Abstract: All published procedures for multi-element preconcentration of trace elements, prior to x-ray fluorescence analysis of water, are reviewed and critically evaluated. Most preconcentration methods applied to the determination of single elements in water are also listed.

Abstract: For multi-element analysis of human urine, 25-ml samples doped with yttrium as internal standard are evaporated gently and then ashed up to 460°C overnight. The residue is pelletized and analysed by energy-dispersive x-ray fluorescence. Acid addition to facilitate the digestion is not mandatory. Recoveries are nearly quantitative for traces of Fe, Ni, Cu, Zn and Sr, to a lesser extent for lead, but not for arsenic or selenium. The standard deviation per measurement is typically around 6%. The detection limits are such that some 10 elements can be determined simultaneously in normal urine, and possibly more in cases of importance to toxicology or industrial hygiene.

Abstract: The relative merits of eight procedures for preconcentrations of trace metal ions from natural water samples and synthetic solutions are evaluated. Spikes (100 μg l−1 ) of Mn, Co, Zn, Eu, Cs and Ba and the corresponding radioactive tracers were added to batches of drinking water, estuarine water, sea water, ground water, twice-distilled water and ahumic material solution. After equilibration for 25 months, the following techniques were applied: passage through columns of Dowex Al chelating resin and ofsilylated silica gel, filtration through laminate membrane filters and chelating diethylenetriamine cellulose filters, precipitation with sodium diethyldithiocarbamate and l-(2-pyridylazo)-2-naphthol, extraction with ammonium pyrrolidinedithiocarbamate, and chelation by 8-quinolinol (oxine) followed by adsorption on activated carbon. The quantitative characteristics of these techniques and the influence of the water matrix effects are discussed, as well as the applicability for x-ray fluorescence analysis.

Abstract: A computer routine was developed for qualitative and quantitative analysis of photographically recorded spark-source mass spectra. Particular attention is given to the case of a graphite matrix. The program starts from the line intensities (expressed as Seidel values) and isotope masses calculated from the densitometer readings by a commercially available routine. From the intensities in the different exposures (typically 15 stages), it computes the parameters for the linear parts of the density curves for each ion. Taking into account mutual interferences of multivalent ions, isotope or C-clusters, oxide, carbide and dicarbide ions, the program automatically identifies and then quantifies the elements present. The precision of the results is around 5%. Reading and complete processing of one photoplate is achieved within 23 h.

Abstract: The determination of silicon via the 28Si(n,p)28 Al reaction by means of 14-MeV neutrons is applied to the analysis of pollution and natural aerosols. A Whatman 41 filter (40 cm2) on which airborne particulate material has been collected is compressed into a 3 × 12.7 mm pellet. Standards are prepared in the same way from clean filters spiked with a silicate solution. After a 50-s irradiation and a 75-s decay time, the sample is counted for 2 min with 5 × 5 NaI(Tl) well detector. The 1.779-MeV photopeak of 28Al is measured with a single channel sealer chain or with a multichannel analyser. The reproducibility, sensitivity and liability to interference from other elements were investigated for both counting systems. The homogeneity of the pellets and the filters was checked. The overall precision of one single-channel determination was estimated to be 3.5% after a 24-h high-volume sampling time. Samples collected in urban, industrial and remote areas with concentrations ranging from 0.05 to 15 μg Si m-3 air were analysed and the results are discussed.

Abstract: A general standardization method is described for the determination of oxygen in solid samples via the 16O(n,p)16N reaction. Two systems of flux monitoring are considered: the sample versus standard comparator method and BF3 monitoring. The average flux in sample and standard, fast neutron shielding, fast neutron scattering, absorption of fast neutrons, absorption of 16N γ-rays and counting efficiency of sample and standard are considered. The influence of the target diameter on the obtained correction factors has also been studied. Total achievable accuracy is believed to be about 1%.

Abstract: The precision of the 14-MeV neutron activation determination of oxygen in steel has been examined as a function of the oxygen content for a large number of steel and cast iron samples. The experimental and the statistically expected standard deviations have been compared. In the conditions used, 2.5 counts from 16N were registered for each p.p.m. of oxygen in the samples. The neutron flux was monitored by a simultaneously irradiated oxygen standard.

Abstract: Non-destructive 14-MeV neutron activation analysis for silicon in steel has been applied with 56Mn as internal standard.56Mn is formed from the iron matrix via the 56Fe(n,p)56Mn reaction. Several methods of internal standardisation via56Mn are discussed. The 0.84-MeV photopeak of 56Mn is recommended if steel samples of about the same composition are to be analysed. Chemically analysed steel samples are used as silicon standards. A precision of 0.7% was obtained for an analysis plus standardisation time of 13 min. Special attention was paid to interferences produced by concentration changes of impurity elements. Several possible sources of errors were investigated.