Dynamic adsorption of phenolic compounds on activated carbon produced from pulp and paper mill sludge: experimental study and modeling by artificial neural network (ANN)

Abstract

A new low-cost activated carbon (AC) was produced from pulp and paper mill sludge through chemical activation by zinc chloride. Different characterization analyses were carried out on developed AC; these demonstrated a carbon material with highly porous structure. Dynamic adsorption of phenolic compounds (i.e. phenol, 2-chlorophenol, and 4-nitrophenol) from simulated aqueous solution was investigated in a fixed-bed adsorption column using the prepared AC. Dynamic behavior of the adsorption column was assessed in terms of breakthrough curves obtained at different key operating conditions such as bed height, feed flow rate, inlet concentration, and temperature. Sharp breakthrough curves were observed at high-feed flow rate, high-inlet concentration, high temperature, and low-bed height which show the correct dynamic behavior of the adsorption column. The breakthrough times followed the order of 4-nitrophenol > 2-chlorophenol > phenol at all key operating conditions. This arrangement was attributed to their relative adsorption capacities. Data-oriented artificial neural network (ANN) technique along with two empirical physical models (Thomas and Yan model) was employed to characterize breakthrough curves for the adsorption of phenolic compounds through the fixed-bed column. Although, both Thomas and Yan models were able to fit well the breakthrough curves obtained at various operating conditions, a nearly perfect match between experimental breakthrough curves and the predicted ones was attained using ANN. The results of the present study demonstrated that the ANN technique can be employed as a powerful technique for modeling of adsorption process.