Weighted PARAFAC and Nonlinear Regression for Handling Intensity Changes in Fluorescence Spectroscopy Caused by pH Fluctuations

Abstract

Fluorescence spectroscopy is a sensitive and selective technique, which can be of great value in bioprocesses to provide online, real-time measures of chemical compounds. Although fluorescence spectroscopy is a widely studied method, not much attention has been given to issues concerning intensity variations in the fluorescence landscapes due to pH fluctuations. This study elucidates how pH fluctuations cause intensity changes in fluorescence measurements and thereby decreases the quality of the subsequent quantification. A photo-degradation process of riboflavin was investigated using fluorescence spectroscopy and used as a model system. A two-step modeling approach, combining weighted PARAllel FACtor analysis (PARAFAC) with weighted nonlinear regression of the known reaction kinetics, is suggested as a way of handling the fluorescence intensity shifts caused by the pH changes. The suggested strategy makes it possible to compensate for uncertainties in the shifted data and thereby obtain more reliable concentration profiles for the chemical compounds and kinetic parameters of the reaction.