A new methodology for the study of the magmatic-hydrothermal transition in felsic magmas: applications to barren and mineralised systems

This study aims to develop a robust research methodology to examine the evolution of magmatic volatile phases during the cooling of felsic magmas via detailed melt- and fluid-inclusion studies, in particular the investigation of inclusions originally containing both melt and aqueous fluid. Then, using these techniques I will examine fluid immiscibility processes in two felsic magmatic systems, one mineralised, the other barren. In particular, I address the constraints on the exsolution of magmatic vapour and aqueous liquids, and how it is manifested in quartz-hosted inclusions, as well as the nature and composition of the exsolved phases. In developing a research philosophy two factors need to be paramount, it needs to be as widely applicable as possible, and the limitations need to be recognised and explored. Thus, the results deriving from these techniques may provide a test of the methodology. The thesis is based on two case studies, Rio Blanco (Chile) and Okataina (New Zealand). The first case study involves sub-volcanic intrusives and associated extrusives from the La Copa Rhyolite, and intrusives from the Don Luis Porphyry, two post-ore rhyolitic suites from the Los Bronces-Rio Blanco Porphyry Cu-Mo deposit. The second case study involves rhyolitic lavas (< 65 Ka) from the Okataina Volcanic Centre in the Taupo Volcanic Zone in New Zealand. This study is not intended to examine the geology of these systems, but rather to use them as examples of felsic systems, in diverse tectonic settings. Both as test cases for developing robust research techniques and for any information that they can provide regarding late-stage magmatic processes, particularly volatile phase exsolution, and the role of melt/fluid and liquid/vapour immiscibility. At Rio Blanco, the melt inclusion populations consist predominantly of glass inclusions and coexisting dark, inhomogeneous crystalline silicate melt inclusions (CSMI's). An important discovery from this study is the recognition that CSMI's trap volatile-rich melt, probably identical to the melt trapped as glass inclusions, and are crystallised, not "devitrified" or the product of post-magmatic alteration. Heating experiments demonstrate that both the glass and CSMI's from Rio Blanco have decrepitated and degassed post-trapping, notwithstanding the apparent lack of petrographic indicators of degassing in the glass inclusions. This coexistence appears to be a common occurrence; however, its significance seems to have been overlooked in a number of previous studies. From an initial volatile-rich melt, aqueous volatile phases (dominantly vapour) exsolved, forming bubbly magmatic emulsions. Inherently, magmatic emulsions are metastable, and disrupt into discrete melt and vapour phases. The vapour-rich phases separated from the melt and escaped, cooling, condensing, and mixing as they did so. Rio Blanco melt inclusions and fluid inclusions trapped all of these phases, in various combinations, both demonstrating the process in fine detail, and sampling the phase compositions. Analysis of the phases demonstrates partitioning of metals (Cu, Zn, and possibly Pb) into the vapour phase, its transport out of some of the magma bodies, and implies concentration by mixing and condensing to form metal-rich hypersaline fluid inclusions in the carapace of the Don Luis Porphyry. The Okataina case study provided an invaluable counterpoint to Rio Blanco. Phenocryst crystallisation pressures were supercritical, although the evidence suggests that volatile phase exsolution (VPE) occurred post- rather than pre-trapping, so that trapped magmatic emulsions are not observed. Okataina also contains coexisting CSMI's and glass inclusions, although many of the samples contains a complex array of partly crystalline silicate melt inclusions. Importantly for this study, many of the inclusions do homogenise during experimental heating, indicating that decrepitation and degassing were not as pronounced as at Rio Blanco. Heating experiments showed that despite coexisting CSMI's and glass inclusions, there was only a single melt trapped. This provides evidence of the post-trapping behaviour of melt inclusions, lacking at Rio Blanco. Although pre-trapping VPE did not occur to a large degree, post-trapping VPE did. Inclusions in which exsolution of an aqueous volatile phase has occurred provide a measure and sample of the amounts of fluids that were exsolved from a known quantity of melt, and may provide a method of determining the actual amounts of hydrothermal fluids that a magma body may exsolve. In evaluating these results some inevitable limitations of the techniques have been uncovered, particularly those relating to the vagaries of melt inclusion formation and preservation, and these have been evaluated. However, qualitative, and to some extent quantitative results have been produced, some of which have been published in research journals. Together, the case studies demonstrate and sample the fine detail of the exsolution of volatile-rich phases from silicate melts, their escape from those melts, and eventual cooling and condensing to form the kinds of hypersaline hydrothermal fluids found as fluid inclusions in ore-bodies. Further, they provide insights into common post-trapping behaviours of melt inclusions, some aspects of which appear to have been misinterpreted in some published melt inclusion studies.