Search

ISBN:978-3-906327-52-5

ISBN:978-3-906327-00-6

ISBN:978-3-906031-69-9

Natural convection of water plays a key role in shaping the thermal structure of the subsurface. As a consequence of water density decreasing with temperature and temperature increasing with depth in the Earth’s crust, an unstable situation develops: cold, dense fluid overlies hot, less dense fluid. Given a sufficiently permeable and vertically extensive host, hot water will buoyantly rise towards the surface as cold water sinks to replace it, forming a convective circulation cell. Rock units with permeability far too low to facilitate convection may, however, contain high- permeability faults or fractures capable of sustaining convective fluid flow. We investigate the effects and emergent behavior of fracture convection by means of numerical simulation using the Complex Systems Modeling Platform (CSMP++). Despite being confined to a narrow, planar feature, natural convection in a fault or fracture significantly impacts the thermal structure of the subsurface via conduction heat exchange with the host rock and enhances fluid mixing. We compare convection in wide, filled fault zones and narrow, open slots and show that convection patterns within the fracture and the strength of this subsequent thermal perturbation are a function of transmissivity – permeability times thickness – rather than either permeability or thickness alone. Fluid circulation may lead to temperature changes of several tens of degrees in the host up to a distance roughly equal to half the height of the fracture, significantly changing the local geothermal gradient. Above the fracture, these thermal perturbations influence conductive heat flow and take on a diffuse pattern, appearing much wider than the originating fracture or fault. By varying host rock permeability, we show that fluid exchange between fracture and host varies with transmissivity, but has little effect on convection patterns. Conversely, Rayleigh convection within the host rock dominates and overprints convection within the fracture. We build on these results by investigating convection in five elliptic fractures in an en echelon array. Thermal perturbations arising from convection in each fracture eventually connect and align, thereby imposing mutual feedback of convection in neighboring fractures, despite a lack of hydraulic connectivity between them. This interaction leads to a self-organizing pattern of convection cells, with regions of upflow and downflow aligning with the upflows and downflows in their neighbors, inducing a regular pattern of alternating cooled and heated host rock regions between the fractures that is aligned (sub-) perpendicular to the en echelon fracture array. By varying fracture spacing, we determine that this “synchronization” effect may occur iii when fracture spacing is less than convection cell width. Even in fractures with heterogeneous permeability, convection cells continue to self-organize despite zones of low-transmissivity. Fractures which alone cannot facilitate convection may nevertheless host weaker fluid circulation cells in response to the thermal perturbation caused by convection in a neighboring fracture. Linear elasticity is then introduced to our fracture convection model. Using the Barton- Bandis relationship, we vary fracture transmissivity within the fracture as a function of effective normal stress and depth and show that convection strength and vertical extent depend on fracture stiffness, with convection becoming increasingly confined to shallow depths as stiffness decreases. We then include thermo-elastic strain, allowing temperature changes due to convection to induce thermal strains and alter effective normal stress and transmissivity, resulting in high-transmissivity channels which facilitate downward flow of cold water. Using a simplified fracture propagation model, we show that these channels may induce contractive thermal strains sufficient to induce tensile failure in the host rock below, propagating and deepening the fracture and allowing further infiltration of cold water. We then model injection and production of cold water into a single fracture, focusing on thermo-elastic strains and their associated stresses. Injection of cold water into a fracture induces contractive thermal strains behind the thermal front, generally resulting in decreased normal stress and enhanced transmissivity. We identify a zone of increased normal stress in a ring around this thermal front, decreasing fracture transmissivity. This “stress-ring” is shown to be responsible for a significant portion of the difference between models with and without thermo- elasticity, with a particularly important role in fractures where flow channeling is minimal. Thermo-elastic effects become stronger with increasing host rock bulk modulus, but the relative effect of the stress-ring remains the same. The work presented here chiefly investigates the thermal impact of natural convection in a fracture (chapter 2), the interactions between multiple convecting fractures (chapter 3), the effect of thermo-elasticity on convection patterns (chapter 4), and the impact of thermal strains on fracture flow patterns between a well doublet (chapter 5). We then conclude by summarizing our key findings and look forward to the ways future studies may build upon the work presented here (Chapter 6).

Fluid inclusions are the only available samples of paleo-fluids responsible for crystallization of hydrothermal minerals including ore phases. Analysis of fluid inclusions implicitly assumes that the inclusions have preserved their chemical composition since the time of their entrapment. There is, however, an increasing evidence from experimental work and analytical studies of natural samples showing that inclusions hosted in quartz – a ubiquitous host in many ore-forming systems – can experience post-entrapment modification. The contents of small monovalent cations, e.g. Cu, can be enhanced up to several orders of magnitude in vapor-dominated fluid inclusions by diffusional exchange for hydrogen. In this study, I focus on chemical analysis of fluid inclusions hosted by various minerals that co-precipitated with quartz and host the same fluid, in order to compare their compositional consistency and address existence and extent of post-entrapment modification. The fluid inclusions were analyzed by laser ablation-inductively coupled plasma mass spectrometry (LA-ICP-MS) and I explore analytical possibilities and challenges when applied to analysis of ultra-trace quantities of elements in small amounts of fluid, such as Au contents of magmatic vapor. I use representative samples from mineralization settings at convergent plate boundaries, namely granite-related Sn-W deposit (Mole Granite, Eastern Australia) and porphyry Cu deposit (Bingham Canyon, Utah, USA). In these systems, highly saline brines are important carriers of chloride-complexed metals and coexist with sulfur-rich vapors containing sulfide-complexing metals, e.g. As, Au, and, arguably, to various degree Cu. In the Mole Granite Sn-W deposit in Eastern Australia brine and vapor coexist in hydrothermal quartz, topaz, and beryl. Fluid inclusions in these three minerals show consistent salinities of 28–39 wt. % NaCleq in brine and 1–6 wt. % NaCleq in vapor at formation temperatures above 380 °C. These two fluid types have nearly identical concentrations of trace elements such as B, Na, K, S, and Cs, but reveal significant discrepancies in Cu abundances. The Cu concentrations in the quartz-hosted vapor inclusions are in the range of 5–50 000 ppm, in contrast to 20–2000 ppm Cu found in topaz- and beryl-hosted vapor inclusions. Partition coefficients of Cu between brine and vapor recorded by the topaz- and beryl-hosted fluids are in agreement with experimental results for magmatic-hydrothermal conditions, and in favor of the brine. The quartz-hosted fluids, however, show partitioning coefficients of Cu that exceed those found in all available experiments in sulfur-bearing systems. The vapor inclusions preserved in quartz are thus interpreted to have undergone significant post-entrapment modification, facilitated by their initially high sulfur content. By contrast, Au in the brine and vapor inclusions has been detected as erratic values only, despite that the detection limits of Au achieved by the LA-ICP-MS analysis were as low as single ppb. This extreme depletion in Au has been addressed by thermodynamic modeling indicating that saturation of the fluid with liquid bismuth can cause effective and essentially complete removal of Au from the fluid, and precipitates both metals as Bi-Au alloy. In the Bingham Canyon Cu-Mo-Au deposit (Utah, USA), garnet, pyroxene and quartz coexist in carbonate-replacement skarns that trapped brine, vapor as well as initial single-phase fluid prior to its separation. The single-phase intermediate-density fluid with an average salinity of 15 wt. NaCleq was trapped above approx. 550 °C and 800 bar. Salinities of brine and vapor inclusions are 32–42 and 2–7 wt. % NaCleq, respectively, and they formed at 420–380 °C and approx. 200 bar. The characteristics of corresponding fluid types in garnet, quartz, and to some degree in pyroxene remain nearly identical including their metal and other trace element contents. With improved analytical configuration, reliable Au signals were obtained from the fluid inclusions. Both Cu and Au have approximately equal content relative to Na in vapor and brine. This contrasts with previous fluid analyses from the porphyry quartz stockwork suggesting that majority of Cu and Au was transported by vapor. The new data suggest that the Cu concentrations in the vapor from the quartz stockwork were likely enhanced after inclusion entrapment, while the Au concentrations in the vapor phase require verification using lower instrumental detection limits. The Cu, Au and S concentration in the fluids from the porphyry quartz stockwork are consistently higher (0.2–5 ppm Au, 0.05–3 wt. % Cu) than those from the skarn minerals (0.02–0.4 ppm Au, 0.006–0.3 wt. % Cu), in contrast to ore grades that vary in an opposite manner (Au/Cu of 10-5 in the porphyry vs. 5∙10-5 in the skarn). This dichotomy was likely caused by chemical reactivity of carbonate precursor to the skarns and fluid focusing promoted by porosity and permeability enhancement during its decarbonation. In broader context, post-entrapment modification of vapor inclusions appears to be related to acidic and sulfur-rich fluids such as those preserved in granite- or porphyry-hosted pathways. By contrast, skarn-forming reactions appear to effectively neutralize acidity and together with low sulfur concentrations, the skarn-hosted fluids lack the principal driving forces that allow chemical-potential coupling and metal diffusion between fluid and host mineral, hence control the extent of post-entrapment modification. This study illustrates novel approaches to analysis of ultra-low metal concentrations in small volumes of fluids and analysis of fluid inclusions in non-traditional silicate minerals. I have identified and addressed several outstanding challenges of the LA-ICP-MS analytics, specifically effects of contamination on reliable analysis of elements close to their detection limits, strategies for reduction of the analytical data, and estimation of average concentrations for fluid inclusion populations.

This thesis comprises three papers describing the formation of the world class Bajo de la Alumbrera porphyry copper deposit in NW Argentina. Combining high precision zircon geochronology with geochemistry we have constrained the timescales of porphyry copper deposit formation and established a temporal and genetic link to nearby explosive volcanism. Finally, this link to explosive volcanism provides a unique opportunity to study the volcanic deposits to reveal key insights into the formation of a world class porphyry copper deposit. Chapter 1 provides a introduction and overview to the objectives of the thesis and details of the current state of research in both geochronology and the genesis of porphyry copper deposits. Chapter 2: From a long-lived upper-crustal magma chamber to rapid porphyry copper emplacement: Reading the geochemistry of zircon crystals at Bajo de la Alumbrera (NW Argentina), presents new high precision zircon U-Pb geochonological data in order to constrain the timing and duration of porphyry copper deposit formation. We calculate the duration of ore-formation by dating zircons from pre-, syn- and post-mineralisation porphyries based on intersection relations between successive intrusion and vein generations, using high precision chemical-abrasion isotope-dilution thermal ionisation mass spectrometry (CA-ID-TIMS). Based on the youngest assemblages of zircon grains, which overlap within analytical error, all four intrusions were emplaced within 29 ka, which places an upper limit on the total duration of hydrothermal mineralisation. Re/Os dating of hydrothermal molybdenite fully overlaps with this high-precision age bracket. By examining the geochemistry of zircon antecrysts, which record ~200 ka of magmatic evolution, we show that that the four porphyry intrusions record a common geochemical and thermal history, and that they were derived from the same upper-crustal magma reservoir. During this upper crustal magmatic evolution, the presence of low cathodoluminescent zones ("dark-zones") in some zircon crystals record zircon crystallisation during a transient event of rapid growth. This rapid growth resulted from mafic magma injection into the base of the magma chamber, releasing a CO2-rich vapour phase into the dacitic crystal mush. This vapour phase then migrated upwards to the apical part of the magma chamber from where it was expelled, together with successive batches of magma, to form the porphyry copper deposit within a short time-span of less than a few 10,000 years. The short duration of host rock emplacement, hydrothermal alteration and mineralisation presented in this study provides critical constraints on fluid storage in magma chambers and the genesis of large porphyry copper deposits. This chapter is published in Earth and Planetary Science Letters. Chapter 3: Zircon petrochronological evidence for a plutonic-volcanic connection in porphyry copper deposits, presents new high-precision U-Pb zircon geochronological evidence linking the formation of the Bajo de la Alumbrera porphyry copper deposit to nearby explosive volcanism. The results indicate that the magmatic-hydrothermal deposit and 245 m of explosive volcanism were derived from a common magma reservoir that evolved over a minimum duration of 217 ± 25 ka before the fi nal eruption. A time-gap of 19 ± 12 k.y. between porphyry mineralization and the onset of explosive volcanism indicate that these distinct events occurred as the result of a common magmatic process. We propose that deposit formation and volcanic eruption were both triggered by the same injection of a volatile-saturated primitive magma into the base of the magma chamber. Catastrophic loss of volatiles by explosive volcanism terminated the ore forming capacity of the upper crustal magma chamber, as evidenced by the intrusion of a syn-eruptive barren quartz-feldspar porphyry. Our results demonstrate that porphyry copper deposits provide critical information to understand how volatiles control the fate of hydrous magmas, between pluton formation and explosive volcanism. This chapter is published in Geology. Chapter 4: Magmatic petrogeneseis leading to the formation of the Bajo de la Alumbrera porphyry copper deposit: Insights from cogenetic explosive volcanism, builds on the findings of Chapter 3 by examining the magmatic petrogenesis of the explosive volcanism that terminated the ore formation at the Bajo de la Alumbrera porphyry copper deposit. We examined amphibole and apatite crystals from the entire volcanic sequence to track the magmatic evolution during the time preceding ore formation. The presence of two distinct generations of amphibole crystals from the same volcanic rocks reveals records both a deep (15 - 25 km) and shallow (~10 km) crustal evolution of porphyry forming magmas. The upper crustal evolution is revealed from amphibole and apatite crystals, which indicate the co-crystallisation of plagioclase, amphibole, biotite, apatite, titanite and zircon in an upper crustal magma reservoir, which evolved over ~200 ka. During the upper crustal evolution, the magma was H2O saturated and exsolving a magmatic fl uid. The melt compositions in equilibrium with apatite reveal the upper crustal magma was not enriched in S, relative to other arc magmas. Conversely, amphibole compositions reveal low Cu concentrations (~10 ppm) in the melt at the time of eruption, which suggest either the upper crustal magma reservoir had inherently low Cu concentrations or the formation of the Bajo de la Alumbrera porphyry copper deposit (~19 ka prior to explosive volcanism) effectively stripped the upper crustal magma reservoir of Cu. The lower crustal record is preserved in mafic crystal clots that appear towards the upper parts of the 7 Ma volcanic sequence. The presence of olivine pseudomorphs, which are replaced by clinopyroxene and amphiboles, together with the absence of plagioclase corroborates their crystallisation in the lower crust under hydrous conditions. High fO2 (ΔNNO > 3), revealed from the amphibole and Cr-rich spinel compositions, likely resulted from extended olivine crystallisation. Injection of this oxidised, hydrous, magma into the upper crustal magma reservoir at a late stage is recorded by abundant dissolution and reverse zoning in the major phenocryst phases towards the upper parts of the 7 Ma volcanic sequence. Our results demonstrate that the interplay between the injection of oxidised magmas from the lower crust and high Cl contents in the upper crustal magma chamber, promoted the efficient extraction of Cu and S from the melt to form the world class Bajo de la Alumbrera porphyry copper deposit. Chapter 5 presents the general conclusions of this thesis including the potential for future work.