Apatite solubility in carbonatitic liquids and trace element partitioning between apatite and carbonatite at high pressure

We have measured apatite solubility in calcic carbonatitic liquids and determined apatite/melt partition coefficients for a series of trace elements, including the rare earth elements (REE), high field strength elements (HFSE), Rb, Sr, U-Th-Pb. Experiments were performed between 4 and 6 GPa, from 1200 to 1380 °C, using the multianvil apparatus. Trace element concentrations were determined by laser ablation ICP-MS and electron microprobe. In addition, a specific protocol was designed to measure carbon concentration in the apatites, using the electron microprobe. Two starting apatite samples were used in order to test for the effect of apatite chemistry on partitioning behavior.Apatite solubility is lower in calcitic melts by a factor 3-5 compared to dolomitic melts (3-5.5 vs. 10-18 wt.% P₂O₅ in melt). We interpret this difference in terms of solubility product in the liquid and propose an empirical model for apatite saturation that takes into account melt calcium content. We conclude that calcitic melts that may form by melting of carbonated eclogites could be saturated with residual apatite, contrary to dolomitic melts formed in carbonated peridotites.Compatibility behavior of the REE depends on apatite silica content: REE are compatible in apatites containing 3.5-5 wt.% SiO₂, with values between 1.5 and 4, whereas REE are incompatible in apatites containing 0.2 wt.% SiO₂. HFSE, U, Th, and Y are compatible in silica-rich apatite, with while . Strontium is always retained in the melt, with of the order of 0.5. Lead appears to be incompatible in apatite, although this finding is weakened by almost complete Pb loss to sample container. High silica concentration favors REE incorporation in apatite by allowing for charged balanced coupled substitution. Sulfur and carbonate may also favor REE incorporation in apatite. Our results allow to reconcile previously published experimental determinations of REE partitioning. We use our experimentally determined partition coefficients to investigate the impact of residual apatite during partial melting of recycled carbonated material (eclogite + sediments) and discuss how the chemical characteristics of the produced liquids can be affected by residual apatite.     

Solubility and solution mechanisms of C-O-H volatiles in silicate melt with variable redox conditions and melt composition at upper mantle temperatures and pressures

Solubility and solution mechanisms in silicate melts of oxidized and reduced C-bearing species in the C-O-H system have been determined experimentally at 1.5 GPa and 1400 °C with mass spectrometric, NMR, and Raman spectroscopic methods. The hydrogen fugacity, f_H2, was controlled in the range between that of the iron-wüstite-H₂O (IW) and the magnetite-hematite-H₂O (MH) buffers. The melt polymerization varied between those typical of tholeiitic and andesitic melts.The solubility of oxidized (on the order of 1-2 wt% as C) and reduced carbon (on the order of 0.15-0.35 wt% as C) is positively correlated with the NBO/Si (nonbridging oxygen per silicon) of the melt. At given NBO/Si-value, the solubility of oxidized carbon is 2-4 times greater than under reducing conditions. Oxidized carbon dioxide is dissolved as complexes, whereas the dominant reduced species in melts are CH₃-groups forming bonds with Si⁴⁺ together with molecular CH₄. Formation of complexes results in silicate melt polymerization (decreasing NBO/Si), whereas solution of reduced carbon results in depolymerization of melts (increasing NBO/Si).Redox melting in the Earth’s interior has been explained with the aid of the different solution mechanisms of oxidized and reduced carbon in silicate melts. Further, effects of oxidized and reduced carbon on melt viscosity and on element partitioning between melts and minerals have been evaluated from relationships between melt polymerization and dissolved carbon combined with existing experimental data that link melt properties and melt polymerization. With total carbon contents in the melts on the order of several mol%, mineral/melt element partition coefficients and melt viscosity can change by several tens to several hundred percent with variable redox conditions in the range of the Earth’s deep crust and upper mantle.     

Crystal-melt partitioning of noble gases (helium, neon, argon, krypton, and xenon) for olivine and clinopyroxene

Mineral-melt partition coefficients of all noble gases (^min/meltD_i) have been obtained for olivine (ol) and clinopyroxene (cpx) by UV laser ablation (213 nm) of individual crystals grown from melts at 0.1 GPa mixed noble gas pressure. Experimental techniques were developed to grow crystals virtually free of melt and fluid inclusions since both have been found to cause profound problems in previous work. This is a particularly important issue for the analysis of noble gases in crystals that have very low partition coefficients relative to coexisting melt and fluid phases. The preferred partitioning values obtained for the ol-melt system for He, Ne, Ar, Kr, and Xe are 0.00017(13), 0.00007(7), 0.0011(6), 0.00026(16), and , respectively. The respective cpx-melt partition coefficients are 0.0002(2), 0.00041(35), 0.0011(7), 0.0002(2), and . The data confirm the incompatible behaviour of noble gases for both olivine and clinopyroxene but unlike other trace elements these values show little variation for a wide range of atomic radius. The lack of dependence of partitioning on atomic radius is, however, consistent with the partitioning behaviour of other trace elements which have been found to exhibit progressively lower dependence of ^min/meltD_i on radius as the charge decreases. As all noble gases appear to exhibit similar ^min/meltD_i values we deduce that noble gases are not significantly fractionated from each other by olivine and clinopyroxene during melting and fractional crystallisation. Although incompatible, the partitioning values for noble gases also suggest that significant amounts of primordial noble gases may well have been retained in the mantle despite intensive melting processes. The implication of our data is that high primordial/radiogenic noble gas ratios (³He/⁴He, ²²Ne/²¹Ne, and ³⁶Ar/⁴⁰Ar) characteristic of plume basalt sources can be achieved by recycling a previously melted (depleted) mantle source rather than reflecting an isolated, non-degassed primordial mantle region.     

The influence of melt composition on the partitioning of REEs, Y, Sc, Zr and Al between forsterite and melt in the system CMAS

Partition coefficients for a range of Rare Earth Elements (REEs), Y, Sc, Al and Zr were determined between forsteritic olivine (nearly end-member Mg₂SiO₄) and ten melt compositions in the system CaO-MgO-Al₂O₃-SiO₂ (CMAS) at 1 bar and 1400 °C, with concentrations of the trace elements in the olivine and the melt measured by laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). The REEs and Sc were added at levels sufficient to ensure that concentrations in the olivine were well above the detection limits. The REE partition coefficients decrease with increasing silica in the melt, indicating strong bonding between REEO_1.5 and SiO₂ in the melt. The variation of as a function of ionic radius is well described by the Brice equation for each composition, although a small proportion of this variation is due to the increase in the strength of the REEO_1.5-SiO₂ interactions in the melt with ionic radius. Scandium behaves very similarly to the REEs, but a global fit of the data from all ten melt compositions suggests that deviates somewhat from the parabolas established by the REE and Y, implying that Sc may substitute into olivine differently to that of the REEs. In contrast to the behaviour of the large trivalent cations, the concentration of Al in olivine is proportional to the square root of its concentration in the melt, indicating a coupled substitution in olivine with a high degree of short-range order. The lack of any correlation of REE partition coefficients with Al in olivine or melt suggests that the REE substitution in olivine is charge-balanced by cation vacancies. The partition coefficient of the tetravalent trace element Zr, which is highly incompatible in olivine, depends on the CaO content of the melt.     

Olivine/melt transition metal partitioning, melt composition, and melt structure—Influence of Al for Si substitution in the tetrahedral network of silicate melts

The influence on olivine/melt transition metal (Mn, Co, Ni) partitioning of substitution in the tetrahedral network of silicate melt structure has been examined at ambient pressure in the 1450-1550 °C temperature range. Experiments were conducted in the systems NaAlSiO₄-Mg₂SiO₄- SiO₂ and CaAl₂Si₂O₈-Mg₂SiO₄-SiO₂ with about 1 wt% each of MnO, CoO, and NiO added. These compositions were used to evaluate how, in silicate melts, substitution and ionization potential of charge-balancing cations affect activity-composition relations in silicate melts and mineral/melt partitioning.The exchange equilibrium coefficient, , is a positive and linear function of melt Al/(Al + Si) at constant degree of melt polymerization, NBO/T. The is negatively correlated with the ionic radius, r, of the M-cation and also with the ionization potential (Z/r², Z = electrical charge) of the cation that serves to charge-balance Al³⁺ in tetrahedral coordination in the melts. The activity coefficient ratio, (γ_M/γ_Mg)^melt, is therefore similarly correlated.These melt composition relationships are governed by the distribution of Al³⁺ among coexisting Q-species in the peralkaline (depolymerized) melts coexisting with olivine. This distribution controls Q-speciation abundance, which, in turn, controls (γ_M/γ_Mg)^melt and . The relations between melt structure and olivine/melt partitioning behavior lead to the suggestion that in natural magmatic systems mineral/melt partition coefficients are more dependent on melt composition and, therefore, melt structure the more alkali-rich and the more felsic the melt. Moreover, mineral/melt partition coefficients are more sensitive to melt composition the more highly charged or the smaller the ionic radius of the cation of interest.     

Oxygen fugacity dependence of Os solubility in haplobasaltic melt

Os equilibrium solubilities were determined at 1350 °C over a wide range of oxygen fugacities (−12 < log fO₂ < −7) applying the mechanically assisted equilibration technique (MAE) at 10⁵ Pa (= 1 bar). Os concentrations in the glass samples were analysed using ID-NTIMS. Additional LA-ICP-MS and SEM analyses were performed to detect, visualize and analyse the nature and chemistry of “nanonuggets.” Os solubilities determined range at a constant temperature of 1350 °C from 0.63 ± 0.04 to 37.4 ± 1.16 ppb depending on oxygen fugacity. At the highest oxygen fugacities, Os³⁺ can be confirmed as the main oxidation state of Os. At low oxygen fugacities (below log fO₂ = −8), samples are contaminated by nanonuggets which, despite the MAE technique, were still not removed entirely from the melt. However, the present results indicate that applying MAE technology does reduce the amount of nanonuggets present significantly, resulting in the lowest Os solubility results reported to date under these experimental conditions, and extending the experimentally accessible range of fO₂ for these studies to lower values. Calculated metal/silicate melt partition coefficients are therefore higher compared to previous studies, making Os more siderophile. Neglecting the as yet unknown temperature dependence of the Os metal/silicate melt partition coefficient, extrapolation of the obtained Os solubilities to conditions for core-mantle equilibrium, results in a , while metallic alloy/silicate melt partition coefficients range from 1.4 × 10⁶ to 8.6 × 10⁷, in agreement with earlier findings. Therefore remains too high by 2-4 orders of magnitude to explain the Os abundance in the Earth’s mantle as result of core-mantle equilibrium during core formation.     

Zn/Fe systematics in mafic and ultramafic systems: Implications for detecting major element heterogeneities in the Earth’s mantle

Oceanic basalts, such as mid-ocean ridge basalts (MORB) and ocean island basalts (OIB), are characterized by large isotopic and trace element variability that is hard to reconcile with partial melting of a peridotitic mantle alone. Their variability has been attributed to the presence of heterogeneities within the mantle, such as recycled crust, metasomatized material or outer core contribution. There have been few attempts to constrain the major element composition of those heterogeneities, most studies focusing on incompatible trace elements and radiogenic isotopes. Here, we report Zn, Mn and Fe systematics in mafic and ultramafic systems (whole-rocks and minerals) and we explore their use for detecting lithological heterogeneities that deviate from peridotitic mantle dominated by olivine and orthopyroxene. We suggest that Zn/Fe ratio is a particularly promising proxy. Zn/Fe fractionates equally between olivine, orthopyroxene and melt (e.g. the inter-mineral exchange coefficients  ∼  is ∼0.9-1), and the distribution of Zn/Fe between minerals appears to be temperature-independent within error. In contrast, clinopyroxene and garnet are characterized by low Zn/Fe ratios compared to co-existing melt, olivine and orthopyroxene, that is, and are both <<1. These partitioning behaviors imply that Zn/Fe ratios are minimally fractionated during partial melting of peridotite and differentiation of primitive basalts, if differentiation is dominated by olivine control. Thus, the Zn/Fe ratios of primitive basalts preserve the Zn/Fe ratio of the primary parental magma, providing insight into the signature of the mantle source region. We also infer that Zn/Fe ratios in melts are unlikely to be fractionated by modal variations in peridotitic material but are highly fractionated if garnet and/or clinopyroxene are the main phases in the source during melting. Similar Zn/Fe ratios between MORB and average upper mantle confirm the lack of fractionation during peridotite melting. However, high Zn/Fe ratios of some OIB cannot be explained by peridotite melting alone, but instead require the presence of high Zn/Fe lithologies or lithologies that have bulk exchange coefficients  < 1. All garnet-bearing or clinopyroxene-bearing lithologies, such as eclogites and garnet pyroxenites, fit the latter requirement.     

Experimental comparison of trace element partitioning between clinopyroxene and melt in carbonate and silicate systems, and implications for mantle metasomatism

Experiments in the systems diopside-albite (Di-Ab) and diopside-albite-dolomite (Di-Ab-Dmt), doped with a wide range of trace elements, have been used to characterise the difference between clinopyroxene-silicate melt and clinopyroxene-carbonate melt partitioning. Experiments in Di-Ab-Dmt yielded clinopyroxene and olivine in equilibrium with CO₂-saturated dolomitic carbonate melt at 3 GPa, 1375 °C. The experiments in Di-Ab were designed to bracket those conditions (3 GPa, 1640 °C and 0.8 GPa, 1375 °C), and so minimise the contribution of differential temperature and pressure to partitioning. Partition coefficients, determined by SIMS analysis of run products, differ markedly for some elements between Di-Ab and Di-Ab-Dmt systems. Notably, in the carbonate system clinopyroxene-melt partition coefficients for Si, Al, Ga, heavy REE, Ti and Zr are higher by factors of 5 to 200 than in the silicate system. Conversely, partition coefficients for Nb, light REE, alkali metals and alkaline earths show much less fractionation (<3). The observed differences compare quantitatively with experimental data on partitioning between immiscible carbonate and silicate melts, indicating that changes in melt chemistry provide the dominant control on variation in partition coefficients in this case. The importance of melt chemistry in controlling several aspects of element partitioning is discussed in light of the energetics of the partitioning process. The compositions of clinopyroxene and carbonate melt in our experiments closely match those of near-solidus melts and crystals in CMAS-CO₂ at 3 GPa, suggesting that our partition coefficients have direct relevance to melting of carbonated mantle lherzolite. Melts so produced will be characterised by elevated incompatible trace element concentrations, due to the low degrees of melting involved, but marked depletions of Ti and Zr, and fractionated REE patterns. These are common features of natural carbonatites. The different behaviour of trace elements in carbonate and silicate systems will lead to contrasted styles of trace element metasomatism in the mantle. Received: 15 July 1999 / Accepted: 18 February 2000     

Crystallization of melilite from CMAS-liquids and the formation of the melilite mantle of Type B1 CAIs: Experimental simulations

Type B CAIs are subdivided into B1s, with well-developed melilite mantles, and B2s, with randomly distributed melilite. Despite intensive study, the origin of the characteristic melilite mantle of the B1s remains unclear. Recently, we proposed that formation of the melilite mantle is caused by depletion of the droplet surface in volatile magnesium and silicon due to higher evaporation rates of volatile species compared to their slow diffusion rates in the melt, thus making possible crystallization of melilite at the edge of the CAI first, followed by its crystallization in the central parts at lower temperatures. Here, we present the results of an experimental study that aimed to reproduce the texture observed in natural Type B CAIs. First, we experimentally determined crystallization temperatures of melilite for three melt compositions, which, combined with literature data, allowed us to find a simple relationship between the melt composition, crystallization temperature, and composition of first crystallizing melilite. Second, we conducted a series of evaporation and cooling experiments exposing CAI-like melts to gas mixtures with different oxygen fugacities (f_O2). Cooling of the molten droplets in gases with logf_O2?IW-4 resulted in crystallization of randomly distributed melilite, while under more reducing conditions, melilite mantles have been formed. Chemical profiles through samples quenched right before melilite started to crystallize showed no chemical gradients in samples exposed to relatively oxidizing gases (logf_O2?IW-4), while the near-surface parts of the samples exposed to very reducing gases (logf_O2?IW-7) were depleted in volatile MgO and SiO₂, and enriched in refractory Al₂O₃. Using these experimental results and the fact that the evaporation rate of magnesium and silicon from CAI-like melts is proportional to , we estimate that Type B1 CAIs could be formed by evaporation of a partially molten precursor in a gas of solar composition with . Type B2 CAIs could form by slower evaporation of the same precursors in the same gas with .     

Redox equilibria of iron and silicate melt structure: Implications for olivine/melt element partitioning

Olivine/melt partitioning of ΣFe, Fe²⁺, Mg²⁺, Ca²⁺, Mn²⁺, Co²⁺, and Ni²⁺ has been determined in the systems CaO-MgO-FeO-Fe₂O₃-SiO₂ (FD) and CaO-MgO-FeO-Fe₂O₃-Al₂O₃-SiO₂ (FDA3) as a function of oxygen fugacity (f_O2) at 0.1 MPa pressure. Total iron oxide content of the starting materials was ∼20 wt%. The f_O2 was to used to control the Fe³⁺/ΣFe (ΣFe: total iron) of the melts. The Fe³⁺/ΣFe and structural roles of Fe²⁺ and Fe³⁺ were determined with ⁵⁷Fe resonant absorption Mössbauer spectroscopy. Changes in melt polymerization, NBO/T, as a function of f_O2 was estimated from the Mössbauer data and existing melt structure information. It varies by ∼100% in melts coexisting with olivine in the FDA3 system and by about 300% in the FD system in the Fe³⁺/ΣFe range of the experiments (0.805-0.092). The partition coefficients ( in olivine/wt% in melt) are systematic functions of f_O2 and, therefore, NBO/T of the melt. There is a -minimum in the FDA3 system at NBO/T-values corresponding to intermediate Fe³⁺/ΣFe (0.34-0.44). In the Al-free system, FD, where the NBO/T values of melts range between ∼1 and ∼2.9, the partition coefficients are positively correlated with NBO/T (decreasing Fe³⁺/ΣFe). These relationships are explained by consideration of solution behavior in the melts governed by Qⁿ-unit distribution and structural changes of the divalent cations in the melts (coordination number, complexing with Fe³⁺, and distortion of the polyhedra).     

Carbon solubility in core melts in a shallow magma ocean environment and distribution of carbon between the Earth’s core and the mantle

The solubility of carbon in Fe and Fe-5.2 wt.% Ni melts, saturated with graphite, determined by electron microprobe analysis of quenched metal melts was 5.8 ± 0.1 wt.% at 2000 °C, 6.7 ± 0.2 wt.% at 2200 °C, and 7.4 ± 0.2 wt.% at 2410 °C at 2 GPa, conditions relevant for core/mantle differentiation in a shallow magma ocean. These solubilities are slightly lower than low-pressure literature values and significantly beneath calculated values for even higher pressures [e.g., Wood B. J. (1993) Carbon in the core. Earth Planet. Sci. Lett.117, 593-607]. The trend of C solubility versus temperature for Fe-5.2 wt.% Ni melt, within analytical uncertainties, is similar to or slightly lower (∼0.2-0.4 wt.%) than that of pure Fe. Carbon content of core melts and residual mantle silicates derived from equilibrium batch or fractional segregation of core liquids and their comparison with our solubility data and carbon content estimate of the present day mantle, respectively, constrain the partition coefficient of carbon between silicate and metallic melts, in a magma ocean. For the entire range of possible bulk Earth carbon content from chondritic to subchondritic values, of 10⁻⁴ to 1 is derived. But for ∼1000 ppm bulk Earth carbon, is between 10⁻² and 1. Using the complete range of possible for a magma ocean at ∼2200 °C, we predict maximum carbon content of the Earth’s core to be ∼6-7 wt.% and a preferred value of 0.25 ± 0.15 wt.% for a bulk Earth carbon concentration of ∼1000 ppm.     

Trace element partitioning between ilmenite, armalcolite and anhydrous silicate melt: Implications for the formation of lunar high-Ti mare basalts

We performed a series of experiments at high pressures and temperatures to determine the partitioning of a wide range of trace elements between ilmenite (Ilm), armalcolite (Arm) and anhydrous lunar silicate melt, to constrain geochemical models of the formation of titanium-rich melts in the Moon. Experiments were performed in graphite-lined platinum capsules at pressures and temperatures ranging from 1.1 to 2.3 GPa and 1300-1400 °C using a synthetic Ti-enriched Apollo ‘black glass’ composition in the CaO-FeO-MgO-Al₂O₃-TiO₂-SiO₂ system. Ilmenite-melt and armalcolite-melt partition coefficients (D) show highly incompatible values for the rare earth elements (REE) with the light REE more incompatible compared to the heavy REE ( 0.0020 ± 0.0010 to 0.069 ± 0.010 for ilmenite; 0.0048 ± 0.0023 to 0.041 ± 0.008 for armalcolite). D values for the high field strength elements vary from highly incompatible for Th, U and to a lesser extent W (for ilmenite: 0.0013 ± 0.0008, 0.0035 ± 0.0015 and 0.039 ± 0.005, and for armalcolite 0.008 ± 0.003, 0.0048 ± 0.0022 and 0.062 ± 0.03), to mildly incompatible for Nb, Ta, Zr, and Hf (e.g. 0.28 ± 0.05 and : 0.76 ± 0.07). Both minerals fractionate the high field strength elements with D_Ta/D_Nb and D_Hf/D_Zr between 1.3 and 1.6 for ilmenite and 1.3 and 1.4 for armalcolite. Armalcolite is slightly more efficient at fractionating Hf from W during lunar magma ocean crystallisation, with D_Hf/D_W = 12-13 compared to 6.7-7.5 for ilmenite. The transition metals vary from mildly incompatible to compatible, with the highest compatibilities for Cr in ilmenite (D ∼ 7.5) and V in armalcolite (D ∼ 8.1). D values show no clear variation with pressure in the small range covered.Crystal lattice strain modelling of D values for di-, tri- and tetravalent trace elements shows that in ilmenite, divalent elements prefer to substitute for Fe while armalcolite data suggest REE replacing Mg. Tetravalent cations appear to preferentially substitute for Ti in both minerals, with the exception of Th and U that likely substitute for the larger Fe or Mg cations. Crystal lattice strain modelling is also used to identify and correct for very small (∼0.3 wt.%) melt contamination of trace element concentration determinations in crystals.Our results are used to model the Lu-Hf-Ti concentrations of lunar high-Ti mare basalts. The combination of their subchondritic Lu/Hf ratios and high TiO₂ contents requires preferential dissolution of ilmenite or armalcolite from late-stage, lunar magma ocean cumulates into low-Ti partial melts of deeper pyroxene-rich cumulates.     

The effect of sulfur on the partitioning of Ni and other first-row transition elements between olivine and silicate melt

The effect of sulfur dissolved as sulfide (S²⁻) in silicate melts on the activity coefficients of NiO and some other oxides of divalent cations (Ca, Cr, Mn, Fe and Co) has been determined from olivine/melt partitioning experiments at 1400 °C in six melt compositions in the system CaO-MgO-Al₂O₃-SiO₂ (CMAS), and in derivatives of these compositions at 1370 °C, obtained from the six CMAS compositions by substituting Fe for Mg (FeCMAS). Amounts of S²⁻ were varied from zero to sulfide saturation, reaching 4100 μg g⁻¹ S in the most sulfur-rich silicate melt. The sulfide solubilities compare reasonably well with those predicted from the parameterization of the sulfide capacity of silicate melts at 1400 °C of O’Neill and Mavrogenes (2002), although in detail systematic deviations indicate that a more sophisticated model may improve the prediction of sulfide capacities.The results show a barely discernible effect of S²⁻ in the silicate melt on Fe, Co and Ni partition coefficients, and also surprisingly, a tiny but resolvable effect on Ca partitioning, but no detectable effect on Cr, Mn or some other lithophile incompatible elements (Sc, Ti, V, Y, Zr and Hf). Decreasing Mg# of olivine (reflecting increasing FeO in the system) has a significant influence on the partitioning of several of the divalent cations, particularly Ca and Ni. We find a remarkably systematic correlation between and the ionic radius of M²⁺, where M = Ca, Cr, Mn, Fe, Co or Ni, which is attributable to a simple relationship between size mismatch and excess free energies of mixing in Mg-rich olivine solid solutions.Neither the effect of S²⁻ nor of Mg#^ol is large enough by an order of magnitude to account for the reported variations of obtained from electron microprobe analyses of olivine/glass pairs from mid-ocean ridge basalts (MORBs). Comparing these MORB glass analyses with the Ni-MgO systematics of MORB from other studies in the literature, which were obtained using a variety of analytical techniques, shows that these electron microprobe analyses are anomalous. We suggest that the reported variation of with S content in MORB is an analytical artifact.Mass balance of melt and olivine compositions with the starting compositions shows that dissolved S²⁻ depresses the olivine liquidus of haplobasaltic silicate melts by 5.8 × 10⁻³ (±1.3 × 10⁻³) K per μg g⁻¹ of S²⁻, which is negligible in most contexts. We also present data for the partitioning of some incompatible trace elements (Sc, Ti, Y, Zr and Hf) between olivine and melt. The data for Sc and Y confirm previous results showing that and decrease with increasing SiO₂ content of the melt. Values of average 0.01 with most falling in the range 0.005-0.015. Zr and Hf are considerably more incompatible than Ti in olivine, with and about 10⁻³. The ratio / is well constrained at 0.611 ± 0.016.     

Experimental study of the incorporation of Li, Sc, Al and other trace elements into olivine

We performed a series of synthesis experiments at 1 atm pressure to investigate the substitution mechanisms of 1+ and 3+ ions into olivine. Forsterite crystals were grown from bulk compositions that contained the element of interest (e.g. Li) and different amounts of additional single trace elements. By working at constant (major element) liquid composition and temperature we eliminated all compositional effects other than those due to the trace elements. Mineral-melt pairs were then analysed to determine the compositional-dependence of the partition coefficient (D), which corresponds to , and where [element] refers to weight concentration of the element in the respective phase.We find that Li forms a stable coupled substitution with Sc and, at above ∼500 ppm Sc in the crystal, Li⁺ and Sc³⁺ ions form an ordered neutral complex ([LiSc]). This complex dissociates at lower trace element concentrations and a second, concentration-independent, mechanism begins to dominate. This second solution mechanism is most likely 2Li⁺ ⇔ Mg²⁺ where one of the Li atoms is in an interstitial position in the crystal lattice. Natural olivines show Li contents slightly greater than Sc (on an atomic basis), indicating that both substitution mechanisms are significant. Unlike Sc, Al does not appear to form a stable complex with Li in the olivine structure.Sodium is highly incompatible in olivine with of ∼0.00015-0.03. Olivine-liquid partitioning of Na⁺ is independent of Sc³⁺ or Al³⁺ concentration. This indicates that the coupled substitution of Na⁺ with any 3+ ions is unlikely. Instead, the relevant substitution mechanism appears to be 2Na⁺ ⇔ Mg²⁺. Although independent of 3+ ion concentration, is inversely correlated with the Li concentration of both melts and crystals, implying that Na competes (unsuccessfully) with Li to replace Mg in the olivine structure.Aluminium is highly incompatible in forsterite . Values of are similar for all phase pairs synthesised from starting materials containing between 10 and 100,000 ppm Al. This suggests that Al is principally incorporated in forsterite by replacing one Mg and one Si atom , where the Al atoms on octahedral (Mg) and tetrahedral (Si) sites are dissociated from one another.The incorporation of gallium into forsterite is influenced by the presence of Li. Where Li concentration in the crystal is much greater than that of Ga (on an atomic basis) we find an excellent correlation between and melt Li content. This relationship indicates that Ga³⁺ and Li⁺ replace 2Mg²⁺ on octahedral sites and that the Ga and Li atoms are, like Sc and Li, strongly associated in the crystal structure.The mechanism by which scandium is incorporated into forsterite is strongly governed by the presence Li. As discussed above, ordered complexes form readily in forsterite in Li-rich experiments. Under Li-absent but Sc-rich conditions (Sc in the crystal >∼500 ppm), is proportional to the concentration of Sc in the melt. This indicates that Sc incorporation is charge-balanced by the formation of magnesium vacancies , and that both species are associated . At lower Sc concentrations (<500 ppm in the crystal), the concentration-dependence of partitioning indicates that the complexes dissociate.Our results demonstrate that partitioning of 1+ and 3+ ions into olivine is complex and involves a range of point defects which yield strongly composition-dependent crystal-melt partition coefficients. Since physical and chemical properties of natural olivine, such as diffusion of ⁶Li and ⁷Li and H₂O solubility, depend on the concentrations of the defects identified in this study, our results provide an important insight into how determining substitution mechanisms can improve our understanding of large-scale mantle processes and properties.     

Trace element partitioning and substitution mechanisms in calcium perovskites

We have determined the partition coefficients of a large number of trace elements between CaTiO₃ perovskite and anhydrous silicate melts at atmospheric pressure and 3 GPa. Determination of the concentration limits of Henrys law behaviour in the CaO-Al₂O₃–SiO₂–TiO₂ system reveals that the incorporation of rare earth elements (REE) and tetravalent large ion lithophile elements (LILE⁴⁺ such as U and Th) at the Ca-site of CaTiO₃ perovskite occurs with charge compensation through Ca-vacancy formation rather than by coupled substitution of Al for Ti. When melt composition is varied, we find that partition coefficients for REE and Th are strong functions of the CaO content of the melt. The observed trends are in excellent agreement with those predicted from the Ca-vacancy model. Given that they adopt the same crystal structure and have similar trace element partitioning behaviour, CaTiO₃ perovskite and the deep mantle phase CaSiO₃ perovskite can be considered analogous to one another. When the analogy is pursued in detail, we find that partitioning into both phases follows the composition-dependence predicted by the Ca-vacancy model. Thus, substitution of REE, U⁴⁺ and Th into CaSiO₃ in the lower mantle also occurs with Ca-vacancy formation to balance charge. Furthermore when 2+, 3+ and 4+ partition coefficients for both phases are plotted as functions of CaO melt content, the trends for CaSiO₃ and CaTiO₃ appear to be continuous. This surprising result means that partitioning into Ca-perovskite is independent of pressure and temperature and also of whether or not the host is CaSiO₃ or CaTiO₃. One implication is that CaSiO₃ crystallising from a peridotitic magma ocean may have partition coefficients for Th and U up to about 400. Crystallisation and sequestration of as little as 0.25 volume% of this phase in the lower mantle early in earth history would make a significant contribution to current mantle heat production.Electronic Supplementary Material Supplementary material is available for this article at     

Phase relations of carbonate-bearing eclogite assemblages from 2.5 to 5.5 GPa: implications for petrogenesis of carbonatites

We have experimentally investigated the phase and melting relations of garnet + clinopyroxene + carbonate assemblages at 2.5–5.5 GPa, to assess the feasibility of carbonated eclogite as a source for some crustally emplaced carbonatites. The solidus of our composition was at 1,125 °C at 2.5 GPa, 1,225 °C at 3.5 GPa and 1,310 °C at 5.0 GPa. Melts were sodic calcio-dolomitic carbonatites, and were markedly more calcic than the dolomitic melts produced by partial melting of carbonated peridotite. Na contents of the experimental carbonatites decreased with increasing pressure when compared at similar degrees of melting, and SiO₂ contents increased with degree of melting. Experiments on a second composition with enhanced Na₂O demonstrated its strong effect in lowering melting temperatures in carbonate eclogite. Natural carbonated eclogite bodies in the peridotitic upper mantle will have a range of solidus temperatures. In many cases, carbonate will be molten in the upper 250 km. Carbonate melt would segregate from its source eclogite at very low melt fractions and infiltrate surrounding peridotitic wall rock. This would result in metasomatic enrichment of the peridotitic wall rock, but its exact nature will depend on the relative P–T positions of the eclogite + CO₂ and peridotite + CO₂ solidii. As a result of these inevitable metasomatic interactions, it is considered unlikely that carbonatite melts derived from carbonated eclogite in the upper mantle could be emplaced into the crust unmodified. However, they may have a role in metasomatically enriching and carbonating parts of the upper mantle, producing sources suitable for subsequent production of silica undersaturated silicate liquids and carbonatites ultimately emplaced in the crust.Editorial responsibility: J. Hoefs     

Trace element geochemistry of Amba Dongar carbonatite complex,India: Evidence for fractional crystallization and silicate-carbonate melt immiscibility

Carbonatites are believed to have crystallized either from mantle-derived primary carbonate magmas or from secondary melts derived from carbonated silicate magmas through liquid immiscibility or from residual melts of fractional crystallization of silicate magmas. Although the observed coexistence of carbonatites and alkaline silicate rocks in most complexes, their coeval emplacement in many, and overlapping initial⁸⁷Sr/⁸⁶Sr and¹⁴³Nd/¹⁴⁴Nd ratios are supportive of their cogenesis; there have been few efforts to devise a quantitative method to identify the magmatic processes. In the present study we have made an attempt to accomplish this by modeling the trace element contents of carbonatites and coeval alkaline silicate rocks of Amba Dongar complex, India. Trace element data suggest that the carbonatites and alkaline silicate rocks of this complex are products of fractional crystallization of two separate parental melts. Using the available silicate melt-carbonate melt partition coefficients for various trace elements, and the observed data from carbonatites, we have tried to simulate trace element distribution pattern for the parental silicate melt. The results of the modeling not only support the hypothesis of silicate-carbonate melt immiscibility for the evolution of Amba Dongar but also establish a procedure to test the above hypothesis in such complexes.     

Copper partitioning in a melt-vapor-brine-magnetite-pyrrhotite assemblage

The effect of sulfur on the partitioning of Cu in a melt-vapor-brine ± magnetite ± pyrrhotite assemblage has been quantified at 800 °C, 140 MPa, f_O2 = nickel-nickel oxide (NNO), logf_S2=-3.0 (i.e., on the magnetite-pyrrhotite curve at NNO), logf_H2S=-1.3 and logf_SO2=-1. All experiments were vapor + brine saturated. Vapor and brine fluid inclusions were trapped in silicate glass and self-healed quartz fractures. Vapor and brine are dominated by NaCl, KCl and HCl in the S-free runs and NaCl, KCl and FeCl₂ in S-bearing runs. Pyrrhotite served as the source of sulfur in S-bearing experiments. The composition of fluid inclusions, glass and crystals were quantified by laser-ablation inductively coupled plasma mass spectrometry. Major element, chlorine and sulfur concentrations in glass were quantified by using electron probe microanalysis. Calculated Nernst-type partition coefficients (±2σ) for Cu between melt-vapor, melt-brine and vapor-brine are , , and , respectively, in the S-free system. The partition coefficients (±2σ) for Cu between melt-vapor, melt-brine and vapor-brine are , , and , respectively, in the S-bearing system. Apparent equilibrium constants (±1σ) describing Cu and Na exchange between vapor and melt and brine and melt were also calculated. The values of are 34 ± 21 and 128 ± 29 in the S-free and S-bearing runs, respectively. The values of are 33 ± 22 and60 ± 5 in the S-free and S-bearing runs, respectively. The data presented here indicate that the presence of sulfur increases the mass transfer of Cu into vapor from silicate melt. Further, the nearly threefold increase in suggests that Cu may be transported as both a chloride and sulfide complex in magmatic vapor, in agreement with hypotheses based on data from natural systems. Most significantly, the data demonstrate that the presence of sulfur enhances the partitioning of Cu from melt into magmatic volatile phases.     

Activities of NiO, FeO, and O in silicate melts

We have measured activity coefficients for NiO and FeO in a variety of silicate melts (SiO₂-CaO-MgO-Al₂O₃) using electrochemical methods similar to square wave voltametry. We report the activity of the oxide ion (a_O²⁻) in one composition. Based on these measurements, we have constructed a model that predicts the variations in activity we observe, and also variations in NiO activity reported in the literature. Activity of metal-oxide components such as NiO and FeO in silicate melts can be understood by considering contributions from both the activity of the oxide ion and the activity of the cation through expressions of the type: