Energetics of a two-phase model of lithospheric damage, shear localization and plate-boundary formation

cited By 102International audienceThe two-phase theory for compaction and damage proposed by Bercovici et al. (2001a, J. Geophys. Res., 106, 8887-8906) employs a nonequilibrium relation between interfacial surface energy, pressure and viscous deformation, thereby providing a model for damage (void generation and microcraking) and a continuum description of weakening, failure and shear localization. Here we examine further variations of the model which consider (1) how interfacial surface energy, when averaged over the mixture, appears to be partitioned between phases; (2) how variability in deformational-work partitioning greatly facilitates localization; and (3) how damage and localization are manifested in heat output and bulk energy exchange. Microphysical considerations of molecular bonding and activation energy suggest that the apparent partitioning of surface energy between phases goes as the viscosity of the phases. When such partitioning is used in the two-phase theory, it captures the melt-compaction theory of McKenzie (1984, J. Petrol., 25, 713-765) exactly, as well as the void-damage theory proposed in a companion paper (Ricard & Bercovici, submitted). Calculations of 1-D shear localization with this variation of the theory still show at least three possible regimes of damage and localization: at low stress is weak localization with diffuse slowly evolving shear bands; at higher stress strong localization with narrow rapidly growing bands exists; and at yet higher shear stress it is possible for the system to undergo broadly distributed damage and no localization. However, the intensity of localization is strongly controlled by the variability of the deformational-work partitioning with dilation rate, represented by the parameter γ. For γ ≫ 1, extreme localization is allowed, with sharp profiles in porosity (weak zones), nearly discontinuous separation velocities and effectively singular dilation rates. Finally, the bulk heat output is examined for the 1-D system to discern how much deformational work is effectively stored as surface energy. In the high-stress, distributed-damage cases, heat output is reduced as more interfacial surface energy is created. Yet, in either the weak or strong localizing cases, the system always releases surface energy, regardless of the presence of damage or not, and thus slightly more heat is in fact released than energy is input through external work. Moreover, increased levels of damage (represented by the maximum work-partitioning f*) make the localizing system release surface energy faster as damage enhances phase separation and focusing of the porosity field, thus yielding more rapid loss of net interfacial surface area. However, when cases with different levels of damage are compared at similar stages of development (say, the peak porosity of the localization) it is apparent that increased damage causes smaller relative heat release and retards loss of net interfacial surface energy. The energetics and energy partitioning of this damage and shear-localization model are applied to estimating the energy costs of forming plate boundaries and generating plates from mantle convection