Response surface modelling and non-destructive evaluation of 100% reclaimed asphalt pavement-based roller-compacted geopolymer concrete

The construction industry requires sustainable alternatives to reduce the environmental impacts associated with cement production and the increasing accumulation of reclaimed asphalt pavement (RAP). This study develops a low-carbon roller-compacted geopolymer concrete (RCGC) incorporating 100% RAP, containing approximately 4.5% residual bitumen, as a full replacement for natural aggregates. Ordinary Portland cement was completely eliminated by using a geopolymer binder based on metakaolin (MK) and granulated blast-furnace slag (GBFS), activated with sodium hydroxide, sodium silicate, and hydrated lime. A response surface methodology based on a central composite design was employed to optimise mixture proportions and evaluate the effects of NaOH molarity, HL/SH ratio, and GBFS content on compressive strength and non-destructive testing responses, including ultrasonic pulse velocity and Schmidt rebound hammer values. Strong relationships were observed between NDT parameters and compressive strength, with direct UPV providing the highest predictive accuracy (R² = 0.899). XRD and FTIR analyses confirmed the formation of N-A-S-H and C-A-S-H gels, explaining matrix densification and strength development. The optimum mixture containing 40% GBFS and no hydrated lime achieved a 28-day compressive strength of 16.8 MPa, consistent with the performance requirements for low-traffic pavements according to the U.S. Army and Air Force pavement design guidelines. Predictive models (R² = 0.87−0.95) were also developed to estimate compressive strength from NDT measurements and mix design variables.