Third-order data generation and time measurement-assisted modeling from Diclofenac kinetics and acquisition of excitation-emission fluorescent signals.

GABRIEL G. SIANO; SOFIA MORA; AGUSTINA V. SCHENONE

Santa Fe

Congreso; XIX Chemometrics in Analytical Chemistry; 2024

Diclofenac (DCF) is a non-steroidal anti-inflammatory drug. It has been reported to exhibit fluorescence and is known to be photosensitive [1]. These two characteristics can be used in experiments based on the monitoring of reaction kinetics for quantitative purposes.This work presents experimental and chemometric details about the generation of third-order data, resulting from the monitoring of DCF kinetics through sequential recording of fluorescence intensities at different Excitation and Emission wavelengths. Additionally, during the fluorescence data acquisition, the times at which these readings occurred were also recorded by connecting the spectrofluorimeter to an Arduino board.It is known that there are potential spectral-temporal dependencies during this type of sequential fluorescence recordings, which sometimes prevents the use of models that rely on the multilinearity of data. Therefore, the fluorescence and time data were modeled together using Parallel Factor Analysis (PARAFAC) with data imputation, localization in the time domain, time-derived modeling constraints (such as smoothing of kinetic profiles) and pseudo Excitation-Emission Matrices (pseudoEEMs) [2].Samples were designed with a calibrated analyte (DCF), in the presence of potential uncalibratedinterferences, such as pyridoxine (PIR) and tryptophan (TRP). A Python script was written to directly control the spectrofluorimeter, circumventing limitations imposed by its commercial Graphical User Interface (GUI) and enabling the implementation of technical optimizations. Data acquisition was achieved through emission scans in alternating/opposite directions for consecutive excitations. This minimizes unnecessary restarts/repositioning of the monochromators, which is a significant cause of time loss in an analysis, even with a fast-scanning spectrofluorimeter. As a result, a significant reduction in analysis time per sample was achieved.Regarding the analysis of multicomponent samples, high degrees of concordance were obtained between the information provided by pure substances and the profiles resolved in each spectral mode, even with moderate or severe overlaps. Furthermore, the description of the kinetic mode was smoother and more physically realistic compared to those obtained without using the measured times. These improvements were due to the use of the actual times, rather than implicitly assuming that all data from each EEM were obtained simultaneously, with no changes in fluorophore concentrations. According to our models and also to results from HPLC analysis, DCF does not fluoresce, at least not conventionally, and it is likely that a derivative is the one that actually fluoresces.References[1] Aparici-Espert I, Miranda MA and Lhiaubet-Vallet V, Molecules, 2018, 23 (3), 673.[2] Siano GG, Vera Candioti L and Giovanni LL, Chemometrics and Intelligent Laboratory Systems, 2020, 199, 103961.