “hyperconjugative beta stabilization” thing, and finally silicon migrates to the former carbonyl oxygen. Neat, no?
For a massive review of allylmetals, check out Chem. Rev. 93, 2207.