The statistical design-of-experiment method was used to identify the significant factors for making a manganese oxide-loaded activated carbon matrix. The carbon matrixes, which were made by reacting KMnO4 with carbon material, were tested as gas-diffusion electrodes for oxygen reduction. Three factors-KMnO4 concentration, reaction temperature, and reaction duration were tested in a two-level full-factorial design-of-experiment. The modification of carbon morphology and its effect on the performance of oxygen reduction are discussed. Temperature, KMnO4 concentration, and the interaction between temperature and reaction time were found to have a significant influence on the catalytic activity of the manganese oxide-loaded carbon electrode. (C) 2008 Elsevier B.V. All rights reserved.