At upper mantle pressures the endmember orthopyroxenes, MgSiO3 and FeSiO3, have been shown experimentally to transform to a denser high-P monoclinic polymorph with the C2/c space group. An analysis of the orthoenstatite to high-P clinoenstatite transition indicates that there is a decrease in molar volume of ~ 3 % and that Δ S and Δ H of the transition are significant. As a result, this transformation has petrologic and geophysical implications particularly for the assumed changes in mineralogy and bulk properties that occur during the subduction of lithospheric slabs. It is, therefore, important to determine the stability and crystal chemical properties of this high-P phase for relevant compositions including the Mg and Fe endmembers.
Reversals of the orthoenstatite to high-P clinoenstatite transition
have been reported in the literature, but for the ortho-high-P clino transition
in FeSiO3 there is only a single reversal bracket. We have,
therefore, performed new reversal experiments in the multi-anvil press
in order to constrain the P-T slope of the orthoferrosilite (OFS)-high-P
clinoferrosilite (HCFS) transition. The starting materials for these experiments
were an 80 % -20 % by weight mixture of OFS and clinoferrosilite (CFS)
respectively. Examination of these run products yielded four full reversal
brackets and several half brackets that span 400 °C in temperature
and 2.0 GPa in pressure (Fig. 3.4-5). In addition, four other experiments
(the squares in Fig. 3.4-5) contained more than one phase, segregated within
the capsule. CFS and OFS were identified by X-ray diffraction in two of
these. The segregated nature of the samples demonstrates that reaction
had occurred. In two other experiments, CFS was identified in part of the
sample. The X-ray diffraction pattern from the other portion contained
a number of peaks with strong intensities along with others that were extremely
weak. The high intensity peaks were found to be consistent with the hk0
reflections of pyroxene and, since these reflections are common to both
orthopyroxene and clinopyroxene, unequivocal identification of this phase
was not possible. The observed preferential enhancement of the hk0 reflections
could be due to either preferred orientation in the X-ray sample or to
strong stacking disorder. Some preferred orientation was confirmed by collecting
several X-ray patterns using different diffraction geometries. However,
a preliminary investigation by TEM revealed that this material was predominantly
OFS with strong stacking disorder and contains narrow domains of CFS (T.G.
Sharp, personal communication). Such structural disorder is interpreted
to represent conditions very close to the transition boundary, which is
supported by the fact that one of these experiments lies within a reversal
bracket.
![]() |
Fig. 3.4-5: Experimental reversals of the orthoenstatite = clinoenstatite phase boundary in FeSiO3. Filled symbols are experiments performed in Bayreuth, the open symbols are results from Akimoto et al. (J. Geophys. Res. 70, pp 5269-5278, 1965). |
Preliminary assessment of the boundary yields a best-fit given by:
The slope is somewhat steeper than previously proposed but is consistent with the one published bracket (Fig. 3.4-5). Combining these results with data for the OFS - P21/c low-P clinoferrosilite (LCFS) boundary, a triple point between OFS, HCFS and LCFS is tentatively located at 827 °C and 4.9 GPa.
It is important to note that the ortho - high-P clino transition in FeSiO3 occurs at much lower pressure than in MgSiO3. This means that the addition of Fe in enstatite will stabilise the C2/c polymorph to lower pressures relative to orthopyroxene.