Volume 11 Issue 3 ( September 2024)

Dye-sensitized solar cell (DSSC) is significantly promising as a third-generation solar cell, offering an inexpensive and simple fabrication process. The performance of DSSC has been proven to improve through the optimization of DSSC parts such as photoanode. Therefore, this research aimed to optimize photoanode part of DSSC using random forest method through machine learning predictions. The optimization process was carried out through combinations of single and double layers to identify top- performing configurations. The two top prediction outcomes were A3 with DN-F05 0.8 mM) and double layer configurations combining A3+A2 with DNF-05 0.8 mM. Furthermore, the two combinations were fabricated and characterized using UV-VIS, FT-IR, and I-V test meters. The results of UV-VIS measurement indicated a typical peak of DNF-05 at a wavelength of 482 nm, while FT-IR spectra showed the presence of TiO₂ groups and dyes. Based on the results, the highest power conversion efficiencies of 3.062% were obtained in single layer.

Keywords: Dye-sensitized solar cell, Random forest, DN-F05 dye.