Optimisation of microalgal starch formation for the biochemical production of biobutanol

The current global energy demands have led to an overuse of petroleum resources. In consequence, biofuels have emerged as a suitable and renewable replacement for fossil fuels. A potential replacement for gasoline is biobutanol due to its high heat of combustion, among several other properties (Abdehagh et al., 2014). Biobutanol is produced through ABE fermentation, a renowned microbial process. An efficient feedstock selection, however, is still necessary to overcome current production challenges (Kumar and Gayen, 2011). Although microalgal biomass has largely been examined as a biodiesel feedstock due to its ability to accumulate oil bodies (Markou et al., 2012), only recently has it been considered a promising substrate for ABE fermentation because its structure contains starch, a polymeric carbohydrate. Simple microalgal cultivation strategies such as nitrogen and phosphorus limitation have been
shown to further enhance starch contents (Markou et al., 2012), but such strategies must be effectively implemented to reduce the expected trade-off in growth. Kinetic models capable of predicting cell dynamics during cultivation represent a robust tool for establishing optimized strategies. Nevertheless, most modelling approaches have been constructed under a single-nutrient basis and/or are only able to predict lipid formation. Thus, this work aims to develop a novel predictive multi-parameter kinetic model for the optimization of starch formation during microalgae cultivation. The algal growth rate is co-limited by nitrogen, phosphorus, and the carbon source, and follows a compartmentalized structure considering three intracellular pools: active biomass, starch, and lipids. The model was fitted and successfully validated against experimental
datasets generated from lab-scale cultures of Chlamydomonas reinhardtii CCAP 11/32C grown mixotrophically under various nutrient concentrations. Fitting of the model parameters was carried out through an in-house developed optimization algorithm linking stochastic and deterministic methods avoiding getting trapped in local optima. The predictive performance of the model is exploited by establishing the optimal conditions for maximum starch formation.