Solar water splitting depends on more than choosing a light-absorbing semiconductor and an active electrocatalyst. The interface between them also determines how effectively photo-generated charge carriers reach the reaction site—and how much energy is lost to recombination. In a metal-insulator-semiconductor (MIS) photoanode, a carefully selected protective-layer thickness can help manage that interface, improving photovoltage even when the semiconductor and catalyst form a relatively low-barrier junction.
This finding challenges a common design assumption: that the protective insulator should always be as thin as possible to ease charge transfer. Results from an n-type silicon (n-Si), hafnium oxide (HfO2) and nickel (Ni) system show why intermediate thickness can be more effective.
Why protective layers matter in solar water splitting
Many semiconductors that absorb sunlight effectively are vulnerable under the operating conditions required for water oxidation. At an oxygen-evolving photoanode, anodic conditions can promote oxidation and degradation of the semiconductor. A protective oxide layer can separate it from the electrolyte while allowing a metal or metal-oxide electrocatalyst to support the surface reaction.
These layered structures are often called MIS systems: the semiconductor absorbs light, the insulator provides protection and influences charge transfer, and the electrocatalyst drives the chemical reaction. System performance depends on the properties of all three materials and on the interfaces between them.
A further challenge is the semiconductor-electrocatalyst barrier. A sufficiently high barrier can help suppress unwanted charge flow and recombination, but catalysts with favorable work-function characteristics may not always be the most practical choice. Nickel, for example, can catalyze oxygen evolution but has less favorable work-function characteristics for forming a high-barrier junction with n-Si than catalysts such as iridium. That mismatch can limit photovoltage.
How insulator thickness controls charge-carrier flux
Charge carriers cross an insulating layer through tunneling. Increasing the layer thickness lowers the probability of tunneling, reducing the transfer rates of both minority and majority carriers. The design objective is not simply to maximize or minimize transfer. It is to manage the two carrier fluxes so that useful reaction current is maintained while parasitic recombination is limited.
With no insulator, or with a very thin one, charge carriers can move readily between the semiconductor and metal. In a system with a low inherent barrier, this can also allow substantial majority-carrier leakage into the electrocatalyst. When majority and minority carriers meet there, recombination consumes charge that could otherwise contribute to the reaction, reducing photovoltage.
A moderate reduction in tunneling can change that balance. Minority carriers may still transfer to the catalyst quickly enough to sustain the reaction, while the flow of majority carriers—and therefore recombination—is reduced. If the insulator becomes too thick, however, minority-carrier extraction is also hindered. The system then requires additional applied voltage to achieve current, and performance falls.
Results from an n-Si/HfO2/Ni photoanode
The study examined n-Si covered with HfO2 layers of varying thickness and a nickel electrocatalyst. The protective-layer thickness was controlled through atomic layer deposition (ALD), which enabled adjustments on the nanometer scale. Electrochemical testing assessed photocatalytic oxygen evolution under illumination, alongside a ferri-/ferrocyanide redox reaction used to examine photovoltage without the same oxygen-evolution overpotential losses.
Performance followed a non-monotonic trend. As HfO2 thickness increased from zero, photovoltage improved, reached a maximum at approximately 2.1 nm, and then declined at greater thicknesses. For the tested system, the reported photovoltage at that intermediate thickness was 479 mV, compared with about 76 mV for n-Si/Ni without the HfO2 layer. The improvement was also reflected in oxygen-evolution performance.
The researchers also tested an HfO2-protected silicon structure with nickel islands under continuous operation. Its current remained stable for more than two hours in the reported test conditions. This supports the protective role of HfO2 in that configuration, while the thickness-dependent measurements demonstrate that protection and electronic performance must be considered together.
What this means for MIS system design
The practical lesson is that the thinnest protective layer is not automatically the best layer. In MIS photoanodes with relatively low inherent semiconductor-electrocatalyst barrier heights, thickness can act as a design variable for balancing carrier extraction against recombination. A suitable layer may let minority carriers reach catalytic sites while limiting the opposing carrier flux that causes losses.
For engineers and technical teams evaluating photoelectrode designs, this suggests several considerations:
- Assess the interface, not just the materials. Semiconductor, insulator and catalyst choices jointly affect charge transfer, protection and photovoltage.
- Optimize thickness experimentally and analytically. A thickness series can reveal whether the system is limited by recombination at low thickness or by charge extraction at high thickness.
- Match the strategy to the junction. The reported benefit applies to systems with relatively poor inherent barrier heights and insulators that transport carriers by tunneling. Systems that already have a high barrier may respond differently.
- Consider fabrication precision. Nanometer-scale control, such as that provided by ALD in the study, is important when small thickness changes affect tunneling and device behavior.
The work also indicates that tuning an insulating layer may offer an alternative route to improving some systems, rather than relying only on more complex junction architectures. It does not establish that one thickness or material will suit every semiconductor-electrocatalyst combination; the optimum depends on the system’s electronic properties and charge-transfer behavior.
Conclusion
In solar water splitting, a protective insulator is more than a barrier against degradation: its thickness can shape the movement of charge carriers at the semiconductor-electrocatalyst interface. The n-Si/HfO2/Ni example shows that an intermediate layer thickness can improve photovoltage by limiting recombination without excessively restricting useful carrier transfer. For MIS photoanode development, optimizing this balance offers a practical way to improve performance and broaden the combinations of semiconductors and electrocatalysts worth evaluating.