471706 Evidence for CO Insertion in Alkane and Alkenes Production through Fischer-Tropsch Reaction

Friday, November 18, 2016: 10:30 AM
Franciscan D (Hilton San Francisco Union Square)
Motahare Athariboroujeny1, Greg Collinge1, Jean-Sabin McEwen1 and Norbert Kruse2, (1)The Gene and Linda Voiland School of Chemical Engineering and Bioengineering, Washington State University, Pullman, WA, (2)Voiland School of Chemical Engineering and Bioengineering, Washington State University, Pullman, WA

Despite extensive studies on the mechanism of the catalytic hydrogenation of CO according to the Fischer (FT) Tropsch synthesis, a detailed understanding of the reaction mechanism responsible for the formation of long chain hydrocarbons is still missing. The so-called C-C coupling and CO-insertion mechanisms remain the two main proposed mechanisms for FT. 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. We provide here experimental evidence for the CO-insertion mechanism via Chemical Transient Kinetics (CTK) studies, developed by our group for operando-type investigations at atmospheric pressures and above.

Chemical transients, triggered by abrupt changes in the molecular fluxes of the reactants, provide relaxation-type information on the response behavior of catalysts relative to these changes. A careful calibration of the method and reactor system allows counting atom amounts of species adsorbing on the surface from the point in time of switching the fluxes till the occurrence of a steady state (build-up phase). Similar information can be gleaned from the reverse procedure by returning from the steady state of the reaction to non-catalytic conditions (back-transients). Furthermore, from the occurrence of delay times in the formation of reactants and products during the build-up, key features of the mechanistic steps can be established. Previous studies with Co/MgO as well as pure Co and Ni allowed the H, C, and O coverages during the build-up to be evaluated for the CO hydrogenation under atmospheric pressure conditions. This way it was shown that the monolayer limit was surpassed even before reaching the steady state, with the amounts of O atoms always exceeding those of C atoms. Thus, the catalytic surface did not provide metallic sites anymore under steady state conditions.

Furthermore, no hydrocarbon chain lengthening was ever observed without gaseous CO being built-up in the reactor. Most interestingly, the “Anderson Schulz-Flory” chain lengthening probability α calculated from the build-up species clearly showed proportionality to the CO pressure which is in agreement with a CO insertion mechanism. On the contrary, a non-monotonous behavior as a function of accumulating carbon (or CHx) was observed as would be expected if CHx insertion or C-C coupling were occurring. With respect to elucidating the CO hydrogenation mechanism toward different products, the target bond of monomer insertion is as important as insertion of the monomer itself. In our previous studies, no water was being produced despite the build-up of considerable oxygen and hydrogen amounts on the surface. Accordingly, it was concluded that hydroxyl (OH) groups are most probably being formed on the surface and CO is being inserted into the O-H or O-R bond of surface hydroxyl or alkoxyl.

Despite the considerable number of arguments favoring a CO insertion mechanism in our model studies, the debate of C-C coupling versus CO insertion is still ongoing and far from being solved. In the present CTK study we provide further evidence for the occurrence of a CO insertion mechanism by choosing bimetallic “CoMn” and ternary “CoMnTi” catalysts. Both these catalysts were prepared according to the oxalate route, as developed in our laboratory, in the absence of a generic support material.

Switching molecular fluxes from H2 to CO+H2 at atmospheric pressure conditions caused reactants and products to be produced with characteristic delay times as compared to the Ar noble gas (as reference). The first species to appear in the gas phase with a delay time of less than 5 s was methane. Interestingly, besides paraffins also olefins were produced in our studies. These species were always observed time-correlated with gaseous CO. For a “Co1Mn1Ti0.1” catalyst (atomic ratios 1:1:0.1) containing Co metal and Mn (as Mn5O8 according to TEM results) in equal amounts and a non-stoichiometric TiOx phase with much smaller abundance, the delay time for the appearance of gaseous CO could be as large as 40 s. Moreover, alkanes and alkenes products occurred in sequence of their carbon number. The large delay times encountered in our study have most probably to be associated with the presence of extended oxidic phases, Mn5O8 and TiOx. Strong Metal Support Interaction (SMSI) effects may be invoked to explain the finding.

In back-transients, the CO outlet flux dropped exponentially with the same time constant as Ar. As the Ar decay is characteristic of the reactor volume emptying, we conclude that CO was irreversibly chemisorbed on the catalyst surface during steady state reaction conditions. We also note that once the CO inlet flow was stopped to initiate the back-transients, C2+ alkanes and alkenes decreased quickly with similar time constants (yet much longer than those for CO and Ar). This is in agreement with CO being the insertion monomer since no chain lengthening happened without gaseous CO. Finally, plotting the chain lengthening probability vs. the CO partial pressure resulted in a linear behavior. Quite differently, a non-monotonous behavior was observed when the chain lengthening probability was plotted vs. the titrated carbon amounts.

It is also remarkable that the formation of H2O and CO2 in our CTK studies with “CoMnTi” were largely delayed in comparison to CO and hydrocarbons. Consequently, while the catalyst took up major amounts of oxygen and hydrogen atoms during the build-up stage, the absence of water production would appear OH groups were being abundantly formed during these initial reaction times. We therefore speculate that surface hydroxyl (or alkoxyl) groups are the target for CO insertion to form formate-type species. Such species could be easily hydrogenated to form paraffins, olefins, or even oxygenates as observed when the CO hydrogenation is being performed at high pressures over “CoMnTi” ternary catalysts.


Fischer–Tropsch Synthesis, Chemical Transient Kinetics, CO Insertion

Extended Abstract: File Not Uploaded