This paper reports the thermodynamic analysis of NixFe1-x oxidation for awaruite, Ni3Fe, which is widespread in serpentinized ultrabasic rocks, and other Ni–Fe phases over a wide range of temperatures (400–1873 K) and crustal pressures up to 2 kbar. It was found that the equilibrium γ(NixFe1 – x)–Fe oxides constrains oxygen fugacity (fO2) at the NNO–IW range. For γ(NixFe1 – x) with an iron mole fraction x ≥ 0.5, fO2 approaches the IW buffer. The reaction of kamacite, α(Ni0.05Fe0.95), with oxygen is close to the IW–IM buffer reactions. The fO2 values of awaruite preservation are no higher than ∆QFM = –(7.8–5.2) at T = 400–600 K and increase to ∆QFM = –(2.7–2.0) at T > 1000–1200 K. The obtained approximations of the temperature dependence of fO2 can be used to estimate the redox conditions in low-sulfide systems containing NixFe1 – x under the conditions of the Earth’s crust.

At high PT parameters of cumulates of ultramafic–mafic intrusions and low oxygen fugacity (below QFM buffer), Pt in the form of a carbonyl complex is dissolved in a CO2-bearing fluid. The high solubility of Pt chloride in brines with NaCl, which is related to the formation of low-sulfide deposits of platinum group elements, is attained only at high oxygen fugacity (above the QFM buffer). It is suggested that native platinum at low oxygen fugacity in low-Н2О СО–СО2 (Н2О) fluid can also transit into a cation-soluble form due to the reaction of chloration. The experimental data are provided for the interaction of NaCl with magnetite and chromite (accessory minerals of ultramafic–mafic intrusions) at Р = 200 MPa, Т = 950°С, and fO2

Raman spectroscopic data of quenching phases in experiments on the dissolution of Pt in reduced carbonic fluid, containing about 30 mol

At high PT parameters of the cumulates of ultramafic-mafic intrusions at low fO2 (below the QFM buffer), platinum dissolves in the fluid with CO as a carbonyl complex of the native metal. The high solubility of platinum as PtCl2 in brines with NaCl, which is associated with the formation of low-sulfide PGE deposits, is achieved at high oxygen fugacity (above the NNO buffer). It is assumed that at low oxygen fugacity in the low water CO–CO2 fluid, native Pt can also be converted into a cation-soluble form by chlorination. Experimental data (Р = 200 MPa, Т = 950oC, fO2 QFM and fluid CO–CO2) on the reaction of NaCl with magnetite and chromite, accessor minerals of mafic-ultramafic intrusions, with the formation of iron and chromium chlorides are presented. As shown by thermodynamic calculations, the equilibrium in the FeCl3–FeCl2 pair provides the high chlorine fugacity (fCl2). This fugacity is only 3–4 orders of magnitude lower than fCl2 in the Pt–PtCl2 equilibrium and 2.5–3 orders of magnitude higher than in the aqueous fluid 1 M HCl at the same P–T–fO2 parameters.

The paper presents materials on experimental study of pargasite stability. On the example of calcic amphibole, experimental modeling of the processes occurring in the conditions of volcanic hearth at pressures up to 5 kbar was carried out. The phase diagram of pargasite was clarified. The occurring reactions and their parameters are revealed. Based on the experimental data obtained, the stability of pargasite is controlled by three reactions. The first reaction takes place in the area of low water pressure less than 1 kbar – dehydration reaction:Prg = Fo + Sp + Di + Ne + An +H2O. The second reaction takes place in the area of water pressure more than 1.2-1.5 kbar and temperature around 1100°C. Pargasite decomposition is controlled by incongruent melting:Prg = Fo + Sp +{Di+Ne +An}L+H2O. The third reaction takes place in the same pressure range as the previous one, but at lower temperatures ~1050°C. This reaction determines the pargasite liquidus in the melt and is related to the interaction between the amphibole and the coexisting melt:Prg + L= Fo + Sp + Di +{Ne +Pl}L+H2O.Presumably, the activity of the melt silicaaSiO2has the greatest impact on the pargasite liquidus.

The paper presents materials on experimental study of pargasite stability. On the example of calcic amphibole, experimental modeling of the processes occurring in the conditions of volcanic hearth at pressures up to 5 kbar was carried out. The phase diagram of pargasite was clarified. The occurring reactions and their parameters are revealed. Based on the experimental data obtained, the stability of pargasite is controlled by three reactions. The first reaction takes place in the area of low water pressure less than 1 kbar – dehydration reaction:Prg = Fo + Sp + Di + Ne + An +H2O. The second reaction takes place in the area of water pressure more than 1.2-1.5 kbar and temperature around 1100°C. Pargasite decomposition is controlled by incongruent melting:Prg = Fo + Sp +{Di+Ne +An}L+H2O. The third reaction takes place in the same pressure range as the previous one, but at lower temperatures ~1050°C. This reaction determines the pargasite liquidus in the melt and is related to the interaction between the amphibole and the coexisting melt:Prg + L= Fo + Sp + Di +{Ne +Pl}L+H2O.Presumably, the activity of the melt silicaaSiO2has the greatest impact on the pargasite liquidus.

Flushing of hydrous silicic magmas with crustal carbonic fluid may be an important factor controlling the dynamics of rhyolitic eruptions. We present combined theoretical and experimental study of the interaction of carbonic fluid with a hydrous silicic melt. The process of diffusional equilibration of a CO2 bubble with a silicic melt was simulated numerically in the spherical shell approximation. The rapid water transfer from the melt to the bubble is followed by a slower diffusion of CO2 into the melt. The water distribution in the melt becomes almost uniform over a period proportional to the diffusional unit of time 0.14τw, determined by the initial inter-bubble distance W equal the distance between neighbor bubbles centers and the water diffusion coefficient Dw in the melt (τw = W 2/Dw), while the CO2 distribution remains strongly contrasting and the melt remains undersaturated in CO2. This process was modelled experimentally with a hydrous albite melt at P = 200MPa and T = 950–1000 °C. In the first series of experiments at T = 950◦C, a glass powder was filled with pure CO2 at the beginning of the experiment, forming numerous bubbles at the run temperature. Micro-FTIR measurements showed that after 40 minutes the water content in the melt decreased from 4.9 down to 1.8 wt. % with the maximum CO2 content of 500 ppm (below saturation). After 4 hours, the crystallinity increased to 85%, and almost all of the fluid bubbles escaped. The second series of experiments CO2 interacted with a 2 mm high column of hydrous albite melt. Diffusion profiles in the quenched glass were measured using EMPA (H2O) and micro-FTIR (CO2 and H2O). The estimated diffusion coefficients in the melt for H2O (1.1 × 10−6 cm2 /s) and CO2 (1.5 × 10−7 cm2 /s) are consistent with published data. Scaling analysis predicts that in the nature, after the influx of CO2 bubbles a few millimeters in size, the maximum dehydration of rhyolitic magma with viscosity near 105 Pa s without a significant increase in CO2 content occurs after 1–30 days, i.e. a period compatible with the minimum duration of pre-eruption processes in the magma chamber.

Formation of graphite was observed in experiments on synthesis of dry carbon-bearing albite glasses in platinum capsules in an Internally Heated Pressure Vessel at 500 MPa and Т = 1200–1250°С. A thermodynamic model is proposed that explains the achievement of low oxygen fugacity near QFM-2 in the melt at low fugacity of hydrogen formed due to the decomposition of trace amounts of water in a compression medium (Ar gas). The unexpectedly low fugacity of oxygen is explained by the shift of equilibrium between the gases dissolved in the melt CO2 + H2 = H2O + CO to the right due to the low activity of molecular water at a low total content of H2O 0.1–0.5 wt

The paper reports results of an experimental study of amphibole crystallization from the highly magnesian andesite melt of Shiveluch volcano, Kamchatka. The experiments were carried out in IHPV at 300 MPa and 940–980°С in iron-saturated platinum capsules, using rapid quenching and temperature oscillations (in some experiments). The redox state of iron in the system was measured before and after the experiments using Mössbauer spectroscopy. The maximum size of the experimental amphibole crystals (up to 200 μm) was close to those of natural amphibole phenocrysts in the volcanic rocks of Shiveluch volcano. The experimental data show that the content of octahedrally coordinated Al (Al6) in the amphibole considerably varies with small variations in the intensive parameters (P, T, and fO2) and composition of the melt, and the maximum Al6 concentration can be evaluated only by using a reasonably large dataset of amphibole analyses. A modified 13eCNK method is suggested to calculate the values of Al6 and Fe3+/Fe2+ with regard for the Ti concentration and the probable partial transfer of Mg into site B in high-Mg amphibole. Calculations with this modified technique yield lower Fe3+/Fe2+ and higher Al6 values. Our experimental data show that the temperature of amphibole liquidus crystallization decreases from about 990 to 960°C when the oxygen fugacity drops from NNO + 1.5 to NNO + 0.4. In view of this, the transition from amphibole-bearing to anhydrous mineral assemblage in the magmas of Shiveluch volcano might have been caused by variations of the oxygen fugacity but not water. The application of our geobarometer to amphiboles from Shiveluch volcano (extrusions Krasnaya and Karan) yields the highest pressure estimate of above 1 GPa, corresponding to the P-T conditions of the melting of garnet-bearing amphibolite in the lower crust.

Fluoro-sodalite was synthesized for the first time at temperatures of 400–800°C and H2O pressures of 1–2 kbar in the Si–Al–Na–H–O–F system. X-ray diffraction and infrared spectroscopic investigations showed that fluorine is incorporated in the sodalite structure as anionic octahedral groups, [AlF6]3–, the number of which can vary from 0 to 1. Correspondingly, the end-members of the F-sodalite series are Na7(H2O)8[Si5Al7O24] and Na8(AlF6)(H2O)4[Si7Al5O24]. Depending on the composition of the system, F-sodalite associates at 500–650°C with nepheline, albite, cryolite, and villiaumite, which are joined by analcime below 500°C and aluminosilicate melt above 650°C. Fluorine-bearing sulfate–chlorine-sodalite was found for the first time in a pegmatite sample from the Lovozero massif. The highest fraction of the fluorine end-member in natural sodalite is 0.2. The incorporation of F into the sodalite structure requires much more energy compared with Cl– and SO 4 2- , because it is accompanied by a structural rearrangement and a transition from tetrahedral Al to octahedral Al.

It is shown experimentally that fluid mostly consisting of CO2 and CO is a good solvent for many petrogenic and ore components under high P–T parameters. Siderite decomposed to oxides and a mixture of CO2 and CO was the source of the fluid in the experiment. It is demonstrated that an ultrapotassic alkaline melt was formed on the oxide matrix as a result of fluid transport of the components of the basaltic melt at 2–5 kbar and 900–1000°C. This melt is characterized by high concentrations of Rb, LREE, Sr, and Ba. The mechanism of solubility of metals in the carbon dioxide–reduced fluid includes the formation of volatile carbonyls. As is evident from thermodynamic modeling, the stability of Ni(CO)4 at high pressure reaches magmatic temperatures. The finding of native nickel in aerosols during eruption of the Tolbachik Volcano in 2012–2013 is explained by decomposition of carbonyl from fluid of a deep origin during magma ascent.

The Zr-Hf geochemical indicator, i.e., the Zr/Hf ratio (in wt %) in granitic rocks is proposed to be used as the most reliable indicator of the fractionation and ore potential of rare-metal granites. It was empirically determined that the fractional crystallization of granitic magma according to the scheme granodiorite → biotite granite → leucogranite → Li-F granite is associated with a decrease in the Zr/Hf ratio of the granites. The reason for this is the stronger affinity of Hf than Zr to granitic melt. This was confirmed by experiments on Zr and Hf distribution between granitic melt and crystals of Hf-bearing zircon (T = 800°C, P= 1 kbar). The application of the Zr/Hf indicator was tested at three classic territories of rare-metal granites: eastern Transbaikalia, central Kazakhstan, and the Erzgebirge in the Czech Republic and Germany. The reference Kukul’bei complex of rare-metal granites in eastern Transbaikalia (J3) is characterized by a uniquely high degree of fractionation of the parental granitic melt, with the granites and their vein derivatives forming three intrusive phases. The biotite granites of phase 1 are barren, the leucogranites of phase 2 are accompanied by greisen Sn-W mineral deposits (Spokoininskoe and others), and the final dome-shaped stocks of amazonite Li-F granites of phase 3 host (in their upper parts) Ta deposits of the “apogranite” type: Orlovka, Etyka, and Achikan. The Kukul’bei Complex includes also dikes of ongonites, elvanes, amazonite granites, and miarolitic pegmatites. All granitic rocks of the complex are roughly coeval and have an age of 142±0.6 Ma. The Zr/Hf ratio of the rocks systematically decreases from intrusive phase 1 (40–25) to phases 2 (20–30) and 3 (10–2). Compared to other granite series, the granites of the Kukul’bei Complex are enriched in Rb, Li, Cs, Be, Sn, W, Mo, Ta, Nb, Bi, and F but are depleted in Mg, Ca, Fe, Ti, P, Sr, Ba, V, Co, Ni, Cr, Zr, REE, and Y. From earlier to later intrusive phases, the rocks become progressively more strongly enriched or depleted in these elements, and their Zr/Hf ratio systematically decreases from 40 to 2. This ratio serves as a reliable indicator of genetic links, degree of fractionation, and rare-metal potential of granites. Greisen Sn, W, Mo, and Be deposits are expected to accompany granites with Zr/Hf < 25, whereas granites related to Ta deposits should have Zr/Hf < 5.

The concept of granitic melt fractionation as the main process in the concentration of rare elements in granites calls for the development of a reliable method to determine the evolutionary sequences of granite series. We propose to use for this purpose a zirconium-hafnium indicator, the Zr/Hf weight ratio in granitic rocks (Zaraisky et al., 1999, 2000). By the example of three classic regions of rare-metal deposits, eastern Transbaikalia, central Kazakhstan, and Erzgebirge (Czech Republic and Germany), it was empirically shown that the Zr/Hf ratio of granites decreases during the fractional crystallization of granite magmas in the sequence granodiorite → biotite granite → leucogranite → lithium-fluorine granite. The reason is the higher affinity of Hf compared with Zr to a granite melt. This implies that the crystallization and settling of accessory zircon will cause the progressive enrichment of Hf relative to Zr in the residual melt. As a result, the Zr/Hf ratio decreases regularly in the series of sequential phases of granite intrusion related to a single magma chamber from granodiorite to biotite granite, leucogranite, and Li-F granite (from 45-30 to 10-2). Our experimental investigations supported the preferential enrichment of haplogranite melt in Hf and zircon crystals in equilibrium with melt in Zr ( T = 800°C and P = 1 kbar). The Zr/Hf indicator was tested by the example of the wellknown Kukulbei rare-metal granite complex of eastern Transbaikalia (J3), which is unique in the degree of fractionation of initial granite melt with the formation of three phases of granite emplacement and vein derivatives. An important feature of the complex is its “short” differentiation trend. It was supposed that the granite magma of the first phase is parental, and the later phases forming small intrusive bodies in large massifs of biotite granites of the first phase are sequential products of its crystallization differentiation in a magma chamber. The biotite granites of the first phase are barren. The leucocratic granites of the second phase are accompanied by tin-tungsten greisen deposits (e.g., Spokoininskoe), and the upper part of cupola-like stocks of Li-F amazonite granites of the third phase host apogranite-type tantalum deposits (Orlovka, Etyka, and Achikan). In addition to three granite phases, the Kukulbei complex includes dikes of ongonites, elvans, amazonite granites, and chamber miarolitic pegmatites. All of the granitic rocks of the complex have similar isotopic ages of 142± 0.6 Ma. The Zr/Hf ratio decreases systematically from phase 1 (40–25), to phase 2 (20–10), and phase 3 (10–2). The ongonites, elvans, and pegmatites have similar Zr/Hf ratios (15-5), falling between the ranges of leucocratic muscovite granites and Li-F granites. Compared with other granite series, the granitic rocks of the Kukulbei complex show specific petrographic and geochemical features: they are strongly enriched in Rb, Li, Cs, Be, Sn, W, Mo, Ta, Nb, Bi, and F but depleted in Mg, Ca, Fe, Ti, P, Sr, Ba, V, Co, Ni, Cr, Zr, REE, and Y. From the early to late intrusion phases, the degree of enrichment and depletion in these element groups increases regularly. This is accompanied by a significant decrease (from 40 to 2) in Zr/Hf, which can be used as a reliable indicator of genetic relations, degree of fractionation, and rare-metal potential of granites. Granites with Zr/Hf values lower than 25 are promising for prospecting for Sn, W, Mo, and Be greisen deposits, whereas the formation of Ta deposits requires Zr/Hf values lower than 10.

The data obtained on the sodic part of the SiO2-Al2O3-Na2O-K2O system with F at 800°C and 1 kbar provide the basis for constructing a phase diagram showing the region of an aluminosilicate melt. In this system, oxide and fluoride phases are identified that control the stability field of the melt and the solubility of F. Liquid immiscibility was detected in aluminous nepheline-and quartz-normative Li-bearing compositions (the latter compositions are characterized by a wider immiscibility field). Solubility of F was determined in an aluminosilicate melt saturated with respect to F, i.e., coexisting with phases rich in this element. The F concentrations in the glasses range from 2 to 20 wt %. The quartz-normative glasses are poorer in F (no more than 5 wt % F) than the nepheline-normative glasses (which contain mostly 5–10 wt % F). The maximum F concentrations (> 10 wt %) in the phase diagram lie on both sides of the albite composition point in the region of ultragpaitic nepheline-normative melts and in the region of normal syenite melts. Changes in the phase relations when Na is substituted for K were determined in the quartz-normative silicate melt.