Carbon in Crystalline Silicon

  • PDF / 262,814 Bytes
  • 2 Pages / 604.8 x 806.4 pts Page_size
  • 9 Downloads / 177 Views



Formation of Icosahedral Aluminum-Manganese by Electron, Laser and Ion Beams /. A. Knapp and D. M. Follstaedt Sandia National Laboratories Albuquerque, NM 87185, USA Last year at the National Bureau of Standards, examination of certain aluminummanganese alloys which had been melted and the rapidly cooled on a spinning copper wheel showed that the atomic arrangement of the resulting solid exhibited long-range order with icosahedral orientational symmetry. This means, for example, that certain directions show five-fold rotational symmetry in electron diffraction patterns. The discovery is surprising because five-fold rotational symmetry is impossible for a crystalline material, and all previously examined solids with long-range order have been found to be crystalline. The problem is illustrated by an analogy in two dimensions: a floor cannot be tiled using a periodic array of pentagons (fivefold symmetry) without either overlapping tiles or leaving gaps. Naturally, the observation of an ordered material exhibiting icosahedral symmetry has generated intense interest. Currently, many workers interpret the experimental observations in terms of a "quasicrystalline" atomic arrangement. Again, the analogy of tiling a floor is helpful: a floor can be covered using two differently shaped tiles in a non-periodic array such that the overall symmetry of the pattern is five-fold. The pattern does not repeat, so no two areas of the floor have exactly the same local arrangement of tiles. Such a pattern is called a "Penrose tiling," after its inventor. A "quasicrystal" would be a three-dimensional Penrose tiling, where two different groupings of atoms are arranged in a non-periodic array to

form the solid. Such a solid would be able to diffract electrons like a crystal, but symmetries such as icosahedral would be allowed. If this explanation is correct, quasicrystals are an entirely new class of ordered materials. In order to determine the properties of the "icosahedral phase" and to explore other techniques which might lead to its formation, we have examined thin surface layers of aluminum-manganese produced by surface alloying techniques. First, alternating aluminum and manganese layers are deposited onto an aluminum or iron substrate to a total thickness of 50-100 nm. This surface layer is then treated with an electron beam, a laser beam, or a highenergy ion beam to form the icosahedral phase. If a laser or fast electron beam treatment is used, the surface layer is melted and then frozen very quickly (within less than one millionth of a second); this treatment produces a layer of small icosahedral particles, typically 30 nm in diameter. If a slower electron beam treatment is used, and the substrate is iron, the molten layer cools more slowly and grains of the icosahedral phase up to a micrometer in diameter are formed. Although generally similar to other liquid quenching techniques which have been used to form the icosahedral phase, the laser and electron beam treatments have well-known temperature histories which allow us to pla