Optimization of solid-phase extraction using artificial neural networks and response surface methodology in combination with experimental design for determination of gold by atomic absorption spectrometry in industrial wastewater samples

Solid-phase extraction (SPE) is often used for preconcentration and determination of metal ions from industrial and natural samples. A traditional single variable approach (SVA) is still often carried out for optimization in analytical chemistry. Since there is always a risk of not finding the real optimum by single variation method, more advanced optimization approaches such as multivariable approach (MVA) should be applied. Applying MVA optimization can save both time and chemical materials, and consequently decrease analytical costs. Nowadays, using artificial neural network (ANN) and response surface methodology (RSM) in combination with experimental design (MVA) are rapidly developing. After prediction of model equation in RSM and training of artificial neurons in ANNs, the products were used for estimation of the response of the 27 experimental runs. In the present work, the optimization of SPE using single variation method and optimization by ANN and RSM in combination with central composite design (CCD) are compared and the latter approach is practically illustrated.