A Nickel Finish Protects Silicon Photoanodes for Water Splitting

See allHide authors and affiliations

Science  15 Nov 2013:
Vol. 342, Issue 6160, pp. 811-812
DOI: 10.1126/science.1246766

The large-scale generation of hydrogen from water with sunlight could provide a sustainable source of this industrially important gas, but could also provide fuel for vehicles and a storage medium for solar energy. The direct photoelectrochemical (PEC) splitting of water into hydrogen and oxygen, which combines a photovoltaic cell and an electrolyzer into a single device, remains an important goal (1). One problem is that some of the materials that work well for photovoltaics, such as n-type silicon (Si), corrode in electrolyzer solutions. On page 836 of this issue, Kenney et al. (2) show that a 2-nm-thick nickel (Ni) film on an n-type silicon semiconductor not only provides some stability against corrosion when used for oxygen evolution in a PEC configuration, but also generates a high voltage via a metal-insulator-semiconductor (MIS) configuration.

Hydrogen is used today primarily in the petroleum refining industry and for ammonia synthesis. More than 50 million tons of hydrogen are produced worldwide every year from fossil fuel feedstocks that generate CO2 emissions. In a carbon-free energy system, however, hydrogen must be produced from water splitting by means of renewable resources such as wind and solar energy. As solar energy is our largest resource, combining photovoltaics with electrolysis would seem to be the clear choice for renewable hydrogen production (3), but to date the only commercially available pathway in this scheme is the electrolysis step. Unfortunately, hydrogen from electrolysis with photovoltaic-generated electricity is far too expensive to be commercially viable, so other pathways must be considered.

Doubling up for solar hydrogen production.

A design configuration is shown where two separate semiconductors with different band gaps are illuminated in series to form a tandem system for water splitting. Sunlight illuminates the p-type electrode, which absorbs the visible light and transmits the red and near-infrared light that then illuminates the n-type electrode. The work of Kenney et al. shows that a thin nickel film can protect n-type silicon from corrosion by the electrolyte.

In a PEC water-splitting system, a semiconductor electrode is immersed in an aqueous solution, and when illuminated it splits water directly at the semiconductor's surface. For the use of n-type silicon, the study of Kenney et al. presents several critical results. The 2-nm Ni film combined with the thin native silicon oxide layer (SiOx) that forms on silicon exposed to air, and the surface of the nickel oxidized in the electrolyte to form nickel oxide (NiOx). The resulting NiOx/Ni/SiOx/Si device generated a voltage of 500 mV when exposed to light, with no need for the thermally grown SiOx layer that has traditionally been required to achieve that voltage (4). This combination stabilized and catalyzed the interface for oxygen evolution. An unexpected finding is that lithium ions from the electrolyte further improved the stability.

Formally, this device could be viewed as a buried junction, where the charge carriers form at the junction buried underneath the 2-nm nickel layer (the MIS structure). However, thicker 5- and 10-nm nickel films did not display the same high voltage, so the aqueous solution must play a role in the operation of this device. The authors attribute this effect to incomplete screening of the solution by the Ni/NiOx layer.

Recent technoeconomic analysis (5) shows that to produce cost-competitive hydrogen via a PEC process, the solar-to-hydrogen efficiency should be at least 15% and perhaps greater than 20% (6). Other studies have shown that to achieve this efficiency, not only must the semiconductor electrode have the same solid-state properties as current photovoltaic devices, it also must have a tandem configuration (79). In a tandem configuration, two semiconductors with different band gaps are illuminated in series, so that the top semiconductor with the higher band gap absorbs the visible light and transmits the rest through to the bottom cell with a lower band gap. Thus far, the only PEC system that shows greater than 10% water-splitting efficiency is a tandem device composed of high-efficiency III-V semiconducting materials (10), such as gallium arsenide. This tandem configuration limits the semiconductors that can be used to pairs of highly crystalline materials that have matching crystal lattices.

An alternative tandem scheme presented by Nozik (11, 12) relaxes these requirements by making use of separated p-type and n-type photoelectrodes with two different band gaps. This separated p-n arrangement eliminates the need to either match lattices or create stacks of dissimilar materials, and further allows the use of polycrystalline materials. Nozik showed that this configuration could perform unassisted water splitting, but the efficiency was limited by the photoanode. There are a number of excellent p-type photoelectrodes, including silicon, that can produce hydrogen with high efficiency, but no known n-type photoelectrodes (photoanodes) can produce oxygen with high efficiency.

The reason why n-type silicon was thought to be unsuitable for oxygen evolution was its instability in basic conditions and the formation of a thick oxide film that blocks the reaction in acidic conditions. However, the results of Kenney et al. show that a thin Ni film can protect the n-Si surface for oxygen evolution, as well as afford a good photovoltage. Thus, their result opens up the possibility of using this electrode in a p-n tandem configuration by coupling it with a photocathode that has a wider band gap, such as p-type copper gallium diselenide (p-CGS).

A tandem configuration of these two materials has a maximum theoretical efficiency greater than 25% (9). As shown in the figure, sunlight first illuminates p-CGS, which has a band gap of 1.68 V. The light that is not adsorbed illuminates the Ni-coated Si photoanode; when CGS is deposited on transparent conducting glass substrates, it shows good transparency for the longer-wavelength light below its band gap (13). Such a configuration illustrates the ability of a PEC system to integrate polycrystalline thin films with single-crystal photoelectrodes into a viable tandem device; this would be more difficult to accomplish with a solid-state device.

The results of Kenney et al. are a long way from being integrated into a viable water-splitting device. However, they do point the way toward reconsideration of a long-held belief about n-type silicon as a photoanode for oxygen evolution. The results open up some additional possibilities for a solar water-splitting system with efficiencies of 15% or greater.


View Abstract

Stay Connected to Science

Navigate This Article