(559c) Fundamentals of Melt Infiltration for Supported Catalyst Preparation. the Case of Co/SiO2 Fischer Tropsch Catalysts

1.      Introduction

Supported metal catalysts are prepared for example via impregnation-drying,[1] electrostatic adsorption[2] or deposition-precipitation.[3] Although it is fair to say that impregnation-drying is still the most commonly applied method due to its convenience and low costs. An alternative preparation method can be melt infiltration by which a molten salt is introduced to a porous support via capillary forces, without the use of a solvent. Transition metal nitrate salts are particularly useful, as they have low melting points due to the high amount of crystal water. Melt infiltration has been applied to load mesoporous supports to prepare single crystal mesoporous metal oxides.[4] However, for supported catalysts a high dispersion of the metal oxide is intended. Alternatively, melt infiltration was applied with Fe(NO₃)₃·9H₂O on MCM-41, but resulted in the formation large extraporous crystallites.[5] Also melt infiltration of Co(NO₃)₂·6H₂O on extrudates was applied to obtain an egg-shell distribution.[6]

In this work we report on the fundamentals of melt infiltration and its application to supported catalyst preparation. First, the extent of pore filling by the melt infiltration step is determined with differential scanning calorimetry. By following the melting behaviour of a transition metal nitrate salt in the presence of a mesoporous support direct information is obtained on the amount of salt and its the location, i.e. intraporous versus extraporous. In addition, the effect of the thermal treatment is evaluated with XRD and TEM. Finally, the calcination atmosphere is used to control the metal oxide dispersion and distribution after calcination. As a test case, Co/SiO₂ catalysts are prepared via melt infiltration for the Fischer Tropsch reaction.

2.       Results and Discussion

 The extent of pore filling of SBA-15 with Co(NO₃)·6H₂O (CoN) by melt infiltration is determined with differential scanning calorimetry. Figure 1 shows the melting of different amounts of CoN relative to the pore volume recorded during two subsequent heating cycles. In the first cycle a melting peak is observed at ~53 ºC for all samples, corresponding to melting of extraporous CoN. In the second cycle, melting at ~53 ºC is only observed for theoretical fillings above 75%, which indicates that the maximum filling has been achieved. Quantification of the residual extraporous CoN in the second heating confirmed a filling limit at 75% V_CoN/V_p, based on the density of crystalline Co(NO₃)₂·6H₂O. The melting peak between 5 ºC and 20 ºC is ascribed to melting of intraporous CoN. The low enthalpy involved in this transition indicates a non-crystalline CoN structure. Therefore the density of CoN inside the pores is likely lower than crystalline Co(NO₃)₂·6H₂O, which results in a higher effective filling than 75%.

The effect of the thermal treatment on the extent of pore filling was evaluated with XRD and TEM. An isothermal step of 12 hrs at 60 ºC was introduced to allow full equilibration of the melt infiltrate in an open crucible as well as in a closed vessel and XRD patterns were recorded. After open melt infiltration, the formation of extraporous Co(NO₃)₂·4H₂O was detected. In the closed vessel, an amorphous intraporous Co(NO₃)₂·6H₂O phase had formed. With TEM many empty support particles were observed after calcination of the open melt infiltrated sample which confirms that precursor decomposition had prevented complete infiltration (data not shown).

The final Co₃O₄ dispersion was controlled by the calcination atmosphere. Plug-like nanoparticles that spanned several pore widths formed after calcination under N₂. However, a highly dispersed system was obtained when the calcination procedure with 1% NO/N₂ was applied.[7] The calcined catalysts were reduced and tested for the Fischer Tropsch reaction. Their activities were comparable to the maximum activity reported recently.[8]

3.      Conclusion

The applicability of melt infiltration for supported catalyst preparation was studied in detail. With differential scanning calorimetry, the extent of pore filling was directly determined. A limit was observed at 75% which is likely related to a lower density of the intraporous phase. In addition, it was established that to obtain a homogeneous pore filling in all support particles, decomposition of the precursor needs to be prevented. The calcination atmosphere allows control over the metal oxide dispersion. Co/SiO₂ catalysts were prepared via melt infiltration and these showed high activity for the Fischer Tropsch reaction.

Figure 1. DSC thermograms of SBA-15/CoN mixtures with increasing amounts of CoN (% V_CoN/V_p), recorded during heating at 2.5 º/min.

4.      References

[1] Neimark, A.V., et al., Ind. Eng. Chem. Prod. Res. Dev., 1981, 20, 439.

[2] Schreier, M., et al., Journal of Catalysis, 2004, 225(1), 190.

[3] van der Lee, M.K., et al., Journal of the American Chemical Society, 2005, 127(39), 13573.

[4] Wang, Y.M., et al., Advanced Functional Materials, 2006, 16, 2374.

[5] Schüth, F., et al., Microporous and Mesoporous Materials, 2001, 44, 465.

[6] Iglesia, E., Applied Catalysis a-General, 1997, 161(1-2), 59.

[7] Sietsma, J.R.A., et al., Angewandte Chemie-International Edition, 2007, 46(24), 4547.

[8] Den Breejen, J.P., et al., Journal of Catalysis, 2010, 270, 146.