May, Matthias M.: GaP- and InP-based Surfaces for Solar Water Splitting. , 2015
Humboldt-Universität zu Berlin
Open Accesn Version (available 01.01.3000)

Solar water splitting is a promising approach to generate hydrogen as a sustainable fuel. In this process, intermittent sunlight is absorbed by a semiconductor. The thereby excited charge-carriers oxidise water to oxygen and reduce protons to hydrogen, effectively storing solar energy in chemical bonds. The challenges of this approach to energy conversion mainly arise due to the peculiarities of the solid-liquid phase boundary, over which the charge carriers have to be transferred. This interface can induce corrosion, charge carrier recombination, or hinder charge carrier extractability to the electrolyte. A detailed understanding of the semiconductor surface and its interaction with the electrolyte is therefore desirable to facilitate an appropriate device design. Gallium phosphide (GaP) and indium phosphide (InP) are two semiconductors which have been investigated in solar water splitting applications as p-type photocathodes for decades, albeit with a limited success for GaP, mainly due to the formation of an unfavourable surface oxide. In this work, I have exposed well-defined (100)-surfaces of GaP and InP to water and oxygen in ultra-high vacuum and investigated their initial interaction with the adsorbates by photoelectron spectroscopy and in situ reflection anisotropy spectroscopy. The results are interpreted in context of recent computational studies in the literature. For InP, the In-rich, (2 x 4) mixed dimer surface, which is more easily accessible and consequently typically employed, is found to behave more favourably upon water adsorption than its counterpart, the P-rich, p(2 x 2)/c(4 x 2) surface. For GaP, the situation is reversed with the P-rich surface, which exhibits an extraordinary stability against the incorporation of oxygen stemming from gaseous water. As the more stable surface reconstruction upon water exposure in comparison to the Ga-rich surface, I conclude that the P-rich surface could be the more suitable starting point for the employment of GaP(100) surfaces in solar water splitting. The dilute nitride GaPN is closely related to GaP and, due to the magnitude and direct-like nature of its band gap, an attractive candidate for a top cell absorber in a tandem configuration with silicon. Here, it was investigated with respect to preparation, electronic structure, and its behaviour upon water exposure. In contact with water adsorbed from the gas phase, its behaviour does not differ from pristine GaP. Step edges of terraces, which are induced by a 2-degree misorientation of the Si substrate, however, do present a point of attack for water molecules on the P-rich surface. Nitrogen-related features of the valence band structure of GaPN are probed and identified with photoelectron spectroscopy. Furthermore, the potential to modify an already mature GaInP/GaInAs photovoltaic tandem by functionalisation of its surface is explored. A transformation of the topmost n+-doped AlInP window layer, followed by in situ Rh catalyst deposition, enables efficient water splitting at a solar-to-hydrogen efficiency of 14%. This surface functionalisation technique could also be suitable for a transfer to other III-V photoabsorbers. The findings presented in this work underline the importance of well-defined surfaces for solar water splitting. Their exact constitution greatly impacts the interaction with water, avoiding or favouring initial corrosion and the formation of unfavourable bond topologies. Well-defined surfaces also facilitate a specific functionalisation reducing charge carrier recombination at the surface as well as corrosion under operating conditions.