Previous   |    Next

doi:10.2204/iodp.proc.314315316.224.2019

Results

Residual shear strength increases from ~2.5 MPa at σ_n′ = 5 MPa to 12–14 MPa at σ_n′ = 35 MPa for most samples (Figure F5; Table T2). Sample 316-C0006F-19R-1 is noticeably weaker than the other samples and increases from 1.3 to 8.6 MPa residual shear strength. Sample 19R-1 is also the only sample to exhibit a significant peak shear strength at σ_n′ = 5 MPa, whereas for the other samples the peak strength is not significantly higher than the residual shear strength (Table T2).

For all samples, residual shear strength increases as a function of effective normal stress, similar to Coulomb-Mohr behavior. However, our samples did not strictly exhibit Coulomb-Mohr behavior but rather sublinear shear strength as a function of effective normal stress. This can be seen in the values of residual coefficient of friction, which decrease as a function of effective normal stress (Figure F6; Table T2). For most samples, residual µ ranges from 0.46 to 0.56 at σ_n′ = 5 MPa and decreases to 0.34–0.40 at σ_n′ = 35 MPa. The residual µ of the weaker Sample 316-C0006F-19R-1 ranges from 0.22 to 0.27.

A feature of the residual coefficient of friction data is that they decrease nonlinearly with increasing effective normal stress. We fit the data for each sample with both power law and logarithmic functions:

where the parameters A, ω, B, and C are empirically determined from the regression. We observe that both power law and logarithmic provided excellent fits to the data for the four stronger samples, with the coefficient of determination R² values of at least 0.978 (Figure F7; Table T3). The power law and logarithmic functions provide nearly identical fits to the data.

In contrast to the four stronger samples, the weaker Sample 316-C0006F-19R-1 exhibits a low R² value for both the power law and logarithmic fits. It is possible that the reason for this is that the coefficient of friction may show a more significant increase at lower effective normal stresses than the other samples. In general, the unique behavior of this sample can be attributed to its clay mineral content, which is the highest of the five samples at 67 wt% (Figure F8; Table T1). For our data set, we observed a general trend of decreasing coefficient of friction with increasing clay content (Figure F8) with the low friction of Sample 19R-1 being clearly related to high clay content. We note that because we tested mostly remolded samples, our experiments isolate the effect of mineral assemblage. In nature, microstructural differences due to different depositional settings (i.e., trench wedge [Unit II]) compared to hemipelagic basin [Unit II]) sediments) could also play a role (e.g., Takahashi et al., 2013).

Top of page   |    Previous   |    Next