Application of Response Surface Methodology and Machine Learning Combined with Data Simulation to Metal Determination of Freshwater Sediment

Abstract

A comparative study between conventional methods (EPA 3050B and ISO 11466.3) of metal extraction and a simple low-cost method, using aqua regia, was carried out in this work. Six elements (Mn, Cu, Zn, Pb, Ni, and Cd) were determined by flame atomic absorption spectrometry (FAAS) in a certified sample of sediment (CNS 392). Central composite design (CCD) and response surface methodology (RSM), as well as machine learning, were used to find the optimal conditions for metal extraction. The influence of the parameters—volume of nitric acid in aqua regia (v), time of extraction (t), and temperature (T)—on Mn, Cu, Zn, and Pb recoveries was investigated. The best condition for the recovery of all the metals was v = 2.5 mL of HNO3, t = 2 h, and T = 90 °C. In comparison with the conventional methods, the aqua regia method was found to present better recovery values and lower standard deviations for all the metals studied.