Premium
Resolution of Heavy Metal Mixtures from Highly Overlapped ASV Voltammograms Employing a Wavelet Neural Network
Author(s) -
Gutiérrez Juan Manuel,
MorenoBarón Laura,
Céspedes Francisco,
Muñoz Roberto,
del Valle Manuel
Publication year - 2009
Publication title -
electroanalysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.574
H-Index - 128
eISSN - 1521-4109
pISSN - 1040-0397
DOI - 10.1002/elan.200804419
Subject(s) - copper , cadmium , analytical chemistry (journal) , resolution (logic) , analyte , deconvolution , metal , materials science , electrode , chemistry , chromatography , computer science , metallurgy , algorithm , artificial intelligence
This work describes the chemometric assisted ASV determination of three heavy metals in water (lead, copper and cadmium) in presence of thallium and indium as interfering species. Stripping was carried out in open atmosphere, employing a graphite‐epoxy transducer as working electrode, without any surface regeneration after each analysis. The concentration range studied was from 0.4 to 20 ppm for both analytes and interferents. Due to the overlapping nature of the signals, a wavelet neural network (WNN) was used for deconvolution of the voltammogram. In order to validate the resolution capability, a k‐fold cross validation procedure was performed. Mixtures of metals could be resolved with good prediction of their concentrations; obtained vs. expected comparison graphs exhibited, for a set of samples not employed for training, correlation values of 0.996 for lead, 0.989 for cadmium and 0.995 for copper.