Some Comments About the Experimental Results

Our intention is to examine the experimental results described in Section II in the light of the general ideas just developed. For two reasons, a limited discussion of the experimental facts should be made. First, the considerations put forward in Section IV do not yet constitute a real theory second, the available experimental results are still incomplete and allow only of a partial justification of the theoretical ideas. 

With regard to thermal decomposition on semiconductors and insulators, the following mechanisms are generally proposed  

The adsorption steps (1) and (la) are generally more rapid than reactions (2) and (2a) which are favored by the presence of positive holes. For this reason, semiconductors of the p-type are generally better catalysts than insulators, whereas n-type semiconductors are the least eflacient ones. 

Comparison of the two elementary processes (2) and (3) shows the latter to be the most endothermic one. Indeed, decomposition of N2O yielding NO and N entails the rupture of the N=N bond, which requires 85 kcal. mole (3.6 e.v. molecule ) whereas for N2O decomposition into Na and 0, the breaking of the NO bond requires only 38 kcal. mole (1.6 e.v. molecule 0-This explains that thermally, in the absence of radiation, reaction (2) is always considerably favored, compared to reaction (3). Under irradiation, however, a great number of excess free carriers are produced in the conduction and valency bands these carriers tend to recombine. In this respect the adsorbed N2O molecule may behave like a recombination center. This phenomenon can be accounted for by considering the adsorbed N2O molecule to be an acceptor level. Under this hypothesis, the N2O chemisorption results from the capture, by the weakly adsorbed molecule, of an electron from the conduction band. At the moment of recombination with a positive hole from the valency band, a variable amount of energy can be recovered, depending on the position of the level constituted by the adsorbed N2O molecule. For the silica and alumina we have utilized, the width of the forbidden region is about 10 e.v. process (3) which only requires 3.6 e.v. may thus become possible. 

In conclusion, the existence of an energy transfer from solid to gas seems to us clearly established in the case of N2O radiolysis but the mechanism of this transfer cannot, as yet, be determined. 

---