H2S and SO2 toxic gases removal using date palm-tree branches based activated carbon: experimental findings and machine learning (ML) modeling

Utilization of carbonaceous waste material as precursors for production of activated carbon (AC) is desired. Hence, this study made use of date palm tree branches (DPB)-a major source of waste-in Saudi Arabia to produce AC that was used for treatment of air contaminated with H2S and SO2 gases. A dynamic adsorption setting was used to explore the uptake of H2S and SO2 onto DPB AC column with the aid of state-of-the-art bench-top gas analyzers. Several process parameters were varied to understand their effect on the breakthrough and exhaustion times of the gases under study. Response surface methodology (RSM) technique was used to further explore the relationship between the variables and machine learning (ML) techniques were used to create models for prediction of breakthrough and exhaustion times. Textural properties of DPB AC were explored using characterization techniques. AC produced from DPB had high BET specific surface area (800.87 m2/g) and a flaky morphology along with several oxygen-containing surface functional groups. The breakthrough profiles of both H2S and SO2 exhibited similar sigmoidal pattern. However, the broader breakthrough pattern of SO2 indicated that DPB has more affinity for SO2. The RSM study indicate that an increase in gas flow and gas initial concentration led to a shortening od both breakthrough and exhaustion time for both gases. However, when the column depth as increased, the breakthrough and exhaustion times lengthened for both gases. Furthermore, machine learning (ML) modeling results revealed that the artificial neural network (ANN) model was predicted H2S/SO2 breakthrough and exhaustion times with greater efficiency relative to other ML models. Our study revealed that AC produced from DPB can be used for treatment of air contaminated with H2S and SO2 gases and mechanistic modelling techniques can aid in exploring the adsorptive removal of harmful gases. It is recommended that a larger sample size should be utilized when conducting mechanistic modelling in future studies and that effective means for desorbing and regenerated contaminated AC should also be explored.