The first page of this article is displayed as the abstract.

The first page of this article is displayed as the abstract.

The direct determination of selenium in steel by flame atomic absorption spectrometry is described. Spectral interference from iron is minimised by instrumental settings and the remaining interferences are eliminated by using a selenium-free steel matrix solution. The sample decomposition procedure, which proved to be critical for obtaining correct results, is described. Selenium contents as low as 0.01% can be determined. The relative standard deviations are 5.12 and 1.03% for 0.014 and 0.096% selenium, respectively.

The concentrations of manganese, chromium, copper and cobalt were accurately determined in various standard steel samples, as were manganese, copper, zinc and iron in tomato leaves and bovine liver (National Bureau of Standards reference materials). Scatter signals were evaluated by means of the two-line method. Various potential spectral interferences were studied including the effect of 5000 µg ml–1 iron solutions on manganese and chromium determinations. The plasma behaved as would be expected of a narrow line source. The two-electrode version of the direct current plasma that was used, limited the precision of the measurements to 10% relative standard deviation.

Based on the static high sensitivity inductively coupled plasma (SHIP) introduced by Buscher et al. (W. Buscher, A. Klostermeier, C. Engelhard, S. Evers, M. Sperling, J. Anal. At. Spectrom., 2005, 20, 308-314), a new torch design for ICP mass spectrometry was developed. The SHIP-torch, including the external air cooling system, was modified in order to allow its application as an ion source in a conventional ICP mass spectrometer. While the torch geometry was adjusted for its use in connection with the sampling interface, the principal shape of the torch remained the same as in the recently developed SHIP-OES-system. The plasma discharge was operated at 0.65 kW rf power with a total plasma gas flow rate of 1.95 L min−1 and the mass spectrum was investigated. A standard pneumatic nebuliser was used as sample introduction system and the ion signals of a multi elemental standard solution were measured for different sample carrier gas flow rates from 0.2 to 1.4 L min−1. Limits of detection were obtained for a number of elements. These results were compared to those achieved with the conventional ICP-MS setting.

Matrix effects in inductively coupled plasma atomic emission spectrometry (ICP-AES) due to metallic elements in solution are found to be largely due to plasma processes. The effects of such matrix elements on the sensitivities of analyte lines are evaluated. Most matrix elements give rise to considerably greater effects than do the frequently studied alkali metals, under operating conditions normally adopted for routine use. Analytes are affected to a degree dependent on the total excitation potential. Lines with high excitation potential are strongly affected. Different matrix elements produce widely varying effects on the sensitivity of a single line, depending on the energy required to dissociate and ionise the matrix. These observations are consistent with the hypothesis that the matrix effects are due to cooling of the plasma by the matrix. Such a model enables matrix effects to be quantitatively predicted for matrices at 0.05 M.

A flow injection (FI) introduction system for flame atomic absorption spectrometry (FAAS) has been constructed for single-standard calibration and sample dilution. Dilution factors ranging from 5.93 to 38.8 in six discrete stages were produced by replicate injection of 12.5-µl volumes into an aqueous carrier stream flowing down lines of different lengths. The dilution factors were measured for five different solutions for each line. No dependence on concentration was found and the relative standard deviations ranged from 0.75 to 3.1%. The manifold was evaluated by the analysis of solutions of magnesium (6, 12.5 and 35 p.p.m.), nickel (180 p.p.m.), calcium (75 p.p.m.) and chromium (180 p.p.m.). Recoveries ranging from 95.3 to 106.8% were obtained with no over-all evidence of bias. The uncertainty in the over-all method, including a contribution from the curve-fitting procedure, was estimated to be 5%.

The application of inductively coupled plasma mass spectrometry (ICP-MS) for trace and ultra-trace metal determination in foodstuffs has been investigated. The ICP-MS technique was compared with electrothermal atomisation atomic absorption spectrometry (ETA-AAS) for Cd and Pb in three biological standard materials. Results for Mo and V by ICP-MS were compared with the certified values. The ETA-AAS gave results that were closest to the certified values, with good precision. The ICP-MS results were encouraging and compared well for Cd and Pb, with both ETA-AAS and the certified results. The precision was acceptable at the higher element concentrations but degraded considerably at the lower levels. An investigation into the causes of the poor precision revealed molecular interferences and instrument hardware were partially to blame. Noticeable improvements in accuracy and dramatic improvements in precision were then obtained. Molecular ion interferences over the 51–81 a.m.u. mass range were investigated using several chemical matrices (HNO3, HCl, H2SO4, H3PO4, H2O2 and NH3) and the results confirmed that dilute HNO3 is the most suitable acid matrix. Data are presented on the problems of the determination of As and V in the presence of 51CIO+ and 75ArCl+ and the possibilities for data correction explored.

The optimum conditions necessary for the determination of trace elements in organic based materials by ICP-MS are discussed. Problems experienced with condensation of particulate carbon on the sampler orifice have been eliminated by the use of a low flow of oxygen introduced into the nebuliser gas and controlled by a mass flow controller. The background spectra obtained are composed of a few simple molecular mass ions at low mass (<80 a.m.u.). These degrade detection limits for a few elements, however the problem of blank contamination in the commonly used solvents is seen as the limiting factor in detection power. Detection limits are in general below ng ml–1 levels however for most elements, comparable to those reported for a similar system running aqueous solutions.

Arrhenius plots were prepared from signals obtained for the atomisation of Pb, Bi and Au from a graphite surface at atmospheric pressure and in vacuum at 10–9 atm. Spatially integrated absorbance was recorded for atmospheric pressure studies and mass spectral intensity for vacuum runs. Comparison of activation energies (Ea) calculated from the leading edges of the signals showed no significant differences between atmospheric and vacuum vaporisation. The Ea values are thus not influenced by potential radial inhomogeneities in the vapour densities within the tube. For the elements and conditions used in this study, Arrhenius data support a vaporisation model that assumes the analyte to be present on the graphite surface in the form of microparticles as opposed to a monolayer.