| Hydrogen (H2) is a promising alternative energy carrier because of its high energy density, feasibility of large-scale production, and free emission of greenhouse-gas upon combustion. The direct conversion of abundant and widespread solar energy into clean and storable H2 via photoelectrochemical (PEC) water splitting represents an attractive way to solve the ever-increasingly serious energy crisis and environmental problems. For practical application, a solar-to-hydrogen (STH)efficiency of at least 10% will be necessary for the PEC cell, by which semiconductors with optical absorption edge of longer than 530 nm are required (the bandgap of semiconductors should be smaller than 2.34 eV), as calculated from the standard solar spectrum of AM 1.5 G 100 mW cm-2. Several n-type semiconductors,such as α-Fe2O3 (15% of theoretical STH efficiency) and Ta3N5 (15.9% of theoretical STH efficiency), and several p-type phosphides, arsenides, sulfides, selenides and Si semiconductors, are promising candidates as photoelectrodes for PEC water splitting.The inefficient or unstable n-type photoanodes are now the limiting factors in PEC cells for practical application, after the exploration of efficient p-type photocathodes.After several decades of intense investigation, the practical STH efficiency achieved by α-Fe2O3 photoanodes is farm from 10%,which may be due to the intrinsic physical and chemical properties of α-Fe2O3. Ta3N5 emerges as a promising photoanode candidate for solar water splitting,with a bandgap of 2.0 eV and suitable band edge positions for unassisted water splitting. Improving the charge separation efficiency, surface charge injection efficiency, lowering the external bias, and improving the PEC stability of Ta3N5 photoanodes are general strategies to engineering Ta3N5 photoanodes to realize their practical applications.In this dissertation, we aim to improve the PEC stability and solar energy conversion efficiency of Ta3N5 photoanodes, by studying the reaction mechanism of PEC water splitting. After that, we aim to improve the PEC performance of other oxynitrides with high theoretical STH efficiency and suitable band edge positions.The main research contents are as follows:Improving the PEC stability of a Ta3N5 photoanode by increasing the surface charge injection efficiency. The photo-generated holes can be consumed by water oxidation and N3- ions oxidation in Ta3N5. When the consumption of holes by water oxidation is inefficient, holes accumulation will be alternatively alleviated through self-oxidative decomposition of Ta3N5. Accelerating the water oxidation kinetics (increasing the surface charge injection efficiency) is pivotal to improve the PEC stability of Ta3N5 photoelectrodes. The most stable Ta3N5 photoanode for PEC water splitting has been experimentally demonstrated in the world at that time, by employing uniformly distributed and cheap Co3O4 nanoparticles as the water oxidation catalysts. A plateau photocurrent of 6 mA cm-2 and a photocurrent of 3.18 mA cm-2 at 1.2 VRHE have been obtained on the Co3O4/Ta3N5 photoanode under AM 1.5 G 100 mW cm-2 simulated sunlight.Improving the plateau photocurrent of a Ta3N5 photoanode by increasing the charge separation efficiency. The presently achieved plateau photocurrent on Ta3N5 photoelectrodes under AM 1.5 G 100 mW cm-2 is far from the theoretical maximum (ca. 12.9 mA cm-2), which is possibly due to the poor charge separation efficiency in Ta3N5 photoelectrodes induced by defect states in the bulk Ta3N5. We show that the introduction of Ge (5%) in Ta3N5 increases the charge separation efficiency by a factor of 2 and increases the plateau current by a factor of 3.2. The intentionally added Ge will volatilize during the synthesis of Ta3N5, which may reduce the amount of N vacancies and thereby promotes the electron transport in Ta3N5 photoanodes by acting as the fluxing agent.Improving the plateau photocurrent of a LaTiO2N photoanode by increasing the charge separation efficiency. LaTiO2N can absorb comparable wavelengths of sunlight and possesses similar band edge positions as Ta3N5, which is a promising photoanode candidate. However, LaTiO2N photoanode is always inefficient. The above study on Ta3N5 photoanodes suggests that poor electron transport associated with insufficient inter-particle connection and abundant grain boundaries contributes to the suppressed PEC activities of LaTiO2N photoanodes. By establishing highly crystalline porous LaTiO2N particles and superior inter-particle connectivity with reduced density of grain boundaries among the film particles, a record plateau photocurrent of 6.5 mA cm-2 and a photocurrent of 4.45 mA cm-2 at 1.23 VRHE have been demonstrated on the Co3O4 modified LaTiO2N photoanode under AM 1.5 G 100 mW cm-2 simulated sunlight.Effects of N contents in TaON on charge separation efficiency. Oxynitrides are non-stoichiometric compounds (derived from O substitutes for N), however, the effects of O contents (or N contents) in oxynitride photoelectrodes on charge separation are unknown. We show that the photoresponse of TaON photoanodes for solar water splitting varies with the N contents in them. The film conductivity and charge carrier concentration of TaON increase with increasing O contents, which are favorable for charge separation. The space charge layer thickness of TaON decreases with increasing O contents, which is harmful for charge separation. Therefore, a moderate carrier concentration in TaON that enables a relatively large volume of space charge layer with respect to that of the bulk while still affords sufficient film conductivity is highly desirable for efficient charge separation. |