Spectrophotometric assay for copper and iron in transformer oil using partial least squares regression (PLS2)

Abstract

A simple and low cost spectrophotometric assay is proposed in this work for simultaneous determination of copper and iron levels in insulating oils. The procedure requires minimum sample volume (250 muL), it is carried out in disposable micro tubes (1.5 mL) and aqueous standards are used for calibration. In brief, after dilution with n-propanol, the wear metals are extracted with hydrochloric acid and, after neutralization, the analytes are reduced to the oxidation states Cu(I) and Fe(II) respectively. Finally, the mixture of bathocuproine-disulfonic acid and bathophenantroline-disulfonic acid is added for selective formation of copper and iron complexes. Then, chloroform and methanol are added, the sample is centrifuged and absorption spectra are registered in the upper-aqueous layer (470-550 nm). Multivariate calibration, covering concentration range 0-4 mug/mL of each analyte in oil is performed using partial least squares regression. The test solutions are prepared by spiking the aqueous standards of copper and iron to metal-free oil diluted with n-propanol (1:1). The prediction results were in agreement with those obtained using univariate spectrophotometric assay (one ligand) and those obtained by electrothermal atomic absorption spectrometry (after dissolution of wear metals and dilution with n-propanol)