By making a solar photovoltaic material more resilient, researchers may have found a way to make artificial photosynthesis—that is, using sunlight to make fuel—cheap enough to compete with fossil fuels.

If you want hydrogen to power an engine or a fuel cell, it’s far cheaper to get it from natural gas than to make it by splitting water. Solar power, however, could compete with natural gas as a way to make hydrogen if the solar process were somewhere between 15 and 25 percent efficient, says the U.S. Department of Energy. While that’s more than twice as efficient as current approaches, researchers at Stanford University have recently developed materials that could make it possible to hit that goal. The work is described in the journal Science.

One way to make hydrogen using sunlight is to use a solar panel to make electricity and then use that electricity to power a commercial electrolyzer that splits water, forming hydrogen and oxygen. But combining the solar panel and the electrolyzer in one device might be cheaper and more efficient. The electrons produced when light hits a photovoltaic material could facilitate chemical reactions, and the capital costs of one machine would likely be lower than the cost of two (see “A Greener ‘Artificial Leaf,’” “Sun Catalytix Seeks Second Act with Flow Battery,” and “Artificial Photosynthesis Effort Takes Root”).

For some time now researchers have known that you could approach 15 to 25 percent efficiency if you combined two solar cell materials in such a system. One solar cell would power half of the water-splitting reaction—forming hydrogen. The other could form oxygen.

The hydrogen part is pretty much solved now, but researchers have had trouble with the oxygen half. The most efficient solar cell materials for this reaction (silicon, for example) quickly corrode. The Stanford researchers discovered that they could make silicon last for days, rather than just a few hours, by coating it with a protective layer of nickel just two-billionths of a meter thick. The materials split water for three days before the researchers stopped the experiment to examine the materials for damage. They found none.

Other materials—such as metal oxides—can last this long, but they split water very slowly. The new materials are an order of magnitude faster, says John Turner, a research fellow at the National Renewable Energy Laboratory in Golden, Colorado. “Over 40 years of work on oxides has not produced a result like this,” he says.

It could be a while before the materials are used in commercial hydrogen production. To achieve the needed efficiencies, the materials would still need to be incorporated into a system that uses two solar cells. And a big remaining question is how long the materials can last. To be economical, a system would have to run for at least five years, Turner says.