Optimization of on-line microwave flow injection analysis system by artificial neural networks for the determination of ruthenium

A methodology based on the coupling of experimental design and artificial neural networks (ANNs) was proposed in the optimization of a new on-line microwave flow injection system (FIA) for the determination of ruthenium, grounded on its catalytic effect on the oxidation of dibromocarboxyarsenazo (DBM-AsA) by potassium periodide under the microwave irradiation. The response function (RF) used was a weighted linear combination of two variables related to sensitivity and sampling rate. A neural network with extended delta-bardelta (EDBD) learning algorithm was applied to predict the maximal RF, according to which the optimized conditions were obtained. The optimized new on-line microwave FIA system is able to determine ruthenium in 5¯200 ng ml-1 range with a detection limit of 2.1 ng ml-1 and a recovery of 94.6%. A sampling rate of 58 h-1 was obtained. In contrast to traditional methods, the use of this methodology has advantages in terms of a reduction in analysis time and an improvement in the ability of optimization.