Multivariate curve resolution-alternating least-squares and second-order advantage in first-order calibration. A systematic characterisation for three-component analytical systems

Fabricio A Chiappini,Licarion Pinto,Mirta R Alcaraz,Nematollah Omidikia,Hector C Goicoechea,Alejandro C Olivieri
DOI: https://doi.org/10.1016/j.aca.2024.343159
2024-11-01
Abstract:Background: Recent interest has been focused on the application of multivariate curve resolution-alternating least-squares (MCR-ALS) to systems involving the measurement of first-order and non-bilinear second-order data. The latter pose important challenges to bilinear decomposition models, due to the phenomenon of rotational ambiguity in the solutions, even under the application of the full set of chemical constraints that is usually employed in MCR-ALS calibration. Results: After the analysis of several simulated and experimental datasets, important conclusions regarding the role of the selectivity patterns in the constituent spectra have been drawn concerning the achievement of the second-order advantage. Theoretical considerations based on the calculation of the areas of feasible solutions helped to support the observations regarding the predictive ability of MCR- ALS in the various datasets. Significance: The understanding of the impact of rotational ambiguity in obtaining the second-order advantage with both first-order and non-bilinear second-order data is of paramount importance in the future development of analytical protocols of complex samples.
What problem does this paper attempt to address?