Building a Theoretical Framework for Melting of Chemically Heterogeneous Mantle
Variations in trace element and isotopic signatures of ocean island basalt (OIB) are often linked with mantle lithological heterogeneities derived from recycled oceanic crust. Recent experimental studies have revealed that partial melting of mafic lithologies such as pyroxenites can also account for OIB major element diversity and have also addressed modification of mafic components via interaction with peridotites. However, the ultimate dynamic fate of mafic bodies interacting with surrounding mantle peridotites, along with the implications for terrestrial magmatism at large, is yet to be explored in a systematic way. As a first step towards building a theoretical framework, we perform a series of calculations that link multicomponent phase equilibria to mass and energy balances during isenthalpic equilibration of mafic and peridotitic compositions. The calculations are based on the self-consistent pMELTS thermodynamic algorithm of Ghiorso et al. (2002). In each calculation, basaltic and peridotitic compositions are first equilibrated separately at a given temperature and pressure to obtain a reference state. These two components are then mixed at a range of mass proportions and reequilibrated under identical conditions of T and P. The nature of chemical reaction between mafic bodies and peridotite matrix can be identified via comparison of the reequilibrated and reference states. We explore the influence of imposed constraints including pressure, mass proportion of basaltic material compared to peridotite, and initial mafic and peridotite compositions. In general, the bulk of modern subducted oceanic crust is of MORB-like composition, but removal of small degree partial melts or siliceous fluids during subduction can lead to silica-deficient compositions dominated by garnet and clinopyroxene. Accordingly, we examine MORB-like and silica deficient garnet pyroxenite compositions. We also test an Fe-rich Archaean basalt. We discuss the results with respect to the partial melting behaviour of the two-component mantle.
The Plumbing System of the Franklin Magmatic Event on Victoria Island, N.W.T., Preliminary Results
The feeder system to the Natkusiak flood basalts (ca 723Ma) is superbly exposed in the Minto Inlier of Victoria Island, as 10 to 100m-thick sills extending laterally over 100s of km. The sills (26-30) are hosted by the Neoproterozoic Shaler Supergroup, composed of clastic metasediments, carbonates, mudstones and evaporites, capped by the Natkusiak lavas. The Neoproterozoic rocks are unconformably overlain by flat-lying Paleozoic sandstones and carbonates. ENE-trending normal faults dissect the unconformity. The upper sills are rather monotonous dolerites and resemble distal Franklin sills. They contain minor cumulus olivine, abundant Fe-Ti-oxides and interstitial granophyre, common magmatic sulphides, and have amphibole-bearing granophyric pods in their upper thirds. Our sampling has focussed on a set of olivine-rich sills, which have more significant potential for Norils=k-type mineralization. Thin (<1m) olivine-rich upper chilled margins (ca 25% phenocrysts) are locally developed, attesting to the influx of olivine-charged slurries. These give way rapidly to olivine-free diabasic rocks beneath, suggesting that the suspended phenocrysts settled out rapidly. Fieldwork, examination of drillcore, and modeling of geochemical data imply that olivine is significantly enriched in the lower third of some sills, and locally reaches ca 55% by weight. Geochemical modeling suggests that the extracted olivine was roughly of constant composition throughout the peridotite, consistent with injection of a slurry followed by en-masse settling of olivine. Some sills have an olivine clinopyroxenite band above the olivine-rich cumulates. The euhedral, sector-zoned clinopyroxene persists into the base of the diabasic zone, and texturally is a cumulus phase.
Experimental Constraints on Lithium Exchange Between Clinopyroxene, Olivine and Aqueous Fluids at Elevated P and T
Lithium partitioning has been experimentally determined between clinopyroxene, olivine and hydrous fluid at 800-1100°C and 1 GPa. Experiments were done by equilibrating natural mineral powders with Li-doped fluids in sealed noble metal capsules for durations of 48 to 142 hours. Lithium concentrations were determined in run-product solids using LA-ICPMS, and partition coefficients determined by mass balance. Lithium was found to be more compatible in olivine than clinopyroxene, and both sets of D-values decrease with temperature following the relationships: ln DLicpx/fluid = -7.38(±0.55) + 7.04(±0.66) × 1000/T and ln DLiol/fluid = -5.93(±1.60) + 6.46(±2.09) × 1000/T. Similar slopes indicate that lithium partitioning between olivine and clinopyroxene is independent of temperature, which has been confirmed in experiments containing both these phases. Preliminary experiments examining the effect of REE content and fO2 suggest that DLiol/cpx may be a function of crystal chemistry. Lithium isotope partitioning between clinopyroxene + fluid and olivine + clinopyroxene has also been determined in experiments in which fluids of known isotopic composition were equilibrated with natural cpx and olivine. Run-product solids were analysed by MC-ICPMS and SIMS. The isotopic fractionation between clinopyroxene and fluid at temperatures greater than 900°C is ∼2.5‰ (±2‰) and the measured isotopic exchange between olivine and clinopyroxene is ∼5‰ (±4‰). These results have been supplemented by lithium diffusion measurements done on a natural clinopyroxene at 800-1000°C and in San Carlos olivine at 1000°C. The experiments were conducted in sealed silica tubes with either a LiCl Li-source or NaCl Li- sink and containing solid oxygen buffers. The lithium diffusion coefficient is independent of the diffusion gradient as values are the same if the flux of lithium is into or out of the crystal. Additionally the lithium diffusion coefficient in clinopyroxene appears to be independent of fO2. A single measurement of lithium diffusion in olivine was made and found to be two orders of magnitude slower than for clinopyroxene at similar conditions. Our results indicate that slab-derived fluids migrating by porous flow will rapidly exchange Li with their mantle wall-rock, effectively buffering the fluid composition close to ambient mantle values, and rapidly attenuating the slab Li signature.
A Method for the Flux Growth of Intermediate Composition Olivine
Though solid solution of iron and magnesium between forsterite (Mg2SiO4) and fayalite (Fe2SiO4) is possible in the olivine crystal structure, the high oxygen fugacity condition of the terrestrial mantle inhibits the widespread crystallization of intermediate (Fo40-Fo60) composition olivine. This limitation is not the same for some other inner solar system bodies (e.g. the Moon and Mars), where conditions are reducing and olivine compositions are wide ranging. Unfortunately, the amount of samples from the Moon and Mars is extremely limited; with only Apollo and Luna mission samples, lunar meteorites, and Martian meteorites available for direct mineralogic and petrologic characterization. These characterizations have provided a useful basis for many spectroscopic and modeling interpretations, but many fundamental questions remain and may only be answerable through either direct observation of rocks or by analog experimentation. The motivation for our work on growth of intermediate olivine crystals, therefore, is to create realistic starting material for use in Mars and Moon analog experiments. A variety of crystal growth methods have been previously used to synthesize olivine, including: the Czochralski-pulling (CZ) method, the floating-zone image furnace (FZ) method, and sol-gel processing techniques. Both the CZ and FZ methods have the advantage of producing large crystals, but the growth apparatus and regulation of reduced atmospheric conditions during growth can make these techniques both time and cost intensive. Sol-gel processing to produces olivine fibers is a useful chemical technique, but obtaining larger grain sizes can be difficult. An alternative method for crystal growth is through the use a flux, which can grow crystals relatively quickly and inexpensively. We have grown synthetic crystals of intermediate composition (Fo30-Fo70) olivine using a lithium borate (B5Li3O9) flux. The starting material was a mixture of magnesite (MgCO3), siderite (FeCO3), and quartz (SiO2) powder in a 1:1:1 ratio. The advantage of using siderite is that the iron is already present in the ferrous form. Upon heating and decarbonation, this mixture represents a bulk composition of Fo50 (FeMgSiO4) olivine. Flux was then added to the starting material mixture so that the final mixture was 50% starting material and 50% flux by weight. This final mix was then placed in a platinum crucible that was heated to 1100 °C in a vacuum furnace for three days. The use of a vacuum furnace ensured that conditions remained reducing during crystal growth. The result was growth of olivine crystals that are generally small (< 1 mm in length) and have euhedral crystal form. These crystals have been analyzed by electron microprobe, and are systematically zoned from core to rim with Mg-rich cores (∼Fo70) transitioning to Fe-rich rims (∼Fo30). This zoning represents an expected heterogeneity due to olivine growth from a finite reservoir of starting material. The flux growth of this intermediate composition olivine was primarily a 'proof of concept' experiment, and showed that olivine crystals can be grown using a flux under sub-solidus conditions. Additional crystal growth experiments would be useful to gauge the response of olivine to changes in temperature, duration, and composition of the flux + starting material mixture.
The Influence of Database Configuration on the Derivation of Trace Element Partitioning Expressions for Clinopyroxene
Clinopyroxene has a large role in controlling the trace element budget in the crust and upper mantle due to its moderate level of compatibility with respect to many trace elements (e.g. REE, HFSE) and its role in melting and crystallization processes. Decades of experimental work has illustrated the high degree to which the partitioning of most elements is dependent on phase composition, temperature and pressure. In order to quantitatively describe clinopyroxene partitioning it is required to derive expressions using theoretical constraints that are calibrated with experimental data [1-6]. The experimental database available for such an analysis has grown considerably in the past decade, both in number of experiments and the range of composition and experimental conditions. Our research is focused on the compilation and evaluation of this newly available experimental clinopyroxene data. This compilation was conducted as part of the ongoing update of the GERM partition coefficient (Kd) database (http://earthref.org/databases/KDD). Previous versions of the partitioning database provided a summary of partitioning information from the literature, however without data on phase compositions. In this current update, data on phase compositional and experimental conditions are being added on an investigation by investigation basis. One of the areas of initial concentration has been information on clinopyroxene, due to the relatively large size of the available data and the significance of the dataset as a whole. As an example of the utility of the database, we have compiled the available clinopyroxene trace element experimental data on natural, doped natural and synthetic materials, focusing on the data for the rare earth and high field strength elements. Our initial analysis documents the significant difference in the numbers of experiments between different elements and groups of elements. In addition, the distribution of pyroxene composition, melt composition, and experimental temperature and pressure varies greatly both between and within the groups. These differences all have an effect on any method of calibration/regression describing trace element partitioning constraints. This compilation is both a contribution to experimental investigators and a method of providing better understanding of the constraints on the numerical models for element partitioning in clinopyroxenes. Understanding the characteristics of the database we are working with when deriving numerical expressions will help us avoid mistakes in application, as well as provide us with helpful guidelines for future experimental designs. Sources 1 - Forsythe et al. (1994) Chem. Geol., 117, 107-126. 2 - Hack et al (1994) Chem. Geol. 117 89-105. 3 - Wood and Blundy (1997) Contrib Mineral Petrol. 129. 166-181. 4 - Wood and Trigila (2000) Chem. Geol. 172. 213-223. 5 - Wood and Blundy (2001) EPSL. 188. 59-71. 6 - Blundy and Wood (2003) EPSL. 210. 383- 397.