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.     

Temperature dependence of Pt and Rh solubilities in a haplobasaltic melt

The temperature dependence of the solubilities of Pt and Rh in a haplobasaltic (anorthite-diopside 1-bar eutectic composition) melt has been investigated at 1 bar and 1300 to 1550°C using the mechanically assisted equilibration technique (Dingwell et al., 1994). The experiments were performed at almost constant oxygen fugacity (log fO₂ = −2.5 ± 0.3) over the entire temperature range. Major element concentrations in the quenched glass samples were determined using an electron microprobe. Pt and Rh concentrations were obtained by laser ablation inductive coupled plasma mass spectrometry. From our data, we obtain the following expressions for the solubilities of pure Pt and pure Rh in anorthite-diopside eutectic melt at 1 bar and log fO₂ = −2.5:

     

Temperature independent thermal expansivities of sodium aluminosilicate melts between 713 and 1835 K

The thermal expansivities of eight sodium aluminosilicate liquids were derived from the slope of new volume data at low temperatures (713−1072 K) combined with the high temperature (1300−1835 K) volume measurements of Stein et al. (1986) on the same liquids. Melt compositions range from 47−71 wt% SiO₂, 0−31 wt% A1203, and 17−33 wt% Na₂O; the volume of albite supercooled liquid at 1092 K was also determined. The low temperature volumes were derived from measurements of the glass density of each sample at 298 K, followed by measurements of the glass thermal expansion coefficient from 298 K to the respective glass transition interval. This technique takes advantage of the fact that the volume of a glass is equal to the volume of the corresponding liquid at the limiting fictive temperature (T′_f), and that T′_f can be approximated as the onset of the rapid rise in thermal expansion at the glass transition in a heating curve (Moynihan, 1995). No assumptions were made regarding the equivalence of enthalpy and volume relaxation through the glass transition. The propagated error on the volume of each supercooled liquid at T′_f is 0.25%. Combination of these low temperature data with the high temperature measurements of Stein et al. (1986) allowed a constant thermal expansivity of each liquid to be derived over a wide temperature interval (763−1001 degrees) with a fitted 1σ error of 0.6–4.6%; in every case, no temperature dependence to dV/dT^liq could be resolved. Calibration of a linear model equation leads to fitted values ± 1σ (units of cm³/mole) for (26.91 ± .04), (37.49 ± .12), (26.48 ± .06) at 1373 K, and (7.64 ± .08 × 10^-3 cm³/mole-K). The results indicate that neither Si02 nor Al₂O₃ contribute to the thermal expansivity of the liquids, and that dV/dT^liq is independent of temperature between 713–1835 K over a wide range of liquid composition. Calculated volumes based on this model recover both low and high temperature measurements with a standard deviation <0.25%, whereas values of dV/dT^liq can be predicted within 5.6%.     

The effect of Cl on Pt solubility in haplobasaltic melt: Implications for micronugget formation and evidence for fluid transport of PGEs

Platinum solubility was determined in a haplobasaltic, diopside-anorthite melt at 1523 K and 0.2 GPa as a function of oxygen fugacity and chlorine content. Synthetic glass powder of an An₄₂Di₅₈ composition was sealed in a platinum or platinum-iridium alloy capsule and equilibrated with a solid CaCl₂ and MgCl₂ chlorine source, water and the noble-metal capsule. All experiments were run in an internally-heated pressure-vessel equipped with a rapid-quench device with oxygen fugacity controlled by the water content and intrinsic hydrogen fugacity of the autoclave (MnO-Mn₃O₄). Resultant glasses were analyzed by isotope dilution ICP-MS and LA-ICP-MS to determine the solubility and distribution of Pt and assess potential Cl-complexation of Pt in the melt.Experiments with run durations longer than 96 h show Pt solubilities consistent with solubilities determined for the equivalent Cl-free diopside-anorthite system, under the same P-T conditions. These results indicate that chlorine has no discernable effect on Pt solubility and there is no evidence of Pt-Cl complexing in the silicate melt from 0.6 to 2.75 wt% Cl (saturation).However, products from short run duration experiments (<96 h) contain Pt concentrations which are orders of magnitude higher than those of the Pt-free starting glass and of the experimental products of the longer run duration experiments. These anomalously high levels are most pronounced in the shortest experiments and Pt concentration decreases with increasing run duration. It is suggested that this excess platinum is dissolved within the Cl-bearing fluid during the heating stages of the experiment and is left behind as the fluid dissolves into the melt leaving small amounts of Pt as “micronuggets”, increasing the bulk Pt concentration. With increasing run duration the platinum appears to migrate out of the melt, back to the capsule walls, decreasing the amount of Pt contained within the glass. This behavior offers compelling evidence that Cl-bearing fluids have the capacity to transport significant amounts of Pt under magmatic conditions.Mass balance calculations on the excess amount of Pt in the glass (above inherent solubility) in short duration experiments lead to an estimation of Pt concentration in the Cl-bearing fluid ranging from tens to a few hundred ppm, versus ppb levels in the melt. The correspondingly high estimated apparent partition coefficients of 10³-10⁴ suggest that Cl-bearing fluids can be highly efficient at enriching and transporting platinum in mafic magmatic-hydrothermal ore-forming systems. These values strongly contrast with recent experimental results in felsic systems, highlighting the potential importance of melt composition on partitioning, the need for composition specific partitioning experiments, as well as a detailed understanding of Pt distribution in experimental products.     

Platinum solubility in a haplobasaltic melt at 1250°C and 0.2 GPa: The effect of water content and oxygen fugacity

Haplobasaltic melts with a 101 kPa dry eutectic composition (An₄₂Di₅₈) and varying water contents were equilibrated with their platinum capsule at 1523 K and 200 MPa in an internally heated pressure vessel (IHPV) equipped with a rapid quench device. Experimental products were inclusion-free glasses representative of the Pt-saturated silicate melts at the experimental conditions. Platinum concentrations were determined using an isotope dilution multicollector inductively coupled plasma mass spectrometer and water contents and distribution by Karl Fischer titration and Fourier transform infrared spectroscopy, respectively.The water content of the melt has no intrinsic effect on platinum solubility, for concentrations between 0.9 wt.% and 4.4 wt.% H₂O (saturation). Platinum solubility increases with increasing water content, but this effect is an indirect effect because increasing water content at fixed f_H2 (imposed by the IHPV) increases the oxygen fugacity of the experiment.The positive oxygen fugacity dependence of Pt solubility in a hydrous silicate melt at 200 MPa is identical to that in anhydrous melts of the same composition determined in previous studies at 101 kPa. This study extends the range of platinum solubilities to oxygen fugacities lower than was previously possible. Combining the data of this and previous studies, Pt solubility is related to oxygen fugacity (in bar) at 1523 K by the equation:

[Pt]_total(ppb)=1389×f_O2+7531×(f_O2)^1/2     

Thermochemistry of sulfide liquids. II. Associated solution model for sulfide liquids in the system O-S-Fe

We present a new formulation to describe the thermodynamics of liquids in the system O-S-Fe. The model is based on an associated regular solution formulation. According to this model, liquids in the O-S-Fe ternary are made up of an equilibrium solution of the six melt species S, Fe, FeO, FeO_1.5, FeS and FeOS. The model presented here represents oxygen and sulfur fugacities as well as phase equilibria with stoichiometric solid phases better than models from the literature on O-Fe and S-Fe binaries. Furthermore, this model represents a substantial improvement on the model of Kress (1997), which is the only other thermodynamic model available in the ternary system. Asymmetric regular solution parameters are required along the FeO join in order to reproduce experimental data with the chosen list of species. Symmetric regular solution parameters are required along the Fe-S binary. Mixing between any of the species considered and FeOS close to ideal. The associated solution model presented here will serve as a more solid foundation for future models in O-S-Fe- Ni-Cu liquids. Efficient and robust strategies for calculating equilibrium speciation and estimating model parameters are presented. Received: 15 June 1999 / Accepted: 5 February 2000     

Pyroxene oikocrysts in troctolitic cumulates — evidence for supercooled crystallisation and postcumulus modification

Studies of pyroxene oikocrysts in plagioclase-olivine cumulates crystallised from high alumina olivine tholeiite magmas in layered intrusions inferred to represent periodically replenished magma chambers show that small plagioclases enclosed within the oikocrysts differ significantly from plagioclases in the host rock. Plagioclases within oikocrysts are very elongate with a high surface area, show extensive mutual attachment at high angles resembling skeletal growth, and tend to be randomly orientated. Some are rounded and embayed, are reversely zoned and more sodic by An 5–10 than some of the plagioclase in the host rock. Plagioclase crystal centres tend to be closer spaced in the oikocrysts than in the host rock. In contrast, plagioclases in the host rock are discrete, coarser grained, less elongate, strongly aligned, and locally polygonised.Plagioclases in the oikocrysts are inferred to be self-nucleated clusters resulting from supercooled crystallisation at up to 25–40° C below equilibrium crystallisation temperatures. The oikocrysts crystallised early and are cumulus, and at least some crystallised in place. They are significant because they probably preserved the early condition of the crystal mush. Comparison of plagioclases within the oikocrysts and in the host rocks allows the textural evolution during solidification to be inferred. Features of supercooled plagioclases in the oikocrysts are generally absent beyond the oikocrysts because they were largely overprinted by postcumulus processes involving grainsize coarsening due to selective resorption, and the development of lamination. Supercooling may thus have been more common than is recognised but the inherent disequilibrium of supercooled products made them susceptible to postcumulus overprint except where preserved in early oikocrysts. Thus, much of the cumulus-looking plagioclase in cumulates may not be truly cumulus, but really the result of substantial postcumulus overprint or magmatic diagenesis.     

Compositions of immiscible liquids in volcanic rocks

Immiscible liquids, preserved as chemically distinct, glassy globules, (Si-rich and Fe-rich) occur in many tholeiitic basalts and some alkaline and calcalkaline lavas. The glasses typically form part of a dark mesostasis containing skeletal magnetite crystals. In thick flows, the Si-rich liquid may crystallize to granophyric patches, and the Ferich one to aggregates of hedenbergite, magnetite, and accessory phases. The mesostases containing these immiscible phases constitute from 20% of a primitive olivine tholeiite (MgO=7.5%) to 50% of a highly fractionated quartz tholeiite (MgO=2.8%), but may be less if the rock is oxidized. Abundant ferric iron promotes early crystallization of magnetite and prevents the iron enrichment necessary to reach the immiscibility field; thus, aa flows rarely exhibit immiscibility, whereas the more reduced pahoehoe ones do.Alumina and alkalis are concentrated in the Si-rich liquid, whereas the remainder of the major elements are concentrated in the Fe-rich melt; but the partitioning of Fe, Mg, Ca, and P is less pronounced in alkaline rocks than in tholeiites. Conjugate liquids have compositions of granite and Fe-rich pyroxenite, though the Si-rich melt in alkaline rocks is more syenitic and the Fe-rich one contains considerable normative alkali feldspar. The liquids coexist with plagioclase and augite of, respectively, An₅₀ and Ca₃₄Mg₁₉Fe₄₇ compositions in tholeiites, and An₄₀ and Ca₄₂Mg₂₉Fe₂₉ in alkaline rocks. Immiscibility is not restricted to K-rich residual liquids, but the miscibility gap is narrower for Na-rich compositions. In tholeiitic basalts with 52% SiO₂, the Na₂O/K₂O ratios in conjugate liquids are equal, but at lower silica contents the Si-rich liquid is relatively more sodic, whereas at higher silica contents it is relatively more potassic. This may explain the association of sodic granites with mid ocean ridge basalts.Immiscible liquids are present in sufficient amounts in so many volcanic rocks that magma unmixing should be considered a viable means of differentiation during the late stages of fractionation of common magmas, at least at low pressures.     

The exsolution of magmatic hydrosaline chloride liquids

Hydrosaline liquid represents the most Cl-enriched volatile phase that occurs in magmas, and the exsolution of this phase has important consequences for processes of hydrothermal mineralization and for volcanic emission of Cl to the atmosphere. To understand the exsolution of hydrosaline liquids in felsic to mafic magmas, the volatile abundances and (Cl/H₂O) ratios of more than 1000 silicate melt inclusions (MI) have been compared with predicted and experimentally determined solubilities of Cl and H₂O and associated (Cl/H₂O) ratios of silicate melts that were saturated in hydrosaline chloride liquid with or without aqueous vapor in hydrothermal experiments. This approach identifies the minimum volatile contents and the values of (Cl/H₂O) at which a hydrosaline chloride liquid exsolves from any CO₂- or SO₂-poor silicate melt. Chlorine solubility is a strong function of melt composition, so it follows that Cl solubility in magmas varies with melt evolution. Computations show that the (Cl/H₂O) ratio of residual melt in evolving silicate magmas either remains constant or increases to a small extent with fractional crystallization. Consequently, the initial (Cl/H₂O) in melt that is established early during partial melting has important consequences for the exsolution of vapor, vapor plus hydrosaline liquid, or hydrosaline liquid later during the final stages of melt ascent, emplacement, and crystallization or eruption. It is demonstrated that the melt (Cl/H₂O) controls the type of volatile phase that exsolves, whereas the volatile abundances in melt control the relative timing of volatile phase exsolution (i.e., the time of earliest volatile exsolution relative to the rate of magma ascent and crystallization history).

Comparing melt inclusion compositions with experimentally determined (Cl/H₂O) ratios and corresponding volatile solubilities of hydrosaline liquid-saturated silicate melts suggests that some fractions of the eruptive, calc-alkaline dacitic magmas of the Bonnin and Izu arcs should have saturated in and exsolved hydrosaline liquid at pressures of 2000 bars. Application of these same melt inclusion data to the predicted volatile solubilities of Cu-, Au-, and Mo-mineralized, calc-alkaline porphyritic magmas suggests that the chemical evolution of dioritic magmas to more-evolved quartz monzonite compositions involves a dramatic reduction in Cl solubility that increases the probability of hydrosaline liquid exsolution. The prediction that quartz monzonite magmas should exsolve a hydrosaline chloride liquid, that is potentially mineralizing, is consistent with the general observation of metal-enriched, hypersaline fluid inclusions in the more felsic plutons of numerous porphyry copper systems. Moreover, comparing the volatile contents of melt inclusions from the potassic, alkaline magmas of Mt. Somma-Vesuvius with the predicted (Cl/H₂O) ratios of hydrosaline liquid-saturated melts having compositions similar to those of the volatile-rich, alkaline magmas associated with the orthomagmatic gold–tellurium deposits of Cripple Creek, Colorado, suggests that hydrosaline chloride liquid should have exsolved at Cripple Creek as the magmas evolved to phonolite compositions. This prediction is consistent with the well-documented role of Cl-enriched, mineralizing hydrothermal fluids at this major gold-mining district.     

Calculated silica activities in carbonatite liquids

Carbonatite magmas precipitate silicates, in addition to the abundant carbonates, oxides, and phosphates. Calculated silica activities for equilibria involving silicates and a silica component in magmatic liquids predict specific assemblages for silicate and oxide phases in carbonatites. These assemblages provide tests of alternative sources (carbonatite magma, coeval silicate magma, or older rock) for silicate minerals in carbonatites. Quartz, feldspars, and orthopyroxene are unlikely to be primary magmatic phases in carbonatites, because the silica activity in carbonatite magmas is too low to stabilize these minerals. Zircon and titanite should be unstable relative to baddeleyite and perovskite, respectively, but they do occur in carbonatites. Liquids dominated by carbonate are strongly nonideal with respect to dissolved silica. Consequently, activity coefficients for a silica component in carbonatite liquids are >>1, so that small mole fractions of SiO₂ translate into silica activities sufficient to stabilize phlogopite, clinopyroxene, amphibole, monticellite, and forsterite, among other silicates. Examination of silicate mineral assemblages in carbonatites in the light of silica activity indicates that many carbonatites are contaminated by solid silicate phases from external sources but these xenocrysts can be discriminated from magmatic minerals.     

In-situ high-temperature Raman spectroscopic studies of aluminosilicate liquids

We have measured in-situ Raman spectra of aluminosilicate glasses and liquids with albite (NaAlSi₃ O₈) and anorthite (CaAl₂Si₂O₈) compositions at high temperatures, through their glass transition range up to 1700 and 2000 K, respectively. For these experiments, we have used a wire-loop heating device coupled with micro-Raman spectroscopy, in order to achieve effective spatial filtering of the extraneous thermal radiation. A major concern in this work is the development of methodology for reliably extracting the first and second order contributions to the Raman scattering spectra of aluminosilicate glasses and liquids from the high temperature experimental data, and analyzing these in terms of vibrational (anharmonic) and configurational changes. The changes in the first order Raman spectra with temperature are subtle. The principal low frequency band remains nearly constant with increasing temperature, indicating little change in the T-O-T angle, and that the angle bending vibration is quite harmonic. This is in contrast to vitreous SiO₂, studied previously. Above Tg, intensity changes in the 560–590 cm^?1 regions of both sets of spectra indicate configurational changes in the supercooled liquids, associated with formation of additional Al-O-Al linkages, or 3-membered (Al, Si)-containing rings. Additional intensity at 800 cm^?1 reflects also some rearrangement of the Si-O-Al network.     

Peralkaline acid liquids: A petrological study

Electron-microprobe analyses of the feldspars and associated ferromagnesian minerals in the peralkaline volcanics, comendites and pantellerites, are presented together with new data on the major and trace-elements of the rocks and residual glasses. The feldspar phenocrysts in the pantellerites span a narrower range (Or₃₃–Or₃₉) than those of the comendites (Or₃₀–Or₄₆); both sets show only limited increase in Or outwards, and the zoning is greatest in quartz-bearing assemblages. The feldspar microlites in the residual glasses are invariably more potassic (2–4% Or) than their associated phenocrysts. In pantellerites the feldspars become more potassic as the residual liquids become more sodic; thus the most potassic feldspar is found in the most sodic (and peralkaline) pantellerite.Of the ferromagnesian phenocrysts, aenigmatite is the most ubiquitous and is commonly associated with hedenbergite±fayalite, or ferrorichterite; in the later stages of crystallization (groundmass), it is associated with acmite, arfvedsonite and tuhualite. Aside from slight variation in Ti/Fe+Ti ratio, aenigmatite is virtually constant in composition. The pyroxenes from the different assemblages have zones which together almost span the range acmitehedenbergite. Both ferrorichterite and arfvedsonite incorporate F but not Cl, and are slightly potassic. Tuhualite exists as two varieties; one blue and potassic, the other violet and sodic; both varieties reject halogens. Using (estimated) free-energy data, a field in fo₂, T space is postulated in which Fe-Ti oxides are absent; their place is taken by pyroxene and aenigmatite. The no-oxide field will be intercepted by a cooling liquid in which peralkalinity is increasing and in which fo₂ is near but above the FMQ buffer.The characteristic pattern of trace-elements in peralkaline volcanics (e.g., high Nb, Ta, Zr, Mo, Zn, Cd, R.E, etc; low Sr, Ba, Mg) are considered to be as much evidence for the peralkaline (salic) condition as of the genetic process. Several lines of evidence suggest that at liquidus temperatures, peralkaline rhyolites are essentially anhydrous.     

Voluminous arc dacites as amphibole reaction-boundary liquids

Dacites dominate the large-volume, explosive eruptions in magmatic arcs, and compositionally similar granodiorites and tonalites constitute the bulk of convergent margin batholiths. Shallow, pre-eruptive storage conditions are well known for many dacitic arc magmas through melt inclusions, Fe–Ti oxides, and experiments, but their potential origins deeper in the crust are not well determined. Accordingly, we report experimental results identifying the P–T–H₂O conditions under which hydrous dacitic liquid may segregate from hornblende (hbl)-gabbroic sources either during crystallization–differentiation or partial melting. Two compositions were investigated: (1) MSH–Yn?1 dacite (SiO₂: 65 wt%) from Mount St. Helens’ voluminous Yn tephra and (2) MSH–Yn?1?+?10% cpx to force saturation with cpx and map a portion of the cpx?+?melt?=?hbl peritectic reaction boundary. H₂O-undersaturated (3, 6, and 9 wt% H₂O) piston cylinder experiments were conducted at pressures, temperatures, and fO₂ appropriate for the middle to lower arc crust (400, 700, and 900 MPa, 825–1100?°C, and the Re–ReO₂ buffer?≈?Ni–NiO?+?2). Results for MSH–Yn?1 indicate near-liquidus equilibrium with a cpx-free hbl-gabbro residue (hbl, plg, magnetite, ± opx, and ilmeno-hematite) with 6–7 wt% dissolved H₂O, 925?°C, and 700–900 MPa. Opx disappears down-temperature consistent with the reaction opx?+?melt?=?hbl. Cpx-added phase relations are similar in that once ~10% cpx crystallizes, multiple saturation is attained with cpx, hbl, and plg, +/? opx, at 6–7 wt% dissolved H₂O, 940?°C, and 700–900 MPa. Plg–hbl–cpx saturated liquids diverge from plg–hbl–opx saturated liquids, consistent with the MSH–Yn?1 dacite marking a liquid composition along a peritectic distributary reaction boundary where hbl appears down-temperature as opx?+?cpx are consumed. The abundance of saturating phases along this distributary peritectic (liquid?+?hbl?+?opx?+?cpx?+?plg?+?oxides) reduces the variance, so liquids are restricted to dacite–granodiorite–tonalite compositions. Higher-K dacites than the Yn would also saturate with biotite, further limiting their compositional diversity. Theoretical evaluation of the energetics of peritectic melting of pargasitic amphiboles indicates that melting and crystallization of amphibole occur abruptly, proximal to amphibole’s high-temperature stability limit, which causes the system to dwell thermally under the conditions that produce dacitic compositions. This process may account for the compositional homogeneity of dacites, granodiorites, and tonalites in arc settings, but their relative mobility compared to rhyolitic/granitic liquids likely accounts for their greater abundance.     

Energetics of water dissolution in trachyte glasses and liquids

Enthalpies of dissolution in HF solutions have been measured at 323 K for a series of hydrous trachyte glasses. Enthalpies of mixing between water and molten trachyte have then been calculated from heat capacity data for the same set of samples and available enthalpy for pure water. The moderately negative enthalpies of mixing suggested at 1 bar by the measurements made on glasses almost disappear when trachyte liquids and water are referred to the same temperature, and particularly so when enthalpies of mixing are calculated for a few kbars pressure. As found for albite and phonolite liquids, trachyte melts thus appear to mix nearly ideally as far as enthalpy is concerned. These results imply that the enthalpy of exsolution of water from magmas is very small or negligible under the P-T-X conditions relevant to trachytic volcanism, even for complete degassing of up to 5 wt% H₂O. Furthermore, the viscosity increase associated with exsolution-driven cooling is negligible compared to the decrease caused directly by water exsolution.     

Heat capacities and entropies of silicate liquids and glasses

The heat capacities of several dozen silicate glasses and liquids composed of SiO₂, TiO₂, Al₂O₃, Fe₂O₃, FeO, MgO, CaO, BaO, Li₂O, Na₂O, K₂O, and Rb₂O have been measured by differential scanning and drop calorimetry. These results have been combined with data from the literature to fit C _pas a function of composition. A model assuming ideal mixing (linear combination) of partial molar heat capacities of oxide components (each of which is independent of composition), reproduces the glass data within error. The assumption of constancy of ¯C _p,iis less accurate for the liquids, but data are not sufficient to adequately constrain a more complex model. For liquids containing alkali metal and alkaline earth oxides, heat capacities are systematically greater in liquids with high field strength network modifying cations. Entropies of fusion (per g-atom) and changes of configurational entropy with temperature, are similarly affected by composition. Both smaller cation size and greater charge are therefore inferred to lead to greater development of new structural configurations with increasing temperature in silicate liquids.     

Dynamics of Na in sodium aluminosilicate glasses and liquids

²³Na NMR measurements on Na₂Si₃O₇, Na₃AlSi₆O₁₅, and NaAlSi₃O₈ glasses from room temperature to 1200°C show that the dynamics and local structure of sodium in silicate/aluminosilicate glasses and melts vary with composition and temperature.The peak positions decrease in frequency between room temperature and 200°C indicating that the Na sees a larger average site as temperature is increased. Between 200°–300° and 700°C, line widths, nutation frequencies and peak positions are consistent with motional averaging of quadrupolar satellites. Above 700°C there is little or no change in the peak positions with temperature. Chemical shifts of the materials at 1000°C (Na₂Si₃O₇: 3.6; Na₃AlSi₆O₁₅:-1.3; NaAlSi₃O₈:-6.4 ppm) indicate a slight change in the average Na coordination number from 6–7 for the silicate to 7–8 for the aluminosilicates.