Valorization of Biodiesel Into Bio-Based Plasticizers: Optimization of Epoxidized Fatty Acid Isobutyl Ester Production Using Response Surface Methodology and Artificial Neural Network

Biodiesel is a promising green chemical feedstock due to its renewability and sustainability. In this study, bio-based plasticizers, epoxidized fatty acid isobutyl esters (Ep-FABEs), were prepared using biodiesel as a feedstock through a combination of transesterification and the formic acid autocatalytic method. The effects of reaction temperature, reaction time, FA/C═C molar ratio, and H₂O₂/C═C molar ratio on the conversion to oxirane (RCO) during the epoxidation process were investigated using a central composite design. Both response surface methodology (RSM) and artificial neural network (ANN) models were developed to model and optimize the epoxidation process. Comparative analysis revealed that the ANN model demonstrated superior predictive capabilities, with lower mean squared error (MSE), lower mean absolute error (MAE), and a higher coefficient of determination (R²) compared to the RSM model. The ANN model predicted an RCO of 92%, which was closely aligned with the experimental value of 91% under optimized conditions (reaction temperature of 66°C, reaction time of 6.7 h, FA/C═C molar ratio of 0.35, and H₂O₂/C═C molar ratio of 2.70). Additionally, the physico-chemical properties of Ep-FABEs were further analyzed. These findings provide valuable insights into the production of bio-based plasticizers using biodiesel as a feedstock.