Model supported catalysts, support effects study

Support Effects Studied on Model Supported Catalysts... 

The purpose of this review is to summarize briefly from the new GPLE perspective what has been learned from experimental studies of supported metal catalysts regarding the kinetics of sintering. Companion reviews [17,18] provide more comprehensive analyses of kinetic data and mechanistic information obtained from model supported catalysts [17] commercially-relevant real supported metal catalysts [18]. The discussion in this paper focuses on the effects of atmosphere temperature and catalyst properties on the kinetics of sintering of the letter group of catalysts. 

An ideal study of support effects requires model catalysts with metal particles that are identical in size and shape (so that only the support oxide varies). This is difficult to achieve for impregnated catalysts, but identical metal particles can be prepared via epitaxial model catalysts [36]. Well-faceted Rh nanocrystals were grown on a 100-cm area NaCl(OOl) thin film at 598 K. One half of a Rh/NaCl sample was covered with Al Oj, and the other half with TiO. The preparation of Rh particles for both Al Oj- and TiO -supported model catalysts in a single step prevents any differences in particle size, shape, and surface structure which could occur if the samples were prepared in separate experiments. Three model catalysts were prepared, with a mean Rh particle size of 7.8, 13.3, and 16.7 mn (the films were finally removed from the NaCl substrate by flotation in water). Activation was performed by O /H treatments, with the structural changes followed by TEM (Fig. 15.6). Oxidation was carried out in 1 bar O at 723 K prodncing an epitaxially grown rhodium oxide shell on a Rh core (cf Fig. 15.5e), whereas the hydrogen reduction temperature was varied. 

Effects ci support Principles associated with metal-support effects and their relationships to sintoing and redispersion phenomena are addressed by Bartholomew. An important and peiht obvious concept in regard to studies of model catalysts is that investigatitni of model catalysts, because of thdr lack rtf porosity and surface roughness, provides a more definitive measure of the relative strmgths of metal-suf rt interactions than study conventional supported metals. 

In the early 80 s, Bachelier et al. [9] demonstrated that catalytic efficiency depends on the Mo loading of the catalyst. We have studied the effect of molybdenum loading on HDS activity and found an optimum metal loading of about 6 wt% for a selected y-alumina support (Figure 1). Such a behaviour has been rationalized by a change in the molybdenum sulfide particle size [9]. With the addition of cobalt, the activity per molybdenum atom increases at low cobalt content and then reaches a plateau as shown in Figure 2. Such a result confirms earlier works done by Bachelier et al. on NiMo catalysts [10, 12]. Today, the preferred interpretation is that cobalt atoms decorate the molybdenum sulfide particles. This hypothesis was predicted by a geometrical model [8] and confirmed experimentally [12],... 

This mechanism includes reversible adsorptions of NO and CO, steps (1) and (2), and the dissociation of adsorbed NO, step (3) as rate determining step. The values of tlie rate constant of step (3) and of the equilibrium adsorption constants of CO and NO determined on these different Pt catalysts were discussed in terms of changes in the adsorption properties of Pt induced by support effects [10]. Hence kinetics could be useful to state on the modifications in the extent of such interactions when Rh is added to Pt, in particular when the deactivation proceeds during the CO+NO reactions. This study reports kinetic data on a fresh and on an aged bimetallic Pt-Rh/AljO, catalyst which have been further interpreted with kinetic models including competitive adsorptions of NO and CO on a single kind of active site as well as non-competitive adsorptions in accordance with preferential adsorptions of the reactants on Pt and Rh sites as suggested by Van Slooten and Nieuwenhuys [11]. 

The majority of published research has concentrated on the preparation of the catalyst - the effect of different supports and different metals, the addition of second metals and the effect of different preparation methods on the selectivity of the catalysts for selective hydrogenation [2,3,5,6-10]. The effects of reaction conditions on selectivity have received considerably less attention. Gallezot and Richard [4] commented on the scarcity of systematic studies on the influence of reaction parameters such as pre-reduction of the catalyst, temperature, pressure, concentration of reactant and nature of the solvent for a given catalyst and reaction. Since then Singh et al. [11] have obtained quantitative kinetic data on the liquid phase hydrogenation of citral over Pt/SiOa catalysts and have used this information to present a kinetic model which fits their data. 

The information on the activity of various carbons for hydroprocessing reactions involving model compounds is usually part of the more comprehensive studies involving carbon-supported catalysts. In such studies, the activity of carbon is only determined as a baseline for investigating catalytic effects of the addition of metals to carbon support. With respect to the catalytic activity of carbon alone, this makes the experimental results less conclusive. Moreover, the accuracy of the experimental data may be limited because of generally low conversions observed when carbons alone are used as catalysts. For the purpose of this review, the results from testing of carbon and conventional supports such as y-Al203 under identical conditions are of primary interest. 

Berland et al. also studied the combination of a model NSR catalyst (1 % Pt/ 10 % Ba0/Al203, denoted as Pt/Ba-Al) with oxides-based SCR samples [84]. WO3 supported over ceria-zirconia oxides (W03/Ce-Zr) were studied as the active NH3-SCR catalysts. The effect of the composition of the ceria-zirconia mixed oxides was studied with a constant WO3 loading (10 wt.% of W, added by impregnation). It is demonstrated that Pt/Ba-Al NSR catalyst can release important amount of ammonia, until over 50 % of selectivity at 300 °C (Fig. 19.8a). SCR materials W03/Ce-Zr, with different Ce-Zr ratio, were associated downstream to the Pt/Ba-Al NSR catalyst. In the NSR -I- SCR combined system, the DeNOx efficiency is strongly improved. An enhancement of 24 points in NOx conversion was obtained at 300 °C for the better SCR sample (W03/Ce-Zr(2o-go)) (Fig. 19.8b) [83]. 

A preliminary modelling analysis involved the parametric study of a multi-tubular externally-cooled fixed-bed reactor for a generic selective oxidation process, where the catalyst load consisted of cylindrical honeycomb monoliths with washcoated square chaimels, made of highly conductive supports. In this early work, the attention was focused on the effect of catalyst design. Simulation results were generated by a steady-state, pseudo-continuous 2D monolithic reactor model, where the catalyst is regarded as a continuum consisting of a static, thermally connected solid phase... 

---