Photochemistry’s rise over the last ten years or so has been one of the big stories in organic chemistry, but there are still some difficulties with using it. The use of photoredox catalysts has brought blue light into a lot of fume hoods, which is certainly more selective and easier to use than than old ultraviolet sources – I will not miss mercury lamps, I can tell you. But the physical problem remains of getting enough photons poured into the reaction. It’s been demonstrated that with any of these sources that you only tend to get photoreactions in the outer few millimeters of the solution. Thus the flow photochemistry rigs that people have used, which are attempts to turn the entire reaction into a 2mm-deep layer around the light source – otherwise, you have to stir vigorously and just keep beaming away until things are finally done.
But physics to the rescue: there are other possibilities. Here’s a paper on one, triplet upconversion. Near-infrared light has far better ability to penetrate solutions (and tissue, for that matter). But taking advantage of that isn’t always easy. One of the reasons it’s penetrating is that it’s not be absorbed by anything along the way, and even when it is, light of those wavelengths isn’t packing enough energy to do a lot of work. It’s definitely not going to break or form any bonds by itself – absorption at those frequencies all goes into things like rotational or vibrational energy, so you’re just going to warm things up. And it’s not enough to send any of your favorite photoredox catalysts into an electronically excited state, either.
Triplet upconversion, though, is a way of pooling this energy and making something out of it. A sensitizer species is chosen that can usefully absorb the NIR light to provide a singlet excited state, and this decays to a triplet form. That interacts with another species (the “annihilator”), turning it into an excited triplet, and then two of those react with each other. One falls back to the ground state, and the other ends up boosted to yet a higher energy state, and then emits a photon with some oomph to it – you’ve upconverted near-IR light into the visible or even UV range by doing a two-for-one exchange. This is a big topic in many fields, not least solar energy production.
You have to pick your species carefully – the sensitizer, the annihilator, and eventual photoredox catalyst that absorbs that last photon – but what’s organic chemistry for if we can’t tune the properties of our molecules by changing their structures? This latest paper has gotten things to line up with a palladium octabutoxyphthalocyanine complex that absorbs down at 730nM, furanyldiketopyrrolopyrrole as the annhilator, and eosin as the eventual catalyst. Eosin is known to (for example) catalyze dehalogenation under blue-light conditions, and in this system it’ll do so under near-IR illumination. But you get comparable yields with the triplet-upconversion system while using a light source that’s only one-thousandth the power (!)
That’s because, physically, this would seem to be equivalent to setting off photochemical illumination sources, molecule-by-molecule, throughout most of the solution. Absorption tests showed that the NIR light was penetrating typical reaction solutions hundreds of times better than blue LED light, so that’s quite an improvement over trying to beat in the photons from the outside. Switching to a different (tetraphenyltetranaphthoporphyrin) palladium complex and a tetra-t-butylperylene annihilator (which emits up into the blue range), they were able to use the popular Ru(bpy)3 catalyst. Interestingly, in one of the systems tried (a cyclobutane-forming reaction), the catalyzed system gave a 48% yield, while leaving out the Ru catalyst still gave a 38% one, so it appears that the excited singlet perylene species is enough by itself to get the reaction to go in some cases.
This idea would seem to have real applications for photochemical scaleup. The higher penetration is a great feature, as is the lower power consumption, and the excess heat of whole setup would be a lot easier to deal with (it’s a major problem as you go to larger rigs). Getting your photons this way might be the photochemistry of the future, if enough good absorber/annihilator pairs can be identified.