oa Kinetics of the Fischer-Tropsch Reaction Over Alumina Supported Cobalt Catalyst in a Slurry Reactor

Depleting oil reserves, environmental pressure, as well as abundant reserves of coal, natural gas and biomass, have all contributed to a revived interest in Fischer-Tropsch (F-T) technology for producing ultra-clean, virtually sulfurfree, transportation fuels and chemicals. F-T technology involves conversion of synthesis gas (i.e., a mixture of H2 and CO) to a wide spectrum of hydrocarbons. In this study, the kinetics of the Fischer Tropsch (FT) synthesis reaction over 0.27 % Ru 25 % Co/Al2O3 catalyst was studied in a 1L stirred tank slurry reactor (STSR). Supported cobalt catalysts and slurry reactors are used in commercial processes for natural gas conversion to liquid fuels (e.g. ORYX GTL in Qatar). With known kinetics one can size reactors and predict their performance as a function of process conditions. Experiments were conducted at reactor pressures of 1.4 MPa and 2.4 MPa, temperatures of 205°C and 220°C, H2/CO feed ratios of 1.4 and 2.1 and gas space velocities ranging from 2 to 15 NL/g-cat*h.

Langmuir-Hinshelwood-Hougen-Watson (LHHW) type rate equations were derived on the basis of a set of reactions originating from carbide, Eley-Rideal and enolic pathways, and two empirical power laws from the literature were used to describe CO disappearance rates.

Model rate laws were fitted to isothermal experimental rates using least-squares nonlinear regression to obtain model parameter values. Physical and statistical tests were used to discriminate between rival models. Optimisations were performed by first applying bounds to obtain realistic values of the parameters and then assumptions were made regarding the degree of adsorption for some species. Finally, nonlinear regression of model rate laws using non-isothermal experimental rates was also performed by applying Arrhenius law-based constraints to obtain physically meaningful results.

Goodness of fit for the most physically significant models were compared using qualitative (parity curves) and quantitative (mean absolute relative residuals (MARR), R-square and F-test) analysis. It was found that the model based on carbide mechanism involving dissociative CO and hydrogen adsorption (M1) and the model based on hydrogen-assisted dissociative adsorption of CO followed by hydrogenation of dissociatively adsorbed CO (M3) provided the best fit to the experimental data.