Choice of Adsorbate
The choice of adsorbate is critical in a TPD experiment. The selected gas should chemisorb selectively on the metal without sorbing on other catalytic components or on the support. Additionally, it should easily form monolayer coverage and not react irreversibly with either the metal surface or the support. An example of the latter is the interaction of CO with nickel metal in which volatile (and poisonous!) nickel carbonyl-Ni(CO)4 - can form.
The interaction of the adsorbate with the metal during a TPD experiment can be viewed in terms of potential energy diagrams, as depicted in Figure 1. The first step in the experiment involves the collision of an adsorbate molecule with the surface and subsequent physisorption. This process involves a small activation barrier (E1), and energy loss (H1), a few kcal/mole. The physisorbed molecule then transfers to a chemisorbed state. This transfer requires the adsorbate to overcome a second activation energy barrier which may be small (E2) or large (E3). It should be noted that the heat of chemi- sorption, depicted as H2 in Figure 1, is independent of the activation energy barrier. In cases where the activation energy barrier is large, the chemisorption process is said to be activated, meaning that it will be kinetically slow, thus requiring either long equilibration times, higher adsorption temperatures, or possibly both, for complete surface coverage. In choosing an adsorbate, it is important to recognize those that may exhibit activated chemisorption and to adjust the temperature or time of chemisorption accordingly. An example of activated chemisorption is that of H2 on supported cobalt metal. Figure 2 shows TPDs of H2 from such a catalyst for two sets of adsorption conditions. Curve A shows a TPD pattern in which the adsorption was conducted at room temperature. Only a small desorption signal was observed since few chemisorption sites were filled. In contrast, curve B shows a similar TPD in which the adsorption was carried out at an elevated temperature and the sample then cooled in hydrogen. Such a procedure allowed all the chemisorbing sites to fill and resulted in a meaningful TPD pattern.
Unfortunately, there is little theoretical basis for choosing an adsorbate and, thus, one must rely on previous experience and the literature. Table 1 summarizes some suitable adsorbates of for a variety of transition metals. This list is by no means all-inclusive and does not take into account the effect of supports or additives to the catalyst, which can markedly affect their chemisorptive properties. As such, this table should be considered a starting point for choosing an adsorbate.
Choice of Adsorption Conditions
The choice of adsorption conditions is just as important as the choice of adsorbate. Adsorption should be carried out at temperature sufficiently high and for a period of time sufficiently long to ensure complete surface coverage. At the same time one must prevent the reaction of certain adsorbates with the surface, such as the disproportionation of CO to CO2 and C, which can occur at elevated temperatures. An additional possible complication is the occurrence of spillover in which a selective adsorbate can chemisorb on a metal crystallite and move onto the support (see Figure 3). Spillover is a kinetically slow process and is primarily observed at high temperatures. Thus, the adsorption temperature and time must be optimized to ensure complete coverage without the on set of spillover. As a rule of thumb, one can start with adsorption temperatures of 100-200°C and adsorption times of 30-90 minutes. Test experiments, however, must be conducted to ascertain that these conditions are suitable. A flushing step, at low temperature, is typically used after adsorption in order to clean the surface of any weakly-held species.
Adsorbate Stoichiometry
In order to effectively interpret the results of a TPD experiment, it is necessary to know the adsorbate-metal stoichiometry. Adsorption stoichiometries cannot be directly obtained from a simple TPD experiment of a powder catalyst, but can be obtained by comparison of chemisorption surface areas4 with BET surface area for a metal powder or metal foil, determined in some cases by interpretation of IR spectra, or by independent measurement of metal crystallite size using techniques such as TEM or XRD. Again, some rules-of-thumb can be applied to Stoichiometric factors. Hydrogen and oxygen will typically adsorb dissociatively, i.e. a H2 (or O2) molecule will dissociate and form an H-M (or 0-M) adsorbate-adsorbent pair. Thus, its stoichiometry is adsorbate/metal=0.5. There are some instances where oxygen will adsorb associatively 02/M=1, particularly at low temperatures. For Rh and Ir, it has been reported that the hydrogen adsorption stoichiometry can vary with crystallite size and, for very small crystallites, stoichiometries of H2/M=1 have been reported. These, however, are rare cases.
The stoichiometries for CO adsorption are less well-defined. CO can adsorb with varying stoichiometries depending on the metal and the crystallite size. CO/M Stoichiometries between 0.5 and 2 are routinely reported. In such cases it is sometimes best to assume an average stoichiometry (say, 1) or simply use the CO uptake as a basis of comparison between catalysts.
