Thermodynamic and kinetic characteristics of a Cu-Mn composite oxygen carrier for low-temperature chemical-looping air separation

The further development of the chemical-looping air separation (CLAS) technology is hindered by the required high redox temperature. Improving the thermodynamic properties of oxides by elemental composition is an effective way to decrease the reduction temperature of oxygen carrier. In this paper, a Cu-Mn composite oxygen carrier was prepared by the addition of Mn₂O₃ to CuO, with Cu_xMn_{3− x}O₄ as the active phase, and the oxygen uncoupling properties of the Cu-Mn/Zr composite oxygen carrier were investigated, focusing on the redox reactivity under different oxygen concentrations of Cu_xMn_{3− x}O₄ ⇋ Cu_xMn_{2− x} + O₂(g). The oxygen transport capacity of this type of oxygen carrier was determined as 0.0463 g O₂/g oxygen carrier. With an increase in the oxygen concentration (0.001–50%) in the carrier gas, the required oxygen uncoupling temperature increases (625.5–921 °C). Based on relationship between initial reduction temperature and equilibrium oxygen concentration, the following thermodynamic characteristics of the redox reaction were obtained: ΔG = –0.119 T + 150.41 kJ/mol and K_p = exp(14.31–18090.8/T). Compared to a Cu-based oxygen carrier, the reduction temperature is significantly reduced and the equilibrium oxygen concentration at the same reaction temperature is greatly increased. The redox reactivity under different heating rates and reaction temperatures were also investigated, and the results show that both the reduction and oxidation reactions are temperature driven at the considered temperatures. The reduction and oxidation kinetic models were established by the iso-conversional method. When the relative conversion X < 0.5, the reduction of the Cu-Mn oxygen carrier follows a shrinking core model (n = 3), and the oxidation follows a shrinking core model (n = 2). When X ≥ 0.5, the reduction proceeds according to a one-dimensional diffusion model, whereas the oxidation follows a shrinking core model (n = 3).