Abstract: Recent development, present state and expected future developments in energy-dispersive X-ray spectrometry are discussed. Attention is paid to the improvements in analytical selectivity, sensitivity, detection limit, quantitative character and applicability range, which are the result of new or better excitation sources, detectors, instrument design, automation, computer software and theoretical developments.

Abstract: Cellulose powder with 2,2′-diaminodiethylamine (DEN) functional groups exhibits efficient complexation of transition metal cations. Collection yields above 85 % are obtained up to a chelation capacity of 1.5 meq per gram. Since a good collection is obtained for a pH up from 5, no pH adjustments have to be made for natural water samples. The cellulose-DEN powder is insensitive to abundant substances like alkali and alkaline earth ions, and humic matter. Some cations can be eluted efficiently in a small volume of HNO3. Blank concentration levels from the cellulose-DEN powder are reported.

Abstract: Although the analysis of thermal water by spark-source mass spectrometry (SSMS) is rather timeconsuming, it allows the detection of about 20 elements of geochemical interest down to the ppb-level. A physical preconcentration is proposed in order to collect elements having quite different chemical properties, e.g. alkalis, transition elements, and elements occurring in anionic form. The relative sensitivity factors appear to be rather independent of the salt content of the graphite electrodes. Contrary to neutron activation analysis, SSMS has a quite uniform elemental sensitivity, and allows to determine elements for which neutron activation is not suitable, e.g. Sn and Pb. The precision of SSMS is however by a factor of about 2 worse than that obtained for neutron activation.

Abstract: The zirconium/hafnium ratios of zircons are determined using proton induced X-ray emission. Submilligram samples, imbedded in a starch layer and deposited on a 50 μg/cm2 polystyrene carrier, are irradiated for 1020 min with a 5 nA beam of 3.7 MeV protons, while the Hf-Lβ and Zr-Kα X-rays are counted with a Si(Li) detector. The standard deviation per analysis is in the 36 % range. Only few interferences are possible. To eliminate errors due to absorption effects the zircon layer thickness should be above 40 μm or reproducibly thin samples should be employed. Measuring the Hf-Lβ/Hf-Lα ratio for samples and standards might provide a practical check for the absence of absorption errors. The sensitivity is so favourable that, in practice, the minimal sample size is only limited by the minimal amount that can be handled properly.