HR: 0830h
AN: S21E-0343 [PDF]
TI: The Transition From Fault-Slip to Cataclastic Flow in Gypsum Under Hydrous and Dehydrating Conditions:
An Experimental Investigation on the Faulting Process at Intermediate Depth
AU: * Milsch, H H
EM: hmilsch@ldeo.columbia.edu
AF: Lamont-Doherty Earth Observatory of Columbia University, 61 Route 9W, Palisades, NY 10964 United States
AU: Scholz, C H
EM: scholz@ldeo.columbia.edu
AF: Lamont-Doherty Earth Observatory of Columbia University, 61 Route 9W, Palisades, NY 10964 United States
AB:
It has been suggested that intermediate-depth earthquakes occur on reactivated preexisting faults that were created at
shallow depth. Fluids released by dehydration of serpentinized rocks within the fault plane could build up pressures that act
against the fault normal stress eventually promoting localized unstable slip.
We performed conventional triaxial friction experiments on natural gypsum samples having saw-cut and surface-ground fault
planes inclined by $37.5\deg$ to the vertical.
At room temperature dry and wet samples were either strained up to 10 % at constant axial strain rates of 4 $\times$
10$^{-6}$ to 1 $\times$ 10$^{-5}$ s$^{-1}$ cycling the confining pressure or statically loaded varying the pore pressure. The
maximum confining and pore pressures were 50 MPa. Argon and water were used as pore fluids for dry and wet samples
respectively. At constant strain rate the samples showed stick-slip only up to 10 MPa (dry) and 15 MPa (wet) confining
pressure. At static load the onset of stress drops occured below 10 MPa and 15 MPa effective pressure for undeformed dry and
wet samples respectively, in agreement with Terzaghi's principle. At higher pressures the faults locked and deformation
occurred through bulk cataclasis as if the samples were intact. A locked fault could not be reactivated by lowering the
confining pressure. Instead, a secondary fault formed at approximately $30\deg$ to the vertical (dry) or bulk deformation
continued (wet).
Undrained static load experiments at 150 $\deg$C and 50 MPa confining pressure on initially dry samples showed that once
dehydration started both un- and predeformed samples weakened drastically. Only the former samples showed stress drops,
stick-slip upon subsequent straining and a visible offset of the two loading blocks. The predeformed samples did not show any
secondary fault development.
The transition from fault-slip to cataclastic flow is interpreted as the point where the shear stress necessary for
frictional sliding becomes larger than the shear stress that corresponds to the yield strength of the samples. The inability
to reactivate locked faults is interpreted to mean that such faults have become welded at high normal stress, increasing
their slip resistance.
Our results demonstrate that dehydration assisted unstable slip on preexisting faults is indeed possible. Although some of
the observations might be particular to gypsum they also indicate that strong mechanical locking of the fault may continue to
promote ductile flow unless new, more favourably oriented faults are formed either by leaking of fluid into the surrounding
rock or a new, yet unknown process.
DE: 3902 Creep and deformation
DE: 7209 Earthquake dynamics and mechanics
DE: 7218 Lithosphere and upper mantle
DE: 8010 Fractures and faults
DE: 8045 Role of fluids
SC: Seismology [S]
MN: 2003 Fall Meeting