RSM and ANN‐based optimization of reactive extraction of propionic acid using tributyl phosphate with both conventional and natural diluents

Propionic acid (PA) has a wide application in various food and chemical industries. In the present work, the reactive extraction of PA from aqueous solutions is done with eco‐friendly natural diluent alsi oil and harmful conventional diluents butanol and benzene, and the tri‐n‐butyl phosphate (TBP) as extractant. Design of experiments was done for the physical and reactive extraction using Box–Behnken design with response surface methodology (RSM). The effect of different factors, such as temperature, initial PA concentration, diluents, extractant, and aqueous‐to‐organic phase ratio, was analyzed. In the physical extraction, the extraction efficiency achieved 75.57% for butanol, 29.91% for benzene, and 47.57% for alsi oil. In the reactive extraction method, the maximum extraction efficiency of 96.89%, 92.54%, and 92.37% for TBP‐butanol, TBP‐benzene, and TBP‐alsi systems, respectively, was achieved. For reactive extraction, optimal conditions were 308 K, 0.1 mol/L initial concentration, a 1:1 volume ratio, and 25% extractant composition predicted by the RSM method and the artificial neural network (ANN) optimization method. ANN shows a better regression parameter ( R 2 = 0.962) than RSM. The higher percentage of extraction efficiency was achieved with conventional diluent butanol; however, the harmless natural diluent alsi oil shows better results, which makes it an alternative to hazardous conventional solvents in the industrial extraction process. These results can help design efficient extraction methods for recovering the PA from the aqueous wastewater stream. The extraction efficiency was achieved as TBP‐butanol > TBP‐benzene > TBP‐alsi oil.