Final Report Summary - MAGMA (Mass transfer of metals and sulfur between mafic silicate melts and volatiles: an interdisciplinary approach)
Project context and objectives
The major goal of the project was to assess the significance of mafic to intermediate magmas in the formation of magmatic-hydrothermal ore deposits of Au and Cu. Such deposits typically form from a volatile phase (sulphur, chlorine and CO2-bearing low-density, high-temperature aqueous solutions with dissolved metals) exsolving from a rising, cooling and crystallising silicate melt.
Work performed
In the first stage of the project, I studied the solubility of Au and Cu in high-temperature sulphur and chlorine-bearing aqueous volatiles at physical-chemical conditions that are characteristic of natural magmas in the earth's upper crust. A series of experiments were conducted in custom-built cold-seal molybdenum-hafnium-carbide (MHC) pressure vessel assemblies to obtain Au and Cu solubilities at such extreme conditions. The volatile phase was trapped in quartz mineral in the form of synthetic fluid inclusions at the imposed high pressure (P) and temperature (T) conditions. Subsequently, they were analysed by laser-ablation inductively coupled plasma mass spectrometry (LA-ICPMS) to obtain the concentrations of dissolved metals. The results revealed that in the case of Au, reduced sulphur species, likely hydrosulphide anions, are the most important complex-forming ligands, whereas in the case of Cu, the relative significance of chloride complexes is much higher. A surprising discovery of this study was the recognition of the significant role of dissolved alkali-chlorides in stabilising Au-hydrosulphide, and, to a smaller extent, Cu-hydrosulphide complexes in low-density magmatic vapours. Ab initio quantum chemical calculations showed that the most likely reason for this effect is ionic interaction between alkali metal cations and Cu- or Au-hydrosulphide anions (Cu(SH)2-, Au(SH)2-), or the formation of mixed Cu- and Au-hydrosulphide-chloride complexes also associated with alkali metal cations.
The exsolution of a low-density, low-salinity vapour, and a dense saline brine phase from an initially single phase magmatic volatiles upon decompression is a typical process in nature and plays an important role in ore deposit formation. Therefore, we extended our investigation with the study of vapour/brine partition coefficients of Cu, Au and Mo, and the temperature dependence of Au and Cu solubilities in these phases. These experiments showed that Mo, Cu and Au partitions into the brine independent of temperature (at T=650-900 °C) and the presence of H2S, SO2 and HCl. These experiments also showed that with decreasing T, the solubility of Au drops slightly, but the solubility of Cu increases significantly.
The second stage of the project focused on studying the distribution of S and Cl between andesite melts and magmatic volatiles; so that the composition of magmatic volatiles present in upper crustal magma chambers can be predicted from silicate melt inclusion trapped in phenocrysts of erupted volcanic rocks. An innovative experimental approach, based on the use of high fluid/melt mass ratios in the capsule load allowed us to obtain precise and accurate partition and exchange coefficients, even for dilute volatiles typical of natural systems. These experiments have shown that the partition coefficients of S are about 1 to 2 orders of magnitude higher than those of Cl, and that pyrrhotite or anhydrite saturation limits the maximum S concentration in the volatile phase to about 3 mol% at 1 000 °C and 200 MPa. Furthermore, with the exchange coefficients of K, Na, Fe and H in hand, we can now predict that Cl is present in sub-equal proportions of FeCl2, NaCl, KCl and HCl in the volatile phase in equilibrium with andesite melts in sub-volcanic magma chambers. This is remarkably different from the composition of high-temperature volcanic gases measured in surface fumaroles and has significant implications for the capacity of the volatile phase to transport economically important metals.
In the third stage of the project, we investigated the solubility of Au, Cu and Ag in silicate melts. This study has shown that Au is highly incompatible in the structure of the silicate melt, and that in hydrous melts it shows similar speciation to that observed in the volatile phase. The presence of S2- in the melt greatly increases the solubility of Au, and Cl has a similar effect but to a much smaller extent. Furthermore, the simultaneous presence of the two anions in significant concentrations yields particularly high Au solubilities, similar to the volatile phase. Most likely, AuSHo is the major dissolved Au species in most natural silicate melts. Therefore the major limiting factor in the Au-transport capacity of silicate melts is the S concentration at sulphide saturation, and in felsic systems, the Al/(Na+K) ratio of the melt. The effect of S and Cl is much smaller on the solubility of Cu and Ag in silicate melts than in the case of Au. Furthermore, the major element composition of the melt itself exerts only a moderate effect on their solubilities. The effect of temperature is more significant, as both Cu and Ag solubilities decrease greatly between 1 000 and 800 °C in rhyolite melts. As the effect of temperature on Cu solubilities in the silicate melt is opposite to that in the volatile phase and pyrrhotite, it is likely that the Cu-budget of natural magmas shifts towards these latter phases with melt evolution.
An auxiliary output of this study was showing that Mo preferentially partitions into andesite melts over magmatic vapours independent of S and Cl concentrations.
The combination of the above results allowed us to model the partition coefficients of Au and Cu between andesite melts and magmatic volatiles with major element compositions typical of natural magmas. These calculations have shown that Au can be very efficiently extracted from mafic to intermediate magmas by degassing if the oxygen fugacity in the system is below the sulphide to sulphate transition, whereas magmatic volatiles are inefficient in the extraction of Cu from such magmas. Thus volatile input into an ore-forming felsic system from an underlying hotter, more mafic magma will increase the characteristic Au/Cu ratio of the forming ore deposit. Nevertheless, the formation of porphyry type Cu deposits may also be promoted by early volatile exsolution, because it shifts the Cu budget in the evolving magma in favour of the silicate melt by destructing stable sulphide minerals through efficient S removal from the system. This way, Cu may remain available for extraction by more Cl-rich volatiles exsolving at lower temperatures from the residual melt. Furthermore, mafic magmas may largely promote the formation of porphyry Cu deposits by supplying large amounts of S for the ore-forming systems, which is also essential for the formation of such deposits.
Main results
The industrialised world relies on the availability of various metals and on a low price of commodities. These results can be used to increase the efficiency of explorating magmatic-hydrothermal Au and Cu ore deposits. The most important characteristics of a sub-volcanic magmatic complex, such as major element composition, metal concentrations and oxidation state can be assessed using traditional petrological methods and silicate melt inclusions in minerals. This information combined with the new metal and volatile solubility and partitioning data allows the estimation of the ore-forming potential of a prospective intrusive complex.
Volatiles released from magmas not only serve as a source of metals for ore deposit formation, but are also the driving force of explosive volcanic eruptions. The monitoring of volcanic gas compositions is one of the methods used in eruption forecasting. Part of the data resulting from this project will help linking the gas composition observed on the surface to processes occurring at depth in the magma chamber, thus improving our capability of forecasting volcanic eruptions.
Naturally, more work is needed for the development of comprehensive models applicable to a wide P-T and melt compositional range. However, the success of this project in expanding our knowledge from felsic magmas to more mafic systems, and the new experimental methods developed are a major step in this direction.
The major goal of the project was to assess the significance of mafic to intermediate magmas in the formation of magmatic-hydrothermal ore deposits of Au and Cu. Such deposits typically form from a volatile phase (sulphur, chlorine and CO2-bearing low-density, high-temperature aqueous solutions with dissolved metals) exsolving from a rising, cooling and crystallising silicate melt.
Work performed
In the first stage of the project, I studied the solubility of Au and Cu in high-temperature sulphur and chlorine-bearing aqueous volatiles at physical-chemical conditions that are characteristic of natural magmas in the earth's upper crust. A series of experiments were conducted in custom-built cold-seal molybdenum-hafnium-carbide (MHC) pressure vessel assemblies to obtain Au and Cu solubilities at such extreme conditions. The volatile phase was trapped in quartz mineral in the form of synthetic fluid inclusions at the imposed high pressure (P) and temperature (T) conditions. Subsequently, they were analysed by laser-ablation inductively coupled plasma mass spectrometry (LA-ICPMS) to obtain the concentrations of dissolved metals. The results revealed that in the case of Au, reduced sulphur species, likely hydrosulphide anions, are the most important complex-forming ligands, whereas in the case of Cu, the relative significance of chloride complexes is much higher. A surprising discovery of this study was the recognition of the significant role of dissolved alkali-chlorides in stabilising Au-hydrosulphide, and, to a smaller extent, Cu-hydrosulphide complexes in low-density magmatic vapours. Ab initio quantum chemical calculations showed that the most likely reason for this effect is ionic interaction between alkali metal cations and Cu- or Au-hydrosulphide anions (Cu(SH)2-, Au(SH)2-), or the formation of mixed Cu- and Au-hydrosulphide-chloride complexes also associated with alkali metal cations.
The exsolution of a low-density, low-salinity vapour, and a dense saline brine phase from an initially single phase magmatic volatiles upon decompression is a typical process in nature and plays an important role in ore deposit formation. Therefore, we extended our investigation with the study of vapour/brine partition coefficients of Cu, Au and Mo, and the temperature dependence of Au and Cu solubilities in these phases. These experiments showed that Mo, Cu and Au partitions into the brine independent of temperature (at T=650-900 °C) and the presence of H2S, SO2 and HCl. These experiments also showed that with decreasing T, the solubility of Au drops slightly, but the solubility of Cu increases significantly.
The second stage of the project focused on studying the distribution of S and Cl between andesite melts and magmatic volatiles; so that the composition of magmatic volatiles present in upper crustal magma chambers can be predicted from silicate melt inclusion trapped in phenocrysts of erupted volcanic rocks. An innovative experimental approach, based on the use of high fluid/melt mass ratios in the capsule load allowed us to obtain precise and accurate partition and exchange coefficients, even for dilute volatiles typical of natural systems. These experiments have shown that the partition coefficients of S are about 1 to 2 orders of magnitude higher than those of Cl, and that pyrrhotite or anhydrite saturation limits the maximum S concentration in the volatile phase to about 3 mol% at 1 000 °C and 200 MPa. Furthermore, with the exchange coefficients of K, Na, Fe and H in hand, we can now predict that Cl is present in sub-equal proportions of FeCl2, NaCl, KCl and HCl in the volatile phase in equilibrium with andesite melts in sub-volcanic magma chambers. This is remarkably different from the composition of high-temperature volcanic gases measured in surface fumaroles and has significant implications for the capacity of the volatile phase to transport economically important metals.
In the third stage of the project, we investigated the solubility of Au, Cu and Ag in silicate melts. This study has shown that Au is highly incompatible in the structure of the silicate melt, and that in hydrous melts it shows similar speciation to that observed in the volatile phase. The presence of S2- in the melt greatly increases the solubility of Au, and Cl has a similar effect but to a much smaller extent. Furthermore, the simultaneous presence of the two anions in significant concentrations yields particularly high Au solubilities, similar to the volatile phase. Most likely, AuSHo is the major dissolved Au species in most natural silicate melts. Therefore the major limiting factor in the Au-transport capacity of silicate melts is the S concentration at sulphide saturation, and in felsic systems, the Al/(Na+K) ratio of the melt. The effect of S and Cl is much smaller on the solubility of Cu and Ag in silicate melts than in the case of Au. Furthermore, the major element composition of the melt itself exerts only a moderate effect on their solubilities. The effect of temperature is more significant, as both Cu and Ag solubilities decrease greatly between 1 000 and 800 °C in rhyolite melts. As the effect of temperature on Cu solubilities in the silicate melt is opposite to that in the volatile phase and pyrrhotite, it is likely that the Cu-budget of natural magmas shifts towards these latter phases with melt evolution.
An auxiliary output of this study was showing that Mo preferentially partitions into andesite melts over magmatic vapours independent of S and Cl concentrations.
The combination of the above results allowed us to model the partition coefficients of Au and Cu between andesite melts and magmatic volatiles with major element compositions typical of natural magmas. These calculations have shown that Au can be very efficiently extracted from mafic to intermediate magmas by degassing if the oxygen fugacity in the system is below the sulphide to sulphate transition, whereas magmatic volatiles are inefficient in the extraction of Cu from such magmas. Thus volatile input into an ore-forming felsic system from an underlying hotter, more mafic magma will increase the characteristic Au/Cu ratio of the forming ore deposit. Nevertheless, the formation of porphyry type Cu deposits may also be promoted by early volatile exsolution, because it shifts the Cu budget in the evolving magma in favour of the silicate melt by destructing stable sulphide minerals through efficient S removal from the system. This way, Cu may remain available for extraction by more Cl-rich volatiles exsolving at lower temperatures from the residual melt. Furthermore, mafic magmas may largely promote the formation of porphyry Cu deposits by supplying large amounts of S for the ore-forming systems, which is also essential for the formation of such deposits.
Main results
The industrialised world relies on the availability of various metals and on a low price of commodities. These results can be used to increase the efficiency of explorating magmatic-hydrothermal Au and Cu ore deposits. The most important characteristics of a sub-volcanic magmatic complex, such as major element composition, metal concentrations and oxidation state can be assessed using traditional petrological methods and silicate melt inclusions in minerals. This information combined with the new metal and volatile solubility and partitioning data allows the estimation of the ore-forming potential of a prospective intrusive complex.
Volatiles released from magmas not only serve as a source of metals for ore deposit formation, but are also the driving force of explosive volcanic eruptions. The monitoring of volcanic gas compositions is one of the methods used in eruption forecasting. Part of the data resulting from this project will help linking the gas composition observed on the surface to processes occurring at depth in the magma chamber, thus improving our capability of forecasting volcanic eruptions.
Naturally, more work is needed for the development of comprehensive models applicable to a wide P-T and melt compositional range. However, the success of this project in expanding our knowledge from felsic magmas to more mafic systems, and the new experimental methods developed are a major step in this direction.
