Carbon adsorption on waste biomass of passion fruit peel: A promising machine learning model for CO2 capture

The alarming increase in the concentration of carbon dioxide (CO2) in the atmosphere, mainly due to human emissions, represents a significant threat to life. In this context, carbon capture and storage (CCS) technologies have emerged as promising solutions, such as adsorption on carbonaceous materials, standing out as a prominent approach. This study aims to quantify the maximum CO2 capture in the laboratory scale using functionalized activated carbon by passion fruit peel biomass (FACPFP) and to develop a simple and improved machine learning model to predict the capture of this greenhouse gas. FACPFP was successfully prepared through chemical activation with K2C2O4 and doping with ethylenediamine (EDA) at 700 °C and 1 h. The samples were thoroughly characterized by thermogravimetric analysis (TGA), scanning electron microscopy (SEM) with energy dispersive X-ray detector (EDX), Fourier transform infrared spectroscopy (FTIR) and X-ray photoelectron spectroscopy (XPS). CO2 sorption was assessed using functional density theory (DFT). For predictive model, multiple linear regression with cross-validation was used. Under CO2 atmosphere conditions, the textural parameters allowed to see the probable presence of ultra-micropores, the BET surface area, the total pore and micropore volume were 105 m²/g, 0.03 cm³ /g and 0.06 cm³ /g, respectively. The maximum CO2 adsorption capacity in the FACPFP reached about 2.2 mmol/g at 0 °C and 1 bar. The predictive model demonstrated an improvement of CO2 adsorption precision, raising it from 53% to 61% with cross-validation. This study also aims to stimulate future investigations in the area of CO2 capture, due to the extreme relevance of this topic.