Determination of the mechanisms and effects of poisons on Cobalt-based Fischer-Tropsch catalysts

10 % Co/ZnO, 10 % Co/SiO2 and 10 % Co/TiO2 catalysts were prepared and screened using a fixed-bed reactor under industrially relevant F-T conditions. These materials have been characterised using a range of analytical techniques, including N2 porosimetry, XRD, TPR and CO pulse chemisorption analysis, in order to better understand how their physical and chemical properties are related to F-T performance. Physical and chemical properties of each catalyst varied significantly depending on the nature of the support material. A positive correlation between Co3O4 crystallite size and the average pore diameter of the support was observed. Therefore, the average Co3O4 crystallite size was thought to be determined by the nature of the support pore structure. Co/SiO2 reduced at a slightly lower temperature than Co/TiO2 and Co/ZnO. TPR profiles were also broader in nature for Co/TiO2 and Co/ZnO indicating increased metal-support interactions during catalyst activation. F-T activity of each catalyst was screened and the nature of the support material significantly influenced performance. Specific activity increased with increasing XRD crystallite size, metal dispersion and average pore diameter. However, turn over frequencies did not vary significantly between the three supported catalysts. Therefore, activity per active site was largely independent of support or dispersion effects. The difference in specific activity was thought to be a result of increased metal-support interactions influencing the extent of catalyst reduction. C7+ selectivity was seen to increase with decreasing XRD crystallite size, Co particle diameter and average pore diameter. Co/SiO2 had the highest C7+ selectivity value under steady state conditions. This trend was reflected in the shape of the liquid hydrocarbon product distributions which showed a shift towards higher molecular weight products. No correlation between C7+ selectivity and o/p ratio was observed in. Therefore, re-adsorption of α-olefins did not appear to account for differences in C7+ selectivity. Wax product distributions, ASF α-values and o/p ratios were consistent for all three catalysts. The primary aim of this thesis was to further develop the understanding of how sulphur poisons typically found as impurities in coal, natural gas and biomass derived syngas alter the performance of cobalt based F-T catalysts. The sulphur tolerance of an industrially supplied 10 % Co/ZnO catalyst was tested in a range of in-situ poisoning studies. Various concentrations of H2S, DMS and C2H5SH were added directly to a purpose built fixed-bed reactor and the effects on catalytic activity and selectivity were monitored. As the concentration of DMS added to the reactor increased, the length of time required for complete catalyst deactivation decreased. At 1 ppm the total DMS inlet required for deactivation was 13.5 μmol.g-1 giving a surface sulphur coverage of 0.4. Therefore at 1 ppm, one molecule of DMS was seen to deactivate approximately 2.5 surface cobalt sites. The total DMS inlet required for complete deactivation decreased with decreasing concentration of S fed to the reactor. Therefore, the effects of DMS on F-T activity of 10 % Co/ZnO were more marked at lower concentrations. Effects of H2S and C2H5SH were also tested and total S inlet required for complete activity loss increased with increasing molecular size. S poisoning also brought about an increase in o/p ratio without influencing chain growth probability. It was concluded that Co/ZnO has specific catalytic sites that independently catalyse the secondary hydrogenation of olefins and their chain-growth to higher molecular weight products. Sulphur was thought to poisons those sites responsible for hydrogenation and not those responsible for chain growth. The extent to which secondary hydrogenation was suppressed was higher using H2S compared to DMS and C2H5SH. H2S pulse chemisorption studies were carried out on all three supported catalysts. The nature of the support material did not significantly influence the amount of H2S required to saturate the catalyst surface. No bulk sulphides were detected to have formed during pulse chemisorption analysis. The ZnO support also offered no selective adsorption of H2S over the active Co metal