Silicon carbide (SiC) is a chemically stable and photocatalytic semiconductor material. SiC can be grown by different atomic configurations resulting in a large number of polytypes. Among SiC polytypes, 3C-SiC has the smallest band gap (2.3 eV), and thus 3C-SiC is capable to absorb significant part of solar light. In addition, conductive type of SiC can be controlled by doping. Photocatalytic materials with p-type conduction operate as cathode and reduce hydrogen in water. Cathodic operation prevents oxidation of photocatalytic materials. Therefore, p-type 3C-SiC will be an efficient and durable photocatalyst.
However, bulk crystal growth of 3C-SiC is very difficult because this polytype is stable only in growth at much lower temperature than sublimation of SiC, which is a typical condition for bulk SiC growth. This difficulty makes 3C-SiC growth only by chemical vapor deposition (CVD) on Si or 4H- or 6H-SiC substrates. Such the heteroepitaxial growth induces large number of structural defects or polytype mixing owing to lattice mismatch or inheritance of the substrate polytype.
For our recent 3C-SiC photocathode fabrication, we employed p+-type 4H-SiC substrates with the (0001) Si-face inclined 0.7° to <1120> direction. Growth on the 4H-SiC substrates with quasi on-axis Si-face prevents polytype mixing and p+-type conductivity of the substrate facilitates ohmic contact formation on the substrate side. We performed 3C-SiC growth on the substrates by CVD with silane and propane as source gases and in an Al acceptor doping condition. A photograph of the grown crystal is shown in Fig. 1(a). After the growth, we fabricated ohmic contact on the substrate side of the crystals. The 3C-SiC surface is active surface as the photocathode. In addition, we formed Pt or Pd cocatalysts on the 3C-SiC surface to enhance hydrogen generation by the photocathodes. The schematic of the photocathode structure is shown in Fig. 1(b).
We measured photocurrent-potential characteristics for the photocathodes in 1 M H2SO4 aqueous solution using chopped solar simulator light with 100 mW/cm2. The results are shown in Fig. 1(c). For bare surface, cathodic photocurrent was observed below 0.6 V vs SCE, and it reached -1.4 A/cm2 at -1 V vs SCE. On the other hand, with cocatalysts, photocurrents were observed at higher potential, and in the case of with Pt cocatalysts, it was observed from 0.9 V vs SCE. In addition, at -1 V vs SCE, the photocurrent was -2.3 A/cm2.
In solar to hydrogen conversion, the Faraday efficiencies of the photocathodes were near to 100% and the photocathodes were not degraded during experiments. Solar to hydrogen conversion efficiency for the photocathode with Pt cocatalysts was 0.52 %. Considering such a high efficiency and possibility of further improvement, SiC is a very promising material for the solar to hydrogen conversion technology.