Rh, Ni, and Ca substituted pyrochlore catalysts for dry reforming of methane Original Research Article

Dry reforming of methane (DRM) has been widely investigated, with most studies showing rapid deactivation due to carbon deposition. This suggests a need to develop catalysts that limit carbon formation while avoiding structural changes at the elevated temperatures typical of this reaction. Here, we report CO2 reforming of methane on four pyrochlore catalysts. First, rhodium was partially substituted for Zr in lanthanum zirconate (La2Zr2O7) to give La2Zr2−xRhxO7−δ (LRZ, x = 0.112, 2 wt% Rh) pyrochlore. A second pyrochlore catalyst was synthesized in which Ca was further substituted into the La-site to give La1.95Ca0.05Zr2−xRhxO7−δ (LCRZ, x = 0.055, 1 wt% Rh). This was done to compare the effect of Ca substitution on the La site along with Rh substitution on Zr site. A third catalyst was synthesized where Ni was substituted in the Zr-site to give La2Zr2−xNixO7−δ (LNZ, x = 0.112, 1 wt% Ni). A fourth catalyst, containing no catalytically active metal, La1.97Sr0.03Zr2O7−δ (LSZ), was synthesized to provide a direct comparison to the substituted pyrochlores. Effects of substitution and kinetic measurements were examined for dry reforming of methane in a fixed-bed reactor. XRD results prior to reaction showed that all the pyrochlore catalysts had a cubic unit-cell lattice. Results from XPS suggested that there were two oxidic phases (RhO2 and Rh2O3) in Rh-substituted LRZ and LCRZ and two phases (NiO and Ni2O3) present in Ni substituted LNZ. TPR of the catalysts confirmed the presence of two reducible Rh species in LCRZ and LRZ, and four reducible Ni species in LNZ. Textural measurements revealed that among the active catalysts, LCRZ had the highest BET surface area (10.0 m2/g) and pore volume (0.10 cm3/g). Temperature programmed surface reaction (TPSR) tests indicated different light-off curves for different catalysts, with LCRZ being the most active by this measure. Steady state tests at 750 °C using an equimolar reactant feed for 450 min showed that the Ni-based pyrochlore (LNZ) deactivated rapidly. LCRZ and LRZ showed similar activity, however, LCRZ showed lower carbon built-up. XRD of the spent catalysts showed that the pyrochlore structure was unchanged during reaction for all catalysts. Carbon deposited on catalyst surface during reaction was characterized by TPO. The Ni-based pyrochlore showed higher carbon deposition (1.4 g/gcat.) than either LCRZ (0.26 g/gcat.) or LRZ (0.44 g/gcat.). These results suggested that Rh substituted into the pyrochlore was more active and selective for synthesis gas compared to a directly comparable atomic loading of Ni. In addition, the replacement of Ca2+ for La3+ may provide improved oxygen mobility of the catalyst (through the introduction of lattice oxygen defects) resulting in the oxidation of carbonaceous species deposited on the active sites on the catalyst during the reaction. Substitution of metals into the crystal lattice might have also lowered the bond energy of La–O and Zr–O lattice bonds resulting in the release of oxygen from the lattice, which probably oxidized surface carbon thus slowing down carbon accumulation.