Investigations into moisture damage resistance of Nano-Modified Asphalt mixtures using granitic aggregates and 50/70 Penetration Grade Binder

Abstract

Moisture-induced damage is a major concern in asphalt mixtures because water infiltration weakens the bitumen–aggregate bond and reduces mixture strength and durability. This study investigated the influence of nano-modification on the mechanical and moisture-resistance properties of hot-mix asphalt (HMA) prepared using 50/70 penetration-grade bitumen and crushed granitic aggregates from three quarry sources in Northern Uganda. Nano-bentonite, nano-calcium carbonate (nano-CaCO₃), and liquid ZycoTherm were evaluated at 2%, 4%, and 6%; 3%, 6%, and 9%; and 0.1%, 0.2%, and 0.3%, respectively, by weight of bitumen. An unmodified mixture served as the control. The experimental program evaluated bulk density, Marshall Stability, Marshall Flow, Marshall Quotient (MQ), air voids, Voids in Mineral Aggregates (VMA), Voids Filled with Asphalt (VFA), indirect tensile strength (ITS), and Tensile Strength Ratio (TSR). Moisture susceptibility was assessed by comparing conditioned and unconditioned specimens, while analysis of variance (ANOVA) was used to determine the significance of the investigated factors. The results showed that nano-modification generally improved the mechanical and moisture-resistance properties of the HMA mixtures, although responses varied with nanomaterial type, dosage, and aggregate source. The most favorable overall responses within the investigated ranges were obtained at 4% nano-bentonite, 6% nano-CaCO₃, and 0.2% liquid ZycoTherm. Liquid ZycoTherm produced the most favorable TSR response, while nano-bentonite and nano-CaCO₃ generally produced stronger mechanical responses. Aggregate source significantly affected TSR but not Marshall Stability or MQ. ANOVA showed that factors with p < 0.05 were statistically significant. However, statistical significance did not imply high explanatory power. The R² values were 0.397 for Marshall Stability, 0.236 for MQ, and 0.229 for TSR, indicating that the models explained 39.7%, 23.6%, and 22.9% of the observed variability, respectively. The relatively low R² values, particularly for MQ and TSR, indicate limited explanatory power and support further investigation using additional material, mixture, conditioning, and performance-related variables.