The Chemisorption of Oxygen

Whilst hydrogen enters into a chemisorptive bond with charcoal at very low temperatures, oxygen remains physically adsorbed unless relatively high temperatures are reached. At liquid-air temperatures the adsorption entropy of oxygens shows that the adsorbed molecules are completely free to move and rotate over the surface 168). 

O7 ions are also formed, as a first step, in electrical reduction processes where 02 is reduced at a cathode 170). 

It is quite reasonable to assume, therefore, that oxygen can be chemi- 

As already stated above, a cesium-metal surface, when exposed to oxygen at liquid-air temperatures, is spontaneously covered with a chemisorbed layer of oxygen. As the fully oxidized product which is formed by oxidation at higher temperatures proves to be cesium superoxide (Cs02), we may assume that the chemisorbed layer formed at — 180°C. also consists of chemisorbed 02 ions. As the work function of cesium has a low value, the difference in energy between levels A and D in Fig. 18 is rather small and, consequently, minimum E of the chemisorption curve will be appreciably lower than level A. 

The formation of some organic hydroperoxides by oxidation with molecular oxygen is catalytically promoted by metals like silver or copper 171). A dissociative chemisorption of oxygen cannot be active in these processes they probably proceed via the chemisorption of O7 ions (or O2 molecules forming a covalent bond resonating with an ionic bond). 

---

Figure 4.5 STM images (6.2 x 6.5 nm) observed in the chemisorption of oxygen at Ni(110) at room temperature (a) the (3 x 1)0 state at 0 = 0.33 (b) the (2 x 1)0 state at 0 = 0.5 (c) the (3 x 1)0 state at 0 = 0.66. Corresponding ball models of these are shown in (d), (e) and (f) and are typical of oxygen-induced reconstructions at metal surfaces. The small black balls represent the O adatoms. (Reproduced from Ref. 12).

From thermodynamic considerations it is evident that bulk nickel cannot be oxidized by CO2. However, it is not justified to conclude from this that dissociative chemisorption of CO2 will not occur. Consider, for example, the chemisorption of oxygen or hydrogen which on several metals takes place under conditions where bulk oxides or hydrides are not at all thermodynamically stable. Dissociative adsorption of CO2 has indeed been observed by Eischens and Pliskin (35). 

Gold forms a continuous series of solid solutions with palladium, and there is no evidence for the existence of a miscibility gap. Also, the catalytic properties of the component metals are very different, and for these reasons the Pd-Au alloys have been popular in studies of the electronic factor in catalysis. The well-known paper by Couper and Eley (127) remains the most clearly defined example of a correlation between catalytic activity and the filling of d-band vacancies. The apparent activation energy for the ortho-parahydrogen conversion over Pd-Au wires wras constant on Pd and the Pd-rich alloys, but increased abruptly at 60% Au, at which composition d-band vacancies were considered to be just filled. Subsequently, Eley, with various collaborators, has studied a number of other reactions over the same alloy wires, e.g., formic acid decomposition 128), CO oxidation 129), and N20 decomposition ISO). These results, and the extent to which they support the d-band theory, have been reviewed by Eley (1). We shall confine our attention here to the chemisorption of oxygen and the decomposition of formic acid, winch have been studied on Pd-Au alloy films. 

Chuvylkin et al. (54) have used this approach to discuss EPR signals arising from weak R02 surface complexes in a number of systems where the g tensor does not fit the pattern expected [Eq. (6) and Fig. 3] from the ionic model. This is not discussed quantitatively, but they conclude that the appearance of covalently bonded oxygen is impossible without a favorable orientation of appropriate electronic orbitals. A similar covalent bonding approach has been considered theoretically for the chemisorption of oxygen on silicon surfaces (55). Examples of weakly bonded oxygen are given in Section IV,E. 

Sato and Akamatu (139) report that alkali metals enhance the chemisorption of oxygen on carbon and weaken the carbon-carbon bonds at the surface so as to accelerate combustion. On the other hand, they report that phosphorus, while catalyzing the adsorption of oxygen on carbon, has a retarding effect on the release of the surface oxide. 

By considering the chemisorption of oxygen, the role of interchange of electrons between the gas and solid will become increasingly evident. 

Equations (30) and (31) could be checked by measurements of the kinetics of the chemisorption of oxygen and hydrogen on oxides. However, we shall first apply these correlations to the kinetics of the chemisorption of oxygen on NiO, and check the results obtained with the experimental observations made by several authors. 

Garner and associates (71-73) investigated the oxidation of CO on crystalline CU2O and also measured the chemisorption of oxygen and... 

It must be emphasized again that these discussions on the conductivity of zinc oxide have been based on the two assumptions (1) that the chemisorption of oxygen has a strong effect on the resistance of the zinc oxide and (2) that the rate-determining step in the chemisorption of oxygen is an electron transfer. The former assumption has been based on indirect arguments, as well as on some direct experimental evidence. The latter assumption is based on the ability of the model to explain in a simple consistent manner the various experimental effects, some of which have been described, and some of which will be, in later chapters. 

The frozen in photoconductivity, as was concluded by Melnick, will arise from effects of the surface barrier layer or, of course, would arise similarly from any other rate-limiting process in the adsorption of oxygen. For our model in this discussion we shall use electron transfer over the surface barrier as the rate-limiting reaction. In this case, the rate at which adsorption occurs is proportional to exp ( —Ei/kT), where E2 is the barrier height. Thus if we measure the decay in photoconductivity (the chemisorption of oxygen) at room temperature, and then suddenly quench the sample to 130°K, it is obvious that the rate of decay in photoconductivity will decrease considerably. The change in the rate will be dependent on Ei and the temperature to which the sample is quenched. 

A good example for sulfides is provided in the chemisorption of oxygen on the edge surface of MoS2. If site densities are calculated using one oxygen atom per molybdenum on the edge, the densities are inexact, particu-... 

Oxidation can be viewed as the chemisorption of oxygen. For example, nickel and silicon are oxidized at ambient conditions. The resulting oxide layer is thermodynamically more stable and passivates the pure material below it. Another important example is the oxidation of aluminum which provides the metal with a very hard roughly 100 nm thick aluminum oxide (AI2O3) layer. To stabilize the aluminum surface even more and to passivate it against reactive chemicals the thickness of the oxide layer can be increased electrochemically. This procedure is called the eloxal process (efectrolytical oxidation of a/uminum). 

Oxygen atoms, chemisorbed on tungsten or molybdenum, seem to prefer 100 faces 293). In the chemisorption of oxygen the metal has to provide the binding electrons, and it might be expected that a low value of the work function would favor the process. Steric factors, however, may also play their role here. 

There are some indications in literature, however, that may endorse the non-dependence. In Sec. VII,6 we learned that the chemisorption of oxygen on iron was not restricted to the surface only iron atoms (ions) moving on top of the chemisorbed oxygen atoms (ions) bring about an... 

Hall effect measurements enable the number of electrons in the conduction band of ZnO to be established during the chemisorption of oxygen. Though this investigation has been carried out in the dark only, the results are very important in this context 26>. Hall effect measurements may be carried out with polycrystalline sintered powders. They are not disturbed by grain boundaries, narrow connecting necks, pores and cavities, as are conductivity measurements. 

It seems very probable that the chemisorption of oxygen and carbon monoxide diagnoses the number of surface sites, (Cr3+) (cus), but the degree to which n is 1 or 2 is uncertain. At lower temperatures of activation, chemisorption of carbon dioxide probably diagnoses the number of strongly basic sites, 02-(cus) or OH (cus). The reliability of the second conclusion is less than that of the first. We suggest that the catalytic reactions which we have studied use various eombinations of these sites. 

---