Bioconversion process of source-separated organic waste for ethanol production

Production of biofuel such as ethanol from lignocellulosic biomass is a beneficial way to meet sustainability, energy security, and environmental goals. Lignocellulosic biomass such as source-separated organic (SSO) waste is particularly attractive since it is widely available, often at a negative cost, reduce the land depletion from using food-based biomass for ethanol production and reduce the amount of generated waste. Therefore, in order to meet the future fuel demands and cope with increasing volume of municipal waste this study was a first attempt to use SSO as a feedstock for ethanol production. The main objectives of the study were: a) to compare standard and modified celluloseorganic- solvent-based lignocellulosic fractionation (COSLIF) pretreatment of SSO waste for ethanol production in terms of enzyme savings, sugar formation and ethanol yields; b) to produce ethanol from SSO by using modified COSLIF pretreatment and fermentation with two different recombinant strains: Z. mobilis 8b and S. cerevisiae DA2416; and c) to develop experimental kinetic model capable of predicting behavior of batch SSCF on SSO waste with different SSO substrate concentrations using Berkeley Madonna program. Based on the obtained results, it was found that SSO is an excellent feedstock material for ethanol conversion. The efficiency of modified COSLIF pretreatment was improved by 20% compared to standard method using ethanol washing of pretreated SSO samples during the experimental procedures instead of acetone. On average, glucose yield from SSO samples pretreated by modified COSLIF was about 90% compared to 10% for untreated samples. S. cerevisiae DA2416 outperformed Z. mobilis 8b on ethanol yields during the fermentation process, with 0.50 g ethanol/g potential sugar fed on SSO in less than 5 days, with a 96% cellulose conversion, totalling in 150 g/L ethanol produced. A kinetic model with newly integrated values of experimentally defined SSO feedstock constants was proven to predict the ethanol yield accurately with substrate concentration ranges of 20 g/L - 50 g/L. Model prediction at higher substrate concentration (e.g. 100 g/L) deviated from the experimental values, suggesting that ethanol inhibition is a major factor in bioethanol conversion.