MONITORING BATCH ETHANOL FERMENTATION USING MID-INFRARED SPECTROSCOPY

A methodology for monitoring the main components of batch ethanol fermentations conducted in Erlenmeyer flasks was developed. The methodology was based on the technique of Fourier transform mid-infrared (FT-MIR) spectroscopy associated with the multivariate calibration by partial least squares (PLS). The supernatant spectra of the centrifuged samples (without cells) were obtained at 25 °C. The calibration was done from data of 8 fermentations (calibration set) and the models validation was performed with a condition not used in the calibration (validation set). The models presented excellent prediction results with root-mean-square error prediction (RMSEP) of 0.58 g L-1 (sucrose), 2.76 g L-1 (fructose), 1.44 g L-1 (glucose), 0.50 g L- 1 (cells), 1.83 g L-1 (ethanol), and 0.76 g L-1 (glycerol). The proposed methodology provided monitoring of the main components during batch ethanol fermentations.