1. Introduction
The presence of some elements in gasoline above certain concentration limits is undesirable, not only by the possibility of damaging motor parts and poor performance of the fuel, but also because of the pollution caused by the release of toxic elements to the atmosphere due to the combustion of the fuel [1]. Anthropogenic activities are of main importance in the mobilization of mercury and its compounds and, among them, the burning of fossil fuels is of special concern [2]. According to the literature [3], different Hg species can be present in petroleum and in its fractions. Metallic elements are normally present at very low concentrations in gasoline, requiring very sensitive techniques or pre-concentration steps for their determination [4].
However, few papers on the determination of total mercury in petroleum products have been published. Excellent limits of detection using cold vapor atomic absorption spectrometry (CV AAS) [2] or cold vapor atomic fluorescence spectrometry (CV AFS) [5] have been reported, but they can be too dangerous for common use when a previous BrCl oxidation of the sample is employed [6]. Mercury is a difficult element to be measured in complex matrices because the matrix must be destroyed at elevated temperature without loss of the analyte or contamination. Furthermore, the determination of Hg in gasoline and petroleum products samples is difficult due to the volatile nature of both mercury and the matrix, which may lead to severe loss. The difficulties of getting this kind of sample into aqueous solution are not unique to mercury determination [7]. In this way, a recent work was published on the determination of mercury in petroleum products by isotopic dilution cold vapor inductively coupled plasma mass spectrometry (ID-CV-ICP MS) using wet ashing in a closed vessel system [8]. Sample preparation as microemulsion has been applied also successfully for the determination of mercury and other elements in petroleum products, such as mercury in naphta by ET AAS with direct pre-concentration in the graphite tube, using palladium as modifier and multiple injections [9], determination of As and Sb also in naphta by GF AAS using Ir as permanent modifier [10] and determination of mercury in gasoline by CV AAS with reduction in the microemulsion [2]. Sample preparation as emulsion has also been employed for vegetable oils for the determination of Hg, As and Cd by flow injection chemical vapor generation ICP-MS [11]. The use of gold trap was well exploited for mercury pre-concentration and speciation in a variety of environmental samples [12].
The lack of certified reference gasoline samples makes difficult the validation of the analysis accuracy. Ways to circumvent this problem are, for example, to carry out analysis by different techniques or by applying recovery tests. Other problems in relation to trace analysis of petroleum products are: (i) calibration using the standard addition technique may lead to errors, since the analyte forms in the material to be analyzed are unknown and may behave differently in comparison to the spiked form; (ii) the potential loss of the analyte during sample preparation and (iii) the low levels of the analytes present in this kind of sample, as already mentioned [7] and [13]. The last point is of special concern, since the size of the production, refining and consumption of petroleum products may result in a relatively large emission of contaminants to the environment