Statistical design-principal component analysis optimization of a multiple response procedure using cloud point extraction and simultaneous determination of metals by ICPOES

A procedure has been developed for the simultaneous determination of traces amounts of Cd, Cr, Cu, Mn, Ni and Pb from saline oil-refinery effluents and digested vegetable samples using inductively coupled plasma optical emission spectrometry (ICP OES). The procedure is based on cloud point extraction (CPE) of these metals as 2-(bromo-2-pyridylazo)-5-diethyl-amino-phenol (Br-PADAP) complexes into a micellar phase of octylphenoxypolyethoxyethanol (Triton X-114). Optimization of the procedure was performed by response surface methodology (RSM) using a Doehlert design. Principal components (PC) were used to simplify the multiple response analysis. A response surface for the first PC score is useful in determining the optimum conditions for the Cd, Cr, Cu, Mn and Pb determinations whereas the second PC is highly correlated with the Ni response. Improvement factors of 22, 36, 46, 25, 65 and 39, along with limits of detection (3 sigma(B)) of 0.081, 0.79, 0.38, 0.83, 0.28 and 0.69 mu L-1, and precision expressed as relative standard deviation (%R.S.D., n = 8, 20.0 mu g L-1) of 1.5, 2.2, 3.5, 2.6, 2.5 and 2.5 were achieved for Cd, Cr, Cu, Mn, Ni and Pb, respectively. The accuracy was evaluated by spike tests in oil-refinery effluent samples and analysis of a vegetable certified reference material (NIST 1571, orchard leaves). Results found were in agreement with certified values. (c) 2006 Elsevier B.V. All rights reserved.