Every time we use fossil fuels to move ourselves around, heat our homes, or generate electricity we're using stale solar energy. Plants took sunlight, water, and carbon dioxide and turned them into hydrocarbons, which then got socked away for anywhere between hundreds of thousands and hundreds of millions of years.
Rather than burn those vintage hydrocarbons, releasing their stored carbon into the atmosphere, wouldn't it be great if we could mimic the process of photosynthesis on our own? Imagine taking sunlight, water, and carbon dioxide and turning them directly into fuel for our machines. Not only could we avoid dumping the Paleozoic's sequestered carbon back into the atmosphere, but we might well be able to avoid the other kinds of pollution involved in extracting and processing fossil fuels.
That's the theory. The practice is more complex and fraught with obstacles. But a group of researchers at Caltech might just have found a way around one of those obstacles, using a material once deemed worthless for technological use.
In order to mimic what plants do with sunlight, CO2, and water, we'd need to split water into its component parts (oxygen and hydrogen), split the CO2into carbon and oxygen, then chemically combine the hydrogen and carbon to make fuels like those we currently refine from petroleum.
(Why not just use hydrogen as a fuel? Mainly because hydrogen is light, takes up a lot of room, and doesn't hold all that much energy per unit of fuel compared to hydrocarbons. While hydrogen might well see some use someday as an automotive fuel, it's unlikely that heavy equipment or aircraft will ever be able to use hydrogen as an energy source.)
Running this long chain of chemical processes on solar energy is obviously possible: plants do it without thinking. Simulating photosynthesis in an industrial setting, though, is tricky. For instance: we know how to split water into oxygen and hydrogen: we run an electric current through water with salts, "electrolytes," dissolved in it. The electrical power will pry water molecules apart, and the resulting hydrogen will migrate through the electrolite solution toward the negatively charged electrode, while the oxygen does the same in the direction of the positively charged electrode.
To make this process (called "electrolysis") run on solar power, we could merely wire the whole thing up to a battery of solar panels on the laboratory roof. But that wouldn't be very efficient. Ideally, the electrodes immersed in the electrolyte solution would themselves act as photovoltaic cells, directly capturing the sun's energy and harnessing it to split water into oxygen and hydrogen.
But the semiconducting materials most appropriate for use as photoelectrodes, like silicon, gallium phosphide, and gallium arsenide, have a little problem: they corrode when exposed to an electrolyte solution, sometimes in just a few seconds. And if you coat the photoelectrodes to keep them from corroding, they either absorb less light or transmit less electrical current, or both.
That's why a paper due to be published Friday in the journal Science is exciting some interest in the Artificial Photosynthesis Community. A team of researchers from Caltech's Joint Center for Artificial Photosynthesis (JCAP) found that by depositing an extremely thin layer of titanium dioxide on the photoelectrodes, they could protect those photoelectrodes from the corrosive effects of the electrolyte solution without cutting down on either light transmission or conductivity.
The team, led by postdoctoral scholar Shu Hu, turned to a form of titanium dioxide once considered and then rejected for use in computer chips. Chipmakers spurned the substance because it didn't hold electric charges the way they wanted it to. But as it turns out, that "leakiness" turns out to be precisely the property you want in a material used to coat an electrode. The team applied the "leaky" titanium dioxide to the semiconductor in layers between 4 and 143 nanometers thick. A nanometer is a billionth of a meter: a human hair is about 700 times thicker than the thickest layer of titanium dioxide the Caltech team used.
Titanium dioxide is commonly used as a white pigment in uses from paint to toothpaste, but in layers one seven-hundredth as thick as a human hair, it's transparent. And that means the photoelectrodes' titanium "raincoats" didn't interfere with their ability to capture light and use it to split water.
The photoelectrodes were also dotted with nickel oxide "islands"; nickel oxide is an effective "electrocatalyst," which reduces the amount of energy needed to split water molecules.
The result? Instead of the semiconductors corroding to the point of uselessness in seconds or minutes, the titanium-coated photoelectrodes continued to function for 100 hours of continuous operation.
"For the better part of a half century, these [semiconductors] have been considered off the table for this kind of use," says Caltech chemistry professor Nate Lewis, the paper's principal investigator. "But we didn't give up on developing schemes by which we could protect them, and now these technologically important semiconductors are back on the table."
It's an incremental development rather than a breakthrough, to be sure. The team doesn't know whether titanium dioxide applied in less-exacting industrial settings, as opposed to atom by atom in a Caltech lab, will work as well. And looming over the whole solar fuel concept, there remains a pesky fact: fossil fuels are cheap because the planet has been saving them up for millions of years, and making fuels out of solar energy in realtime, with just this year's sunlight as opposed to sunlight that hit the planet before there were dinosaurs, will almost certainly be less convenient even when we master the process.
But it's a step in an interesting direction, nonetheless.