Interlayered graphitic and nongraphitic schists from the Tauern Window, Eastern Alps, show very different microstructures developed during extension/isothermal decompression. Garnets in graphitic schists are pervasively fractured perpendicular to the stretching direction, whereas garnets in nongraphitic (NG) schists have fewer, more randomly oriented fractures. Secondary fluid inclusion planes in matrix quartz are abundant in graphitic samples and are continuous with and/or parallel to the garnet microcracks. Pre-crack FIs have XCO2<0.1, whereas XCO2=0.2-0.6 in microcrack FIs. Crack spacings in 42 garnets from graphitic samples are tightly clustered at 200±50 microns; 56 garnets in NG samples are uncracked or have spacings of 200 to >1000 microns. These observations suggest that the presence or absence of graphite played a key role in controlling microcracking after garnet growth ended at 500-550°C and 6-7 kbar.

Metamorphic COH fluids in equilibrium with graphite must change composition continuously to maintain equilibrium during changes in P and T. Specifically, an initially aqueous fluid will consume graphite and produce an increasingly carbonic fluid during decompression. Fluids in NG rocks have no such constraints and may remain aqueous throughout decompression. Calculations in a model graphitic rock at fO2=10e-23 show that the fluid evolves from XH2O=0.92, XCO2=0.07, XCH4=0.01 to XH2O=0.62, XCO2=0.38, XCH4=0 during decompression from 7 kbar to 4 kbar at 500°C; below 4 kbar, the fluid is dominantly CO2. The molar volume of the model fluid would expand 140% (23 to 32 cm3/mol) from 7 to 4 kbar or 195% (23 to 45 cm3/mol) from 7 to 3 kbar. In contrast, pure H2O in an NG horizon would expand only 115-120% over the same P intervals. Pore fluid expansion is thus greater in graphitic than nongraphitic rocks, and is likely to promote microcracking in graphitic rocks. This effect may be enhanced by a shift from wetting to nonwetting behavior as the fluid becomes more carbonic during decompression.

These results predict that microcracking should be an inevitable consequence of decompression in graphitic schists. Rocks that lack graphite are less likely to undergo microcracking. Microseismicity is thus predicted to be more common in graphitic than nongraphitic rocks during unroofing of mountain belts.