Functionality of Rotational Ambiguity in Self-Modeling Methods to Signal Contribution of Chemical Component

Self-modeling curve resolution (SMCR) methods are a powerful tools in chemometrics, used for the decomposition of the measured data matrix to the product of two chemically meaningful matrices, containing the pure profiles. But, the results of these methods may be ambiguous. There are two types of ambiguities in SMCR methods, intensity and rotational ambiguities. Rotational ambiguity is the more difficult source than intensity ambiguity in SMCR methods. In The presence of rotational ambiguity, SMCR methods resolve data sets to a range of feasible solutions. Implementing the adequate information and constraints, eg, non-negativity for constituent concentrations and for spectral signals, unimodality for evolving signals such as those from chromatographic and zero regions where the contributions of certain components are known to be absent and etc, can reduce or remove the rotational ambiguity. In some special cases, the minimal information of the system, constraint of non-negativity is sufficient for obtaining unique solutions [1]. Tauler and et al, showed that the degree of rotational ambiguity depends on the selectivity and the similarities between the pure concentration and spectral profiles of components [2]. In this work we are showing that rotational ambiguity does not depend only on the selectivity. Increase or decrease of contribution of each component that present in the data matrix effects the rotational ambiguity of all species. Several simulated second‐order chromatographic two‐component systems are studied regarding the comparison of rotational ambiguity with increasing of contribution of each component, using computing volume of the area feasible solutions. Also the effect of contribution of components on accuracy of quantitative analysis was analyzed with experimental second-order excitation-emission fluorescence data sets. These data sets are several two component systems include Tyrosine as analyte and Tryptophan as interferent which is present in the unknown samples. Analysis of experimental data sets with grid search method and calculation the concentration range and evaluating the error prediction range showed that increasing or decreasing of contribution of components are effective on accuracy of quantitative analysis using SMCR methods.