Now this right here is a pretty weird paper. As an organic chemist, the most elemental and magical reaction has always been the direct functionalization of a just-plain-alkane. Unactivated C-H bonds are the pennies of the science – they’re everywhere, and they aren’t worth much. Turning one into an actual functional group just seems bizarre, even though there are several reactions that do it, starting with free-radical halogenation and moving on to more strenuous reactions like carbene and nitrene insertions and (more recently) redox photochemistry. But the fundamental energetic problem remains: the C-H bond is a pretty good one, and you have to pay a substantial enthalpic cost to break it, which means that lots of other things in your molecule will generally break before the C-H does. Another complication (as the penny analogy illustrates) is that most organic molecules have quite a few C-H sites to pick from.
You can get around that last problem by making the reaction about bringing in a relatively simple alkane and getting it to react with something else, though, rather than making it about picking off particular CH bonds in a more complex framework, and that’s what this current paper does. They show that if you take (the simplest case) a benzaldehyde of some sort and a cycloalkane, you can make the cycloalkane come in like a carbanion nucleophile and add to the aldehyde. In other words, if you have (say) cyclopentane in there, you get the product that you’d get if it were cyclopentyl bromide instead and you made the Grignard reagent out of it first. You just skip having to have the bromoalkane and get right down to the organometallic-ing.
That’s accomplished by using chromium (II) chloride and copper (II) chloride and hitting the reaction with purple 390nm light. There’s methylboronic acid in there, too, to complex the resulting alcohol products and keep them from interfering with the reaction. The proposed reaction is a bit of a three-ring circus – the copper chloride produces a photochemical chloride radical that abstracts a C-H to make a cycloalkyl radical (which you can also get by another pathway). It’s a chromium (III)-cycloalkyl species in this cycle (formed from that cycloalkyl radical) that reacts with the aldehyde, but the aldehyde is involved in cycling around for more radical formation on the Cu end of things via photoexcitation, too (if you’re really into this stuff, you’ll have to check out their Figure 5; I’m not going to deal with all its points. If you leave out the copper, leave out the chromium, or leave out the purple light, you get no product.
Nitriles, esters, and thioethers all seem to survive the conditions just fine, and there are a couple of example with an N,N-disubstituted aniline where the N-methyl turns into the nucleophile, which I found interesting and would have liked to have seen explored more. In fact, the whole paper (to the eyes of this medicinal chemist) is rather nitrogen-poor: I would have like to see more N-methyls fed into the reaction, and more nitrogen heterocycles (saturated and unsaturated). Can you use N-methylpyrrolidine instead of cyclopentane, for example? Outside of nitrogens, what happens if you use tetrahydrofuran? Does the reaction still go (and off of which carbon?) or does it open the ring and do something else?
So if they’d sent this one to me for review, I would have been one of those annoying referees that liked the paper but wanted to see more experiments with more stuff. Mostly, I’d like to see this neat reaction moved out of the world of adding cycloalkanes to aldehydes, which is interesting in a mechanistic sense but not so exciting for doing real-world chemistry. This system deserves more than being a new catalytic kind of Barbier or Grignard-ish reaction, and I hope it can be.
