Novel application and comparison of multivariate calibration for the simultaneous determination of Cu, Zn and Mn at trace levels using flow injection diode array spectrophotometry

Three different calibration approaches: Partial Least Squares, Unfold Partial Least Squares and n-way Partial Least Squares are compared in terms of explained variance, root mean square error of calibration and root mean square error of cross-validation. Attention is also focused on the application of genetic algorithms to spectral data as a way to obtain an improvement in calibration accuracy. Influence of initial starting conditions for the genetic algorithms (population size, mutation probability, % initial terms etc.) was investigated by means of factorial experimental designs. A simple flow injection manifold coupled to a diode array spectrophotometer and multivariate calibration were employed in order to determine Ni, Cd, Pb, Cu, Zn and Mn. Calibration was not successful for the three first elements. Rapid determination of the microamounts of Cu, Zn and Mn was performed in the presence of concomitant ions with a sample throughput of 180 samples h(-1). Since three metals were simultaneously determined, this corresponds to 540 determinations h(-1). The effect of different matrix interferences (Fe, Al, Ca, Mg, Cr, Sr, Co) was studied in natural and synthetic water samples. Chemical masking with NaF was assayed for the interferent ions Fe and Al. The typical recoveries ranged from 91 to 94% with typical relative standard deviation between 5 and 10 %. (C) 1999 Elsevier Science B.V. All rights reserved.