Performance evaluation and prediction of activated carbon for VOCs via experiments and LFER methods

Due to the diversity and fluidity of volatile organic compounds (VOCs), it is difficult to conduct experimental studies on activated carbon (AC) used to capture VOCs in industry. Therefore, predicting the adsorption capacity of AC for VOCs is of great significance for evaluating the adsorption performance of AC and developing its promising applications. Herein, based on the experimental data of dynamic adsorption, the poly parameter linear free energy relationship (pp-LFER) model was developed to quantify the interactions between VOCs and AC. Through the dynamic adsorption curves of 25 VOCs on a fixed bed packed with the AC, the partition coefficient (PC) and adsorption capacity (q) of AC for VOCs were calculated, and the pp-LFER equations used to predict them were established. For the PC values, the pp-LFER model consists of the excess molar refractive index (E), molar polarizability (S), hydrogen-bond acidity (A), and logarithmic hexadecane-air partition coefficient (L), and the R2 of it was 0.86. For the ln q values, the developed model can predict it in R2 of 0.83, and the increased adsorption capacity was predominantly attributed to the dispersive interaction (lL). Finally, the internal and external validation results confirmed that the developed models have reliability, robustness, and predictability.