Abstract: Accuracy and precision of silicon determination in biological matrices (serum, urine, water, beer and spinach) by spectrometric techniques (when necessary after acid destruction) were assessed by means of a collaborative interlaboratory trial. The trial was set up in accordance with ISO 5725-2 (1994). The relative overall repeatability standard deviation was acceptable. It varied between 4% for spinach powder (mean content: 176 mg kg(-1)) and 11% for serum (mean content: 5.33 mg L-1). On the other hand, the relative overall between-laboratory standard deviation was found to vary from a satisfactorily 15% for spinach after destruction (mean content: 3.32 mg L-1) to an unacceptable 107% for spinach powder (mean content: 176 mg kg(-1)). The overall conclusion of the trial was that silicon determination in biological matrices can properly be performed by spectrometric techniques. However, when sample pretreatment (i.e., acid destruction) is needed prior to silicon determination problems still remain.

Abstract: Eight silicate glasses were prepared by directly fusing and stirring 50-100 g each of basalt, andesite, komatiite, peridotite, rhyolite, and quartz-diorite. These are referred to as MPI-DING glasses and were made for the purpose of providing reference materials for geochemical, in-situ microanalytical work. Results from various analytical techniques indicate that individual glass fragments are well homogenised with respect to major and trace elements at the mu m to mm scale. Heterogeneities due to quench crystallisation of olivine have been observed in small and limited areas of the two komatiitic glasses. In order to obtain concentration values for as many elements as possible, the glasses were analysed by a variety of bulk and microanalytical methods in a number of laboratories. From the analytical data, preliminary reference values for more than sixty elements were calculated. The analytical uncertainties of most elements are estimated to be between 1% and 10%.