Carbon dioxide (CO₂) is a major greenhouse gas contributing significantly to global warming and climate change, posing a serious threat to environmental and human systems worldwide. As anthropogenic emissions continue to rise, the development of efficient and sustainable CO₂ capture technologies has become a global priority. This study focuses on the synthesis of activated carbon from Pterospermum acerifolium fruits using three chemical activators&mdashH₃PO₄, KOH, and K₂CO₃&mdashfor CO₂ adsorption applications. The process parameters, including impregnation ratio, pyrolysis temperature, and activation time, were optimized using Response Surface Methodology (RSM) based on a Box-Behnken Design (BBD). The optimized conditions yielded iodine numbers of 1197.05 mg/g (H₃PO₄), 1124.12 mg/g (KOH), and 1252.10 mg/g (K₂CO₃), with respective yields of 35%, 11%, and 12.68%. Comprehensive characterization was performed using TGA, CHNS, FTIR, Raman, FESEM, TEM, and BET analyses. To improve handling properties, KOH-activated carbon was further converted into granules using potato starch and polyvinyl alcohol as binders. The maximum CO₂ uptake was recorded at 303 K: 18.53 mmol/g for KOH powder and 12.18 mmol/g for starch-based granules. Adsorption data best fit Freundlich, Sips, and Dubinin&ndashAstakhov isotherm models, indicating heterogeneous surface adsorption. Kinetic analysis followed a pseudo-first-order model, suggesting a physisorption mechanism, while thermodynamic studies confirmed an exothermic and spontaneous process. The adsorbents retained ~99% adsorption capacity over five cycles. An Artificial Neural Network (ANN) model effectively predicted CO₂ uptake under varying conditions, demonstrating high correlation with experimental results.

Keywords: Activated carbon CO₂ adsorption capacity Box-Behnken design Artificial Neural Network Isotherm models Kinetic models