Skip to Content

Chemical News

Photoreactors For All

I’ve been hearing about this for a while now, so I’m glad to link to the manuscript (open access, now with a working link!) on a new photoreactor that should find wide use in organic synthesis. It was developed by the MacMillan group at Princeton and researchers at Merck, and it’s now commercially available. It has higher-power LEDs than most homebrew setups, and the reaction chamber is optimized for exposure, so in many reactions it leads to shorter times and higher yields. Some of the differences are dramatic – 3 minutes versus an hour, 20 minutes versus 3 hours, etc. A variety of transformations are shown to improve: N-arylations, alkyl decarboxylative coupling, fluorination, trifluoromethylation, and more.

I think that this is just what the photoredox field needs – an inexpensive, standardized piece of apparatus that people can just throw reactions into and not worry about. This should lead to more experimentation and reaction discovery, and there’s a lot to discover. It’s a young area, so there are still things being worked out: for example, one of the reactions (cross-electrophile coupling of alkyl halides and aryl halides) actually seemed to work more poorly in the new reactor. But the team found that the problem was that the reaction was generating HBr faster than the inorganic base (sodium carbonate) could neutralize it. Switching to a soluble base (lutidine) and turning down the intensity of the lamps led to a faster, higher-yielding reaction than the literature comparisons.

The ACS Pharma Roundtable companies have been evaluating this system, and have endorsed it as a standard to improve reproducibility of published procedures. Penn Optical Coatings is commercializing the device (and is taking orders for it), and I’m told that Sigma-Aldrich will also be selling the units. If you haven’t run one of these photoredox transformations, try it out and get ready to make bonds in ways you never have before.


34 comments on “Photoreactors For All”

  1. me says:

    “The authors declare no competing financial interest.”

    1. SimonL says:

      We use Hepatochem device with Kessil H150-Blue. It’s giving us reliable data and shorten the reaction time form 24 hours to 30 minutes.

  2. 🐀 💡 says:

    But how can I fit my rats into the photoreactor?

    1. CAprof says:


      1. Anonymous says:

        I actually attended an NSF funded, week long program in high school where one of the lecturers put a live mouse in a blender. I think it was supposed to be about entropy but I’ve never trusted biologists since then.

        1. Nick K says:

          Urgh, that really spoiled my evening…

        2. CR says:

          And then turned the blender on? How does that pass IACUC? Not sure I believe a live mouse went into a blender and it was turned on story.

          1. Anonymous says:

            University of Washington, December 27, 1962.

          2. CR says:

            1962, say no more. I’m sure the presenter was smoking as well.

          3. Design Monkey says:

            Heck, in 1962 presenter quite possibly was not only smoking, but also popping OTC benzedrine pills by half a bottle. Totally legal.

  3. Anon says:

    Yay, now we can study the effects of ortho vs meta vs para substituents on any given photoredox reaction in a streamlined fashion and in scale. Beware DMac, now it will be competition for the Science papers.

    1. Truthsayer says:

      Well, even if you beat him to a Science-worthy project, he will pretend that he was first. Not sure these reactors will help for that.

      Feel bad for the junior faculty and students.

      1. Mufasa says:

        Everything the blue light touches is his kingdom, so it doesn’t count as stealing

      2. MLCT says:

        Well to be fair he and nicewise did invent the field…

  4. CR says:

    Out of curiosity, anyone know what price it’s selling for?

  5. ScientistSailor says:

    The link is broken

    1. Derek Lowe says:

      Fixed, thanks! Interesting that it took until 5:30 in the afternoon for someone to notice that, though (!)

      1. ScientistSailor says:

        It’s only 2:30 in the afternoon…

      2. Mark Thorson says:

        Now I get the no cookie error. That’s too bad, I won’t lower security just to look at that. I’d like to know what LEDs it’s using — what wavelength and how much power. I suppose the ideal machine would have several types, so you could try hitting different absorption bands, e.g. to promote specific reactions vs. side reactions or to improve efficiency. If you only had a few types, I suppose you’d concentrate on blue, UV, and maybe red, because you can get efficient high-power LEDs at those wavelengths — green and yellow not so good. UV and blue would be more likely to be absorbed efficiently by most reactants and catalysts, though they also wouldn’t penetrate the reaction medium as far. That brings up another question — how thick is the layer of reactants? That’s a problem in photobioreactors for growing algae. In a dense culture, light only penetrates about a millimeter. The usual solution is to keep the growth medium moving and turbulent, so every cell gets its time in the sun.

        Just poking around in the patent literature, I don’t see anything that looks like this work. There are a bunch of patents from a German group on a reactor which divides the reactant flow among capillary tubes which are irradiated. I suppose that’s one way to do it, but how do you clean something like that? Maybe they plan on selling you a fresh set of capillary tubes for every reaction. That’s the way to make money on these machines. 🙂

  6. Bern Yee Iris says:

    Hope they have covers and no stray light peaks out. Blue light is toxic and the targets are your eyes first, skin second, and brain last.

    1. Mark Thorson says:

      Your cellphone will get you first.

    2. Another blue light special says:

      If you’re doing that much photoredox, the brain is already gone any way…

      1. Bern Yee Iris says:


    3. b says:

      If you look at DMac’s group website, they list recommended orange safety glasses to purchase.

    4. Derek Lowe says:

      There are indeed orange covers for the device. Nice user name!

    5. Chris Phoenix says:

      This reactor uses 4 LED’s of about 1.1 watt apiece.

      Blue light exposure in the tens of watts is readily available in any room well-lit by “white” LEDs. White LEDs are blue LEDs with a partially transparent orange phosphor.

      You wouldn’t want to look directly at the light-emitting surface of a high-intensity LED. But unless I’m missing something, diffused light from this thing should be more innocuous than a modern LED TV or an ordinary LED lightbulb.

      1. Morten G says:

        Doesn’t the number of photons that hit your retina depend on the size of the pupil?

  7. ayrfield says:

    LED’s have been a massive boon for synthetic photochemistry – high intensity and efficiency, low waste heat – but the real breakthrough will be when UV LEDs become affordable. Most of the work I did for my PhD was with reactions in the the 300 nm – 350 nm range.

    1. Quantum says:

      Wavelength distribution is fairly narrow with LEDs, which is also a huge advantage over Xe light bulb setups and the like.

      I’m missing proper actinometry in the photoreactor assessment, by the way. Without fully characterizing the wavelengths and number of photons that hit the sample there can’t be a proper standartization.

  8. Sergey says:

    We’ve reported a simpler design more than a year ago:
    This one can be really built by everyone in half a day.

    1. Ir(WTF)bpy says:

      Yeah but this was in Central Science ™ so it’s obviously much better

    2. John says:

      Hi Sergey, away from my computer, but are the blueprints for the 3d printer in the si as an electronic attachment? This looks like exactly what I am looking for.

  9. Wallace Grommet says:

    Regarding UW labaratory practices, irregularities, to put it mildly, continue to this day….

Comments are closed.