Diffusion Creep Regime Research Articles

Lateral and radial viscosity structure beneath Fennoscandia inferred from seismic and magnetotelluric observations

Fennoscandia is continuously uplifting in response to past deglaciation, termed glacial isostatic adjustment or GIA, and its mantle viscosity is well constrained from ice sheet and sea level data. Here, we compare those GIA-constrained viscosities for the Fennoscandian upper mantle with geophysically-constrained viscosities. We construct the upper mantle viscosity structure of Fennoscandia by inferring temperature and water content from seismic and magnetotelluric (MT) data. Using a 1-D MT model for Fennoscandian cratons together with a global seismic model, we infer an upper mantle viscosity (below 250 km) of ∼1021±2 Pa·s, which encompasses the GIA-constrained viscosities of 1020 − 1021 Pa·s. The GIA viscosities are better matched if the Fennoscandian upper mantle is a wet harzburgite or a dry pyrolite, where pyrolite is ∼10 times more viscous than harzburgite. Using the average temperatures and water contents for harzburgitic upper mantle, the GIA viscosities require 1–4 mm grain sizes indicating a diffusion creep regime. In northwestern Fennoscandia, where a high-resolution 2-D resistivity model is available, greater inferred mantle water content implies viscosities that are 10–100 times lower than those for the Fennoscandian Craton. Our work suggests that the combination of seismic and MT observations can improve upper mantle viscosity estimates, especially for regions with laterally-varying viscosity structures or where GIA constraints are not available. Although our method represents an important step forward, viscosity uncertainty can be further reduced by incorporating additional constraints on rock composition, grain size and mantle stress, as well as more accurate geophysical data, into the viscosity calculation.

Deformation of Al-Cr spinel in the diffusion creep regime: mechanical results and flow laws

Deformation of Al-Cr spinel in the diffusion creep regime: mechanical results and flow laws

Influence of grain size on the high-temperature creep behaviour of M5Framatome1 zirconium alloy under vacuum

The effect of grain size on the viscoplastic behaviour of M5Framatome zirconium alloy thin sheets was investigated at high temperature under uniaxial tension, using a variety of equiaxed microstructures with controlled grain sizes. In the α phase domain, a Coble diffusional creep regime and a dislocation creep regime were observed, in agreement with the literature. A negative sensitivity of the strain rate to temperature was highlighted in the upper part of the α+β two-phase temperature range, consistently with the literature. For the first time, a linear creep regime was evidenced in the β phase domain. In this regime, a sensitivity of the strain rate on the third power of the grain size is observed, suggesting a Coble regime with diffusion along grain boundaries. Modelling with multi-mechanism Norton power-law rate equations, including dependence on grain size and a homogenous strain-rate assumption (Taylor model), enabled to satisfactorily reproduce the experimental results over the 700–1100 °C temperature range, especially the negative sensitivity of the strain rate to temperature between 880 and 930 °C. Very good agreement was obtained with a second order self-consistent and full field homogenization schemes.

Low‐Temperature Plasticity and Dislocation Creep of Fangshan Dolomite

AbstractTo explore the rheology of dolomite and investigate recent findings regarding the so‐called inversion of activation energy between dislocation and diffusion creep, we compressed medium‐grained Fangshan dolomite (113 ± 42 µm) at effective confining pressures of 50–300 MPa, temperatures of 27°C–900°C, and strain rates of 10−6 to 2 × 10−4 s−1 using a Paterson gas‐medium apparatus. Two end‐member deformation regimes with corresponding diagnostic flow laws and microstructures were identified. At temperatures ≤500°C, low‐temperature plasticity (LTP), which is characterized by microstructures of predominant abrupt undulatory extinctions and f‐twins, was determined to dominate the deformation of Fangshan dolomite. The corresponding flow behavior can be described by an with and (Regime 1). At temperatures ≥800°C, dislocation creep, which shows characteristic microstructures of smooth undulating extinction and new recrystallized grains, dominated the deformation of Fangshan dolomite. The corresponding flow behavior can be expressed by a power law equation, with , , and (Regime 2). At temperatures between ∼500 and 800°C, a transition regime between LTP and dislocation creep was identified (Regime 3) with the dependence of flow stress on strain rate increasing gradually with increasing temperature. When extrapolated to natural conditions, our flow law of dislocation creep for dolomite in combination with that of diffusion creep reported by Davis et al. (2008) suggests that the dislocation creep regime of dolomite is limited to a relatively narrow region of high temperature and relatively high stress, whereas the diffusion creep regime dominates the deformation of dolomite in tectonic settings with low stress levels.

Rheological property of H2O ice VI inferred from its self-diffusion: Implications for the mantle dynamics of large icy bodies

The volume diffusion coefficient of water in ice VI was determined in the pressure-temperature range of 1.3–1.9 GPa and 300–320 K by in situ isotope tracer diffusion experiments. We determined the activation energy of the volume diffusion to be 61.9 ± 9.5 kJ/mol. The viscosity of polycrystalline ice VI under diffusion creep was estimated from the diffusion coefficients based on the theory of the diffusion creep. From a compilation of viscosity values in the current diffusion creep regime and the viscosity previously determined by plastic deformation experiments of ice VI in a high stress regime (Durham et al. 1996), here we provide the relationships between viscosity, stress, and the average grain size of polycrystalline ice VI. The most plausible deformation mechanism of the layers of ice VI underneath the internal oceans in large icy bodies was inferred from the viscosity-stress-average grain size relationship. We also discuss the critical thickness of the ice VI layer, which determines the onset of thermal convection.

Rock Strength and Texture Evolution During Deformation in the Earth's Ductile Lithosphere: A Two‐Phase Thermodynamics Model

AbstractBuilding upon the two‐phase and grain damage theory, we propose a new formulation allowing to track the evolution of phase mixing/segregation during ductile deformation of a two‐phased aggregate. Our model is based on a set of variables characterizing a rock texture: the mean grain sizes and the phase proportion. During ductile deformation, activation of different micromechanical processes impacts the aggregate texture. Dislocation and diffusion creep are the two main deformation processes considered. We only account for the effect of Zener pinning in slowing down grain growth and allow for active grain‐size reduction mechanisms in the diffusion creep domain. For this purpose, an equation is proposed to track the phase mixing evolution during ductile deformation. Numerical application using anorthite rheology shows that any grain reduction mechanisms that could be active in the diffusion creep regime requires a very high partition fraction in order to reach the grain size predicted by the feldspar piezometer. Application of this model to gabbroic composition, relevant for the ductile crust, demonstrates that the strong coupling between phases grain sizes and interface evolution results in steady‐state grain sizes far below the field boundary. This effect is coeval with an important increase of mixing between the two phases. In addition, accounting for the phase mixing results in a drop of the global aggregate stress during deformation. This model allows for further comparison of mylonitized textures evolution with natural shear zones at the local and regional scales.

Diffusion Creep of Enstatite at High Pressures Under Hydrous Conditions

AbstractMantle convection and large‐scale plate motion depend critically on the nature of the lithosphere‐asthenosphere boundary and thus on the viscosity structure of Earth's upper mantle, which is determined by the rheological properties of its constituent minerals. To constrain the flow behavior of orthopyroxene, the second most abundant constituent of the upper mantle, deformation experiments were carried out in triaxial compressive creep on fine‐grained (~6 μm) samples of enstatite at high pressures (3.8–6.3 GPa) and high temperatures (1323–1573 K) using a deformation‐DIA apparatus. Based on results from this study, the deformation behavior of enstatite is quantitatively presented in the form of a flow law that describes the dependence of deformation rate on differential stress, water fugacity, temperature, and pressure. Specifically, the creep rate depends approximately linearly on stress, indicating deformation in the diffusion creep regime. A least squares regression fit to our data yielded a flow law for diffusion creep with an activation energy of ~200 kJ/mol and an activation volume of ~14 × 10−6 m3/mol. The magnitude of the water‐weakening effect is similar to that for olivine with a water fugacity exponent of r ≈ 0.7. This strong dependence of viscosity on water fugacity (concentration) indicates that the viscosity of an orthopyroxene‐bearing mantle varies from one geological setting to another, depending on the large‐scale water distribution. Based on the rheology contrast between olivine and enstatite, we conclude that olivine is weaker than enstatite throughout most of the upper mantle except in some shallow regions in the diffusion creep regime.

Grain‐ to multiple‐grain‐scale deformation processes during diffusion creep of forsterite + diopside aggregate: 1. Direct observations

AbstractWe uniaxially deformed fine‐grained (~ 1 μm) forsterite + diopside (5 and 20 vol %) aggregates in the diffusion creep regime. Prior to deformation, line markers were milled on a lateral surface of a cylindrical sample to detect single‐ to multiple‐grain‐scale deformation. We performed deformation experiments and observations of the marker‐etched surface after sample cooling multiple times on the same specimens. The strain measured at the scale of several tens of grains from macroscopic shortening of the markers parallel to the compression axis is consistent with the total strain of the sample. However, microscopically, the markers are intensely segmented and rotated at the grain scale increasing with the sample strain. Meanwhile, essentially, no deformation is observed within the grains in most of the samples. The surface microstructures, including the deformation of the markers, reveal the serial operations of grain boundary migration, grain boundary sliding, rigid body grain rotation, and grain neighbor switching, which correspond well to processes expected in diffusion‐controlled superplasticity. This sequence is commonly observed in both samples consisting of forsterite grains of tabular and equiaxed grain shapes, which have been shown to develop notable crystallographic preferred orientation (CPO) and random (or weak) CPO, respectively, during diffusion creep. Intragranular regions of relatively larger forsterite grains in the specimens deformed at stresses near the transition between deformation mechanisms from diffusion creep to dislocation creep reveal marker deformation and formation of surface creases and subgrain boundaries, which indicate intragranular dislocation processes. Overall, the surface microstructures reflect the deformation state of the materials well.

Grain-damage hysteresis and plate tectonic states

Shear localization in the lithosphere is an essential ingredient for understanding how and why plate tectonics is generated from mantle convection on terrestrial planets. The theoretical model for grain-damage and pinning in two-phase polycrystalline rocks provides a frame-work for understanding lithospheric shear weakening and plate-generation, and is consistent with laboratory and field observations of mylonites. Grain size evolves through the competition between coarsening, which drives grain-growth, and damage, which drives grain reduction. The interface between crystalline phases controls Zener pinning, which impedes grain growth. Damage to the interface enhances the Zener pinning effect, which then reduces grain-size, forcing the rheology into the grain-size-dependent diffusion creep regime. This process thus allows damage and rheological weakening to co-exist, providing a necessary positive self-weakening feedback. Moreover, because pinning inhibits grain-growth it promotes shear-zone longevity and plate-boundary inheritance. However, the suppression of interface damage at low interface curvature (wherein inter-grain mixing is inefficient and other energy sinks of deformational work are potentially more facile) causes a hysteresis effect, in which three possible equilibrium grain-sizes for a given stress coexist: (1) a stable, large-grain, weakly-deforming state, (2) a stable, small-grain, rapidly-deforming state analogous to ultramylonites, and (3) an unstable, intermediate grain-size state perhaps comparable to protomylonites. A comparison of the model to field data suggests that shear-localized zones of small-grain mylonites and ultra-mylonites exist at a lower stress than the co-existing large-grain porphyroclasts, rather than, as predicted by paleopiezometers or paleowattmeters, at a much higher stress; this interpretation of field data thus allows localization to relieve instead of accumulate stress. The model also predicts that a lithosphere that deforms at a given stress can acquire two stable deformation regimes indicative of plate-like flows, i.e., it permits the coexistence of both slowly deforming plate interiors, and rapidly deforming plate boundaries. Earth seems to exist squarely inside the hysteresis loop and thus can have coexisting deformation states, while Venus appears to straddle the end of the loop where only the weakly deforming branch exists.

Field theory and diffusion creep predictions in polycrystalline aggregates

In polycrystals, stress-driven vacancy diffusion at high homologous temperatures leads to inelastic deformation. In this work, a novel continuum mechanics framework is proposed to describe the strain fields resulting from such a diffusion-driven process in a polycrystalline aggregate where grains and grain boundaries are explicitly considered. The choice of an anisotropic eigenstrain in the grain boundary region provides the driving force for the diffusive creep processes. The corresponding inelastic strain rate is shown to be related to the gradient of the vacancy flux. Dislocation driven deformation is then introduced as an additional mechanism, through standard crystal plasticity constitutive equations. The fully coupled diffusion-mechanical model is implemented into the finite element method and then used to describe the biaxial creep behaviour of FCC polycrystalline aggregates. The corresponding results revealed for the first time that such a coupled diffusion-stress approach, involving the gradient of the vacancy flux, can accurately predict the well-known macroscopic strain rate dependency on stress and grain size in the diffusion creep regime. They also predict strongly heterogeneous viscoplastic strain fields, especially close to grain boundaries triple junctions. Finally, a smooth transition from Herring and Coble to dislocation creep behaviour is predicted and compared to experimental results for copper.

Low oxygen fugacity dependency for the deformation of partially molten lherzolite

We present the results of an experimental investigation of the influence of oxygen fugacity on the deformation of a partially molten spinel lherzolite using a 0.1MPa gas-media creep rig under temperatures of 1160–1190°C, stresses of 4–74MPa and well-controlled oxygen fugacities of 10−7 to 10−11MPa. The partially molten spinel lherzolite was deformed by a dislocation-mediated creep process such as dislocation creep or dislocation-accommodated grain boundary sliding (DGBS) (n=3.5±0.3) as well as in the diffusion creep regime (n=1.2±0.2). The average oxygen fugacity exponent of lherzolite is 0.04±0.02, which is significantly smaller than the values measured for olivine single crystals (0.10–0.36) and dunite rocks (0.20) deformed by dislocation creep. We attribute the low oxygen fugacity exponent of partially molten lherzolite samples either to the operation of grain boundary sliding, if DGBS dominates the deformation, or to the presence of pyroxenes whose deformation has a weak or no dependency on oxygen fugacity, if dislocation creep dominates the deformation. In the latter case, the oxygen fugacity exponent decreases rapidly with decreasing volume fraction of olivine. The low oxygen fugacity exponent for our partially molten lherzolite samples implies that dislocation creep or DGBS of mantle peridotite will most likely be insensitive to variations of oxygen fugacity in the upper mantle.

Mechanisms for the generation of plate tectonics by two-phase grain-damage and pinning

Weakening and shear localization in the lithosphere are essential ingredients for understanding how and whether plate tectonics is generated from mantle convection on terrestrial planets. We present a new theoretical model for the mechanism of lithospheric weakening and shear-localization and hence plate generation through damage, grain evolution and Zener pinning in two-phase (polycrystalline) lithospheric rocks. Grain size evolves through the competition of coarsening, which drives grain growth, with damage, which drives grain reduction. However, in a two-phase medium the interface between phases induces Zener pinning, which impedes grain growth and facilitates damage. The size of the pinning surfaces is given by the roughness of the interface, and damage to the interface causes smaller pinning surfaces, which in turn drive down the grain-size, forcing the rheology into the grain-size-dependent diffusion creep regime. This process allows damage and rheological weakening to co-exist, which is normally considered impossible in single phase assemblages. Moreover pinning greatly inhibits grain-growth and shear-zone healing, which is much faster in single phase materials. Hence, the resulting shear-localization is rapid (less than 1Myr), but the healing time for a dormant weak zone is very slow (greater than 100Myrs); these effects therefore permit rapidly forming and long-lived plate boundaries. The model therefore provides a key ingredient and predictive theory for the generation of plate tectonics on Earth and other planets.

Microstructural and metamorphic evolution of a high‐pressure granitic orthogneiss during continental subduction (Orlica–Śnieżnik dome, Bohemian Massif)

AbstractA microstructural and metamorphic study of a naturally deformed medium‐ to high‐pressure granitic orthogneiss (Orlica–Śnieżnik dome, Bohemian Massif) provides evidence of behaviour of the felsic crust during progressive burial along a subduction‐type apparent thermal gradient (∼10 °C km−1). The granitic orthogneisses develops three distinct microstructural types, as follows: type I – augen orthogneiss, type II – banded orthogneiss and type III – mylonitic orthogneiss, each representing an evolutionary stage of a progressively deformed granite. Type I orthogneiss is composed of partially recrystallized K‐feldspar porphyroclasts surrounded by wide fronts of myrmekite, fully recrystallized quartz aggregates and interconnected monomineralic layers of recrystallized plagioclase. Compositional layering in the type II orthogneiss is defined by plagioclase‐ and K‐feldspar‐rich layers, both of which show an increasing proportion of interstitial minerals, as well as the deformation of recrystallized myrmekite fronts. Type III orthogneiss shows relicts of quartz and K‐feldspar ribbons preserved in a fine‐grained polymineralic matrix. All three types have the same assemblage (quartz + plagioclase + K‐feldspar + muscovite + biotite + garnet + sphene ± ilmenite), but show systematic variations in the composition of muscovite and garnet from types I to III. This is consistent with the equilibration of the three types at different positions along a progradeP−Tpath ranging from <15 kbar and <700 °C (type I orthogneiss) to 19–20 kbar and >700 °C (types II and III orthogneisses). The deformation types thus do not represent evolutionary stages of a highly partitioned deformation at constantP−Tconditions, but reflect progressive formation during the burial of the continental crust. The microstructures of the type I and type II orthogneisses result from the dislocation creep of quartz and K‐feldspar whereas a grain boundary sliding‐dominated diffusion creep regime is the characteristic of the type III orthogneiss. Strain weakening related to the transition from type I to type II microstructures was enhanced by the recrystallization of wide myrmekite fronts, and plagioclase and quartz, and further weakening and strain localization in type III orthogneiss occurred via grain boundary sliding‐enhanced diffusion creep. The potential role of incipient melting in strain localization is discussed.

Stress-driven Melt Segregation in Partially Molten Feldspathic Rocks

High-strain torsion experiments were performed on a series of samples composed of anorthite plus51 to 12% melt to investigate the formation of melt-rich bands produced by stress-driven melt segregation. Fine-grained (3^4 � m) samples were deformed in the diffusion creep regime at a temperature of 1450 K and a confining pressure of 300 MPa at shear strain rates of 1 � 10 ^4 to 16 � 10 ^4 s ^1 and shear stresses of 15^150 MPa to shear strains between � ¼ 1� 9a nd 6� 6. The dependence of viscosity, � , on melt fraction, � , for these partially molten aggregates can be expressed as � ¼ 2� 6 � 10 12 exp (^24 � ) Pa s. In each sample, melt-rich bands develop by a shear strain of � ¼ 1, forming a population of bands at an angle of 5^158 to the shear plane and 40^308 to the applied maximum principal stress. The spacing between and width of the meltrich bands increases as melt fraction increases from50� 01 to 0� 06, then roughly levels off as melt fraction increases to 0� 12. This band spacing, � s, increases linearly with increasing compaction length,

Crystallographic preferred orientation produced by diffusional creep of harzburgite: Effects of chemical interactions among phases during plastic flow

Experimental deformation experiments have been conducted on fine‐grained, two‐phase aggregates of olivine and orthopyroxene to investigate the role of grain and phase boundary sliding on rheology and fabric development. A suite of large‐strain (γ ≥ 1) general shear experiments conducted at T = 1200°C and P = 1.6 GPa on aggregates ranging from 65 to 0 vol % orthopyroxene, to characterize the evolution of fabric, were complemented by small‐strain axial compression experiments at T = 1200°C and P = 0.3 GPa, to better constrain the rheology. Microstructural and rheological data suggest that deformation of these two‐phase aggregates in the diffusion creep regime occurs via interface‐reaction‐controlled diffusion creep that is accompanied by extensive migration of olivine‐orthopyroxene phase boundaries. The resulting rheologies suggest that olivine + orthopyroxene composites are weaker than the olivine end‐member at the conditions tested. Physically, this behavior arises because long‐range, i.e., grain‐scale, diffusion of Si4+ is unnecessary in these pseudobinary two‐phase aggregates. We further demonstrate that interface‐controlled diffusion creep leads to strong crystallographic preferred orientations (CPO) of the component minerals, which develops in the near absence of dislocation activity. The CPO formed in these anhydrous, low‐stress experiments has the olivine a axis aligned perpendicular to the flow direction (“type B” fabric) argued by some to be the unique result of deformation under conditions of high differential stress and high water fugacity. Phase boundary dynamics, thus, are argued as a significant factor in the accumulation of strain in polyphase aggregates.

The effect of dissolved magnesium on creep of calcite II: transition from diffusion creep to dislocation creep

We extended a previous study on the influence of Mg solute impurity on diffusion creep in calcite to include deformation under a broader range of stress conditions and over a wider range of Mg contents. Synthetic marbles were produced by hot isostatic pressing (HIP) mixtures of calcite and dolomite powders for different intervals (2–30 h) at 850°C and 300 MPa confining pressure. The HIP treatment resulted in high-magnesian calcite aggregates with Mg content ranging from 0.5 to 17 mol%. Both back-scattered electron images and chemical analysis suggested that the dolomite phase was completely dissolved, and that Mg distribution was homogeneous throughout the samples at the scale of about two micrometers. The grain size after HIP varied from 8 to 31 μm, increased with time at temperature, and decreased with increasing Mg content (>3.0 mol%). Grain size and time were consistent with a normal grain growth equation, with exponents from 2.4 to 4.7, for samples containing 0.5–17.0 mol% Mg, respectively. We deformed samples after HIP at the same confining pressure with differential stresses between 20 and 200 MPa using either constant strain rate or stepping intervals of loading at constant stresses in a Paterson gas-medium deformation apparatus. The deformation tests took place at between 700 and 800°C and at strain rates between 10−6 and 10−3 s−1. After deformation to strains of about 25%, a bimodal distribution of large protoblasts and small recrystallized neoblasts coexisted in some samples loaded at higher stresses. The deformation data indicated a transition in mechanism from diffusion creep to dislocation creep. At stresses below 40 MPa, the strength was directly proportional to grain size and decreased with increasing Mg content due to the reductions in grain size. At about 40 MPa, the sensitivity of log strain rate to log stress, (n), became greater than 1 and eventually exceeded 3 for stresses above 80 MPa. At a given strain rate and temperature, the stress at which that transition occurred was larger for samples with higher Mg content and smaller grain size. At given strain rates, constant temperature, and fixed grain size, the strength of calcite in the dislocation creep regime increased with solute content, while the strength in the diffusion creep regime was independent of Mg content. The results suggest that chemical composition will be an important element to consider when solid substitution can occur during natural deformation.

Effects of solute segregation on the grain-growth kinetics of orthopyroxene with implications for the deformation of the upper mantle

The kinetics of static grain growth in a pure synthetic orthopyroxene and orthopyroxene of natural composition were investigated to determine the effects of solute segregation on grain growth. In all experiments the grain size of the synthetic orthopyroxene increases considerably; in contrast, the natural orthopyroxene shows no perceptible grain growth at the same conditions. High-resolution X-ray compositional mapping confirms that the natural orthopyroxene has large concentrations of aluminum and calcium cations segregated on and near its grain boundaries. This segregation is believed to be responsible for the slow grain-growth kinetics in the natural orthopyroxene. Slow grain growth may permit long-term deformation in the diffusion creep regime, as has been observed in some deep mantle xenoliths. As such, it may provide a means of weakening the lithosphere.

On convection in ice I shells of outer Solar System bodies, with detailed application to Callisto

It has been argued that the dominant non-Newtonian creep mechanisms of water ice make the ice shell above Callisto's ocean, and by inference all radiogenically heated ice I shells in the outer Solar System, stable against solid-state convective overturn. Conductive heat transport and internal melting (oceans) are therefore predicted to be, or have been, widespread among midsize and larger icy satellites and Kuiper Belt objects. Alternatively, at low stresses (where non-Newtonian viscosities can be arbitrarily large), convective instabilities may arise in the diffusional creep regime for arbitrarily small temperature perturbations. For Callisto, ice viscosities are low enough that convection is expected over most of geologic time above the internal liquid layer for plausible ice grain sizes (⩽a few mm); the alternative for early Callisto, a conducting shell over a very deep ocean (>450 km), is not compatible with Callisto's present partially differentiated state. Moreover, if convection is occurring today, the stagnant lid would be quite thick (∼100 km) and compatible with the lack of active geology. Nevertheless, Callisto's steady-state heat flow may have fallen below the convective minimum for its ice I shell late in Solar System history. In this case convection ends, the ice shell melts back at its base, and the internal ocean widens considerably. The presence of such an ocean, of order 200 km thick, is compatible with Callisto's moment-of-inertia, but its formation would have caused an ∼0.25% radial expansion. The tectonic effects of such a late, slow expansion are not observed, so convection likely persists in Callisto, possibly subcritically. Ganymede, due to its greater size, rock fraction and full differentiation, has a substantially higher heat flow than Callisto and has not reached this tectonic end state. Titan, if differentiated, and Triton should be more similar to Ganymede in this regard. Pluto, like Callisto, may be near the tipping point for convective shutdown, but uncertainties in its size and rock fraction prevent a more definitive assessment.

Water weakening of clinopyroxenite in diffusion creep

We performed triaxial compressive creep tests to study water weakening of clinopyroxenite in the diffusion creep regime. All tests were carried out on fine‐grained samples at confining pressures between 100 and 300 MPa at temperatures of 1321 to 1421 K for water‐saturated and 1398 to 1508 K for anhydrous conditions. Samples were prepared by hot‐pressing ground powder of Sleaford Bay clinopyroxenite. Water was added to the sample during the run by dehydration of a talc sleeve. Water‐saturated aggregates crept ∼30 times faster than aggregates under anhydrous conditions at a temperature of 1400 K and a pressure of 300 MPa. The stress exponent in both cases was ∼1, and the activation energies under water‐saturated and anhydrous conditions were 340 ± 30 and 760 ± 20 kJ/mol, respectively. Under water‐saturated conditions creep rate increased with increasing water fugacity to the power 1.4 ± 0.2 with an activation volume for creep of 14 ± 6 × 10−6 m3/mol. Comparison of the water weakening in aggregates of clinopyroxene, olivine, and anorthite deformed in the diffusion creep regime indicates that the creep strength of clinopyroxene has a stronger dependence on the water fugacity than do the creep strengths of olivine and anorthite. In water‐rich environments, clinopyroxene will be the weakest component in polycrystalline rocks composed of these phases.

The effect of grain size and melt distributions on the rheology of partially molten olivine aggregates

Abstract Grain size distribution affects the distribution of melt in partially molten rocks, as melt is partitioned into the finer grained regions. As the creep strength of partially molten rocks depends on melt fraction, grain size distribution affects rheology in two ways: in the diffusion-creep regime the strength of the rock decreases with decreasing grain size; and by affecting the melt distribution, grain size distribution affects rock strength in the dislocation and the diffusion-creep regimes. To investigate this effect on partially molten dunites, we deformed synthetic aggregates of olivine + mid-ocean ridge basalt (MORB) with varying grain size distributions and melt fractions. We found that samples with bimodal grain size distributions did not obey the flow law derived for samples with a uniform grain size distribution. In the samples with bimodal grain size distributions the finer grained parts are enriched in melt, reducing the relative melt fraction in the coarser grained parts. The stress exponent is close to 1 for lower stresses for samples with non-uniform grain size distributions, while it is close to 3 for samples with a uniform grain size distribution. Electron backscatter diffraction (EBSD) analysis shows crystallographic preferred orientation (CPO) well developed in all samples. Microstructural investigation shows that partitioning of the liquid out of the coarser grained part of the aggregate into the finer grained part leads to a localization of the strain into the weaker, finer grained and melt-enriched part. By treating the two parts as mechanically different phases that deform at a uniform stress, we are able to approximate the rheological behaviour of the aggregates with bimodal grain size distribution.