Abstract:
1. Distribution of trace elements between biotite and hydrous granitic melt.
The synthetic system haplogranite-anorthite-biotite-H2O, doped with a suite of trace elements (V, Cr, Co, Ni, Cu, Zn, Rb, Cs, Sr, Ba, Mo, W, Nb, Ta, REE, Pb and Zr) was experimentally investigated over the temperature interval of 700-800°C at 200 MPa (H2O) to study the distribution of trace elements between biotite and coexisting hydrous melt. The starting materials were a series of volatile-free synthetic glasses obtained by heating a mixture of appropriate reagent grade chemicals (oxides, carbonates or hydroxides) corresponding to ~76 wt% haplogranite (35 wt% SiO2, 40 wt% NaAlSi3O8 and 25 wt% KAlSi3O8), ~8 wt% anorthite, ~15 wt% biotite (with a 1:1 molar ratio of Mg/Fe) and ~1 wt% total trace elements, so selected as to avoid peak interference during microprobe analysis. Experiments were carried out under water-saturated conditions in cold-seal autoclaves with run durations of between 35 and 45 days. The oxygen fugacity was close to the Ni-NiO buffer. Run products were analysed by electron microprobe.
Besides biotite, run products contained other phases including amphibole (hornblende), clino-pyroxene, plagioclase, magnetite, allanite and glass.
Measured partition coefficients indicate that most alkalis, alkaline earth metals and transition metal elements favoured biotite over the melt (e.g. DRb ~ 2.6, DCs ~0.6, DBa ~1.8, DSr ~0.6, DCo ~34, DNi ~84, DZn ~8.1, DV ~26, DCr ~15, and DTi ~10). The converse is, however, true for the REEs, with D-values decreasing from DLa ~0.6 to DGd ~0.04 and again rising to DLu ~0.4. The variation of trace element partition coefficients with ionic radius was described by an elastic strain model. The alkalis (Cs+, Rb+, and Na+) and alkaline earths (Ba2+, Sr2+, and Ca2+), including Pb2+, follow K+ into the largest interlayer site of the biotite crystal, each group fitting onto separate parabolic curves. Most of the other trace elements, including 2+ (Ni, Co, Zn, Cu, and Fe), 3+ (V and Cr), 4+ (Ti and Zr), 5+ (Nb and Ta), and 6+ cations (W and Mo) follow Mg2+ into the M2 site, in 6-fold co-ordination, with each group following its own parabolic curve. However, only the divalent group, with more than two trace elements, has been treated onto the elastic strain model with fit parameters ~90, ~6850 kbar and ~0.67 Å.
The REEs show a partitioning pattern, which cannot be explained by substitution onto a single lattice site. They apparently partition into two crystallographic sites of the biotite (i.e. K and M1). The LREEs predominantly partition into the K-site, the MREEs show mixed partitioning character in both the K and M1 sites and the HREEs predominantly partition into the M1 site. The measured partition coefficients, therefore, fit onto two parabolic curves.
The accessory mineral allanite was found in run products belonging to some samples doped with REEs in the starting mixtures. The REEs are highly compatible in allanite with their partitioning behaviour possibly being influenced by temperature (e.g. DCe ~390, DPr ~340, DNd ~290, DEr ~40, at 800°C and DPr ~650, DNd ~630, DEr ~60, at 775°C).
Brice model fit-curves show that all REEs partition into the octahedral Ca-site of the allanite crystal.
Biotite is an early crystallising mineral in magmatic systems and tends to fractionate a greater proportion of the compatible trace elements from the melt, rendering them unable to form viable ore deposits. On the other hand, incompatible trace elements are progressively enriched in the melt with increasing crystallisation and fractionation of biotite, with the result that at fluid saturation, a good proportion of these elements will be extracted from melt by exsolved and migrating hydrothermal fluids, to be later concentrated and deposited to form ore deposits. Allanite-melt partition coefficients for REEs are so high that a very small fraction of crystallisation and fractionation of allanite may drastically deplete the melt of its REE content close to zero.
2. Speciation and oxidation state of copper in silicate melts
The oxidation state of copper in silicate melts of variable composition was studied at 1 bar at temperatures of between 1250 and 1400°C under controlled oxygen fugacity.
Synthetic silicate glasses (each doped with 1 wt% Cu, as CuO) of haplogranites, tholeiite, alkali-basalt and andesite were prepared from appropriate amounts of reagent-grade oxides, carbonates and hydroxides. With regard to the haplogranites, a series of 4 separate glasses were prepared with varying molar ratios of [Na2O+K2O]/Al2O3 (i.e. 0.6, 0.93, 1.42 and 1.75) but keeping the SiO2 content and the molar ratio Na2O/K2O relatively constant, to represent a range of naturally occurring melts, including peraluminous, subaluminous, and peralkaline varieties.
Experiments were conducted using a gas-mixing technique to investigate the effect of oxygen fugacity on the redox ratio of copper (Cu2+/Cu+) in silicate melts, at various equilibrium temperatures (1250-1400°C). Oxygen fugacity inside the furnace (1 bar total pressure) was imposed by pure CO2 and CO2/H2 gas mixtures with varying ratios. A Y- doped ZrO2 oxygen sensor assembly was installed in the furnace as a means of monitoring the imposed oxygen fugacity inside the furnace.
A Bruker IFS 125 HR Fourier-Transform-Spectrometer (FTIR) was used to measure absorbance due to crystal field bands of Cu2+ in the run product glasses. The Cu+ content in the glasses was determined as the difference between the total Cu determined by electron microprobe and the Cu2+ measured by optical spectroscopy.
Glass standards containing only Cu2+ were also prepared under high oxygen partial pressure generated by the decomposition of PtO2 in sealed Pt/Rh capsule at 1250°C.
The extinction coefficients e (l/mol/cm) in glass standards were calculated to be 12.8 for alkalibasalt, 16.8 for tholeiite, 11.6 for andesite, 7.26 for haplogranite with the molar ratio (Na + K)/Al = 0.6, 33.2 for haplogranite with (Na + K)/Al = 0.93, 15.6 for haplogranite with (Na + K)/Al = 1.42 and 15.4 for haplogranite with (Na + K)/Al = 1.75.
Peraluminous haplogranite glass with molar ratio (Na+K)/Al = 0.60 entirely contained Cu+ ions, which could not be converted into Cu2+ ions even under the most oxidising conditions using PtO2.
For a given glass composition under the same gas atmosphere, absorbance (A), colour intensity and ratio Cu2+/Cu+, increase with decreasing run temperatures. On the other hand, if temperature is kept constant and gas atmosphere varied, absorbance (A), glass colour intensity and the ratio Cu2+/Cu+, increase with increasing oxygen fugacity.
Generally the maximum absorption band due to Cu2+ ions at any run temperature appeared at ~ 13000 cm-1 (wavenumber) for andesite, alkalibasalte, and tholeiite glass and ~ 12500 cm-1 for haplogranite glasses.
For a silicate melt of a given composition, plots of log (Cu2+/Cu+) against log fo2 at any run temperature exhibit a linear relationship, with graphs for different temperatures trending almost parallel to each other with slopes in the range of 0.239 to 0.256. The lowest and highest temperature curves lie uppermost and lowermost, respectively. At a fixed run temperature T, plots of log (Cu2+/Cu+) against log fo2 for a silicate melt of any given composition exhibit a similar linear relationship, with curves for the granitic , dioritic and basaltic melts lying upper-most , intermediate and lowest, respectively.
?H for the oxidation of Cu+ to Cu2+ was estimated to be -531.3, -392.8, -386.9, -394.3, -473.9 and -521.9 kJ/mol, for tholeiite, alkali-basalt, andesite, haplogranite with (Na + K)/Al = 0.93, haplogranite with (Na + K)/Al = 1.42 and haplogranite with (Na + K)/Al = 1.75, respectively.
The redox ratios Cu2+/Cu+ for granitic and dioritic magmatic systems that often crystallise at temperatures in the range 700 - 800°C, under oxygen fugacities between those of QFM (quartz-fayalite-magnetite), Ni-NiO, and H-M (hematite-magnetite) was estimated. The results show a predominance of Cu2+ in granitic and dioritic plutons, i.e. the ratio Cu2+/Cu+ is always greater than one. Basaltic melts, however, seem to behave differently. They crystallise in the temperature range (1100 - 1200°C), under oxygen fugacities between those of QFM and Ni-NiO buffer systems, with a predominance of Cu+ over Cu2+ ions in their melts, i.e. with very low ratios of Cu2+/Cu+ (usually less than one).
3. Solubility of copper in common rock-forming minerals
A range of synthetic systems was experimentally investigated (in two options) at isobaric and isothermal conditions (2 kbar and 600-700°C) to determine the solubility of copper in some common rock-forming minerals (muscovite, phlogopite, orthoclase, and albite) which crystallized from hydrothermal solutions in equilibrium with excess metallic Cu and Cu2O. Samples for option one experiments were prepared simply by mixing reagent grade chemicals corresponding to ~65 wt% mineral and ~ 35 wt% Cu2O, whereas samples for option two experiments consisted of mixtures corresponding to ~90 wt% mineral and ~ 10 wt% hypothetical Cu-bearing mineral (e.g. for the mineral muscovite 90 wt% KAl2 [AlSi3O10] [OH]2 + 10 wt% Cu+Al2(AlSi3O10)(OH)2).
Experiments were carried out in cold seal bombs at 600 - 700°C with run durations of 15 - 20 days using capsules made of pure copper.
Depending on the composition of the starting mixtures, different copper solubilities were observed as different mixtures stabilised different mineral assemblages. For instance, the solubility of Cu (as wt% Cu2O) in orthoclase (CuOr) is in the range 0.863-0.049 and 0.089-0.049 in run products of option 1 (Or1) and 2 (Or2) mixtures, respectively. In albite, CuAb ranges from 1.175-0.630 in Ab1 and 0.041 in Ab2. In muscovite, CuMu ranges from 0.965-0.634 in Mu1 and 0.08 in Mu2. In phlogopite, CuPhl ranges from 1.142-0.555 in Phl1. In quartz, CuQtz ranges from 1.336-1.173 in Qtz1 and 0.054-0.049 in Qtz2. The upper and lower values of the solubility ranges correspond to 600 and 700°C, respectively. Many other phases crystallised in the run products of both experimental options remain to be identified whilst others simply exist in very low proportions. So there is seemingly still more work left to be done. In principle, once the experimental samples are fully characterised, the results obtained in these experiments can be combined with those of copper solubility studies in hydrothermal fluids, available from thermodynamic data bases, to calculate partition-coefficients of copper between fluid and minerals existing in equilibrium.