Sequential acquisition of fluorescence signals with changing fluorophore concentrations. Multivariate Curve Resolution with time measurements assistance

TL;DR

This paper introduces a time-assisted multivariate curve resolution method that effectively analyzes fluorescence data with changing fluorophore concentrations, maintaining accuracy in dynamic systems like chromatography and reaction kinetics.

Contribution

It proposes a novel approach using time-based localization to organize partial EEM data for improved third-order data modeling in dynamic conditions.

Findings

Time-assisted MCR effectively resolves dynamic fluorescence data.

Predictions from MCR are comparable to higher order models.

Spectral and kinetic profiles are accurate and physically meaningful.

Abstract

Sequential registering of fluorescence signals in conventional Excitation-Emission Matrices (EEMs), followed by modeling based on multilinear properties of the data, requires stable fluorophore concentrations throughout the acquisition of each EEM. Rapid concentration changes, as seen in chromatography or certain kinetics, can disrupt the conventional bilinearity of EEMs. This deviation depends on the relative rates of concentration changes versus spectral scanning speeds. Although entire scans can be slow, individual data points are acquired almost instantaneously, maintaining linear dependencies. Within specific time intervals, concentrations can be assumed constant. By using time-based localization, partial EEM data can be organized into partially filled cubes that allow for the correct modeling of third-order data. Additionally, Multivariate Curve Resolution requires reshaping operations. Two third-order datasets were analyzed using a time-assisted MCR implementation, following a strategy similar to one reported for chromatographic data (LC-EEM) with Parallel Factor Analysis. The second dataset comes from Diclofenac reaction kinetics (Kin-EEM). The results suggest that time-assisted MCR resolves such data effectively. The solutions found were similar to those obtained when processing the same data as cubes. Predictions from calibration models based on MCR results were comparable to those obtained when deriving higher order models from the same data. High degrees of similarity were achieved between resolved and reference spectral profiles. Both chromatographic and kinetic profiles were accurate and physically meaningful. The proposed strategy highlights the potential of incorporating time-based localization to enhance the analysis of fluorescence data in dynamic systems.