Experimental measurements and modelling of CO2 solubility in aqueous mixtures of benzylamine and N‐(2‐aminoethyl) ethanolamine

Abstract In this work, carbon dioxide solubility in N‐(2‐aminoethyl) ethanolamine (AEEA) activated aqueous benzylamine (BZA) solutions are studied using a stirred‐cell reactor in the temperature and pressure range of 313.15–333.15 K and 0.2–219 kPa, respectively. AEEA is a linear diamine with a primary and secondary amine groups, and BZA is a primary cyclic amine. The concentration of the aqueous blends used are (20mass% BZA + 10mass% AEEA), (24mass% BZA + 6mass% AEEA), and (28mass% BZA + 2mass% AEEA). Density and viscosity of unloaded aqueous amine blends are also measured in experimental temperature and concentration ranges and correlated using Joubian–Acree mathematical model. Model‐predicted density and viscosity data are in good agreement with experimental results showing 0.05% and 3.41% AAD, respectively. To correlate experimental vapor–liquid‐equilibrium data, Kent–Eisenberg (KE), artificial neural network (ANN), and soft models are used. Equilibrium constants of monocarbamate formation reaction of BZA and AEEA are regressed as a function of temperature and CO2 loading to fit the experimental data with KE model expression. KE model is also utilized to estimate the pH of CO2 loaded aqueous amine solutions. ANN model is found to predict CO2 solubility with better accuracy (1.56% AAD) in comparison of KE model (8.27% AAD) and soft model (15.5%). The CO2 absorption capacity of (20mass% BZA + 10mass% AEEA) solvent (~0.8 mol CO2/mol amine) is higher than that of monoethanolamine (~0.5 mol CO2/mol amine). Heats of absorption values of (BZA + AEEA) solvents (~25 kJ/mol CO2) predicted from Gibbs–Helmholtz relationship are found to be lower than that of MEA (~87 kJ/mol CO2) and PZ (~66 kJ/mol CO2).