When submitted to gentle mechanical taps a granular packing slowly compacts until it reaches a stationary state that depends on the tap characteristics. The properties of such stationary states are experimentally investigated. The influence of the initial state, taps properties and tapping protocol are studied. The compactivity of the packings is determinated. Our results strongly support the idea that the stationary states are genuine thermodynamic states.

We report on experiments to measure the temporal and spatial evolution of packing arrangements of anisotropic and weakly confined granular material, using high-resolution γ-ray adsorption. In these experiments, the particle configurations start from an initially disordered, low-packing-fraction state and under vertical solicitations evolve to a dense state. We find that the packing fraction evolution is slowed by the grain anisotropy but, as for spherically shaped grains, can be well fitted by a stretched exponential. For a given type of grains, the characteristic times of relaxation and of convection are found to be of the same order of magnitude. On the contrary, compaction mechanisms in the media strongly depend on the grain anisotropy.

This Letter reports experimental and numerical results on particle dynamics in an out-of-equilibrium granular medium. We observed two distinct types of grain motion: the well known cage motion, during which a grain is always surrounded by the same neighbors, and low probability ``jumps,'' during which a grain moves significantly more relative to the others. These observations are similar to the results obtained for other out-of-equilibrium systems (glasses, colloidal systems, etc.). Although such jumps are extremely rare, by inhibiting them in numerical simulations we demonstrate that they play a significant role in the relaxation of out-of-equilibrium systems.

The carbon and water contents and the corresponding isotopic compositions have been measured on a set of glassy samples collected by dives on the Mid-Atlantic Ridge (MAR) near 34°50′N where volcanoclastic deposits are present. The volatile phases have been extracted by crushing under vacuum and stepwise heating.The δ18O of the glasses shows that N-MORB are depleted in 18O (down to 5.14‰) whereas all the other lavas fall in the mantle range, 5.4–5.8‰. These data preclude strong interaction between seawater and magmas before eruption.The posteruptive contents of dissolved water and carbon measured on N-, T-, E-MORB and alkali basalts range from 1125 to 5253 ppm and from 20 to 119 ppm, respectively. The vesicle gas is dominated by CO2 in N- and T-MORB. Water is present as small amounts in E-MORB vesicles and represents up to 17 vol.% of the total gas in alkali-basalt vesicles. The preeruptive water and carbon concentrations of these magmas have been reconstructed. They range from 1130 to 8497 ppm and from 343 to 15677 ppm, respectively. The isotopic data demonstrate that seawater contamination is detectable only when most of the vesicles have been broken during eruption and quenching. Otherwise, all the δ13C and δD values obtained in vesicles or dissolved in glasses, fall in a typical mantle range, −4‰ to −7‰ and −50‰ to −88‰, respectively.Most of the lavas have degassed under conditions close to chemical and isotopic equilibrium, except enriched magmas for which kinetic processes were predominant during the last eruptive event and precluded the attainment of chemical and isotopic equilibrium.Taking into account the effects of crystal fractionation, the primitive magmas had initial concentrations of water ranging from 1100 to 8000 ppm. Because of the important CO2 degassing (80–99.9%), a reconstruction, based on the two-steps degassing model of Pineau and Javoy [Earth. Planet. Sci. Lett. (1994) 123] was made. As the enriched character increases, carbon saturation is reached at increasing depths and pressures: between 2.6 and 6.6 kbars for N- and T-MORB, between 6 and 12.8 kbars for E-MORB and between 15 and 27.5 kbars for alkali basalts.The high water and carbon contents found in the alkali-rich lavas are likely linked to an enriched fertile mantle source beneath this ridge segment. That source provides the additional supply of incompatible elements needed to produce (1) The E-MORB by magma mixing and (2) The explosive eruption characteristics.

We report experimental results on granular compaction under consecutive vertical taps. The evolution of the mean volume fraction presents a slow densificaiion until a final steady state and follows a KWW law. This result does not depend on the shape of the grains. The link between compaction and convection is also studied. Finally, a local analysis of the packing is discussed.

In this work, a digital imaging technique is used to study the superficial fluctuations observed when a granular packing is slowly driven to the threshold of instability. The experimental results show the presence of three types of events. Small superficial rearrangements of grains are observed during all the experiments. They present a power-law behavior although the system is not in a critical state as predicted by self-organized criticality models. In thick granular piles, large rearrangements are detected at regular angular intervals. They are related to the threshold of instability of the contact network that relaxes to stable configurations producing internal rearrangements of the grains. Finally, an avalanche is triggered when the superficial beads that are set in motion acquire enough momentum to destabilize grains from layers below.

A granular medium submitted to vertical tapping reveals simultaneously compaction and convection. The two phenomena are directly coupled and their dynamics can be quantified by a characteristic compaction time and by an estimation of the convective downhill speed along the wall. A remarkable change of behavior is observed around the liftoff acceleration threshold of the whole packing, with a drastic slowing down of both dynamics below this threshold. Above it, a collective shock wave densifies the packing at each tap, whereas, below it, cumulative localized rearrangements will compact the entire system in the long time range.

The collision of an impacting bead on a 2D bead packing is experimentally investigated. It is found that the energy dissipated in the collision is strongly dependent on the density of inter-granular contacts in the packing. We show indeed that the energy released by the packing after the impact is increased when the density of inter-grain contacts is lowered. This increase manifests itself by means of an important augmentation of the number of grains ejected from the packing. In addition, we propose a phenomenological description of the propagation of the shock wave (induced by the impacting bead) through the packing. This description is inspired from percolation models and provides a plausible explanation for the increase of the energy released by a packing with a low density of inter-grain contacts.

A packing of equal glass beads is placed on a box with a rough bottom. The box is slowly inclined till an avalanche begins at a critical angle. The avalanche dynamics decreases the surface slope until a second critical angle is reached. In previous works [1–3], the stability of the packing was found to be affected by the number of layers, the packing length and the surrounding humidity.

We report here an experimental study on surface granular flows. The granular packing is made of monosize glass beads initially poured in a box with controlled humidity. We have first studied the angles before and after the avalanche, and the mass displaced out of the box during the avalanche as a function of the number of layers of the packing. For small packings, up to approximately 10 layers, the stability of the system is significantly affected by the bottom rough surface. On the contrary, for thicker systems, the critical angles are unchanged with variations of the number of layers. In a second experiment, in the same apparatus in 3d and also in 2d, we have studied flow of a mass of grain on a surface near the situation where no accretion nor erosion exist (at the neutral angle). The influence of the geometry of the system is investigated.

Mantle temperature variation and plate spreading rate variation have been considered to be the two fundamental variables that determine the extent of mantle melting and ocean crust production. Along the length of a ∼200 km portion of the Mid-Atlantic Ridge (MAR) between the Oceanographer (35°N) and Hayes (33°N) transforms, the mantle potential temperature is the same, the plate spreading rate is the same, but the extent of mantle melting and crustal production vary drastically. In addition to the typical crustal thickness variation on ridge segment scales at the MAR, i.e. thicker at segment centers and thinner at segment ends, there exist between-segment differences. For example, the ∼90 km long segment OH-1 is magmatically robust with a central topographic high, thick crust, and a large negative gravity anomaly whereas the ∼45 km long segment OH-3 is magmatically starved with a deep rift valley, thin crust and a weak negative gravity anomaly. We demonstrate that the observed differences in the extent of mantle melting, melt production and crustal mass between segments OH-1 and OH-3 are ultimately controlled by their fertile mantle source compositional difference as reflected by the lava compositional differences between the two segments: >70% of OH-1 samples studied (N=57) are enriched MORB with [La/Sm]N>1, but >85% of OH-3 samples studied (N=42) are depleted MORB with [La/Sm]N<1. Calculations show that the mean OH-1 source is more enriched in incompatible elements, total alkalis (∼0.36 wt% Na2O and ∼0.09% K2O) and H2O content (∼280 ppm) than the mean OH-3 source, which is depleted of incompatible elements, total alkalis (<0.17% Na2O and <0.01% K2O) and H2O content (∼70 ppm). These fertile compositional differences result in significantly reduced solidus temperature of OH-1 source over that of OH-3 source, and allows melting to begin at a significantly greater depth beneath OH-1 (∼90 km) than beneath OH-3 (<60 km), leading to a taller melting column, higher degrees of decompression melting, greater melt production, thus thicker crust and more negative gravity anomaly at OH-1 than at OH-3. We emphasize that fertile mantle source compositional variation is as important as mantle temperature variation and plate spreading rate variation in governing the extent of mantle melting, crustal production, and MORB chemistry. The buoyancy-driven focused mantle upwelling model better explains the observations than the subcrustal melt migration model. Future mantle flow models that consider the effect of fertile mantle compositional variation are expected to succeed in producing along-axis wavelengths of buoyant flow comparable to the observed size of ridge segments at the MAR. We propose that the size and fertility of the enriched mantle heterogeneities may actually control the initiation and evolution of ridge segments bounded by non-rigid discontinuities at slow-spreading ridges.

We study the diffusion process in a granular gas. We first show that for finite size systems the choice of boundary conditions is of crucial importance. With periodic boundary conditions, the coefficient of diffusion is found to depend on the system size and does not saturate for large systems, which is of course not physical. The problem is solved by using reflecting boundaries. In that case, we find good agreement between numerical results and the Langevin theory. We also study the influence of an external random force on the diffusion process for a forced system. In particular, we analyze differences in the mean square velocity and displacement between the elastic and inelastic case.

A simple numerical model is used to simulate the effect of vertical taps on a packing of monodisperse hard spheres. Our results are in good agreement with an experimental work done in Chicago and with other previous models, especially concerning the dynamics of the compaction, the influence of the excitation strength on the compaction efficiency, and some aging effects. The principal asset of the model is that it allows a local analysis of the packings. Vertical and transverse density profiles are used, as well as size and volume distributions of the pores. An interesting result concerns the appearance of a vertical gradient in the density profiles during compaction. Furthermore, the volume distribution of the pores suggests that the smallest pores, ranging in size between tetrahedral and octahedral sites, are not strongly affected by the tapping process, in contrast to the largest pores which are more sensitive to the compaction of the packing.

We present several numerical results on granular mixtures. In particular, we examine the efficiency of diffusion as a mixing mechanism in these systems. The collisions are inelastic and to compensate the energy loss, we thermalize the grains by adding a random force. Starting with a segregated system, we show that uniform agitation (heating) leads to a uniform mixture of grains of different sizes. We define a characteristic mixing time tau(mix), and study theoretically and numerically its dependence on other parameters like the density. We examine a model for bidisperse systems for which we can calculate some physical quantities. We also examine the effect of a temperature gradient and demonstrate the appearance of an expected segregation.

We report an experimental study on the collision of a bead on a two-dimensional hexagonal granular packing. This collision process is of crucial importance in aeolian transport of grains. We have investigated the kinematic properties of the incident bead before and after the collision, and the resulting deformation of the packing. A typical collision is characterized by the rebound of the impacting bead and the ejection of a few beads of the packing. We have shown that the properties of the rebound bead depend weakly on the impact speed and that the rebound process involves only a few bead layers of the packing. On the contrary, the ejection mechanism depends strongly on the impact speed. In particular, it is found that the number of ejected grains increases with the impact speed whereas the most likely value of their energy is practically independent of the impact speed. Furthermore, we have given evidences that the ejection process involves a great number of packing layers and therefore is extremely sensitive to the height of the packing. For small packing heights, one observes additional ejected grains which can be interpreted as being produced by the reflection of the shock wave on the bottom of the pile.

We analyze the influence of boundary conditions on numerical simulations of the diffusive properties of a two-dimensional granular gas. We show in particular that periodic boundary conditions introduce unphysical correlations in time that cause the coefficient of diffusion to be strongly dependent on the system size. On the other hand, in large enough systems with hard walls at the boundaries, diffusion is found to be independent of the system size. We compare the results obtained in this case with Langevin theory for an elastic gas. Good agreement is found. We then calculate the relaxation time and the influence of the mass for a particle of radius R(s) in a sea of particles of radius R(b). As granular gases are dissipative, we also study the influence of an external random force on the diffusion process in a forced dissipative system. In particular, we analyze differences in the mean-square velocity and displacement between the elastic and inelastic cases.

We report an experiment on a granular packing: a box filled with glass beads is tilted very slowly up to the maximum angle of stability where a big avalanche is produced. During the build-up period many rearrangements occur on the free surface of the packing. Digital imaging was used to study these rearrangements. The probability distribution of sizes for the observed mass fluctuations follow a power-law behavior, which is the signature of self-organized criticality. However, this description breaks down in the limit of big rearrangements where inertia effects are not negligible.

Submersible observations and sampling were carried out in the rift valley of the Mid-Atlantic Ridge (MAR) near 34 degrees 40'N-35 degrees N. The 4-km-wide rift valley consists of a Neo Volcanic Zone (NVZ) (<1 km wide) bounded at the west by a Median Ridge (MR) (5 km wide and 20 km long) and at the east by the first scarps of the eastern wall. The MR and the eastern wall are characterized by Volcanic cones about 200-300 m height culminating at depths of 1500-1900 m which are made up of volcaniclastic deposits (pyroclasts and hyaloclasts) suggestive of explosive volcanism. Based on their surface morphology, degree of vesicularity, and composition, the erupted deposits are classified into four groups: (1) poorly vesicular (<15% vesicles) N- and T-MORBs (K/Ti <0.25, Na2O + K2O < 2.9%) consisting of sheet hows and pillows formed during fissure eruptions in the NVZ at 2000-2300 m depths; (2) vesicular (15-30% vesicles) E-MORBs (K/Ti = 0.25 - 0.45, Na2O + K2O > 2.8 - 3.2%) and alkali basalts (K/Ti = 0.45 - 0.70, Na2O + K2O > 3.3 - 4) made up mainly of pillows; (3) highly vesicular (>35% vesicles) pillow lava and pyroclastic (scoria-like) alkali basalts (K/Ti > 0.45 - 0.80, Na2O + K2O > 3 - 4%); and (4) hyaloclastites consisting of glassy shards of alkali basalt composition. The total water and carbon contents of the deposits increase with the incompatible element concentrations. The estimated initial H2O content for the N- and T-MORBs is less than 3500 ppm, whereas for the E-MORBs and alkali basalts the H2O content is near 4000 and 7000 ppm, respectively. While the H2O is mainly in the melt, the carbon is in the form of CO2 filling vesicles. The vesicles are formed from magma with an initial carbon content of 1000-3000 for the N- and T-MORBs, 3000-6500 ppm for the E-MORBs and higher than 1 wt% for the alkali basalts.The various lava types were derived from a heterogeneous mantle source composed of enriched and depleted components during sequential eruptions of N-, T- and E-MORBs and alkali basalts (K/Ti > 0.7). The amount of CO2 and H2O in equilibrium with the dissolved species present in the vesicles indicates that CO2 (XCO2 = 1 - 0.84) was the main exsolved compound responsible for bubble nucleation. The increase in the degree of vesicularity and pressure of the volatile phases is mainly due to the early exsolution of CO2 from an alkali melt. The exsolution of significant amounts of dissolved water occurred only for the alkali basalt a few hundred meters beneath the seafloor and contributed to late bubble expansion. This subsequent addition of magmatic water to the vesicles increased the gas pressure and triggered explosions. An alternative hypothesis for the explosive volcanism is based on field observations. During crater collapsed, seawater could have been trapped in fractured volcanic conduits and later sealed by hydrothermal fluid circulation and precipitation. In such an environment, this seawater will be heated and vaporized during renewed magmatic upwelling. Both scenarios give rise to fragmented debris (hyaloclasts and pyroclasts) and the explosive events create turbulent flows followed by differential gravity settling of the particles (shards versus lapilli) through the seawater. (C) 2000 Elsevier Science B.V. All rights reserved.

Mixing of granular materials is difficult because commonly used mixing methods can lead to undesired segregation. We are concerned here with diffusion as a mechanism allowing mixing.

We study drainage in a horizontally oriented rough fracture joint filled with glass beads. The shape and structure of the drained areas is the result of competition between two effects: (1) variations in the capillary thresholds necessary to be overcome in order to drain the pores and (2) the height variations due to the roughness of the fracture joint. These height variations have long range correlations due to the self-affine nature of the fracture. The capillary thresholds are uncorrelated. We tune the relative strength of these two effects by performing experiments in a centrifuge and thus changing the "strength of gravity." As gravity is increased, the structure of the drained areas change from that of invasion percolation to a structure composed of compact blobs linked together by threadlike links. We study both the geometry and the effect of trapping while changing acceleration of gravity from zero to 6g(0). At high centrifugal acceleration we further observe fragmentation, migration and coalescence of bubbles of fluid inside the drained areas.