The MCR-ALS procedure, including selection of components and
used restrictions, was similar to that carried out for the LC-DAD matrices. Additionally, in the first evaluated region, which includes TBZ,
FBZ, and CBZ, it was essential to apply the correspondence constraint
for the correct differentiation between TBZ and FBZ spectra [19]. The latter constraint forces analyte sub-profiles in the augmented elution time
mode to be zero for samples where it is absent.
The residual fits for the LC-FLD analysis were about 0.01 UF (arbitrary units of fluorescence), which is ca. 1% with respect to the maximum measured intensity. After convergence of the ALS optimization,
quantification of each analyte was performed through the corresponding pseudo-univariate calibration curves. Fig. 4B shows the satisfactory
chromatographic and spectral profiles retrieved by MCR-ALS for a typical validation sample, and Fig. 3B displays the predictions for the fluorescent agrochemicals in validation samples. The good results are also
corroborated by the statistical parameters of Table 2. In this table, the
positive influence of the FLD in the sensitivity of the method is evident,
as reflected in LODs between 0.2 and 3 ng mL−1.