Microdroplets rev up chemical reactions

Electrospray ionization (ESI) is best known as a way to introduce chemical compounds and mixtures into mass spectrometers. The machines apply a voltage to a stream of liquid containing a sample, creating an aerosol of charged, microsized droplets that then fly into the detector for analysis.

But what ESI might one day be just as well-known for is as a tool to accelerate chemical reactions, enabling scientists to test out catalysts, reagents, and conditions faster and more efficiently.

Under the right conditions, the microdroplets generated by ESI can serve as reaction vessels. Chemists don’t yet know exactly why the tiny droplets accelerate the reaction of chemical reagents contained within, but they do know the microdroplets sometimes even allow chemical transformations to proceed that wouldn’t normally occur in bulk solutions—in flasks and other larger vessels.

R. Graham Cooks’s group at Purdue University stumbled upon microdroplet reaction acceleration while trying to analyze samples with desorption ESI, or DESI, a variant of the usual technique in which a stream of electrosprayed droplets hits a surface and picks up molecules that are adsorbed there before flying into a mass spec for analysis. The researchers noticed that some molecules they were interested in didn’t show up. “They were presumably being extracted and released,” Cooks says. But they weren’t becoming charged, which meant they couldn’t be detected, he adds.

So the team decided to add a reagent to the droplets that would derivatize the molecules within, creating compounds that could more readily take on a charge and be detected. And it worked: Cooks and his team converted ketones to easily charged hydrazone versions of their molecules of interest. As the researchers monitored the reaction with DESI, they realized that the product could be observed almost immediately. This meant that the molecules were reacting far faster than they could in a flask. The researchers determined that the amount of product depended on droplet size and on the distance between where the droplets were created and the mass spectrometer (Chem. Sci. 2011, DOI: 10.1039/c0sc00416b).

Since that first reaction, Cooks’s group and Richard N. Zare’s group at Stanford University, working independently, have demonstrated that many other reactions are similarly accelerated. So far, almost every reaction they’ve tried speeds up in electrospray microdroplets relative to bulk conditions.

The most a reaction can be accelerated is approximately 10 million times the rate it would occur in a bulk solution, Cooks notes. That’s the acceleration you’d get by moving a reaction from the liquid phase to the gas phase, in which reagents don’t have to contend with solvent molecules.

Gas-phase reactions may be speedy, but they don’t produce enough product for practical applications in most cases. Microdroplets provide a compromise: Because the reactants aren’t actually in the gas phase, they don’t reach that top speed, but more product is formed.

Zare’s group has seen larger rate accelerations than Cooks’s group. In the case of the Pomeranz-Fritsch reaction, an acid-catalyzed synthesis of isoquinoline, the reaction in microdroplets was more than a million times as fast as the bulk reaction (Angew. Chem. Int. Ed. 2015, DOI: 10.1002/anie.201507805). In addition, Zare’s team detected previously unobserved intermediates.

More recently, Zare’s group used microdroplets to carry out two-phase reactions requiring both an aqueous phase and an organic phase without the need for a catalyst that helps reactants migrate between the two (Angew. Chem. Int. Ed. 2017, DOI: 10.1002/anie.201612308). The researchers introduce reagents in separate streams of aqueous and organic droplets. The reactions happen when the droplets collide and fuse, allowing the reagents to mix. Such reactions are of practical interest because they’re widely used in chemical, pharmaceutical, and polymer manufacturing. If these reactions can be accelerated, then droplet chemistry could become more useful for preparative applications. “We were able to make a few milligrams per minute of purified compound,” Zare says. “I think it can be scaled up further.”

Zare is also using droplets to further speed up reactions that have previously been accelerated by other means. For example, chemists already know that some reactions can be accelerated by “on water” chemistry that occurs at the interface between aqueous and organic solutions. Zare wanted to see whether he could speed up such reactions even more in microdroplets. Zare’s team discovered that a [2σ + 2σ + 2π] cycloaddition of diethyl azodicarboxylate and quadricyclane, previously studied by Chemistry Nobel laureate K. Barry Sharpless of Scripps Research Institute and coworkers, is further accelerated in microdroplets by a factor of 100 (Angew. Chem. Int. Ed. 2017, DOI: 10.1002/anie.201708413).

One of Zare’s microdroplet reactions might even shed light on prebiotic chemistry. Scientists have long thought that life had to have started when certain genetic building blocks came together in small, cell-like compartments. Using charged aqueous microdroplets, Zare and his team demonstrated phosphorylation of sugars, including ribose, and the subsequent formation of uridine ribonucleoside, one of the building blocks of RNA, in the absence of any enzymes—proteins typically needed to cobble together genetic material (Proc. Natl. Acad. Sci. USA 2017, DOI: 10.1073/pnas.1714896114). This reaction is thermodynamically unfavored in bulk solution; microdroplets overcome those thermodynamic barriers. Zare’s observation suggests that phosphorylation reactions on prebiotic Earth might have happened in aqueous microdroplets from sea spray.

Although researchers are achieving all sorts of reactions and increased reaction rates with microdroplets, they don’t yet know how the phenomenon works. Cooks thinks that because reactants in the microdroplets are only partially surrounded by solvent molecules at the interface between the droplet and the air, the activation energy needed for them to react is lower. Concentration effects caused by evaporation of the tiny droplets could play a part too.

Zare agrees. He thinks the fact that reactants aren’t surrounded by solvent molecules may even lead to different chemistry from what happens in bulk conditions.

For example, Zare’s group found that a Diels-Alder reaction that occurs in bulk solution doesn’t happen in microdroplets. Instead, an acid- or base-catalyzed hydrolysis occurs between the reactants, he says. That Diels-Alder reaction is one of the few identified so far that isn’t accelerated in microdroplets (Analyst 2017, DOI: 10.1039/c6an02225a).

Abraham Badu-Tawiah, a former grad student with Cooks who is now a chemistry professor at Ohio State University, uses reaction acceleration in microdroplets to test catalysts and find new reactions.

“The traditional way is to do a reaction in bulk, wait for a long time, come back tomorrow, see if anything worked, and then change it,” Badu-Tawiah says. “In this case, within a few minutes, even seconds, we know whether or not it’s working.” If something’s not working, he can quickly make changes. “We are making picomolar quantities of catalysts, and they are proving to be effective,” Badu-Tawiah says. “It speeds up the whole cycle of creating new chemicals.”

In this way, Badu-Tawiah has identified previously unknown reactions that he was then able to perform in bulk as well. For example, his group discovered a photochemically driven dehydrogenation pathway for converting tetrahydroquinolines into quinolines (Angew. Chem. Int. Ed. 2016, DOI: 10.1002/anie.201603530).

“All the reactions I’ve found with droplets can be reproduced in the bulk phase,” Badu-Tawiah says.

He acknowledges that the microdroplet experiments as they are currently done are unlikely to pin down the mechanism of acceleration. The exact size of the droplets is difficult to control, so it’s hard to measure precisely the effect of various parameters, he says. The microdroplet experiments can be used semiquantitatively to understand what is happening chemically, but he thinks getting quantitative physical models of how the acceleration works will be difficult.

Kevin R. Wilson, who studies aerosols and interfacial chemistry at Lawrence Berkeley National Laboratory, is interested in tackling the challenge anyway. The acceleration rates that Cooks, Zare, and Badu-Tawiah have been citing are “eye-popping,” Wilson says. “My group has been interested in gas-surface reactions on droplets and nanoparticles, mainly in the context of atmospheric chemistry and organic aerosol chemistry. We began to ask ourselves, ‘If this is really true, how might this acceleration impact the way we think about reactions in atmospheric or environmental compartments, aerosols, or cloud droplets?’ ”

Before the scientists can answer that question, they need to figure out the mechanism. To do that, they want to use droplets with precisely measured sizes and charges instead of the range of sizes and charges usually produced by electrospray. The best way to control the process is to react pairs of droplets together.

Wilson’s team has designed equipment that allows them to react pairs of droplets under tightly controlled conditions. In one device, a branched quadrupole trap holds a microdroplet in place while the researchers introduce a second one to react with the first (Anal. Chem. 2017, DOI: 10.1021/acs.analchem.7b03704). Another device allows them to measure reactions resulting from collisions of airborne droplets (Anal. Chem. 2017, DOI: 10.1021/acs.analchem.7b03601).

They tested the branched quadrupole trap using the reaction of an aldehyde and a primary amine in the presence of a thiol to give a fluorescent product. The quadrupole trap used 30-µm droplets, whereas the droplets used in other reaction acceleration experiments are typically smaller than 10 µm. They didn’t see much acceleration, Wilson says, because their droplets were at the upper end of the range in which acceleration has been observed.

“We think we have the apparatus to pick apart the mechanism,” Wilson says. “We just need to use smaller droplets.”

Wilson is still “agnostic” about exactly which mechanism underlies reaction acceleration in microdroplets. “Hopefully in the next few months, we can say more about that. What we really want to do is make a whole series of measurements as a function of droplet size,” he says. “Then we might be able to say very clearly how rates scale with droplet size.”

Wilson wants to apply what his group learns to atmospheric chemistry. “There’s a lot we don’t understand about chemistry in organic aerosols,” he says. Some reactions that run slowly in beaker-scale experiments might be occurring in the atmosphere at much higher rates. Those reactions could be “much more important than we would imagine looking at an organic chemistry textbook,” he says.

In addition to reaction discovery, the microdroplet approach is finding application in chemical education. At Purdue, where Cooks teaches, instructors have introduced four experiments involving electrospray synthesis to sophomore organic labs, including Claisen-Schmidt base-catalyzed condensation (J. Chem. Educ. 2014, DOI: 10.1021/ed500288m) and the haloform reaction (J. Chem. Educ. 2015, DOI: 10.1021/acs.jchemed.5b00263).

“The first time, we did the accelerated reaction and the conventional reaction side by side,” Cooks says. “The conventional reaction took an hour. In that same hour, students easily did six different substituents in the spray reactions.”

Cooks, Zare, and Badu-Tawiah all hope the microdroplet approach for accelerating organic reactions will move beyond their labs to those of organic chemists.

To that end, Badu-Tawiah is developing what he calls a “contained electrospray” device. Its multiple inlets enable controlled introduction of reagents that then mix in microdroplets. As a first application, he used the device to study protein folding and unfolding (Analyst 2017, DOI: 10.1039/c7an00362e).

“I hope to make the device available to nonexperts,” Badu-Tawiah says. “I’m developing this contained electrospray to minimize sample prep, so you don’t need a lot of training to use this process.”

Cooks is optimistic that people will adopt the approach. “I’ll be surprised if this doesn’t catch on,” he says. “It’s easy to do, it’s very fast, and you get results that are comparable to slower, more conventional ways.”