Multivariate near infrared spectroscopy for predicting polyhydroxybutyrate biosynthesis by mixed microbial consortia cultured on crude glycerol

Polyhydroxybutyrate (PHB) is a bioplastic biosynthesized by microorganisms as an intracellular granule. In this study, PHB was produced by a mixed microbial culture fed crude glycerol in a sequencing batch bioreactor. Bioreactor operations were monitored by near infrared (NIR) spectroscopy (908-1708 nm) with a direct insertion probe to develop a model for predicting microbial PHB concentration. Multivariate statistical methods, principal component analysis (PCA) and partial least squares (PLS), were applied for modeling of the NIR spectral data. The results showed that NIR spectra can (i) be used to predict PHB content in-situ, with a calibration model and prediction model exhibiting a good fit (R-2 = 0.88 and R-2 = 0.89, respectively) and (ii) also be used to monitor substrate, glycerol, concentration. (C) 2015 Elsevier Ltd. All rights reserved.