In the system Na2O-Al2O3-SiO2-H2O-TiO2, the behavior of Ti-containing structural complexes has been determined in H2O-saturated silicate melts and in coexisting silicate-saturated aqueous fluids as well as in silicate-rich supercritical fluids to 900 degrees C and 2225 MPa. Titanium speciation in aqueous fluids in the system TiO2-H2O was also characterized. All measurements were carried out in situ at the desired temperature and pressure using confocal microRaman and microFTIR spectroscopy. The experiments were carried out in an Ir-gasketed hydrothermal diamond-anvil cell (HDAC) with K-type thermocouples for temperature measurement and the Raman shift of C-13 synthetic diamond to monitor pressure.

Structural characterization of silicate melts and aqueous fluids equilibrated at pressures and temperatures corresponding to the Earth's interior requires measurements in-situ while the samples are at the pressure and temperature of interest. To this end, structure and structure-property relations of melts and coexisting fluids in silicate-COH systems have been determined at temperatures up to 1000 degrees C and at pressures to similar to 2.0 GPa.

Hydrogen isotope fractionation between water-saturated silicate melt and silicate-saturated aqueous fluid has been determined experimentally by using vibrational spectroscopy as the analytical tool. The measurements were conducted in situ with samples at the high temperature and pressure of interest in an externally heated diamond cell in the 450-800 degrees C and 101-1567 MPa temperature and pressure range, respectively. The starting materials were glass of Na-silicate with Na/Si = 0.5 (NS4), an aluminosilicate composition with 10 mol% Al2O3 and Na/(Al+Si) = 0.5 (NA10), and a 50:50 (by volume) H2O:D2O fluid mixture. Platinum metal was used to enhance equilibration rate. Isotopic equilibrium was ascertained by using variable experimental duration at given temperature and pressure.

The C-O-H-N solubility and solution mechanisms in silicate melts and C-O-H-N speciation in coexisting fluid to upper mantle temperatures and pressures and with redox conditions from the MH to the IW buffer are discussed. Focus is on in-situ structural characterization of coexisting melt and fluid. In fluid + melt-CON, fluid + melt-NOH, and fluid + melt-OH systems, volatiles are dissolved in molecular form (CO2, CH4, NH3, N-2, H2O, H-2) and as complexes that form chemical bonding with the silicate network (CO3, CH3, NH2, OH). In silicate-OH systems molecular H2O (H2O) and OH-groups exist in silicate- and aluminosilicate-saturated fluids and coexisting water-saturated melts above similar to 400 degrees C and similar to 0.5 GPa with their OH/H2O-ratio positively correlated with temperature. The extent of hydrogen bonding in both fluids and melts diminishes with temperature so that above similar to 400 degrees C it cannot be detected. The Delta H of hydrogen bonding in aqueous fluid (22 +/- 1 kJ/mol) is about twice that in silicate melts (10 2 kJ/mol). Silicate speciation in silicate-saturated fluid and hydrous silicate melts comprises similar Q-species with Delta H of the solution reactions in silicate-saturated fluid, water-saturated melt, and supercritical fluid similar to 400 kJ/mol. In COH-silicate systems methane solubility in melt increases from 0.2 wt% to similar to 0.5 wt.% in the melt NBO/Si range from 0.4 to 1.0 at 1-2.5 GPa and 1400 C. The solubility increases by similar to 150% between the redox conditions of the IW and MH buffers. At the NNO buffer conditions and more oxidizing, carbon exists as carbonate complexes in melts and as CO2 in fluid. Reduced (C + H)-bearing species in melts (CH3-groups and molecular CH4) are stable at f(H2)(MW) and more reducing conditions, whereas the species in coexisting fluid are CH4, H-2, and H2O. In NOH-silicate systems, the N solubility in melt decreases from 0.98 to 0.28 wt.% in the melt NBO/Si-range from 0.4 to 1.18 at the redox conditions of the IW buffer. The solubility decreases by about 50% between the redox conditions of the IW and MH buffers. At IW, nitrogen occurs in silicate melts amine groups, NH2, bonded to the silicate network, and as molecular NH3, whereas in coexisting NOH fluids the dominant species are NH3, N-2, H-2 and H2O. The NH2-/NH3 abundance ratio varies by similar to 55 between melt compositions with NBO/Si = 1.18 and 0.4. In fluids and melts, decreasing hydrogen fugacity leads to oxidation of nitrogen to form molecular N-2 so that at the MH redox conditions, the dominant N-bearing species is N-2. The redox-dependent solution mechanisms of COHN volatile components in silicate melts affect their structure differently, which results in redox-dependent thermodynamic and transport properties of magmatic liquids in the interior of the Earth and terrestrial planets. These properties include mineral/melt minor and trace element partitioning, melt/fluid isotope fractionation, and transport and thermodynamic properties of melt saturated with variably-oxidized COHN volatile components. 2012 Elsevier B.V. All rights reserved.

Hydrogen isotope partitioning (as H2O and D2O) between silicate-saturated aqueous fluid and water-saturated aluminosilicate melt has been determined with vibrational spectroscopy (Raman and infrared) in situ with the samples at high temperature and pressure by using a hydrothermal diamond-anvil cell (HDAC) for sample containment. To assess the effects of pressure and, therefore, different silicate speciation in fluids and melts, on the D/H partitioning behavior, two pressure/temperature experimental trajectories (450-800 degrees C1155-754 MPa, and 450-800 degrees C/562-1271 MPa) were used. In these temperature and pressure ranges, the fluid/melt partition coefficients are temperature (and pressure) dependent with the average enthalpy change, Delta H = -6.6 +/- 15 kJ/mol and -10.3 +/- 1.1 kJ/mol for H2O and D2O, respectively. The Delta H-values for the lower-pressure trajectory (and, therefore, lower fluid density) were 15-20% higher than for the higher-pressure (and higher fluid density) trajectory. The (D/H) ratios of fluids and melts, (D/H)(fluid) and (D/H)(melt), are also temperature dependent with a small negative All for (D/H)(fluid) (average: -2.4 +/- 0.8 kJ/mol) and a positive Delta H-value for (D/H)(melt) (2.3 +/- 1.4 kJ/mol). The (D,H) exchange equilibrium between fluid and melt is also temperature (and pressure) dependent so that for the low-density P/T trajectory, the Delta H = -4.2 +/- 0.6 kJ/mol, whereas for the higher-density trajectory, Delta H = -5.4 +/- 0.7 kJ/mol. The difference between the H2O and D2O fluid/melt partition coefficients and the temperature- and pressure-dependent D/H fractionation behavior in and between hydrous silicate melts and silicate-saturated aqueous fluid in part is because pressure increases with increasing temperature in the HDAC experiments and the volume difference between fluid and melt differ for H2O and D2O. In addition, the silicate speciation in fluids and melts are temperature and pressure dependent, which also leads to significantly temperature- and pressure-dependent D/H fractionation within and between silicate melts and fluids at high temperature and pressure. In the Earth's deep crust and upper mantle, hydrogen isotope partitioning between condensed phases and aqueous fluid can differ substantially from that between condensed phases and pure H2O because the aqueous fluid in the Earth's interior is a concentrated silicate solution wherein the silicate speciation affects the isotope partitioning.

Raman spectroscopy of glasses and melts is a powerful technique which can provide information about the silicate network connectivity via the study of Q(n)-species distribution and also on other tetrahedrally coordinated cations such as aluminium, phosphorus, ferric iron and titanium. Recently, Raman spectroscopy of glasses and melts has focused on information obtained about the speciation of volatile compounds. After a brief introduction to the definition of the glassy state and on the relationships between glass and melt, the main steps of Raman spectral analysis are described, discussing the main approaches and the extent to which this method is quantitative. Finally, examples of the application of Raman spectroscopy in the field of Earth sciences are given.

Speciation of D-O-H-C-N volatiles in alkali aluminosilicate melts and of silicate in D-O-H-C-N fluid has been determined in situ to 800 degrees C and >2 GPa under reducing and oxidizing conditions by using an externally heated hydrothermal diamond-anvil cell with Raman spectroscopy as the structural probe. Reducing conditions were near those of the IW oxygen buffer, whereas oxidizing conditions were obtained by conducting the experiments with oxidized components only and with Pt as a catalyst. Raman bands assigned to C-H stretching in CHxDy isotopologues and CH4 groups (including CH3) were employed to determine the CH4/CHxDy ratio in fluids and melts. This ratio decreases from 1.5-2 at 500 degrees C to between 1.2 and 1 with 800 degrees C with Delta H-values of 13.6 +/- 2.1 and 5.5 +/- 1.1 kJ/mol for melt and fluid, respectively. The CHx/CHxDy fluid/melt partition coefficient ranges between similar to 16 and similar to 3 with Delta H= 33 +/- 6 kJ/mol assuming no pressure effect. This behavior of deuterated and protonated complexes is ascribed to speciation of volatile and silicate components in fluids and melts in a manner that is conceptually similar to D/H partitioning among complexes and phases in brines and hydrous silicate systems.

Nitrogen is the dominant gas in Earth atmosphere, but its behavior during the Earth' differentiation is poorly known. To aid in identifying the main reservoirs of nitrogen in the Earth, nitrogen solubility was determined experimentally in a mixture of molten CI-chondrite model composition and (Fe, Ni) metal alloy liquid. Experiments were performed in a laser-heated diamond-anvil cell at pressures to 18 GPa and temperatures to 2850 +/- 200 K. Multi-anvil experiments were also performed at 5 and 10 GPa and 2390 +/- 50 K. The nitrogen content increases with pressure in both metal and silicate reservoirs. It also increases with the iron content of the (Fe, Ni) alloy. Sieverts' formalism successfully describes the nitrogen solubility in metals up to 18 GPa. Henry's law applies to nitrogen-saturated silicate melts up to 4-5 GPa. Independently of these solubility models, it is shown that the partition coefficient of nitrogen between metal and silicate melts changes from almost 10(4) at ambient pressure to about 10-20 for pressures higher than 1 GPa. The pressure-dependence of the nitrogen partitioning can explain the recently suggested depletion of nitrogen relative to other volatiles in the bulk silicate Earth. (c) 2013 Elsevier Ltd. All rights reserved.

Hydrogen-rich and water-rich fluids exert different control on dissolution mechanism of oxide and silicate minerals in the Earth's interior. With Mg-silicate-H-2 fluids, dissolution tends to be incongruent with the (Mg/Si)(fluid) < (Mg/Si)(Mg-silicate) with formation of SiOH and SiH4 complexes in the fluid (Shinozaki et al. 2013, 2014). In contrast, in Mg-silicate-H2O systems, Mg-silicate minerals in the mantle (pyroxene and forsterite) dissolve stoichiometrically (congruently) in aqueous fluids to at least 10 GPa pressure. Metasomatic alteration by H-2-rich fluids enriches, therefore, the mantle in SiO2 compared with alteration by H2O fluid. This difference becomes increasingly important with mantle depth because the environment becomes more reducing, which results in an increase of H-2/H2O fluids (Shinozaki et al. 2014, this issue). Chemical gradients with depth of the Earth could be affected by increased H-2/H2O of mantle fluids whereby Mg/Si ratios, for example, will become variable. Silicate-H-2 alteration processes likely also played major roles during the early, core-forming stages of the Earth. Such a process could be responsible for Mg/Si changes in the early silicate Earth.

Experiments to determine silicate structural species in silicate-saturated aqueous fluids in equilibrium with silica polymorphs (quartz and coesite), enstatite, and enstatite+forsterite in the SiO2-H2O and MgO-SiO2-H2O systems have been carried out in situ in the 0.4-5.4GPa and 700-900 degrees C pressure and temperature ranges, respectively. MicroRaman spectroscopy was the structural probe. In the SiO2-H2O system (1.6-5.4GPa/700-900 degrees C), the detected silicate species are Q(0) (SiO44-), Q(1) (0.5 Si2O76-), and Q(2) (SiO32-). The expression 2Q(1)Q(0)+Q(2) describes the equilibrium among these species with H and V values from the isochoric temperature and isothermal pressure dependence of its equilibrium constant, K=XQ0XQ2/(X-Q1)(2), range from -23 to -69kJ/mol and -1 to -2cm(3)/mol, respectively. In the system MgO-SiO2-H2O the calculated silica solubility, using literature algorithms, is approximately 50% of that in the SiO2-H2O system at similar temperature and pressure. Only Q(1) and Q(0) species were detected in the MgO-SiO2-H2O fluids, whether in equilibrium with enstatite+forsterite (P<3GPa) or enstatite only (P>3GPa). The temperature and pressure dependence of the equilibrium constant, K=X-Q1/X-Q0, for this system yields average values of H=405kJ/mol and V=-2.30.4cm(3)/mol. The speciation of silicate in aqueous fluids resembles that in hydrous melts as a function of temperature and pressure at deep crustal and upper mantle temperature and pressure conditions, and they become increasingly similar with depth. As the silicate speciation and solubility in the aqueous fluid depend on silicate composition, the pressure and temperature at which complete miscibility occurs will also vary with silicate composition. The structural similarity between fluids and melts will also lead to fluid/melt element partition coefficients trending toward 1 and mineral/fluid partition coefficients trending toward mineral/melt values in the upper mantle as the silicate-H2O systems approach complete miscibility with increasing temperature and pressure.