V24C-01 INVITED
The Effect of Growth Kinetics on the Development of Element- and Isotope Profiles in Single Mineral Grains
The formation of a rock texture is the consequence of multiple interacting processes controlling the crystallization history of minerals, and is typically recorded in several ways and on different scales. Recent advances in microanalysis enable observation of compositional (elemental as well as isotopic) variations on progressively decreasing (sub-mm) scales, and theoretical approaches for simulating realistic, disequilibrium kinetic effects are becoming increasingly sophisticated. Here we present some new perspectives arising from the continuing development of our numerical approach for describing kinetically controlled element uptake and isotope fractionation during diffusion-controlled mineral growth. We assume a spherical grain growing under local equilibrium at the interface with a spherical matrix of a given size and composition. It has been shown previously that the uptake of a component in a growing crystal depends critically upon the ratio of the crystal growth rate (R) to the diffusivity (D) of that component in the growth medium. Highly compatible (or incompatible) elements potentially will be depleted (or accumulated) near the surface of the growing crystal, forming a disequilibrium compositional boundary layer against the advancing interface. Equilibration of this boundary layer with the bulk reservoir depends mainly on D of the species of interest, and because D is now known to vary with mass of an isotope of a given element, signficant isotope fractionation is predicted to occur during progressive mineral growth under some circumstances. In the present study we focus specifically on the role of crystal-growth kinetics [R=f(time)] in controlling isotope profiles recorded in crystals. We show that the amount of fractionation, and especially the shape of the resulting radial isotope profiles, is quite sensitive to the assumed growth law or history. Differences in the initial grain size – as might apply in the case of xenocrysts in magma – have a significant effect on the amount of fractionation along a similar growth path due to the radial 'volume effect' in a spherical grain. A growth history that involves accelerating R (as in a rapidly cooling small intrusion) produces a radial isotope profile that differs significantly from that produced by a constant R, even for crystals grown to the same final size. In combination with microanalytical methods and experimentally determined rate data, this modeling approach may provide new insights into the mechanisms of mineral growth, as well as timescales and textural evolution in a variety of geological settings.
V24C-02
Combining textural and micro-geochemical analysis in igneous rocks: A Santorini case study.
Here we look at the role of textural and micro-geochemical analysis in the quantification of Igneous textures.
Textural analysis provides constraints on the crystal population in terms of size and shape and can help
constrain mixed crystal populations and crystal residence times. Micro-geochemical analysis can be used to
constrain isotopic fingerprints that crystals pick up in the magma system and also important timescale
information through diffusion and chemical exchange between crystals and magmas. In Santorini volcano we
have applied textural and micro-geochemical constraints to constrain magmatic processes at scales that run
over 10s of thousands of years in eruptive products from its main cycles, to processes occurring over a few
weeks in its recent eruptive products.
http://www.dougalearth.com
V24C-03
High Resolution Neutron Computed Tomography of Vesicular Pyroclasts: Interplay of Vesicles and Crystalline Phases
In the last decade great efforts have been made in experimental volcanology as well as in the field of analytical and numerical modelling in order to understand explosive volcanic eruptions. 2D textural analysis, e.g. bubble- and/or crystal-size distribution analysis has been an essential component of this progress. Yet reliable information on the spatial distribution of connected vs. isolated vesicle structures, especially with respect to the crystalline phase distribution remains poorly constrained. Image analysis of thin sections has provided insights, but such analysis always implies the transformation from a 2D to a 3D structure. That however is especially problematic if the analysed volcanic rocks exhibit a strongly varying size and shape of pore space and phenocrysts resulting in a complex 3D structure; as is often the case in silicic pyroclasts. Especially for large samples (15 - 50 cm3) neutron computed tomography provides the first non- destructive method to analyse this complex 3D structure. We applied high resolution (50 μm) neutron tomography to investigate the 3D structure of vesicular (45 - 72 %), silica rich pyroclastic material from different explosive, hazardous volcanoes. The experiments were performed at the ANTARES beamline of the research reactor FRM II of Technische Universität München in Garching, Germany. Volume reconstructions of the pore space relative to the different crystal phases were calculated with special software from the 2D image slices obtained by tomographic scanning. The reconstructed volumes enabled us to test the applicability of this technique, novel in the field of volcanology, to volcanic rocks with different textural characteristics. Neutron tomography reacts sensitively to hydrogen (and thus water) and permits the distinction of different mineral phases. Within the samples analysed, interplay of the vesicle distribution with the crystal distribution could be observed. We will argue that this type of analysis yields valuable constraints on the degassing of active volcanoes.
V24C-04 INVITED
Mechanism of Formation of Snowball Garnets Revealed Through EBSD and Chemical Analyses
For two decades, investigators have sought to determine whether or not garnets with spiral inclusion trails rotate with respect to an external, geographic reference frame. Prior studies have mainly relied upon microstructural observations, but have failed to resolve the question, mainly because most geometries observed in snowball garnets can be explained by both models. Clearly, additional data from different approaches are needed to breach the impasse. In this study, chemical analysis and electron backscattered diffraction analysis (EBSD) were used on two populations of snowball garnets from the Lukmanier area (central Swiss Alps) to generate new classes of data that can be used to better identify the different stages associated with the formation of such microstructures. The garnets from the first population are characterized by a spiral geometry exhibiting an amount of apparent rotation of 360 degrees. In this population, the spiral geometry of the inclusion trails evolves towards a crenulated inclusion pattern in the last 90 degrees of the spiral curvature indicating that the last part of the garnet growth history was not accompanied by the rotation of the garnet crystals. EBSD maps performed on the same sections reveal a crystallographic central domain exhibiting 270 degrees of rotation and smaller garnet grains in the last 90 degrees of the spiral curvature. Chemical data indicate that the modification of the inclusion pattern occurred for a XMn of 0.009, contemporaneously with the modification of the crystallographic orientation. In contrast, the second population exhibits a spiral geometry that does not exceed 270 degrees. These garnets do not record any evidence for a modification of the stress-field regime during garnet growth, and a single crystallographic orientation is observed for the entire spiral. Microstructural observations, chemical mapping, and EBSD data all support the idea that the crystallographic orientation in the snowball garnets is controlled by the local stress conditions operating during garnet nucleation. This interpretation leads in turn to two related inferences: (1) Snowball garnets that form by a mechanism involving continuous rotation of the porphyroblast during growth exhibit a single crystallographic domain for the whole spiral; (2) Conversely, the incremental superimposition of foliations resulting from changes in the stress field (as postulated by the non rotational model) leads to the formation of snowball garnets characterized by several domains of different crystallographic orientation, reflecting variations in the operative local stress during growth.
V24C-05
Integrating crystallographic data and phase equilibria to quantify P-T-X evolution during reaction texture formation
Coronal symplectitic reaction textures occur as a result of changes in intensive variables. This variation can arise as a result of changes in pressure and/or temperature or result from modification of bulk composition due to an influx of fluids. These processes lead to development of chemical potential gradients that drive diffusion and are responsible for the vermicular nature of symplectitic reaction textures. Deducing the P-T conditions of reaction and the P-T-X path responsible for texture formation is a difficult but critical step in interpreting the crystallization history of symplectites as well as providing appropriate boundary conditions for modeling texture development. Symplectite textures in gedrite-cordierite rocks from Thor-Odin gneiss dome in British Columbia, Canada preserve spl+crd, an+crd, and crn+crd two-phase assemblages after sillimanite porphyroblasts. These two-phase assemblages are not present as a consistent progression of layers as in other examples of symplectitic textures, but occur in a variety of locations with respect to the central sillimanite porphyroblast that are also unrelated to adjacent matrix mineral assemblages. This suggests that the chemical potential gradients responsible for symplectite formation are not consistent around the texture. The two-phase symplectitic assemblages are encased by a rim of polygonal cordierite. These inconsistent relationships make proper interpretation of the relative timing of symplectite and cordierite rim growth, as well as establishing the P-T-X conditions and kinetics of reaction difficult using traditional methods. The integration of mineral chemistry, phase-equilibria, crystallographic analysis and image analysis has provided a method of determining the P-T conditions at which symplectite formation began as well as providing information on how the size and nature of the chemical system evolved during reaction and growth. EBSD data from cordierite rims and the sillimanite porphyroblasts show a consistent c-axis alignment around the texture providing evidence that the cordierite grew epitaxially on Al2SiO5 and prior to symplectite formation. Pseudosection analysis of the gedrite-cordierite rocks places this crystallization relationship as occurring at ~6.5 kbar and 750 C. The initiation of symplectite formation also occurred at these conditions. The shift from monophase coronal growth to the subsequent inconsistent development of different two-phase symplectitic assemblages around sillimanite suggests a fundamental shift in the kinetics (e.g. cation diffusion rates) and/or size of the chemical system responsible for reaction texture formation.
V24C-06
Textural Evidence for Metastable and Non-equilibrium Processes in Meta- Pelites
Many metapelites contain textures with porphyroblasts that surround well shaped earlier crystals which should be consumed according to reactions that are typically invoked to explain the origin of the whole-rock assemblage. A common example is staurolite porphyroblasts (or poikiloblasts) enclosing well formed garnets that do not appear to have dissolved during staurolite formation, even though the typically inferred equilibrium reaction is staurolite growth at the expense of garnet and chlorite. Another example is idioblastic staurolite surrounded by andalusite when the inferred andalusite forming reaction should have consumed staurolite. These relationships suggest that the reaction mechanisms which formed the porphyroblasts are not the reactions inferred from calculations based on bulk chemistry that presume equilibrium assemblages are present in the rock throughout its history. However, the textures can be explained by the processes that occur due to strongly overstepped reactions in contact and regional-contact environments. In these types of situations, reaction mechanisms controlled by meta-stable equilibria can lead to growth of porphyroblasts from reactions that are fundamentally different from those inferred using an equilibrium approach.
V24C-07
Kinetics of Nucleation and Intergranular Diffusion Determined From Numerical Simulations of Crystallization in Regionally Metamorphosed Rocks
Numerical simulations of porphyroblast crystallization in regionally metamorphosed rocks constrain key, immeasurable parameters that control rates of nucleation and intergranular diffusion. The relative rates of these two processes are fundamental determinants of the size and spatial disposition of crystals in metamorphic rocks. Modeling these features of natural porphyroblastic textures allows quantification of these rates. We simulate prograde crystallization in the case of diffusion-controlled nucleation and growth by modeling diffusion through the intergranular fluid of a single rate-limiting component (Al) from a reactant mineral (chlorite) to the growing crystals (garnet). Nucleation rates depend on temperature and Al supersaturation in the intergranular fluid, which is buffered by reactants but depleted near products. Porphyroblast size depends on nucleation time, local reactant abundance, and intergranular diffusion of nutrients from distant sources. The rate of effective intergranular diffusion is given by an Arrhenius expression and additional parameters for Al concentration (1 mol cm-3 at 550 °C and 0.4 GPa) and porosity along interconnected grain edges (1x10-5). Nucleation rate at maximum supersaturation of Al is calculated from a theoretical expression based in transition state theory, which produces an exponential increase in nucleation rate to a maximum value. Agreement between a simulation and the natural rock is evaluated from statistical measures of mode, mean crystal size, crystal number density, crystal-size distribution, the degree of ordering of crystal centers, and the correlation between crystal size and isolation. By simulating crystallization in three samples, initial values have been determined for the kinetic parameters governing effective diffusion and nucleation. Diffusivity is 1.0×10-10 to 6.7×10-10 m2 s-1, maximum nucleation rate is between 1.50×10-13 and 1.15×10-11 nuclei cm-3 s-1, and nucleation acceleration is 0.001 to 0.080 (dimensionless). These results provide the first constraints on the key parameters that control nucleation and growth of crystals in regionally metamorphosed rocks.
V24C-08
Experimental Determination of Mechanisms and Rates of Fe-Mg Exchange Between Spinel Grains Mediated by a Fluid Phase
The overall mechanism and kinetics of mineral reactions results from a complex interaction of several processes such as surface reaction kinetics, volume diffusion and net transfer. In order to quantify the kinetics of reactions involving multiple phases in multicomponent systems, it is necessary to understand and characterize the nature and rates of each of these processes. Most laboratory experiments up to now have focused on kinetics of reactions where the reactants and products are in direct physical contact with each other. However, there is abundant textural evidence in rocks that reactions occurred between mineral grains that are physically separated from each other, frequently mediated by a fluid phase. We have devised an experimental setup to study the mechanism and kinetics of such reactions in the laboratory. Polished single crystals of two spinels (synthetic MgAl2O4 and a natural spinel with 44 mol% Hercynite component), 2mm on a side, were placed in a gold capsule (length: 2cm, diameter: 4mm) separated from each other by a 5mm long tube of Au or alumina. The capsule was welded shut after adding distilled water (80-100μl). Such capsules were annealed (2 Kbar, 700-750°C, up to 21 hours) in hydrothermal cold seal vessels. After annealing the crystals were cleaned in an ultrasonic bath in order to rinse them of possible quench products. The surfaces were examined optically and near surface chemistry was determined using Rutherford Backscattering Spectroscopy (RBS). We observe time dependent changes in the morphology as well as the chemistry of the crystals, as follows: After short times, the surface of the Mg spinel shows scattered etch pits while terraces form on the Fe spinel. After longer anneals, the etch pits disappear and the surface of the Mg spinels appear polished. Surface compositions are found to be different, depending on whether a Au or alumina separator was used in the experiments. The Fe rich spinel composition remains unchanged whereas the Mg spinel shows erratic gains in Fe (He1-30) when a Au separator is used. The results are more systematic with an alumina separator and the surfaces of both spinels show the same equilibrated compositions (He20-22). No concentration – depth profile could be resolved in the Fe rich spinel, whereas systematic diffusion profiles develop in the Mg rich spinel. Fits to these concentration profiles yield Fe-Mg diffusion coefficients of about 5e-19 m2/s (experiments with Au-separator) and 2e-18 m2/s (alumina separator), respectively, at 750°C. We interpret the difference in behavior between the Au-separator and the alumina-separator experiments to result from a more controlled activity of Al2O3 in the fluid in the latter experiments. Based on these observations we conclude: (i) The initial fluid that is out of equilibrium with both spinels attacks the surface of both but in different manners – forming etch pits in one case and terraces in the other. (ii) Once a steady state concentration distribution is attained in the fluid and it is in local equilibrium with each crystal, further element exchange appears to operate by diffusion in the Mg spinel and surface reaction in the Fe spinel. (iii) The diffusion profile that develops in the Mg spinel is a measure of reaction progress and allows characterization of the kinetics of the overall reaction. While more data are required to fully characterize this particular system, the success of this set up opens the possibility of exploring the kinetics and mechanism of various exchange and net transfer reactions.