Inorganic semiconductors are promising materials for driving photoelectrochemical water-splitting reactions. However,
there is not a single semiconductor material that can sustain the unassisted splitting of water into H<sub>2</sub> and O<sub>2</sub>. Instead, we
are developing a three part cell design where individual catalysts for water reduction and oxidation will be attached to
the ends of a membrane. The job of splitting water is therefore divided into separate reduction and oxidation reactions,
and each catalyst can be optimized independently for a single reaction. Silicon might be suitable to drive the water
reduction. Inexpensive highly ordered Si wire arrays were grown on a single crystal wafer and transferred into a
transparent, flexible polymer matrix. In this array, light would be absorbed along the longer axial dimension while the
resulting electrons or holes would be collected along the much shorter radial dimension in a massively parallel array
resembling carpet fibers on a microscale, hence the term "solar carpet". Tungsten oxide is a good candidate to drive the
water oxidation. Self-organized porous tungsten oxide was successfully synthesized on the tungsten foil by anodization.
This sponge-like structure absorbs light efficiently due to its high surface area; hence we called it "solar sponge".