SCANNING VO~. 14,160-168 (1992) OFAMS, Inc. Received: February 19, 1992 SIMS Analyses of Rare Earth Elements in Natural Minerals and Glasses: An Investigation of Structural Matrix Effects on Ion Yields PIER0 BOTTAZZI. zyxwvutsrqpo LL'lSA ~TTOLINI. zyxwvutsrqpo AND RICCARDO VANNUCCI* C.N.R. Centro di Studio per la Cri5tallochimica e la Cristallografia, and "Dipartimento Scienze della Terra, Via A. Bassi, 4, 1-27 zyxwvutsrqp 100 Pavia. Italy Summary: An investigation of the influence of matrix structure upon secondary ion signals for rare earth elements (REEs) in various silicates (amphiboles, clinopyroxenes, garnets, and plagioclases) was carried out by means of con- ventional energy filtering technique (CEF). Such matrices were analysed both as natural crystals and fused glasses. Of the selected REEs for the different minerals, only La showed an intensity relatively higher than Si in glassy amphiboles and augites. The hypothesis proposed of an ionization behaviour of La that differs with respect to other REEs in silicates has not been confirmed in plagioclases. All other relative REE ion yields from crystals and glasses agreed within the statistics of measurement for all investi- gated matrices. Moreover, any difference between crystal and glass REE ionization appeared not to be attributable to varying charging conditions during analysis under zyxwvu 0- pri- mary beam bombardment. Laboratory contamination dur- ing glass preparation andor the presence of residual light REE-enriched micron-scale accessory phases (e.g., apatite) could be responsible for the higher LdSi signal in augite and amphibole glasses. Results obtained show that the effi- ciency of REE ionization compared with that of Si, gener- ally is not affected by the sample structure as far as high energy secondary ions are concerned; therefore, glasses obtained by fusing silicate minerals are suitable for analyt- ical work and can be employed to calibrate REE ion signals for quantitative purposes. Address for reprints: Luisa Ottolini C.N.R. Centro di Studio per la Cristallochimica e la Cristallografia Via A. Bassi: zyxwvutsrqp 1 1-27 I00 Pavia. Italy Introduction In recent years, microanalysis applied to geochemistry and cosmochemistry has greatly benefitted from the capa- bilities of secondary ion mass spectrometry (SIMS) for its potential of in situ elemental and isotopic investigation, high sensitivity (ppm-ppb range), and spatial resolution at the micron scale. Molecular and multiply charged ions, which constitute a major source of complications in sec- ondary ion mass spectra, have been largely resolved by means of high mass resolution (Compston zyx et zyx al. 1984, Huneke et zyxwvu al. 1983) and various energy-filtering tech- niques (Metson et al. 1984, Shimizu et al. 1978, Zinner and Crozaz 1986). These latter procedures are particularly useful for analysis of elements having mass number 60 amu or higher and generally have been preferred to a high mass resolution technique for trace element investigations in minerals and rocks (Crozaz and Zinner 1985, 1986, MacRae 1987, MacRae and Russel 1987, MacRae and Wu 1990, Ray and Hart 1982, Shimizu 1978, Shimizu and Allegre 1978, Yurimoto et al. 1989). The existence of SIMS matrix effects (Deline et al. 1978, Shimizu 1985) has introduced another difficulty for quantitative analysis: secondary ion intensities are not always linearly related to their elemental concentrations, but dependent to various extents on the concentration of other element(s) present in the matrix. This effect is evi- dent for major constituents rather than for elements dilut- ed in the matrix (Havette and Slodzian 1980, Ray and Hart 1982. Shimizu et zyxw al. 1978). Such phenomena, not yet entirely understood, have led to the necessity to adopt an empirical approach, which relies upon external standards to calibrate the elemental secondary ion signals. As Ganjei and Morrison (1978) point out, the accuracy of sensitivity factors derived from external standards is critically dependent on the matrix effects. If the sample matrix is sufficiently different from the standard, the change in secondary ion signals will result in large analyt- ical errors. In case of trace element analysis in minerals, such a standard should have, in principle, the same major element chemistry as the investigated matrix and possess well-characterised concentrations of the trace elements of interest. Evidence indicates that due to the extreme variety