Valorization of fish bone waste as novel bioflocculant for rapid microalgae harvesting: Experimental evaluation and modelling using back propagation artificial neural network

Abstract

Harvesting of microalgae biomass is identified as one of the bottlenecks in microalgae biofuel industry due to expensive and energy-intensive dewatering technologies. Alternatively, flocculation process using bioflocculants have given much attention in recent years as green substitutes over chemical flocculants. In this study, bioflocculant was extracted from waste fish bone using mild acid to harvest the freshwater microalgae, Chlorella vulgaris. The optimum flocculation occurred at pH of 9.8 and 50 °C using fish bone bioflocculant which led to flocculation efficiency of 97.65%. To predict complex processes such as microalgae flocculation, artificial neural network (ANN) was employed. Bayesian regularization model with a topology of 2-10-1 showed high correlation coefficients, R2 of more than 0.98, which indicated that the model was significant and robust in identification of the optimum conditions. Characterizations of fish bone bioflocculant and biofloc confirmed the involvement of potassium and other cations as well as carbohydrate and protein substances to flocculate C. vulgaris cells, employing sweeping and charge neutralization as key mechanisms. This finding proposed a valuable reference for practical and rapid harvesting of microalgae using low-cost bioflocculant and the ANN algorithm can be applied in microalgae processing industries for making crucial assessments regarding the process operating conditions.