Genesis of intrusive-related hydrothermal gold deposits

Abstract

The nature of gold transport and deposition in deep, intrusive-related hydrothermal systems has been the topic of this PhD-research. Specifically, the role of I-type granites in the formation of gold deposits that occur in contact metamorphic aureoles has been addressed by an integrated field study of gold deposits in the aureole of the Cullen Batholith in the Proterozoic Pine Creek Inlier, northern Australia. This investigation shows that gold was deposited in quartz vein which formed near the thermal peak of contact metamorphism at 550 ≤ T ≤ 620°C at a pressure around 200 MPa. Gold is localized within quartz veins that occupy fracture arrays in a variety of structural types of major antiforms. The gold veins formed in two distinct structural settings. In the first, the veins indicate fluid flow focusing while the host structure was actively deforming. In the second, predominantly bedding-conformable vein systems and randomly oriented stockworks indicate the focusing of fluids into anticlines in the absence of deformation. On a regional scale, gold mineralization occurs where the vein systems intersect carbonaceous units in the metapelitic host-rock sequence or where veins formed near carbonaceous units (≤100 m to contacts). Although abundant, veins further from carbonaceous beds are usually barren. This association prompted the question of what unique chemical conditions existed within or near the carbonaceous slates and what was their effect on gold solubility. The chemical conditions of gold transport and deposition have been inferred by interpretation of paragenetic relationships, a detailed study on the host-rock sequence (Matthiii and Henley, in prep.), and predictions about the nature of magmatic hydrothermal fluid. Throughout the central Pine Creek Inlier gold-quartz veins are associated with pyrrhotite, arsenopyrite and potassium silicate alteration (K-feldspar + biotite in paragenesis with andalusite and cordierite). Potassium metasomatism accompanied gold mineralization while N a, Ca was leached from the hydrothermally altered rocks. In the absence of evaporites in the metasedimentary sequence, high fluorine contents in biotite from the alteration selvages of the goldquartz veins ( 4,300≤F≤49,000 ppm) indicate that a magmatic fluid was involved in ore genesis. This involvement is consistent also with ( 1) mutual cross-cutting relations of gold-quartz veins, pegmatites, and aplite dykes in many of the deposits, (2) the location of the deposits in the roof of the I-type granites at a vertical distance less than 2 km (typically ≤1 km) from the sub-horizontal intrusive contact, and (3) the high salinity of the mineralizing fluid as estimated from the Cl-content of biotite in most of the deposits. A magmatic derivation of the sulphur in the epigenetic sulphides is consistent with sulphur isotopic data from the Cosmo Howley deposit (-0.3 to +7.2 per mil 8348). Here an early-magmatic, oxidizing fluid could explain the heavy sulphur isotopic signatures of sulphides in structurally early gold-quartz veins which constitute the high grade stage of mineralization ( 834S arsenopyrite: +3. 7 to + 7.2 per mil). The I-type granitoids have I-SC characteristics (Ague and Brimhall, 1988): hornblende, pink-K-feldspar, accessory shene, and fluorite; whole rock Fe⁺³ /Fe⁺² ≥ 0.24 which indicate that the redox state of the magmatic fluid was at or above the NNO buffer at the temperature of crystallization. But the pyrrhotite-arsenopyrite assemblage in the veins and the association of ore with carbonaceous units indicates that gold was deposited under fairly reducing conditions {10⁻²²·⁵ ≤ fO₂ ≤ 10⁻²⁰ ). Thus, local reduction of initially more oxidized fluid passing in the vicinity of carbonaceous metasediments appears to have been responsible for localized gold deposition in the deposits in the Pine Creek Inlier. That this mechanism is highly efficient, and may explain the concentration of gold in many gold deposits globally, has been confirmed by speciation and reaction-progress calculations. These calculations indicate that AuCl₂⁻ was more important than Au(HS)₂⁻ and AuHS⁰ at the high temperature of gold precipitation in the Pine Creek Deposits. However, even at lower temperatures (400-500°C), total gold solubility has a minimum near the maximum stability of graphite, so that reaction with carbonaceous matter could be an important precipitation mechanism. Fluid-phase immiscibility consequent upon fluid contamination by hydrocarbons could also have occurred in the gold deposits, but its effects on gold solubility at high temperatures have yet to be determined. In the Cosmo Howley gold deposit, most of the vein-hosted ore is located in hornfels and biotitic slate within 50 meters of the footwall of carbonaceous slates. For this zone, it was inferred (from the field evidence of near-lithostatic fluid pressure and the nature of chemical alteration) that fluid flow was upward and down a temperature gradient. Flow focusing into the vein-fault conduits should have entailed mixing of mineralizing fluid rising from greater depth with hydrocarbon-rich fluid which ascended along the steeply-dipping beds of carbonaceous slate prior to being focused laterally in the subvertical vein-fault conduits. This mixing hypothesis has been corroborated by results from steady-state modelling of pressure-driven fluid flow for a cross-sectional geometry of the Cosmo Howley deposit with the writer's Monte Carlo method for modelling flow processes in complex geometries with order-of-magnitude variations in permeability. Another numerical method has been designed to track fluid batches on their migration paths through the flow models. This tool has been applied in a simplistic flow model of the Cosmo Howley deposit. Using boundary conditions determined from the field study and from chemical calculations, preliminary results indicate that hydrocarbon concentrations in the vein-fluid in the footwall of the slate can exceed the threshold value for fluid immiscibility upon fluid mixing which occurs within the vein conduits near the carbonaceous slate units. Flow patterns which could lead to this type of mixing have also been identified for a range of steady-state models based on other structural geometries which typify many gold deposits in a metamorphic setting. Dynamic patterns of fluid flow near actively-deforming faults have been simulated in laboratory experiments using a newly-designed electric analog apparatus that models fluid flow in metamorphic environments using analogs between charge transport and fluid flux, storativity and capacity, and permeability and impedance. The electric analog experiments model the field evidence of ( 1) episodic faulting in the deposits and (2) the common presence of crack-seal textures in the gold-bearing veins which suggest that fluid flow was intermittent and that the permeability magnitude and distribution changed incrementally with time. The boundary conditions for the electric analog model experiments comprise measurements of permeabilities, porosities and storativities of appropriate finegrained rock at relevant PT-conditions of metamorphism. The electric analog simulations show that the steady-state models provide upper bounds on the dynamic focusing of fluids by fault/vein conduits, because the hydraulic conductivity of the faults has to be increased for years to thousands of years before fluid is drawn into the fault from a wider region upstream. This applies under the proviso that the permeability of the metamorphic rocks is in the micro to nano-darcy range. Electric analog experiments also show that after short-lived events of fault dilation (months to tens of years) during which the hydraulic conductivity of the fault was increased by up to eight orders of magnitude, the fluid pressure upstream in the fault recovers instantaneously to nearly the pre-failure pressure as soon as fault compaction and sealing is simulated. This indicates that equilibrium between the fluid pressure in the fault and the fluid pressure in the wallrock is not attained. Furthermore, although decompression upon fault dilation is complete, extremely little inflow occurs in the subsequent tens to hundreds of years, because the rate of inflow is determined by the hydraulic conductivity of the wallrock. These electric analog experimental results represent a range of manifestations of the slow diffusivity of fluid pressure which is inferred to describe metamorphic lowpermeability environments. This diffusivity decreases proportional to the hydraulic conductivity and the storativity of the rock. Small diffusivity values (e.g., tens of meters per year) imply that single fault-slip events cannot decrease fluid pressure in the host rocks significantly if the fault-slip-induced increase in the hydraulic conductivity of the fault is sustained for less than tens to hundreds of years. For the case in which the fault undergoes dilation during failure, the decompression within the fault is complete, so that the effective stress on the faulted rock is very high. Under those conditions and temperatures above 400°C, healing and sealing of the fault should occur within months to hundreds of years, as will be shown on the basis of experimental data from the literature. Summarizing these results, a sustained or frequently regenerated high hydraulic conductivity seems to be the pre-requisite for substantial focusing of fluid flow by faults. Alternatively, the escape of overpressured fluids by hydrofracturing can be focused by the regional permeability distribution. This has occurred in those gold deposits in the Pine Creek Inlier which were not actively deformed during oregenesis.