On modifications of fluid inclusions in quartz : re-equilibration experiments and thermodynamical calculations on fluids in natural quartz

Fluids in rocks can be traced to great depths, and are found in crustal rocks as well as in mantle rocks. Information about the deep fluid which is obtained from fluid inclusions must be handled with care, for the way up after entrapment in a crystal is long and full of interferences at different pressures and temperatures. The fluid system may be a very complex mixture (H₂0, CO₂ , CH₄ , N₂, CO, NH₃, and H₂ ) with distinct physical and chemical proporties. Chapter 1 indicates that many types of fluids have been recognized in rocks, and that each type may be related to a distinct grade of metamorphism (eg. Sorby 1858). For example, the type of fluid recognized in inclusions from very deep rocks, like granulites and upper mantle rocks is pure CO₂, Many interpretations of fluid inclusions are based on the assumption that fluid inclusions do not change in density and composition after entrapment. In recent years it was recognized that fluid densities may change during the long way up, and some voices were heard argueing compositional changes. In this study (Chapter 2 and Chapter 3) it is experimentally proven that CO₂-H₂0 fluid inclusions, which are synthesized in Brazilian Quartz at 835 K and 200 MPa, leak preferentially H₂0 during reequilibration at different conditions. The inclusions are submitted to an internal overpressure and underpressure. The remaining inclusions become relatively enriched in CO₂, and the H₂0 must have diffused through the crystal, for there are no traces of cracks. Several underpressurized inclusions have "implosion halos" after re-eQuilibration, which consist of many small secondary inclusions. Transmission electron microscopy (TEM) observations of the reequilibrated Brazilian Quartz and of Quartz lenses from Naxos, Greece, (Chapter 4) indicate that dislocations adjacent to fluid inclusions account for the non-decrepitative preferential H₂0 leakage. Ductile strain of the Quartz neighbouring inclusions, which are overpressurized and underpressurized, is expressed in the localized occurrence of many dislocations. Selective diffusion of H₂0 along dislocations, which are assumed to be inaccessible for CO₂, results in the observed preferential H₂0 leakage. Additionally, a solubility gradient of quartz in water between an underpressurized fluid inclusion and a dislocation cause the growth of small secondary bubbles at dislocations to optically visible ftimplosion halos" at the expense of the fluid inclusion. The fluid which should occur at great depths can be indirectly obtained through thermodynamical calculations. The theoretical background of these calculations is described in Chapter 5. The well documented metamorphism from Rogaland (SW Norway) provides enough information to proceed these fluid calculations (Chapter 6). The calculated gaseous fluid inclusions are similar to the observed natural fluid inclusions. However, the calculated aqeous fluids are not recognized or registered by natural fluid inclusions. The absence of H₂0 in low density metastable CO₂ -CH₄ -rich inclusions with graphite, and the occurrence of CH₄-N₂-rich inclusions are both proposed to result from preferential H₂0 leakage. The occurrence of CO₂ rich inclusions in granulites and mantle rocks may partly be caused by the fact that before exposure, these deep rocks have usually suffered more intensive recrystallization during uplift than shallow rocks. In future studies on fluid inclusions the effect of deformation on aqueous inclusions can not be ignored