NATURAL AND EXPERIMENTAL CONSTRAINTS ON LITHIUM INCORPORATION AND DIFFUSION IN GARNET AS AN INDICATOR OF FLUID ROCK INTERACTION

Abstract

Fluid flow in the deep crust is a major avenue of mass transport and impacts rheology, the generation of arc magmas, and contributes to the formation of ore deposits. Evidence for fluid infiltration events can be recorded by elemental zoning within garnet. One example is lithium, a fluid-mobile element that diffuses rapidly and can act as a tracer of fluid-rock interactions in metamorphic terrains. While whole-rock lithium concentrations have been used to determine the timescales of short-lived events including time-integrated fluid event durations in metamorphic settings, resolving individual fluid pulses requires detailed in situ measurements of lithium concentrations across metamorphic minerals. Interpreting timescales of fluid infiltration in metamorphic rocks on the basis of variations in lithium concentration in garnet requires exploration of lithium in metamorphic garnet as well as quantitative knowledge of lithium in garnet diffusion coefficients (D) that have not yet been experimentally determined. This study takes a two-pronged approach at (1) deciphering the distribution of lithium in garnets from metamorphic terrains that have experienced fluid-rock interaction and (2) constraining the diffusivity of lithium in garnets of varying composition at different temperature and oxygen fugacity conditions. For the natural rock record, four metamorphic localities were selected to represent a range of pressure and temperature conditions, tectonic settings, and fluid histories: Waits River, VT; Monviso, Italy; Erzgebirge, Germany; and Catalina, CA. Lithium concentration data were collected on garnets from these localities via LA-ICP-MS, in the form of traverses and quantitative mapping. The results of those analyses reveal supporting evidence of fluid infiltration events and show correlations between lithium and other trace and rare earth elements. To determine the diffusivity of lithium in garnet, two types of powder source diffusion experiments were performed; box furnace experiments using a sealed silica tube method, and gas mixing furnace experiments using a suspended, open platinum capsule. Garnet starting material and experimental run products were analyzed for their lithium concentration using LA-ICP-MS and SIMS depth profiling. The sealed silica tube experiments used polished almandine-pyrope, grossular, and spessartine grains surrounded by a lithium rich powder (spodumene or Li-doped crushed garnet) at temperatures ranging from 600 to 800 °C for durations of 1 to 8 weeks. The garnet starting material (control) displayed a uniform lithium concentration from surface to interior. The experimental run-product garnets exhibited an enrichment in lithium of up to 100s ppm at the rim followed by a smooth decrease towards the interior over length scales on the order of 1-8 μm, indicative of diffusive uptake. One-dimensional diffusion modeling was performed to calculate a best-fit D for each temperature. The data were best fit by a model with two independent mechanisms, which has been postulated for Li in olivine and pyroxene and theorized for lithium in garnet. The diffusivities of both mechanisms are slower than lithium diffusion in olivine along interstitial sites, comparable to diffusion in olivine through metal vacancies, and faster than lithium+REE coupled substitution into garnet. The gas mixing furnace experiments were conducted at 800 °C with ratios of CO:CO2 that correspond to FMQ+2 and FMQ-2, as well as in air (FMQ+13). The FMQ+13 run products display a typical, exponentially decreasing diffusion profile that can be modeled using the same two-mechanism model as the box furnace experiments. Results from the FMQ+2 and FMQ-2 experiments show a more complex diffusion behavior, where the lithium concentration is elevated at the surface of the grain, increases until it reaches a local maximum and then decreases until it reaches the internal lithium concentration of the garnet. This “peaked” curve can best be explained by a “lithium incorporation and loss” model, where lithium is moving both in and out of the garnet along different pathways activated at reduced fO2 conditions. These results appear to be the first experimentally derived diffusion coefficients for lithium in garnet and can be used to constrain the duration of individual fluid flow events in metamorphic settings.