A search for alternative Deacon catalysts

Abstract

High-throughput methods, emissivity-corrected infrared thermography (ecIRT) and a sequential 10-fold fixed bed gas phase reactor have been used to accelerate the development of new catalysts for the Deacon process (4 HCl + O₂ ⇌  2 Cl₂ + 2 H₂O). Both setups were modified to withstand the corrosive process conditions. Besides the reactor equipment also the catalysts themselves often suffered corrosion during the reaction. To consider the catalyst corrosion the tested oxides were typically aged for 24 h under a HCl-O₂ atmosphere at reaction temperature. The experimentally observed stability of selected binary mixed-metal oxides is correlated with literature data for the bulk chlorination tendency of the pure oxides and the corresponding melting/sublimation point of the chloride. The starting binary mixed-metal oxides have been selected based on a set of thermodynamic data for bulk chlorination and chloride oxidation and doping of TiO₂ and SnO₂, which are preferentially used as catalyst supports in Deacon reactions. The best catalysts discovered were optimized through doping and composition variations. Characterization of the best materials is provided.

Introduction

Chlorine is one of the most important chemical elements in the production of numerous objects for our high standard of living, either directly as component but more often latently in production intermediates; hence the Cl₂ production is an indicator of the development status of a country's chemical industry [1]. However, the production of one ton of Cl₂ by the current membrane chlor-alkali electrolysis needs 2790 kWh electric power equivalent to around 2.1 tons CO₂ [2]. Due to the present energy crisis and increasing concern about anthropogenic CO₂ emissions contributing to the global warming, the huge energy consumption of this process becomes even more important [3]. Apart from technology improvements as by the development of the oxygen depolarized cathode (ODC) [4], the recovery of Cl₂ from the by-product HCl is one piece in the puzzle for more efficient Cl₂ industry, e.g. in the manufacture of isocyanates as intermediates for polyurethanes [3]. The recycling of HCl by electrolysis saves one third of energy compared to the chlor-alkali electrolysis [5], [6]. The most efficient way to generate Cl₂ from HCl is the Deacon process. This gas phase process, today based on heterogeneous RuO₂ catalysts, needs only 20% of the energy needed for HCl-electrolysis [7].

The Deacon process was invented and patented by Henry Deacon in 1868. He used Cu compounds on clay as catalysts for the oxidation of gaseous HCl to Cl₂ [8]. The intrinsic problem of Cu based Deacon catalysts is their tendency to evaporate in the presence of chlorine. In 1915, Neumann reported that the undesired catalyst vaporization during the reaction can be reduced by the substitution of Na⁺ or K⁺ ions into CuCl₂ [9]. The original Deacon catalyst has been systematically studied and refined by Shell in the 1960s [10]. In these studies the highest stability and Cl₂ production showed a catalyst composed of 5 wt% Cu, 5 wt% rare earth metal (preferential is a mixture of Pr and Nd) and 3.1 wt% K on a silica support. Further progress in the HCl oxidation has been reported by Mitsui Chemical [11] with a Cr₂O₃ on SiO₂ catalyst in a fluidized-bed reactor (MT-process). This Cr based catalyst is more stable and more active than the Cu based ones. However, there is concern about the release of hazardous Cr(VI)compounds, which is supported by the findings of Amrute et al. [12]. With the discovery of RuO₂ on rutile TiO₂ as long term stable and highly efficient Deacon catalyst by Sumitomo [13], new interest in the research for heterogeneous HCl oxidation catalysts was triggered. Recently Bayer developed Ru based catalysts on cassiterite SnO₂ for this process [14], [15], [16]. Both supports, TiO₂ and SnO₂, have the rutile structure in common. This crystal structure favors the epitaxial growth of RuO₂ on the support and is considered a requirement for a high Cl₂ yield [17], [18]. Zweidinger et al. have shown that during the catalysis only the surface of RuO₂ is chlorinated and that the process is self-regulating, which is assumed to explain the impressive stability [19]. Furthermore, Sumitomo and Bayer added SiO₂ or γ-Al₂O₃ to their commercial Ru catalysts. These additives reduce agglomeration of the active phase and enhance the long term stability [17], [20]. For a deeper understanding of the atomic scale of the Deacon process reviews by Pérez-Ramírez et al. and Over [3], [21] can be recommended. Mondelli et al. claimed stable copper-based catalysts as an alternative to noble metal (Ru) containing catalysts [22], while Amrute et al. reported noble metal free CeO₂ as suitable catalysts for the Deacon reaction at 430 °C [23].

In the Deacon process the success of the catalyst is closely related to its corrosion resistance. Due to the extraordinarily high and varying commodity price of the metal Ru there is a need for alternative catalyst compositions which are low in Ru content, active and stable under these harsh conditions. Since broad screening by one at a time experimentation is time consuming and costly, high-throughput (HT) methods had to be developed and applied to accelerate screening and testing. This combination of combinatorial chemistry and HT methods is a valuable approach for the discovery of materials under complex reaction conditions [24], [25]. In our search we applied (1) emissivity corrected IR-thermography (ecIRT) for primary screening and (2) a sequential 10-fold gas phase reactor setup with online mass spectrometry for secondary screening. Our ecIRT setup offers possibilities for the testing of up to 206 catalysts in parallel under identical reaction conditions. EcIRT relies on the spatial resolution of small temperature changes, which can be associated to heats of reaction [26]. However, competitive or parallel reactions, such as side-reactions of the catalyst with feed gases or change of emissivity due to reaction of the catalyst surface with feed gas components may hamper the interpretation of the measurement results [27]. More reliable data can be obtained by analysis of the product gas. Such an analysis has been implied in our sequential 10-fold gas phase reactor setup by online mass spectrometry. However this approach reduces the sample throughput strikingly, especially when long-time aging experiments are performed.

Herein we report the results of our HT search for new Deacon catalysts with a focus on stable catalysts compositions.