Simultaneous Determination of Trace Manganese, Iron, Copper and Zinc Using Spectrophotometry Assisted by Artificial Neural Network Algorithm

Jin HOU,Guo-song CHEN,Zhen-pu WANG

DOI: https://doi.org/10.3969/j.issn.1004-4957.2001.02.003

2001-01-01

Abstract:All of Mn(Ⅱ ),Fe(Ⅲ ),Cu(Ⅱ ) and Zn(Ⅱ ) react sensitively with 2 _(5 _ bromo _ 2 _ pyridylazo) _ 5 _ diethylaminophenol and p _ octylpolyethyleneglycol phenylether,a non _ ionic surfactant,at pH 9.5 to form the stable ternary micellar complexes.Their λ max are at a wavelength of 566,560,562 and 559 nm,respectively.This indicated that their absorption spectra were seriously overlapped.Their apparent molar absorptivities are 1.13× 105,7.32× 104,1.02× 105 and 1.04× 105 L· mol－ 1· cm－ 1,respectively at their corresponding λ max.The simulated annealing－ back propagation of error－ artificial neural network algorithm has been applied to simultaneously determine trace manganese,iron,copper and zinc in simulated samples and grain samples by spectrophotometry without separation.The optimum computational conditions were studied in detail.Although additivity of the system is poor,the satisfactory results have been still achieved because of the excellent feature of the algorithm.