Geometrical evolution of interlocked rough slip surfaces: The role of normal stress

Abstract

International audienceWe study the evolution of slip surface topography using direct shear tests of perfectly mating surfaces. The tests are performed under imposed constant normal stress and constant slip rate conditions, to a sliding distance comparable to the roughness scale of the studied surfaces. Prismatic limestone blocks are fractured in tension using four-point bending and the generated surface topographies are measured using a laser profilometer. The initially rough fracture interfaces are tested in direct shear while ensuring a perfectly mating configuration at the beginning of each test. The predetermined sliding distance in all tests is 10 mm and the sliding velocity is 0.05mm/s. A constant normal stress is maintained throughout the tests using closed loop servo control. The range of normal stresses applied is between 2MPa and 15MPa. After shearing, the surface topographies are re-scanned and the geometrical evolution is analyzed. We find that surface roughness increases with increasing normal stress: under normal stresses below 5MPa the surfaces become smoother compared to the original geometry, whereas under normal stresses between 7.5 MPa and 15 MPa the surfaces clearly become rougher following shear. Statistical spectral analyses of the roughness profiles indicate that roughness increases with length-scale. Power spectral density values parallel to the slip orientation are fitted by power-law with typical power value of 2.6, corresponding to a Hurst exponent of 0.8, assuming self-affine roughness. This power value is consistent for the post-sheared surfaces and is obtained even when the original surface roughness does not follow initially a power-law form. The value of the scaling-law prefactor however increases with increasing normal stress. We find that the deformation associated with shearing initially rough interlocked surfaces extends beyond the immediate tested surface, further into the intact rock material. The intensity of the damage and its spatial distribution clearly increase with increasing normal stress. Wear loss is measured by subtracting the post-shear surface from the pre-shear surface matrices using known reference points. Our measurements indicate that wear loss and roughness evolution are both positively correlated with the mechanical shear work applied during the experiments. We argue, therefore, that normal stress plays a significant role in the evolution of interlocked surfaces, such as geological faults, and strongly affects the energy partitioning during slip

On the debris-level origins of adhesive wear

“…On the whole, the behavior observed in the simulations is consistent with the classical picture of adhesive wear hypothesized from experimental observations (21)(22)(23)(24). Recent small-scale wear experiments on ceramics and rocks (25,26) confirm the formation of cylindrical and spherical wear debris particles. We make the distinction that this work is focused on wear by fracture-induced debris formation, as opposed to surface folding and delamination (23) mechanisms that may occur at different scales and/or under different wear conditions, e.g., abrasive wear.…”

“…Inspired by (i) our recent study (19) that revealed the critical importance of the junction shear strength (τ ) on debris particle formation and (ii) macroscopic laboratory observations of a linear correlation between the tangential work and wear volume (7,25,27,28), we examined the relationship between tangential work and the volume of resultant debris particles. Remarkably, we found that, at the debris level, these two quantities are related with a proportionality constant of 1/τ across the wide range of simulations performed in this work (Fig.…”

Smoothing of Fault Slip Surfaces by Scale‐Invariant Wear

“…Our results show with statistical certainty that within the first tens of meters of slip faults smooth in both the slip‐parallel and slip‐perpendicular directions. The weak sensitivity to slip reported in Brodsky et al (), along with experimental evidence fault reroughening (e.g ; Badt et al, ), has led previous studies to suggest that fault roughness is a steady state between smoothing processes and reroughening (e.g., Brodsky et al, ; Candela et al, ; Fang & Dunham, ; Shervais & Kirkpatrick, ). However, we report an RMS‐smoothing exponent of 0.5 B (from equation , which is 5 times greater than that reported in Brodsky et al ().…”

On the Development of Shear Surface Roughness

“…Earlier frictional slip experiments on solid materials of different rheological properties demonstrated contrasting roughness on slip surfaces. Experiments with marble (Badt et al, ) and granite (Amitrano & Schmittbuhl, ) showed that the slip in the brittle regime produces greater roughness (Hurst exponent values >0.7), attributed to plow‐related processes, such as gouge formation and striations. Similar slip experiments on halite and limestone, in contrast, develop relatively smoother roughness (Hurst exponent < 0.5), where asperities cause plastic strain localization in their neighborhood (Renard et al, ; Sagy et al, ).…”

“…Increasing inclination of fractures to the compression direction results in a transformation from localized to global development of slickensides, characterized by one dimensional, multiordered wavy irregularities when θ = 60°. This effect of θ in promoting linear irregularities probably results from increasing normal stress on the shear surfaces (Badt et al, ; Davidesko et al, ). Larger compressive normal stresses favor the wavy mechanical instability in our models, which eventually produces linear geometry (details of the mechanics discussed later) to facilitate the slip motion.…”