Response Surface Methodology and Artificial Neural Network Optimization and Modeling of the Saccharification and Fermentation Conditions of the Polyhydroxybutyrate from Corn Stover

Abstract

In this study, a response surface methodology (RSM) and artificial neural network (ANN) were used for the modeling and optimization of the enzymatic hydrolysis and fermentation conditions of the polyhydroxybutyrate (PHB) production from pretreated corn stover. Using three factors in each process, a face-centered central composite Design (FCD) was employed to optimize the reducing sugar yield (RSY), biomass, and PHB production of the saccharification and fermentation, respectively. The optimal conditions (RSY of 0.539 g/g) for enzymatic saccharification were 9.776% w/v solids loading, 83.656 FPU/g enzyme loading, and 18 h saccharification time. It was found that solids loading had the greatest impact on RSY. On the other hand, the optimum fermentation medium conditions were 19.727 g/g C/N ratio, 18.421 g/g C/P ratio, and 9.623 mL of trace elements solution (TES) per liter of fermentation medium, with a maximum PHB concentration of 7.083 g/L. During the fermentation of corn stover hydrolysate medium, the C/N ratio positively affected the growth and the PHB productivity of the microorganism. Moreover, the reliability of the RSM and ANN methods in modeling and predicting the RSY, biomass, and PHB concentrations were compared. The ANN showed higher accuracy compared to the RSM models. This study proved that the microorganism could synthesize PHB under optimal fermentation conditions using the reducing sugars from the enzymatic saccharification of pretreated corn stover.