Since titania (TiO2) was first shown to be capable of splitting water into hydrogen and oxygen, in the presence of an external electric field and super band gap irradiation, it has become synonymous with semiconductor photochemistry [1]. In the years since this discovery over 50 semiconductor systems have been studied, yet to date, none have proven suitable for the commercial solar generation of fuel. In essence we are yet to mimic plants and reproduce photosynthesis using semiconductor materials in an economically viable manner.

The importance of addressing this issue has been highlighted across the globe, with over 1000 papers on water splitting being published in 2010. In January 2011 the MRS Bulletin was dedicated to semiconductor strategies for splitting water [2]. In 2010 the Engineering and Physical Research Council (EPSRC) of the UK announced multimillion pound funding for research in this area [3]. There are also other examples from around the world where there is a high level of commitment and support for investigation in this area.

There are, arguably, three main problems associated with the use of semiconductor systems when splitting water to provide a fuel source [4].

The first is that, to be economically viable, the semiconductor must absorb and become active under visible light illumination. It must also produce excited species that are capable of driving a chemical reaction in the absence of external power supplies. These factors and characteristics are relatively well understood for most semiconductor systems and to a certain extent limit the systems of interest. The second problem is that the photoexcited carriers can simply recombine within the semiconductor and produce no useful work. This is termed internal recombination and must be prevented as far as possible to enhance performance. And, the third is that reactants and products are held in close proximity at the surface of the catalyst, as they are produced very closely to each other. This means that the equilibrium of the reaction is not completely pushed toward products, due to the possibility of back reactions.

All three of these factors combine to reduce the overall efficiency of the system.

What is being done today to address these issues? Currently the research community is blending semiconductor materials with other semiconductors or metals to increase the amount of sunlight absorbed or reduce internal recombination. Research groups are also focusing on nanostructuring materials to enhance the reactivity of the semiconductor.

Can we learn from photosynthesis? Can we start to use materials to increase the performance of the system? Can we use materials that naturally help, by not suffering from some of the problems highlighted above? In nature, plants separate the process of producing excited carriers with the process of reacting them to form a product. I'd like to present the case for materials that can address the second two problems with splitting water, and so maybe learn a little from the natural world.

There are a class of materials that are well known, well characterized, and commercially well used. They can sustain an internal dipole, indefinitely, and this dipole can be measured as a surface charge density of many coulombs per unit area. The large and sustainable dipole interacts with the photoexcited species of electrons and holes, effectively separating them across the surface. This has the impact of spatially separating surface photochemistry: the surface exhibits reduction in one location and oxidation in another. It also reduces the ability of the excited carriers to internally recombine. It's as if there's a p-n junction within a single material pulling photoexcited carriers apart [5].

How much control is there over this selective photochemistry? It has been shown that patterns as small as 100 nm can be drawn and show good fidelity for selective photoreduction of metal cations to metal clusters. Conversely it has also been shown that surfaces as large as many square centimeters can also exhibit selective photochemistry [6]. One surface acts as the cathode, the other as the anode [7]. A further additional benefit is that the large surface dipole can interact with the species on the surface [8], and in the case of molecules, disrupt the electrons in the bonding orbitals [9]. This could reduce the energy required to break those bonds and enhance reactivity.

What are these materials? They are ferroelectric materials, such as BaTiO3 and LiNbO3.

The work being undertaken around the globe on blended, core shell, nanostructured, and other forms of semiconductor structure will continue to move forward to address some of the challenges we are currently facing in developing efficient semiconductor systems. But perhaps one answer might lie in thinking outside of the box and investigating new types of materials with inherent properties that can directly address some of the problems associated with systems currently being developed.

Further Reading
[1] A. Fujishima, K. Honda, Nature, 238 (1972), p. 37
[2] MRS Bulletin, 36 (2011), p. 1
[4] F.E. Osterloh, B.E. Parkinson, MRS Bulletin, 36 (2011), p. 17
[5] S. Dunn et al. J Mater Chem, 17 (42) (2007), p. 4460
[6] S. Dunn et al. J Am Chem Soc, 129 (28) (2007), p. 8724
[7] J.L. Giocondi, G.S. Rohrer, J Phys Chem B, 105 (2001), pp. 8275–8277
[8] S. Dunn et al. Appl Phys Lett, 85 (16) (2004), p. 3537
[9] A.L. Cabrera et al. Surf Sci, 336 (3) (1995), p. 280

Read full text on ScienceDirect

DOI: 10.1016/S1369-7021(11)70151-1