Researchers have been working on a process that uses sunlight to produce hydrogen by splitting water directly. To understand photoelectrolysis, think of a PV cell underwater, where the electrochemical energy produced is immediately used to electrolyze water, instead of creating an external current. The light hits the cell, and hydrogen bubbles appear on one side of the cell, while oxygen appears on the other side, just as in electrolysis. (Of course one could use a PV cell to power an electrolyzer, but the idea here is to make a simpler and more economical system.)

The interface between the water (electrolyte) and certain semiconductor materials forms a diode junction that generates power–and thus does the electrolysis. The presence of catalysts at the surface can also help with the energetics and kinetics of the reactions that form the hydrogen and oxygen, respectively.

One of the problems is that the minimum voltage for splitting water (1.3 volts) is higher than a photocell can easily produce, and high-bandgap materials capable of generating enough voltage can utilize only ultraviolet light, which is a small fraction of the solar spectrum.

Work at NREL and the University of Hawaii has focused on developing multijunction cells which use more of the solar spectrum. These additional layers are sandwiched inside the basic cell that does the photolysis, and provide a boost to the electro potential available to do the water splitting. The electrochemistry and solid state physics of these devices are very complex. One of the main challenges has been to come up with materials and configurations that will be less susceptible to corrosion from the electrolyte and which will last long enough to be practical. Efficencies above 12% have been seen (i.e., the energy value of the hydrogen produced vs. the amount of incident sunlight. (See the 2002 H2 DOE Program Reviews–ref. below. Also, the 2003 meeting in May will have new updates.)

Researchers at the University of Duquesne published an important development in Science Magazine last September. Titanium dioxide is known to be a cheap and stable photocatalyst for splitting water, but hydrogen yields were always less than 1% (due to the high band gap of the material). The new development involved preparing the material in a flame, introducing carbon into its structure. Cells using this new material saw a factor of 10 increase in hydrogen production. The University is actively seeking licensees or partners to pursue this technology. (Contact me for details).

The design goal at NREL and Hawaii is to come up with a monolithic device that needs no external electrical connections. The simple version of the Duquesne cell requires an external bias power source (which could be powered by a fuel cell using some of the hydrogen produced), but which would still be a net producer of power. Net yields are already at 8.5%, and are expected to improve.

Though commercial devices are a ways off, photosplitting of water is another process that could supply hydrogen by purely renewable means.

References:

2002 Hydrogen Program Review Meeting – Renewable Production Electrolytic Processes
http://www.eere.energy.gov/hydrogenandfuelcells/hydrogen/annual_review2002.html#Renewable

Science…27 Sept 02
“Efficient Photochemical Water Splitting by a Chemically Modified n-TiO2”

Science 17 April 98
“A Monolithic Photovoltaic-Photoelectrochemical Device for Hydrogen Production via Water Splitting”

( I can provide pdf copies of the Science articles).