In‐situ monitoring of Saccharomyces cerevisiaeITV01 bioethanol process using near‐infrared spectroscopy NIRS and chemometrics

Abstract

The application feasibility of in-situ or in-line monitoring of S. cerevisiae ITV01 alcoholic fermentation process, employing Near-Infrared Spectroscopy (NIRS) and Chemometrics, was investigated. During the process in a bioreactor, in the complex analytical matrix, biomass, glucose, ethanol and glycerol determinations were performed by a transflection fiber optic probe immersed in the culture broth and connected to a Near-Infrared (NIR) process analyzer. The NIR spectra recorded between 800 and 2,200 nm were pretreated using Savitzky-Golay smoothing and second derivative in order to perform a partial least squares regression (PLSR) and generate the calibration models. These calibration models were tested by external validation and then used to predict concentrations in batch alcoholic fermentations. The standard errors of calibration (SEC) for biomass, ethanol, glucose and glycerol were 0.212, 0.287, 0.532, and 0.296 g/L and standard errors of prediction (SEP) were 0.323, 0.369, 0.794, and 0.507 g/L, respectively. Calibration and validation criteria were defined and evaluated in order to generate robust and reliable models for an alcoholic fermentation process matrix. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 32:510-517, 2016.

Real‐time monitoring of ethanol production during Pichia stipitis NRRL Y‐7124 alcoholic fermentation using transflection near infrared spectroscopy

“…2A). These parameters are similar to those in related studies [18,[23][24][25]. These parameters are similar to those in related studies [18,[23][24][25].…”

“…For glucose, the results of the model parameters (R 2 = 0.998 and RMSEP = 4.81 g/L) can be compared to other published work [26]. In the same row, comparing glucose model with these in another study [24], there is an improvement in the R 2 value due to the application of MSC, instead of adaptive calibration strategy (in this work, 0.995; in [23], 0.987). In our experience, application of the first derivative does not show advantageous results with glucose, since deconvolution of overlapping peaks is not efficient enough to reveal the spectral information of glucose, taking into consideration that it is not the only analyte in the matrix, and furthermore, it exhibits structural and optical similarity with the other analytes of interest.…”

“…These results can be compared with those in a previous study [24], comparing the validation models that were constructed using a single window for each component, while in this study, there are at least two windows for each component and in the case of biomass five, resulting in models with an R 2 closer to unity and with relatively the same number of factors. It is possible to think that by increasing the number of wavelengths introduced into the model, a forced fit occurs and model overfitting would occur.…”

“…According to model parameters in Table 2, this six-factor biomass model presents 4.62% error when predicting the concentration of biomass in the medium, with an RMSEP of 0.231 and a determination coefficient (R 2 ) of 0.982. There is an improvement in the prediction model by using the combination of the spectral regions for the biomass model described in the materials and methods section of this study, obtaining a %e smaller than those mentioned elsewhere [24,26], where S. cerevisiae and a spectral region for biomass of 1100-1500, 1760-2010, 2330-2500, and 908-936, are used. Although the majority of these studies use S. cerevisiae, the NIR calibration region for yeasts is almost the same in general, with slight variations, allowing the possibility of comparing spectral information www.els-journal.com between different yeast.…”

Two‐dimensional fluorescence as soft sensor in the monitoring of biotransformation performed by yeast

“…However, both methods are not as sensitive compared to fluorescence spectroscopy. Moreover, in the case of NIR and Raman spectroscopy, useful information is spread over almost the entire frequency range, and there can also be problem with spectra repeatability …”