| The steady-state behavior of an adiabatic fixed bed in which carbon monoxide or carbon dioxide was methanated has been studied. The two methanation reactions are highly exothermic and special precautions were taken to ensure that the reactor operated approximately adiabatically.;Over a wide range of inlet concentration and temperature, a lower, quenched steady state with incomplete conversion and an upper, ignited steady state with complete conversion were found. For both reactions an inverse relationship existed between conversion and inlet concentration along the lower branch. Due to the different kinetic expressions for CO and CO(,2) methanation, rather different changes of the quenched branch were observed with respect to changes in inlet temperature.;For CO-methanation at higher temperatures only an upper branch existed. When the lower branch existed, no ignition was observed with increasing CO inlet concentration, and ignited steady states were obtained only through preheating of the reactor exit. The reaction zone, once formed, invariably moved to the entrance of the bed.;For CO(,2)-methanation, the same behavior was observed at low inlet temperatures. However, at a higher inlet temperature, a tendency to ignite occurred with increasing inlet concentration along the quenched branch. Ignition could be avoided by changing the CO(,2) concentration slowly and the quenched branch could be extended to high inlet concentrations.;The catalyst was a 25% nickel on (alpha)-alumina. The rate expressions for CO and CO(,2) methanation were found to be of Langmuir-Hinshelwood type. The apparent order of reaction changes from +1 to -1 and from +1 to 0 for CO and CO(,2), respectively.;At yet higher inlet temperatures, multiplicity disappeared and the graph of conversion versus concentration had the shape of a hook. At low concentrations, the conversion decreased with increasing CO(,2)-concentration. The conversion reached a minimum and for a further small increase in CO(,2)-concentration complete conversion was obtained.;The experimental results were described with a two-phase model in which the catalytic reaction rate was computed using an effectiveness factor. A rational approximation to the effectiveness factor was developed to facilitate fast computations.;The model could fit the experimental results only by changing two of its parameters arbitrarily. It was necessary to assume that the axial dispersion of heat was four times greater than predicted from established correlations. It was also necessary to assume that the adsorption coefficient in the kinetic rate expression was larger than that found in the kinetics study. With these changes, all the experimental observations could be predicted.;Pellet multiplicity was not found during the numerical simulation. Thus, the multiplicity was due to the strong axial dispersion of heat. |
