The Mechanics of Deep Earthquakes: An Experimental Investigation of Slab Phase Changes

Tuesday, 16 December 2014
James R Santangeli, David P Dobson, Simon A Hunt and Philip G Meredith, University College London, London, United Kingdom
The mechanics of deep earthquakes have remained a puzzle for researchers since 1928 when they were first accurately identified by Kiyoo Wadati1 in Japan. Deep earthquakes show a split distribution, with peaks centered around ~370-420km and ~520-550km. As these events are limited to subducting slabs, it is accepted that they may be due to phase changes in metastable slab material. Indeed, conditions at ~350km depth are nominally appropriate for the olivine – wadsleyite transition, consistent with the anticrack mechanism previously observed in (Mg,Fe)2SiO42. The additional peak around 520km suggests that there is another siesmogenic phase change; candidates include Ca-garnet -> Ca-perovskite, wadsleyite -> ringwoodite and enstatite -> majorite or ilmenite. Importantly, for large scale seismogenesis to occur candidate phase changes must be susceptible to a runaway mechanism. Typically this involves the release of heat during exothermic reactions, which acts to increase reaction and nucleation rates. It is worth noting that the post-spinel reaction (sp -> pv + fp) marks the cessation of deep earthquakes; possibly as a result of being endothermic.

This research aims to identify which of these candidates could be responsible for seismogenesis.

We use high-pressure split cylinder multi-anvil experiments with acoustic emission detection. Low-pressure analogue materials have been used to allow greater cell sizes and thus sample volumes to enable accurate location of AE to within the sample. The candidate phase is annealed below its phase boundary, and then taken through the boundary by further compression. Acoustic emissions, if generated, are observed in real time and later processed to ensure they emanate from within the sample volume. Initial results indicate that the pryroxene -> ilmenite transition in MgGeO3 is seismogenic, with several orders of magnitude increase in the energy of AE concurrent with the phase boundary.


1) Wadati, K. (1928) Shallow and deep earthquakes. Geophysical Magazine. 1, 162-202

2) Green, H., W., & Burnley, P., C. (1989) A new self-organizing mechanism for deep-focus earthquakes. Nature. 341, 733-737