The surface roughness of soil grains affects the mechanical behaviour of soils, but the characterization of real soil grain roughness is still limited in both quantity and quality. A new method is proposed, which applies the power spectral density (PSD), typically used in tribology, to optical interferometry measurements of soil grain surfaces. The method was adapted to characterize the roughness of soil grains separately from their shape, allowing the scale of the roughness to be determined in the form of a wavevector range. The surface roughness can be characterized by a roughness value and a fractal dimension, determined based on the stochastic formation process of the surface. When combined with other parameters, the fractal dimension provides additional information about the surface structure and roughness to the value of roughness alone. Three grain sizes of a quarzitic sand were tested. The parameters determined from the PSD analysis were input directly into a Weierstrass–Mandelbrot function to reconstruct successfully a fractal surface.

The pore size, shape and orientation of an illite-dominant clay were mapped during one-dimensional compression, using mercury intrusion porosimetry (MIP), scanning electron microscopy and gas adsorption. The total porosity was found to spread over the three International Union of Pure and Applied Chemistry classes of pores sizes: micropores (below 2 nm), mesopores (2–50 nm) and macropores (above 50 nm), and all three pore classes were observed during the compression. The clay structure is aggregated, with visible inter-aggregate pores (about 80% of the total porosity), and the remaining intra-aggregate pores of size approximately equal to the thickness of illite platelets (50–100 layers). During compression the largest pores first collapsed, followed by a progressive collapse, in an orderly manner, of smaller and smaller pores. MIP data suggest that the macroscopic deformation mainly translates at the pore scale into changes of inter-aggregate porosity, while intra-aggregate pores spread over the micro- to mesopore size range. Gas adsorption tests show that the volume of intra-aggregate pores decreases with loading, probably due to rearrangement of particles composing the aggregates, while the specific surface area reduces. Examination of the pores’ orientation on both vertical and horizontal planes confirms a preferential orientation of pores normal to the loading direction, with a gradual flattening of the pores.

In this paper, the evolution of nanoscale pores, covering inter-particle pores and inter-layer pores, in illite-dominant clay during consolidation is monitored using small angle X-ray scattering (SAXS) and nitrogen gas adsorption (N₂GA) techniques. No obvious change observed in the characteristic peaks of SAXS intensity curves during consolidation suggests that the intra-particle structure of the clay, including interlayer spacings, is not affected by mechanical loading, at least up to 4 MPa. The N₂GA test results show that the volume of inter-particle pores inside the aggregates does decrease gradually as the compression proceeds, which is accompanied by a gradual reduction in specific surface area, probably due to the rearrangement of the particles composing the aggregates. The inter-particle pores are compressed as a whole during consolidation instead of the progressive collapse in an ordered manner, from the larger to the smaller. By comparing the pore-size distributions of illite-dominant clay obtained by MIP (mercury intrusion porosimetry) and N₂GA techniques, it is found that the shapes of the two distributions in the common measurement range are obviously not matched, essentially due to the sequential nature of the drying and wetting processes. While filling the research gap in the evolution of intra-aggregate pores during consolidation, this study also shows that the N₂GA technique and SAXS measurement used in conjunction with each other appear as a powerful approach for clay nano-pores identification.

The surfaces of soil grains are not perfectly smooth, especially examined at small scale. In geotechnical engineering, surface roughness has been found to be able to influence the inter-particle friction angle at micro scale and small-strain stiffness at macro scale. However, the quantity and quality of the studies on surface roughness of natural soils are still limited. In this study, the evolution of surface roughness of natural sand grains with increasing normal load was investigated by a single-particle compression apparatus. Thirty Leighton Buzzard sand (LBS) grains coarser than 2·36 mm were tested, and the surface roughness was measured before and after compression by an optical interferometer. The deformations of the asperities and of the bulk of the sand grains in the vicinity of the contact were mapped. Three stages were identified as the normal load increased: (a) plastic deformation of the asperities; (b) asperities and bulk plastic deformation; and (c) bulk only plastic deformation. At very small normal load, only the asperities were found to deform plastically, and the surface roughness of the sand grains decreases due to the flattening of the asperities. Within this regime, the load–displacement relationship of LBS grains under compression could be simulated by the modified Hertz model, which takes surface roughness into consideration. With increasing normal load, the bulk of the sand grains began to yield near the contact. The geometry of the surfaces of LBS grains in contact with the loading platen is the main factor that influences the plastic deformation of the bulk. Differently from the plastic deformation of the asperities, the plastic deformation of the bulk could both smoothen and roughen the surfaces. When plastic deformation of the bulk occurred, both Hertz and modified Hertz theory could not predict the load and displacement relationship of sand grains. Through analysing the cumulative distributions of surface roughness of 30 LBS grains at different normal loads by the Weibull function, the surface roughness was found to decrease dramatically with increasing normal load at first and then tended to be constant.

This paper attempts to bridge the gap between the well-known characteristics of strain rate-independent elastic stiffness and of the time-dependent behaviour of clays at large strains. Effects of acceleration, strain rate and ageing are examined in the very small to small strain region, using results from tests on reconstituted London Clay and kaolin. Static measurements of the shear modulus at small strains, using local instrumentation, indicate that the shear modulus is rate-independent when the soil is subjected to elastic deformations only. Most often, in practice, soil specimens are left to rest between the compression and shearing stages, until deformations associated with creep become negligible. The testing strain rate imposed upon shearing induces a temporary acceleration of strains, which results in very high measured stiffnesses. Test data show, however, that when the shearing axial strain rate is of the same order of magnitude as that measured at the end of the compression/creep stage, the remaining plastic creep strains influence the measured stiffness, which is then measured to be comparatively low. Test data also seem to show that temporary strain acceleration may hide effects of recent stress path rotation. Additionally, bender element tests were performed during isotropic compression tests with changes in stress rate and creep stages. The data indicate that the strain rate affects the value of the dynamic shear modulus of the soil. It is suggested that the increase in dynamic shear modulus during short-term creep is caused by mechanisms linked to the gradual decrease in strain rate during creep, but further investigation would be needed to clarify which physical mechanisms. A new method to account for strain rate in evaluating the dynamic shear modulus of normally consolidated reconstituted clays is finally proposed.