At 08:47 AM 24/09/2001 -0500, Eric Essene wrote:
>Bruce, Dug, and all,
> A higher grade reaction that is similar to Dugald's is:
> almandine + O2 = Al2SiO5 + quartz + magnetite.
This reacton limits almandine rather than almandine+annite; hence it must
be at _lower_ grade than M Joshi's reaction (other conditions being equal).
On the Ni+NiO buffer, the corresponding reaction is
Alm + NiO = Al2SiO5 + 2 Qtz + Mt + Ni .
Again using the Holland+Powell 1998 database, the endmember curve slopes
negatively across the Ky field from 5.3 kbar 250 deg C to 2.1 kbar 365 deg
C on the Ky=And curve. From there it crosses the And field less steeply to
0 kbar 550 deg C. Almandine (+NiO) is on the high-T, high-P side of the
curve. Hence, the very common subassemblage Sil+Qtz+Mt indicates f(O2)
higher than the Ni+NiO buffer (and much higher than Qtz+fayalite+Mt)
whether or not garnet is present.
>...There will be similar reactions for any ferromagnesian silicate in
Extending this point, we may note that Mt generally represents saturation
of the "ideal" metapelitic system KFMASH with the additional component
Fe2O3. Considering that Mt indicates relatively high f(O)2, the
KFMASH-Fe2O3 univariant P-T curves must be substantially displaced, and the
P-T "petrogenetic grid" for Mt-bearing metapelites must be significantly
different from that for graphite-bearing metapelites, especially at high