Diferentes estratégias espectroanalíticas para a determinação de metais em cosméticos

Abstract

In this work, it was developed a procedure for the determination of metals in cosmetics, employing the techniques ICP OES, GF AAS, LIBS and LA-ICP OES; in addition, it was explored the potential of chemometric tools. For the determination of Cd, Co, Cr, Cu and Ni by ICP OES and Pb by GF AAS in eyeshadow and lipstick samples, it was developed a sample preparation method using diluted nitric acid and a hot block. For the optimization of chemical parameters of sample preparation and operating parameters of ICP OES, it was used factorial design and for data analysis and selection of the best working conditions, the desirability functions were employed. In total, it was analyzed 30 samples of cosmetics (21 eyeshadow and 9 lipstick), of different colors, brands and from different countries (Brazil, USA and China). These samples were analyzed in triplicate and with standard addition to perform the recovery calculations. The analyzed samples showed concentration of Cd and Pb within the limits set by the laws, however, 7 eyeshadow samples had concentration of Cr and Ni above than 5 g g-1. In the case of lipstick samples, all values were below this safe limit. All samples analyzed by ICP OES and GF AAS were also analyzed without sample pretreatment by LIBS. The values obtained in the quantitative analysis were employed in the development of supervised classification models such as the KNN and SIMCA. For the obtainment of these models, data of children´s cosmetics samples were included. The classes considered in the models were: below and above the legislation, with a limit set at 5 mg kg-1, according to the recommendations in the literature. In addition to analysis of eyeshadow and lipstick samples, it was explored the potential of the techniques LIBS and LA-ICP OES along with chemometric tools for exploratory analysis of sunscreen samples. The results were obtained for three different equipment (two LIBS and one LA-ICP OES), and several models employing principal component analysis (PCA) were generated. The signals profile of Al, Ni, Ti, and Zn were analyzed and the results obtained between the three equipment were similar.