A partial model for the structure of anodized amorphous Al2O3

Abstract

The anodic Al2O3 films that were used in this study were obtained by anodizing commercially pure (2S) aluminum in a sulphuric acid electrolyte bath. Four different types of samples of this anodic film were analyzed: (1) powder samples which were made by grinding the oxide film until it was fine enough to pass through a 400 mesh screen and pressing it into a one inch disc in a hydraulic press; (2) heated powder samples which were prepared in the same manner as the powder samples except that they were heated to various temperatures; (3) and (4) samples which were oriented so that x-ray scattering data could be gotten from the front and edge respectively of the anodic Al2O3 films. The x-ray scattering data was taken using a bent crystal monochromator and point counting techniques with both CuK-alpha and MoK-alpha radiation. X-ray scattering data was taken for all four kinds of samples. The experimental x-ray scattering data was corrected for: (1) polarization, (2) sample penetration by beam, (3) multiple scattering, (4) normalization, and (5) termination of data effects. The x-ray data was then Fourier inverted and analyzed using the approximate atomic radial distribution method of Warren as well as the pair function and electronic radial distribution methods. This analysis indicated that the amorphous anodic films were isotropic. In addition, by examining the radial distributions for the nonheated amorphous film samples, it was concluded that each aluminum atom had 4.54 nearest oxygen atoms in either a tetrahedral or octahedral arrangement and therefore it is concluded that the amorphous film has three tetrahedra for each octahedron. Charge balance is obtained for this system by connecting two tetrahedra to each corner of each octahedron. Analysis of the heated powder samples indicated that the amorphous Al2O3 film had converted to gamma-Al2O3 at a temperature of 850°C and to corrundum by 1200°C. Since the gamma-Al2O3 has a cubic close packed oxygen structure as opposed to the hexagonal close packed structure of corrundum, one is led to conclude the amorphous film probably has an essentially cubic close packed structure. Using all of the above data, one finds this is adequate to build a short range model but a long range structure does not appear feasible using these criteria.