In-situ Heating Electron Microscopy on Cu/SnO2 Bilayer Nanoribbons

Abstract

In-situ electron microscopy provides a means to directly visualize interfacial processes in real time and with high spatial resolution in real and in reciprocal spaces. However, the traditional preparation of electron-transparent samples from thin films often alters the interface(s) of interest. An innovative approach is to deposit a layer of a chosen material onto a miniature substrate that is already electron-transparent and mounted for TEM imaging. Cu/SnO 2 bilayer nanoribbons were made by coating about 10 nm-thick face centered cubic Cu layers on tetragonal single crystal SnO 2 nanobelts so to study as-made, extending hetero-structure interfaces. The SnO 2 ribbons typically have width and thickness ranging from 10 nm to 1 µm and width/thickness ratios as high as ten [1].The thick and thin sides of a given ribbon can be either of the SnO 2 (101) or ( 010 ) surfaces.TEM observations of bilayers made at room temperature reveal the existence of distinct structural Cu types for growth on SnO 2 (101) and ( 010 ) surfaces, respectively. Cu on SnO 2 ( 010 ) always forms flat and epitaxial Cu(111) films, Fig.1a. In contrast, growth on SnO 2 (101) produces dense and continuous films of Cu grains with no preferred orientation relative to the substrate, Fig.1b. The response of the bilayers to in-situ TEM heating is studied using a 300kV JEOL 3010 transmission electron microscope equipped with a double-tilt heating stage. When subjected to repeated heatingcooling temperature cycles between 25 and 200 o C, the epi-bilayers bent reversibly governed by the theory for macroscopic bimetallic strips, Fig.2. In contrast, the untextured bilayers always displayed a degree of plastic deformation.The Cu layers became unstable when heated to above ~225ºC. Thermodynamic considerations indicate that Cu does not wet SnO 2 (010) or (101) at equilibrium. Both the untextured and epitaxial Cu films irreversibly converted to thick, flat, pure Cu islands between 225 and 500ºC, Fig.3. The onset of island formation was followed by a rapid but brief increase in island number and then a sustained period of slow island thickening through surface diffusion of Cu.Above 550ºC the Cu islands underwent a series of solid-state reactions with the SnO 2 substrate, leading to various phases and major changes in morphology. Sn was first diffused into many of the thickening Cu islands. On the SnO 2 (101) surface, these alloy islands began to etch rapidly into the ribbon substrate at about 600ºC, Fig.4, while the SnO 2 (010) surface proved to be more chemically resistant. As etching continued, many of the flat islands lost their faceting, became quasi-spherical in shape. By 725ºC the majority of these particles had transformed into Cu-Sn phases. Further heating fused the nanoribbons.