Grain size distribution does not affect the residual shear strength of granular materials: An experimental proof

Granular materials are used in several fields and in a wide variety of processes. An important feature of these materials is the diversity of grain sizes, commonly referred to as polydispersity. When granular materials are sheared, they exhibit a predominant small elastic range. Then, the material yields, with or without a peak shear strength depending on the initial density. Finally, the material reaches a stationary state, in which it deforms at a constant shear stress, which can be linked to the residual friction angle ϕ_{r}. However, the role of polydispersity on the shear strength of granular materials is still a matter of debate. In particular, a series of investigations have proved, using numerical simulations, that ϕ_{r} is independent of polydispersity. This counterintuitive observation remains elusive to experimentalists, and especially for some technical communities that use ϕ_{r} as a design parameter (e.g., the soil mechanics community). In this Letter, we studied experimentally the effects of polydispersity on ϕ_{r}. In order to do so, we built samples of ceramic beads and then sheared these samples in a triaxial apparatus. We varied polydispersity, building monodisperse, bidisperse, and polydisperse granular samples; this allowed us to study the effects of grain size, size span, and grain size distribution on ϕ_{r}. We find that ϕ_{r} is indeed independent of polydispersity, confirming the previous findings achieved through numerical simulations. Our work fairly closes the gap of knowledge between experiments and simulations.