S.H. Sie

Quantitative pixe microanalysis of geological matemal using the CSIRO proton microprobe

Abstract The proton microprobe opens up new areas of geoscience research, that demand an efficient quantitative PIXE analysis method for trace element levels down to the limits of detection. The method developed at the HIAF laboratory reliably treats statistical fluctuations to provide quantitative PIXE microanalysis down to these levels. The software package (GEO-PIXE) that supports this standardless analysis procedure includes secondary X-ray fluorescence, and is structured to process the large numbers of analyses required in many geoscience applications. Analyses of standard rocks demonstrate the accuracy of the method, and highlight the problems associated with the use of standards in trace element analysis.

SNIP, a statistics-sensitive background treatment for the quantitative analysis of PIXE spectra in geoscience applications

Abstract Statistical fluctuations in X-ray spectra must be treated properly for reliable quantitative PIXE analysis. A background approximation that provides reliable treatment of fluctuations, the Statistics-sensitive Non-linear Iterative Peak-clipping (SNIP) algorithm, is described. Monte Carlo simulation demonstrates the stability of this background approximation, and hence the deduced trace element concentrations, over a wide range of counting statistics.

Proton microprobe-determined partitioning of Nb, Ta, Zr, Sr and Y between garnet, clinopyroxene and basaltic magma at high pressure and temperature

Abstract Application of a proton microprobe to experimentally produced mineral-liquid pairs has enabled determination of Nb and Ta partition coefficients ( D ) between garnet, clinopyroxene, and a Nb- and Ta-enriched hydrous basaltic liquid at 2.5 GPa, 1100°C. For garnet D Nb =0.02, D Ta =0.06, and for clinopyroxene D Nb =0.005 and D Ta =0.013. Both minerals significantly fractionate Nb and Ta from each other and may have an important role in modifying the Nb/Ta ratio of evolving magmas from that of their parent magma or their source region. For eclogite melting up to 40% change could occur in the Nb/Ta ratio. Additional partition coefficients obtained using the proton microprobe include values of: for garnet, D Zr (0.4–0.7), D Sr ( D Y (>2.5–9); for clinopyroxene, D Zr (0.10), D Sr (0.06) and D Y (0.9); and for orthopyroxene D Zr (0.18), D Sr (0.04) and D Y (0.18). These were determined for natural element abundance levels, and are consistent with data obtained for artificially enriched compositions, verifying Henrys law behaviour and the valid application of these data to geochemical modelling of natural systems.

Proton microprobe determined partitioning of Rb, Sr, Ba, Y, Zr, Nb and Ta between experimentally produced amphiboles and silicate melts with variable F content☆

Abstract A proton microprobe was used to measure partition coefficients for Rb, Sr, Ba, Y, Zr, Nb and Ta between experimentally produced amphiboles and hydrous basaltic melts. A limited amount of data was also obtained for the distribution of trace elements in clinopyroxene and mica. Partition coefficients for trace elements in amphibole and basanite melts are (at 1σ): Rb 0.34 ± 0.14; Sr 0.33 ± 0.07; Ba 0.46 ± 0.16; Y 0.6 ± 0.2; Ti 0.95 ± 0.19; Zr 0.25 ± 0.06; Nb 0.08 ± 0.01; and Ta 0.09 ± 0.03. Only small, generally non-systematic differences in these values are observed with variation in pressure (10–20 kbar) and temperature (1000–1050°C), but large differences accompany changes in melt composition and F content. For a (F-free) basaltic andesite melt, at 20 kbar and 950°C, amphibole/melt distribution coefficients are: Rb 0.07 ± 0.01; Sr 0.35 ± 0.03; Y 1.3 ± 0.1; Ti 1.75 ± 0.12; Zr 0.35 ± 0.06; Nb 0.21 ± 0.01; and Ta 0.19 ± 0.02. The data support proposals that residual amphibole in mantle source regions for some nephelinites explains their relatively high HFSE/LILE ratios. In contrast, the data do not favour amphibole as the cause of characteristically low HFSE/LILE observed in mantle-derived island arc basalts. The increases in partition coefficients (excepting for Rb) for the more SiO2-rich melt are consistent with similar trends observed in phenocryst-matrix pairs from volcanic rocks. The compositional dependence of HFSE partition coefficients increases with increasing field strength. This trend can be related to steric effects within polymerised aluminosilicate units of the melt phase. Amphiboles grown from F-enriched melts are relatively depleted in TiO2, Al2O3, CaO and incompatible trace elements. These effects are only large, however, at high F concentrations (> 2 wt%). The effects of F on incompatible elements will be least in melts containing high concentrations of Al2O3, FeO, MgO and CaO. For these reasons, it is unlikely that concentrations of HFSE and other incompatible elements in natural magmas (with the possible exception of some rare F-and SiO2-rich magmas) are significantly affected by F.

Ni in chrome pyrope garnets: a new geothermometer

Proton microprobe analyses of the minerals in garnet-peridotite xenoliths from kimberlites show that the partitioning of Ni between chrome pyrope garnet and olivine is strongly temperature(T)-dependent. The range of Ni contents in olivines is small relative to that in the analyzed garnets; a geothermometer therefore can be derived, based only on the Ni content of garnet. This allows estimation of T for single Cr-pyrope grains, such as the inclusions in diamonds, if these can be assumed to have equilibrated with olivine.

Trace-element zoning in garnets from sheared mantle xenoliths

Proton-microprobe analyses of garnets from sheared high-temperature ultramafic xenoliths reveal marked zonation of trace elements, paralleling trends in major and minor elements. Garnet rims (600–1000 μm wide) are enriched in Fe, Ti, Zr, Y and Ga, and either enriched or depleted in Cr, relative to cores. Zoning profiles for Ti and Zr are S-shaped and extend further into the grains than the Cr and Ga gradients. The profiles are consistent with the formation of Ti, Zr, Y-enriched garnet overgrowths, followed by diffusive equilibration between rim and core over years to hundreds of years. This enrichment in Fe, Ca, Al and incompatible elements is ascribed to melt infiltration and consequent melt-crystal exchange and garnet growth, shortly before eruption. Zr/Y is 1 to 2 in garnet cores but 4 to 5 in rims, and so the infiltrating melt may have been relatively alkalic. Major and trace element concentrations in such high-temperature sheared xenoliths are not likely to resemble those of primative mantle or of residual mantle depleted by melt extraction.

Quantitative analysis of PIXE spectra in geoscience applications

Abstract Geoscience research using the proton microprobe demands an accurate analysis method for all element concentrations, particularly those at low levels where the proton probe is providing unique trace element data. A method, and its integration into the GEO-PIXE suite of computer codes, is described which meets these requirements, with particular emphasis on elements near the detection limit. Results using the method are compared with accepted analyses of standard rocks and electron microprobe analyses. Monte Carlo tests indicate that the method is accurate for low level trace elements near the detection limit, and for demonstrably different spectra for which standards comparisons cannot be readily made.

Conditions of diamond growth: a proton microprobe study of inclusions in West Australian diamonds

Crystalline primary inclusions in diamonds from the Argyle and Ellendale lamproites have been analyzed for Mn, Ni, Cu, Zn, Ga, Pb, Rb, Sr, Y, Zr, Nb, Ta, Ba and Mo by proton microprobe. Eclogite-suite inclusions dominate at Argyle and occur in equal proportions with peridotite-suite inclusions at Ellendale. Eclogitic phases present include garnet, omphacitic clinopyroxene, coesite, rutile, kyanite and sulfide. Eclogitic clinopyroxenes are commonly rich in K and contain 300–1060 ppm Sr and 3–70 ppm Zr: K/Rb increases with K content up to 1400 at 0.7–1.1% K. Rutiles have high Zr and Nb contents with Zr/Nb=1.5–4 and Nb/Ta ∼16. Of the peridotite-suite inclusions, olivine commonly contains > 10 ppm Sr and Mo; Cr-pyropes are depleted in Sr, Y and Zr, and enriched in Ni, relative to eclogitic garnets.Eclogite-suite diamonds grew in host rocks that were depleted in Mn, Ni and Cr, and enriched in Sr, Zn, Cu, Ga and Ti, relative to Type I eclogite xenoliths from the Roberts Victor Mine. Crystallization temperatures of the eclogite-suite diamonds, as determined by coexisting garnet and clinopyroxene from single diamonds, range from ∼1085 to ∼1575° C. Log KD (Cicpx/Cignt) varies linearly with 1/T for Zr, Sr and Ga in most of the same samples. This supports the validity of the temperature estimates; Argyle eclogite-suite diamonds have grown over a T range ≥400° C. Comparison with data from eclogite xenoliths in kimberlites suggests that KDSrand KDZrare mainly T-dependent, while KDGamay be both temperature-and pressuredependent. KDNi, KDCuand KDZnshow no T dependence in these samples.In several cases, significant major-and/or trace-element disequilibrium is observed between different grains of the same mineral, or between pyroxene and garnet, within single diamonds. This implies that these diamonds grew in an open system; inclusions trapped at different stages of growth record changes in major and trace-element composition occurring in the host rock. Diamond growth may have been controlled by a fluid flux which introduced or liberated carbon and modified the composition of the rock. The wide range of equilibration temperatures and the range of composition recorded in the inclusions of single diamonds suggest that a significant time interval was involved in diamond growth.

Trace element geochemistry of ilmenite megacrysts from the Monastery kimberlite, South Africa

Abstract Ilmenite megacrysts in the Monastery kimberlite occur both as discrete monomineralic crystals and intergrown with all the other phases of the Cr-poor megacryst suite (cpx, opx, garnet, Fe-rich olivine, phlogopite and zircon). The ilmenites show systematic variations in trace element content which are interpreted in terms of a fractional crystallization model. Covariation of major and trace elements in the ilmenites with respect to their Nb content defines smooth curves, with breaks in the trends corresponding to changes in the inferred cumulate assemblage. Nb apparently behaved as an incompatible element in the megacryst magma throughout its crystallization history, and the Nb content of ilmenite serves as a useful fractionation index. After the appearance of ilmenite, the crystallization sequence of the Monastery megacryst suit is ilmenite+cpx+garnet+opx, followed by ilmenite+phlogopite, then ilmenite+zircon±phlogopite and finally by ilmenite+zircon+olivine+phlogopite. The incompatible behaviour of Nb indicates that ilmenite was overall never a predominant phase in the cumulate assemblage. Elevated Cr contents in late-stage ilmenites cannot be explained by the simple fractional crystallization model, and may require another process such as magma mixing or magma reaction with wall rock. The parent megacryst magma must have been highly magnesian and enriched in incompatible trace elements and in this respect may have been similar to meimechite. The late differentiate of this magma cannot be kimberlite, but must be undersaturated and high in iron, potassium, titanium and incompatible elements.

Quantitative PIXE microanalysis of fluid inclusions based on a layered yield model

Abstract The PIXE microanalysis of geological material using the proton microprobe often involves the treatment of layered or complex three-dimensional samples. While many problems remain intractable, some specific cases lend themselves to quantitative analysis. In addition to thick targets, the two sample geometries that have been treated quantitatively at the CSIRO are fluid inclusions and mineral grains in polished thin sections. The analysis of synthetic targets of known thickness layers has been used to assess the accuracy of this calculation. Also, the effects of inclusion geometry and uncertainties in depth and thickness, on the deduced concentrations in the fluid, have been investigated. The method was applied to the in situ microanalysis of fluid inclusions in quartz. Direct optical measurements of depth and thickness were used, and the quartz matrix was analyzed separately. For many inclusions, the observed Cl K α K β ratio was used as a direct measure of inclusion depth. The results compared well with analyses obtained by the bulk crush-leach method.