Applied Catalysis A: General, Vol.151, No.1, 267-287, 1997
Mathematical-Modeling of Transient Diffusion and Adsorption of Cyclopropane in Nax, Ni/Nax and Eu/Nax Zeolites
The transient uptake of cyclopropane gas in NaX zeolite under isothermal conditions has been simulated by a mathematical model. The model has been curve-fit to the experimental data obtained in a stainless steel microreactor with a small bed of catalyst (NaX zeolite). Intra-crystal line diffusion of cyclopropane gas was assumed to play a significant role in the uptake of the adsorbate gas in the zeolite. Based upon this assumption, the response to a switch from a stream of pure non-adsorbing argon to one of 0.5% cyclopropane/argon (a step function) was simulated by assuming an effective intracrystalline diffusivity of the cyclopropane/argon mixture and also taking into account the CSTR conditions in the microreactor. The Langmuir adsorption isotherm was also used to explain the adsorption of the cyclopropane gas in the active sites of the zeolite catalyst. From the simulation results and subsequent curve-fit of the experimental data, the effective diffusivity of the cyclopropane/argon was estimated to be about 2 x 10(-11) cm (2)/s. The diffusivity of the cyclopropane in Ni/NaX was estimated to be about 2 x 10(-12) cm(2)/s and 3 x 10(-12) cm(2)/s for the Eu/NaX system. These values of the diffusion coefficient seem reasonable in comparison to the results of diffusion coefficients obtained by similar methods such as Zero Length Chromatography and gravimetric techniques for other organic components in different zeolites. A non-isothermal model taking into account the heat of adsorption of cyclopropane and the activation energy of diffusion of the gas mixture was also formulated to observe any possible temperature rise in the catalyst bed. The temperature of the bed rose not more than 2 degrees C, causing very little changes in the effective diffusivity of the gas mixture, thereby justifying the assumption of isothermality of the uptake process, according to the modeling data.