Poisoning experiments

Studies of the IR spectra of adsorbed hydrogen as a function of catalyst oxidation state showed it had little effect in the intensity of the IR bands. 

Infrared studies show that when water is adsorbed on the surface, the background intensity in the hydroxyl region increases new bands may appear but hydrogen-bonding effects make such conclusions uncertain. If such a catalyst is then exposed to hydrogen (or deuterium), no bands due to adsorbed hydrogen (or deuterium) are observed. Thus, adsorption of water apparently occurs on the active sites and blocks out type I chemisorption. 

2 Although oxygon poisoning of hydrogenation at room temperature appears to 

Over zinc oxide it is clear that only a limited number of sites are capable of type I hydrogen adsorption. This adsorption on a Zn—O pair site is rapid with a half-time of less than 1 min hence, it is fast enough so that H2-D2 equilibration (half-time 8 min) can readily occur via type I adsorption. If the active sites were clustered, one might expect the reaction of ethylene with H2-D2 mixtures to yield results similar to those obtained for the corresponding reaction with butyne-2 over palladium That is, despite the clean dideutero addition of deuterium to ethylene, the eth- 

6 Analysis showed that equilibration was not complete. The initial composition was (H2-HD-D2) 34 36 30. After the run, the composition was 33 40 27. 

---

The nature of such sites seems consistent with the behavior shown In the pyridine and lutldlne poisoning experiments. The acidic nature of the reduced metal sites which hold the nitrogen bases seems established. Difference In the extent to which the exposed metal cation Is accessible to the nitrogen atom of the organic nitrogen base could explain the selective poisoning seen with different substituted pyrldlnes. Presumably the cations In an active pair are somewhat less accessible than most exposed Co and Mo cations, which, because they normally hold two MO molecules, are probably exposed In Incomplete tetrahedral sites. 

The essential features of the above picture are shown schematically in Fig. 6. Hydrogen or deuterium adsorbs and desorbs rapidly on these sites (half-time less than 1 min) hence, the slower H2-D2 equilibration (half-time about 8 min) appears to be determined by the site-to-site migration required for exchange. Poisoning experiments show that water also prefers these sites in agreement with the limited IR data, Fig. 6 shows the adsorbed water yields surface hydroxyls (10). 

The organoactinide surface complexes exhibited catalytic activities comparable to Pt supported on sihca [at 100% propylene conversion at —63°C, >0.47s (U) and >0.40 s (Th)], despite there being only a few active sites (circa 4% for Th, as determined by CO poisoning experiments and NMR spectroscopy) [92]. Cationic organoactinide surface complexes [Cp An(CH3 ) ] were proposed as catalytic sites. This hypothesis could be corroborated by the use of alkoxo/hydrido instead of alkyl/hydrido surface ligands, which led to a marked decrease of the catalytic activity, owing to the oxophilic nature of the early actinides [203, 204]. Thermal activation of the immobihzed complexes, support effects, different metal/ligand environments and different olefins were also studied. The initial rate of propylene conversion was increased two-fold when the activation temperature of the surface complexes under H2 was raised from 0 to 150°C (for Th 0.58 0.92 s" ). 

The experiments with reversible poisoning of alumina by small amounts of bases like ammonia, pyridine or piperidine revealed [8,137,142,145, 146] relatively small decreases of dehydration activity, in contrast to isomerisation activity which was fully supressed. It was concluded that the dehydration requires only moderately strong acidic sites on which weak bases are not adsorbed, and that, therefore, Lewis-type sites do not play an important role with alumina. However, pyridine stops the dehydration of tert-butanol on silica—alumina [8]. Later, poisoning experiments with acetic acid [143] and tetracyanoethylene [8] have shown the importance of basic sites for ether formation, but, surprisingly, the formation of olefins was unaffected. 

Poisoning experiments usually show very selective adsorption. As the first increments of poison are added, activity declines rapidly. Then with later increments the effect of the poison is less severe, probably because the adsorption is less discriminating between active and inactive chromium. This suggests that the active site population is that portion of the reduced chromium which, due to surface heterogeneity, is most coordinatively insaturated. Presumably this would correspond to some of the CrA population identified by the Turin group, which under favorable conditions comprised almost half the total chromium (76, 40). 

Poisoning experiments. Poisoning of acidic sites by adsorption of bases is another technique that has been applied to identify the centers of catalytic activity for various reactions. The results of many of the earlier studies were interpreted in terms of interaction of the adsorbed bases with... 

Na+-poisoning experiments. The presence of a larger number of weakly acidic protons is also consistent with the infrared results. 

After the type and strength of interaction of a potential poison with the catalyst surface has been studied and the number of adsorption sites estimated, its effect on the rate of a given catalytic reaction can be studied. Any kind of catalytic reactor may in principle be used for these studies, that is, static as well as dynamic methods are suitable, and the various forms of pulse techniques are applicable. The real distinction between the two types of poisoning experiments that have been performed lies in the fact that the poison is either fed together with the reactant and is present in the gas phase throughout the run or the surface is poisoned by irreversibly preadsorbing the poison while the gas phase is kept free of it. If the poison is present in the gas phase, a larger number of modes of interaction with different surface sites may be possible than for the... 

Ammonia seems to be too strong a base if specific adsorption is required. A characterization of the chemical nature and a determination of the number of catalytically active sites by means of poisoning experiments with ammonia will, therefore, not readily be possible. Ammonia can thus not be recommended as a simply acting specific poison. Conclusive results may, however, be obtained by stepwise poisoning, adding successive small quantities of ammonia, provided that the modes of interaction with the catalyst of this ammonia are controlled by spectroscopic techniques under the reaction conditions. 

PyL. Infrared spectroscopy in combination with poisoning experiments is a recommendable technique and has already been proved to permit meaningful conclusions to be drawn on the nature of active sites and reaction intermediates (47,226,230) (see also Section V). 

It thus follows from the foregoing discussion that C02 adsorption on alumina is a key compound for the study of the chemical nature of a variety of distinct surface sites. However, it is also apparent that unambiguous conclusions can hardly be drawn from C02-poisoning experiments due to manifold, simultaneously existing surface species. 

Nevertheless, C02 is an extremely valuable probe molecule because the infrared spectra of the chemisorbed species respond very sensitively to their environments. Thus, the frequency separation of the typical band pairs of the carbonate structures may be taken as a measure of the local asymmetry at the chemisorption site. The application of 13C-FT-NMR should be extremely valuable for a still more extensive study of the nature of sites by C02 adsorption. Due to the very detailed information on the structure of sites on oxide surfaces that can be obtained by C02 chemisorption studies, this compound should in some cases also be applicable as a specific poison. A very careful study of the type of interaction with the surface, however, has to be undertaken for each particular system before any conclusive interpretation of poisoning experiments becomes meaningful. 

Two reactions for which specific poisoning experiments have contributed to the elucidation of the reaction mechanisms and permit evaluation of the possibilities and pitfalls of the technique are discussed as examples in this section. The first example is the dehydration of alcohols on alumina catalysts, and the second, the isomerization of olefins on the same type of catalyst. 

Early poisoning experiments using nitrogen bases such as ammonia, pyridine, and piperdine have shown that the secondary isomerization of the primary ole-finic products can be completely suppressed, whereas the dehydration activity of the alumina catalyst was only slightly influenced by these poisons (30, 31,341-344). This is a typical example of selective poisoning, where a consecutive reac-... 

Poisoning experiments with varying amounts of preadsorbed pyridine have recently been carried out by KnOzinger and Stolz (47). Pyridine is solely held by Lewis acid sites under the experimental conditions as shown by infrared spectroscopy. The rate of isobutylene formation from f-butanol was essentially independent of the degree of poisoning, and the true activation energy of the reaction remained constant at 25 kcal/mole, when the number of preadsorbed pyridine molecules varied between 3 and 9X 10n/m2. It thus, appears that Lewis sites which retain pyridine at temperatures between 550° and 150°C, respectively, do not interfere in this reaction. 

Poisoning experiments have thus shed some light on the chemical nature of the... 

The existence of various types of active sites all involved in the isomerization of olefins-their creation depending on the activation temperature—has already been postulated (365, 373, 379). Recent poisoning experiments with pyridine on 77-AI2O3 (activated at 500° and 600°C) seem to give some evidence for the existence of at least two chemically distinct active sites. Their true nature is still obscure (378), but that they are blocked by pyridine indicates that Lewis sites should be involved. It had been shown that the 77-AI2O3 was deactivated... 

---