Redox cycles with doped calcium manganites for thermochemical energy storage to 1000 °C

Redox cycles of doped calcium manganite perovskites (CaMnO3−δ) are studied for cost-effective thermochemical energy storage at temperatures up to 1000 °C for concentrating solar power and other applications. If the thermodynamics and kinetics for heat-driven reduction can be tailored for high temperatures and industrially accessible low O2 partial pressures (PO2⩾10-4 bar), perovskite redox cycles can offer high specific energy storage at temperatures much higher than state-of-the-art molten-salt subsystems. To this end, a range of A-site and B-site doped CaMnO3−δ were screened for their reducibility at 900 °C and PO2≈10-4 bar via thermogravimetric analysis. For compositions with high reducibility, notably A-site doped Ca1−xSrxMnO3−δ (x=0.05 and 0.10) and B-site doped CaCryMn1−yO3−δ (y=0.05 and 0.10), oxygen non-stoichiometry δ with respect to temperature and PO2 were measured and used to fit thermodynamic parameters of a two-reaction, point-defect model of the redox process for the two prominent crystalline phases (orthorhombic and cubic) that the perovskites occupy during the cycle. The fits compare favorably to differential scanning calorimetry measurements with the magnitude of the overall reduction enthalpies decreasing as the degree of reduction increases and the perovskites shift from orthorhombic to cubic crystalline phases. Based on thermodynamic limits, redox cycles of both Ca1−xSrxMnO3−δ compositions between air at 500 °C and PO2≈10-4 bar at 900 °C can store and release up to ≈700 kJ kg−1 with over 50% of the total energy stored as chemical energy. This is approximately 140 kJ kg−1 more chemical energy than the thermodynamic limits for CaCryMn1−yO3−δ compositions under the same cycle conditions. Approaching these thermodynamic limits for the specific energy storage of these redox cycles in a concentrating solar plant requires fast kinetics for perovskite reduction in the solar receiver and for reoxidation in the heat recovery reactor. Isothermal packed-bed redox cycling experiments of Ca1−xSrxMnO3−δ and CaCryMn1−yO3−δ compositions at temperatures up to 1000 °C show that reoxidation is fast compared to reduction. Thus, specific thermochemical energy storage is limited by residence times available for high-temperature reduction. The Sr-doped compositions approach higher fractions (≈90% or more) of the specific energy storage equilibrium limit after 300 s of reduction in the packed bed configuration above 800 °C and completely reoxidize in ⩽20 s in air. Non-isothermal cycling with heating from 500 °C to 900 °C in low PO2≈10-4 bar and subsequent reoxidation during cooling in air back to 500 °C demonstrate excellent chemical stability over 1000 cycles for all doped CaMnO3−δ compositions tested. The results suggest that these redox cycles may offer a viable energy storage subsystem with long-term stability for future concentrating solar plants and other high-temperature energy storage applications.