Genetic-algorithm-based wavelength selection in multicomponent spectrophotometric determination by PLS: application...

Area of Science:

• Analytical Chemistry
• Environmental Chemistry

Background:

• Spectrophotometric analysis of mixtures with similar spectra presents challenges.
• Accurate quantification of copper and zinc in water is crucial for environmental monitoring.

Purpose of the Study:

• To develop and validate a method for simultaneous determination of copper and zinc.
• To assess the efficacy of a genetic algorithm (GA) for wavelength selection in partial least squares (PLS) calibration.

Main Methods:

• Utilized a spectrophotometric method based on the Zincon reaction at pH 9.
• Employed a genetic algorithm (GA) for optimal wavelength selection in partial least squares (PLS) models.
• Validated the GA-PLS model using synthetic solutions of copper and zinc.

Main Results:

• The GA-PLS model significantly improved prediction accuracy compared to PLS without GA.
• Root-mean-square deviation (RMSD) for copper and zinc were 0.0407 and 0.0865, respectively, with GA.
• The method demonstrated good reliability for simultaneous determination of copper and zinc in natural, tap, and waste waters.

Conclusions:

• GA-PLS is a robust approach for wavelength selection in spectrophotometric analysis of complex mixtures.
• The developed method provides accurate and reliable simultaneous quantification of copper and zinc in diverse water matrices.