Effects of roundness and relative particle size on extreme void ratios of binary coarse-grained soil mixtures

The packing behavior of binary mixtures is critical and has been extensively studied. However, previous research rarely included quantitative roundness and aspect ratio analysis of both the finer and coarser fractions. Using rounded and angular sand particles of three uniform sizes, we prepared 60 mixtures and measured their maximum and minimum void ratios. We incorporated image-based roundness and aspect ratio analysis to better explain the packing behavior. The void ratio decreases to a minimum value and then increases as the smaller size fraction is gradually added to the larger particles. The threshold for achieving this lowest void ratio is about 40% of the volume fraction of finer particles. However, this threshold drops below 40% when the finer particles are less rounded than the larger ones. Conversely, when the finer particles are substantially more rounded than the larger particles, the threshold exceeds 40%. A set of relationships linking differences in measured particle shape descriptors between larger and smaller particles and the packing properties of the binary mixture is proposed. In addition, we developed linear and nonlinear predictive models to estimate the minimum void ratio of binary mixtures by incorporating the roundness-based threshold fine content and the size ratio-based void ratio reduction from the upper bound packing state. The nonlinear model better captures the measured packing response and provides higher accuracy (R² = 0.955) than the linear model (R² = 0.850). This research approach (e.g., testing binary size-shape mixtures) has great potential to deepen our understanding of soil behavior in the future.