Response surface methodology (RSM) and artificial neural network (ANN) approach to optimize the photocatalytic conversion of rice straw hydrolysis residue (RSHR) into vanillin and 4-hydroxybenzaldehyde

Abstract

Abstract Effective use of waste lignin is always a challenging task, technologies have been applied in the past to get value-added compounds from waste lignin. However, the existing technologies are not economical and efficient to produce the value-added chemicals. Alkali soluble lignin from rice straw hydrolysis residue (RSHR) is subjected to photocatalytic conversion into value-added compounds. Photocatalysis is one of the multifarious advanced oxidation processes (AOPs), carried out with TiO2 nanoparticles under a 125 W UV bulb. Gas chromatography mass spectroscopy (GCMS) confirmed the formation of vanillin and 4-hydroxybenzaldehyde. RSM and ANN techniques are adopted to optimize the process conditions for the maximization of the products. The response one (Y 1) vanillin (24.61 mg) and second response (Y 2) 4-hydroxybenzaldehyde (19.51 mg) is obtained at the optimal conditions as 7.0 h irradiation time, 2.763 g/L catalyst dose, 15 g/L lignin concentration, and 14.26 g/L NaOH dose for alkali treatment, suggested by face-centered central composite design (CCD). RSM and ANN models are statistically analyzed in terms of RMSE, R 2 and AAD. For RSM the R 2 0.9864 and 0.9787 while for ANN 0.9875 and 0.9847, closer to one warrant the good fitting of the models. Therefore, in terms of higher precision and predictive ability of both models the ANN model showed excellence for both responses as compared to the RSM model.