The creep in the Earth’s interior is dominated either by diffusion creep or by dislocation creep results. If it is dislocation creep, the strain rate will be proportional to stress to the power of 3.0-3.5, leading to non-Newtonian viscosity. The dominant slip system in dislocation creep will produce lattice-preferred orientation in olivine, which causes seismic anisotropy. In contrast, if diffusion creep dominates, both lattice-preferred orientation and shape-preferred orientation will be produced in olivine and strain rate proportional to stress will result in Newtonian viscosity. A series of experimental deformation studies indicate that the dominant creep mechanism in the upper mantle transition from dislocation creep in the shallow lithosphere to diffusion creep in the asthenosphere with a transition depth of 200-250 km (e.g. summarized in Hirth and Kohlstedt, 2003; Karato and Wu, 1993). However, based on silicon and oxygen diffusion experiments, Fei et al. (2012; 2013; 2014) suggest that the aforementioned deformation studies have overestimated the water and pressure effects on creep rates owing to experimental limitations. It is therefore necessary to systematically re-examine the creep mechanisms in the upper mantle.
Since olivine creep is controlled by Si self-diffusion within grain interior and along the grain boundary, we systematically measured the pressure, temperature, and water content dependence of silicon grain boundary diffusion coefficient in iron-free forsterite aggregates (Fig. 1) and calculated the diffusion and dislocation creep rates from the Si grain-boundary diffusion coefficients determined in this study and the lattice diffusion coefficients from our previous studies (Fei et al. 2012; 2013). The results suggest that diffusion creep dominates in the shallow and cold lithospehre, whereas dislocation creep dominates in the deeper asthenosphere. There is a transition from diffusion to dislocation creep at 50-80 km depth beneath oceans and at 100-150 km depth beneath continents. This transition well explains the seismic anisotropy jumps observed at the mid-lithosphere discontinuity in the continental mantle and at the Gutenberg discontinuity near the oceanic lithosphere-asthenosphere boundary (Fig. 2).
Fig. 2 Creep mechanisms in the upper mantle inferred from DSilat and DSigb. “Diffusion” and “dislocation” indicate regions where diffusion and dislocation creep dominate, respectively.
H.-Z. Fei, S. Koizumi, N. Sakamoto, M. Hashiguchi, H. Yurimoto, K. Marquardt, N. Miyajima, D. Yamazaki & T. Katsura, New constraints on upper mantle creep mechanism inferred from silicon grain-boundary diffusion rates, Earth and Planetary Science Letters 433, 350-359, 2016.