DOE OSTI2013
The objectives of this project, covering two phases and an additional extension phase, were the development of thin film-based hybrid photovoltaic (PV)/photoelectrochemical (PEC) devices for solar-powered water splitting. The hybrid device, comprising a low-cost photoactive material integrated with amorphous silicon (a-Si:H or a-Si in short)-based solar cells as a driver, should be able to produce hydrogen with a 5% solar-to-hydrogen conversion efficiency (STH) and be durable for at least 500 hours. Three thin film material classes were studied and developed under this program: silicon-based compounds, copper chalcopyrite-based compounds, and metal oxides . With the silicon-based compounds, more specifically the amorphous silicon carbide (a-SiC), we achieved a STH efficiency of 3.7% when the photoelectrode was coupled to an a-Si tandem solar cell, and a STH efficiency of 6.1% when using a crystalline Si PV driver. The hybrid PV/a-SiC device tested under a current bias of -3~4 mA/cm 2 , exhibited a durability of up to ~800 hours in 0.25 M H 2 SO 4 electrolyte. Other than the PV driver, the most critical element affecting the photocurrent (and hence the STH efficiency) of the hybrid PV/a-SiC device was the surface energetics at the a-SiC/electrolyte interface. Without surface modification, the photocurrent of the hybrid PEC device was ~1 mA/cm 2 or lower due to a surface barrier that limits the extraction of photogenerated carriers. We conducted an extensive search for suitable surface modification techniques/materials, of which the deposition of low work function metal nanoparticles was the most successful. Metal nanoparticles of ruthenium (Ru), tungsten (W) or titanium (Ti) led to an anodic shift in the onset potential. We have also been able to develop hybrid devices of various configurations in a monolithic fashion and optimized the current matching via altering the energy bandgap and thickness of each constituent cell. As a result, the short-circuit photocurrent density of the hybrid device (measured in a 2-electrode configuration) increased significantly without assistance of any external bias, i.e. from ≤1 mA/cm 2 to ~5 mA/cm 2 . With the copper chalcopyrite compounds, we have achieved a STH efficiency of 3.7% in a coplanar configuration with 3 a-Si solar cells and one CuGaSe 2 photocathode. This material class exhibited good durability at a photocurrent density level of -4 mA/cm 2 (“5% STH” equivalent) at a fixed potential (-0.45 V RHE ). A poor band-edge alignment with the hydrogen evolution reaction (HER) potential was identified as the main limitation for high STH efficiency. Three new pathways have been identified to solve this issue. First, PV driver with bandgap lower than that of amorphous silicon were investigated. Crystalline silicon was identified as possible bottom cell. Mechanical stacks made with one Si solar cell and one CuGaSe 2 photocathode were built. A 400 mV anodic shift was observed with the Si cell, leading to photocurrent density of -5 mA/cm 2 at 0V RHE (compared to 0 mA/cm 2 at the same potential without PV driver). We also investigated the use of p-n junctions to shift CuGaSe 2 flatband potential anodically. Reactively sputtered zinc oxy-sulfide thin films was evaluated as n-type buffer and deposited on CuGaSe 2 . Ruthenium nanoparticles were then added as HER catalyst. A 250 mV anodic shift was observed with the p-n junction, leading to photocurrent density at 0V RHE of -1.5 mA/cm 2 . Combining this device with a Si solar cell in a mechanical stack configuration shifted the onset potential further (+400 mV anodically), leading to photocurrent density of -7 mA/cm 2 at 0V RHE . Finally, we developed wide bandgap copper chalcopyrite thin film materials. We demonstrated that Se can be substituted with S using a simple annealing step. Photocurrent densities in the 5-6 mA/cm 2 range were obtained with red 2.0eV CuInGaS 2 photocathodes. With the metal oxide compounds, we have demonstrated that a WO 3 -based hybrid photoelectrode was feasible. Specifically, we showed that WO 3 paired with an a-Si tandem solar cell can generate short circuit photocurrent density of 2.5 mA/cm 2 , equivalent to STH efficiency of 3.1%. Long-term durability tests demonstrated WO 3 ability to split water over extended periods, for up to 600 hours at current density levels of 2.0-2.5 mA/cm 2 . Efforts have been done to decrease WO 3 bandgap using foreign elements incorporation. We did not manage to reduce the bandgap of WO 3 with this method. However, more promising results have been achieved with bilayered systems, where only the top part of WO 3 films was modified. Also, we have demonstrated that alloying WO 3 with CuO can form 2.2eV bandgap CuWO 4 . Incorporating conductive carbon nanotubes in CuWO 4 reduced its intrinsic bulk resistance. Saturation photocurrent densities in the 0.4-0.5 mA/cm 2 range were achieved. Recently, in collaboration with University of Texas at Arlington, we have identified new quaternary metal oxides with CuWO 4 as primary material host. Our experimental work on ceramics confirmed the theoretical calculations that crowned bismuth as a possible candidate to improve CuWO 4 water splitting efficiency.