Segregation and mixing are very current processes in nature and industry: they are generally present together in these processes. The first part of this talk will be devoted to a presentation of the different mechanisms, essentially at the microscopic scale, which can lead to segregation in granular media. Then, on the basis of recent experiments and with a particular attention to segregation due to size difference, we analyse experimental situations where segregation is observed in dry granular media, with an effort to identify the mechanisms responsible for it: segregation under shaking, in surface flow and in rotating drums. Finally we try to explore ways to obtain homogeneous mixtures of different grains.
Volcanogenic high-concentration mass flows, such as high-concentration pyroclastic flows, volcanic debris avalanches, and lahars, are essentially highly energetic flows of volcaniclastic material. The mixture of volcanic material, highly fragmented by the eruption processes, and a fluid, e.g., air, volcanic gas, or water, is intensely sheared in the fast moving mass flow. This kind of flow is dominated by the "grain-inertia" or "rapid granular flow" regime: due to the high solids-concentration and high velocity gradients stress development and transfer is controlled solely by particle-particle collisions. This determines the bulk mechanical behaviour, the development of local shear zones, segregation processes, and, hence, the appearence of the final deposit. To study the general properties of the rapid granular flow regime and its applications to volcanogenic high-concentration mass flows a discrete particle model has been developed, which allows to investigate the processes described above in detail. The results of this investigation show that many phenomena observed in the field, e.g., flow behaviour, gradings, development of flow units, can be explained by the processes acting in the rapid granular flow regime.
Sturzstroms are giant landslides that travel kilometers within minutes. Data from terrestrial examples of0.001-10 cu. km suggest that their runout lengths increase in proportion to the square root of their volume. This trend is investigated assuming that runout is controlled by fragmental flow. The results indicate that runout lengths depend on the potential energy available for motion after initial collapse (itself a function of the dimensions of the collapse zone), on the degree of rock fragmentation, and on the rate of momentum loss by fragment collisions in a basal boundary layer, assumed to thicken with time by diffusion. A dependence on initial conditions during collapse would explain claims that a minimum volume is required for sturzstroms to form: Beneath a critical volume, insufficient energy is available to initiate fragmental flow, and so the unstable mass slumps downslope. The drop height (H) is often similar to the vertical extent of the collapse zone and so is linked by geometry to sturzstrom volume. The ratio of H to L (the horizontal distance of effective transport), normally interpreted as a measure of frictional resistance, is thus reinterpreted as an inverse measure of the energy available for runout after collapse. By providing a physical basis for observed trends, the analysis justifies use of empirical limits for forecasting the runout lengths of major landslides.
Landslides or debris avalanches regularly cause large amounts of human and material damages so that a good description of these gravitational flows is of great importance. Since the characteristic length of the flowing material is generally much larger than the fluid thickness, we have developed here a numerical debris avalanche model solving the vertically integrated long wave equations in a coordinate system linked to the topography (Mangeney et al., submitted). This model takes into account a friction law at the base of the flow (Savage and Hutter, 1989) and is based on a shock-capturing numerical method, similar to those currently used to simulate compressible flows with shock waves. One of the major problems for volcanic events forecasting is to assess the overflowing of the topographic relief. The sensitivity of the flow to the topography is investigated here theoretically for different values of the friction angle. The model is then applied to 3D real topography inorder to simulate the 26th December 1997 debris avalanche in Montserrat, Lesser Antilles (Mangeney et al., in press). Results are compared with data on debris deposits given by the Montserrat Volcano Observatory. Debris avalanche path prediction is then discussed referring to the previous theoretical analysis.
Mangeney A, Heinrich Ph & Roche R, Pure Appl. Geophys, (1998).
Savage SB & Hutter K, J. Fluid Mech, 199, 177-215, (1989).
Mangeney A, Heinrich P, Roche R, Boudon, G & Cheminee JL, J. Phys. Chem. Earth, (1998).
Debris avalanches are voluminous (>106 m3) fast moving and long-runout gravity flows of granular materials resulting of large-scale flank collapse of a volcano or a mountain. These flows are highly mobile as indicated by low apparent internal friction coefficients (H/L<0.3), where H is the fall height and L the horizontal runout distance. The Flims deposits of Holocene age result from the most voluminous debris avalanche in the Alpine belt (12 km3, H/L = 0.13) that mobilized Jurassic limestones. Deposits occur in the Upper Rhein valley on a distance of 16 km, and they reach up to 300 m in thickness.
In the deposits, various facies were recognized, and are related to the lithology of the transported material and to their location within the deposits. (1) The bedded facies constitutes the main part of the deposits. It is characterized by fractured bed packets which are plurimetric in thickness, and are limited by major cataclasic surfaces (surfaces 1). These bed packets are fragmented by oblique fractures with decimetric spacings (surfaces 2) that dip upslope. Close to the base, the material is pervasively microfracturated whereas in the overlying levels the deposits are more massive. (2) This facies grades vertically and laterally into the granular facies. It corresponds to poorly sorted breccias in which shattered clasts up to meters in size float in a fine grained matrix. It is sometimes stratified into layers of various grain size. Normal faulting can affect these deposits close to the sommital surface. The avalanche deposits directly overly fluvial gravels of the valleys floors, but poor mixing occurs between both materials.
The transport of the debris avalanche involves major layer-parallel slip movements along surfaces 1 that induce a stretching and thinning of the bed packets by clasts movements along surfaces 2. The granular facies results of the destruction of the fabric that characterizes the bedded facies by a flow transformation suggesting a turbulent behavior. This transformation is inhibited in proximal domains and at the base of the avalanche by topographic confinement in the valley and the load of the moving mass. Then, the preserved internal fabric of the bedded facies indicates that it results from early evolution stage of the moving debris avalanche.
Contribution n° 138 of the PNRN (INSU, Paris)
Debris avalanches are generally modelled by a continuum medium assuming that the grain diameters are small compared to the avalanche height. Moreover, the Navier-Stokes equations are generally depth-averaged and the flow of debris is described by solving long wave equations in a coordinate system linked to the topography. In this paper, the validity of the long wave equations is studied by comparing results calculated by a Navier-Stokes model and a long wave model. The comparisons are carried out first for a volume with an initial parabolic shape flowing down an inclined plane in a 1D geometry. The effects of an obstacle are then studied showing the limits of the long wave approximation for such problems. For both cases, friction is neglected on the slopes. Finally, a parabolic shape of the slope is tested in both numerical models and the approximation of long waves is corrected in the long wave model for a small curvature.
Recent thermal and fluid dynamical models have shown that density-driven ascent of isoviscous granitic melt through the crust in narrow dykes is geologically instantaneous, provided the role of suspended crystals is ignored. However, it is well known that solids in suspension can affect significantly the rheology and hence flow properties of magmas. Using granular flow theory, we show how fluctuations in particle velocity in the shear field can be used to predict the rheological behaviour of magmas with crystal contents (solidosity) in the range 0.1 < <phi>< 0.25. Specimen calculations for the solidosity and fluctuation on particle velocity for ascending granitic magmas show that the energy associated with these irregular movements is much greater than that expected on purely Brownian grounds. We find no clear correlation between viscosity m and the yield strength, and while it is possible that magma has curious, unexpected bonding properties, magmatic suspensions where the grain size is on the scale of centimetres (eg. some coarsely porphyritic magmas), low estimated Bagnold numbers (ca. 10-8) render Bingham flow models inappropriate. Self-organisation of the flow into a crystal-poor margin and crystal-rich core (flow differentiation) by diffusion of particles towards the region of minimal shear results in the familiar plug-type flow profiles traditionally associated with an inherent yield strength. Shear-enhanced diffusion is an inevitable consequence of granular flow that leads to flow differentiation, providing a simple mechanistic explanation for high concentrations of pheoncrysts in the cores of porphyritic dykes (Bagnold effect). Deviation from simple Newtonian flow is unlikely to be significant until the magma has stopped ascending and emplacement begins.
The rheological behaviour of near-solidus magmatic systems is a matter of current debate. Inferences from rock structures and application of physical models help us to understand the particular phenomena that occur during magma emplacement and flow through the continental crust. The study of particular structures appearing in granitic (sensu lato) rocks suggest the existence of phenomena of rheomorphism and fluidification of near-solidus magmatic systems, in which the rheological critical threshold (e.g. RPT of Vigneresse et al. 1996) was previously passed. The term rheomorphism was introduced in the geological literature by Backlund (1937) to describe complex processes of transformation of rocks into a totally or partially fluid mass. These particular structures are characterised by the following features: (1) they form tabular bodies, several cm width and tens of m length, with the resemblance of intruding dikes, (2) the contacts with the host granites are sharp at the scale of the crystal size, (3) there is no plastic intracrystalline deformation, and (4) the whole chemical composition is identical to that of the host granites. The only appreciable difference is the texture: the dikes have a bimodal crystal size distribution and the host granite is essentially unimodal. These dike-like structures, that we call rheomorphic dikes, are very common in different plutonic complexes and appear in two different situations: (1) associated to intramagmatic shear zones, and (2) not associated to shear movement but dilatant structures in the near-solid magma. Both cases have been studied with the result that they show similar textural relationships and, consequently, may have a similar origin. An important aspect is the common presence of clusters of several crystals in the host granitoid. These polycrystalline aggregates are frequently formed by plagioclases and biotites. We have used image analysis of electronic pictures (BSE, Z-contrast) to determine the crystal size distributions for the main crystalline phases (Qtz, Bt, Pl, Kfs) in both dikes and magmatic shear zones as well as in the host granitoid. The textural analysis indicates that these rheomorphic structures may be the result of fluidification by disruption of polycrystalline aggregates in a near-solidus magmatic system. Magma from these rheomorphised zones may intrude as dikes into the more viscous, near-solid, surrounding magma. These results may have implications for the understanding of intrusions and autointrusions of granitic magmas in the continental crust, mainly for the late process associated to magma emplacement.
Backlund HG, Bull. Geol. Inst. Univ. Uppsala, 27, 219-269, (1937).
Vigneresse JL, Barbey P & Cuney M, Journal of Petrology, 37, 1579-1600, (1996).
We present a model of partial cooling and crystallization for intrusions of acid magmas in the continental crust. Our model is calibrated in shape and size by data obtained from real plutons, after gravity data inversion. It also takes into account the effects of latent heat, as well as a linearized rate of crystallization based on experimental studies of the respective rate of mineral crystallization. We examine the rate to which three major stages are reached during cooling : the development of a fabric (25% crystals), rigidity onset (50% crystals) and the total locking of the system (75% crystals). We specifically focus on the rigid percolation threshold (RPT, 50% crystals) which marks the onset of rigidity, i.e. the point at which the framework formed by the crystals can react to ambiant stress field. The parameters that control cooling are the difference in temperature between the intrusion and the surrounding rocks, as well as the interval between liquidus and solidus temperature for magma crystallization. The time to reach rigidity increases as a function of the square of the intrusion size. However, this parameter can be constrained by data on magma volume delivery. The anisotropy of the intrusion also reveals to be important since the smaller side of the intrusion control the bulk cooling rate. We introduce a rigidity rate which corresposnds to the velocity of crystallizing 50% solids in the whole pluton. It represents an averaged rate of cooling. The rigidity rate compares favorably (50-200 mm/y) with tectonic rates (1-5 cm/y). It implies that crystallizing granites may sustain stress, react by compression to regional stress field and locally alter the surrounding stress pattern.
The geochemical equilibrium and disequilibrium melting accompanied by fractionation of partial melts from residual crystal matrices is a phenomena believed to result in a layered features of igneous rocks (Langmuir, 1989). The gabbroid rock layering has been modelled in laboratory experiments with a help of a powerful centrifuge and high-temperature furnace assembly. A time series of centrifuge experiments with varying rate of melt crystal-fractionation have been performed on partially molten fine-grain pyroxen gabbro. Initial samples of gabbro consist of 60-65 vol.% of plagioclase (Pl) with An-index 90-95, 25-30 vol.% augite (Cpx), 5-7 vol.% of titanomagnetite (Ti-Mt), <1 vol.% of olivine (Ol) have been partially annealed at 1235 and 1290°C in air for 24-30 h. Initial samples of partially molten gabbro contained 55-60 vol% (1235°C) and 65-70 vol.% (1290°C) of partial melt plus Pl and Ti-Mt crystals (Px and Ol have been absent). Cylindrical samples were loaded in Al2O3 capsules (diameter 5 mm, height 20 mm) and centrifuged for 2 h at accelerations (a) 100, 200, 500, and 1000 g at 1235°C, and (b) 1000 g at 1290°C. Chemical composition was estimated from microprobing of samples, local proportion of melt and crystals was determined from image analysis of thin sections. In the first case (a) the upper layer is under all conditions consisted of Pl-crystals + melt, whereas the mode of the bottom layer, containing both Pl and Ti-Mt, is dependent on acceleration. At low centrifuge accelerations (100-200 g) the bottom layer is almost devoid of Pl whereas at high accelerations the lower layer contains significant proportion of buried Pl-crystals below the Pl-free cap of Ti-Mt pile. Density of melt has been calculated from the chemical composition (Bottinga et al., 1982). The calculated density contrast in the case of slow fractionation is larger than in the case of fast centrifuging. The bottom layer of heavy crystals under conditions of slow fractionation has a significant chemical and density gradient. Under conditions of fast fractionation the density contrast are smeared out and the overall density gradient is negligible.In the second case (b) at 1290°C two layers of liquid of different composition and density were observed + a layer of Pl-crystals between them + a layer of Ti-Mt-crystals on the bottom. The top layer of melt with the density less than the density of Pl-crystals overlays the pile of Pl-crystals. Between layer of Ti-Mt on the bottom and layer of Pl in the centre of capsule a layer of melt with a gradient of chemical composition and density was observed.Observed acceleration and temperature dependence of modal layering depends on relative contributions from collective settling/floating of crystals, from their fractionation according to size and buoyancy. Higher acceleration, larger the difference in time scales for the fractionation of small and large crystals and as a result of this density current of heavy Ti-Mt crystals is more efficient resulting in locking of more small-size Pl-crystals in a cumulus layer. In the performed experiments the centrifuging results in a stable density gradient of melt, thus, any convection current of melt in the capsule has been avoided. Fast fractionation and melting at 1290°C results in melt segregation in two distinct liquids which may be explained in terms of the continuos melt withdrawal model in a three component eutectic system (Yoder, 1976).
Langmuir CH, Nature, 340, 199-205, (1989).
Bottinga Y, Weil D & Richet P, Geochemica et Cosmochemica Acta, 46, 909-919, (1982).
Yoder HS, Generation of Basaltic Magma, National Academy of Sciences, Washington DC, 205 pp, (1976).
Understanding why and how strain-localization occurs in rocks is a major goal of rheology of the crust, with application in basement and economic geology. For a given rock type, localization occurs when strain-rate is locally and persistently amplified, mainly in response to variation in differential stress and temperature. Variations in rock-types (mineral compositions, grain-sizes, modification of rheology with strain amount) and the presence of fluids (volatiles, melt) play also a role at initiating and enhancing strain-localization. Strain-localization in solid-state deformed rocks is well documented at every scale. It occurs when displacements are accomodated within domains getting progressively narrower and leads to strain intensities that are heterogeneously distributed: - at the millimeter scale, at tips of «hard» grains (porphyroclasts) in response to stress concentration, at a given temperature; - at the kilometer scale, within crustal-scale shear zones in response to the interplay of several softening factors (stress concentration, input of heat and fluids, lithological and geometrical softening).
By contrast to solid rocks, non-localization of strain is expected in high-enough liquid-fraction bearing magmas, due to the almost Newtonian behaviour of melt at a given temperature, and relative independence of melt viscosity with temperature. Since strain intensity in a deforming granitic magma can be evaluated by its resulting magnetic fabric pattern, strong evidence is given that magma is a non-localizing material. Homogeneous deformation in granitic magmas, attesting for non-localization of strain, will be illustrated from the scales of >>100 km2 to a few m2 (map views), down to the scale of a few dm3 (3D fabric description).
In conclusion, when magma is present in the crust, since magma is less viscous than «solid rock», displacements are achieved mainly within the magma body. As a consequence, in an orogenic domain magma localizes strain within its proper liquid-fraction window, but no further strain-localization occurs in the magma, i.e. strain distributes evenly in the body. Finally, below a given liquid fraction (not clearly defined: 25%?), or at the fully crystallized stage, strain-localization recovers, provided that a minimum applied stress remains, as a result of the great sensitivity of strain-rate with mostly stress and temperature. Hence, cross-cutting or peripheral shear zones appear and characterize the late magmatic emplacement stage.
Fragmental rocks, such as conglomerates, breccias and tillites, are types of concentrated suspensions that geologists simplify for modelling purposes as object-matrix systems. Deformed rocks of this kind are frequently used for geological strain analysis, in view of their common occurrence throughout the geological succession. The usual method of determining strain in such rocks is by Rf-phi analysis, but this assumes that clasts are approximately ellipsoidal, and that the strain determined from a statistically significant number of clasts is the strain for the rock as a whole. Neither assumption is likely to be correct. By their name, conglomerates are aggregates of many different rock types, that must be expected to deform differently. Furthermore, the fragments are commonly far from 'round', and their shapes are likely to have an effect on the deformation.
Treagus et al. (1996) presented preliminary modelling of deformation of non-ellipsoidal objects in a matrix. This reiterated theory for deformation of isolated objects of initially circular/elliptical shape, with a viscosity contrast to the matrix: i.e. that the objects would deform homogeneously to ellipses, but by a different strain to the bulk strain. A corollary is that non-elliptical objects would not deform homogeneously, and their shapes would therefore change irregularly. Finite-element modelling shows how initially square-shaped objects with a viscosity contrast to the surrounding matrix can deform to a variety of different shapes: concave-ended barrels for stiffer objects; bone shapes or smoothly elongate lobes for incompetent objects. Yet in a diagonal orientation, the squares deform virtually into rhombs of varying diameters. Further FE modelling shows the wide variety of object strains and shapes can be derived from one initial shape (e.g. square, triangle, rectangle) in different orientations, and with various values of viscosity contrast and progressive strain. These results suggest that shape changes of irregular objects, such as the changes of straight edges to concave or convex, could provide a new source of information on viscosity contrasts in object-matrix systems. Current research is considering multiple objects in a matrix, that are more realistic models for fragmental rocks and concentrated suspensions.
In a parallel study of naturally deformed fragmental rocks, we are analysing clast shapes and strain on different scales, in conglomerates, breccias and tillites from localities in Britain, Ireland and N France. Using the results of FE modelling in conjunction with these analyses, we aim to find new ways of computing whole-rock strain from clast shapes and strain, and quantifying the strain variations in terms of viscosity ratios among component rock types.
Treagus, SH Hudleston, PJ & Lan, LJ, Struct. Geol, 18, 1167-1172, (1996).
1- What is the problem? In the brittle frictional upper sedimentary beds of earth's crust, lateral compression causes appearance of thrusts according to a process of localization. To understand how oil can move along this network of discontinuities, geologist has to reconstitute the complete evolution of the geological structure. But, he only has at his disposal in situ seismic profiles which give information about present underground architecture. So, he has to use either analogical modelling whose sandbox experiment is the most current example, or numerical modelling (Finite Element Method, Distinct Element Method) to reconstitute the chronology of thrusts.
2 - Mechanical analysis: In experimental models, similarity conditions impose the choice of granular materials (sand, Pyrex powder, glass microbeads) with very low cohesion and internal frictional angle equal to natural rock's one. These materials show, when constrain, fault zones whose width is about ten times the mean diameter of grains and that are then perfectly visible by tomography.
Bibliographical study show that mechanical parameters of the granular material, in particular its macroscopic friction angle, and the initial conditions, that is to say the thickness of the model, have respectively influence on thrusts orientation and on the width of the pop-up structures at the top. Only frictional conditions on the horizontal base and on the vertical mobile backstop change deformation mechanisms activated during the shortening of the sand bed. In other words, rigid boundaries cause disturbances that must not be integrated in the analysis of the prototype strains.
Thus, we realised a sandbox model constituted of three parts. Each of them correspond to different frictional conditions (model A: frictional backstop, smooth base; model B: smooth vertical wall, rough base; model C: all the faces are frictional). We used an equipment of the French Institute for Petrol called STRUCTURATOR and adapted to computerized X-ray tomography. The predominant parameter is the ration between frictional angle at the interface and internal frictional angle of the material.
We note that the deformation is made in three steps: the first one is the initiation of fault zones. Then, the sliding of the hanging wall on the footing one is accompanied by the formation of out-of-sequence thrusts. Finally, strains spread in a piggyback way toward foreland. Each step is dependent on the boundary conditions.
In this way, more frictional is the base, more the root point of thrusts is distant from the backstop: in the model where the base is rough, the first fault zone is a curved synthetic thrust initiated at the bottom of the piston while in the model with smooth base, two conjugate thrusts delimit a pop-up structure. In the two cases, thrusts merge at the surface in a singular point where an increase in height ends. That give evidence of the diffusion of the compression more or less ahead in the model: when shortening goes on, as the stresses are highly concentrated near the vertical wall when the basal friction is important, closely spaced synthetic and antithetic thrusts are also more numerous in this case. More, in accordance with Davis and Dahlen theory, topographic slope depends on basal detachment strength. But we note in our experience that it also depends on the parietal friction whose role was up to now neglected.
We propose to mechanically explain how the major and out-of-sequence fault zones form, this on the base of the Mohr-Coulomb law linking the critical shear stress with the normal stress and the internal friction angle. In the model with rough base, the first thrust appears at the foot of the backstop because of the high concentration of shear stresses due to friction in this zone. Under the shortening, thrust sheet slides in the weakly perturbed material. But the climbing of this sheet interfere with parietal friction whose consequence is to jam the dilatancy and so to increase the mean stress in a wedge built on the vertical wall and consequently protected from localization. This rigid wedge is perfectly visible with scanner. It is limited in its front face by an antithetic out-of-sequence thrust starting in the top corner and resulting of an increase in shear stress in this point. On the contrary, in the models with smoother base, stresses are diffused in a more homogeneous way. It results that localization depends on the backstop friction: if it is low, localization will happen in internal singularities of the material; in the opposite case, as the same process takes place than in the case of high basal strength, lateral friction introduces shear stresses along the wall whose consequences are the development of a first backthrust and the increase of the confinement stress.
However that may be, the most important fault zone is the last created antithetic thrust. Sliding along this discontinuity leads to an increase in the height of the model. In the top zone, sand will progressively undergo the compression, what explains why out-of-sequence thrusts nucleates at the point corresponding to the initial thickness. The second consequence of the sliding is the increase in the mean stress that causes the lock of the frictional sliding.
3 - Numerical modelling by the Distinct Element Method: When applied to reconstitute sandbox experiment, Finite Element Method do not respect the discrete nature of the granular material. So, we have tried to simulate this experimental test with the Distinct Element Method (PFC2D software) in which particles are independent of each other and interact according to the Newton's second law and a local contact law. These particles are in fact rolls because of the two dimensional limit of PFC2D.
In geology, this method has been used until now in the case of structures in extension and with regular assemblies of particles. We have tried to adapt this way of modelling for structures in compression. But it is not acceptable for two reasons: on the one hand, discontinuities depend on the geometry of the initial arrangement, on the other hand, repulsion strengths due to the overlap of two particles can not be dissipated because of perfectly horizontal contact stresses.
Consequently, we construct a collection of rigid rolls in an randomly way. For reasons relative to computation time and computer's performances, mean diameter of grains has to be increased. Once created, particles are not in contact. So we let the sedimentation take place before applying velocity to the vertical wall. More, preliminary studies show us the great importance of the way of representing the boundary conditions to simulate well the architecture of experimental thrusts.
We succeed in modelling 1/ the diffusion of efforts in the assembly of rolls and 2/ the discontinuity of velocity underlining the sliding of one block over another one.
In conclusion, PFC2D, if boundary conditions are correctly simulated, could be a good tool for geometrical and cinematical interpretation of sandbox experiment. But the time required for calculation and resultant approximations makes this software not very useful for the determination of mechanical data (stresses, strains) compared with the Finite Element Method.
Data on phase composition of the oval and flatterned metal-bearing microinclusion (MBM) which are disposed in the cavity of orogenic rhyolites of South Sickhote-Aline and Okhotsk-Chukotka volcanic belts recieved using a electron microprobe techniques and a electron microscope confirm the preexisted concept that these cryptocrystolline mineral aggregates (size to 0,25 mm.) are a carbon-bearing condensation-crystallization dispersion systems (DS) or aerogels. Their dispersed particles (DP) consist of grains (size from 0,01 to 50 mm.) of native metals (Pb, Sn, Zn) and alloys, shapeless relict minerals (Ti-amphibole, orto- and clinopyroxenes, sphene, andesine, zircon, epidote, quartz, albite), newly formed minerals (quartz, albite, clinopyroxenes, calcite, sericite) and multicomponents poor crystallize phases. The dispersion medium (DM) consists of a carbon-bearing substance. The MBM dissipative (space-temporary) structures are represented by the two type elements. One of them are the MBM main share, which enrich of the metal DP. Another structure elements are enriched of the carbon-bearing substance, the newly formed minerals. A chemical composition of DP, its relationships with DM, the dissipative structure availability in the MBM suggests that aerogels had been formed in result of special development of nonlinear oscillatory processises (coagulation, condencation, crystallization and polymerization) under nonequillibrium conditions from deep aerozoles, suspensions (or flows of reduced fluids included of hydrocarbons and organometallic compounds, relict grains of minerals) during their circulation through fracture within metabasic rocks and ascent to the Earth surface as gas bubbles together with rhyolite magmas. The considerable activety of basic compounds in deep fluids (aerozoles, suspensions) promotes to migration of the metallic ultradispersed particles and acid compounds - to their coagulation.
Magmas flowing during their emplacement are considered to be suspensions of rigid particles crystallising progressively in a viscous medium submitted to a simple shear deformation. The question is, how flow history in terms of direction and intensity of shear, can we inferred from the directional properties of mineral shape fabric developed during emplacement? To answer this question, three-dimensional analogue modelling of magnetite fabrics developed under high simple shear (up to <gamma>=20) were performed using a ring shear apparatus. The deformed solid suspension is made of a fluid (56% of silicone, 44% of wax) and 1% volume fraction of magnetite grains. Experimental magnetite fabrics are analysed by Anisotropy of Magnetic Susceptibility method and compared to shape fabrics measured by image analysis on oriented thin-sections using Inertia Tensor method. Directional properties - foliation and lineation - as well as strength and symmetry of the magnetic fabric ellipsoid are comparable to those obtained for shape fabric. It confirms that magnetic fabrics in low concentration suspensions are mainly controlled by the shape anisotropy of magnetite grains. During progressive deformation and low shear strains (<gamma><4), directional properties of fabric, already acquired from 1 g, evolves progressively toward the shear direction according to the theoretical fabric evolution of homogeneously shaped ellipsoidal particles population. At high shear strains (<gamma>>4), foliation stabilise close to the shear plane, lineation forming an angle with the shear direction. This stabilisation of fabric is induced by non-homogeneously distributed aspect ratios of magnetites whereas few cyclic contact-interactions developed between particles are sufficient to generated obliquity between shear direction and lineation. These results demonstrate that shape fabric properties of natural particles and deformation intensity become unconnected for simple shear strain higher than <gamma>>4.
Einstein (1906) developed a theory to describe the viscosity of a suspension of rigid spheres in a viscous fluid. Jeffery (1922) generalised Einstein's approach to ellipsoidal particles. He also developed equations describing the rotation of ellipsoids suspended in a newtonian viscous fluid. From the sixties, these equations has been widely used by geologists for modelling the development of shape-preferred orientations (SPO) and try to derive finite strain and/or flow directions in various types of rocks, from unconsolidated sediments to igneous rocks. In this contribution we propose a review of the basic theoretical (e.g. Gay, 1968; Ghosh and Ramberg, 1976; Fernandez, 1987; Jezek et al., 1996), numerical (e.g. Reed and Tryggvason, 1974; Willis, 1977; Fernandez and Laporte, 1991; Jezek et al., 1994; Ildefonse et al., 1997), and experimental (e.g. Fernandez, 1987; Ildefonse et al., 1992; Ildefonse and Mancktelow, 1993; Arbaret et al., 1997; Fernandez and Fernández-Catuxo, 1997) results concerning the development of SPO, in various types of flow (coaxial to progressive simple shear), from Jeffery's original paper to today's applications. We also discuss and review the limits of Jeffery's model when applied to Geological material (variability of particle size and/or shape, mechanical interactions, etc...). Some examples of three-dimensional numerical modelling will be shown, as well as several examples of analogue experiments.
Fernandez A & Fernandez-Catuxo J, in: Granite: from segregation of melt to emplacement fabrics. Kluwer, 145-157, (1997).
Fernandez A & Laporte D, J. Struct. Geol, 13, 337-347, (1991).
Ildefonse B, Launeau P, Fernandez A & Bouchez JL, J. Struct. Geol, 14, 73-83, (1992).
Jeffery GB, Proc. Roy. Soc. London, 102, 161-179, (1922).
Jezek J, Melka R, Schulmann K & Venera Z, Tectonophysics, 229, 165-180, (1994).
Jezek J, Schulmann K & Segeth K, Tectonophysics, 257, 203-221, (1996).
Institute of Geology, 54, Pervomaiskaya St. Syktyvkar, 167610, Russia
Numerous hipabissal intrusions of acid rocks, that have evident pyroclastic structure were found last years on western slope of the Polar Urals. These intrusions are build up of tuffisits (intrusive tuffs) and intrusive ignimbrits. These rocks are genetically associated with extended fault pluton composed A-granite (Lemva Massif) and with volcanic comagmatits of these granites. Tuffisits form the "mantle covers" veins and small diatrems placed into granite and rhyolite bodies. Intrusive ignimbrits form veins or thick dykes, sometimes up to 300 m in diameter and up to 5 km in length. Small ignimbrite dykes and veins are located within tuffisite diatremes or other tuffisite bodies, whereas the larger dykes of these rocks are placed within Paleozoic shists, which are country rocks for all volcanites of the Polar Urals. The intrusion of tuffisits was always precedes ignimbrite magmatism. Both tuffisits and ignimbrits usually contain well rounded fragments of granitoids, very similar to cobbles and pebbles. Intrsive pyroclastic rocks and associated granites and rhyolits are connected to system of conjugate arc-shaped faults. The zone of development of intrusive pyroclastic rocks in the Polar Urals is traced now up to 100 km. These rocks were formed in this region as the result of fluidisation of clastic material of explosive origin, possible together with drops of a melt, be expanding gas flows. A source of a gas was the decompressed granite magma. The gas escape was violent, frequently explosive, that caused crushing walls of magmatic camera and cooled consolidated magma, splashing of fluid magma and formation of tuffisite and ignimbrite material. Just this material formed dykes, veines, diatremes and all other intrusive pyroclastic bodies.
Microstructural characteristics of high temperature mylonites of two different localities, both inferred to be deformed by granular flow, have been compared for their mineral assemblage, grain size, phase distribution of the constituent phases, and deformation mechanisms.
The temperature conditions recorded in both mylonites indicate high temperature conditions around 600-700°C based on the Ti-content of recrystallized hornblende. However, in both mylonite series the metamorphic mineral assemblage is only locally equilibrated. Temperature estimates can only be obtained from newly formed phases, which yield the thermal conditions during certain episodes of the deformation history.
Both mylonite series consist of polyphase mixed material. In the Jotun Nappe mylonites it consists of hornblende + plagioclase ± orthopyroxene ± clinopyroxene, in the Ivrea Zone mylonites it consists of orthopyroxene + clinopyroxene + k-feldspar + plagioclase ± biotite ± quartz. In the mylonites of the Jotun Nappe the distribution of phases is linked to metamorphic processes, i.e. the consummation of metamorphically metastable primary hornblende and nucleation of new hornblende with slightly different chemical composition. In the samples of the Ivrea Zone mylonites the chemical disequilibrium is manifested in the occurrence of biotite in the pyroxene-bearing matrix, suggesting that the high grade mineral assemblage is only preserved due to water deficient conditions during deformation. The most striking difference is recorded in the grain sizes of the recrystallized matrix which is one order of magnitude smaller in the sample series from the Ivrea Zone compared to the mylonites from the Jotun Nappe. The very small grain sizes in the mylonites from the Ivrea Zone (2-10 µm) can be either explained by very high differential stresses at similar temperature conditions or by a lower temperature which may have been overestimated due to the water deficient conditions, preserving the high grade mineral assemblage during the deformation.
We present preliminary results of analogue experiments on compaction and shearing of wet granular media. We use a very simple experimental device: suspensions of rigid particles (PVC cylinders, glass beads, ...) are placed in a glass tube (diameter 15 cm), and compressed and/or sheared by a piston. As the piston goes down, the fluid is progressively expelled at the edges of the model. The displacement and the rotation of the piston are continuously recorded by transducers, and the experimental set-up also allows one to capture series of video images of the model during the experiments.
The purpose of this work is to re-examine the rheological threshold that corresponds to the locking of the rigid phase, i.e. when the rigid backbone canot be disrupted anymore, and particles have to be strained for the deformation to continue. In our experiments, this threshold is generally between 30 and 40% particle volume fraction. However, it varies significantly with 1) the variation of shape and size of the particles in the model, 2) with the deformation regime. The locking threshold is lower for cylinders than for beads, and lower for a mixture of cylinders and beads than for cylinders. Mixing particles of different sizes also reduces the locking threshold. For a given population of particles, the threshold is lower when the model is sheared than when it is compressed axially. It is also lower when the piston rotates in alternate directions (oscillatory rotation) than when it rotates in one direction (continuous rotation). The latter effect is due to the dilation of the granular medium when undergoing continuous shearing in one direction.
We will discuss the implications of these results on some physical properties of crystallising magmas (rheological threshold; transition from magmatic flow to hypersolidus, "solid-state" flow assisted by melt; seismic properties of magmatic mush).
The main genetic control of the Colombian emerald mineralization is a polymict hydrothermal breccia outcroping in all the deposits in three morphological types: firstly, it occurs along faults limiting structural horses in duplex systems, secondly, it forms stratiform horizons generaly concordant to the bedding of hosting cretaceous black shales, finaly, this breccia fills extensional fractures crosscutting the bedding.
Two classes of fragments compose the breccia: i) black fragments from the enclosing black shales, ii) white fragments from albitized shales. The matrix is either a black flour when it is black shales derived or a white flour when it results from the crushing of albitized shales. The cement is mainly composed of carbonate, pyrite, albite and some emerald mineralization. This polymict hydrothermal breccia is either matrix-supported or cement-supported depending on the rate of cementation.
The polymict hydothermal breccia is a transported breccia because one class of fragment is allochtonous relatively to the enclosing rock.
These observations are suggestive of a fluidization process during which suspended rocks fragments float in an escaping fluid phase, the whole mass behaving as a pulp (Branquet et al., 1999). The fluidization induces elutriation within the breccia. The ballistic behaviour of the fragments during the transport yields to impact brecciation resulting in abundant rock flour matrix. Injection of this slury fluid in faults and fractures attests to the synchronism between deformation and mineralizing fluid flows. Thus, the breccia can be considered as a major channelway for hydrothermal fluid during the mineralizing process. Multistage brecciation is observed and appears to correspond to successive faulting-fluid flow pulses. The opening of dilatant sites during compressive tectonics and the combination of fluidization and hydraulic fracturing suggests that the faulting-fluid flow pulses may be related to successive build-up and drop of the fluid pressure.
Branquet Y, Cheilletz A, Giuliani G, Laumonier Band Blanco O, Fluidized hydrothermal breccia in dilatant faults during thrusting: the Colombian emerald deposits case. In McCaffrey K. (ed) Geological Society of London, Special Publication, in press, (1999).
Index of EUG 10 Volume
Further EUG 10 Information
Index of the Journal of Conference Abstracts
Cambridge Publications Home Page
Last Updated on Wednesday, March 17, 1999.
© 1997 Cambridge Publications