(753a) Mechanistic Understanding of the Highly Selective Oxidative Dehydrogenation of Ethane Via Novel Supported Alkali Chloride Catalysts

The oxidative dehydrogenation of ethane is a promising process for the selective production of ethylene, one of the most essential building blocks in the chemical industry. While steam cracking is the state of the art process for the industrial ethylene production, it entails crucial drawbacks: Steam cracking leads to a wide range of olefins with only mediocre ethene selectivity, it requires high reaction temperatures, and coking is a serious problem which imposes a frequent regeneration of the reactors. By contrast, the oxidative dehydrogenation allows higher ethane selectivity levels and lower reaction temperatures. Due to the lower reaction temperature, alkyne formation can almost completely be avoided. As coking is not a problematic issue, ODH of ethane allows a continuous plant operation.

However, ODH of ethane can only be an alternative to steam cracking if high ethene selectivity levels can be guaranteed which is only possible with advanced catalysts. Tailoring highly efficient catalysts for the oxidative dehydrogenation is a challenging task: The thermodynamically favored total oxidation leads to carbon oxides as byproducts, mainly initiated by the re-adsorption of ethylene on metal cations. Thus, an efficient catalyst has to minimize the re-adsorption of olefins and simultaneously allow high activities for the alkane activation.

In this work, supported alkali chloride catalysts are investigated, mainly focusing on the ODH reaction mechanism of those materials. Essential feature of those advanced materials is the presence of a solid core (magnesium oxide partially doped with Dy₂O₃) and an alkali chloride shell that is molten under reaction conditions. It is shown that especially using eutectic mixtures of LiCl with other alkali / alkaline earth metals is a highly promising approach to tailor highly efficient materials.

To better understand reactivity and selectivity of this novel catalyst concept, it is imperative to elucidate the reaction mechanism, including information about elementary steps and the rate determining step. The mechanistic understanding is fundamental and forms the basis for improving this class of catalysts. Such information for those catalysts is extracted from kinetic measurements, both in steady state and in transient configuration. Furthermore, isotopic labelling studies (especially isotopic oxygen scrambling) were performed and kinetic experiments were coupled with in-situ spectroscopy to identify the catalytically active site.

Experimental

Catalysts were prepared by wet impregnation, using MgO (partially doped with Dy₂O₃) as a support and alkali chlorides (LiCl and related eutectica containing K⁺ or Na⁺) as shell layer. Characterization involved in-situ XRD, in-situ Raman and IR spectroscopy. Steady state and transient reaction kinetic studies were performed in tubular fixed-bed quartz reactors equipped with GC and MS gas analysis systems.

Results and discussion

In steady state configuration, it could be shown that a co-feeding of CO₂ did not result in a change of the ethene formation rates. Thus, it is can be ruled out that  CO₂  interacts with the chloride melt and blocks active sites.

Step experiments showed a significant oxygen uptake by the catalysts. The retained oxygen reacted quantitatively with ethane at nearly 100% selectivity to ethylene. Thus, oxygen chemically bound in the melt is active and selective for ODH. By contrast, it is shown that ethane does not dissolve into the highly polar chloride melt. The C-H bond activation per se must thus take place at the surface of the molten chloride shell.

The reactive oxygen species is assumed to form at the interface between support and overlayer, while its concentration additionally depends on the properties of the overlayer. This has been deduced from the following facts:

i)     The concentration of stored oxygen depends on the properties of the support material: Less oxygen is stored in catalysts with MgO support compared to MgO doped with Dy₂O₃, even though the addition of Dy₂O₃decreases the specific surface area of the support.

ii)    Reactivity is strongly influenced by the chemical and physical properties of the support.

iii)  The amount of stored oxygen increases with the melt thickness and chloride loading, respectively.

iv)  The composition of the overlayer plays a critical role for the oxygen-storage properties. A shell of Li-Na-Cl in eutectic mixture enables the uptake of more oxygen compared to a layer of Li-K-Cl.

v)    The oxygen uptake increases with temperature, implying that the absorption of oxygen in the chloride melt is an activated process.

Therefore, an intermediate must be formed and the activation process relates to the oxygen dissociation. However, an inert purging period between oxygen loading and ethane exposure in step experiments shows that part of the oxygen uptake is reversible. This indicates that part of the oxygen is dissolved molecularly in the melt, while the other part is dissociated and retained in a chemically bound form.

Step experiments varying the ethane concentration reacting with the oxygen loaded catalyst showed a dependence of the initial ethene formation rate on the ethane partial pressure in the feed stream. Thus, it can be concluded that the C-H bond activation has to be the rate determining step of the ODH reaction.

The concentration of stored oxygen can be correlated as well with the steady-state activity, while the viscosity of the melts mainly influences the selectivity towards ethene. Regarding the selectivity, we assume that the molten overlayer re-arranges quickly, preventing the formation of defect sites that are known to favor total oxidation reactions. Furthermore, we assume that molten surfaces tend to weakly adsorb ethene, as the mobility and the diffusion of metal cations reduce the adsorption of olefinic double bonds. Thus, the re-adsorption of ethene is suppressed due to the rearrangement of the surface.

This concept of self-healing surfaces is assumed to be the key factor to control the ODH selectivity.

Mechanistically, the functionality of the catalyst can be explained in the following way: If a molten overlayer forms under reaction conditions, oxidation and reduction sites are spatially separated. The catalytically active species, i .e. hypochlorite as claimed in literature, is formed at the interface of the support and the melt by oxidation of a chloride anion. Ethane activation, however, occurs on the surface of the molten overlayer where the hypochlorite is reduced, avoiding that ethane is exposed to the oxygen reduction site at the surface of the support.

The catalysts used in this study are supposed to work via a Pseudo-Mars-van-Krevelen mechanism, with oxidation and reduction sites being not identical. The kinetics of the oxidation of the catalytically active intermediate species is different from the reduction kinetics, as the hypochlorite ion is not stable at elevated temperatures and thus supposed to form only in small concentrations. As the ethane activation takes place at the surface of the catalyst, the kinetics of the intermediate reduction is determined by surface reactions, which confirms our findings on the rate-determining step.

Detailed information about the mechanism allows a comprehensive understanding of the functionality of the catalyst and forms the basis for tailoring new, even more efficient catalysts for ODH, which could help to make ethane ODH economically more attractive.