In photoelectrochemical (PEC) hydrogen production, the semiconductor must absorb light and generate charge, while its surface supports hydrogen evolution and withstands contact with electrolyte. Those demands can conflict: a protective coating may impede charge transfer, while a catalyst alone may leave the semiconductor vulnerable to corrosion. A study of a MoS2 photocathode built on gallium indium phosphide (GaInP2) demonstrates how one thin-film interface can address both challenges.
The tested MoS2 layer acted as both a hydrogen-evolution catalyst and a protective barrier. Compared with a PtRu-catalyzed GaInP2 device, it had a more positive photocurrent onset potential and remained operational much longer under the study’s acidic test conditions. The results also point to an important design consideration for engineers: the structure of the catalyst–semiconductor interface can matter as much as the catalyst’s intrinsic activity.
Why combine protection and catalysis?
GaInP2 has optical and electronic properties suited to use as a large-bandgap absorber, including as the top absorber in a tandem solar water-splitting device. Its limitation is durability: in acidic electrolyte, the semiconductor can corrode. Protecting it typically involves adding a coating, and if that coating is not sufficiently active for the hydrogen evolution reaction (HER), a separate cocatalyst may also be needed.
Multiple layers bring practical trade-offs. They add interfaces to engineer and can reflect or absorb light that would otherwise reach the semiconductor. In a tandem device, that lost light can constrain the current available from the absorber. A thin, earth-abundant material that provides both protection and catalytic activity could therefore simplify the surface architecture and reduce reliance on precious-metal catalysts. MoS2 is a candidate because its edge sites can catalyze HER and the material can be used as a protective layer.
How the MoS2 photocathode was structured
The study compared two photocathodes made from the same type of pn+-GaInP2 semiconductor: one coated with MoS2 and one carrying platinum-ruthenium (PtRu) nanoparticles. The MoS2 coating was formed by depositing a thin molybdenum film and then converting it through sulfidization. The resulting structure was a conformal coating, meaning it covered the semiconductor surface continuously rather than appearing as separated catalyst particles.
By contrast, the PtRu catalyst consisted of nanoparticles with partial surface coverage. Surface analysis supported this distinction: the MoS2-modified sample showed no detectable Ga, In, or P signals before testing, consistent with a continuous overlayer, while these semiconductor elements were detectable on the PtRu sample.
This comparison goes beyond asking which catalyst is more active in isolation. A PEC device’s measured performance also depends on how charges move across the semiconductor–catalyst interface and whether the coating protects the absorber without imposing excessive optical losses.
Performance and durability in acidic electrolyte
In three-electrode tests using 3 M sulfuric acid under illumination calibrated to one sun, the MoS2-coated photocathode reached an onset potential of 0.91 V versus the reversible hydrogen electrode (RHE), measured at a photocurrent density of 1 mA/cm2. The PtRu comparison reached 0.83 V versus RHE. A more positive onset indicates that the photocathode begins producing appreciable hydrogen at a more favorable potential.
The measured light-limited current densities were approximately 13 mA/cm2 for MoS2 and 15 mA/cm2 for PtRu. This difference matters: the MoS2 coating did not maximize the current in this comparison, and the study identified parasitic absorption by the more conformal Mo/MoS2 layer as a possible contributor. The result illustrates a recurring engineering balance between surface protection, catalyst coverage, and light reaching the absorber.
Durability separated the two structures more sharply. At a fixed operating potential, the MoS2 photocathode maintained its light-limited current for 100 hours, when testing was stopped, with about a 10% loss in initial current density. The PtRu device experienced a sharp current drop after about six hours. Post-test X-ray photoelectron spectroscopy found no exposed GaInP2 or underlying GaAs signals on the MoS2 sample. On the PtRu sample, the analysis indicated Ga oxide formation and exposure of the GaAs substrate, consistent with corrosion of the semiconductor layer.
What interface measurements revealed
The researchers used transient photoreflectance spectroscopy (TPR) to investigate charge behavior near the photocathode surface. TPR tracks changes in reflected light after optical excitation, providing information about photo-generated carriers and the electric fields at the interface.
The spectra indicated that electrons remained localized near the surface in the uncoated and PtRu-coated GaInP2 structures. In the MoS2-coated device, a feature associated with those surface electrons was barely visible, which the study interpreted as evidence of more effective electron transfer from GaInP2 into MoS2. The proposed explanation is that the conformal MoS2 interface avoids a surface energy well that can retain electrons, allowing them to reach the catalyst more readily.
This helps explain why the MoS2 device showed a better onset potential even though MoS2 has lower intrinsic dark catalytic activity for HER than PtRu. Photocathode performance is not determined by catalyst activity alone; charge separation and transfer across the complete interface also influence the operating result.
Practical implications for PEC design
For industrial automation and energy technology teams evaluating solar-hydrogen systems, the findings are a materials-design result rather than evidence of a production-ready hydrogen source. The tests were conducted on single-absorber photocathodes in a laboratory PEC setup, and the 100-hour result establishes performance only over that test period and under those conditions.
Even so, the comparison highlights several considerations for development work:
- Assess the full interface. Measure charge transfer and surface behavior alongside catalyst activity.
- Balance coverage with optical throughput. A continuous protective layer can improve stability and electron transfer, but may absorb light and reduce current.
- Compare durability at relevant operating conditions. A strong initial current is not sufficient if corrosion rapidly limits operating life.
- Optimize the architecture for integration. The study identifies antireflective coatings, light-trapping approaches, and thinner protection layers as possible routes to improve current.
The GaInP2/MoS2 result shows that an earth-abundant catalytic protection layer can combine favorable onset behavior with substantially longer stability than the PtRu comparison, while also revealing an optical trade-off that remains to be addressed. For future PEC development, the practical lesson is to engineer the semiconductor–catalyst interface as an integrated system: protection, charge transfer, catalytic activity, and light management all contribute to useful hydrogen production.