Previous   |    Next

doi:10.2204/iodp.proc.314315316.224.2019

Experimental methods

We conducted laboratory shearing experiments under true-triaxial stress conditions and controlled pore pressure, using a biaxial testing apparatus with servo-hydraulic control (Figure F3). Samples were constructed as two layers with an area of 5.4 cm × 5.7 cm and initial thicknesses ranging from ~2 to 6 mm under an initial normal stress of 5 MPa. Most of the samples were built remolded, meaning core material was lightly cold-pressed into the sample assembly. One sample, 316-C0006E-40X-8, was tested as two intact wafers trimmed perpendicular to the core axis. The two subsamples were sandwiched in a three-block assembly outfitted with porous metal frits, allowing fluid access, jacketed in rubber. The jacketed assembly was placed in the pressure vessel and subjected to confining pressure and then saturated with 3.5 wt% NaCl brine as pore fluid. The effective normal stress includes the combined effects of externally applied normal load, confining pressure, and two independent pore pressures. One of the pore pressures was applied to the inner faces of the sample layers and designated as the inlet pressure; the other pore pressure accesses the outer faces of the samples and was designated as the outlet pressure (Figure F3). The confining pressure was held constant at 6 MPa and the inlet pore pressure was held constant at 5 MPa. The outlet pressure was set to a no-flow (undrained) condition to monitor pore pressure in the layer during shearing, following Ikari and Saffer (2011). Pore pressure fluctuations recorded by the outlet pressure are accounted for in calculating the effective normal stress; these fluctuations are small and have little effect on the experiment (Figure F4). With the confining and pore pressures held constant, the effective normal stress was raised from 5 to 15, 25, and 35 MPa by increasing the externally applied normal load (Figure F4).

In each experiment, we sheared our samples at a constant driving rate boundary condition of 11 µm/s, as measured at the load cell on the vertical piston (i.e., load point velocity). The shear stress τ is measured continuously, from which we calculate a coefficient of sliding friction µ (Handin, 1969):

where µσ_n′ is the effective normal stress. Note that in calculating a sliding coefficient of friction we assume that the shear strength results entirely from frictional strength and there is no cohesive strength component. This facilitates comparison with previous studies, but we acknowledge that the cohesion in sheared materials may be significant (Ikari and Kopf, 2011).

Peak shear strength was measured only at the lowest effective normal stress (5 MPa) because the higher effective normal stresses tested in this study are much higher than the in situ effective stresses experienced by the samples. Residual shear strength was measured at every effective normal stress. In some cases, the sample reached a steady state; however, in many cases a slight slip strengthening or weakening trend was superimposed on the data (Table T2). In the case of slip weakening, we measured the maximum strength. In slip strengthening cases, we calculated a strengthening rate as dτ/dx, where x is displacement, and picked the strength value where dτ/dx became constant (see Ikari et al., 2011).

Top of page   |    Previous   |    Next