Metal crystallites, formation

Grubbs and coworkers (35) while examining Rh and Co catalysts derived from 14 reported the loss of infrared CO stretches and visual darkening of the catalysts after use for hydrogenation of olefins, aldehydes or ketones, cyclohexene disproportionation to benzene and cyclohexane or the cyclotrimerization of a wide variety of acetylenes. Stille (36) using a rhodium catalyst prepared from 14 observed activity for the hydrogenation of benzene that increased with reuse, a phenomenon usually associated with metal crystallite formation. Rhodium catalysts of 15 and 16 used to hydroformylate octene-1 revealed a loss of carbonyl adsorptions and a loss in catalytic activity upon reuse (37). 

There are two views on the origin of enantiodifferentiation (ED) using Pt-cinchona catalyst system. In the classical approach it has been proposed that the ED takes place on the metal crystallite of sufficient size required for the adsorption of the chiral modifier, the reactant and hydrogen [8], Contrary to that the shielding effect model suggest the formation of substrate-modifier complex in the liquid phase and its hydrogenation over Pt sites [9],... 

The formation of H2, which involves electrons, would be favored by the metal crystallites... 

Taking, for instance, Al, with a melting point of 660 °C and a web substrate temperature of 50 °C, zone I formations will be created (porous structure, pointed crystallites, large voids) and up to 250 °C, formations in the transitional area (densely packed fibers) will appear. Up to 450 °C zone II (pillar-shaped crystallites), and above this temperature zone III (conglomerate-type crystallites) formations will be seen. Because of the relatively low maximum thermal stress that may be applied to polymer webs, the growth in metallized layers on polymer webs mainly occurs in Zone I or in the transitional zone. The different growth is also evident from comparison of cooling drum and free-span coater methods. 

Lopez et al. [27] prepared Pd/SiC>2 catalysts under both acidic (pH = 3) and basic (pH = 9) conditions in the sol-gel step and reported that an acid medium promotes the formation of small metal crystallites. This finding is consistent with the formation of a micro-porous silica gel network at a low pH. By comparing samples prepared by the sol-gel method and impregnation, these authors found in the former a stronger metal-support interaction which they ascribed to the square planar palladium complex used as a precursor. Finally, their results showed that the method of preparation as well as the conditions used in each method impact on how these catalysts deactivate in the hydrogenation of phenylacetylene. 

The role of tbe additives, which may be used in bi- or muitnnetaltic combinations (Re, Ir, Ru, Ag, Au, Ge etc.) is poorly umterstood. They help to enhance the properties of platinum hy keeping it in a suitable state of dispersion, and to modulate the acidic character of the support. They appear to oppose the sintering of metallic crystallites by the formation of alloys or polymetallic dusters. 

Removal of lattice oxygen from the surface of nickel oxide in vcumo at 250° or incorporation of gallium ions at the same temperature [Eq. (14)] causes the reduction of surface nickel ions into metal atoms. Nucleation of nickel crystallites leaves cationic vacancies in the surface layer of the oxide lattice. The existence of these metal crystallites was demonstrated by magnetic susceptibility measurements (33). Cationic vacancies should thus exist on the surface of all samples prepared in vacuo at 250°. However, since incorporation of lithium ions at 250° creates anionic vacancies, the probability of formation of vacancy pairs (anion and cation) increases and consequently, the number of free cationic vacancies should be low on the surface of lithiated nickel oxides. Carbon monoxide is liable to be adsorbed at room temperature on cationic vacancies and the differences in the chemisorption of this gas are related to the different number of isolated cationic vacancies on the surface of the different samples. 

Reduction of the oxide begins with some difficulty, in the absence of metal nuclei, and this accounts for the slow exothermic phenomenon whose intensity is maximum at the beginning of reduction and which results probably in the formation of metal nuclei on the oxide surface. Since the intensity of the fast exothermic phenomenon increases when the extent of reduction is larger, it must be related to a reduction process now occurring at the metal-oxide interface, carbon monoxide being adsorbed on metal crystallites. All carbon monoxide in dose G is adsorbed on the metal and reacts with nickel oxide at the metal-oxide interface since the slow exothermic phenomenon does not appear on curve G (Fig. 34). Calorimetric curves similar to curve G are obtained during the reaction of subsequent doses of carbon monoxide. Finally, it appears from curves B to G (Fig. 34) that desorption of carbon dioxide is a slower process than the adsorption of carbon monoxide and its interaction with nickel oxide. 

These observations are satisfactorily explained by a mechanism in which salt decomposition is completed through two consecutive reactions during which there is stepwise cation reduction Cu " Cu" -> Cu . The appearance of a liquid phase during acetate formation, when a is less than 0.5 and Cu " - Cu, accounts for the first acceleratory rate process which fits the autocatalytic exponential law (da/df = ka) with = 200 15 kJ mol" between 468 and 505 K. Reaction occurs in an intracrystalline fiised material but the maintenance of an outer unreactive surface prevents particle coalescence or comprehensive fusion. The product of this rate process, believed to be predominantly copper(I) malonate, then decomposes by a first order reaction for which = 188 6 kJ moT between 477 and 528 K. This probably occurs in the solid state to give a residue identified as a dispersion of copper metal crystallites on a coherent carbon substrate. Structures and properties, including the catalytic activity of this material, have been reported [124]. 

With Pd- or Pt-containing catalysts the problem arises how to discriminate between reduced Ni and the reduced metal. Temperature-programmed reduction experiments ( 5) have shown that Pd is reduced arond 80 C. Reduction of Ni starts at 200 to 300 C. Reoxidation and rereduction point to a possible Pd-Ni alloy formation. We have studied Pd-NiSMM and Pt-NiSMM samples after reduction at 350 and 450 C by TEM combined with electron microprobe analysis. Metal crystallites with a maximum diameter of 20 nm are formed. Part of them contain Pd and Pt, respectively. Because of the background of lattice Ni2+, reduced Ni is difficult to distinguish by this technique. Since, moreover, large metal crystallites are observed from which no Pd or Pt signal is obtained at all, it seems reasonable to assume that these crystallites are reduced Ni. The presence of alloys cannot be ruled out. 

---