Energy & Fuels, Vol.33, No.5, 3916-3924, 2019
Extraction Induced by Microemulsion Breaking: A Novel Strategy for Mg, Mn, and Zn Extraction from Ethyl Alcohol-Containing Gasoline
The determination of metals in petroleum derivates is an important issue related to quality control of these products. Therefore, we conducted a research work to propose a simple strategy for the extraction of Mg, Mn, and Zn from Brazilian automotive gasoline, which contains ethyl alcohol, and followed their quantification by flame atomic absorption spectrometry (FAAS). The extraction method consisted of two steps. In the first step, we formed a microemulsion with the gasoline sample and the acid extractant solution using 1-propyl alcohol as the dispersant agent. Afterward, we induced microemulsion breaking with addition of water, yielding two immiscible phases: (i) a gasoline phase on top, and (ii) a water/alcohol acid solution at the bottom. Mg, Mn, and Zn extracted from the sample were in the bottom phase. The water/alcohol phase was collected for the determination of the analytes by FAAS. The experimental parameters of the method were optimized, as well as the calibration conditions. Under optimum conditions, we dispersed 100 mu L of the extractant solution (3.5 mol L-1 HNO3) through the sample (9.20 mL) using 700 mu L of 1-propyl alcohol. After that, we induced microemulsion breakdown with 300-600 mu L of deionized water, depending on the analyte. The extracts (bottom phase) were collected and analyzed using a matrix-matching strategy, and the adjustment of the nebulizer aspiration flow rate was shown to be very important to achieve satisfactory accuracy. The method presented the following limits of detection: 0.02, 0.03, and 0.03 mg L-1 for Mg, Mn, and Zn, respectively, and the limits of quantification of 0.05, 0.1, and 0.09 mg L-1, respectively. Five samples of Brazilian gasoline were tested using the developed method, yielding concentrations of Mg, Mn, and Zn between 0 and 0.05, 0.11 and 0.17, and 0.22 and 0.46 mg respectively. The analysis of spiked samples was carried out to evaluate the method accuracy, and recovery percentages between 90 and 120% were verified.