The structural role and homogeneous redox equilibria of iron in peraluminous,metaluminous and peralkaline silicate melts

The compositional dependence of the redox ratio (FeO/FeO_1.5) has been experimentally determined in K₂O-Al₂O₃-SiO₂-Fe₂O₃-FeO (KASFF) and K₂O-CaO-Al₂O₃-SiO₂-Fe₂O₃-FeO (KCASFF) silicate melts. Compositions were equilibrated at 1,450° C in air, with 78 mol % SiO₂. KASFF melts have from 1 to 5 mol % Fe₂O₃ and include both peraluminous (K₂O2O₃) and peralkaline (K₂O>Al₂O₃) compositions. KCASFF melts have 1 mol % Fe₂O₃ encompassing peraluminous, metaluminous (CaO+K₂O>Al₂O₃) and peralkaline compositions. Peralkaline KASFF melts with 1 mol % Fe₂O₃ have low and constant values for the redox ratio, whereas in peraluminous melts the redox ratio increases with increasing (K₂O/Al₂O₃). Increasing total iron concentration increases the redox ratio in peraluminous melts and slightly decreases the redox ratio in peralkaline melts. Substituting CaO for K₂O at fixed total iron (1 mol %) increases the redox ratio in both peraluminous and metaluminous KCASFF melts; however, the redox ratio in peralkaline KCASFF melts is not affected by this exchange. These data indicate that Fe³⁺ is in four-fold coordination, with K⁺ or Ca²⁺ providing local charge balance. The tetrahedral ferric species is most stable in peralkaline melts and least stable in peraluminous melts, due to the competition between Al³⁺ and Fe³⁺ for charge balancing cations in the latter melt. Tetrahedral Fe³⁺ is also less stable when Ca²⁺ provides local charge balance. The data are consistent with a network modifying role for Fe²⁺ in the melt.The data are interpreted to reflect the effects of melt composition on the partitioning of K⁺ and Ca²⁺ and Fe³⁺ and Al³⁺ between various species in the melt. These relationships are discussed in terms of homogeneous equilibria between various iron-bearing and iron-free melt species. The results also reflect the effect of liquid composition on the exchange potentials Fe³⁺ Al_–1 and Ca_0.5K_–1. The exchange potentials are relatively constant in peralkaline melts, but decrease in metaluminous and peraluminous melts as both (CaO+K₂O)/(CaO+K₂O+Al₂O₃) and K₂O/CaO decrease. These qualitative observations imply that minerals exhibiting these exchanges will also be similarly affected as liquid composition changes. Present address: Department of Geological Sciences, Virginia Tech, Blacksburg, VA 24061, USA     

Shear viscosities of CaO-Al2O3-SiO2 and MgO-Al2O3-SiO2 liquids: Implications for the structural role of aluminium and the degree of polymerisation of synthetic and natural aluminosilicate melts

The shear viscosity of 66 liquids in the systems CaO-Al₂O₃-SiO₂ (CAS) and MgO-Al₂O₃-SiO₂ (MAS) have been measured in the ranges 1-10⁴ Pa s and 10⁸-10¹² Pa s. Liquids belong to series, nominally at 50, 67, and 75 mol.% SiO₂, with atomic M²⁺/(M²⁺+2Al) typically in the range 0.60 to 0.40 for each isopleth. In the system CAS at 1600°C, viscosity passes through a maximum at all silica contents. The maxima are clearly centered in the peraluminous field, but the exact composition at which viscosity is a maximum is poorly defined. Similar features are observed at 900°C. In contrast, data for the system MAS at 1600°C show that viscosity decreases with decreasing Mg/(Mg + 2Al) at all silica contents, but that a maximum in viscosity must occur in the field where Mg/2Al >1. On the other hand, the viscosity at 850°C increases with decreasing Mg/(Mg + 2Al) and shows no sign of reaching a maximum, even for the most peraluminous composition studied. The data from both systems at 1600°C have been analysed assuming that shear viscosity is proportional to average bond strength and considering the equilibrium:

Al^[4]-(Mg,Ca)_0.5⇔(Mg,Ca)_0.5-NBO+Al_XS     

Columbite solubility in granitic melts: consequences for the enrichment and fractionation of Nb and Ta in the Earth's crust

The behaviour of niobium and tantalum in magmatic processes has been investigated by conducting MnNb₂O₆ and MnTa₂O₆ solubility experiments in nominally dry to water-saturated peralkaline (aluminium saturation index, A.S.I. 0.64) to peraluminous (A.S.I. 1.22) granitic melts at 800 to 1035 °C and 800 to 5000 bars. The attainment of equilibrium is demonstrated by the concurrence of the solubility products from dissolution, crystallization, Mn-doped and Nb- or Ta-doped experiments at the same pressure and temperature. The solubility products of MnNb₂O₆ (K_sp ^Nb) and MnTa₂O₆ (K_sp ^Ta) at 800 °C and 2 kbar both increase dramatically with alkali contents in water-saturated peralkaline melts. They range from 1.2 × 10⁻⁴ and 2.6 × 10⁻⁴ mol²/kg², respectively, in subaluminous melt (A.S.I. 1.02) to 202 × 10⁻⁴ and 255 × 10⁻⁴ mol²/kg², respectively, in peralkaline melt (A.S.I. 0.64). This increase from the subaluminous composition can be explained by five non-bridging oxygens being required for each excess atom of Nb⁵⁺ or Ta⁵⁺ that is dissolved into the melt. The K_sp ^Nb and K_sp ^Ta also increase weakly with Al content in peraluminous melts, ranging up to 1.7 × 10⁻⁴ and 4.6 × 10⁻⁴ mol²/kg², respectively, in the A.S.I. 1.22 composition. Columbite-tantalite solubilities in subaluminous and peraluminous melts (A.S.I. 1.02 and 1.22) are strongly temperature dependent, increasing by a factor of 10 to 20 from 800 to 1035 °C. By contrast columbite-tantalite solubility in the peralkaline composition (A.S.I. 0.64) is only weakly temperature dependent, increasing by a factor of less than 3 over the same temperature range. Similarly, K_sp ^Nb and K_sp ^Ta increase by more than two orders of magnitude with the first 3 wt% H₂O added to the A.S.I. 1.02 and 1.22 compositions, whereas there is no detectable change in solubility for the A.S.I. 0.64 composition over the same range of water contents. Solubilities are only slightly dependent on pressure over the range 800 to 5000 bars. The data for water-saturated sub- and peraluminous granites have been extrapolated to 600 °C, conditions at which pegmatites and highly evolved granites may crystallize. Using a melt concentration of 0.05 wt% MnO, 70 to 100 ppm Nb or 500 to 1400 ppm Ta are required for manganocolumbite and manganotantalite saturation, respectively. The solubility data are also used to model the fractionation of Nb and Ta between rutile and silicate melts. Predicted rutile/melt partition coefficients increase by about two orders of magnitude from peralkaline to peraluminous granitic compositions. It is demonstrated that the γNb₂O₅/γTa₂O₅ activity coefficient ratio in the melt phase depends on melt composition. This ratio is estimated to decrease by a factor of 4 to 5 from andesitic to peraluminous granitic melt compositions. Accordingly, all the relevant accessory phases in subaluminous to peraluminous granites are predicted to incorporate Nb preferentially over Ta. This explains the enrichment of Ta over Nb observed in highly fractionated granitic rocks, and in the continental crust in general. Received: 9 August 1996 / Accepted: 26 February 1997     

Iron and titanium solution properties in peraluminous and peralkaline rhyolitic liquids

The saturation surface of pseudobrookite (Fe₂TiO₅) was determined for melts in the system SiO₂-Al₂O₃-K₂O-FeO-Fe₂O₃-TiO₂ at 1400° C and 1 atm. The variation in concentrations of Fe₂O₃, TiO₂ and Fe₂TiO₅ in liquids can be used to infer relative changes in activity coefficients of these components with changing K₂O/(K₂O+Al₂O₃) of the melts. Saturation concentrations of these components are low and relatively constant in the peraluminous melts and increase with increasing K₂O/(K₂O+Al₂O₃) in peralkaline liquids. The activity coefficients of Fe₂O₃ and TiO₂ and Fe₂TiO₅, therefore, are higher in peraluminous liquids than in peralkaline liquids in this system. In addition, the iron redox ratio was measured as a function of K₂O/(K₂O+Al₂O₃) for liquids just below the saturation surface; was fixed so all variations in redox ratio are entirely due to changes in melt composition. The redox ratio from unsaturated liquids was applied to saturated liquids where redox analysis of the glass is impossible. The homogeneous equilibrium experiments indicate that the activity coefficient of Fe₂O₃ relative to that of FeO is significantly greater in peraluminous melts than peralkaline melts. Both the heterogeneous and homogeneous equilibria suggest that in peralkaline liquids K⁺in excess of that required to charge balance tetrahedral Al⁺³ is used to stabilize both Fe⁺³ and Ti⁺⁴. Calculations show that ferric iron and titanium compete equally effectively for charge-balancing potassium but neither can outcompete aluminum. The observed changes in solution properties of Fe₂O₃ and TiO₂ in the synthetic melts are used to explain variations in Fe-Ti oxide stabilities in natural peraluminous and peralkaline rhyolites and granites. Since the activity coefficients of both ferric iron and titanium are significantly higher in peraluminous liquids than in peralkaline liquids, Fe-Ti oxides should occur earlier in the crystallization sequence in peraluminous rhyolites than in peralkaline rhyolites. In addition, iron will be reduced in peraluminous granites and rhyolites relative to peralkaline ones under comparable P, T, and . Finally, observed crystallization patterns for minerals containing highly charged cations other than ferric iron and titanium are evaluated in the context of this and other experimental studies.     

P–T–fO2 controls on a partly inverse chromite-bearing ultramafic intrusive: an evaluation from the Sukinda Massif, India

The chromites from the alpine type ultramafic intrusive of Sukinda, India, display a typical partly inverse spinel form and occur in two distinct zones: Brown Ore Zone (BOZ) and Grey Ore Zone (GOZ). The host ultramafites are mostly altered and are represented by the serpentinite, tremolite-talc(chlorite) schist, talc-serpentine schist and chlorite rock. The less altered variants are dunite, harzburgite and websterite. A dyke of orthopyroxenite runs through the main ultramafic body.The composition of olivine (Fo₉₂), orthopyroxene (En_92–89) and Al₂O₃ contents of the parental liquid (10.40–11.45%) determined from chromites, suggest that the parent melt is of boninitic affinity. The chemical plot of TiO₂ content against cr# of chromites corroborates a boninitic parental melt. The Fe–Mg partitioning in olivine and chromite depicts the temperature for chromitites as 1200 °C. A compositional plot of mg# and cr# suggests crystallization at high pressure conditions, corresponding to the kimberlite xenolith field. From the P–T diagram of pyrolite melting and mineral assemblage, the pressure of crystallization is stipulated to be ≥1.2 GPa. The f_O2 values estimated from Fe³⁺/Cr+Al+Fe³⁺ ratios range from 10^−8.3 to 10^−9.3 for the GOZ and 10^−7.1 to 10^−7.3 for the BOZ. The f_O2 values together with the pressure range suggest crystallization at upper mantle conditions. The heterogeneity in chemical composition and f_O2 conditions for the GOZ and BOZ could be linked to heterogeneity in the upper mantle.     

Experimental Constraints on the Relationships between Peralkaline Rhyolites of the Kenya Rift Valley

Crystallization experiments on three comendites provide evidencefor the genetic relationships between peralkaline rhyolitesin the central Kenya rift valley. The crystallization of calcicclinopyroxene in slightly peralkaline rhyolites inhibits increasein peralkalinity by counteracting the effects of feldspar. Fractionationunder high fO₂ conditions produces residual liquids that areless, or only slightly more, peralkaline than the bulk composition.In contrast, crystallization under reduced conditions (<FMQ,where FMQ is the fayalite–magnetite–quartz buffer)and at high fF₂ inhibits calcic clinopyroxene and yields residualliquids that are more peralkaline than coexisting alkali feldspar,whose subsequent crystallization increases the peralkalinityof the liquid. A marginally peralkaline rhyolite [molar (Na₂O+ K₂O)/Al₂O₃ (NK/A) = 1·05] can yield a more typicallycomenditic rhyolite (NK/A = 1·28) after 95 wt % of crystallization.This comendite yields pantelleritic derivatives (NK/A >1·4)after 25 wt % crystallization. Upon further crystallization,extreme peralkaline compositions (NK/A     

Solubility of manganotantalite, zircon and hafnon in highly fluxed peralkaline to peraluminous pegmatitic melts

The behavior of tantalum and zirconium in pegmatitic systems has been investigated through the determination of Ta and Zr solubilities at manganotantalite and zircon saturation from dissolution and crystallization experiments in hydrous, Li-, F-, P- and B-bearing pegmatitic melts. The pegmatitic melts are synthetic and enriched in flux elements: 0.7–1.3 wt% Li₂O, 2–5.5 wt% F, 2.8–4 wt% P₂O₅ and 0–2.8 wt% B₂O₃, and their aluminum saturation index ranges from peralkaline to peraluminous (ASI_Li = Al/[Na + K + Li] = 0.8 to 1.3) with various K/Na ratios. Dissolution and crystallization experiments were conducted at temperatures varying between 700 and 1,150°C, at 200 MPa and nearly water-saturated conditions. For dissolution experiments, pure synthetic, end member manganotantalite and zircon were used in order to avoid problems with slow solid-state kinetics, but additional experiments using natural manganotantalite and zircon of relatively pure composition (i.e., close to end member composition) displayed similar solubility results. Zircon and manganotantalite solubilities considerably increase from peraluminous to peralkaline compositions, and are more sensitive to changes in temperature or ASI of the melt than to flux content. A model relating the enthalpy of dissolution of manganotantalite to the ASI_Li of the melt is proposed: ∆H _diss (kJ/mol) = 304 × ASI_Li − 176 in the peralkaline field, and ∆H _diss (kJ/mol) = −111 × ASI_Li + 245 in the peraluminous field. The solubility data reveal a small but detectable competitivity between Zr and Ta in the melt, i.e., lower amounts of Zr are incorporated in a Ta-bearing melt compared to a Ta-free melt under the same conditions. A similar behavior is observed for Hf and Ta. The competitivity between Zr (or Hf) and Ta increases from peraluminous to peralkaline compositions, and suggests that Ta is preferentially bonded to non-bridging oxygens (NBOs) with Al as first-neighbors, whereas Zr is preferentially bonded to NBOs formed by excess alkalies. As a consequence Zr/Ta ratios, when buffered by zircon and manganotantalite simultaneously, are higher in peralkaline melts than in peraluminous melts.     

The role of phosphorus in rhyolitic liquids as determined from the homogeneous iron redox equilibrium

The redox ratio of iron is used as an indicator of solution properties of silicate liquids in the system (SiO–Al₂O₃–K₂O–FeO–Fe₂O₃–P₂O₅). Glasses containing 80–85 mol% SiO₂ with 1 mol% Fe₂O₃ and compositions covering a range of K₂O/Al₂O₃ were synthesized at 1400°C in air (fixed fO₂). Variations in the ratio FeO/FeO_1.5 resulting from the addition of P₂O₅ are used to determine the solution behavior of phosphorus and its interactions with other cations in the silicate melt. In 80 mol% SiO₂ peralkaline melts the redox ratio, expressed as FeO/FeO_1.5, is unchanged relative to the reference curve with the addition of 3 mol% P₂O₅. Yet, the iron redox ratio in the 85 mol% SiO₂ potassium aluminosilicate melts is decreased relative to phosphorus-free liquids even for small amounts of P₂O₅ (0.5 mol%). The redox ratio in peraluminous melts is decreased relative to phosphorus- free liquids at P₂O₅ concentrations of 3 mol%. In peraluminous liquids, complexing of both Fe⁺³–O–P⁺⁵ and Al⁺³–O–P⁺⁵ occur. The activity coefficient of Fe⁺³ is decreased because more ferric iron can be accommodated than in phosphorus-free liquids. In peralkaline melts, there is no evidence that P⁺⁵ is removing K⁺ from either Al⁺³ or Fe⁺³ species. In chargebalanced melts with 3 mol% Fe₂O₃ and very high P₂O₅ concentrations, phosphorus removes K⁺ from K–O–Fe⁺³ complexes resulting in a redox increase. P₂O₅ should be accommodated easily in peraluminous rhyolitic liquids and phosphate saturation may be suppressed relative to metaluminous rhyolites. In peralkaline melts, phosphate solubility may increase as a result of phosphorus complexing with alkalis. The complexing stoichiometry may be variable, however, and the relative influence of peralkalinity versus temperature on phosphate solubility in rhyolitic melts deserves greater attention.     

Thermodynamic analysis of the system Na2O-K2O-CaO-Al2O3-SiO2-H2O-F2O?1: Stability of fluorine-bearing minerals in felsic igneous suites

Thermodynamic analysis of the system Na₂O-K₂O-CaO-Al₂O₃-SiO₂-H₂O-F₂O_–1 provides phase equilibria and solidus compatibilities of rock-forming silicates and fluorides in evolved granitic systems and associated hydrothermal processes. The interaction of fluorine with aluminosilicate melts and solids corresponds to progressive fluorination of their constituent oxides by the thermodynamic component F₂O_–1. The chemical potential (F₂O_–1) buffered by reaction of the type: MO_n/2 (s)+n/2 [F₂O_–1]=MF_n (s, g) where M=K, Na, Ca, Al, Si, explains the sequential formation of fluorides: carobbiite, villiaumite, fluorite, AlF₃, SiF₄ as well as the common coexistence of alkali- and alkali-earth fluorides with rock-forming aluminosilicates. Formation of fluorine-bearing minerals first starts in peralkaline silica-undersaturated, proceeds in peraluminous silica-oversaturated compositions and causes progressive destabilization of nepheline, albite and quartz, in favour of villiaumite, cryolite, topaz, chiolite. Additionally, it implies the increase of buffered fluorine solubilities in silicate melts or aqueous fluids from peralkaline silica-undersaturated to peraluminous silica-oversaturated environments. Subsolidus equilibria reveal several incompatibilities: (i) topaz is unstable with nepheline or villiaumite; (ii) chiolite is not compatible with albite because it only occurs only at very high F₂O_–1 levels. The stability of topaz, fluorite, cryolite and villiaumite in natural felsic systems is related to their peralkalinity (peraluminosity), calcia and silica activity, and linked by corresponding chemical potentials to rock-forming mineral buffers. Villiaumite is stable in strongly peralkaline and Ca-poor compositions (An_<0.001). Similarly, cryolite stability requires coexistence with nearly-pure albite (An_<2). Granitic rocks with Ca-bearing plagioclase (An_>5) saturate with topaz or fluorite. Crystallization of topaz is restricted to peraluminous conditions, consistent with the presence of Li-micas or anhydrous aluminosilicates (cordierite, garnet, andalusite). Fluorite is predicted to be stable in peraluminous biotite granites, amphibole-, clinopyroxene- or titanite-bearing calc-alkaline suites as well as in peralkaline granitic and syenitic rocks. Fluorine concentrations in felsic melts buffered by the coexistence of F-bearing minerals and feldspars increase from peralkaline through metaluminous to mildly peraluminous compositions. At low-temperature conditions, the hydrothermal evolution of peraluminous granitic and greisen systems is controlled by white mica-feldspar-fluoride equilibria. With decreasing temperature, topaz gradually breaks down via: (i) (OH)F_–1 substitution and fluorine transfer to fluorite by decalcification of plagioclase below 600 °C, (ii) formation of muscovite and additional fluorite at 475–315 °C, and (iii) formation of paragonite and cryolite, consuming F-rich topaz and albite below 315 °C. These equilibria explain the absence of magmatic fluorite in Ca-bearing topaz granitic rocks; its abundance in hydrothermal rocks is due to: (i) closed-system defluorination of topaz, (ii) open-system decalcification of plagioclase or (iii) hydrolytic alteration. These results provide a complete framework for the investigation of fluorine-bearing mineral stabilities in felsic igneous suites.Electronic Supplementary Material Supplementary material is available in the online version of this article at . A link in the frame on the left on that page takes you directly to the supplementary material.Editorial responsibility: T.L. Grove     

Lamellar nigerite in Zn-rich spinel from the Falun deposit,Sweden

Lamellar nigerite is found in Zn-rich spinel from a sample that contains chiefly anthophyllite + spinel + cordierite, lesser amounts of quartz and chlorite, as well as sphalerite, pyrite, pyrrhotite, and galena, and rare cassiterite and rutile. Nigerite can be described as interlayering of spinel-like (R²⁺Al₂O₄) and nolanite-like ((Sn, Ti)Al₄O₈) structures. In nigerite, the spinel-like part is also compositionally related to spinel, but in the nolanite-like part only the structural analogy exists. Stoichiometric assumptions that relate the anhydrous cation sum to the amount of R⁴⁺ cations present, allow Fe³⁺ estimates from microprobe analyses, and a representative analysis gives the following anhydrous formula: Mg_1.30Fe _0.65 ²⁺ Zn_3.03Mn_0.03Al_11.65Fe _0.35 ³⁺ Sn_0.32Ti_0.18O₂₄.The nigerite is Zn-rich with a Zn ratio (Zn/(Zn+Mg+ Fe²⁺)) of about 0.59, and the Sn ratio (Sn/(Sn+Ti)) that ranges from about 0.63 to 0.41. The Fe³⁺ content in these samples ranges from 0.35 to 0.52 (24 oxygen basis).Textures suggest that the nigerite could have formed by the breakdown of R ₂ ²⁺ (Sn, Ti)O₄ and R²⁺Al₂O₄ spinel components during more complex reactions. An experimental investigation of the MgAl₂O₄-Mg₂SnO₄ join indicates that the solubility of Mg₂SnO₄ component in spinel over the T interval 500 to 900° C is about 0.5 to 3.0 mole %. This, coupled with the increased solubility expected from the presence of Ti, gives good agreement with the 2.4 to 2.6 mole % R ₂ ²⁺ (Sn, Ti)O₄ component in spinel that is estimated to be the maximum necessary to form the compositions and amounts of observed nigerite.     

Na2O solubility in CaO-MgO-SiO2 melts

The sodium solubility in silicate melts in the CaO-MgO-SiO₂ (CMS) system at 1400 °C has been measured by using a closed thermochemical reactor designed to control alkali metal activity. In this reactor, Na_(g) evaporation from a Na₂O-xSiO₂ melt imposes an alkali metal vapor pressure in equilibrium with the molten silicate samples. Because of equilibrium conditions in the reactor, the activity of sodium-metal oxide in the molten samples is the same as that of the source, i.e., aNa₂O_(sample) = aNa₂O_(source). This design also allows to determine the sodium oxide activity coefficient in the samples. Thirty-three different CMS compositions were studied. The results show that the amount of sodium entering from the gas phase (i.e., Na₂O solubility) is strongly sensitive to silica content of the melt and, to a lesser extent, the relative amounts of CaO and MgO. Despite the large range of tested melt compositions (0 < CaO and MgO < 40; 40 < SiO₂ < 100; in wt%), we found that Na₂O solubility is conveniently modeled as a linear function of the optical basicity (Λ) calculated on a Na-free basis melt composition. In our experiments, γNa₂O_(sample) ranges from 7 × 10⁻⁷ to 5 × 10⁻⁶, indicating a strongly non-ideal behavior of Na₂O solubility in the studied CMS melts (γNa₂O_(sample) ? 1). In addition to showing the effect of sodium on phase relationships in the CMS system, this Na₂O solubility study brings valuable new constraints on how melt structure controls the solubility of Na in the CMS silicate melts. Our results suggest that Na₂O addition causes depolymerization of the melt by preferential breaking of Si-O-Si bonds of the most polymerized tetrahedral sites, mainly Q⁴.     

Phosphorus solubility mechanisms in haplogranitic aluminosilicate glass and melt: effect of temperature and aluminum content

The solubility behavior of phosphorus in glasses and melts in the system Na₂O-Al₂O₃-SiO₂-P₂O₅ has been examined as a function of temperature and Al₂O₃ content with microRaman spectroscopy. The Al₂O₃ was added (2, 4, 5, 6, and 8 mol% Al₂O₃) to melts with 80 mol% SiO₂ and ∼2 mol% P₂O₅. The compositions range from peralkaline, via meta-aluminous to peraluminous. Raman spectra were obtained of both the phosphorus-free and phosphorous-bearing glasses and melts between 25 and 1218 °C. The Raman spectrum of Al-free, P-bearing glass exhibits a characteristic strong band near 940 cm⁻¹ assigned to P=O stretching in orthophosphate complexes together with a weaker band near 1000 cm⁻¹ assigned P₂O₇ complexes. With increasing Al content, the proportion of P₂O₇ initially increases relative to PO₄ and is joined by AlPO₄ complexes which exhibit a characteristic P-O stretch mode slightly above 1100 cm⁻¹. The latter complex appears to dominate in meta-aluminosilicate glass and is the only phosphate complex in peraluminous glasses. When P-bearing peralkaline silicate and aluminosilicate glasses are transformed to supercooled melts, there is a rapid decrease in PO₄/P₂O₇ so that in the molten state, PO₄ units are barely discernible. The P₂O₇/AlPO₄ abundance ratio in peralkaline compositions increases with increasing temperature. This decrease in PO₄/P₂O₇ with increasing temperature results in depolymerization of the silicate melts. Dissolved P₂O₅ in peraluminous glass and melts forms AlPO₄ complexes only. This solution mechanism has no discernible influence on the aluminosilicate melt structure. There is no effect of temperature on this solution mechanism. Received: 7 October 1997 / Accepted: 11 May 1998     

Redox equilibria and the structural role of iron in alumino-silicate melts

The relationship between the redox ratio Fe⁺²/(Fe⁺²+Fe⁺³) and the K₂O/(K₂O + Al₂O₃) ratio (K₂O^*) were experimentally investigated in silicate melts with 78 mol% SiO₂ in the system SiO₂-Al₂O₃-K₂O-FeO-Fe₂O₃, in air at 1,400° C. Quenched glass compositions were analyzed by electron microprobe and wet chemical microtitration techniques. Minimum values of the redox ratio were obtained at K₂O^*0.5. The redox ratio in peralkaline melts (K₂O^*>0.5) increases slightly with K₂O^* whereas this ratio increases dramatically in peraluminous melts (K₂O^*<0.5) as K₂O is replaced by Al₂O₃. These data indicate that all Fe⁺³ (and Al⁺³) occur as tetrahedral species charge balanced with K⁺ in peralkaline melts. In peraluminous melts, Fe⁺³ (and Al⁺³) probably occur as both tetrahedral species using Fe⁺² as a charge-balancing cation and as network-modifying cations associated with non-bridging oxygen.     

The compressibility of silicate liquids containing Fe2O3 and the effect of composition,temperature, oxygen fugacity and pressure on their redox states

Ultrasonic longitudinal acoustic velocities in oxidized silicate liquids indicate that the pressure derivative of the partial-molar volume of Fe₂O₃ is the same in iron-rich alkali-, alkaline earth- and natural silicate melt compositions at 1 bar. The dV/dP for multicomponent silicate liquids can be expressed as a linear combination of partial-molar constants plus a positive excess term for Na₂O−Al₂O₃ mixing. Partial-molar properties for FeO and Fe₂O₃ components allow extension of the empirical expression of Sack et al. (1980) to permit the calculation of Fe-redox equilibrium in a natural silicate liquid as a function of composition, temperature, fo₂ and pressure; a more formal thermodynamic expression is presented in the Appendix. The predicted equilibrium fo₂ of natural silicate melts, of fixed oxygen content, closely parallels that defined by the metastable assemblage fayalite+magnetite+β-quartz (FMQ), in pressure-temperature space. A silicate melt initially equilibrated at 3 GPa and FMQ, will remain within approximately 0.5 log₁₀ units of FMQ during its closed-system ascent. Thus, for magmas closed to oxygen, iron-redox equilibrium in crystal-poor pristine glassy lavas represents an excellent probe of the relative oxidation state of their source regions.     

Calculated phase equilibria in K2O-FeO-MgO-Al2O3-SiO2-H2O for sapphirine-quartz-bearing mineral assemblages

Sapphirine, coexisting with quartz, is an indicator mineral for ultrahigh‐temperature metamorphism in aluminous rock compositions. Here a new activity‐composition model for sapphirine is combined with the internally consistent thermodynamic dataset used by THERMOCALC, for calculations primarily in K₂O‐FeO‐MgO‐Al₂O₃‐SiO₂‐H₂O (KFMASH). A discrepancy between published experimentally derived FMAS grids and our calculations is understood with reference to H₂O. Published FMAS grids effectively represent constant a_H2O sections, thereby limiting their detailed use for the interpretation of mineral reaction textures in compositions with differing H₂O. For the calculated KFMASH univariant reaction grid, sapphirine + quartz assemblages occur at P–T in excess of 6–7 kbar and 1005 °C. Sapphirine compositions and composition ranges are consistent with natural examples. However, as many univariant equilibria are typically not ‘seen’ by a specific bulk composition, the univariant reaction grid may reveal little about the detailed topology of multi‐variant equilibria, and therefore is of limited use for interpreting the P–T evolution of mineral assemblages and reaction sequences. Calculated pseudosections, which quantify bulk composition and multi‐variant equilibria, predict experimentally determined KFMASH mineral assemblages with consistent topology, and also indicate that sapphirine stabilizes at increasingly higher pressure and temperature as X_Mg increases. Although coexisting sapphirine and quartz can occur in relatively iron‐rich rocks if the bulk chemistry is sufficiently aluminous, the P–T window of stability shrinks with decreasing X_Mg. An array of mineral assemblages and mineral reaction sequences from natural sapphirine + quartz and other rocks from Enderby Land, Antarctica, are reproducible with calculated pseudosections. That consistent phase diagram calculations involving sapphirine can be performed allows for a more thorough assessment of the metamorphic evolution of high‐temperature granulite facies terranes than was previously possible. The establishment of a a‐x model for sapphirine provides the basis for expansion to larger, more geologically realistic chemical systems (e.g. involving Fe³⁺).     

Energetics of multicomponent diffusion in molten CaO-Al2O3-SiO2

The energetics of multicomponent diffusion in molten CaO-Al₂O₃-SiO₂ (CAS) were examined experimentally at 1440 to 1650°C and 0.5 to 2 GPa. Two melt compositions were investigated: a haplodacitic melt (25 wt.% CaO, 15% Al₂O₃, and 60% SiO₂) and a haplobasaltic melt (35% CaO, 20% Al₂O₃, and 45% SiO₂). Diffusion matrices were measured in a mass-fixed frame of reference with simple oxides as end-member components and Al₂O₃ as a dependent variable. Chemical diffusion in molten CAS shows clear evidence of diffusive coupling among the components. The diffusive flux of SiO₂ is significantly enhanced whenever there is a large CaO gradient that is oriented in a direction opposite to the SiO₂ gradient. This coupling effect is more pronounced in the haplodacitic melt and is likely to be significant in natural magmas of rhyolitic to andesitic compositions. The relative magnitude of coupled chemical diffusion is not very sensitive to changes in temperature and pressure.To a good approximation, the measured diffusion matrices follow well-defined Arrhenius relationships with pressure and reciprocal temperature. Typically, a change in temperature of 100°C results in a relative change in the elements of diffusion matrix of 50 to 100%, whereas a change in pressure of 1 GPa introduces a relative change in elements of diffusion matrix of 4 to 6% for the haplobasalt, and less than 5% for the haplodacite. At a pressure of 1 GPa, the ratios between the major and minor eigenvalues of the diffusion matrix λ₁/λ₂ are not very sensitive to temperature variations, with an average of 5.5 ± 0.2 for the haplobasalt and 3.7 ± 0.6 for the haplodacite. The activation energies for the major and minor eigenvalues of the diffusion matrix are 215 ± 12 and 240 ± 21 kJ mol⁻¹, respectively, for the haplodacite and 192 ± 8 and 217 ± 14 kJ mol⁻¹ for the haplobasalt. These values are comparable to the activation energies for self-diffusion of calcium and silicon at the same melt compositions and pressure. At a fixed temperature of 1500°C, the ratios λ₁/λ₂ increase with the increase of pressure, with λ₁/λ₂ varying from 2.5 to 4.1 (0.5 to 1.3 GPa) for the haplodacite and 4 to 6.5 (0.5 to 2.0 GPa) for the haplobasalt. The activation volumes for the major and minor eigenvalues of the diffusion matrix are 0.31 ± 0.44 and 2.3 ± 0.8 cm³ mol⁻¹, respectively, for the haplodacite and −1.48 ± 0.18 and −0.42 ± 0.24 cm³ mol⁻¹ for the haplobasalt. These values are quite different from the activation volumes for self-diffusion of calcium and silicon at the same melt compositions and temperature. These differences in activation volumes between the two melts likely result from a difference in the structure and thermodynamic properties of the melt between the two compositions (e.g., partial molar volume).Applications of the measured diffusion matrices to quartz crystal dissolution in molten CAS reveal that the activation energy and activation volume for quartz dissolution are almost identical to the activation energy and activation volume for diffusion of the minor or slower eigencomponent of the diffusion matrix. This suggests that the diffusion rate of slow eigencomponent is the rate-limiting factor in isothermal crystal dissolution, a conclusion that is likely to be valid for crystal growth and dissolution in natural magmas when diffusion in liquid is the rate-limiting factor.     

Effects of fO2, fS2, temperature, and melt composition on Fe-Ni exchange between olivine and sulfide liquid: implications for natural olivine-sulfide assemblages

The apparent equilibrium constant for the exchange of Fe and Ni between coexisting olivine and sulfide liquid (K_D = (X_NiS/X_FeS)_liquid/(X_NiSi12O2/X_FeSi12O2)_olivine; X_i = mole fraction) has been measured at controlled oxygen and sulfur fugacities (fO₂ = 10^−8.1 to 10⁻¹⁰ and fS₂ = 10^−0.9 to 10^−1.7) over the temperature range 1200 to 1385°C, with 5 to 37 wt% Ni and 7 to 18 wt% Cu in the sulfide liquid. At log fO₂ of −8.7 ± 0.1, and log fS₂ of −0.9 to −1.7, K_D is relatively insensitive to sulfur fugacity, but comparison with previous results shows that K_D increases at very low sulfur fugacities. K_D values show an increase with the nickel content of the sulfide liquid, but this effect is more complex than found previously, and is greatest at log fO₂ of −8.1, lessens with decreasing fO₂, and K_D becomes independent of melt Ni content at log fO₂ ≤ −9.5. The origin of this variation in K_D with fO₂ and fS₂ is most likely the result of nonideal mixing of Fe and Ni species in the sulfide liquid. Such behavior causes activity coefficients to change with either melt oxygen content or metal/sulfur ratio, effects that are well documented for metal-rich sulfide melts.Application of these experimental results to natural samples shows that the relatively large dispersion that exists in K_D values from different olivine + sulfide-saturated rock suites can be interpreted as arising from variations in fO₂, fS₂, and the nickel content of the sulfide liquid. Estimates of fO₂ based on K_D and sulfide melt composition in natural samples yields a range from fayalite-magnetite-quartz (FMQ)-1 to FMQ-2 or lower, which is in good agreement with previous values determined for oceanic basalts that use glass ferric/ferrous ratios. Anomalously high K_D values recorded in some suites, such as Disko Island, probably reflect low fS₂ during sulfide saturation, which is consistent with indications of low fO₂ for those samples. It is concluded that the variation in K_D values from natural samples reflects olivine-sulfide melt equilibrium at conditions within the T-fO₂-fS₂ range of terrestrial mafic magmas.     

Shear modulus, heat capacity, viscosity and structural relaxation time of Na2O–Al2O3–SiO2 and Na2O–Fe2O3–Al2O3–SiO2 melts

The configurational heat capacity, shear modulus and shear viscosity of a series of Na₂O–Fe₂O₃–Al₂O₃–SiO₂ melts have been determined as a function of composition. A change in composition dependence of each of the physical properties is observed as Na₂O/(Na₂O + Al₂O₃) is decreased, and the peralkaline melts become peraluminous and a new charge-balanced Al-structure appears in the melts. Of special interest are the frequency dependent (1 mHz–1 Hz) measurements of the shear modulus. These forced oscillation measurements determine the lifetimes of Si–O bonds and Na–O bonds in the melt. The lifetime of the Al–O bonds could not, however, be resolved from the mechanical spectrum. Therefore, it appears that the lifetime of Al–O bonds in these melts is similar to that of Si–O bonds with the Al–O relaxation peak being subsumed by the Si–O relaxation peak. The appearance of a new Al-structure in the peraluminous melts also cannot be resolved from the mechanical spectra, although a change in elastic shear modulus is determined as a function of composition. The structural shear-relaxation time of some of these melts is not that which is predicted by the Maxwell equation, but up to 1.5 orders of magnitude faster. Although the configurational heat capacity, density and shear modulus of the melts show a change in trend as a function of composition at the boundary between peralkaline and peraluminous, the deviation in relaxation time from the Maxwell equation occurs in the peralkaline regime. The measured relaxation times for both the very peralkaline melts and the peraluminous melts are identical with the calculated Maxwell relaxation time. As the Maxwell equation was created to describe the timescale of flow of a mono-structure material, a deviation from the prediction would indicate that the structure of the melt is too complex to be described by this simple flow equation. One possibility is that Al-rich channels form and then disappear with decreasing Si/Al, and that the flow is dominated by the lifetime of Si–O bonds in the Al-poor peralkaline melts, and by the lifetime of Al–O bonds in the relatively Si-poor peralkaline and peraluminous melts with a complex flow mechanism occurring in the mid-compositions. This anomalous deviation from the calculated relaxation time appears to be independent of the change in structure expected to occur at the peralkaline/peraluminous boundary due to the lack of charge-balancing cations for the Al-tetrahedra.     

Vanadium in peridotites as a proxy for paleo-fO2 during partial melting : prospects, limitations, and implications

The compatibility of vanadium (V) during mantle melting is a function of oxygen fugacity (fO₂): at high fO₂’s, V becomes more incompatible. The prospects and limitations of using the V content of peridotites as a proxy for paleo-fO₂ at the time of melt extraction were investigated here by assessing the uncertainties in V measurements and the sensitivity of V as a function of degree of melt extracted and fO₂. V-MgO and V-Al₂O₃ systematics were found to be sensitive to fO₂ variations, but consideration of the uncertainties in measurements and model parameters indicates that V is sensitive only to relative fO₂ differences greater than ∼2 log units. Post-Archean oceanic mantle peridotites, as represented by abyssal peridotites and obducted massif peridotites, have V-MgO and -Al₂O₃ systematics that can be modeled by 1.5 GPa melting between FMQ − 3 and FMQ − 1. This is consistent with fO₂’s of the mantle source for mid-ocean ridge basalts (MORBs) as determined by the Fe³⁺ activity of peridotitic minerals and basaltic glasses. Some arc-related peridotites have slightly lower V for a given degree of melting than oceanic mantle peridotites, and can be modeled by 1.5 GPa melting at fO₂’s as high as FMQ. However, the majority of arc-related peridotites have V-MgO systematics overlapping that of oceanic mantle peridotites, suggesting that although some arc mantle may melt under slightly oxidizing conditions, most arc mantle does not. The fact that thermobarometrically determined fO₂’s in arc peridotites and lavas can be significantly higher than that inferred from V systematics, suggests that V retains a record of the fO₂ during partial melting, whereas the activity of Fe³⁺ in arc peridotitic minerals and lavas reflect subsequent metasomatic overprints and magmatic differentiation/emplacement processes, respectively.Peridotites associated with middle to late Archean cratonic mantle are characterized by highly variable V-MgO systematics. Tanzanian cratonic peridotites have V systematics indistinguishable from post-Archean oceanic mantle and can be modeled by 3 GPa partial melting at ∼FMQ − 3. In contrast, many South African and Siberian cratonic peridotites have much lower V contents for a given degree of melting, suggesting at first glance that partial melting occurred at high fO₂’s. More likely, however, their unusually low V contents for a given degree of melting may be artifacts of excess orthopyroxene, a feature that pervades many South African and Siberian peridotites but not the Tanzanian peridotites. This is indicated by the fact that the V contents of South African and Siberian peridotites are correlated with increases in SiO₂ content, generating data arrays that cannot be modeled by partial melting but can instead be generated by the addition of orthopyroxene through processes unrelated to primary melt depletion. Correction for orthopyroxene addition suggests that the South African and Siberian peridotites have V-MgO systematics similar to those of Tanzanian peridotites. Thus, if the Tanzanian peridotites represent the original partial melting residues, and if the South African and Siberian peridotites have been modified by orthopyroxene addition, then there is no indication that Archean cratonic mantle formed under fO₂’s significantly greater than that of modern oceanic mantle. Instead, the fO₂’s inferred from the V systematics in these three cratonic peridotite suites are within range of modern oceanic mantle. This also suggests that the transition from a highly reducing mantle in equilibrium with a metallic core to the present oxidized state must have occurred by late Archean times.     

Partition coefficients of Nb and Ta between rutile and anhydrous haplogranite melts

Partition coefficients (D) for Nb and Ta between rutile and haplogranite melts in the K₂O-Al₂O₃-SiO₂ system have been measured as functions of the K₂O/Al₂O₃ ratio, the concentrations of Nb₂O₅ and Ta₂O₅, the temperature, in air and at 1 atmosphere pressure. The Ds increase in value as the K^* [K₂O/(K₂O + Al₂O₃)] molar ratio continuously decreases from highly peralkaline [K^* ∼ 0.9] to highly peraluminous [K^* ∼ 0.35] melts. The D values increase more dramatically with a unit decrease in K^* in peraluminous melts than in peralkaline melts. This compositional dependence of Ds can be explained by the high activity of NbAlO₄ species in peraluminous melts and the high activity of KONb species (or low activity of NbAlO₄ species) in peralkaline melts. A coupled substitution, Al⁺³ + Nb⁺⁵ (or Ta⁺⁵) = 2Ti⁺⁴, accounts for the Ds of Nb (Ta) being much greater in peraluminous melts than in peralkaline melts because this substitution allows Nb (Ta) to enter into the rutile structure more easily. The Ds of Ta between rutile and melt are greater than those of Nb at comparable concentrations because the molecular electronic polarizability of Ta is weaker than that of Nb. The Nb⁺⁵ with a large polarizing power forms a stronger covalent bond with oxygen than Ta⁺⁵ with a small polarizing power. The formation of the strong bond, Nb-O, distorts the rutile structure more severely than the weak bond, Ta-O; therefore, it is easier for Ta to partition into rutile than for Nb. These results imply that the utilization of the Nb/Ta ratio in liquid as a petrogenetic indicator in granitic melts must be done with caution if rutile (or other TiO₂-rich phases) is a liquidus phase. The crystallization of rutile will increase the Nb/Ta ratio of the residual liquid because the Ds of Ta between rutile and melts are greater than those of Nb. Received: 28 December 1998 / Accepted 27 September 1999