| Abstract | In the last decade, there have been major developments in the generation of green fuel sources to replace fossil fuels with the aim of reducing greenhouse gas emissions. Hydrogen is one of these sources that can be used either in the power generating fuel cells or as a direct combustion fuel. To produce large hydrogen on a large scale successfully, the whole process plant ought to utilise clean energy and avoid fossil derived sources as much as possible.
This study focuses on a large scale solar driven hydrogen production in form of a closed sulphur-ammonia thermochemical water-splitting cycle. The cycle utilises both the thermal and quantum components of free and widely available solar irradiation to drive the oxygen and hydrogen generation respectively. The only fluid input to the cycle is water, and only output is hydrogen and oxygen gases. However, in the most common type of sulphur-ammonia cycle, the hydrogen producing step is undertaken by a high energy demanding electrolyser at high operating temperature.
In this research, the electrolyser step is replaced by a solar driven photocatalytic reactor step, as a mean to make the cycle more economical and greener. To begin with, an energy efficient and economically viable photoreactor based on LED lightning, which could mimic the quantum part of solar irradiation, was designed and benchmarked. Light intensity and flux analysis of the LED lightning indicated that cool
LEDs would be the most suitable option for a reactor using visible light driven cadmium sulphide (CdS) photocatalyst particles. To further increase the photostability of the CdS catalyst and improve its photonic yield, a suitable cocatalyst was searched for. Among the cocatalysts under investigation, cobalt phosphides (CoxP) was found to be most promising for an increased hydrogen evolution at a reduced cost, if compared with more common noble metals such as Pt and Pd. Different synthesis method and synthesis parameters showed different CoxP composition could be tuned and optimised, resulting in different hydrogen evolution yields. The CoP synthesised through an organometallic method loaded on CdS by sonication, showed exceptionally high hydrogen yield compared to the other tested CoxP/CdS, and even more than the most referred Pt/CdS, when an aqueous lactic acid
solution was used in the reactor. All investigations involved optimising different working parameters such as; cocatalyst loading on CdS, reactor particle loading, solution concentration variation and irradiation flux. For a solar driven sulphur ammonia thermochemical cycle, the working fluid would be ammonium sulphite rather than lactic acid and therefore, the same various CdS photocatalytic composites and various working parameters were tested again but using ammonium sulphite solution instead. Once again, the organometallic synthesised CoP/CdS composite performed excellent with a comparable hydrogen yield than the more
understood and elaborated Pt/CdS.
For the photocatalytic reactions of both lactic acid (aq.) and ammonium sulphite (aq.) a mathematical predictive model of the hydrogen production was developed. The model was based on a pseudo-steady state approach of the Langmuir-Hinshelwood
adsorption isotherm, which further incorporated the total radiative flux effect into the model. The Langmuir-Hinshelwood part reflected the mechanism of the surface adsorption/desorption and the surface reaction rates whereas the irradiation part of the model was based on an approximate solution of radiation field theory (RTE) and the derivation of geometrical positions of the particles in a cross-section of the reactor which was further expanded to incorporate the total visible radiative flux density received by the particles in the whole photoreactor volume. One of the main advantages of the model was the successful incorporation of both the measured photocatalyst’s optical scattering and absorption coefficients in the relevant LED output range of 410-500 nm, increasing the accuracy of the model. The obtained model parameters were then successfully validated for a range of photocatalytic hydrogen generation experiments conducted at various reaction conditions. These results show that the developed model can be used to predict photocatalytic hydrogen production satisfactory regardless of the radiation type, reactor size or catalyst particle, with minor computational effort or use of any commercial software.
The obtained model information is suggested to work as a supporting aid for any photocatalytic reactor scale-up, which can easily be altered to either reduction product (hydrogen) or oxidation product if the most dominating initial reactions are known. A feasibility study of a complete thermochemical cycle was also done, where implementing several groups of parallel hydrogen photoreactors was suggested as the best option for large scale hydrogen production with the main purpose of replacing the energy demanding electrolyser in the thermochemical cycle. The parallel photoreactor configuration can be easily implemented and be operationally efficient in any thermochemical cycle configuration. Finally, it was shown that an increased hydrogen efficiency (due to an increased photonic energy conversion) and an improved hydrogen economy of the cycle (due to use of cheaper catalyst & reactor materials and a reduced power consumption) can be achieved. |
|---|