Previous page     Contents     Next page

3.1 f. The search for deep earthquakes in the laboratory (W. Durham, D.C. Rubie, T.G. Sharp and C. Dupas, in collaboration with S. H. Kirby/Menlo Park)

Deep earthquakes (those with epicenters greater than approximately 300 km deep) are common, and often quite large, seismic events in subducting slabs, but they are enigmatic in that we know they do not involve brittle failure of rock as in the case of shallow earthquakes. Their great depth, and hence high overburden pressures, assures that they cannot involve any rock dilation, and hence that they must occur by a different mechanism than crustal earthquakes. Because they are invariably located in cold, olivine-rich subducting lithospheric plates, at pressures near and beyond the phase transition pressure of olivine to wadsleyite and ringwoodite (ß- and γ-phase), geophysicists have long suspected that deep earthquakes are somehow related to a kinetically delayed phase transformation of olivine. There is supporting evidence from other materials, namely water ice and the germinate analogue of silicate olivine, that very high-pressure failure can occur, but such failure has never been unambiguously observed in silicate olivine. This recently started project, a collaborative effort between the Bayerisches Geoinstitut and two laboratories in California (Lawrence Livermore National Laboratory and the U.S. Geological Survey) is attempting to produce deep earthquakes in olivine in the laboratory. Taking advantage of the large sample volume, and very high pressures attainable in the multi-anvil apparatus, we are subjecting polycrystalline olivine to very high differential stresses at confining pressures that simulate the transition zone (>14 GPa), but at relatively low temperatures where the bulk phase transformation will be kinetically suppressed. Current thinking suggests that the earthquake occurs when very small amounts of olivine transform due to the combined differential and hydrostatic stresses. Since we do not know under what conditions this instability will occur, we will perform experiments over a range of pressures and temperatures. The "failure", if it occurs, will not be audible, so samples will have to be examined after depressurization for such features as macroscopic offset and microscopic occurrence of high-pressure phases in the neighbourhood of the failure plane.

Bayerisches Geoinstitut, Universität Bayreuth, 95440 Bayreuth, Deutschland
Tel: +49-(0) 921 55 3700 / 3766, Fax: +49-(0) 921 55 3769, E-mail: bayerisches.geoinstitut(at)