Size-selective chemisorption

C-H Bond Activation. We have also examined the chemisorption of various hydrocarbons on different transition metal clusters. In this section we describe results obtained for methane activation on neutral clusters. First, we note that under our experimental conditions (low pressure, near room temperature, short contact time) methane activation readily occurs only on specific type and size metal clusters. For instance, we detect no evidence that methane reacts with iron(12), rhodium(26) or aluminum(2Z) clusters, whereas as shown in figure 5 strong size selective chemisorption is... 

In summary, reactions with ligands that requi re a bond to be broken before chemisorption is completed show size selective behavior. This applies to hydrogen and nitrogen and likely... 

In summary a few "generalizations" have been found. First, size selective chemistry is strongly associated with chemisorption that requi res bond-breaking. Second, metal clusters react rapidly with ligands that molecularly chemisorb even when the eventual products involve dissociation of the ligand. Dehydrogenation of Cg-alkanes on small platinum clusters take exception to this. 

The chemical composition can be measured by traditional wet and instrumental methods of analysis. Physical surface area is measured using the N2 adsorption method at liquid nitrogen temperature (BET method). Pore size is measured by Hg porosimetry for pores with diameters larger than about 3.0 nm (30 A) or for smaller pores by N2 adsorp-tion/desorption. Active catalytic surface area is measured by selective chemisorption techniques or by x-ray diffraction (XRD) line broadening. The morphology of the carrier is viewed by electron microscopy or its crystal structure by XRD. The active component can also be measured by XRD but there are certain limitations once its particle size is smaller than about 3.5 nm (35 A). For small crystallites transmission electron microscopy (TEM) is most often used. The location of active components or poisons within the catalyst is determined by electron microprobe. Surface contamination is observed directly by x-ray photoelectron spectroscopy (XPS). 

Selective chemisorption methods have been used with success for the determination of metal surface area and particle size in supported catalysts, and for titration of acid sites on silica-alumina and zeolite catalysts. The chemisorption methods are sometimes neglected in the quest for a more physical description of the catalyst surface, possibly with the penalty of missing an important and quantitative piece of information about the catalyst surface. 

For supported metal catalysts, no simple calculation is possible. A direct measurement of the metal crystallite size or a titration of surface metal atoms is required (see Example 1.3.1). TWo common methods to estimate the size of supported crystallites are transmission electron microscopy and X-ray diffraction line broadening analysis. Transmission electron microscopy is excellent for imaging the crystallites, as illustrated in Figure 5.1.5. However, depending on the contrast difference with the support, very small crystallites may not be detected. X-ray diffraction is usually ineffective for estimating the size of very small particles, smaller than about 2 nm. Perhaps the most common method for measuring the number density of exposed metal atoms is selective chemisorption of a probe molecule like H2, CO, or O2. 

Another way to change concentration of active material is to modify the catalyst loading on an inert support. For example, the number of supported transition metal particles on a microporous support like alumina or silica can easily be varied during catalyst preparation. As discussed in the previous chapter, selective chemisorption of small molecules like dihydrogen, dioxygen, or carbon monoxide can be used to measure the fraction of exposed metal atoms, or dispersion. If the turnover frequency is independent of metal loading on catalysts with identical metal dispersion, then the observed rate is free of artifacts from transport limitations. The metal particles on the support need to be the same size on the different catalysts to ensure that any observed differences in rate are attributable to transport phenomena instead of structure sensitivity of the reaction. 

Nickel reacts readily with hydrogen, but without significant cluster-size selectivity.Robata et al. found that hydrogen chemisorption on Ni(lll)... 

The chemisorption of CO onto 12 different transition metal clusters °° containing more than a few atoms is facile and exhibits little evidence of the dramatic size-sensitive behavior observed for chemisorption of Hj or of N2, even though Nj and CO are isoelectronic. Whether reactions are observed for the atom and the smaller clusters depends on the metal. Similarly the chemisorption reactions of CO with the cluster ions Nb and COn " exhibit little size-selective behavior. Chemisorption is faster onto the cluster ions than onto the respective neutral clusters. Aluminum is the only metal examined thus far for which CO chemisorption is significantly cluster-size sensitive, with Alg being the most reactive cluster. " ... 

Iron clusters exhibit facile chemisorption toward methanol, the reaction proceeding with little or no cluster-size selectivity. An interesting feature of this system is that the chemisorption rate constants are nearly identical toward various isotopic sjjecies (CH3OH, CH30D,CD3 0H). If dissociation of a C—H or O—H bond was the initial step, then this should be manifested in an observable kinetic isotope effect. Thus the initial chemisorption step most likely involves the lone-pair orbital localized on the oxygen atom. More extensive studies of the chemistry of the Fe methanol system have been explored using infrared multiple-photon dissociation spectroscopy. These results are discussed in detail in Section Vlll. 

Table 1 summarizes the information required for a detailed characterization of a supported metal catalyst for supported bimetallics there are additional questions, e.g., the distribution of atoms in bimetallic clusters and the surface composition of larger alloy crystallites. For the support and the prepared catalyst, the total surface area, pore size distribution, and surface acidity are routinely measured, if required, while other characteristics, e.g., thermal and chemical stability, will have been assessed when selecting the support. The surface structure of alumina, silica, charcoal, and other adsorbents used as catalyst supports has been reviewed. Undoubtedly, the most commonly measured property is the metal dispersion, often expressed in terms of the specific metal area and determined by selective chemisorption or titration but, as discussed (Section 2), there is the recurring problem of deciding the correct adsorption stoicheiometry. 

Metal Dispersion by Chemisorption and Titration Selective Chemisorption. - This is the most frequently used technique for determining the metal area in a supported catalyst and depends on finding conditions under which the gas will chemisorb to monolayer coverage on the metal but to a negligible extent on the support. Various experimental methods, conditions, and adsorbates have been tried and studies made of catalyst pre-treatment and adsorption stoicheiometry, viz, the (surface metal atom)/(gas adsorbate) ratio, written here as Pts/H, Bh jQO,etc., and reviews to about 1975 are available. A summary is given in Table IV of ref. 2 of methods used to confirm the various adsorption stoicheiometries proposed, sometimes from infrared studies. These include chemisorption on metal powders of known BET area or, more satisfactorily, one of the instrumental methods reviewed in Section 3 for the determination of crystallite size distributions. For many purposes, a relative measurement of metal dispersion is sufficient, conveniently expressed as the ratio (number of atoms or molecules adsorbed)/(totfl/ number of metal atoms in the catalyst), e.g., H/Ptt. 

The characterization of supported metal catalysts is a matter of some complexity and supported bimetallic catalysts even more so. Nevertheless the development and application of methods for determining catalyst structure is essential for an understanding of why the performance of a selected combination of metal(s) and support varies as a function of preparative variable, activation procedure, reaction conditions, or time. Although some aspects of catalyst structure can be routinely determined, the basic measurement of absolute metal dispersion by selective chemisorption/gas titration is still the subject of many publications and the necessity of cross-checking by instrumental methods is generally appreciated. The characterization of supported metal catalysts also involves some less accessible properties, e.g., the sites available on crystallites as a function of size, high-temperature... 

Boreskov et al. (4, 5) were the first to complete a systematic investigation of the relationship between particle size and catalytic activity, after their development of a technique for measuring the surface area of platinum catalysts by means of the selective chemisorption of hydrogen (6). They showed that the specific activity of platinum in the oxidation of sulfur dioxide (4) and of hydrogen (5) varied by less than one order of magnitude for catalyst samples differing in platinum surface area by four orders of magnitude. A few years later, Kobosev s ideas were further... 

Effect of particle size on turnover rate for ammonia sjmthesis. Small particles of metallic iron supported on magnesia were prepared by Boudart et The iron particle size could be changed between 1.5 run and 30 run and determined, in part, by electron microscopy. X-ray diffraction, magnetic susceptibility and Moss-bauer spectroscopy. Agreement was satisfactory with particle size values obtained by selective chemisorption of carbon monoxide (if 2 Fe for 1 CO). Two results are noteworthy, (i) The turnover rate for ammonia synthesis increases by a factor of 35 as the iron particle size increases (Table 2.12). (ii) A pretreatment of the iron catalyst with ammonia increases the turnover rate by only 10% for the larger particles, but quite appreciable for iron clusters (Table 2.13). 

---