The brilliant history of photoredox catalysis

Many stories of scientific discovery start with a lightbulb moment, a flash of insight that illuminates a new area of research. For photoredox chemistry, the branch of catalysis lighting up industry and academia, the story begins with an actual lightbulb.

In 2008, David MacMillan at Princeton, and his postdoc, Dave Nicewicz, were pondering a specific synthetic organic chemistry problem: how to asymmetrically form a new carbon-carbon bond adjacent to an aldehyde. This long-sought transformation would be invaluable for efficiently assembling pharmaceuticals, agrochemicals, and other complex organic molecules.

The team had the reaction working using high-energy UV light. But UV photoreactors are expensive, complex, and not widely embraced by organic chemists.

“When I asked Dave if he could think of a way of doing this reaction without UV light, I didn’t mean use low energy light,” MacMillan recalls. But Nicewicz found a way to drive the reaction using a household compact fluorescent lightbulb. The trick was to use a visible-light-absorbing ruthenium bipyridyl catalyst, which was already a hot topic in solar cell research.

Remarkably, Tehshik Yoon at the University of Wisconsin-Madison, and Corey Stephenson at the University of Michigan both independently had the same lightbulb moment. In late 2008 and early 2009, all three researchers^1,2,3 separately published papers using visible light photoredox catalysis, which solved completely different synthetic problems.

A spotlight had been shone on a dark corner of synthetic chemistry. Scattered references to organic chemists using photoredox catalysts can be found as far back as the 1970s, but until 2009 it was an area most eschewed. “There was this idea you couldn’t control these really reactive intermediates,” Yoon says, “and that nobody would want to use a light source to power a reaction.”

This time, with three groups near-simultaneously publishing diverse, powerful, and practical examples of what photoredox could do, “[It] really helped ignite the whole thing,” MacMillan says. “You could immediately start to see all these different directions you could take this chemistry,” he says.

Electrons are the chemical glue that hold organic molecules together. When excited by light, photoredox catalysts can give and take single electrons, catalysing the creation of new bonds that previously needed multi-step procedures to make. “It’s not just the fact you can make molecules faster; you can make all these molecules you wouldn’t even dream of making before, and you can do it in one step,” MacMillan says.

Even so, running light-driven reactions will always be more complex and costly than a conventional reaction, says Daniel DiRocco, a principal process chemist at MSD. “For us to consider utilising something like that, it had to revolutionise the way we can think about putting the molecule together,” he says.

Photoredox offers exactly that. In one early example DiRocco worked on, the fluorination of an advanced pharmaceutical intermediate (API) was reduced from a multi-step transformation to a single step using photoredox.⁴. “It completely changed the way you could think about putting that API together,” DiRocco says, and it drove MSD’s uptake of photoredox.

Meanwhile, as academic labs cottoned on to photoredox and began applying it in myriad new ways, an exponential rise in research papers published on organic photoredox catalysis ensued. Following the initial three papers, tens of papers were being published each year by 2011. By 2014 it was hundreds⁵. In one particularly fruitful branch of research, academics showed photoredox could be combined with other types of catalysis to spectacular effect. Yoon recalls back-to-back papers in Science in 2014 ^6,7, one co-led by MacMillan and the other by Gary Molander at the University of Pennsylvania, that paired photoredox with transition metal catalysis for “very difficult cross-coupling reactions”. According to Yoon, “Those two papers were absolutely astonishing. The power of these reactions is what the pharma industry really jumped on.”

Despite its increasingly obvious synthetic potential, photoredox catalysis faced two barriers to widespread adoption. Most chemists are inexperienced with photochemistry. “People were afraid of how to put together a small reactor with an LED strip,” DiRocco recalls. Also, iridium and ruthenium catalysts can be hard to find. “In the first few years, you couldn’t buy any of them,” DiRocco says.

“The synthesis of some photoredox catalysts remains very challenging,” says John Nguyen from Merck⁴. He should know: Nguyen was a graduate student in Stephenson’s group when the company first recognised the value of supplying photoredox catalysts to the synthesis community and approached Stephenson’s group for assistance to produce them. Nguyen synthesized the first batches of photoredox catalyst ever sold by Merck.

Suppliers have undoubtedly removed the hurdle of photoredox catalyst availability. “An organic chemist is probably not going to spend a few days doing organometallic chemistry to make this new complex just to try it,” says Ben Glasspoole, head of Emerging Chemical Synthesis at Merck. “But if you can click your mouse three times and have it arrive in your lab you are much more likely to try it.”

The same has happened with the photoreactors, custom-made lab equipment for photochemical reactions. Chemists put off by having to jerry rig their own photochemical set up, or having difficulty reproducing a published procedure, could now simply order in a standardised, purpose-built reactor. MacMillan worked with MSD to develop a standardized single-well photoreactor⁸. Nicewicz set up a company to design and manufacture a 16-well reactor, to run multiple reactions in parallel, which came on to the market in mid-2018. “Initial feedback is that people really like it,” Nicewicz says.

“Every pharmaceutical company I visit is now doing photoredox in medicinal chemistry. There are thousands of drug-like compounds being generated this way,” MacMillan says. The next milestone will be when photoredox is used at process scale to produce a marketed drug. Due to issues of light penetration, photoredox cannot be carried out in very large flasks, so process chemists are investigating technologies such as flow reactors. It is an engineering problem that will soon be overcome, DiRocco says. “I don’t think it is that much of a hurdle.”

The cost of precious metal photoredox catalysts is also a potential issue at scale, although the metal-free photoredox catalysts Nicewicz’s lab has developed could help. Organic photoredox catalysts can offer complementary reactivity to the metal complexes, and are now sold as part of many suppliers’ photoredox catalyst suite. But they can also be designed to act as low cost, high-performance replacements for the metal complexes, Nicewicz and DiRocco have shown⁹. “I think there will be an equivalent organic catalyst that can do what the metal bipyridal complexes can do,” Nicewicz says.

In academia, interest in photoredox shows no sign of waning. Combining photoredox with organometallic and other types of catalysis still has high potential for new discoveries, MacMillan and Yoon say.

The next frontier MacMillan and others are beginning to explore is in chemical biology. “Single electron processes can work in cells, you can get light into cells. It’s an area we are really excited about,” MacMillan says. “I think it is the best stuff we have done in 10 years.

“I see photoredox amplifying to get even bigger,” he adds. “The opportunities are vast.”

To learn more about how Merck continues to enable advances in photoredox catalysis, visit: www.sigmaaldrich.com/chemistry/chemical-synthesis/photocatalysis.html