Dc. Rubie et al., LOW DIFFERENTIAL STRESS AND CONTROLLED CHEMICAL ENVIRONMENT IN MULTIANVIL HIGH-PRESSURE EXPERIMENTS, Physics and chemistry of minerals, 20(5), 1993, pp. 315-322
Studies of mass transport and kinetics in minerals at high pressure of
ten require a sample environment in which the stress is near-hydrostat
ic and the chemical environment is carefully controlled. We report her
e details of a multianvil sample assembly in which these requirements
are fulfilled and which has been used to study the effect of pressure
on the kinetics of dislocation recovery in olivine up to 10 GPa. Annea
ling experiments have been performed on single crystals of San Carlos
olivine at 8.5 GPa and 1400 degrees C in a 1200 tonne split-sphere mul
tianvil apparatus. The sample assembly consists of an 18 mm MgO octahe
dron with a LaCrO3 heater of variable wall thickness to give a small t
emperature variation (similar to 20 degrees C) along the 3 mm length o
f the sample capsule. To minimize the differential stress on the sampl
e, the olivine single crystal is surrounded by NaCl and both pressuriz
ation and depressurization are performed slowly at a temperature of 60
0 degrees C (to minimize the strength of the NaCl). The silica activit
y (a(SiO2)) is buffered by orthopyroxene powder in contact with the ol
ivine and the oxygen fugacity is buffered by Ni+ NiO within the sample
capsule. The H2O-content of the sample assembly is minimized by dryin
g all components at 230 degrees C under vacuum. Olivine single crystal
s recovered after annealing at 1400 degrees C and 8.5 GPa show no evid
ence of deformation, either ductile or brittle. Dislocation densities
of 10(9)-10(10) m(-2) are similar to those observed prior to high-pres
sure annealing and indicate differential stresses of <10 MPa. Infrared
spectroscopy indicates that the hydrogen content of a sample annealed
at 10 GPa, 1500 degrees C for 21 h is similar to 13 H/10(6)Si, which,
although low, is higher than that of the crystals prior to high-press
ure annealing. Finally, the effectiveness of the fO(2), buffer has bee
n verified by estimating the fO(2) at the surface of the sample from t
he solubility of Fe in Pt metal in equilibrium with the olivine and or
thopyroxene.