This paper presents a synthesis of the works performed by various teams from France, Italy and Canada around the question of second‐order work criterion. Because of the non‐associative character of geomaterials plastic strains, it is now recognized that a whole bifurcation domain exists in the stress space with various possible modes of failure. In a first part these failure modes are observed in lab experimental tests and in discrete element modelling. Then a theoretical study of second‐order work allows to establish a link with the kinetic energy, giving a basis to explain the transition from a prefailure (quasi)static regime to a postfailure dynamic regime. Eventually the main features of geomaterials failure are obtained by applying second‐order work criterion to five different constitutive rate‐independent models—three being phenomenological and two micromechanical. As a whole this paper tries to gather together all the elements for a proper understanding and use of second‐order work criterion in geomechanics. Copyright © 2010 John Wiley & Sons, Ltd.

A stress dilatancy model with embedded microstructural information, originally developed by the writers, is used to illustrate the pivotal importance of dilatancy and fabric on the behavior of sand. Fabric, as a second-order tensor, enters into the stress dilatancy equation obtained from a microscopic analysis of an ensemble of rigid particles. Model simulations of sand behavior are carried out in triaxial stress conditions along strain paths with varying degrees of controlled dilation (or compaction) including isochoric deformations as a particular case. Under particular strain paths and fabric conditions, it is shown that a relatively dense sand can succumb to instability or liquefaction under other than isochoric (undrained) conditions. This phenomenon is in accord with laboratory experiments in which dilation or compaction is controlled by modulating the amount of water flowing in or out of a sand specimen during shearing. Mixed drained–undrained loading paths are also simulated with particular refere...

Variations of the Kelvin equation [W. Thomson, Philos. Mag. $42,$ 448 (1871)] to describe the freezing point depression of water in capillaries exist in the literature. The differing equations, coupled with the uncertainty in input parameters, lead to various predictions. The difference between the predictions may become substantial when the capillary size decreases much below micron dimensions. An experiment was designed to investigate the predicted values using a customized directional solidification stage. The capillary freezing point depression for glass tubes with radii of $87\mu \mathrm{m}–3\mu \mathrm{m}$ was successfully measured. The image of the ice-water interface at equilibrium was also digitally captured and analyzed to examine the contact angle and the interface shape as well. Both are important for examining the hemispherical interface assumption that was exclusively used in the theoretical derivations. Finally, an equilibrium analysis of the thermodynamic system leads to a theoretical discussion of the problem. The effect of the temperature gradient on the interface shape is addressed, and an engineering criterion for the critical temperature gradient above which the effect must be considered for the interface shape calculation is derived.