Artificial Leaf Is 10 Times Better at Generating Hydrogen from Sunlight

A new "artificial leaf" that can turn sunlight and water directly into hydrogen and oxygen is shattering performance and stability records

4 min read

Artificial Leaf Is 10 Times Better at Generating Hydrogen from Sunlight
Photo: Lance Hayashida/Caltech

A "hydrogen economy" sounds just about as green and eco-friendly as it gets. Fuel cells that combine hydrogen with ambient oxygen in the air can generate electricity with naught but pure water as a byproduct—which is great if you hate pollution and are thirsty. The problem we face now is the source of our hydrogen: the vast majority of it comes from fossil fuels, specifically natural gas. And while transforming methane into hydrogen is 80 percent efficient, that other 20 percent is carbon dioxide.

The vast majority of the accessible clean hydrogen on Earth is locked up with oxygen in water, but breaking apart H2O into an O and a useful H or two isn't a particularly environmentally-friendly or efficient process to get involved in. The fantasy is an "artificial leaf," a passive, inexpensive thing that you can stick in water and expose to sunlight, then watch as it bubbles off all the hydrogen and oxygen you need. People have been working on these, but Caltech has just made an enormous amount of progress with an artificial leaf that, according to the researchers, "shatters all of the combined safety, performance, and stability records for artificial leaf technology by factors of 5 to 10 or more."

Caltech's Joint Center for Artificial Photosynthesis has been working on ways of turning sunlight, water, and CO2 into useful chemical fuels for the last half decade. They've just published a paper in Energy and Environmental Science describing a new design that's efficient, simple, relatively inexpensive, and reliable enough to suggest a path towards industrialization.

The Caltech design uses two electrodes (a photoanode and a photocathode) separated by a membrane. The photoanode is made of gallium arsenide, which is an excellent light absorber, but tends to oxidize when exposed to water. A layer of titanium dioxide helps keep the photoanode stable and protected, and a layer of nickel on top of that acts as a catalyst. When exposed to sunlight, the photoanode oxidizes water molecules, generating oxygen (O2) as well as protons and electrons, which pass through the membrane and are recombined by the photocathode into hydrogen (H2). Put all of this together, and you end up with a single, fully integrated system that cracks water into oxygen and hydrogen when you put it in the sun:

The 1-square-centimeter prototype shown in the video operated at above 10 percent efficiency for 40 hours straight, producing about 0.8 microliters of hydrogen per second. It's all one piece, there's no wiring, and the active electrocatalysts are all “earth abundant” (another way of saying affordable).

It's obviously nowhere near ready for production, although the Caltech team is already working on methods for cost-effectively manufacturing full systems. But watching it effortlessly produce fuel from water and sunlight is pretty awesome.

Having said that, it's important to note that right now, you can get significantly more energy from sunlight (about 20 percent efficiency) by using a conventional solar cell to convert it into electricity rather than hydrogen. It's also way cheaper to do this, since solar cells are an established industry. But, using that electricity to make hydrogen through electrolysis isn't particularly efficient: with a non-concentrator silicon photovoltaic cell and a commercial electrolyzer, you're looking at a complex system that can convert sunlight into hydrogen with an efficiency of about 12 percent. An artificial leaf that could approach the efficiency of existing sunlight-to-hydrogen converters in a much simpler (and eventually cheaper) way is appealing. But if it does so at something like half the energy efficiency of just using existing technology to make electricity directly, why bother with the hydrogen at all?

In a bunch of important ways, hydrogen is a much better medium for energy storage than batteries. As an inert chemical fuel, hydrogen can store energy for decades without degrading, and offers high energy density, especially in compressed or solid form (bound up in a metallic lattice or powder). For electric vehicles, for example, hydrogen could offer much more range than several tons of lithium ion batteries currently do. And refueling hydrogen is much more like putting gasoline in your car than sitting at a rest stop while your car sips from a charger.

imgIllustration: Joint Center for Artificial Photosynthesis/Darius Siwek

The current incarnation of the device isn't completely stable. In other words, there is some degradation, such that you couldn't leave it running indefinitely. The initial solar-to-hydrogen conversion efficiency was almost 11 percent, decreasing to 10 percent after 40 hours of operation, and 9 percent after a total of 80 hours of operation. Eventually (although the paper doesn't say exactly when) the device experienced "catastrophic failure;" it's left up to us to imagine exactly what that meant. The source of this failure seems to be defects in the protective titanium oxide film originally caused by dust particles on the semiconductor surface when the film was being deposited. The electrolyte was able to sneak in through these defects, slowly eating away at the TiO2 and resulting in a linear decrease in conversion efficiency to the point of failure.

The good news is that the researchers have identified the cause of the degradation, and they're pretty sure that it's a solvable issue. They suggest that it might be possible to use sacrificial arrays of nanowires to electrochemically isolate the defects, allowing the device to function longer, and results for such systems will be reported as soon as they've gotten them to work. We hope they’ll do it soon, because being able to turn sunlight directly into hydrogen efficiently on a residential scale could abruptly make a hydrogen economy much more realistic.

The Conversation (0)