Search

Recently a model based on an attractive square well and impulsive repulsive potential between a gas and a solid surface has been employed to estimate the fraction of gas atoms trapped in the potential well at the solid surface. A gas atom was considered trapped if it transferred enough energy to the solid (as judged by the thermal accommodation coefficient) to fall into the potential well, i.e., to have negative total energy where the zero of energy is assumed to be the infinitely separated gas atom and solid with the gas atom at rest. The criterion for gas trapping when considering experimental scattering data has been the assumption that the trapped gas is emitted from the surface with a random (cosine) distribution. To assess the possibility that rare-gas atoms may fall into the interatomic well with the metal surface (i.e., be trapped according to the definition of Ref. 1) and still be re-emitted from the surface in a nonrandom distribution, some estimates of surface residence times for the gas should be made. If such a phenomenon as suggested above were to occur, then the calculated trapping probability of the square well model would be expected to be larger than those found experimentally, as has been observed to be the case for krypton and xenon interactions with a tungsten (110) surface. The idea that a gas particle may execute several vibrations at a surface and still not reach thermodynamic equilibrium is not a new one. In this paper the pertinent theory describing this effect is presented, and empirical calculations estimating the magnitude of the effect are presented.

In a recent paper (1) Mueller and Eyring have discussed the normal state of the hydrogen molecule with use of a variation function constructed in the usual Heitler-London manner, except with replacement, for each electron, of the hydrogen-atom function about atom a, exp(-Z'ra/ao), byt the more complex function φa = exp(-Z'ra/ao)exp(-Z''rb/a0). The authors call one-electron functions of this sort semilocalized orbitals. They have compared the value of the bond energy calculated for the wave function [...], with those given by the simple atomic orbital method and the molecular orbital method, and have pointed out that there is an improvement to 4.20 ev, from the values 3.76 ev and 3.60 ev, respectively

In our discussions of the electronic structure of intermetallic compounds during the last three years brief mention has been made from time to time of the phenomenon of electron transfer [1-3]. The interpretation of the observed interatomic distances in many metallic phases seems to require the assumption that electron transfer has taken place; an example is Al9Co2, for which the distances [4] support the charge distribution [5] Al-2/99Co2+1. The indication by physical properties of a filled-Brillouin-zone structure for FerZn21 also led to the suggestion of electron transfer [2]: about 1/4 electron is indicated to have been removed from each zinc atom and about one electron added to each iron atom. In this paper it is pointed out that the analysis of interatomic distances shows that electron transfer takes place in a great many interatomic compounds, and that the numbers of electrons involved are reasonable, in relation to the changes in valence resulting from loss or gain of electrons and to the partial ionic character of the bonds between unlike atoms and the striving of atoms toward electroneutrality.

When monoenergetic atoms are continuously introduced into a thermal gas, they can undergo deactivating, activating, and reactive collisions. The net result of such collisions is to establish a steady-state distribution of laboratory energies which, although not as sharp as the initial distribution, preserves some of its features, such as being centred at about the initial energy. The reactive collisions which occur under these conditions are characterized by the associated relative energy distribution function and the energy-dependent reaction cross section. As a result, as the initial laboratory energy of the atoms is experimentally varied, the relative energy distribution function can be made to sample appropriately the reaction cross section curve. Therefore, from measurements of the competition between reaction and thermalization processes as a function of initial atom laboratory energies, and from a knowledge of the non-reactive differential scattering cross section, it is possible to obtain information about the dependence on relative energy of the rotationally averaged reaction cross section. The appropriate Boltzmann steady-state equation needed to obtain this information is derived in this paper and solved for an assumed set of reactive and non-reactive cross sections. Distribution functions of relative energies are thereby obtained and used to indicate the usefulness of the suggested measurements.