C–H bond breakers seek smarter tools

In graduate school, John F. Hartwig decided to study something that many at the time thought was a solved problem.

It was 1986, and he had just joined the lab of Robert Bergman at the University of California, Berkeley. Four years earlier, Bergman had published one of the first examples of a metal activating a carbon-hydrogen bond on another molecule. In this reaction, an iridium-containing compound broke the C–H bond on an alkane and then formed a new bond with the carbon. The reaction showed it was possible to turn a relatively inert C–H bond directly into a carbon-metal bond. Bergman thought that the more reactive carbon-metal bond could allow chemists to add new functional groups to the alkane. Before this paper, chemists had seen this kind of C–H activation only with bonds in the same molecule as the metal, a reaction that has less synthetic potential than the intermolecular reaction Bergman reported.

Bergman had discussed beginning a C–H activation project with Hartwig. Bergman’s group was working on other projects as well, and Hartwig’s new lab mates pushed him to pick one of those instead. “They were saying, ‘You don’t want to work on C–H activation, because that’s already been done,’ ” Hartwig says. Some people thought Bergman and others had solved the problem. What’s more, many chemists saw C–H activation as having limited applications beyond reactions involving simple hydrocarbons, like turning methane into methanol. “People had no idea what the potential was for C–H activation in synthetic chemistry,” Hartwig says.

That potential is now clear. C–H activation has become a widely used synthetic method. In the past decade, the number of papers on the topic has increased by 230%, from 362 papers in 2010 to 1,196 in 2020, according to CAS, a division of the American Chemical Society, which publishes C&EN.

Chemists have realized that the C–H bond, ubiquitous in organic molecules, can be used as a handle to add new functional groups and ultimately build complex molecules in ways scientists hadn’t imagined. “In one step, in an afternoon, you could put a group in a position on a molecule that was inaccessible before,” says Hartwig, who is now an organometallic chemist at UC Berkeley. As a result, C–H activation methods have changed the way chemists plan their syntheses, including how they make pharmaceutical and agrochemical compounds.

Today the field has reached middle age. “In some regard, the field is not young anymore,” says Vy Dong, an organometallic chemist at UC Irvine. “There’s a lot of history and a lot of players now.”

Many in the field see a need for fresh ideas and new approaches. For example, chemists have long realized that not all C–H bonds are the same—some are easy to break; others, not so much. And researchers still haven’t determined how to activate nearly all the C–H bonds they want to reach. They want to build a toolbox of chemistries to allow researchers to select the right approach for the specific bond they want to attack in a molecule.

To fill this toolbox, some chemists are looking at new ways to target specific C–H bonds, such as using enzymes, and trying to improve older methods, such as photochemical and radical chemistries. Meanwhile, other groups have set their sights on some of the hardest-to-bust C–H bonds, hoping to find ways to functionalize them and control the stereochemistry of the products. And the work isn’t just in academic labs. Industrial and medicinal chemists have taken notice of the wide scope of C–H activation chemistry and are looking for ways to adapt it from small lab scales to larger industrial ones.

“It’s not one reaction; it’s a whole field of science,” says Jin-Quan Yu, an organic chemist at Scripps Research in California. He says C–H activation contains a rich set of problems to solve, with implications not just in synthesis but also in other fields, such as energy and materials research.

The (not so) humble C–H bond

If you took organic chemistry in college, you may remember not thinking much about reactions with C–H bonds. That’s because once upon a time, scientists thought that C–H bonds didn’t really do any chemistry. “You’re taught at the beginning that a C–H bond is really inert,” UC Irvine’s Dong says. “It’s not really considered a functional group; it’s not reactive.”

Chemists now know this explanation is an oversimplification. An entire genre of organic chemistry is devoted to developing reactions that target these bonds. These reactions can be described in one of two ways: C–H bond activation, which is when the bond breaks to form a new metal-carbon bond, and C–H functionalization, which is when a new functional group is added to the targeted carbon.

C–H bond activation can happen through oxidative addition when the metal is oxidized as it inserts itself into the C–H bond or via electrophilic metalation when the metal’s oxidation state doesn’t change. Both these pathways often rely on transition-metal catalysts, largely iridium for the oxidative additions and palladium for the metalation reactions. In terms of C–H functionalization, chemists have taken the metal-carbon bonds and added a range of different groups, such as those containing carbon, oxygen, and nitrogen.

But while chemists have found multiple ways to conquer the "inert" C–H bond, a significant challenge remains: a lot of C–H bonds look alike. In particular, they often have similar strengths. Take quinoline, for example, which has seven C–H bonds. The energy required to break its strongest C–H bond is 113 kcal/mol, while the energy to break the weakest is 104.5 kcal/mol. However, most of the C–H bonds on quinoline’s aromatic backbone have similar bond dissociation energies, all between 110.8 and 111.5 kcal/mol.

Say a chemist throws a catalyst into a pot with quinoline. Now what? “How do you selectively make one do what you want it to do” when multiple C–H bonds look so similar? Dong asks. David Nicewicz, a photochemist at the University of North Carolina at Chapel Hill, points out another challenge: “How can you activate a very strong C–H bond in the presence of a lot of other weaker C–H bonds?” That is, if you have a method that can chop down C–H bonds of a certain strength, how do you stop that chemistry from acting on all the bonds weaker than the targeted one?

That’s the challenge at the heart of C–H activation chemistry, according to many researchers in the field. “We must find a new way to identify the differences among C–H bonds,” Nicewicz says. Because of the complexity involved in activating a wide range of C–H bonds, some chemists argue that the field needs to think about building a toolbox instead of just one tool. “There isn’t going to be one catalytic method that solves every single problem and synthesis,” Nicewicz says.

Chemists want to be able to access each C–H bond in a molecule, says Shane Krska, a distinguished scientist at Merck & Co. who designs reactions for drug synthesis. For example, he says, “I want to hit that C–H on that heteroaromatic ring and then pick a different chemistry, and it’ll hit this aliphatic portion of the molecule, selectively.” And if chemists then turn to a different molecule, they need the same toolbox to give them access to every site on the new one. “You don’t want 10 screwdrivers and no hammer,” Krska says.

Finding new tools

When trying to fill the C–H activation toolbox, some chemists have turned to nature for inspiration. “It’s very hard for simple catalysts or reagents to be able to differentiate C–H bonds,” says Hans Renata, a biosynthetic chemist at Scripps Research in Florida. But, he says, enzymes can be superselective by orienting molecules in their protein pockets such that specific bonds line up to undergo a reaction. So why not steal some of these enzymes’ tricks?

Renata’s group develops enzymes that can help make potential drug compounds. They’ve focused on enzymes that add oxygen groups, such as hydroxyls, at C–H bonds, because highly oxidized natural products have proved difficult to make in the past.

To make these C–H functionalizing enzymes, the chemists first scan the genomes of microbes to look for genes for hydroxylases, which are enzymes that naturally add hydroxyls at C–H bonds. The researchers then use an iterative process called directed evolution, in which they introduce mutations to the enzyme genes and then screen for resulting hydroxylases that are particularly good at adding OH groups to specific C–H bonds. Recently, the group developed three enzymes that could take complex terpene molecules and synthesize potential drug precursors that have been difficult to access using other synthetic methods (Science, 2020, DOI: 10.1126/science.abb8271).

The researchers want to find more enzymes to add to their toolbox so that they have a library of options to oxidize different bonds in a range of molecules. “In my lab, we probably have 50 or so, but other labs have their own collections of enzymes,” Renata says. And chemists can find more enzymes in the genomes of microbes, which have their sequences published in public databases, he says. For example, Alison Narayan from the University of Michigan has used a combination of directed evolution and computational chemistry to understand how a fungus-based enzyme called TropC catalyzes the synthesis of tropolones, a class of natural product seven-membered aromatic rings with unusual electronic properties (ChemRxiv 2020, DOI: 10.26434/chemrxiv.12780044). The work appears on a preprint server and has not been peer-reviewed.

While some chemists look to nature for new C–H activation tools, others are revisiting older techniques that didn’t get much attention previously, such as photochemical and radical approaches. Chemists have long known that radicals can easily pull off hydrogens from carbons, UNC’s Nicewicz says. “But it was never to the point where it was super practical and maybe as selective as people need it to be now.”

Chemists have recently found that radicals are easier to control than people thought. Nicewicz works on photoredox catalysis, in which chemists first excite a catalyst with light, allowing it to then perform redox chemistry from its excited state. Chemists using these methods can produce catalytic amounts of radical species that are difficult to generate in other ways. They can also use cocatalysts to tune the radicals available to pluck off hydrogens. For example, chemists might be working with a catalyst that absorbs light well but has a short-lived excited state. This catalyst can react with a cocatalyst to produce a relatively long-lasting radical in solution. Because this radical is stabler than the excited catalyst, it also can be more selective and more reactive toward a specific C–H bond, Nicewicz says.

Erik Alexanian, Nicewicz’s colleague at UNC, used this photo-induced radical chemistry to synthesize the cytotoxic natural product chlorolissoclimide (J. Am. Chem. Soc. 2016, DOI: 10.1021/jacs.5b12308). This radical approach by Alexanian's team targeted a specific unsaturated C–H bond that other activation techniques had not been able to target.

In a more recent photoredox chemistry example, Tristan Lambert from Cornell University used a combination of photo- and electrochemistry to activate two neighboring C–H groups at once to make dihydroimidazole and aziridine products from simple substituted ethane molecules (Science 2021, DOI: 10.1126/science.abf2798). The reaction relied on oxidation followed by irradiation, opening up a previously unknown way to activate C–H bonds.

Hard-to-reach bonds

Another challenge with C–H activation is stretching the chemistry to access elusive C–H bonds. For example, Scripps’s Yu and his team have determined how to selectively activate C–H bonds on the hard-to-reach side of a heterocycle (Nat. Chem. 2020, DOI: 10.1038/S41557-020-0424-5). These bonds are far from any groups that might direct a catalyst or reactive molecule to target the specific site, and they don’t have enough electron density to attract positively charged metal catalysts. Yu’s team designed a palladium complex that can effectively wrap around a quinoline and, like an enzyme’s active site, hold the heterocycle so that the C–H bond they want to activate ends up next to the palladium center. Through this approach, the group was able to activate bonds selectively at two positions.

Meanwhile, Hartwig’s group at UC Berkeley reported a reaction that specifically targets the strongest, hardest-to-break C–H bond on an alkane and does so without a functional group directing catalysts or reagents to the bond (Science 2020, DOI: 10.1126/science.aba6146). The team used catalytic borylation, known to target strong primary C–H bonds, with an iridium metal catalyst. In past reactions, chemists had to use solvent amounts of the starting material to get the reaction to proceed. This limited what kinds of starting materials chemists could modify. The new catalyst allowed the team to diversify the starting materials to include solids and complex compounds.

What else do chemists want to add to their C–H activation toolboxes? Merck’s Krska points out that while several chemistries target sp² C–H bonds, there are far fewer, and less developed, tools for attacking sp³ bonds. This deficiency limits the usefulness of C–H activation in making natural products, which often have more sp³ carbons than sp² ones. Natural products are a prime target for synthetic chemists because of their potential pharmacological properties.

Krska would also like to see more C–H activation chemistries that use catalysts instead of elements within the targeted molecule to control which bond to attack. Reactions that don’t rely on specific structural features of molecules would be more versatile and more generally applicable, he says.

Nicole Goodwin, a medicinal chemist at GlaxoSmithKline, thinks that having better computational tools to predict reactivity and selectivity would help find these catalysts. “I’m really optimistic that with the work that’s being done at the interface of organic chemistry and molecular informatics . . . that there will be significant progress in this area,” she says.

Attracting industry’s interest

At first, C–H activation research belonged almost exclusively to academic chemists. Compounds that academic chemists made with C–H activation chemistry looked too simple to industrial and medicinal chemists, nothing like you’d see in the structure of a drug, Merck’s Krska says. “People in industry assumed it wouldn’t work on a complex chemical model.” But in the past decade, the field has advanced, and chemists are now synthesizing molecules that look like drugs. So chemists in industry have started to get interested in C–H activation.

For example, some chemists at companies have started to build libraries of C–H activation reactions. Researchers use these libraries of chemical transformations to map a synthetic path toward a molecule they want to make. To explore these libraries in the lab, chemists have assembled reaction kits—multiwell plates, tiny amounts of reagents, and varied reaction conditions. They provide a way for chemists to see if these reactions can make the compounds they’re interested in, without wasting the starting materials they work hard to make.

Scripps’s Yu says that 3 to 5 years ago, medicinal chemists used his group’s C–H activation chemistry mostly to make small amounts of compounds, on the scale of 1 g, to add to their libraries for high-throughput drug screening efforts. “But in the last 3 years, this has changed,” Yu says. “Now we have more emphasis on process chemistry. We’re trying to work on chemistry that can be scaled up.”

GSK’s Goodwin has observed similar changes. She’d like chemists developing C–H activation tools to consider synthesis on scale, meaning hundreds of grams or larger. Such scales need catalyst systems that are cost effective and environmentally safe. To meet both those needs, chemists should move away from the workhorse palladium catalysts in favor of less costly, less toxic metals. “We are seeing many more reports and groups investigating catalysts derived from first-row transition metals, like iron, cobalt, or nickel. And that will address those concerns,” she says. Goodwin is also excited by organocatalytic C–H activation methods that ditch the metal altogether. “And so I think we’ll continue to see more of those advances from the field in the future.”

More collaboration between industry and academic labs would fuel those efforts to keep scale-up considerations in mind when developing new C–H activation tools, Krska says. What really shot his team forward in incorporating C–H activation into synthesis planning was an academic collaboration with Mitch Smith and Robert Maleczka from Michigan State University. “Mitch and Rob’s students were coming to Merck, using our high-throughput experimentation facilities to screen reactions and to improve on the methods,” Krska says. When the industry team members saw that chemistry up close, they realized that these methods were robust, and a light bulb went off that “this is something we could envision using here at Merck,” Krska says.

The National Science Foundation has started a Center for Selective C–H Functionalization to spur more industry-academic collaborations, as well as academic-academic partnerships. Scripps’s Yu has students involved in the center and thinks that getting people to work together and share ideas is a good way to keep moving C–H activation chemistry forward.

At the start of his career, Yu studied C–H activation for the challenge, he says. “But now, I actually do believe C–H activation will transform synthetic chemistry in the not-too-distant future.”