Determination of copper in glasses by flame atomic absorption and plasma emission spectrometry

Ground samples (150μm, 0.5 g) are decomposed in a mixture of HF and HNO3 acids followed by evaporation to dryness, dissolution of the residue in dilute HNO3 acid, and dilution to 50 ml. Spectral interferences of 14 tested elements are negligible for AAS but in some cases (Nb, Ti, Zr) minor spectral corrections must be made for plasma techniques depending on the level of interferent and analyte. Nonspectral (matrix) interferences at the 1000μg/ml interferent level are all less than 3.5% relative for AAS and less than 5% relative for ICP (except for Ca, Li, and Na where the effects are depressions of 7 to 10% relative). Matrix effects for DCP at the same interferent level are generally more pronounced and amount to enhancements of 5 to 15% relative for 11 of the 14 elements tested. Plasma matrix effects, especially alkali effects, can be compensated by the use of an alkali-aluminum buffer medium (1000μg Na/ml plus 500μg Al/ml as used in this study), a procedure which is most useful for the DCP approach. Copper can be determined down to 10 ppm in the original sample. A comparison of results shows AAS and ICP to be in good agreement with the DCP being slightly less accurate, although the differences in results may be of little practical significance. In terms of freedom from spectral and non-spectral interference, AAS was judged to have a slight advantage for copper.