Harvesting sunlight, the ultimate renewable power source, in a cost-effective way has been long recognized as a necessary route to meet the global energy challenges. Solar energy can be transformed into electricity by means of photovoltaic devices to supply the electricity grid or alternatively it can be stored in batteries. However, given the limitation of current battery technology, methods to store sunlight for long duration and at large scale are still under development. Instead, solar energy can be directly converted into chemical fuels, for instance hydrogen from water splitting or carbon-based products from carbon dioxide reduction, using a semiconductor-liquid junction (SCLJ) in a photoelectrochemical (PEC) cell. PEC water splitting for sustainable hydrogen production has garnered growing attention ever since the first demonstration of photocatalytic water splitting on semiconductor surface. However, light harvesting through a single-absorber system is limited by the large semiconductor band gap energy required to afford overall water splitting. In contrast, a dual absorber PEC tandem cell consists of photoanode and photocathode with different band gaps that can complementarily utilize a wide range of solar spectrum. Despite considerable efforts to boost the solar-to-hydrogen (STH) conversion efficiency for unassisted solar water splitting, the performance and stability of PEC tandem cell devices are still insufficient for commercialization. To further improve the performance of a PEC tandem cell, it is of great importance to have rational design of photoelectrodes and comprehensive understanding of the SCLJ where the photo-synthetic reaction is happening. The work presented in this thesis aims at developing and utilizing novel operando spectroelectrochemistry to characterize the semiconductor/electrolyte interface. These interfacial characterization techniques provide fundamental understanding of the SCLJ and identify the performance bottleneck of the photoelectrodes, which establish the roadmap for optimizing the solar fuel production process. Taking hematite as a model photoanode material, we first explore the potential of intensity-modulated photovoltage spectroscopy (IMVS) for characterization surface recombination process in Chapter 2. Then in Chapter 3 we study the effect of electrolyte pH on the performance of metal oxide photoanode by a set of operando spectroelectrochemical tools. In Chapter 4 we evaluate the charge carrier behavior of the emerging spinel copper ferrite photoanode whose features limiting the solar water oxidation response are revealed for the first time. In addition, to deal with the late onset potential that restrain spinel zinc ferrite photoanodes for tandem cell application, we demonstrate that a sub-nanometer aluminum oxide layer operates as an efficient surface passivation agent improving the onset potential by 100 mV in Chapter 5. Likewise, guided by spectroelectrochemical measurements, we establish a new benchmark on lanthanum iron oxide photoanode by rational surface and bulk engineering in Chapter 6. Finally, combining computational methods and operando spectroelectrochemistry, we identify for the first time the catalytic active sites and surface trap states on chalcopyrite photocathodes in Chapter 7.