On the direct analysis of solid samples by graphite furnace atomic absorption spectrometry

Summary We have studied some limitations of the solid sampling, cup-in-tube technique by comparison with a constant temperature two-step atomiser and found consistently lower vapour-phase temperatures and greater interferences in the former. With the former, higher vapour-phase temperatures and improved analytical results for lead and cadmium were found using Pd(NO₃)₂+Mg(NO₃)₂ instead of NH₄H₂PO₄+Mg(NO₃)₂ as modifier. Investigations of methods to extend the useful calibration range revealed reduced vapour-phase temperatures in the presence of a convective gas flow during atomisation, and a greater potential for errors using non-resonance lines. With respect to the latter, the absorbance signal is much more sensitive to matrix induced changes in the temperature interval in which atoms are formed compared to resonance lines. Furthermore, at the lead 261.4 nm non-resonance line we observed overcompensation errors caused by cobalt when using a Zeeman-effect system, and undercorrection due to the AlCl(g) molecule with a continuum source background corrector.