New nanostructures enhance solar water splitting with hematite

The ultimate goal of the photoelectrochemist is to split water into hydrogen and oxygen using the power of the sun and an inexpensive, safe, and durable device. Hematite (alpha-Fe₂O₃) is a promising material for this application because of its terrestrial availability, stability, and ability to absorb a significant fraction of solar irradiation (band gap = 2.1eV). With this band gap, an impressive 15% solar-to-hydrogen (STH) conversion efficiency is possible, making the use of hematite an extremely attractive method for storing energy from our largest renewable power source, the Sun.

However, many challenges must be overcome if we are to exploit the advantages of hematite. Because of its non-ideal electronic properties, an additional electrical bias must be applied to achieve water reduction. We have shown that this limitation can be overcome by using the tandem cell approach illustrated in Figure 1 to achieve unassisted solar hydrogen production from water.¹ In addition, enhancements including substitutional cation doping and nanostructuring are required to control the charge transport and surface area for water oxidation. Previously, nanostructured and silicon-doped hematite photoelectrodes fabricated by a chemical vapor deposition technique were shown to give record photocurrents corresponding to 3% STH efficiency.² Unfortunately, the morphology of these films is inadequate for obtaining higher efficiencies. In addition, to benefit from the low cost of iron oxide, inexpensive methods of preparing electrodes that yield high efficiencies are required.

Figure 1. Scheme for the tandem cell approach to solar water splitting. A nanostructured hematite photoanode and a second photocell that absorbs complementary wavelengths of sunlight achieve unassisted water cleavage with only sunlight as an input. H₂: Hydrogen. O₂: Oxygen. H₂O: Water. e⁻: Electron. I⁻: Iodide ion, I₃⁻: Triiodide ion. λ: Wavelength.

We have sought to control the electronic properties and morphology of hematite photoelectrodes using two parallel approaches: using extremely thin absorbers and making high-surface-area films using a solution-based colloidal approach. We demonstrated the effectiveness of both methods in using scalable, inexpensive methods to prepare hematite that deliver high photocurrents under standard solar illumination.

The extremely thin absorber (ETA) approach overcomes the poor electronic properties of hematite by decoupling light absorption from charge transport. This is accomplished by depositing an ultrathin layer of hematite on a high-surface-area scaffold. Using a tungsten oxide (WO₃) scaffold with a thin (60nm) layer of hematite, we demonstrated a 20% increase in photocurrent over electrodes prepared on a flat surface with the same amount of hematite (see Figure 2).³ While this was the first demonstration of the effectiveness of this approach using hematite for solar water splitting, we found that the performance was limited by high recombination at the WO₃/hematite interface. Further investigation revealed the importance of crystallinity in the ultrathin films.

Figure 2. The extremely thin absorber (ETA) approach to nanostructuring hematite photoanodes. (a) Cross-sectional scheme and top-down scanning electron micrograph of the electrode. (b) Photoelectrochemical performance of the ETA electrode (solid blue curve) compared to that of a control electrode (dashed red curve) under standard illumination testing conditions. Curves for the electrodes in the dark are shown as black lines. WO₃: Tungsten oxide. Fe₂O₃: Hematite.

We subsequently established a method for controlling the crystallinity of ultrathin hematite films using a SiO_x buffer layer, which was found to change the growth mechanism of hematite. Using this layer, we were able to obtain remarkable photoactivity in films only 12.5nm thick. These films exhibited a water oxidation photocurrent onset potential at 1.1V versus a reversible hydrogen electrode (RHE). The plateau photocurrent was 0.63mA/cm² at 1.5V versus an RHE under standard illumination conditions, representing the highest reported performance for ultrathin hematite films.⁴ In contrast, we observed almost no photoactivity in the photoanode with the same amount of hematite deposited without the buffer layer. In addition, since both the buffer layer and the ultrathin hematite films were prepared using a simple solution-based approach, this method can potentially support the cost-effective preparation of high-performance electrodes.

An even more economical approach to enhancing the performance of hematite photoelectrodes is to prepare nanostructured, porous films using a solution-based colloidal approach, as illustrated in Figure 3(a). Many groups have pursued this approach, and while they have obtained phase-pure and adequately structured hematite, only minimal photoactivity was observed. However, we recently obtained highly photoactive hematite photoelectrodes using this system with a breakthrough discovery of the importance of the sintering temperature.⁵ We found water-splitting photocurrents of 1.0mA/cm² (at 1.55V vs an RHE) with electrodes sintered at 800°C, whereas no photocurrent was observed with electrodes sintered at 400°C to 700°C, as shown in Figure 3(b). Increasing the sintering temperature increased the average particle size and affected the activation and incorporation of dopant atoms. In addition, we found a considerable change in the absorption coefficient with increasing sintering temperature, a critical feature for hematite as a solar-energy converter. We performed a detailed investigation into hematite's crystal structure using powder X-ray diffraction with Rietveld refinement to account for this effect and found a correlation between an increase in C_3v-type crystal lattice distortion and improved optical properties. We are currently working to decouple the photoactivity and the particle size increase in this approach.

Figure 3. The colloid-based approach to nanostructuring hematite photoanodes. (a) Cross-sectional scheme and top-down scanning electron micrograph of the electrode. (b) Photoelectrochemical performance of a porous electrode sintered at 800°C (blue curve) compared to that of electrodes sintered at 400°C and 700°C (green and red curves, respectively) in the dark (broken lines) and under standard illumination (solid lines) testing conditions. E: Potential. J: Current density of photocurrent.

Both of these approaches to controlling the form and function of hematite have contributed to the understanding of the important factors that determine the photoactivity of this promising material for solar energy conversion. We expect that further work that optimizes both the ETA and the solution-based colloidal approaches will eventually realize the 15% conversion efficiency expected from hematite using low-cost photoelectrodes. This would enable economical storage of solar energy in the form of hydrogen.

We thank the Marie Curie Research Training Networks (contract number MRTN-CT-2006-032474) and the Swiss Federal Office of Energy (project number 102326) for financial support.