Abstract

In this study, hybrid of artificial neural network-genetic algorithm (ANN-GA) was used for the development of a microwave-assisted extraction method for determination of target element (zinc, copper, iron and manganese) in tea samples using flame atomic absorption spectrometry (FAAS). A multiple response function (R_m) was applied to describe the experimental conditions for simultaneous extraction of the target element. The power, temperature, extraction time and volume of solvents were the input variables, while the R_m was the output. Optimum conditions were 360 W, 103 °C, 27 min and 2.7:7.3 mL for power, temperature, time and volume of nitric acid:hydrogen peroxide (as solvents), respectively. High determination coefficient between the actual and the predicted data by ANN model (R²= 0.983) indicated the goodness of fit. The developed procedure was then applied to the extraction and determination of these elements in the some tea samples.

Keywords: Artificial neural network, copper, flame atomic absorption spectrometry, genetic algorithm, iron, manganese, microwave- assisted extraction, sensivity analysis, tea, zinc.