(226d) Chemical Transient Kinetics in Studies of the CO Hydrogenation Mechanism over Co-Based Catalysts

Despite extensive studies on the mechanism of the catalytic hydrogenation of CO a generally accepted reaction mechanism responsible for the formation of chain-lengthened paraffins, olefins and oxygenates is still not available. C-C coupling and CO-insertion remain the two main mechanisms considered for the reaction. The distinguishing feature of these two mechanisms is whether cleavage of the C-O bond in the CO molecule occurs before (leading to C-C coupling) or after incorporation of the C₁ monomer (CO-insertion) into the growing hydrocarbon chain. The present contribution is aimed at scrutinizing which of the mechanistic suggestions are compatible with the experimental results obtained by quantitative Chemical Transient Kinetics (CTK) [1]. In particular, we will compare results obtained using model catalysts such as Co/MgO with those obtained for bimetallic CoMn-based ones which recently showed promising catalytic performance in the high-pressure synthesis to tunable product classes, including aldehydes and alcohols [2].

CTK experiments start with a catalyst in dynamic adsorption-desorption equilibrium with hydrogen. A “build-up” transient is then initiated by switching to CO/H₂ mixtures. Similarly, switching back to H₂-only conditions, after establishing steady-state reaction conditions, allows monitoring how the FT catalytically active phase is being scavenged. Build-up and back-transient measurements are usually performed at atmospheric pressure conditions in the presence of reference noble gas to allow a quantitative evaluation in terms of molecular fluxes. All catalysts were prepared using the oxalate method of co-precipitation [1, 2]. Oxalates provide a metal organic framework (MOF) which may be thermally decomposed into nanosized particles with core-shell structure [1].

Our studies with Co/MgO as well as pure Co and Ni model catalysts allowed the atomic amounts of carbon, oxygen and hydrogen during the “build-up” to be determined. This way we showed that the monolayer limit was surpassed even before reaching the steady state. Thus, the catalyst surface did not provide metallic sites under steady state conditions. Such surface atom counting turned out to be difficult for CoMn catalysts. According to electron microscopy studies [2], major amounts of Mn in these catalysts were present as Mn₅O₈ which may undergo partial reduction under overall reducing conditions of the FT process.

Furthermore, for all of the investigated catalyst systems, chain-lengthened hydrocarbons appeared in sequence with considerable delay times (about 20 s) when switching from dynamic hydrogen adsorption-desorption to reactive FT conditions. Besides alkanes, CoMn catalysts also produced olefins under atmospheric pressures. A quantitative evaluation of the delay times during build-up clearly showed olefins form later as compared to alkanes. We therefore conclude that olefins cannot be the precursors of alkane formation. In any case, gaseous CO was necessary for chain lengthening to occur. This behavior is in agreement with a CO insertion mechanism. The provided interpretation was supported by the finding that during back-transients from FT to non-reactive conditions CO disappeared with the same time constant as Ar reference gas. Strikingly, plotting the Anderson-Schulz-Flory” (ASF) chain lengthening probability α for hydrocarbons clearly showed proportionality to the CO pressure but non-monotonous behavior as a function of accumulating carbon (or CH_x) as would be expected if CH_x insertion or C-C coupling were occurring.

The largely delayed appearance of gaseous CO (time-correlated with chain lengthening) after switching molecular fluxes from H₂ to CO+H₂ is considered to be due to the irreversible chemisorption of CO. Since no oxygen containing species (CO₂ and H₂O) appeared either, at these short reaction times, we consider the formation of surface hydroxyl groups as target of a CO insertion is key to the formation of the C1 “most abundant surface intermediate”. Combing transient measurements with DRIFTS (Diffuse Reflection Infra-red Fourier Transformation Spectroscopy) clearly revealed the formation of formate-type species.

As to the back-transients, alkanes are observed to pass through a maximum intensity when switching from CO+H₂ to H₂. On the other hand, olefins do not show this behavior and fade away exponentially immediately after switching gases, i.e. their time constant in a first-order desorption process is shorter than that for alkanes. The consequences of this observation with regard to the reaction mechanism will be discussed.

In summary, we provide clear evidence for the occurrence of a CO insertion mechanism leading to chain-lengthened hydrocarbons in the Fischer Tropsch reaction. The surfaces of Co, Ni, Co/MgO and Co/Mn₅O₈ catalysts do not provide metallic sites during the synthesis. Chain lengthening occurs time-correlated with CO gas. The ASF chain lengthening probability is directly proportional to the CO transient pressures.

 References

1. Schweicher, J., A. Bundhoo, and N. Kruse., J. Am. Chem. Soc. 134 (2012) 16135−16138.

2. Xiang, Y., Kruse, N., Nature Commun. 7, 13058 (2016).