Incorporation of OH in olivine at high pressure and temperature:
new experimental results
Some of Jedley's favorite natural olivines, from Tanzania. Olivines
from this locality have unusually high concentrations of boron (~60 ppm) and
hydrogen (~20 ppm). For details see Kent and Rossman, 2002, American Mineralogist,
87: 1432-1436.
Numerous studies of natural and experimentally hydrated samples show that
olivine can incorporate substantial amounts of hydrogen at high pressures,
which has important implications for the rheology, melting behavior and
transport properties of the mantle. In collaboration with Natalia Deligne (Caltech undergrad
and SURF fellow) and professors Paul Asimow and George Rossman,
I have been investigating the incorporation of OH in olivine by conducting
new experiments at high pressures (2-10 GPa) and temperatures (1100-1300
°C) and varying oxygen fugacity and silica activity. In contrast to previous
experiments, we calculate the concentration of OH in olivine by integrating
the total absorbance from three FTIR spectra polarized in orthogonal directions
and applying the new calibration of Bell et al. (in press), based on the nuclear
reaction analysis technique. This procedure is inherently more accurate
than the use of nominally unpolarized spectra and the generic calibration
of Paterson (1982).
|
|
Reflected light image of an experiment
at 8 GPa, 1150 °C, showing large olivine grains up to ~300 microns in
diameter. The orientation of the grains was determined by Raman spectroscopy.
We got lucky on this one: three b-axes and two c-axes are sticking up at us,
enabling us to obtain all three polarized spectra needed for quantitative
analysis of OH by FTIR.
|
Raman spectra of oriented olivine crystals. Crystals
which are sectioned perpindicular to one of the optic axes can be located
using the presence or absence of three peaks (denoted by arrows) in this region
of the spectrum. Orientations are verified using optical microscopy and polarized
infrared spectroscopy.
These spectra were originally taken by Liz Arredondo and George
Rossman and are located on the Mineral
Spectroscopy Server.
|
Infrared spectra in the OH-stretching
region of olivine from the 8 GPa experiment, polarized with the E-vector
parallel to the three optic axes.
Our preliminary results include the following:
1) The spectra in all samples are dominated by sharp peaks in the range
of ~3650-3400 cm-1, regardless of estimated silica activity. This
is consistent with most previous experimental studies and, broadly speaking,
with the spectra of most natural olivines.
2) The shapes of the spectra in all our samples are similar regardless
of whether grain growth occurred under hydrous conditions or diffusion of
H into existing large crystals took place. This strongly suggests that the
mechanism of incorporation is identical in both cases. However, in experiments
at 1100 °C, large single crystals exhibit strongly zoned concentration
profiles, consistent with the most recent diffusion data for OH in olivine
(Kohlstedt and Mackwell, 1998). This complicates interpretation of these
experiments and casts uncertainty on interpretation of previous experiments
that used the same technique.
3) Comparison of experiments using Ni-NiO or Re-ReO as buffers shows little
effect of oxygen fugacity (in this range, ~2 log units in fO2)
on either concentration or incorporation mechanism.
4) Concentrations measured in the samples that used fine-grained starting
materials are 2.5-3 times higher than the previous estimates for OH solubility
measured by Kohlstedt et al. (1996). This difference is primarily attributed
to use of the new calibration and measurement of polarized spectra.
These results suggest that olivine could be a more important reservoir
for hydrogen in the mantle than previously thought. We are conducting further
experiments to understand more quantitatively the roles of silica activity,
oxygen fugacity, and starting point defect structure on incorporation of
OH in olivine.
