Atomic Structure of Gold and Copper Boundaries

  • PDF / 4,703,647 Bytes
  • 6 Pages / 417.6 x 639 pts Page_size
  • 14 Downloads / 272 Views

DOWNLOAD

REPORT


The technique for preparing the boundaries studied in this work involves depositing gold (or copper) on a single crystal semiconductor wafer [I]. More than one orientation may grow if the epitaxial alignment of the thin film on its substrate has multiple variants [2]. The resulting thin film will then form as a polycrystal and the maze of boundaries separating the crystals may become perpendicular to the plane of the thin film on annealing. The imaging of these films is, on the one hand, routine in that it requires only straightforward HREM. On the other hand, the resolution required for these particular specimens is closer to IA than 2A and that is possible only on machines built very recently. This paper describes an extension to previous work on the films [3], including the observation of a new interface type in gold, and the first HREM observation of a copper bicrystal down the [Ill] direction. EXPERIMENTAL The preparation method of the I1111 gold films and the TEM specimens was covered in detail in [3]. In brief, gold was deposited onto a germanium [111] substrate which was later removed chemically, leaving the film to be floated off onto a fine-mesh grid. The key element in this work was the preparation of a gold film thin enough to have the facets extend from the top surface to the bottom surface, without diverting into the in-plane {1111 twin boundary.

359 Mat. Res. Soc. Symp. Proc. Vol. 589 © 2001 Materials Research Society

The nominal thickness of the film was 8nm, in contrast to a thickness of 30nm in the films studied in the earlier work. The film thickness was estimated by tilting a foil to 45° and measuring the projected width of the boundaries confirming a figure of around 8nm or less. The thickness, of course, may vary across the film. Estimates of film thickness may also be made from the [111] HREM images if we measure the projected width of inclined interfaces and make assumptions about the identity of the inclined plane. The rotation from the film normal also induces diffraction contrast between the (111) and (1 T 1) grains. Any overlapping (111) and (T T T) grains, brought about by an in-plane 111 }twin boundary, would show up with an intermediate contrast level, but very few were seen. I 110} films of gold were formed by deposition onto a germanium [001] substrate. The copper ( 1111 film preparation was more complicated and required intermediate layers of cobalt and silver to be deposited first on the silicon [I ll] wafer [2], followed by ion-milling from the substrate side. Microscopy was performed on the JEOL-ARMII in Stuttgart operating at 1250kV with the sideentry holder installed [4], giving a point-to-point resolution limit of 1.2A, and the JEOL 201 OF in Sheffield for the gold 1110} film. RESULTS Gold 1111) film. Y3 bicrystal

I Fig. 1. HREM image of single (112) facet separating (11) and (T T T) grains, spacing of{220} fringes is 1.44A. The moird fringes, generated digitally at lOx magnification, reveal the displacement between the two grains.

360

As in the earlier work, facetting

Data Loading...