Monitoring and control of fermentation processes is important to ensure high product yield,
product quality and product consistency. More knowledge on on-line analytical techniques
such as near infrared and fluorescence spectroscopy is desired in the fermentation industry to
increase the efficiency of on-line monitoring systems.
The primary aim of this thesis is to elucidate and explore the dynamics in fermentation
processes by spectroscopy. Though a number of successful on-line lab-scale monitoring
systems have been reported, it seems that several challenges are still met, which limits the
number of full-scale systems implemented in industrial fermentation processes. This thesis
seeks to achieve a better understanding of the techniques near infrared and fluorescence
spectroscopy and thereby to solve some of the challenges that are encountered.
The thesis shows the advantages of applying real-time monitoring of bioprocesses and it also
highlights that the applied techniques with different measurement orders deliver specific but
also complementary sources of information. Furthermore, it was shown that valuable process
information can be obtained both by near infrared spectroscopy and fluorescence
spectroscopy, which provide indirect and direct measurements, respectively.
Based on the measurements obtained by near infrared spectroscopy it was found that variation
in scatter and in the absorption can be obtained from the same near infrared spectrum. By
kinetic modelling, it was possible to capture both physical and chemical changes appearing in a
lactic fermentation process. The physical changes were associated with the textural
transformation appearing during the gel formation and chemical changes were associated with
the biological conversion reactions, which take place during the fermentation process.
The results presented in this thesis also highlight that pH changes have a major effect on the
fluorescence intensities, which can influence the quantifications of the relevant components
negatively. When the pH was either increased or decreased, manually, during the measured
process, a clear increase or decrease was observed in the fluorescence landscapes. This thesis
presents a correction strategy based on a chemometric modelling approach, where weighted
non-linear regression and weighted PARAFAC analysis are combined.
Based on the research conducted in this PhD project, it is concluded that near infrared
spectroscopy can provide valuable physical and chemical real-time information during yoghurt
fermentation. Also, it is concluded that fluorescence data must be evaluated carefully if pH
changes appear in the measured system. Furthermore, this thesis concludes that such data still
can be applied for on-line monitoring if corrections or preventive measures during the
quantification are carried out. The findings presented in this thesis have enabled the possibility
of obtaining a better process understanding and to ease monitoring and controlling of
fermentation processes.