The influence of cobalt particle size in the range of 2.6-27 nm on the performance in Fischer-Tropsch synthesis has been investigated for the first time using well-defined catalysts based on an inert carbon nanofibers support material. X-ray absorption spectroscopy revealed that cobalt was metallic, even for small particle sizes, after the in situ reduction treatment, which is a prerequisite for catalytic operation and is difficult to achieve using traditional oxidic supports. The turnover frequency (TOF) for CO hydrogenation was independent of cobalt particle size for catalysts with sizes larger than 6 nm (1 bar) or 8 nm (35 bar), while both the selectivity and the activity changed for catalysts with smaller particles. At 35 bar, the TOF decreased from 23 x 10(-3) to 1.4 x 10(-3) s(-1), while the C5+ selectivity decreased from 85 to 51 wt % when the cobalt particle size was reduced from 16 to 2.6 nm. This demonstrates that the minimal required cobalt particle size for Fischer-Tropsch catalysis is larger (6-8 nm) than can be explained by classical structure sensitivity. Other explanations raised in the literature, such as formation of CoO or Co carbide species on small particles during catalytic testing, were not substantiated by experimental evidence from X-ray absorption spectroscopy. Interestingly, we found with EXAFS a decrease of the cobalt coordination number under reaction conditions, which points to reconstruction of the cobalt particles. It is argued that the cobalt particle size effects can be attributed to nonclassical structure sensitivity in combination with CO-induced surface reconstruction. The profound influences of particle size may be important for the design of new Fischer-Tropsch catalysts.

The effects of metal particle size in catalysis are of prime scientific and industrial importance and call for a better understanding. In this paper the origin of the cobalt particle size effects in Fischer-Tropsch (FT) catalysis was studied. Steady-State Isotopic Transient Kinetic Analysis (SSITKA) was applied to provide surface residence times and coverages of reaction intermediates as a function of Co particle size (2.6-16 nm). For carbon nanofiber supported cobalt catalysts at 210 degrees C and H(2)/CO = 10 v/v, it appeared that the surface residence times of reversibly bonded CH(x) and OH(x) intermediates increased, whereas that of CO decreased for small (<6 nm) Co particles. A higher coverage of irreversibly bonded CO was found for small Co particles that was ascribed to a larger fraction of low-coordinated surface sites. The coverages and residence times obtained from SSITKA were used to describe the surface-specific activity (TOF) quantitatively and the CH(4) selectivity qualitatively as a function of Co particle size for the FT reaction (220 degrees C, H(2)/CO = 2). The lower TOF of Co particles <6 nm is caused by both blocking of edge/corner sites and a lower intrinsic activity at the small terraces. The higher methane selectivity of small Co particles is mainly brought about by their higher hydrogen coverages.

Cobalt on carbon nanofiber model catalysts with very small dispersed cobalt particles of 5 nm were subjected to H(2)O/H(2) treatments at 20 bar and 220 degrees C. Using in situ Mossbauer spectroscopy we could unambiguously prove that oxidation of the nanoparticles by water will not occur when hydrogen is present. Only in a water/argon atmosphere did oxidation take place. This rules out oxidation as the deactivation mechanism in Fischer-Tropsch synthesis. Even more important, we define the relative humidity (RH) as a key parameter to understanding deactivation by water. At a RH below 25% sintering was absent even when measuring for 4 weeks, whereas at a high RH of 62% as much as half of the small super paramagnetic cobalt particles (<5 nm) sintered into larger particles in 1 week. Activity loss as measured at Fischer-Tropsch conditions amounted to 73%, which could be directly related to the metal dispersion loss 77% due to sintering as evidenced by detailed TEM analysis of the spent sample.