Sintering behaviour of model cobalt crystallites: Effect of support phase on the structure and size of crystallites

Abstract

Cobalt-based catalysts are currently being used in the Fischer-Tropsch synthesis (FTS) process. These catalysts however deactivate with time on stream. Sintering has been identified as one of the most significant deactivation mechanisms for cobalt-based catalysts. Therefore understanding the sintering mechanism of cobalt crystallites is important since this can lead to the preparation of stable catalysts. The type of support has been shown to have an influence on the activity and stability of the FTS catalysts. Typical catalyst supports for FTS catalysts include Al2O3, SiO2 and TiO2. In this study, TiO2 was used as support for model cobalt crystallites. This material occurs in different crystalline phases such as rutile, anatase and brookite. These phases have different bulk and surface properties. Three titania supports were obtained. The supports were identified as anatase-, P25- and rutile-support. The anatase-support was shown to consist of 100% anatase phase by XRD. The support was shown to have the highest surface area and mesoporous pore structure. The P25-support was shown to consist of ~85% anatase and ~15% rutile phases. The support was shown to have a surface area lower than that of the anatase-support, but higher than that of the rutile-support. P25-support also has a mesoporous pore structure. The rutile-support was shown to consist of ~2% anatase and ~98% rutile phases. This support had the lowest surface area, although still had a mesoporous pore structure. Cobalt catalysts were prepared by wet impregnation with constant cobalt loading of ~10 wt% onto the supports, followed by calcination at temperatures ranging from 200?C ? 400?C. It was found that calcination temperature in this range did not affect the size of the Co3O4 crystallites for each support. The titania support influenced the phase composition of the calcined catalysts. In this case, the titanate phase CoTiO3 was preferentially formed in rutile-supported catalysts and not in anatase- and P25-supported catalysts. Calcination in air at a temperature range from 200?C ? 400?C did not significantly affect the pore structure of the calcined catalysts. The mesoporous nature of the catalysts was retained after calcination. The calcined catalysts were reduced in H2/Ar at 450?C to convert the cobalt oxide into metallic species. The starting metallic cobalt crystallite size was influenced by the type of titania support phase as evidenced by TEM analysis. Smaller crystallites were obtained in Anatase- and Rutile-supported catalysts. P25-support gave slightly larger cobalt crystallites. The smaller cobalt crystallites in the anatase-support were shown to undergo agglomeration in the reduction step. In a prior TPR experiment, the support itself was shown to interact with hydrogen, leading to a possible reduction of the surface. It is postulated that this interaction with hydrogen might have led to the formation of surface defects and/or oxygen vacancies which in turn may influence the anchoring of cobalt crystallites. The reduction behaviour of cobalt crystallites was shown to be influenced by the titania-support. TPR showed different reduction profiles in the different supports. The titania support also influenced the presence of fcc and hcp cobalt. Both fcc and hcp cobalt were obtained in Anatase- and Rutile-supported catalysts. Only fcc cobalt was obtained in P25-supported catalysts. Sintering studies were conducted in a reducing environment and at higher temperature than usually used in Fischer-Tropsch synthesis. This was to accelerate the sintering process in order to observe changes that would otherwise occur over longer operating time. The crystallite size data from the TEM were fitted well to a lognormal distribution function to obtain crystallite size distribution (PSD) plots. The PSD curves changed with sintering time, confirming the occurrence of sintering of the cobalt particles. The crystallite size data from XRD and TEM showed a power-law type of growth when plotted against time. The sintering kinetics was also studied. In this case, the data was fitted well to the simple power law expression (SPLE) and generalized power law expression (GPLE) models. The sintering data in all three supports were fitted well to the first order GPLE model. The sintering was shown to be different in the three supports as shown by different sintering rate constants. The titania support phase influences the material science characteristics of model calcined, reduced and sintered crystallites. This effect may stem from the different metal-support interaction that occurs in the different titania phases. Sintering was seen to be more prevalent in anatase-supported catalyst compared to P25- and rutile-supported catalysts. The support porosity is also postulated to play a role in the sintering of cobalt crystallites. It was found that sintering tendency increases with decreasing pore size. From an industrial application point of view P25 is the most preferred support showing the least sintering probability.