Abstract: Calibration curves with two quasi-linear sections ("double sloping'') were observed for the medium volatile elements, Cr and Cu, with the use of a SIMAA 6000 graphite furnace atomic absorption spectrometer under interrupted internal gas flow conditions. If a standard transversally heated graphite atomizer (THGA) tube was shortened by 0.5 mm at both of its ends, (i.e. the gaps were enlarged between graphite furnace housing and tube ends), a stronger declination of the calibration curves resulted. Elements with fairly high diffusion coefficients (>5.8 cm(2) s(-1)) and with short appearance time of their transients (<0.6 s), such as Cr and Cu, have shown the most characteristic sensitivity drop towards higher concentrations. This anomalous feature could be eliminated in three different ways; (1) by applying end-capped THGA tubes, (2) using mini-flow (50 cm(3) min(-1)) conditions during the atomization stage, and (3) by adding Pd-Mg chemical modifier. For the low volatile Mo and V, the calibration curves had no irregular shape. For Ag, Co, Cr, Cu, Mn and Ni, the mini-flow settings improved the linearity of the calibration curves and extended the upper limit of the linear calibration range by a factor of 1.5-2.0. The irregular characteristic of the analytical curves was interpreted as an increased vapour loss at higher analyte concentrations through the opened ends of the standard THGA tubes. This vapour loss was associated with the significantly diverse expulsion velocities of atoms, caused by the difference in temperature and concentration gradients, when evaporating amounts of analytes with more than one order of magnitude difference.

Abstract: Size-fractionated aerosol samples, collected over the Atlantic Ocean, were analyzed for up to 20 elements by PIXE. Using a sailboat as sampling platform, duplicate samples were taken for two-day periods by means of battery operated 6-stage cascade impactors, positioned about 8 m above the sea surface. In the PIXE analysis of the fine particle stages (stages 3 to 5) a 5 times smaller beam size was used than for stages 1 and 2. This led to significant improvement in the detection limits for the former stages. The results from the duplicate impactor samples were normally in good agreement, indicating that the combined uncertainty of sampling and PIXE analysis was of the order of 20%. The precision of the PIXE analysis alone was investigated by rebombarding some samples six months after the first analysis. The trends with time of the fine particle sulfur and the coarse particle iron concentrations are discussed in some detail.

Abstract: All published procedures for multi-element preconcentration of trace elements, prior to x-ray fluorescence analysis of water, are reviewed and critically evaluated. Most preconcentration methods applied to the determination of single elements in water are also listed.