Colin J. Peach

Experimental determination of constitutive parameters governing creep of rocksalt by pressure solution

Abstract Theoretical models for compaction creep of porous aggregates, and for conventional creep of dense aggregates, by grain boundary diffusion controlled pressure solution are examined. In both models, the absolute rate of creep is determined by the phenomenological coefficient Z* = Z0exp (−ΔH/RT), a thermally activated term representing effective diffusivity along grain boundaries. With the aim of determining Z0, ΔH and hence Z* for pressure solution creep in rocksalt, compaction creep experiments have been performed on wet granular salt. Compaction experiments were chosen since theory indicates that pressure solution creep is accelerated in this mode. The tests were performed on brine-saturated NaCl powder (grainsize 100–275 μm) at temperatures of 20–90°C and applied stresses of 0.5–2.2 MPa. The mechanical data obtained show excellent agreement with the theoretical equation for compaction creep. In addition, all samples exhibited well-developed indentation, truncation and overgrowth microstructures. We infer that compaction did indeed occur by diffusion controlled pressure solution, and best fitting of our data to the theoretical equation yields Z0 = (2.79 ± 1.40) × 10−15 m3s−1, ΔH = 24.53 kJ mol−1. Insertion of these values into the theoretical model for conventional creep by pressure solution leads to a preliminary constitutive law for pressure solution in dense salt. Incorporation of this creep law into a deformation map suggests that flow of rocksalt in nature will tend to occur in the transition between the dislocation-dominated and pressure solution fields.

On dynamic recrystallization during solid state flow: Effects of stress and temperature

A hypothesis is advanced that dynamic recrystallization of Earth materials undergoing solid state flow may represent a balance between grain size reduction and grain growth processes occurring directly in the boundary between the dislocation and diffusion creep fields. Accordingly, the recrystallized grain size (D) and flow stress (σ) at steady state will be related by the equation delineating the field boundary, which in general is temperature dependent. Creep experiments on a metallic rock analogue, Magnox, yielded D=10 1.12 exp[29.3/RT]σ 1.2:3 and demonstrated that D (μm) decreases with increasing σ (MPa) and increasing temperature (T) in a manner which is in agreement with the field boundary hypothesis. If the model applies to rocks, the widely accepted idea that dynamic recrystallization can lead to major rheological weakening in the Earth may not hold. Moreover, empirical D-σ relations, used in paleo-piezometry, will need to be modified to account for temperature effects.

Influence of crystal plastic deformation on dilatancy and permeability development in synthetic salt rock

While the fluid transport properties of rocks are well understood under hydrostatic conditions, little is known regarding these properties in rocks undergoing plastic deformation. In this study the influence of macroscopic plastic deformation on permeability has been investigated experimentally using synthetic salt rock. Dilatometric triaxial deformation experiments performed on this material, at room temperature, confining pressures (Pc) in the range 5–20 MPa and strain rates of ∼ 4 × 10−5 s−1 to total strains of ∼ 15%, exhibited work hardening behaviour with minor amounts of dilatancy at Pc < 18 MPa. Microstructural observations confirmed that deformation occurred by dislocation glide, with grain boundary microcracking in the dilatant field. At the lowest pressures, deformation-induced dilatancy of only 0.1–0.2 vol.% produced extremely rapid initial increases in permeability (from ≤ 10−21 m2 to ∼ 2 × 10−16 m2), suggesting critical behaviour of the type described by percolation theory. This rapid permeability development with dilatancy is well described by crack linkage models based on percolation theory, provided a broad range of fluid conductance is incorporated. The study has shown that minor dilatancy (< 0.2 vol.%) during plastic deformation of salt rock can lead to very large increases in permeability. This is of direct interest with regard to the behaviour of salt rock in waste disposal and storage systems, and may have important implications for the transport of fluids through, and the interaction of fluids with, crystalline rocks undergoing crystal plastic deformation in nature.

Frictional-viscous flow of simulated fault gouge caused by the combined effects of phyllosilicates and pressure solution

Abstract It is widely believed that conventional brittle–plastic strength envelopes may significantly overestimate crustal strength, in part because they fail to account for the effects of fluid-assisted deformation mechanisms. In particular, pressure solution has been suggested to allow fault creep behaviour at a shear stress well below that predicted by conventional strength envelopes. In addition, many natural fault zones contain significant amounts of phyllosilicates, often defining a foliation. Phyllosilicates are believed to inhibit grain contact healing and increase the rate of pressure solution in quartzose rocks. Hence, the interaction between phyllosilicates and pressure solution may strongly influence fault rheology under the hydrothermal conditions pertaining around the brittle–ductile transition. The aim of this research is to assess the combined effects of pressure solution and phyllosilicates on fault slip behaviour. To this end, we performed high strain, rotary shear experiments on experimental faults containing brine-saturated mixtures of halite and kaolinite as simulated gouge. The experiments were done under conditions where pressure solution and cataclasis dominate over dislocation creep, thus simulating the brittle–ductile transition. Halite was chosen as a rock analogue because of its well-constrained pressure solution kinetics. Experiments were done at room temperature and atmospheric pressure, under drained conditions. In the experiments we explored the effect of varying sliding velocity, normal stress and clay content on fault strength. The results showed frictional-viscous behaviour, i.e. shear strength depending on both normal stress and shear strain rate, in brine-saturated halite/kaolinite mixtures. In contrast, purely frictional (i.e. normal stress dependent and shear strain rate insensitive) behaviour was observed when an inert pore fluid (silicone oil instead of brine) or an inert solid (quartz instead of halite) was used. Monomineralic halite and kaolinite gouges showed purely frictional behaviour as well. This demonstrates that the observed frictional-viscous behaviour was caused by the combined effects of pressure solution and phyllosilicates. The microstructures, which strongly resemble natural mylonites in many respects, suggest that deformation of the gouge involved sliding along kaolinite-rich foliation planes, accommodated by pressure solution and dilatation/cataclasis, the relative amounts of which varied with sliding velocity. If similar behaviour occurs in natural phyllosilicate-rich fault zones, such zones are expected to be significantly weaker than predicted by traditional brittle–ductile strength envelopes. In addition, our microstructures suggest that pressure solution may play an important role next to dislocation creep in microstructural evolution in mylonitic rocks.

Slip behavior of simulated gouge‐bearing faults under conditions favoring pressure solution

Geophysical observations as well as deformation experiments indicate that under hydrothermal conditions, crustal faults can be significantly weakened with respect to conventional brittle-plastic strength envelopes. Pressure solution has long been proposed as a mechanism leading to fault weakness. However, pressure solution has also been proposed as contributing to interseismic fault healing, and the competition between the weakening and healing effects of pressure solution is unclear. To investigate this issue, we have conducted rotary shear experiments on synthetic faults containing granular halite (NaCI) gouge using NaCI-saturated mixtures of water and methanol as pore fluid. The NaCI-water-methanol system was chosen as a rock analogue because pressure solution is known to be important in this system at ambient conditions. We explored the influence of varying pore fluid composition (hence pressure solution rate), gouge grain size, and wall rock surface roughness, as well as normal stress and sliding velocity on slip behavior. All experiments were done under drained conditions. An acoustic emission detection system allowed detection of brittle events in the gouge. The results show no evidence for steady state pressure solution-controlled fault slip. Frictional, rate-insensitive behavior was observed, whereas the microstructures and compaction behavior clearly demonstrated that pressure solution was active in the gouge. Our data show that fluid-assisted healing effects dominated over weakening, causing fault strength to be controlled mainly by brittle-frictional processes. Existing models describing pressure solution-controlled fault creep may not be applicable to a porous gouge undergoing compaction as well as slip.

Diffusive properties of fluid-filled grain boundaries measured electrically during active pressure solution

Abstract Diffusion through ‘wetted’ grain boundaries is often the rate limiting process during rock deformation by intergranular pressure solution. However, the underlying processes operative within such boundaries are poorly understood. In this contribution we have studied the diffusive properties of wetted grain boundaries by measuring the electrical resistivity of single, annular halite–glass contacts undergoing active pressure solution. Optical observation shows continuous growth (i.e. widening) of the annular contacts by pressure solution. From the resistivity measurements and making use of the Nernst–Einstein equation, it was possible to calculate the apparent grain boundary diffusion coefficient Z=DδC (i.e. the product of grain boundary diffusion coefficient D, grain boundary film thickness δ and the solubility C of the diffusing species in the grain boundary fluid) during the pressure solution process. The Z-values obtained lie in the range 3×10−20–2×10−18 m3/s, show an inverse dependence on normal stress (σn) and agree well with values inferred previously from single contact and polycrystalline compaction experiments.

Effect of varying enstatite content on the deformation behavior of fine-grained synthetic peridotite under wet conditions

The effect of varying enstatite content on the deformation behavior of synthetic, fine-grained (1 to 2 μm) forsterite-enstatite rock with ∼0.5 wt% added water was investigated at temperatures of 900° to 1000°C, strain rates between 10−7 and 10−5 s−1, and a confining pressure of ∼600 MPa. The samples exhibited approached steady state flow at stresses ≤60 MPa. The results show that, at constant strain rate, increasing enstatite content is associated with a sharp decrease in flow strength in the range 0–2.5% vol% enstatite, with little further change up to 20 vol%. The observed power law n value of ∼1.7 and microstructures are similar to those obtained in previous work on material with 2.5 wt% enstatite and are consistent with a water-enhanced grain boundary sliding (GBS) dominated deformation mechanism. Significantly, a negative correlation was found between grain size and enstatite content, indicating that enstatite content played a role in controlling the grain size of the starting materials. Moreover, a high correlation between measured flow strength and grain size was found, consistent with a grain size exponent of −3 in a conventional grain size sensitive flow equation. A water-enhanced deformation mechanism involving GBS accommodated probably by grain boundary diffusion and/or dislocation activity is therefore implied, with the effect of enstatite content on flow strength being an indirect physical effect caused by grain size control. Other effects of second-phase content, such as weakening caused by interphase boundary diffusion and/or migration processes, seem to be unimportant under the conditions investigated.

Compaction experiments on wet calcite powder at room temperature: evidence for operation of intergranular pressure solution

Abstract Dead weight uniaxial compaction creep experiments were carried out on fine-grained, super-pure calcite (<74 μm) at room temperature and applied effective stresses of 1–4 MPa. All samples were pre-compacted dry at a stress of 8 MPa, for 30 minutes, to obtain a well-controlled initial porosity. The samples were then wet-compacted under ‘drained’ conditions with pre-saturated solution as pore fluid. Control experiments, which were done either dry or with chemically inert pore fluid, showed negligible compaction. However, samples tested with saturated solution as pore fluid showed easily measurable compaction creep. The compaction strain rate decreased with increasing strain and increasing grain size, and increased with increasing applied stress. Addition of Mg2+ ions to the saturated solution dramatically inhibited compaction. From the literature, Mg2+ ions are known to inhibit calcite precipitation. By comparison with a theoretical model for intergranular pressure solution in calcite, the observed mechanical behaviour and the way that compaction responded to the pore fluid chemistry suggest that, under our experimental condition, intergranular pressure solution is the mechanism of the deformation and that precipitation is likely to be the rate-limiting step.

Kinetics of precipitation of gypsum and implications for pressure‐solution creep

In order to model the role of gypsum in crustal deformation and in trapping hydrocarbons, a quantitative, mechanism‐based understanding of the deformation and compaction behaviour of gypsum is needed. Previous laboratory experiments indicate that intergranular pressure solution is an important deformation mechanism in gypsum and may be controlled by the kinetics of gypsum precipitation. To examine this further, the growth kinetics of gypsum were investigated using seed crystals and super‐saturated aqueous solutions prepared from natural gypsum powders. The results showed both nucleation and growth dominated behaviour. Results relating purely to seed (over)growth showed that at low driving forces (<1 kJ/mole) precipitation follows a second order growth law, while at larger driving forces the order is 3–4. The growth rates are 1 to 2 orders of magnitude slower than those reported in studies on pure systems. However, pressure‐solution creep rates, predicted by coupling the present growth data with a microphysical model for pressure solution, are 1 to 2 orders of magnitude too fast to be consistent with experimentally obtained creep rates. On the other hand, the predicted and measured creep rates show an almost identical dependence on applied stress and grain size, supporting the hypothesis that pressure‐solution creep in gypsum is controlled by the kinetics of precipitation.

Flow behavior of fine-grained synthetic dunite in the presence of 0.5 wt% H2O

The deformation behavior of synthetic dunite (Mg-forsterite plus 2.5 vol% enstatite) with a grain size of ∼1 μm has been investigated in combined constant displacement rate and temperature stepping experiments at temperatures of 850° to 1000°C, strain rates of 10−7 to 10−5 s−1 and a confining pressure of 600 MPa. Nominally dry material behaved purely elastically. In contrast, samples with 0.5 wt% added water flowed at stresses between 9 and 81 MPa reaching strains up to 12%. Fitting a Dorn-type power law to the wet data yielded stress exponents (n) of 1.7±0.4 and apparent activation energies in the range 302±22 kJ mol−1. The wet-deformed samples showed polygonal grains, no crystallographic preferred orientation, no subgrains, low dislocation densities, and evidence for grain boundary cavitation. Grain growth was minor. It is concluded that the wet samples probably deformed by a grain size sensitive, grain boundary sliding dominated mechanism, though the nature of the accommodation process remains unclear, and the possibility of transitional behavior between dislocation creep and diffusion creep cannot be eliminated. Application of the results to upper mantle shear zones supports previous speculation that the presence of such zones can lead to a significant weakening of the top 10 to 20 km of the upper mantle during extension of the continental lithosphere.