An improved graphite furnace electrothermal vaporization device equipped with an autosampler for inductively coupled plasma atomic emission spectrometry is presented. The transport losses of eight selected analytes in the individual segments of the device were determined by means of the radiotracer technique by applying amounts traced comparable to those to be determined in real samples. The results obtained from the radiotracer study were the basis for further improvement of the interface design, leading to considerable increase of the total transport efficiency, which finally was found to be between 26 (for Cr) and 57% (for Ga). The whole system consists of a graphite furnace vaporizer, a power supply, a gas flow box, and an autosampler with incorporated microbalance. The temperature program, gas flows, and autosampler functions are controlled by a data station which also provides the data acquisition and processing of the transient signals. The performance parameters of the developed system were evaluated using aqueous standard solutions. Absolute limits of detection for most analytes were between 0.1 and 1 ng, and for As, K, Ni and Pb, they were between 2 and 3.2 ng.

Solid sampling electrothermal atomic absorption spectrometry using the boat technique and a transversely heated graphite tube was applied to direct analysis of graphite and silicon carbide powders for 14 and 12 impurity elements, respectively. With graphite, for all analytes under investigation, a very effective in-situ analyte/matrix separation was achieved. That was the case also for analytes in silicon carbide requiring atomization temperatures below 2400 degrees C. At higher atomization temperatures, the decomposition products of silicon carbide give rise to significant background, which can still be corrected. Sample amounts of up to 4 mg graphite and 8 mg silicon carbide per analysis cycle were applied. For all analytes in both materials, limits of detection at the lower ng g(-1) and sub-ng g(-1) level were achieved, excluding arsenic for which they were 50 ng g(-1) and 23 ng g(-1) for graphite and silicon carbide, respectively. Quantification was performed using calibration curves measured with aqueous standard solutions. The accuracy was checked by comparison of the results with those obtained by instrumental neutron activation analysis and by other independent methods.

No abstract

No abstract