Influence of preexisting damage on dynamic fragmentation behavior of sandstone

Abstract

Preexisting damage in upper crustal rocks modifies their constitutive behavior and reduces strength, altering fragmentation during near-surface dynamic loading processes such as earthquake rupture, impact cratering, and landslides. To investigate the influence of preexisting damage on brittle fragmentation under these conditions, we study the behavior of heat-treated sandstone under dynamic tensile loading. Heat-treatments induce microfractures and mineralogical changes, simulating preexisting damage that can exist due to processes related to tectonic deformation, exhumation, and faulting-induced damage. A modified sample configuration for a Split Hopkinson Pressure Bar (SHPB) apparatus is used to induce tensile fragmentation in Berea Sandstone samples heat-treated at 250 °C, 450 °C, 650 °C, and 850 °C, and we analyze deformation in post-mortem samples using optical and scanning electron microscopy. Permanent strain during experiments is accommodated by the formation of mode-I fractures, dilation bands, and pore space expansion that can result in localized porosity increases of up to 25 %. Tensile strength increases with heat-treatments up to 450 °C and then decreases for heat-treatments at temperatures above the α-β-quartz transition. Elastic properties of Berea Sandstone also change with heat-treatments, and in a single orientation the Poisson's Ratio becomes negative at heat-treatments above 250 °C. Whereas fractures induced by SHPB experiments are primarily intergranular for sandstone heat-treated at temperatures up to 450 °C, intragranular fractures become more prevalent in sandstone heat-treated at 650 °C and 850 °C. Intragranular cracks emanating from grain contacts are observed in samples loaded under tension, but the mechanism of their formation is different than morphologically similar “Hertzian” fractures formed under compression. We propose a model for contact emanated tensile fractures formed under a remote tensile stress field that resemble Hertzian fractures formed under compressive loading. Finally, we compare our experimentally damaged sample to naturally deformed sandstones from the damage zone of the San Andreas Fault and Serpent Mound Impact Structure in Ohio, and we show distinct features that are likely indicative of brittle damage formed under macroscopic tensile loading. Our results have implications for dynamic tensile brittle fragmentation of sandstone with preexisting damage during earthquake rupture, landslides, and impact cratering. Importantly, these results provide new insights for linking dynamic loading conditions and processes with field observations of brittle damage in sandstone.