V22B-01 INVITED
PtFe Nano and Micro-Nuggets in Experimental Silicate Glasses
Noble metal nuggets (<20nm - 2μm) are ubiquitous in the quenched silicate glass of experiments that equilibrate solid or liquid noble metals (or Fe-metal alloys) with liquid silicate under reducing conditions. Inclusion or exclusion of nuggets from the analysis of 1-atm experiments alters the solubility estimates by up to 12 orders of magnitude at low fO2. Most studies have attributed the presence of nuggets to mechanical contamination from fragments of the equilibrating metal. In contrast, Cottrell and Walker (2006) presented evidence that nuggets of Pt emerge from solution in the silicate upon thermal quench of the experiment. The former scenario requires judicious removal of the nugget signature from the solubility estimate; the latter does not. Here we show that the inclusion/exclusion of nuggets from analysis is largely irrelevant to the solubility at high T and P using data from two very similar sets of high pressure experiments: one which included nuggets (Cottrell and Walker, 2006) and one which excluded nuggets (Ertel et al., 2006). The pivotal issue of "to count or not to count" nuggets appears to be moot at very high T. Both studies document the positive T dependence of Pt2+ solubility in excess of that extrapolated from 1-atm data. If Pt is neutrally speciated in solution at high T, and if this is the nugget-forming species, then elimination of nuggets from the analysis only eliminates the contribution from Pt0. Therefore, in a 60 second LA-ICPMS trace, inclusion of only the 2-3 seconds yielding the lowest Pt concentration, only attempts to return the Pt2+ solubility, which demonstrably covaries systematically with T and fO2, and does not document the full solubility including Pt0. Nevertheless, above 1550 °C, the correlation between Pt concentration and T is far more systematic for the nugget-bearing analyses than the nugget-excluded ones in Ertel et al.'s experiments. Moreover, in our very high T experiments, in which the silicate melt quench rate varies with respect to charge margins, areas of silicate glass in which no nuggets are present return the same Pt concentration as areas in which Pt is present in the melt as Pt2+ and as exsolved Pt0 nuggets. Evidently, rapidly quenched nugget-free areas have both species still dissolved at the atomic level. Our conclusion is that both sets of experiments suggest Pt partition coefficients relevant to magma ocean settings that are orders of magnitude lower than values extrapolated from high fO2 experiments at 1 bar. This conclusion significantly reduces the need for a late veneer contribution to the mantle's Pt budget. http://mineralsciences.si.edu/staff/pages/cottrell.htm
V22B-02
Processing micronuggets-containing LA-ICP-MS data
In many experimental studies of solubility of metals in silicate melts, there are metal nuggets in the glass at the end of the run regardless of the technique used. They are often so small that the only way of discovering them is by examining the spectra obtained by mass spectrometry technique such as LA-ICP-MS, where the spectra (cps of element of choice vs. ablation time) show spikes and bumps. There is no common agreement on how to process data that contains micronuggets, as it is unclear how and when they form. If the nuggets are in equilibrium with the melt during the run, they need to be removed from the signal as including them would lead to an overestimated solubility. Yet if they are exsolved on quench, the metals were dissolved in the melt at run conditions and the nuggets need to be included in the data in order to yield the total solubility data. Most authors have considered them to be quench phase and included them in their analysis. The study presented here was done on a series of Cu, Au and Pd solubility experiments in a minimum An-Di melt as a function of oxygen fugacity (fO2). All runs were conducted in a 1-atm furnace at 1300°C using a standard loop technique, over a range of logfO2's from 0 to -11. 3 to 6 spots per sample were analysed with LA-ICP-MS. The spectra show that the nugget density increases from no nuggets at high fO2 to high density at low fO2. The data was evaluated both by integrating all nuggets and by determining a baseline on the spectra. Each set was plotted against fO2, assuming that whichever dataset appears to be best would produce a strong argument in favour of the correct method. The data agreeing best with the general trend of decreasing solubility with decreasing fO2 was obtained when choosing the overall lowest point of a series of analysis on a given sample; the results indicate that including the nuggets in the laser signal overestimates the solubility by over an order of magnitude at high nugget density. This result implies that the nuggets are equilibrium phases at run conditions. The fact that nugget formation is so strongly fO2 dependent may present an additional argument for this hypothesis, as quench processes are unlikely to be very strongly affected by fO2.
V22B-03 INVITED
Mechanically assisted equilibration of Siderophile elements in silicate melts
Solubilities of a wide range of siderophile elements (SE: Ni, W, Re, Ir, Os, Pt, Rh) in an analog basaltic melt (corresponding to the 1 atm anorthite-diopside eutectic) have been determined using the mechanically assisted equilibration (MAE) technique of Dingwell et al. (1994). Here we present a review of the data obtained and the experience gained in this decade of work. In particular, the major experimental and analytical challenges of the nanonugget problem as well as implications for core/mantle equilibria and core formation scenarios in terrestrial planets are reviewed. To accomplish this, a comprehensive and detailed description of the MAE technique is provided. A general background overview of solubility experiments regarding siderophile elements (SE) is also supplied. In these studies, major element composition was routinely determined by electron microprobe analyses (EMP), whereas trace elements were determined using a wide variety of analytical techniques (Ni, W: INAA, EMP, ICP-AES; Re, Ir, Pt, Rh and Os: INAA, SIMS, dissolution- (diss-ICP-MS) and Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICPMS)). The last technique, in particular, has demonstrated its powerful capabilities as a truly micro-analytical technique supplying information both on absolute trace element concentrations and on small scale heterogeneities in run products. All investigated SEs exhibited solubilities consistent with their presence as stoichiometrically dissolved oxide species in the melt phase. There was no indication for zero-valent species dissolved at any oxygen fugacity (fO2) condition. In the case of highly-SE (HSE: Pt, Rh, Re, Os, Ir), INAA results appear to indicate a decrease of HSE solubility with decreasing fO2 down to a fO2 limit which depends on the investigated HSE. Below this limit, bulk HSE concentrations remain either constant with large variations or increase with further fO2 decrease. Duplicate analyses of samples by LA-ICPMS reveal increasing amounts of so-called nanonuggets with decreasing fO2, which lead to high HSE concentrations in the glass samples obtained by bulk analytical methods such as INAA. The formation of HSE (and potentially some SE) nanonuggets in low fO2 samples raise the question of whether nanonuggets are formed either during the quench by precipitation from precursor species dissolved homogeneously in the melts, or are precipitated in situ at high temperature due to true thermodynamic oversaturation. Through the combination of MAE technique with LA-ICPMS micro-analytics it has been possible to extend our knowledge of the solubility behaviour of HSE to unprecedentedly low fO2 values. Clarification of the solubility mechanism for SE as well as the nanonugget issue, however, will undoubtedly require further novel experimental designs.
V22B-04
Disequilibrium Experiments and Micro-XANES Analysis: Novel Tools to Unravel the Speciation of Sulfur in Silicate Melts.
Sulfur is an element of interest in magmatic processes for several reasons, some of which are related to the oxidation state of S during magma generation and evolution. For example, S as sulfide (S2-) controls the behavior of chalcophile and highly siderophile elements, whereas S as sulfate (S6+) is responsible for high- S explosive volcanic eruptions, which can cause global cooling by increasing the Earth's albedo. An adequate understanding of the speciation of S in magmatic systems and the transition from S2- to S6+ is therefore needed to understand these processes. Data from natural samples is incomplete and experimental data are required to link natural data with oxygen fugacity (fO2). However, the change in speciation from sulfide to sulfate in silicate melts is difficult to simulate experimentally because: (a) common capsule materials react with S (e.g. Pt) or have low melting points (e.g. Au); (b) the change in speciation occurs over a very narrow fO2 interval (FMQ to FMQ+2) and common buffering techniques (e.g. "double capsule" technique with FMQ or NNO buffers) are not sufficient to investigate a wide-enough range in fO2; (c) sulfur solubility in silicate melts in the fO2 range of interest is too low at atmospheric pressures, limiting the use of gas-mixing furnaces to either very oxidized or very reduced conditions. We have used disequilibrium experiments in which sulfate-saturated (i.e. oxidized) basaltic and andesitic melts were reacted with graphite (a reductant) and quenched before the system reached equilibrium. Quenching of the experiments before complete re-equilibration (i.e. complete reduction by graphite) preserved reduction profiles in which sulfate-saturated glass (in the center of the capsule) coexisted with sulfide-saturated glass (at the edge of graphite capsule). We used the ID-21 beamline at the European Synchrotron Radiation Facility (ESRF) to perform micro-XANES analysis at the S K- edge to determine the speciation of sulfur along the reduction profile. Tests with a combination of broad (50 μm to 200 μm) and narrow beams (0.8 μm) showed that beam-sample interaction caused partial reduction of S6+ to S4+ (sulfite) in the glass. However, sulfate reduction under the beam is time- dependent and we modified our analytical protocol to avoid beam damage. By performing rapid scans across the samples by resonant excitation at the energy position of each species of interest we were able to record the transition from sulfide to sulfate in a single experiment over a reduction profile of approximately 120 μm in length.
V22B-05 INVITED
Comparison of synthetic fluid inclusion and quartz-trap methods for determining platinum- group element (PGE) solubility in hydrous salt melts at magmatic conditions
Layered intrusions preserve magmatic inclusions containing hydrous halide melt phases composed of transition metal chlorides with less than 5 wt% H2O. Previous attempts at measuring platinum-group element (PGE) solubility in such high salinity volatiles through the analysis of synthetic fluid inclusions have shown that, although the PGE appear to be highly soluble at geologically-realistic conditions, it is difficult to determine if trapped fluids represent equilibrium fluid compositions. Wide ranges in trapped metal content may result from (i) premature formation of fluid inclusions owing to rapid rates of silicate mineral healing (i.e., hours at magmatic conditions in a saline fluid), (ii) premature entrapment of fluids in which the approach to equilibrium metal solubility requires the generation of the metal-complexing ligands (e.g., HCl) from a slow fluid-mineral or fluid- melt buffer reaction, (iii) the dissolution and reprecipitation of PGE in opened inclusions due to subtle temperature or chemical gradients in run capsules, or (iv) heterogeneous fluid compositions due to the presence of PGE micronuggets or colloids that are more abundant near metal-fluid or metal-buffer interfaces. To circumvent some of the problems associated with synthetic inclusion methods, the solubility of platinum in a hydrous salt melt (S-free; 75 wt% CaCl2+MgCl2, 25 wt% H2O) was investigated at elevated T (700oC) and low crustal pressure. The salt melts were reacted with PtAs2 (natural sperrylite) in Pt capsules buffered at an oxygen fugacity of FMQ-1 using a mixed gas buffer and by the solid mineral assemblage tremolite- diopside-enstatite-quartz which fixes the concentration of relevant metal-complexing ligands at run conditions. Salt melts were trapped (simultaneously) in the matrix of a quartz trap (granulated natural quartz partially isolated in a smaller gold capsule) and in synthetic melt inclusions trapped in pre-fractured quartz. After quenching, the melt inclusions and quartz trap material (in a frozen state) were analyzed by laser ablation ICP-MS (ETH Zürich). Platinum solubility in the presence of the mineral sperrylite is in the low ppm range. Analysis of different portions of the quartz traps using a 90 micron pit size yielded relatively consistent platinum concentrations of 4.9 ppm +/- 1.4 ppm (2 sigma, n=28) with a routine detection limit of 80-90 ppb for Pt. Analyses of traps at different experiments run durations indicate that Pt equilibrium is reached after 150 hours and is approached from over- saturation. Based on these observations, it is suggested that the measured concentrations are representative of Pt solubility at run conditions and that the laser sampling size sufficiently overcomes local heterogeneity in the distribution of Pt quench products. By contrast, Pt concentrations in hydrosaline melt inclusions from different areas of a 3 mm x 10 mm quartz cylinder range from below detection limits (0.2 ppm) to 4.7 ppm (n=102). The range in observed metal concentrations in the fluid inclusions varies between different areas in the quartz cylinder. The data demonstrates that measurements of PGE solubility in hydrous salt melt phases using a frozen quartz trap eliminates some uncertainties concerning the true equilibrum metal solubility that are associated with synthetic inclusion methods. Additionally, the detection limits achievable using the quartz trap method are up to an order of magnitude lower than for synthetic inclusions, owing to the much higher mass of halide melt that may be analyzed in a single ablation.
V22B-06 INVITED
Experimental problems with the measurement of the metal-sulfide-silicate partition coefficients of the Highly Siderophile Elements
The Highly Siderophile Elements (HSEs: the six PGEs plus Re and Au) are defined by their extreme metal/silicate and sulfide/silicate partition coefficients. Accurate values of these partition coefficients as well as of liquid/solid silicate, sulfide and metal partitioning are needed to understand the origin of magmatic sulfide ore deposits, and to constrain models of of core formation in planetary bodies. Partitioning between metal or sulfide and silicate is expected from thermodynamic theory to depend on temperature, pressure, fO2 and fS2, as well as the compositions of all phases, and, unlike the geochemically more familiar partitioning of lithophile trace elements between crystals and silicate melts, these parameters, especially fO2 and fS2, may cause the partition coefficients to vary by orders of magnitude. This large variation makes it difficult to compare studies done under different conditions, so it is essential that the thermodynamics controlling the partitioning be understood. Another problem is when partition coefficients are high, the slightest contamination of the low-HSE phase by the high- HSE phase can obscure the true partitioning behaviour completely. A particular problem has been the presence of `micronuggets' in quenched run products, which may often be due to contamination, for example, in silicate melts due to a persistent suspension of HSE particles at the ppm level, but in other situations may result from exsolution during quenching. In the former case the micronuggets need to be subtracted out to obtain true partition coefficients, in the latter they should be included. It is therefore not always clear whether great differences in reported values of experimentally determined partition coefficients are due to different conditions or micronugget contamination. While advances in microanalytical techniques, particularly laser-ablation ICP-MS, have certainly been helpful, the solution lies in designing experiments that provide an independent test of the results, for example by varying parameters such as fugacities or activities systematically and seeing if the results conform with relations expected from simpler experiments or theory. Conventional reversals are not particularly useful. Some examples will be presented and discussed.
V22B-07
Serpentine dehydration kinetics up to 1023 K and 4 GPa
Serpentine minerals are ubiquitous within the oceanic lithosphere and are hypothesized to be a significant source of water, released during thermal dehydration, in subduction zones. Additionally, serpentinized oceanic mantle has been hypothesized to represent a fluid and incompatiable element reservoir that could be a major part of element cycling in subduction zones. Data detailing the kinetics of serpentine breakdown are difficult to obtain and require in situ analyses. Candela et al., 2007 ( Am. Min., In Press) and Frank et al., 2005 ( GSA Abs. Programs) illustrated that the breakdown of serpentine minerals (chrysotile and lizardite, respectively) is complex and results in the variable release of H2O. Further, it has been noted that the reaction products proceed in a step-wise fashion as a function of temperature and these ‘‘steps'' are not the same for all serpentine minerals. In order to understand the release of volatile phases from serpentine minerals in subduction zones, we studied the thermal decomposition of lizardite over a range of temperature, 500- 750°C in 25° increments, at atmospheric pressure. High-pressure experiments ranged from 2-5 GPa and were conducted from 550-700°C. The thermal decomposition of lizardite was monitored by using a hydrothermal diamond anvil cell and synchrotron X-ray radiation. The experiments were conducted at the GSECARS 13-BM-D beam line (Advanced Photon Source at Argonne National Laboratory), using monochromatic X-ray radiation, with a wavelength of 0.3344 Å, and a MAR 345 online imaging system. The diffraction data were used to monitor the dehydration reactants and products as a function of temperature and time. The reaction products of serpentine dehydration are dominantly forsterite (at all temperatures and pressure of dehydration) with minor amounts of talc-like and anthophyllite-like phases. Enstatite could not be identified definitively. Further, forsterite nucleates before more silica-rich phases, at generally lower temperatures and persists throughout the duration of the experiments, and at lower temperatures than have been reported previously in the bulk of the literature. Increasing temperature, at any given pressure, increased the rate of reaction up to complete dehydration of serpentine in less than five minutes at 700-750°C. Our results constrain the products of serpentine thermal decomposition and the potential mechanism of H2O release within serpentine-rich subduction zones.
V22B-08
New Experimental Measurements of Zircon/Melt Trace Element Partition Coefficients
Zircon was grown from trace element-doped peralkaline melt of the same composition used by Watson (1980). Cold-seal experiments at 1 and 2 kbar and 800°C and piston cylinder experiments at 15 kbar and 900- 1300°C were run for 160-960 hours. Oxygen fugacity was buffered at NNO, HM or IW. Zircon and glass were present in all run products, and small monazite crystals were present in 7 of the 13 experiments. Diameters of run product zircon crystals ranged from 5-15 um at 800°C to 30-50 um at 1300°C. Zircons were analyzed by electron microprobe and run product glass by LA-ICP-MS. Zircon crystals show slight zoning in CL images and in core-to-rim EMP traverses. Trace element partition coefficients D(zircon/melt) were measured for Sc, Ti, V, Y, Zr, Nb, La, Ce, Pr, Nd, Eu, Gd, Ho, Yb, Lu, Hf, Ta, Th and U. At 800°C D values for some elements change slightly between 216 and 336 h, then change little between 336 and 792 h, suggesting that near-equilibrium is achieved after 336 h; less time should be required to reach equilibrium at higher temperatures. Replicate experiments show good agreement, with differences in D values smaller than those between experiments performed at different P-T-fO2-t conditions. In most cases Lu was the most (D 15-59) and Nb (< 0.05 - 0.37) or La (0.01 - 0.97) the least compatible elements. Our measured D values are lower than those estimated by Watson (1980) using mass balance, and unlike Watson (1980) our values usually do not peak in REE diagrams but smoothly increase from La to Lu. Measured D values for REE fall in the middle of other previously published values. D values for Zr (9.8-25) and Hf (9.6-19) are lower than previously measured values because we purposefully chose a melt composition in which zircon is highly soluble to promote crystal growth. Many elements show a steady decrease in D from 1100-1300°C. A positive Ce anomaly is evident at all fO2 values but a Eu anomaly is slight or non-existent. Ce and Eu do not show a strong dependence of D on fO2, but D(U) increases strongly as fO2 decreases from HM to NNO to IW, indicating that U4+ must fit better in the zircon lattice than U6+. The Th/U fractionation factor f = D(Th)/D(U) ranges from 0.43 to 4.8; it is > 1 for all experiments at HM but < 1 for all experiments at NNO and IW, suggesting that high oxygen fugacity may be partly responsible for the relatively high Th/U of magmatic zircons. Watson, E. B., 1980. Note: Some experimentally determined zircon/liquid partition coefficients for the rare earth elements. Geochim. Cosmochim. Acta 44, 895-7.