RESUMO
In this study, low-cost and high adsorption capacity hydrogels based on gum tragacanth biopolymer (GT or TG) and TiO2 nanoparticles were produced by using glutaraldehyde as the crosslinking agent. These hydrogels were applied in photocatalytic process to remove methylene blue from simulated colored solution. TiO2-Gum Tragacanth hydrogels (TGTH) were characterized by FESEM and FTIR to investigate the surface morphology and functional group of the synthesized hydrogel. Contact angle measurements showed that, the hydrophilicity nature of crosslinked TGTH decreased compared to GT films. The effect of particle size, initial dye concentration, pH of the solution and adsorbent/photocatalyst dosage on the removal efficiency was assessed. The obtained results demonstrated that lower dosage of the prepared TGTH (0.15g/L) outperformed GT (0.2â¯g/L) reaching 87% of dye removal, while GT resulted in 69% of removal. In order to model the cationic dye removal process, a Radial Basis Function Neural Network (RBFNN) was investigated. This network was applied to predict dye removal based on the time duration, initial dye concentration, pH of the solution and TiO2dosage in gum tragacanth hydrogel structure ([TiO2/gum tragacanth hydrogel]0 (g/L)). The performance of the proposed model was validated by several training data. The RBFNN model mostly overlapped with the experimental data due to selecting proper structure and training algorithm.
