A principle of statistical factorial design was applied to characterize behaviour of mass-transfer performance in the carbon dioxide (CO2) absorption process using monoethanolamine (MEA) and blended MEA/methyldiethanolamine (MDEA) solutions. Over a hundred experiments were carried out using a bench-scale absorber packed with structured packing. The absorption performance was analyzed in terms of overall mass-transfer coefficient (K(G)a(e)). The results show that CO2 loading of the solution is the most influential process parameter on K(G)a(e), followed by alkanolamine concentration, CO2 partial pressure in feed gas, temperature, and liquid circulating rate, respectively. The temperature interacts with all tested process parameters. A mixing ratio between MEA and MDEA in the blended solution significantly affects K(G)a(e). Second-order correlations for the prediction of K(G)a(e) were developed for both MEA and MEA/MDEA systems. A relative packing factor was derived as a function of packing surface area and eventually integrated into the developed mass-transfer correlations for the purpose of scale-up. (C) 2008 Elsevier B.V. All rights reserved.