Formulation of a carbon sink binder through multi-objective optimization using response surface methodology

• Cement-free binder formulated from iron-rich industrial waste with >75 % upcycled content. • CO₂ curing and oxalic acid activation optimize iron carbonate formation for high strength. • Achieves 62.8 MPa compressive strength with multi-objective RSM optimization. • Statistical model with R² > 0.98 enables reliable strength prediction across curing regimes. • Supports circular economy goals with reduced embodied carbon in construction materials. This study presents the development and multi-objective optimization of a cement-free, carbon-sequestering binder system formulated entirely without Ordinary Portland Cement. The binder integrates iron-rich industrial waste, fly ash, metakaolin, and limestone, activated through oxalic acid to promote iron carbonate formation during CO₂ curing. Response Surface Methodology was employed to model and optimize the combined effects of oxalic acid dosage, CO₂ curing pressure, CO₂ and air curing durations, water-to-binder ratio, and specimen geometry on compressive strength. The statistical model demonstrated high predictive reliability R² = 0.9847; predicted R² = 0.949 with a desirability score of 1.000. An optimized formulation comprising 2 % oxalic acid, 3 bar CO₂ curing pressure, 14 days of CO₂ curing, 5 days of air curing, and a water-to-binder ratio of 0.17 achieved an experimental compressive strength of 62.8 MPa with only 3.41 % absolute error from the predicted value. This strength exceeds typical neat cement paste ranges 25–35 MPa, highlighting the system's potential as a viable cement paste substitute. Microstructural analyses XRD, FTIR, FESEM confirmed the formation of siderite, calcite, goethite, and dense low-porosity matrices, while TGA-DTG validated CO₂ uptake via carbonate formation. Over 75 % of the binder consists of upcycled industrial waste, supporting circular economy goals and significantly reducing embodied carbon. The generalized regression model enables predictive strength estimation across curing regimes and mix designs, offering a reproducible, scalable approach for developing high-performance, low-carbon construction materials.