GSA Annual Meeting in Seattle, Washington, USA - 2017

Paper No. 139-6
Presentation Time: 2:55 PM

H2O IN THE UPPER MANTLE: INSIGHTS FROM MINERAL EQUILIBRIA


LAMB, William M., Department of Geology and Geophysics, Texas A&M University, College Station, TX 77843 and HUNT, Lindsey, Deptartment of Geosciences, Austin Peay State University, Clarksville, TN 37044, w-lamb@geos.tamu.edu

H2O plays an important role in various mantle processes (e.g., melting and deformation), and the H2O content of the upper mantle has been estimated from the H2O contents of nominally anhydrous minerals (NAMs). However, for many common NAMs, the relation between H2O activity (aH2O) and H2O content is not well known, and certain NAMs may be prone to H2O loss during emplacement on Earth’s surface. Thus, it may be advantageous to constrain the activities of various fluid species using mineral equilibria. Amphibole-bearing samples may be particularly useful because equilibria involving this phase can be used to estimate values of aH2O and hydrogen fugacity (ƒH2). Spinel equilibria can be applied to estimate values of oxygen fugacity (ƒO2) that, when combined with ƒH2, will provide a second estimate of aH2O.

Examination of 11 amphibole-bearing peridotite xenoliths from the southwestern U.S.A. reveals various amphibole textures (disseminated grains, rimming spinel, and poikilitic), indicating that amphibole in subsets of these samples may have different origins, including those that crystallized from a magma. The compositions of co-existing pyroxenes ± garnet yield P-T conditions for the 11 samples ranging from 1.1 to 1.5 GPa, and 880 to 980˚C. Application of pargasite dehydration equilibria yields values of aH2O ranging from 0.05 to 0.26. The compositions of coexisting spinel + olivine + opx yield DlogƒO2(FMQ) of -1 to +0.7. Values of ƒH2, estimated using amphibole dehydrogenation equilibria (6 to 91 bars) were combined with values of ƒO2 to estimate aH2O in nine samples (≈ 0.02 to 0.12). The difference between these two values of aH2O, estimated using dehydration equilibria versus ƒH2 + ƒO2, is < 0.08 for all samples, suggesting that the amphiboles experienced little or no retrograde H-loss.

Low values of H2O are indicated for all samples, even though amphibole is present and more than one amphibole forming process is represented. Thus, if an amphibole is stable in the mantle it may act to consume much of the available H, and equilibrate at low values of aH2O. Mantle rocks at the appropriate P, T and bulk composition may, therefore, grow amphibole even when H is present in low amounts, and the presence of a significant oxy-component in the amphibole may be indicative of low aH2O rather than elevated values of ƒO2.