Application of a four-variables Doehlert design for the multivariate optimization of copper determination in petroleum-derived insulating oils by GFAAS employing the dilute-and-shot approach

A method for simple and rapid determination of Cu in petroleum-derived insulating oils used in power transformers by GFAAS is proposed in this work. The method was optimized regarding its instrumental conditions through a multivariate strategy based on the screening of the variables, to identify the significant ones, followed the application of a response surface methodology. A two-level full factorial design was employed for the screening of the variables and a four-variables Doehlert design was applied for full optimization of the procedure. Minimum treatment was imposed to the samples, which were only diluted with hexane (1: 1) before introduction into the spectrometer in order to decrease the viscosity. The parameters related to the temperature program that were optimized in this work were the temperatures of pyrolysis and atomization, the pyrolysis time, the final temperature of the drying step and the heating rate in the drying step. In the optimized conditions, the method could be applied in the determination of Cu in the samples with a limit of detection of 1.6 mu g L (1) and a limit of quantification of 5.3 mu g L (1). The quantification was always performed by standard addition method because of the interferences of matrix that were present, even at the optimized conditions. (C) 2012 Elsevier Ltd. All rights reserved.