Comparison of Bioethanol Production of Starches from Different Andean Tubers

Bioethanol from starchy biomass represents 60% of the global generation. In Ecuador, there is an annual production of potato, cassava and sweet potato of over 600 thousand tons. However, there are no comparative studies that analyze the effects of starch source. Thus, the present work aims to compare the potential for bioethanol production from starches of potato (Solanum tuberosum), cassava (Manihot esculenta), and sweet potato (Ipomea batatas). Starches were subjected to liquefaction with α-amylase. For bioethanol production, two different processes were assessed 1) sequential (SeqSF) saccharification, with gluco-amylase, and fermentation at 30C for 72h, with previously activated yeast (S. cerevisiae, a wine strain), and 2) simultaneous saccharification-fermentation (SSF), where glucoamylase and yeast were added at the same time, at fermentation conditions. Total reducing sugars, after enzymatic digestion and during fermentation, were quantified through 3,5 dinitrosalicylic acid (DNS), and ethanol was quantified through GC. After liquefaction, potato and cassava starches resulted in approx. 44 g/L of sugars, while SP resulted in 33 g/L. For the SeqSF process, after saccharification, potato and cassava presented similar values of sugar concentration, close to 90 g/L (58% conversion), while sweet potato yielded a lower value of 66 g/L (52% conversion). During fermentation, in the case of cassava and potato, over 90% of the sugars were consumed in the first 24h, and less than 1% remained after 48h. For sweet potato, sugar consumption was lower, with about 18% remaining after 72h. In SSF, sugar consumption was slower. For potato and cassava, around 80% of the sugar had been metabolized by 24h, and for SP, it was 68%. However, after 72h, all fermentations presented less than 4% sugars. Ethanol production in general was below 10% v/v, and it was affected by sugar consumption kinetics. Bioethanol production was greater in the SSF process, when compared to the SeqSF.