Microscale processes in experimental serpentine dehydration: implications for deep earthquake mechanisms

Deep focus earthquakes offer insights into Earth’s mantle and supports plate tectonics theory. Because high pressures and temperatures hinder brittle failure, their mechanisms differ from shallow quakes. Dehydration embrittlement, proposed as dominant at 100-350 km depth, involves fluid release from minerals like serpentine, increasing pore pressure and triggering failure. However, serpentine dehydration has a net decrease in pressure, requiring low-permeability layers to trap fluids to enable seismic failure. Experiments also show that serpentine dehydration often leads to ductile weakening without acoustic emissions.

To better understand the micro mechanisms involved in the dehydration of serpentinite, especially in the incipient stage, we have performed high pressure-temperature experiments under isostatic and non-isostatic conditions. Cores of serpentinite with 2 mm diameter were mounted in cubic assemblies with 12 mm edge. Experiments were carried out with the 6-Ram multi anvil press at the Bayerisches Geoinstitute, at pressure of 5 GPa, to a maximum strain of 15% at strain rates between 1.67x10^-4 s^-1 to 2.91x10^-6 s^-1. Temperature during isostatic and non-isostatic conditions was kept constant. Isostatic experiments were conducted at 550 °C and 784°C. non-isostatic experiments were conducted at ~650 °C.

Results show that isostatic dehydration of antigorite at 5 GPa starts at ~ 550 °C and is completed at ~ 800°C. Between 550-650 °C incipient dehydration of antigorite is evidenced by the growth of olivine and phyllosilicate at antigorite grain boundaries.  At these conditions, no failure microstructure is observed. Pores are present between olivine and enstatite grains of fully dehydrated serpentine. When deformation is imposed at incipient dehydration conditions, olivine and phyllosilicate start to cluster and form microscopic shear bands oblique to the main stress direction. These results demonstrate that at microscopic level, dehydration and failure of serpentine is complex. Pre-existing microstructural heterogeneities may influence nucleation of olivine and phyllosilicates. Pore overpressure may not be the only mechanism involved in serpentinite failure. Further work is required to determine the importance of the strength of the dehydration products in leading to localized failure.