Mode I cracking versus dilatancy banding: Experimental constraints on the mechanisms of extension fracturing

Abstract

Fractures or discontinuities perpendicular to the least stress σ3 were generated in a synthetic rock analogue (granular, frictional, cohesive, and dilatant) material in axisymmetric extension tests. These fractures are of two types defined by the mean stress σ. When σ is very small, the fractures form through the mode I cracking at tensile σ3 equal to the material tensile strength. The fracture walls have smooth surfaces. At higher σ, these surfaces become rougher, with the topography features forming faint or delicate plumose patterns. The amplitude of the plumose topography increases with σ. The axial stress σ3 at fracturing reduces in magnitude and changes in sign with the σ increase. Thus σ3 orthogonal discontinuities can form at compressive σ3. SEM observations show that these discontinuities are deformation localization bands where the material is characterized by the heterogeneous decohesion and volume and porosity increase due to dilatancy. The band thickness is several grain sizes. At formation, the bands are not opened, so they are not mode I fractures. They become fractures with plumose fractography after the separation of the sample parts along the band. The formation mechanism of these discontinuities or fractures is not completely clear, but it is suggested that it represents a running constitutive instability in the form of dilatancy banding (with further σ increase the bands become inclined to σ1, i.e., shear). The similarity between the experimentally generated plumose surface fractures and natural joints is discussed, and it is suggested that they can be formed as propagating constitutive instabilities.

Origin of Regular Networks of Joints: Experimental Constraints, Theoretical Background, and Numerical Modeling

“…The adjacent layers are totally coupled. The model parameters are close to those in GRAM experiments in Chemenda et al () and Jorand et al (). The elastic properties of all the layers in the presented models are the same, E = 6.7 × 10 8 Pa and ν = 0.25 and are those for GRAM1 ( E is Young's modulus and ν is Poisson's ratio); , , , and α c are defined from the plot in Figure a to be Pa, Pa, and Pa, α c = 2 (see also Figure ).…”

“…Chemenda et al () conducted similar tests on Granular Rock Analogue Material (GRAM1) made of very small bonded TiO 2 grains (Nguyen et al, ). The mechanical properties of this material have been extensively studied (Mas & Chemenda, , ; Nguyen et al, ).…”

“…Minimum principal stress σ 3 versus confining pressure P c = σ 1 at fracturing/rupture in conventional (axisymmetric, σ 1 = σ 2 ) extension tests of GRAM1 (a) and Berea sandstone (b). Data for sandstone are from Bobich () and for GRAM1 from Nguyen et al () and Chemenda et al (). ψ is the angle between the fracture plane and σ 1 direction; and are the initial tensile and uniaxial compression strengths; is the abscissa of the intersection point of the horizontal and inclined blue lines approximating the experimental data.…”

“…Surfaces of σ 3 ‐normal fractures generated in GRAM1 specimens in axisymmetric extension tests under different confining pressures P c (from Chemenda et al, ).…”

“…Nominal stress‐strain curves from axisymmetric extension tests on GRAM1 (Chemenda et al, ) and Berea sandstone (Bobich, ) at different confining pressures P c = σ 1 indicated on the plots in MPa. E is the Young modulus for GRAM1, E = 6.7 × 10 8 Pa (Mas & Chemenda, ).…”