The preparation of an allylic halide from the alcohol in a regio-and stereospecifically predictable fashion has been of concern to synthetic and mechanistic organic chemists.
Numerous methods (of varying generality) have been developed, among which are the following:
(a) reaction with conventional halide-producing reagents like SOC12l or PX3;2 (b) formation of a sulfonate ester3s4 or other reactive group5 followed by displacement with halide ion in an aprotic solvent; (c) reaction with dimethyl sulfide and an ~-halosuccinimide.6~2b~4C
The reagent system triphenylphosphine/carbon tetrachloride and its several variants' have proved very versatile for the conversion of alcohols and carboxylic acids into halides,* the dehydration of amides and oximes to nitriles,' the dihalomethylenation of carbonyl groups,1ยฐ and condensations leading to esters, amides, and peptides.l' Application of this method to the production of allylic halides seemed most promising in light of the report by Snyder-l2 that Ph,P/CC14 transforms 2-buten-l-01 exclusively into unrearranged chloride, and that only 11% of rearranged material is formed from 1-buten-3-01. Similarly high regioselectivity has been reported for other allylic alcohols in their reactions with Ph3P/CX4.13 Nevertheless, one's enthusiasm is tempered by the fact that low-boiling allylic chlorides such as 1-chloro-2-butene (bp 85' (E), bp 84" (z>) and 3-chloro-1-butene (bp 65") are only with difficulty separable from reagent CC14 (bp 77") and product CHC13 (bp 62'); similar isolation troubles have been noted by others.2gs3ca14
We felt that one could circumvent this problem and, at the same time, facilitate the reaction by letting Ph3P attack a higher-boiling halogen compound which also possesses a better leaving group than yCC13.15 Such reasoning led us naturally to an investigation of hexachloroacetone (HCA) as a substitute for CC14. We have, in fact, found that Ph3P/HCA is a very useful system for the production of allylic chlorides under very mild conditions, resulting in excellent yields, high regio-and stereoselectivity, and great ease of product purification. The recent report16 on the use of HCA as a source of positive halogen prompts us now to communicate our preliminary observations on the reactions of the uand y-methylallyl alcohols.
Reaction of an allylic alcohol in HCA with a slight excess of Ph3P is rapid (~20 min at lo-15O) and exothermic. Flash distillation at ambient temperature produces allylic chlorides as the only volatile products (accompanied by small amounts of CC14 in some cases). Significantly, the high boiling reagent HCA (bp 202O) and product pentachloroacetone (bp 192') are