From Nature’s Pathways to Future Clean Energy Technologies

“My lab is in the business of knowledge creation,” said David Beratan, R.J. Reynolds Distinguished Professor of Chemistry. “Our tools are pencil, paper and computers.” 

Beratan’s theoretical chemistry research group works closely with collaborators like Duke colleagues Michael Therien and Weitao Yang to realize ideas that come out of their research. “We're trying to understand how highly efficient biological energy systems work, so that our experimental collaborators can translate that knowledge into man-made structures that might provide new energy options,” Beratan said. “But we have to create knowledge first, before we can bring those ideas into service to humanity.”

All biological systems manipulate energy. Plants and other photosynthetic organisms use air and sunlight to make food. Animals don’t make their own food, but they are extremely efficient in extracting energy from the food they eat. Understanding how these biological systems convert energy from one source into another might show us better ways to produce the energy we need to power our world.

One clue lies in the movement of subatomic particles.

Living systems transform energy from one form to another by moving electrons and protons around. Beratan explained that all organisms use a particular type of reaction to move electrons, called a bifurcation reaction. It starts with two electrons occupying the same space. When these electrons are divided, their combined energy isn't split down the middle: one of them carries with it a lot of energy — imagine it shooting upwards — while the other one drops with little energy. The high energy electrons can then be harnessed into a pool that is used to make energy-storing compounds.

“What fascinates us is the ability of bifurcation reactions to deliver these electrons at just the right energy levels so that there's not a lot of heat generated,” he said. Loss of energy though heat is one of the main sources of inefficiency in human-created energy systems. 

Beratan’s research is supported by the U.S. Department of Energy. He and Michael Therien have received additional funding from the Keck Foundation to build bio-inspired systems. “These are synthetic constructs my friends make in the lab that mimic the rules we figured out for the living electron bifurcation systems,” Beratan said. One of their goals is to use sunlight as the spark that triggers electron bifurcation. “This has never been demonstrated in man-made systems,” he explained.

Beratan admits that recreating nature’s pathways is difficult, but his research group is making progress. “We think, with Mike Therien’s group, we have a strategy for building a light-driven bifurcation system to test and demonstrate some of our ideas. It’s still in the early stages,” he cautioned. “We're not yet ready to build functioning solar panels or solar cells that use this technology.”

Still, Beratan imagines that in the near future there might be a working prototype of an extremely high-efficiency solar cell based on this research. “These ideas may be five years away from being fully tested and realized, but when they are, the results could be transformative.”

Basic research, which underpins all of science’s great advances, takes time and nurturing to mature. Beratan pointed to a picture on his wall. “This is the research group I worked in as a graduate student at Caltech,” he said. “My Ph.D. advisor, John Hopfield, was just beginning to build mathematical models for associative memory.

“Associative memory works like this: if you remember part of something, like a face, then you might be able to reconstruct the person’s phone number or address from that partial memory,” Beratan said. “Hopfield's research evolved into the modern theory of artificial intelligence and machine learning.”

In 2024, decades after immersing himself in this research, John Hopfield won the Nobel Prize in Physics. 

“When I worked for him as a graduate student — there were no desktop computers in those days — I would see him down in the computer terminal room, programming and developing models that looked very cartoonish 45 years ago, but they evolved into modern AI.”

With Hopfield as a guiding example, Beratan takes the long view of his own research. 

“It took 30 or 40 years before his research all unfolded,” he said. “There's value to be found in basic research, but sometimes it takes decades of filling in the gaps of understanding before the technology follows. You have to be patient.”