Abstract

In excitation-emission fluorescence spectroscopy, the simultaneous quantitative prediction and qualitative resolution of mixtures of fluorophores using chemometrics is a major challenge because of the scattering and reabsorption effects (turbidity) presented mainly in biomaterials. The measured fluorescence spectra are distorted by multiple scattering and reabsorption events in the surrounding medium, thereby diminishing the performance of the commonly used three-way resolution methods such as parallel factor (PARAFAC) analysis or multivariate curve resolution-alternating least squares (MCR-ALS). In this work we show that spectral loadings and concentration profiles from model mixtures provided using PARAFAC and MCR-ALS are severely distorted by reabsorption and scattering phenomena, although both models fit rather well the experimental data in terms of percentage of the explained variance. The method to correct the fluorescence excitation-emission matrix (EEM) consisted in measuring the optical properties (absorption parameter μ<i>_a</i> , scattering parameter μ<i>_s</i>, and anisotropy factor <i>g</i>) of samples and calculating the corresponding transfer function by means of the Monte Carlo simulation method. By applying this transfer function to the measured EEM, it was possible to compensate for reabsorption and scattering effects and to restore the ideal EEM, i.e., the EEM that is due only to fluorophores, without distortions from the absorbers and scatterers that are present. The PARAFAC and MCR-ALS decomposition of the resulting ideal EEMs provided spectral loadings and concentration profiles that matched the true profiles.