Özet:
The objective of this thesis is to develop active and stable dye sensitized and ionic liquid (IL) encapsulated photocatalysts for hydrogen production, and to investigate the roles of photocatalysts constituents on hydrogen production rate. First, comprehensive experimental datasets for photocatalytic (PWS) and photoelectrochemical (PECWS) water splitting were constructed from literature while a large dataset for water solubility in ILs were created computationally using DFT and COSMO-RS. Then, these datasets were analyzed by machine learning (ML); association rule mining (ARM), decision tree (DT), random forest (RF) and deep learning (DL) were implemented in R and Python environment. In PWS analysis, the clear trends and the high fitness of the models constructed, especially those involved the band gap, indicates that PWS over perovskites is predictable but the sound solutions like ion doping, use of co-catalyst, or use of sacrificial donors did not help as much as desired. The prediction accuracies of DL and classification accuracy of DT models for the water solubility in ILs were also high, and the properties of anionic parts of ILs are more influential for water solubility. The band gap of semiconductors used in PECWS were also predicted successfully even though the prediction accuracy of photocurrent density was not satisfactory. In the experimental part of the thesis, the enhancement of photocatalytic hydrogen production over 1 wt% Pt/TiO2 sensitized by N719 dye and coated by a thin layer of ionic liquid ([BMIM][BF4]) has been investigated; SEM-EDAX, CTEM, FTIR, XRD, and UV-Vis characterization of the photocatalysts, and electrochemical analysis of the respective photoelectrodes were performed. The IL coating increased the performance of 1%Pt/TiO2 by providing better charge transfer between the photocatalyst and the aqueous reaction medium while simultaneously preventing the recombination of photogenerated electron-hole pairs; the improvement was much higher with the use of IL and N719 together. The performance of 1% Pt/SrTiO3 was also tested but the reproducible results could not be obtained. Finally, the factors related to the structure of the reactor and operational conditions (like dead volume, gas-liquid interfacial area, sweep gas flowrate and stirring speed) were also found to be influential over the hydrogen production rate.
