Cold-seal Vessels Research Articles

Rapid Ag diffusion in granitic melt: Implications for Ag mineralization and melt inclusion record

Element diffusion in silicate melts is an important process in the transport and enrichment of metals in magmatic and magmatic-hydrothermal systems, but the diffusivity of Ag, unlike other metals such as Cu, Mo, and W, is still unknown. In this study, Ag diffusion in granitic melts with up to 6.9 wt% H2O was investigated at 973–1673 K and 0.1–1 GPa in a piston cylinder apparatus and a cold seal vessel, using both a diffusion couple approach and an argentite (Ag2S) dissolution approach. Silver concentration profiles were analyzed by LA-ICP-MS on recovered samples. Silver solubility, determined by Ag concentration at melt-argentite interface, increases with increasing temperature. Silver diffusivity is found to increase with increasing temperature and water content but to decrease with pressure. At 1 GPa, Ag diffusivity in granitic melt can be described by the following expression:lnDAg,1GPa=-13.77±0.46-13341±534-700±39wTwhere D is Ag diffusivity in m2/s, w is the water content in wt%, and T is the temperature in K. The pressure effect over 0.1–1 GPa gives an activation volume of 11.3 cm3/mol for Ag diffusion. While not being on a par with the mobility of Na and Li, Ag diffusion is among the most rapid, comparable to Cu and faster than other species including H2O and Fe. This implies that Ag diffusion in melt is unlikely to be a rate-limiting step for Ag enrichment into exsolved fluid bubbles or sulfide blobs. Furthermore, melt inclusions in silicate minerals are expected to record Ag concentration in the bulk melt instead of an elevated concentration in the presence of an Ag-rich boundary layer. Our Ag diffusivity data also suggest that a few tens of hours are long enough to homogenize Ag concentration in melt in solubility or partition experiments at temperatures above 800 °C.

Effect of gold and magnetite on the decomposition kinetics of formic acid at 200 °C under hydrothermal conditions

Formic acid and formate ions are intermediate compounds in the water gas shift reaction as well as in the abiotic formation of organic molecules from dissolved CO2 and H2 under hydrothermal conditions. The decomposition kinetics of a 0.1 M formic acid solution has been investigated using PTFE-lined reactors at 200 °C on the liquid-vapour equilibrium of water for calculated in-situ pH comprised between 2.8 and 3.6. Under these conditions, formic acid is the dominant aqueous species. Formic acid aqueous decomposition follows a first-order kinetics with a constant k (s−1) such that ln(k) = −13.6 ± 0.6 at 200 °C. The addition of gold chips in the reactor is found to promote formic acid decomposition and the kinetics constant of the reaction is a linear function of the exposed gold surface by unit volume of solution (S/V). Consequently, experiments run in gold tubes as those typically used in cold-seal vessels can be by more than two orders of magnitude faster than those performed in PTFE reactors under the same conditions. The addition in the PTFE-lined reactor, of particles of magnetite (Fe3O4) with sizes centred on 300 nm or the presence of aqueous Fe at the millimolal level does not catalyse the hydrothermal decomposition of formic acid at 200 °C.

An experimental investigation of the role of microfracture surfaces in controlling quartz precipitation rate: Applications to fault zone diagenesis

We present the results of quartz growth experiments, which were designed to assess the role of microfracture surfaces in controlling quartz-precipitation rates during fault-zone diagenesis. Experiments were run in hydrothermal cold-seal vessels at 300–450 °C and 150 MPa confining pressure for up to 1344 h. Microfractures routinely form at grain contacts during these experiments. Microfracture kinematic-aperture distributions indicate that microfractures form within the first 48 h of each experiment. Regardless of experimental temperature or duration, microfracture-sealing cements account for approximately the same amount of new quartz cement in each experiment. With increasing experimental duration, sealed microfractures were progressively overgrown by grain-boundary overgrowth cements. Spatial and temporal trends in the distribution of overgrowth- and microfracture-sealing cements indicate that precipitation rates on newly formed microfractures greatly exceed those on detrital-grain boundaries. This effect persists regardless of natural iron-oxide grain coatings present in a subset of our experiments. While our results agree with previous research that demonstrated increased growth rates on fracture surfaces in faults in fully lithified rock, fundamental differences in the nature of deformation in our experiments provide insight into quartz cementation in cataclastic deformation bands in faults offsetting high-porosity sandstones.

The system Al2O3-P2O5-H2O at temperatures below 200 C: Experimental data on the stability of variscite and metavariscite AlPO4{middle dot}2H2O

The system Al2O3-P2O5-H2O contains many phosphate minerals that occur in various geologic environments. The natural occurrence of variscite (AlPO4·2H2O, orthorhombic), including its monoclinic polymorph metavariscite, is largely restricted to soils and aluminous rocks like Al-rich igneous rocks and shales, which interacted with P-rich hydrothermal solutions or groundwater. Variscite dehydrates to berlinite (AlPO4) and a hydrous P-Al-rich fluid. This dehydration reaction AlPO4·2H2O = AlPO4 + 2H2O is, however, metastable, at low concentrations of P in the fluid, because berlinite breaks down at lower temperature to augelite [Al2PO4(OH)3] + H3PO4 + H2O and trolleite [Al4(PO4)3(OH)3] + H3PO4 + H2O. The variscite/metavariscite = berlinite equilibrium has been investigated by synthesis experiments from mixtures of γ-Al2O3 and excess phosphoric acid at pressures between 1 and 5 kbar and temperatures of 100–200 °C using standard cold-seal vessels. The hydration-dehydration equilibrium is mainly controlled by temperature and only weakly by pressure. At 4 and 5 kbar, variscite/metavariscite were found at temperatures of ~150 °C, at lower pressure at 115–125 °C in accordance with their natural mode of occurrence. In this T range, however, variscite/metavariscite is not the sole phase but is always accompanied by variable but generally minor amounts of wavellite and trolleite Al4(PO4)3(OH)3. Secondary wavellite Al3(PO4)2(OH)3·5H2O, hydrated Al-phosphate AlPO4· x H2O (1.1 ≤ x ≤ 1.3) and Al-metaphosphate hexahydrate Al(H2PO4)3 formed during quenching and/or drying of the runs. Berlinite is the reaction product at temperatures 200 °C/4–5 kbar and 150 °C/1–3 kbar and may be associated with augelite and trolleite.

Mechanisms of hydrogen incorporation and diffusion in iron-bearing olivine

The incorporation and diffusion of hydrogen in San Carlos olivine (Fo90) single crystals were studied by performing experiments under hydrothermal conditions. The experiments were carried out either at 1.5 GPa, 1,000°C for 1.5 h in a piston cylinder apparatus or at 0.2 GPa, 900°C for 1 or 20 h in a cold-seal vessel. The oxygen fugacity was buffered using Ni–NiO, and the silica activity was buffered by adding San Carlos orthopyroxene powders. Polarized Fourier transform infrared (FTIR) spectroscopy was utilized to quantify the hydroxyl distributions in the samples after the experiments. The resulting infrared spectra reproduce the features of FTIR spectra that are observed in olivine from common mantle peridotite xenoliths. The hydrogen concentration at the edges of the hydrogenated olivine crystals corresponds to concentration levels calculated from published water solubility laws. Hydrogen diffusivities were determined for the three crystallographic axes from profiles of water content as a function of position. The chemical diffusion coefficients are comparable to those previously reported for natural iron-bearing olivine. At high temperature, hydrogenation is dominated by coupled diffusion of protons and octahedrally coordinated metal vacancies $$ {\left( {V^{\prime\prime}_{{{\text{Me}}}} } \right)}, $$ where the vacancy diffusion rate limits the process. From the experimental data, we determined the following diffusion laws (diffusivity in m2 s−1, activation energies in kJ mol−1): $$ D_{{V_{{{\text{Me}}}} [100],[010]}} = 10^{{ - 4.5 \pm 4.1}} \exp [ - (204 \pm 94)/RT] $$ for diffusion along [100] and [010]; $$ D_{{V_{{{\text{Me}}}} [001]}} = 10^{{ - 1.4 \pm 0.5}} \exp [ - (258 \pm 11)/RT] $$ for diffusion along [001]. These diffusion rates are fast enough to modify significantly water contents within olivine grains in xenoliths ascending from the mantle.

Lithium isotopes in the system Qz-Ms-fluid:An experimental study

The fractionation of lithium isotopes among quartz, muscovite, and a chloride-bearing aqueous fluid has been investigated experimentally at 400°–500°C and 50–100 MPa. Experiments were performed for 15–60 days in cold seal vessels with natural mineral specimens. Lithium was introduced primarily through the fluid, which also contained KCl and HCl. In most runs, the fluid was prepared with the L-SVEC standard (δ 7Li = 0) and was 1 M in total chloride with K/Li/H = 100/10/1. In two experiments, a 6Li spike was employed. The experiments demonstrate that quartz and muscovite are susceptible to pronounced, rapid shifts in Li isotopic composition by diffusion through interaction with a Li-bearing fluid, particularly at 500°C. At 500°C, fractionation factors were determined to be Δ quartz-aqueous ≅ +8 to + 12‰ and Δ muscovite-aqueous ≅ +18 to + 20‰. An intermineral fractionation factor is given by Δ muscovite-quartz ≈ +9‰. At 400°C, the results suggest Δ quartz-aqueous ≈ +4 to + 6‰. The study provides evidence of systematic fractionation in lithium isotopes at the temperatures of some magmatic processes, such as those associated with porphyry-type ore systems and pegmatites.

Experimental transformation of Na,Ca-smectite under basic conditions at 150 °C

The aim of this study is to examine the effect of various high pH solutions on smectites. The starting materials were a Na,Ca-smectite and two homoionic Na- and Ca-smectites. The experimental solutions were 0.01 m NaOH (pH 12), 1 m K 2CO 3 and mixtures of 1 or 3 m KCl with 10 −4 or 10 −2 m KOH (pH 10 or 12). To enhance the reaction rate, all experimental investigations were carried out at 150 °C (±5°C max.), during 2 months. The internal pressure was 5 bars (liquid–vapour equilibrium pressure) when experiments were conducted in warm seal autoclaves and fixed to 150 bars, using standard cold seal vessels. The liquid/clay mass ratio was fixed to 10/1. Integrated data obtained by a multitechnique analytical approach [XRD, electron microprobe, transmission electron microscopy (TEM), scanning electron microscope (SEM) on run products, and inductively coupled plasma atomic emission (ICP-AES) and mass spectrometry (ICP-MS) on experimental solutions] show a distinct behaviour of smectite as a function of experimental fluid composition. In NaOH solution, run products remain low-charge smectites. In KCl+KOH solutions, crystal chemical changes concern mostly the interlayer. A significant replacement of Na by K, and a partial substitution of Ca by K are observed. The tetrahedral Si content does not change, but the total interlayer charge in smectite increases, balancing the changes occurring in the octahedral occupancy (release of Al and Fe). Low-charge smectite transforms into high-charge smectite and byproducts are quartz and feldspars. In the presence of 1 m K 2CO 3, homoionic Na- and Ca-smectites are unstable and in part replaced by zeolites (merlinoite), feldspars and calcium silicate hydrates (CSH), the latter being dominated by a tobermorite-like phase. The main consequences of the cement or concrete like pore fluid–bentonite contact at elevated temperature induce dissolution of smectite (at various rates), cation exchange and the formation of byproducts (quartz, feldspars, zeolites and CHS). These mineralogical changes affect the expandability and surface area of minerals, and therefore may significantly change the porosity and fluid flow diffusion in clay barriers exposed to high pH fluids.

Water solubility in carbonatite melts

Water solubility in Ca-Na-Mg carbonate melt was measured at 900 °C from 0.25 to 2.25 kbar using a modified double-capsule technique and rapid-quench cold-seal vessels. The data can be described by the empirical relationship W = 8.705 P 0.635 , where W is water solubility in weight percent and P is pressure in kbar. The relative error in water solubility calculated by this equation is estimated to be about 15%. Already at 1 kbar, water solubility reaches nearly 10 wt%, which is two to three times the value observed for most silicate melts under similar conditions. The exceptionally high water solubility in carbonatite magmas together with the preferential partitioning of alkalis into a coexisting fluid phase may explain the massive aqueous metasomatism (fenitisation) usually observed around intrusive carbonatite bodies.

Water diffusion in synthetic iron-free forsterite

The kinetics of hydrogenation of dry synthetic forsterite single crystals was determined by performing experiments under hydrothermal conditions. The experiments were performed at 1.5 GPa, 1000 °C for 3 h in a piston-cylinder apparatus, or at 0.2 GPa, 900–1110 °C, for 3–20 h in TZM cold-seal vessels. The oxygen fugacity was buffered using Fe–FeO or Ni–NiO powders. Polarized Fourier transform infrared spectroscopy was utilized to quantify the hydroxyl distributions in the samples after the experiments. Hydrogenation rates were measured parallel to the three crystallographic axes from profiles of water content as a function of position in the samples. The chemical diffusion coefficients are marginally slower than in natural iron-bearing olivine for the same diffusion process, but the anisotropy of diffusion is the same, with the [001] axis the fastest direction of diffusion and [100] the slowest. Fits of the diffusion data to an Arrhenius law yield similar activation energies for each of the crystallographic axes; a global fit to all the diffusion data gave 211 ± 18 kJ mol−1, in reasonable agreement with the previous results for natural olivine. Thus hydrogenation most likely occurs by coupled diffusion of protons and octahedrally coordinated metal vacancies. The diffusion rates are fast enough to modify water contents within xenoliths ascending from the mantle, but probably too slow to permit a total equilibration of forsterite or olivine crystals.

H 2O diffusion in rhyolitic melts and glasses

H 2O diffusion plays a major role in bubble growth and volcanic eruption. We report a comprehensive study of H 2O diffusion in rhyolitic melts and glasses. This new study and previous investigations together cover a wide range of conditions: 400–1200°C, 0.1–810 MPa, and 0.1–7.7 wt.% total H 2O content (H 2O t). In order to constrain how the diffusivity depends on H 2O t, both the diffusion-couple experiments and the dehydration experiments are carried out in a cold-seal vessel (CSV), an internally heated pressure vessel, and a piston cylinder. H 2O concentration profiles are measured by infrared (IR) spectroscopy. Although there are still some experimental and analytical difficulties, our data represent a major improvement over earlier data. The diffusion data have been used to quantify H 2O diffusivity as a function of temperature, pressure, and H 2O t. Assuming that molecular H 2O (H 2O m) is the diffusing species, the H 2O m diffusivity (in μm 2/s) can be expressed as: D H 2 O m = exp[(14.08−13,128/T−2.796P/T)+(−27.21+36,892/T+57.23P/T)X], where T is in Kelvin, P is in mPa, and X is the mole fraction of H 2O t on a single oxygen basis. The pressure dependence is not so well-resolved compared to the dependence on T and X. The dependence of D H 2O m on X increases with increasing pressure. The results are consistent with the data of Nowak and Behrens (1997) [Nowak, M., Behrens, H., 1997. An experimental investigation on diffusion of water in haplogranitic melts. Contrib. Mineral. Petrol. 126, 365–376.], but different from the assumption of Zhang et al. (1991a) [Zhang, Y., Stolper, E.M., Wasserburg, G.J., 1991a. Diffusion of water in rhyolitic glasses. Geochim. Cosmochim. Acta 55, 441–456.], because the dependence cannot be resolved from their low-H 2O t diffusion data, and because the dependence is not so strong at low pressures. The activation energy for H 2O m diffusion decreases as H 2O t increases and depends on P (increases with P at X<0.05 and decreases with P at X>0.05). The results roughly reconcile the different activation energies of Zhang et al. (1991a) and Nowak and Behrens (1997). The total (or bulk) H 2O diffusivity ( D H 2O t ) can be calculated from D H 2O t = D H 2O m d X m/d X, where X m is the mole fraction of H 2O m. This approach can reproduce the D H 2O t values to within a factor of 2 in the range of 400–1200°C, 0.1–810 MPa, and 0–7.7% H 2O t. An explicit formula for calculating D H 2O t at H 2O t≤2% is: D H 2 O t = C C 0 exp 10.49− 10,661 T − 1.772P T , where C is H 2O t content by weight, and C 0 equals 1% H 2O t. A formula for calculating D H 2O t at all conditions covered by this work is: D H 2 O t =X exp(m) 1+ exp 56+m+X −34.1+ 44,620 T + 57.3P T − X 0.091+ 4.77×10 6 T 2 , where m=−20.79−5030/ T−1.4 P/ T. The diffusivities obtained in this work can be used to model bubble growth in explosive and nonexplosive rhyolitic volcanic eruptions in all commonly encountered T, P, and H 2O t conditions.

Crystal chemistry of K-richterite-richterite-tremolite solid solutions a SEM, EMP, XRD, HRTEM and IR study

Solid solutions in the ternary K-richterite-richterite-tremolite (K x Na y □ 1-x-y )(Na x+y Ca 2-x-y ) 2 Mg 5 [Si 8 O 22 /(OH) 2 ] have been synthesized at 200 MPa and 800°C using cold-seal vessels, and at 1800 MPa and 800°C using a piston-cylinder apparatus. The amphiboles were synthesized from oxide and hydroxide mixtures in the presence of a 2-molal aqueous chloridic solution. Solid run products have been investigated by optical, electron scanning and high-resolution transmission electron microscopy, electron microprobe, X-ray powder diffraction and fourier transform infrared spectroscopy. The synthesized amphiboles are up to 1000 μm x 50 μm in size. The crystals are chemically homogeneous and structurally well-ordered. Complete solid solutions are observed in the ternary Kri-ri-tr, but compositions very close to pure richterite and K-richterite end-member could not be synthesized, always showing small but significant amounts of tremolite component. In addition, small amounts of cummingtonite component are present in most amphiboles. There are no indications for crystal chemical restrictions for complete solid solutions, however. The lattice parameters of the solid solutions correlate linearly with the cation occupancies on the A-and M4-sites and are therefore a linear combination of the lattice parameters of the end-members tremolite, cummingtonite, richterite and K-richterite. Two systems of OH-stretching bands are observed. The first band system at wavenumbers between 3669 and 3678 cm -1 is due to a vacant A-site and the second between 3721 and 3737 cm -1 due to a filled A-site. The observed fine structure of the bands can be attributed to distinct M4-site occupancies by Ca 2+ , Mg 2+ and Na + . Using pure tremolite as a standard, the vacancy concentration was determined quantitatively from normalized integral absorbances of the band system at 3669-3678 cm -1 . The derived vacancy concentrations are consistent with those derived by electron microprobe. Additional bands at 3659, 3695 and 3710 cm -1 are probably due to triple-chains or higher chain multiplicity faults in the amphibole. The integral absorbances of these bands may be used to determine, the concentration of chain multiplicity faults.

Hydrothermal Synthesis of Corrensite: A Study of the Transformation of Saponite to Corrensite

AbstractHydrothermal synthesis experiments were conducted to study the transition from smectite to corrensite. A mixture of oxides with the bulk composition of corrensite—Na0.4(Si6.4Al1.6)(Mg7.8Al1.2)-O20(OH)10—was sealed in platinum capsules with 29–37 wt. % water. One set of samples was treated in cold-seal vessels at 500°C and 2 kbar for durations of 2, 3, 6, 12, and 24 h; the other set was treated at 350°C and 2 kbar for periods of 12 to 89 d. X-ray diffraction patterns (XRD) of oriented aggregates from treated products were obtained from ethylene glycol-solvated and air-dried preparations. Samples were also heated to 350°C either in a calibrated muffle furnace, removed and quickly placed in a nitrogen filled chamber on the diffractometer, or were heated at 350°C by using a calibrated heating stage mounted on the diffractometer.Initial mineral assemblages at both temperatures contained only saponite and serpentine. In experiments at 500°C, saponite transformed to corrensite within 6 h; in experiments at 350°C, the transformation occurred as early as 22 d. Increased experiment times at both temperatures produced increasing amounts of well-crystallized corrensite, as indicated by several well-defined XRD peaks. No evidence of a randomly interstratified chlorite-smectite (C-S) precursor to corrensite was found. The identification of pure smectite, as opposed to highly-expanded randomly interstratified C-S, was possible only when clays were dehydrated on a heating stage on the diffractometer.These results call for a new examination of hydrothermally-altered basalt that has been reported to contain randomly interstratified C-S as an intermediate step in the reaction of smectite to corrensite or chlorite. These results also strengthen the view held by increasing numbers of investigators that corrensite should be regarded as a single phase, not as a mixed-layered phyllosilicate.

Chemical interaction between a simulated nuclear waste glass and different backfill materials under a thermal gradient

The initial stage of a HLW disposal will be dominated by a strong thermal gradient that will exist between the hot canister and the “cold” geological medium. In the case of hot wastes (MOx, direct disposal of used spent fuel, short interim storage), nuclear glass can begin to dissolve in groundwater and to interact with other materials in the presence of a thermal gradient. Moreover, it has already been demonstrated that thermal gradients strongly influence the hydrothermal corrosion of nuclear glass by creating large elemental segregation and subsequent mass transport processes. The aim of this article is to experimentally test whether the presence of a simulated engineered clay barrier affects these mass transport processes. In parallel, experiments were performed with different clays of various Si/Al ratios (smectite, kaolinite and allophane) as Si and Al are often invoked as possible rate limiting species for nuclear glass aging. A simulated nuclear glass was disposed between two layers of various clays, enclosed in a sealed gold tube filled with deionised water. The gold tube was put in a cold-seal vessel and submitted to a thermal field under high temperatures (300–200°C, 140 bars, 100 days). High temperatures were used in order to enhance transformation kinetics as mass transport is supposed to be independent of the temperature. At the end of the experiments, the newly formed phases along the tube as well as the mineralogical transformations of the reactants were studied by integrated in-situ analyses (SEM, TEM, EDS and XRD). Precise descriptions of the mineralogical transformations are presented. The major result is a systematic re-crystallization of the initial clay under the influence of the leached elements from the glass (in particular Si and alkalies). The nuclear glass aging is strongly affected by the composition of the surrounding clay and consumes Al from the clay. The aluminous-rich medium seems to limit the glass corrosion and the crystallinity of the newly formed phases. Furthermore, the cold end of the system is devoted to almost pure Al–Si bearing minerals associated with borate. This finding attests to (i) the migration capacities of Si, Al and B through the clay barrier and (ii) the confining properties of the clay barrier for the heaviest elements (Fe, Zn, Zr) which were mineralogically trapped at the clay–glass interface. In conclusion, the study reveals that interactions between the nuclear glass and the backfill materials are important. Thus, these interactions have to be taken into account in order to predict the long term evolution of the system. Furthermore, backfill materials are an efficient chemical barrier which favors the immobilization of the heaviest elements (Fe, Zn and Zr in our experiments).

Temperature gradients in large cold-seal pressure vessels

AbstractRadial and longitudinal temperature gradients within cold-seal pressure vessels can contribute significantly to the total temperature unccrtainty in any one experiment and it is important to calibrate such gradients before experiments are undertaken. We have measured temperature gradients in vertically mounted large cold-seal vessels at 1 atm and 1 kbar and in the temperature range 500–800°C Radial gradients were measured with the vessel at various positions within the furnace by recording the temperature difference between a thermocouple in the internal bore of the vessel and one in a well in the top of the vessel. We find that when the vessel is at a certain position in the furnace, the temperatures read by these two thermocouples are equivalent to within 1°C. The contribution of radial gradients to temperature uncertainty can be thus be substantially minimized if experiments are undertaken with the vessel in this position. Longitudinal gradients of 2.8°C/cm at 500°C and 3°C/cm at 800°C were recorded at 1 kbar. These profiles are significantly steeper than those recorded at 1 atm. Not to calibrate temperature gradients at the experimental conditions or failure to calibrate vessels at all could lead to large temperature uncertainties.

The decarbonation and heat capacity of ZnCO 3

The decarbonation curve for ZnCO 3 has been determined in the pressure range 3–20 kbar by using a combination of cold-seal vessels and piston-cylinder apparatus with NaCl assemblies. Heat capacities for both synthetic and natural ZnCO 3 samples were measured by differential scanning calorimetry at temperatures ranging from 340 to 497 K. The results of these experiments indicate that the enthalpy of formation for smithsonite, ΔH f (1,298.15), is approximately −817. kJ/mol. which is about 4 kJ more negative than most tabulated values.

The Solubility of H2O in Melts in the System SiO2-Al2O3-Na2O-K2O at 1 to 2 Kbars

The solubilities of water in six melts in the system $$K_{2}O-Na_{2}O-Al_{2}O_{3}-SiO_{2}$$ were determined at 970-1630 bars and 800°C. Melts were synthesized hydrothermally in Pt capsules in cold-seal vessels and then quenched isobarically. The solubilities were determined by micromanometric measurement of $$H_{2}O$$ evolved during vacuum fusion of vesicle-free glass wafers and are reproducible to ±2.8% (2 s.d.) of the concentration. The solubilities at 970 bars for the granitic and phonolitic minimum melts are $$2.88 \pm 0.10$$ wt % and $$5.01 \pm 0.14$$ wt %, respectively. Both peralkaline and peraluminous granitic melts have higher $$H_{2}O$$ solubilities than the 1 kbar $$P(H_{2}O)$$ minimum melt and indicate the existence of minima in solubilities at molar (Na + K)/Al = 1 along joins of constant $$SiO_{2}$$. Two melts at constant $$SiO_{2}$$ and (Na + K)/Al ratio, but varying in Na/(Na + K) (0.57, 0.69 molar), have the same solubility ($$2.88 \pm 0.10, 2.74 \pm 0.10$$ wt %). The compositional depend...