Enabling unassisted solar water splitting by iron oxide and silicon

Abstract

Photoelectrochemical (PEC) water splitting promises a solution to the problem of large-scale solar energy storage. However, its development has been impeded by the poor performance of photoanodes, particularly in their capability for photovoltage generation. Many examples employing photovoltaic modules to correct the deficiency for unassisted solar water splitting have been reported to-date. Here we show that, by using the prototypical photoanode material of haematite as a study tool, structural disorders on or near the surfaces are important causes of the low photovoltages. We develop a facile re-growth strategy to reduce surface disorders and as a consequence, a turn-on voltage of 0.45 V (versus reversible hydrogen electrode) is achieved. This result permits us to construct a photoelectrochemical device with a haematite photoanode and Si photocathode to split water at an overall efficiency of 0.91%, with NiFeOx and TiO2/Pt overlayers, respectively.

Figure 1. Haematite with radically improved turn-on characteristics.

(a) Steady-state current density-potential behaviours of various haematite photoelectrodes. The current densities of Si photocathode placed behind the haematite photoanode are shown to illustrate the meeting points. (b) Band diagram of unmodified haematite (grey) and NiFeO_x-decorated haematite after re-growth treatments (red) under flat-band, quasi-equilibrium conditions. The Fermi level shift (denoted as 1) is a direct result of the re-growth treatment. The hole quasi-equilibrium potential shift (denoted as 2) is due to the application of NiFeO_x. (c) Open circuit potential measurements of aH, sdH, rgH I, rgH II, rgH III and NiFeO_x-decorated rgH II under 8-sun (red, triangle), 1-sun (blue, square) and dark (black, circle) conditions. Throughout this manuscript, sdH refers to solution-derived haematite; rgH I, rgH II, and rgH III denote haematite samples subjected to the regrowth treatments one, two and three times, respectively. Haematite prepared by atomic layer deposition (ALD) and then annealed at 500 °C and 800 °C are labelled aH and aH 800, respectively. NiFeO_x/rgH II represent rgH II haematite decorated with amorphous NiFeO_x catalysts. The error bars were obtained by taking s.d. values of measurements on at least three different samples for each data point.

Figure 2. X-ray diffraction, Raman and X-ray absorption analysis.

(a) X-ray diffraction patterns, (b) Raman shift spectra and (c) Oxygen K-edge X-ray absorption spectra of aH, aH 800 (ALD-grown haematite annealed at 800 °C in air), sdH, rgH I, rgH II and rgH III. The details of sample IDs can be found in the captions for Fig. 1.

Figure 3. Morphology evolution of haematite as a result of the re-growth treatments.

Scanning electron micrographs image of (a) sdH, (b) rgH I, (c) rgH II and (d) rgH III; scale bars, 100 nm. Magnified views of selected areas in the main frames are presented in the insets. Transmission electron micrographs of focused ion beam prepared cross-sectional samples of (e) sdH, (f) rgH I, (g) rgH II and (h) rgH III, scale bars, 500 nm.

Figure 4. Overall unassisted water splitting.

(a) Schematics of overall unassisted water splitting by haematite photoanode (right) and amorphous Si photocathode (left) in a tandem configuration. (b) Net photocurrent during the first 10 h of operation using NiFeO_x-modified rgH II with TiO₂/Pt-loaded amorphous silicon photocathode in 0.5 M phosphate solution (pH 11.8) in a two-electrode, tandem configuration (no external bias).