DI21B-06
Breakdown of phase D by formation of aluminous bridgmanite

Tuesday, 15 December 2015: 09:15
303 (Moscone South)
Tomoo Katsura1, Takayuki Ishii2, Nobuyoshi Miyajima2, Sylvain Petitgirard2 and Catherine A McCammon1, (1)University of Bayreuth, Bayreuth, Germany, (2)Bayerisches Geoinstitut, Universitaet Bayreuth, Bayreuth, Germany
Abstract:
A certain amount of water is considered to be transported into the deep mantle through subduction in the forms of hydrous minerals. Although the major constituent minerals can host significant amounts of water in the upper mantle, bridgmantite (Brm) and ferropericlase (fPc) can incorporate only a limited amount of water. Therefore, if water is transported into the lower mantle through subduction, it must be hosted by dense hydrous magnesium silicates (DHMSs) such as phase D.

However, it is not clear whether the DHMSs can exist in the lower mantle. Brm contains significant amounts of Fe3+ (Fe3+/ΣFe > 0.1) even in coexistence with metallic iron. The amounts of Fe3+ increase with increasing Al2O3 content in Brm, and the Fe3+/ΣFe ratio exceeds 0.5 in the lower-mantle Brm. Since the Fe3+/ΣFe ratios are only 0.01-0.03 and 0.10-0.14 in the upper-mantle peridotite and MORB, respectively, formation of Brm would take oxygen out of other minerals. Hence, we expect breakdown of DHMS by reduction of OH into H2 through the Brm formation.

To test this hypothesis, we conducted the following experiments. The starting material consisted of an assemblage of a pre-synthesized phase D together with an natural olivine single crystal or a pyroxene aggregate with bulk composition of pyrolite minus fPc (Pyr-fPc) synthesized at an PO2 ≈ IW, T = 1100 K and ambient pressure. The two-phase samples were pressurized to 28 GPa and heated to 1100 K for 1 hour. First SEM observation of recovered samples shows that a 2-µm thick SiO2-rich fine-grained layer has formed on the boundary between phase D and Pyr-fPc Brm indicating the breakdown of phase D, while no evidence of breakdown of phase D was observed in the experiments with the olivine single crystal.

The present result suggests that, although water can be transported into the top of the lower mantle in a form of phase D through subduction, phase D will decompose by reduction at the depth to 720 km. The produced H2 should be re-incorporated into the upper mantle. Thus, water cannot be transported through subduction to the majority of the lower mantle.