Grain Size and localization of deformation in the lithosphere

In this thesis, I studied the impact of grain size on the rheology of the mantle of the Earth. We propose a new model of grain size evolution and show that the equilibrium states it predicts are in good agreement with the experimental data obtained on olivine samples. This model is illustrated by numerical simulations of mantle convection and shear zones. I show that our formalism has a non-negligible effect on the localization of deformation in shear zones and strongly modifies the convection regimes of planets. In our formalism, we propose that the energy of a grain size distribution is fully constrained by the energy of the grain boundaries. This energy diminishes during grain growth and increases during dynamic recrystallization. Our model stipulates that the amount of energy required to nucleate new grains, which is a dislocation-assisted process, is substracted to the energy dissipated in the material. We also show that even when the mean grain size is located in the diffusion regime, the biggest grains of the distribution can remain in the dislocation regime and continue to nucleate smaller grains. Thus, the mean grain size can decrease, even in the diffusion regime. When the growth and recrystallization processes are equivalent, the grain size distribution reaches an equilibrium state which can be compared with experimental data (various published olivine piezometers). The experimental calibration of our model shows that the partitioning of energy is principally temperature-dependent. I test this new model in a fully viscous numerical simulation of shear zone with a non-Newtonian- grain size-dependent viscosity. In this system, depending on the temperature, grain size or thermal state, the deformation can locally operate in diffusion or dislocation regime. Paradoxally, my model shows that the deformation is more localized when the shear zone is in the dislocation regime, where the viscosity does not depend on grain size. We observe that the grain size parameter tends to favor the apparition of shear zones because it minors the stress field but is insufficient to localize deformation in the case of a constant plate velocity. A grain size-activated localization of deformation seems to require a time-dependent stimulation which is typically obtained in visco-elasto-plastic configurations. Finally, I have tested our grain size evolution model, in its static form, in a set of numerical simulations of mantle convection. Our model shows that the grain size reaches its equilibrium very quickly in sub-lithospheric conditions. Thus, I have considered that the grain size distribution is always in the equilibrium state predicted but our model. I show that the consideration of the grain size parameter can fully modify the convection regimes of telluric planets. When surface stresses are high enough to reach the diffusion regime, the stagnant lid usually obtained with very temperature-dependent viscosities is broken by a very non-Newtonian and grain size-dependent top layer. This thesis shows that grain size strongly influences the dynamics of planets in localizing the deformation in their surfaces. The model we propose brings a new view of the creation of plate tectonics at the surface of the Earth.