Production, characterization and applications of n-type and p-type Si wafers for electronic devices

In this thesis study, production and characterization studies of Silicon (Si)-based single junction solar cells were carried out from Si substrate production to the final product. In addition, regression analyzes were performed on cutting, lapping and texturing studies to investigate how process parameters affect the results. Moreover, optimization studies were carried out for the surface roughness minimization in cutting and lapping processes and reflectance value in texturing process by using Differential Evolution, Nelder-Mead, Random Search and Simulated Annealing algorithms. The lowest surface roughness values were obtained using 4 m/s wire speed and 0.5 mm/min cutting speed parameters in cutting process experimentally. In lapping process, the optimum surface roughness values were obtained with the parameter set as 50 RPM rotation speed and 60 min lapping time. In texturing studies, the lowest reflectance value was obtained in samples textured for 40 min in 3.5 % sodium hydroxide-5 % isopropyl alcohol solution. After the texturing process, p-n junction was formed by using 1 % boric acid solution and then the Si chips were coated with ITO thin film by plasma-assisted chemical/physical vapor deposition system. Finally, current collectors were deposited on both sides of the Si chips with a screen printing device. After this process, current-voltage measurements of the solar cells were performed by solar simulator and open-circuit voltage, short-circuit current, fill factor and efficiency were obtained 494 mV, 13.95 mA, 80.1 and 8.83, respectively. In the regression analysis, the highest R2 values were obtained with the second order nonlinear rational models in the cutting and lapping studies and the third order nonlinear polynomial model in the texturing process. For all linear and nonlinear regression functions prepared in optimization studies, the best results were obtained based on Differential Evolution and Sumulated Annealing algorithms while worst results obtained based on Random Search algorithm.