It is well recognized today that the complex behavior of geological materials at the macro scale is closely related to or controlled by the grain-scale interactions. Additionally, the advancements in the application of the discrete-based numerical simulations in geomechanics have necessitated more systematic micro-mechanical studies to be carried out in the laboratory to provide reliable inputs for the development of accurate and sophisticated models. The natural soils/rocks encountered in engineering projects are always subjected to complicated interface conditions due to weathering, erosion, transportation, climate, and the intervention of human activities. This can be reflected as the existence of non-uniform bonding, discrete coating, heterogeneous surface fabric and varying contact types as well as the presence of fluid with varying viscosities at the contacts. However, a large literature gap needs to be filled in the fields of investigating the micromechanical behavior of geological materials with natural types of bonding and coating as well as the soil/rock-fluid interactions.

In this study, micromechanical tests including normal loading and shearing tests are conducted on various types of materials with a custom-built inter-particle loading apparatus. Emphasis is given on the assessment of the normal/tangential force-displacement relationship, interface coefficient of friction, microslip regime, contact stiffness and Young’s modulus. Multiple factors such as the magnitude of normal load, number of loading cycles, loading path and the presence of fluid are systematically evaluated. The materials are mainly categorized into four groups, namely (i) calcium carbonate-bonded sand grains, (ii) iron oxides-coated sand grains, (iii) raw/polished tailing particles, and (iv) analog proppant (ceramic)-mudrock (compressed kaolinite), proppant-proppant, mudrock-mudrock interfaces. The experimental results are compared with those of some representative geomaterials in the literature and also theoretical models are applied to depict the experimental force-displacement curves. The major findings and contributions of this study are concluded as follows:

i. This study proposed new methods to develop weak calcium carbonate bond at the interfaces of sand particles with two types of cementation and discrete iron oxides coating on the surfaces of sand particles.

ii. Compared to the engineered materials and standard quartz sand particles with relatively smooth and consistent surface characteristics, the granular materials with rough surfaces, ununiform surface fabric and coating of silt- to clay-sized microparticles (e.g., iron oxides-coated sands and iron tailing particles) in general had higher frictional values and lower normal/tangential stiffness, which showed a certain degree of scatter. Thus, it is recommended to apply a probabilistic selection of input parameters in the numerical simulations.

iii. The preloading and preshearing history (abrasion) had a considerable effect on the change of the tribological behavior of materials with relatively high surface roughness and low hardness (e.g., iron oxides-coated grain, iron tailing particles as well as kaolinite blocks). This was attributed to the occurrence of surface damage or asperity breakage during shearing, which would even take place at a lower magnitude of normal load.

iv. In terms of the soil/rock-fluid interaction, from the grain-scale test results of iron tailing particles, it is revealed that the presence of the viscous fluid at the interface showed a promising application in the ground improvement and stabilization of tailings dams. However, appropriate concentrations/viscosities need to be considered. For kaolinite blocks, the results showed that the presence of water in the repeating shearing tests significantly decreased the interface friction of the kaolinite samples and both water and biopolymer-based fluid had a softening effect by analyzing the initial tangential stiffness.

v. The results of theoretical model fitting demonstrated that in the normal direction, the Hertz model was suitable to depict the normal loading-displacement response of quartz-based hard grains with a relatively smooth texture, while the Yimsiri-Soga model was more proper for coated grains or natural grains with impurities on their surfaces which had a notable initial soft response. In the tangential direction, the Mindlin-Deresiewicz model could achieve a good match with the experimental shearing load-displacement response by applying the experimental initial tangential stiffness and adjusting the model parameter which had a positive relationship with the normalized microslip displacement.