Sustainable biodiesel production from Brachychiton populneus seed oil using Fe-doped ZnO nanocatalyst: process optimization, thermodynamics, and machine learning modeling
摘要
With the increasing demand for sustainable and renewable energy resources, the production of biodiesel from non-edible feedstocks with the help of efficient, heterogeneous nanocatalysts has become more relevant. The present study aimed to synthesize Fe-doped ZnO nanoparticles and use it as a nanocatalyst for the production of biodiesel from non-edible seed oil of Brachychiton populneus according to response surface methodology (RSM) to optimize the synthesis. The catalyst structure, optical, and thermal properties were evaluated by XRD, FTIR, UV–Vis, TGA, SEM, TEM, and EDX techniques. The XRD data were used to confirm the phase formation of hexagonal wurtzite structure of ZnO with an average crystallite size of ~ 10.6 nm and lattice strain. Initially, UV–Vis spectroscopy was used to determine an optical band gap of 3.28 eV in the material and thermogravimetric analysis was used to show that the synthesized nanocatalyst has improved thermal stability. Suitable conditions for the highest yield of biodiesel (96%) were identified as 6:1 methanol-to-oil ratio, 2 mass% catalyst loading, 75 min reaction time, and reaction temperature of 55 °C. Pseudo-first-order reaction was suggested by kinetic analysis and the reaction activation energy was calculated as 34.5 kJ mol–1, and thermodynamic studies confirmed the endothermic and feasibility of transesterification reaction. The conversion of the fatty acids to fatty acid methyl esters (FAMEs) was confirmed by GC–MS, FTIR, 1H-NMR, and 13C-NMR analysis. Additionally, the machine learning models, especially random forest, performed better in predicting data (R2 = 0.935) compared with the traditional RSM. The results show that Fe-doped ZnO nanocatalyst is an alternative sustainable and efficient process for producing high-quality biodiesel using a non-edible feedstock. Recyclability of the catalyst, engine performance evaluation, and pilot-scale industrial applications should be investigated in the future.