H. Bureau

Complete miscibility between silicate melts and hydrous fluids in the upper mantle: experimental evidence and geochemical implications

The phase relationships between silicate melts and hydrous fluids were studied by direct visual observation in an externally heated diamond-anvil cell. Complete miscibility of silicate melt and water was observed for a wide range of melt compositions, including nepheline, jadeite, dacite, haplogranite and Ca-bearing granite. Some evidence for complete miscibility was also observed in the system basalt–H2O. The critical temperatures in all systems decrease rapidly with pressure. At 15 kbar, the critical temperature for nepheline is around 550°C, for jadeite around 800°C and for granitic compositions it is close to 900°C. In general, the critical temperatures appear to increase with silica content in the system. Our results suggest that there is complete miscibility between silicate melts and water in most of the upper mantle, except at very shallow depths. This means that a water-saturated solidus cannot be defined any more in the deeper parts of the upper mantle. Very silica-rich melt inclusions found in spinel lherzolites associated with fluid inclusion are probably the result of the unmixing of a supercritical fluid containing comparable amounts of water and silicate components. The decomposition of amphibole in a subducted slab occurs at conditions where the miscibility gap between fluid and silicate melt is not yet closed, while the decomposition of phengite and lawsonite occurs far beyond the critical curve. Accordingly, the fluids released by the breakdown of these minerals should have very different properties. Highly mobile, hydrous fluids containing little dissolved silicate should be produced by amphibole breakdown, while the decomposition of lawsonite and phengite will lead to much more silicate-rich and less mobile fluid phases.

Melting of Peridotite to 140 Gigapascals

Under Pressure In order to understand the behavior of materials in the solid deep Earth, it is important to be able to estimate how a material melts at high pressure. To this end, Fiquet et al. (p. 1516) performed experiments using a laser-heated diamond anvil cell coupled to in situ synchrotron measurements of peridotite rock—a mixture of minerals thought to represent Earths upper mantle—across a wide pressure range. The results suggest that liquid phases may exist at very high pressure values, such that seismically anomalous zones near the boundary between the core and the mantle may result from isolated pockets of melt. Along similar lines, the base of primitive Earths mantle may have acquired its trace element signature from partial melting of certain mineral phases higher up in the mantle. High-temperature and -pressure experiments reveal details about how and where the mantle melts. Interrogating physical processes that occur within the lowermost mantle is a key to understanding Earth’s evolution and present-day inner composition. Among such processes, partial melting has been proposed to explain mantle regions with ultralow seismic velocities near the core-mantle boundary, but experimental validation at the appropriate temperature and pressure regimes remains challenging. Using laser-heated diamond anvil cells, we constructed the solidus curve of a natural fertile peridotite between 36 and 140 gigapascals. Melting at core-mantle boundary pressures occurs at 4180 ± 150 kelvin, which is a value that matches estimated mantle geotherms. Molten regions may therefore exist at the base of the present-day mantle. Melting phase relations and element partitioning data also show that these liquids could host many incompatible elements at the base of the mantle.

Volcanic degassing of bromine and iodine: experimental fluid/melt partitioning data and applications to stratospheric chemistry

Abstract In order to understand the degassing behavior of heavy halogens in volcanic processes, we experimentally studied the distribution of Cl, Br, and I between albite melt and hydrous fluids containing 0.01–2 wt% of NaCl, NaBr, or NaI, respectively. All experiments were carried out at 2 kbar and 900°C in rapid-quench cold-seal autoclaves with a run duration of 7 days. The major element compositions and Cl contents of the glassy run products were measured by electron microprobe. Bromine and iodine were measured by proton-induced X-ray emission. Fluid compositions were obtained by mass balance. All halogens investigated were found to partition strongly into the fluid phase. Over the range of concentrations studied, the halogen contents in the melt are proportional to the concentrations in the fluid. The fluid/melt partition coefficients, Df/m, are 8.1±0.2 for Cl, 17.5±0.6 for Br, and 104±7 for I. The logarithm of Df/m is linearly correlated with the ionic radius of the halogenide ion. On the basis of our experimental data, we estimate the amount of bromine injected into the stratosphere by major volcanic explosions. For the 1991 Mount Pinatubo eruption, we obtain Br yields of 11–25 kt as minimum estimates. These numbers are comparable to the total annual influx of bromine into the stratosphere from all other natural and anthropogenic sources (about 100 kt/year). Since bromine is much more efficient in destroying stratospheric ozone than chlorine, it could at least be partially responsible for the massive ozone depletion observed after the 1991 Mount Pinatubo eruption.

Fluid‐magma decoupling in a hot‐spot volcano

Melt and fluid inclusions in olivines from the last 1998 eruption of Piton de la Fournaise (PdF), have recorded a range of volatile partial pressures (350–420 MPa), the highest, so far for this volcano, are found in Fo-rich olivines from a vent remote from the main eruptive activity. Such pressures indicate that fractionation of olivine (and other crystals) occur below the crust-mantle boundary in a CO2-rich volatile saturated environment. Together with previously published data, the model that emerges for PdF is one where olivines which have crystallized from multiple past magma injection events [see Albarede et al., 1997], may be picked up by newly intruding magma, anywhere from upper-mantle depths to the surface. This model is likely to apply to other shield volcanoes. Magma production and transport to the surface is accompanied by continuous open-system degassing through the permeable volcanic pile. The calculated H2O content of primary basalts (MgO ≈ 12–14 wt.%) from PdF may reach 0.7–1 wt.% implying a rather H2O-rich hot spot mantle source.

Experimental investigation of the stability of Fe-rich carbonates in the lower mantle

The fate of carbonates in the Earths mantle plays a key role in the geodynamical carbon cycle. Although iron is a major component of the Earths lower mantle, the stability of Fe-bearing carbonates has rarely been studied. Here we present experimental results on the stability of Fe-rich carbonates at pressures ranging from 40 to 105 GPa and temperatures of 1450-3600 K, corresponding to depths within the Earths lower mantle of about 1000-2400 km. Samples of iron oxides and iron-magnesium oxides were loaded into CO2 gas and laser heated in a diamond-anvil cell. The nature of crystalline run products was determined in situ by X-ray diffraction, and the recovered samples were studied by analytical transmission electron microscopy and scanning transmission X-ray microscopy. We show that Fe-(II) is systematically involved in redox reactions with CO2 yielding to Fe-(III)-bearing phases and diamonds. We also report a new Fe-(III)-bearing high-pressure phase resulting from the transformation of FeCO3 at pressures exceeding 40 GPa. The presence of both diamonds and an oxidized C-bearing phase suggests that oxidized and reduced forms of carbon might coexist in the deep mantle. Finally, the observed reactions potentially provide a new mechanism for diamond formation at great depth.

Magma–conduit interaction at Piton de la Fournaise volcano (Réunion Island): a melt and fluid inclusion study

Major-element, Cl, S, F analyses have been performed on a wide selection of melt inclusions trapped in olivine (Fo81–87) from scoria and crystal-rich lapilli samples of Piton de la Fournaise volcano. As a whole, they display a transitional basaltic composition. The melt inclusions (8–9 wt.% MgO, 0.62–0.73 wt.% K2O) are in equilibrium with olivines (Fo81–85) in the samples from the Central Feeding Zone and the South-East Feeding Zone and show a slight alkaline affinity. The melt inclusions in olivines (Fo85–87) from the North-West Rift (NWR) contain 9.3–9.7 wt.% MgO and 0.54–0.58 wt.% K2O, with a more tholeiitic tendency. In oceanitic lavas and crystal-rich lapilli, the olivine xenocrysts are recognisable by the presence of one or more secondary shear plane fracture(s) filled up with CO2 and alkali-rich basaltic melt inclusions. In dunite nodules, olivines present also contain several secondary shear plane fracture(s) filled up with CO2 and high-SiO2 melt inclusions. Secondary CO2-rich fluid inclusions in olivine (Fo85–87) from the NWR samples indicate PCO2 up to 500 MPa whereas, PCO2 ranges from 95 MPa to few tenths of bars in the other samples. Both the primary melt inclusions and the secondary fluid inclusions strongly suggest that the olivine crystallises and accumulates over a wide depth range (15 km). It is envisioned that cumulative pockets with low residual porosity are repeatedly percolated with a CO2-rich fluid phase, possibly associated with basaltic to SiO2-rich melts, and are finally disrupted and entrained to the surface when vigorous magma transfer occurs. The SiO2-rich residual melts in early-formed dunitic or gabbroic bodies may have acted as contaminant agents for the more alkali character of magmas vented through the central feeding system, where a well-developed cumulative system is thought to exist. Finally, the existence of secondary fluid and melt inclusions in olivines implies that the dunitic bodies are weakened on the micrometric scale.

In situ bubble vesiculation in silicic magmas

Abstract Volatile degassing is a major process driving volcanic eruptions. Therefore, a full understanding of mechanisms ranging from bubble nucleation, growth, coalescence, to magma fragmentation is required. We have simulated magma degassing during ascent in the volcanic conduit by depressurizing hydrated haplogranite melts in high-pressure and high-temperature optical cells (a hydrothermal diamond-anvil cell and an internally heated pressure vessel fitted with sapphire windows). This allowed the whole process of bubble nucleation, growth, and coalescence to be directly observed in situ through images captured from the recording videos. Bubble nucleation pressures, number densities, growth laws, and characteristics of coalescence were estimated as a function of melt water content, decompression rate, and temperature. Melt/vapor surface tension during bubble nucleation and coalescence was calculated. Our data show good agreement with those previously obtained in classical vessels. Methodological improvements are proposed for the experimental simulation of magma degassing in volcanic conduits.

In situ high-pressure and high-temperature bubble growth in silicic melts

We present the first investigation of in situ high-pressure and high-temperature bubble growth in silicic melts. In a hydrothermal diamond-anvil cell, a haplogranite melt (79 wt% SiO2) is hydrated then subjected to cooling and decompression. With decreasing pressure, water exsolves from the melt and bubbles grow. The whole experiment is observed through an optical microscope and video-recorded, so that bubble nucleation, bubble growth, and the glass transition are directly monitored. Bubbles nucleate and expand in melt globules having radii from 15 to 70 μm. Bubbles reached 3.6–9.1 μm in radius within 6.1–11.7 s (until the glass transition is attained) while temperature decreases from 709–879°C to 482–524°C, corresponding to decompressions from 7.0–21.9 to 3.4–15.2 kbar. Bubbles nucleated either in a single event occurring within the first second or in successive pulses over a period of up to 7 s when the melt globules are in contact with a diamond culet of the cell. In these experiments, bubble growth can be fitted to the cube root or a logarithm of time, mainly ascribable to the combination of large water oversaturations due to rapid cooling and decompression. At pressures of 3.4–15.2 kbar, we measure glass transition temperatures that are 20–80°C higher than those calculated at atmospheric pressure.

Determination of the concentration of water dissolved in glasses and minerals using nuclear microprobe

In Earth Sciences, the global water cycle is of fundamental importance. For this reason, the H2O content of volcanic glass and mantle minerals must be analysed: usually by micro-infrared spectroscopy (FTIR) or secondary ion mass spectrometry (SIMS). However, both of these methods require calibration using standards of known water content. To avoid matrix effects, the standards and unknowns must be otherwise identical in composition. In this study we have determined the water content of geological samples, in the range 10 ppm–5wt.%H2O, using an absolute analytical technique: a combination of elastic recoil detection analysis (ERDA) and Rutherford backscattering spectrometry (RBS). We compared the results obtained by this method to data obtained by FTIR on the same samples. We discuss the limitations of the method and use the results to calibrate IR extinction coefficients for FTIR spectroscopy.

In situ mapping of high-pressure fluids using hydrothermal diamond anvil cells

We present new results combining high pressures and temperatures attainable in a diamond anvil cell with in situ synchrotron radiation induced micro-X-ray fluorescence measurements. Hydrothermal diamond anvil cells experiments have been performed by measuring the partitioning of Pb between aqueous fluids (pure water or NaCl-enriched water) and hydrous silicate melts of haplogranite composition using synchrotron X-ray fluorescence. The in situ measurements were performed in the range 0.3–1.2 GPa and 730–850 °C both in the aqueous fluid and in the silicate melts being in equilibrium. Pb is strongly partitioned into high-pressure–temperature hydrous melts when Cl is present in either the hydrous melt or the aqueous fluid. Moreover, our comparisons of in situ results with post-mortem results show that significant changes take place during rapid quenching especially when samples are small (few hundred of microns in diameter). Water exsolution is induced by the quench in the silicate melt showing the high mobility of Pb which immediately partitions into the water vapor phase during the quench. The current in situ approach offers thus a pertinent complementary method to the classical experimental petrology investigations.