| dc.description.abstract | The effect of potassium salts (K2SO4, KNO3, KCl, K2CO3) on cobalt based Fischer-Tropsch catalysts has been studied experimentally and theoretically. In order to better understand the deactivation effect of potassium species, a new poisoning methodology was adopted where it is possible to deposit potassium salts on cobalt catalyst from the gas phase. This simulates a possible poisoning route as it would be in the potential Biomass to Liquids plant based on gasification and Fischer-Tropsch synthesis. The reference catalyst (20%Co/0.5%Re/γ Al2O3) was poisoned by potassium salts with loadings up to 3500 ppm. The aerosol deposition results in a lower catalytic activity ad changes in the selectivity (lower selectivity to CH4 and higher selectivities to C5+ and CO2). The results are compared with the previous results where the incipient wetness impregnation (IWI) technique was used as a method of alkali deposition, and similar response is found. Standard characterization techniques (H2-chemisorption, BET, TPR) did not show any difference between the poisoned and the reference catalyst regardless of potassium level or salt used. Fischer-Tropsch synthesis experiments were carried out at 483 K, 20 bar and a H2/CO ratio of 2.1. The catalyst activity decreased with increasing the potassium level for all the treated samples. The deactivation behavior was fitted to simple deactivation models that can be used to describe catalyst decay upon potassium addition. All treated catalyst showed minor increase in the selectivity towards heavier hydrocarbons SC5+ and SCO2 while methane production SCH4 was slightly decreasing with increasing potassium concentration. The results are in agreement with the previous work when incipient wetness impregnation was used as a method of alkali deposition. This implies that no matter how K is introduced, it acts as a severe poison. The produced aerosol particles of potassium salts are much larger than the pore size of the catalyst, indicating deposition on the external surface of the catalyst. Since there is no difference in dispersion measurements before and after deposition, it is concluded that potassium salts are not situated on the cobalt particles. It is likely that K species are mobile, probably when reaction conditions are reached but transport to the cobalt surface during pretreatment (reduction) cannot be ruled out. It is possible that water facilitates K mobility to reach the Co active sites. Experiments with ash salts, produced from burning charcoal, gave similar trend as pure potassium salts. The adsorption parameters were again unchanged, but the activity results showed that the ash salts also are poisons for the Co-based catalyst, although the effect is not as strong as it is for the pure potassium salts.
The effect of potassium on Co-based Fischer-Tropsch catalyst was further investigated experimentally, using XPS. It was found that a small amount of Co is re-oxidized in the presence of potassium compared to the reference catalyst, which might indicate that potassium promotes re-oxidation of metallic Co. XPS also confirmed that the K species are in the form of K+. Finally, theoretical work was done using DFT calculations where both Co crystallographic structures (FCC (111) and HCP (0001)) were investigated to better understand the interaction of K with the Co surfaces. Both Co surfaces showed that the adsorption energy of potassium was reduced (-2.3 eV to -1.1 eV) with increasing potassium coverage (0.11-1.00 ML). The HCOH* dissociation in the proposed hydrogen assisted mechanism was used to study the influence of potassium on the activation barriers. The calculations showed that K decreased the activation barrier compared to the clean surfaces. The enhanced HCOH* dissociation indicates that K increases the reaction rate for this elementary step. This is not in agreement with the experimental findings, but during the FT experiments the total reaction rate was measured while in the present DFT calculation the single step is studied. Therefore, the DFT results could indicate that K changes the reaction mechanism, but further investigations are needed. | nb_NO |
