Studies on the Applicability of Artificial Neural Network (ANN) in Evaluation of Photocatalytic Performance of TiO2 thin Film Doped by SiO2
Ehsan Rahmani, Dariush Jafari, Ali Ahmadpour, Mojtaba Zebarjad and Hossein Rahmani
Affiliation: Department of Chemical Engineering, Bushehr Branch, Islamic Azad University, Bushehr, Iran.
Keywords: Photocatalytic activity, thin film, Artificial Neural Network, Sol-gel, Multi-layer Perceptron, Simulation
Nanocrystalline films of TiO2 and TiO2:SiO2 with high photocatalytic activity were prepared on glass substrates
by the application of sol-gel method. Then the films were subjected to a high temperature treatment at 500°C,
which resulted in growth of TiO2 crystals. Afterwards the TiO2:SiO2 films were in contact with an aqueous solution (10
mg.L-1) of methyl orange (MO) and irradiated under UV. The resulted films showed a high photocatalytic activity. In the
current study the photocatalytic activity of TiO2 crystals was studied by an Artificial Neural Network (ANN). This was
achieved by predicting the concentration of MO in various values of SiO2 concentration and time of degradation. In order
to perform the modeling, Multi-layer Perceptron (MLP) network was used in this work, with its learning algorithm being
Levenberg-Marquardt (LM). The outcome of modeling showed that there was an excellent agreement between the results
of simulation and the data obtained from the experiments. It is worth noting that in the current work, the methods applied
in recent papers and patents for the preparation of nanocrystalline films and determination of their photocatalytic performance
and also modeling of such processes have been studied.
Rights & PermissionsPrintExport