Experimental investigation on the rheology of fault rocks

The aim of this study is to explore and identify physical and chemical processes occurring during deformation of fault rocks under elevated pressures and temperatures where an interplay between brittle/frictional and viscous processes is expected. The mechanical response of a crushed granitoid rock placed between forcing blocks pre-cut at 45° was studied under a broad range of temperatures (T = 300 – 600°C), confining pressures (Pc ∼ 300 – 1500 MPa) and displacement rates (d ̇, of ∼ 10−6, 10−7 and 10−8 m s−1) to different sample bulk shear strains (γ ∼ 0 – 5). Microstructural observations show, that the crushed fault rock undergoes compaction accommodated by short, closely spaced R1 Riedel shears producing large amounts of fine-grained (< 100 nm) material. This is accompanied by strain hardening in the mechanical record. Compaction is complete around a finite shear strain, γ, of ∼ 1.5, well before peak strength and deformation starts to partition into interconnected, anastomozing slip zones (SZ). Around peak strength (γ ∼ 2) a S-C fabric is well developed and the fine-grained material (< 100 nm) in the SZ is being transformed into a nanocrystalline (mean grain size ∼ 35 nm), partly amorphous material with an intermediate chemical composition between potassium-feldspar and plagioclase. High peak shear strength (τ ∼ 0.6 – 1.4 GPa) is reached around a γ, of ∼ 2.5 where the SZ form 7-12 % of the fault rock volume. Very fine layering and laminar flow structures are observed in the SZ. Fracturing and grain size reduction continues in lenses of material delimited by individual SZ indicating strain partitioning. After peak strength, the fault rocks weaken slightly (40 – 140 MPa) and continue to deform at approximately constant and high stress levels (τ ∼ 0.45 – 1.2 GPa) up to a finite shear strain of ∼ 5. Peak stress as well as the stress level during steady-state deformation exhibits a strong temperature dependence and a weak strain rate dependence indicating a viscous component of deformation.
With increasing strain and temperature, the amount of the slip zones increases (up to ∼ 25 %) indicating either strain hardening in the nanocrystalline, partly amorphous material, or that geometrical constraints do not allow continuous operation of the SZ. After peak strength, the SZ form a through going interconnected, anastomozing network. Deformation continues to localize within the SZ and the material changes its microstructure further. Around 90% of this material is amorphous to the TEM beam in zones, which accommodated high local strains (γ > 5). Turbulent flow structures and a very heterogeneous chemical composition develops in the high strain SZ which cover roughly ∼ 1% of the sample volume exploiting some of the pre-existing SZ and seem to form a multiply-connected topology, in contrast to SZ which accommodated lower strains. Crystalline fragments and nanocrystalline material is often incorporated into the high strain SZ and at highest stresses and lowest temperatures (300°C), small bubbles (∼ 15 nm – 1 μm) appear. This material is preferentially observed at high stress sites and shows intrusive relationships with the surrounding fine-grained material forming injection veins.
The calculated temperature increase at the fault is small (max ∼ 2.5°C) as the displacement rate was slow and the heat conduction high. Based on inspection of the microstructures and mechanical data, it is inferred that the fragment loaded, amorphous material exhibited a fluid-like rheology, i.e. shear stress was proportional to shear strain rate. However the microstructural record down to ångstöm scales is not compatible with the material being a liquid (in the sense of being molten above its liquidus temperature) and indicates that the loss of long-range order was achieved by mechanical work.
Our results indicate, that faults can build up significant stresses at the brittle-viscous transition leading to extreme grain comminution and amorphization. By comparison of the experimentally produced microstructures to microstructures observed in natural pseudotachylites, we conclude that the material produced during the experiments could be identified as a pseudotachylite in nature. However, pseudotachylites are currently being interpreted as high-temperature frictional melts that can form exclusively during earthquakes. Nevertheless, the fragment loaded amorphous material produced during the experiments did form neither fast nor at high temperatures, causing a conundrum. Some natural pseudotachylites are found under conditions, which are considered “paradoxical” under the assumption that all pseudotachylites originated as frictional melts. Our observations open new possibilities how to resolve these paradoxes. It is concluded that the use of pseudotachylites as evidence for ancient earthquakes should be reconsidered.