Steam reforming of methane in a fast fluidized membrane reactor

A novel process is described for the steam reforming of methane in a fast fluidized membrane reactor (CATFORMING). Experiments were performed in a modified Riser Simulator to assess Ni/α-alumina, Ni/NaY and Ni/USY zeolite catalysts. Composite Pd/Inconel H₂ permeation membranes were prepared and characterized. Experimental conditions of temperature (750–850°C), total pressure (2070–2760 kPa) and molar steam to methane ratio (2–6) were selected to match those in use at the industrial scale.; Four composite membranes were produced, three by electroless deposition and one by thermal spraying. Permeation of H₂ was assessed over the range 750–850°C using a 156 μm Pd/Inconel membrane. The permeability coefficient was 2.033 × 10−6 mol/(m s kPa 0.5) with an activation energy of 22.3 kJ/mol. Membranes prepared by the electroless technique possessed H₂/Ar molar selectivities of 400–1200:1. The thermal spray membrane possessed excellent adherence but low H₂/Ar molar selectivity.; Two membranes prepared by electroless deposition (80 μm, 156 μm) were used to assess the effect of H₂ permeation from the reactor on conversion of methane over 20 wt% Ni/α-alumina catalyst. H₂ permeation from the reactor produced conversions in excess of those attainable at equilibrium. Removal of H₂ from the reactor also modified favourably the composition of the synthesis gas.; Comparison was made between USY and NaY zeolite and α-alumina containing 20 wt% Ni. Catalytic activity of the zeolites was not stable due to steam induced collapse of zeolite crystal structure. Over Ni/α-alumina, catalytic activity was related to the available metal surface area. Maximum area was produced at 2.5 wt% Ni (0.535 m2/g). Metal loading of 4 and 20 wt% Ni produced lower metal surface areas, 0.416 and 0.406 m2/g respectively. Below 2.5 wt% catalysts possessed little or no activity. Decay of catalytic activity with time-on-stream was similar for each of the α-alumina catalysts. Coke formation was under kinetic rather than equilibrium control and followed the Voorhies relation. Propensity for coke formation was related to the nickel crystal size.; The Box-Hill criterion was used to design sequential experiments for model discrimination. The rate of methane consumption was best described using a Langmuir-Hinshelwood model for the adsorption of methane. Parameters assessed using a large number of observations collected over 4 and 20 wt% Ni/α-alumina catalyst indicated that the contribution of adsorption was small at elevated temperature. A first order dependency on the partial pressure of methane was found to be adequate. Over the 4 wt% Ni/α-alumina catalyst, the pre-exponential factor and apparent activation energy were 5.14 × 10−10 mol/(kPa g_cat s) and 52.4 kJ/mol. Over the 20 wt% Ni/α-alumina catalyst, the pre-exponential factor and apparent activation energy were 1.15 × 10−11 mol/(kPa g_cat s) and 86.5 kJ/mol.