PRECONCENTRATION OF ANALYTES BY AEROSOL DEPOSITION IN GRAPHITE-FURNACE ATOMIC-ABSORPTION SPECTROMETRY AT THE PG ML(-1) LEVEL

Different experimental configurations were considered, using an atomic absorption spectrometer, in order to obtain a suitable preconcentrati on of analyte inside the graphite furnace via aerosol deposition. The aim was to optimize the procedure in order to measure concentrations d own to 1-5 pg ml(-1) in matrices with a low chemical content, such as snow or ice. At these concentration levels, the risk of contaminating the sample is one of the main problems to be considered when designing the aerosol deposition set-up. In addition, since the deposition time (t(dep)) may be up to several tens of minutes, all the parameters whi ch are decisive in controlling both deposition and atomization steps m ust be time independent and show good repeatability among measurements . With the procedure proposed in this paper it was found that the abso rbance was linearly dependent on the product C-s . t(dep), where C-s i s the analyte concentration in the sample solution, down to the pg ml( -1) level. The regression parameters of the calibration curve were fou nd to be reproducible for a period of weeks, indicating that the instr umental variables presented a high repeatability and, in particular, t hat the yield of atoms per unit mass in the sample was not influenced by the life-time of the graphite furnace (usually about 1000 firings). At a concentration of 3 pg ml(-1) of cadmium, the coefficient of vari ation of the absorbance (5 replicates) was found to be as low as 3 - 5 %, when a 15 min deposition time was applied. At this concentration, t he overall sample standard deviation, which depends on all the steps f rom sampling to analysis, remains the most significant figure of merit . (C) 1997 Elsevier Science B.V.