The active role of CO₂ at low temperature in oxidation processes: the case of the oxidative dehydrogenation of propane on NiMoO₄ catalysts

Abstract

The effects of 3 vol.% of CO₂ added in the gas feed are investigated in the oxidative dehydrogenation of propane (ODP) to propene on NiMoO₄ catalysts. In spite of the low reaction temperature range (400–480 °C), it turns out that CO₂ is not inert during the reaction. CO₂ behaves as a strong oxidant with respect to the catalyst. CO₂ helps to maintain the catalytic surface in a high oxidation state under conditions for which NiMoO₄ undergoes a reduction. CO₂ clearly promotes the non-selectivity of the catalyst and can inhibit deactivation. Under particular conditions it can improve propene formation. CO₂ is able to oxidize propane to propene. We suggest that CO₂ dissociates during the reaction to CO and an active oxygen species, which is responsible for the oxidizing effect of carbon dioxide. This work opens promising perspectives in using CO₂ to tune dynamically the oxidation state of “working” oxide surfaces in their most active and selective configurations for oxidation processes.

Introduction

It is well established that dynamic processes occur at the surface of oxide catalysts during hydrocarbon oxidation reactions. Beside the fact that catalytic surfaces often undergo a reorganization at the atomic scale when molecules adsorb on them [1], [2], [3], the dynamic character of oxide catalysts in oxidation processes mainly comes from their working mechanism, namely the continuous exchange of oxygen atoms between the gas phase and the catalytic surfaces. Related to the dynamic character of a typical Mars and van Krevelen cycle, one must also mention the frequent tendency of the slightly reduced oxide surfaces to rearrange their structures to more oxygen-thrifty ones. The new structures are more stable but also less active [4]. At the macroscopic level, the dynamic behavior of the oxides at work materializes as structuration and reconstruction of the catalytic sites which lead to modifications with time-on-stream of the catalysts, of the kinetics (oscillations, bistabilities, etc.), and to the emergence of synergetic effects between phases [5], [6], [7], [8], [9].

All of this is contrary to the old “static” view that catalysts do not undergo any modification of their characteristics under the conditions of reaction. Precisely, it has been shown that higher activities and selectivities are obtained on oxides with their superficial atoms stabilized in an oxidation state promoting easy oxygen exchanges and reconstruction. Previously we mainly have focused our investigation on Mo-containing oxide catalysts, but our conclusions certainly meet those drawn by other authors with catalysts based on other metal oxides such as VPO and FePO catalysts [6], [10]. Precisely, in the case of the oxidation of hydrocarbons on Mo-containing oxides (MoO₃ and molybdates), high performances are obtained with catalysts having their surfaces maintained in a slightly reduced suboxide state [11].

The stabilization of the superficial atoms of the catalytic surface in their most efficient oxidation state can be achieved through a global approach by tuning the conditions of reaction, namely by adjusting the ratio of the partial pressures of oxygen and hydrocarbon in the gas feed, and the temperature, and by adapting the catalyst formulations, e.g. through the addition of elements in tiny quantities. The understanding of the selectivity of an oxidation process thus requires the knowledge of the effects induced by these factors on the architecture of the active sites and on the dynamic phenomena occurring at the surface of the catalyst during the reaction. But, in order to further improve the catalytic performance of a process, it is also crucial to develop in parallel new strategies leading to a better control of the dynamic behavior of the catalytic surfaces “at work”, with the objective of stabilizing in situ the active sites in a more active and selective configuration during the catalytic reaction.

A very promising approach to master in situ the dynamic phenomena at the surface of oxides at work consists of the addition of gaseous “dopes” (also termed “promoters” hereinafter) in the reaction gas feed. In fact, in fixed-bed reactors, both the partial pressure and the temperature may vary along the catalyst bed and it may thus be very difficult to maintain, in a controlled manner, the surface in a desired oxidation state. This is nearly impossible by simply modifying the global operation conditions. For these systems, the injection of dopes in the gas feed appears as a very useful tool. It is a very flexible method that will allow either to incorporate the dope directly in the gas feed at the reactor inlet or at the level of the different sections along the catalyst bed. In addition, in industry the addition of dopes could be realized easily and practically without any modification of the existent commercial operation unit. Then, the efficiency of the modulation of the oxidation state of the active sites could thus be largely improved thanks to this approach.

Presently, the addition of gas promoters is already used in some industrial applications. The partial oxidation of aromatic compounds, in particular of benzene to phenol, has been performed in the presence of a mixture of oxygen and hydrogen on supported oxide catalysts of V, W, La, Mo, etc. enriched in Pt or Pd [12]. Hydrogen has been used as a dope added to the oxygen-containing gas feed to perform the direct oxidation of propylene to propylene oxide on Ti catalysts [13]. At low temperatures (<500 °C), CO₂ has been used as a dope in the oxidation of alkylaromatics over Fe/Mo/borosilicate molecular sieve [14], in the oxidative dehydrogenation of isobutane over LaBaSm oxide catalysts [15], in the oxidation of butane to maleic anhydride [16] and principally in the oxidative coupling of methane [17], [18], [19], [20], [21]. The effects caused by the introduction of CO₂ have been explained by: (i) a faster, more extended formation of oxycarbonate phase and its regeneration; (ii) the formation of inactive carbonate species; (iii) the formation of peroxocarbonate intermediate which serves as a promoter for the gas phase oxygen oxidation; (iv) a poisoning caused by competitive adsorption on the sites where oxygen (and possibly the hydrocarbon) adsorbs and by the inhibition of molecular oxygen adsorption; (v) the decrease of the formation of coke; and (vi) the lower tendency for hydrocarbon to undergo deep oxidation. On the other hand, based on indirect observations, it can be suggested that co-adsorbates and gas dopes could modify the surface properties of the catalysts by: (i) changing the coordination and the electronic properties of the superficial metal atoms; (ii) a direct participation in the reaction mechanism; (iii) changing the acido-basicity of the oxides; (iv) blocking chemisorption sites; or (v) inhibiting the diffusion of surface species towards reacting molecules [22]. However, only a few experimental verifications of these conclusions are available in the literature and more detailed explanations are very difficult to get.

A detailed experimental approach is thus required in order to understand the exact effects of gaseous dopes during catalytic reactions. This is the aim of our investigations. In all of these cases, important and interesting effects have been observed justifying the aim of the present work.

The present communication reports on the effects brought about by the introduction of small amounts of CO₂ in the oxidative dehydrogenation of propane (ODP). Results concerning the use of N₂O as a dope are given elsewhere [23]. In this work we focus exclusively on the role exerted by CO₂ on the behavior, the characteristics of active sites and the performance of NiMoO₄.

The use of CO₂ as a dope leads to obvious difficulties of interpretation related to its extremely unfavorable thermodynamics. The two C–O bonds in the CO₂ are quite stable and a high amount of heat must be supplied for their dissociation. The entropy is not favorable because CO₂ is a small gas molecule. Both terms of the Gibbs free energy function disfavor the conversion in the gas phase of CO₂ to O₂ and CO (under a pressure of 101.325 kPa and at 25.15 °C, ΔH is 293.0 kJ mol⁻¹). It is estimated that not more than 2% of CO₂ transforms to CO and O₂ at 2000 °C. These are the reasons why CO₂ is traditionally considered as inert in oxidation reactions, which are indeed usually performed at temperatures low enough (<500 °C).

However, the interesting and surprising is that our results show exactly the contrary. We plan to demonstrate that, if all the dynamic processes occurring on the catalytic surfaces are taken into account, the dissociation of CO₂ happens at lower temperature at the surface of our catalyst and must thus be considered to interpret the experimental data correctly and exhaustively.

The strategy of this work is based on reactions in the presence and in the absence of CO₂, and on a detailed physico-chemical characterization of the NiMoO₄ catalysts before and after reaction. In order to verify the oxidant character of CO₂ and to investigate the possibility that CO₂ oxidize propane, a catalytic test with a high amount of CO₂ and a low amount of O₂ was performed. The comparison of the oxidant role of CO₂ and molecular oxygen was performed by oxidizing a molybdenum suboxide. Astounding results are obtained. Contrary to what is traditionally believed, CO₂ is not inert in catalytic oxidation reactions. Our suggestion is that the adsorbed oxygen species, resulting from the dissociation of CO₂ at the surface of the catalyst induces the oxidation of reduced Mo-containing oxides at temperatures around 400–450 °C. In addition, these oxygen species play the role of a powerful oxidant, stronger than those given by molecular oxygen. As an oxidant, CO₂ influences the behavior of the catalysts in oxidation processes. Indeed, CO₂ increases conversion but decreases selectivity. However CO₂ is able to oxidize propane to propene. Under particular conditions, CO₂ can promote the formation of propene but with a loss in the selectivity.

Our results show that the addition of CO₂ could be an important and useful tool to modulate the nature of the active sites during the catalytic oxidation reaction. They give a coherent and never reported explanation about the mechanisms by which such modulation could be realized. They give strong arguments revealing a new role attributed to CO₂ during the catalytic reaction. In addition, they also give guidelines for a utilization of CO₂ as a dope in industrial applications.