Researchers are using a new high-throughput method of identifying new materials. Photo: Caltech

The hunt for “needle in a haystack” materials that could help efficiently produce fuel from just water, sunlight and carbon dioxide has, over four decades, yielded only 16 prospects, none of which led to the creation of a commercially viable solar fuels generator.

But the number of materials that could serve as catalysts for creating solar fuel is now significantly expanding thanks to the development of a new “discovery pipeline” by a team of researchers in California—a breakthrough that puts more options on the table for scientists trying to develop a renewable energy source.

In two years, the research team identified nearly double the number of applicable materials using a new method developed through a partnership between the Joint Center for Artificial Photosynthesis at Caltech and Lawrence Berkeley National Laboratory’s Materials Project.

The scientists say they have also discovered, but not yet reported, dozens of additional materials that may be capable of splitting water using energy from the sun. After splitting water, the extracted hydrogen atoms can be used to create hydrogen gas or combined with carbon dioxide to create hydrocarbon fuel.

High-throughput theory
By mimicking the natural process of photosynthesis, scientists aim to convert and store the energy of the sun for on-demand use in cost-effective and scalable systems. But since water doesn’t separate into hydrogen and oxygen in the presence of sunlight, a material known as a photoanode is needed to facilitate that reaction.

John Gregoire, principal investigator and research thrust coordinator with the Joint Center for Artificial Photosynthesis, said researchers face a significant challenge when trying to identify potential photoanodes. While metal oxides are “very promising materials” for photoanodes due to their stability, they usually don’t absorb visible light.

“The trick is finding the special kind of metal oxides that absorb visual light or, more technically, have a band gap energy in the visible range,” Gregoire told Laboratory Equipment. “These are really kind of needle in a haystack materials. There are many known metal oxides, but very few of them exhibit all the necessary properties to be solar fuels photoanodes.”

For example, there are only 16 known oxide photoanodes, despite the thousands of metal oxides chemists work with on a daily basis.
To accelerate the process of identifying potential photoanodes, Gregoire and a team of researchers co-led by Lawrence Berkeley National Laboratory’s Jeffrey Neaton and Qimin Yan developed a discovery pipeline that integrates theory and experiment.

The researchers started by selectively mining a database of roughly 66,000 compounds with a well-defined hypothesis and identified 174 potentially promising vanadates, which contain vanadium, oxygen and one other element. They then screened those vanadates using a computational method aimed at predicting which materials would exhibit the properties of a photoanode.

“High-throughput theoretical materials discovery has been on the rise recently, but a shortcoming is that such efforts do not involve experiments and therefore are limited by the accuracy of the computational methods,” Neaton, also a principal investigator with the Joint Center for Artificial Photosynthesis, told Laboratory Equipment. The center is a U.S. Department of Energy Innovation Hub dedicated to the development of solar fuels.

But, in the integrated discovery pipeline developed by Gregoire and Neaton, promising materials identified by the computational methods were passed directly to experiment, where researchers measured the materials’ optical and photocatalytic properties.

While high-throughput techniques are commonplace in pharmaceutical and biological labs, materials science tends to be more complex in that running high-throughput theory and high-throughput experiments separately has yielded ineffective results.

“By combining them and coming up with a screening pipeline that uses both experiment and theory together is how we’ve been able to really accelerate the discovery process,” Gregoire said.

While the researchers acknowledge the 12 materials they identified and reported in their study are far from appearing in any type of commercially available solar fuel generator, they say one of the keys to accelerating the development of solar technology is identifying more potential photoanodes. And the integrated pipeline does just that—quickly.

“This is a discovery that they have any activity at all,” Gregoire said.

Additional calculations and experiments are next to figure out how the newly discovered materials can be optimized. Several questions about the materials need to be answered in the process, including, what is their maximum activity? and what is the limit of their efficiency?

As if identifying 12 new materials suitable as photoanodes isn’t accomplishment enough, Gregoire, Neaton and colleagues said they are not done yet.

“The bigger picture is that we only used one of our design criteria to identify these 12. We have a number of other design criteria. So we are exploring very different metal oxides and are continuing to rapidly discover photoanodes in these other spaces, as well. Now that we have this high-throughput discovery pipeline working, we’re getting very good at finding all the needles in the haystack. We have dozens more discoveries.”

Artificial leaf
Other researchers at the Joint Center of Artificial Photosynthesis have developed an artificial leaf, a solar-driven system that splits water to create hydrogen fuels. The system is made up of two electrodes—one photoanode and one photocathode—and a membrane.

The photoanode uses sunlight to oxidize water molecules, generating protons and electrons as well as oxygen gas. The photocathode recombines the protons and electrons to form hydrogen gas. A key part of the JCAP design is a plastic membrane, which keeps the oxygen and hydrogen gases separate. If the two gases are allowed to mix and are accidentally ignited, an explosion can occur; the membrane lets the hydrogen fuel be separately collected under pressure and safely pushed into a pipeline.

In studies, the system was shown to convert 10 percent of the energy in sunlight into stored energy. During natural photosynthesis, plants convert about 1 percent of the sunlight’s energy into stored energy. Additionally, the artificial leaf system was proven capable of continuously operating for more than 40 hours.

“Our work shows that it is indeed possible to produce fuels from sunlight safely and efficiently in an integrated system with inexpensive components,” said Caltech’s Nate Lewis in August 2015, when the system was announced. “Of course, we still have work to do to extend the lifetime of the system and to develop methods for cost-effectively manufacturing full systems, both of which are in progress.”

Part of what Neaton, Gregoire and their team is trying to do is find replacements for the more expensive materials in prototype devices, like the artificial leaf, to drive down costs and create efficient, affordable and scalable solar fuel generation technology.

“This photoanode material is just one material in the overall device that involves having to integrate many materials and having them all work together,” Gregoire said. “We don’t know yet whether any of these discoveries will be able to be integrated with the other known components and create an efficient technology that is deployable and cost competitive with existing methods for producing fuels, but it gives us a lot more options to explore.”

The timeline for when solar fuels may be powering people’s cars and homes is not clear, but researchers say their new method is giving a boost to that effort.

“It’s not going to be next week, tomorrow, even two years from now. It’s still down the road,” said Neaton. “We need to be patient, but I do think that the way to accelerate this process is going to be identifying new materials that could realistically be used in devices.”