Evaluation of pistachio shells as solid wastes to produce activated carbon for CO2 capture: Isotherm, response surface methodology (RSM) and artificial neural network (ANN) modeling

Abstract

The adsorbent material selected for this study was microporous activated carbon prepared by physical activation from pistachio shells (PS). X-ray diffraction (XRD), Fourier transform infrared spectrometer (FTIR), scanning electron microscopy (SEM), and analysis of N2 isotherms was used to determine the textural and structural character of this carbon material. The activated carbon from pistachio shells (AC-PS) indicted a specific surface area (Brunauer–Emmett–Teller specific area: SBET: 715.34 ​m2/g, and total pore volume: 0.346 ​cm3/g) and good sorption of CO2 (3.29 ​mmol/g at 25 ​°C and 5.682 ​bar). The effect of CO2 adsorption capacity as the response function was evaluated and optimized using response surface methodology (RSM) based on pressure and adsorption temperature as independent variables. A back-propagation training approach is used to create feed-forward multilayer perceptron (MLP) artificial neural network models to predict CO2 adsorption in this study. The experimental data of the adsorption experiments were analyzed and evaluated based on the isotherm and thermodynamic models. Adsorption data from the experiments followed the Hill isotherm model. Adsorption was an exothermic process, and there was mainly physical interaction. Minimum performance of 0.0011456 was determined by performing mean square error (MSE) validation at 22 epochs. The R2 values for the RSM and artificial neural network (ANN) models were 0.9957 and 0.999905, respectively. The results showed that the ANN and RSM models could be effectively used to estimate CO2 adsorption. According to these results, AC-PS would be a potential adsorbent for CO2 adsorption.