Redox energetics of novel perovskite-type oxygen carriers for chemical looping reforming

Abstract

The present work focuses on the redox energetics of novel perovskite-type oxygen carriers for chemical looping reforming. The aim of this study is to increase the level of knowledge on the redox characteristics of materials for possible applications as the oxygen carriers for the chemical looping processes. Here we focus on the perovskite-type oxides (ABO3) with lanthanum on the A-site and first row transition metals on the B-site since first row transition metals normally have more than one oxidation sate, non-stoichiometry in the perovskite oxides with such metals on the B-site is common while keeping the same structure. The partial substitution of the cation on the Bsublattice is studied as a measure to adjust the redox energetics. In the present study partial substitution of cobalt in LaCoO3 with Mn and Fe is chosen. The redox behavior of nonstoichiometric compounds may be assessed from the variation of the oxygen nonstoichiometry with temperature and oxygen partial pressure. Thermogravimetric analyses (TGA) is used for most of the systems but in order to reach higher accuracy, specialized instruments are needed. One of the most accurate techniques to measure the oxygen nonstoichiometry is coulometric titration (CT). A novel CT setup was designed, constructed and validated. This setup was subsequently used to study the oxygen non-stoichiometry of the LaMn1-xCoxO3-δ system at 1223, 1273, and 1373 K. For the LaMn1-xCoxO3-δ system it is found that the observed oxygen non-stoichiometry curve is due to the simultaneous reduction of both manganese and cobalt on the B-sublattice and the enthalpy of oxidation values show a linear dependence by x (portion of Co). The oxygen non-stoichiometry and redox energetics of the second studied system, LaFe1-xCoxO3-δ, just similar to the previous system show no indication of the sequential reduction of the cations occupying the B-sublattice. The absolute value of the enthalpy of oxidation increases as the iron portion on the B-sublattice increases, which approves this finding. In contrast to the experimental approaches, DFT calculations might provide a cost effective tool to examine complex systems and obtain an approximation for the redox thermodynamic properties; according to the following reaction: 4LaBO2.5 + O2 = 4LaBO3. It is however necessary to examine the sources of error and the accuracy of the calculations. Therefore a benchmark study was conducted on the formation energetics of lanthanide first row transition metal perovskite-type oxides (LaBO3). The benchmark shows that although fundamental errors in the GGA affect the energetics, still with addition of ad hoc corrections, it is possible to obtain values which are in good agreement with the experiment. The benchmark also confirms that as long as the calculations are spin polarized (with any magnetic structure) and performed with the relaxation of the experimental crystal structure, the total energies are reproduced within 15 kJ/(mol B). Turning to the reduced phases, LaBO2.5, the calculations proved to be much more difficult than the oxidized phases. In contrast with LaBO3 phases, a bigger fraction of total energy of the reduced phases is attributed to the magnetic configuration, meaning that a simple magnetic structure (e.g. ferromagnetic) does not predict the ground state reasonably. Still, the biggest challenge is due to the configuration of oxygen vacancies on the oxygen sublattice. Simultaneous presence of oxygen vacancy and spin configurations necessitates a computational algorithm based on statistics, e.g. Monte Carlo method, which was out of the scope of this study given the three year time frame. This however opens up an opportunity for the future works.