I've written several times about how important metal-catalyzed coupling reactions are to organic synthesis - they're the single biggest change since my grad school days in the 1980s, when they were considered sort of squirrely and exotic. Now they're everywhere, and the literature on them is almost beyond counting.
A lot of work gets done trying to extend these reactions to starting materials that are more easily available but don't tend to work as well, to make the catalysts cheaper and more robust, and to find replacements for the palladium that's so often at the center of things. But people have been scorched in the attempt - several "palladium-free" couplings using other metals have turned out to be actually catalyzed by ridiculous trace amounts of palladium contamination instead.
Now there's a paper in JACS that's getting a lot of attention, and a lot of raised eyebrows. The authors claim that they can couple aryl iodides with plain unfunctionalized aromatic compounds with either amines or alcohol as catalysts - and no transition metals at all - just potassium t-butoxide as base. Organic chemists will recognize that this is a very unusual reaction indeed, since carbon-carbon bonds between aryl groups are not supposed to be so easy to form. This reaction, in fact, would suggest that a lot of the palladium-catalyzed work is some sort of odd detour to get to a process that happens fairly easily anyway.
But that doesn't seem right, somehow. The mechanism for the metal-catalyzed reactions is pretty well worked out (in its broad strokes, anyway), and the metal really is crucial. How can these things be going? The authors suggest that since they're using iodides that a free radical mechanism is operating. Addition of radical scavengers, they say, shuts the reaction down. And while it's true that iodides are great radical precursors, these couplings seem too clean for that mechanism - unless you take care to give them limited opportunities, free radicals tend to react with most everything in sight. (The fact that they don't tend to get regioisomers rules out another possible mechanism through benzyne intermediates).
The other problem I have with that is that potassium t-butoxide is not the sort of thing one generally needs in a radical reaction, although they are proposing a radical anion. Lithium and sodium t-butoxide don't work, interestingly, and I'm not sure what to make of that, either - these sorts of counterion effects can certainly be real (I've seen some myself), but they do call for an explanation.
And what's more, just this morning I've heard from a reader, an experienced chemist in a good lab, who's tried to reproduce this work and (so far) failed. I'd be very interested in hearing from others who've taken a crack at it, too. Real or not? Let's find out.