Diffusion welding of silver interlayers coated onto base metals by planar-magnetron sputtering

Abstract

Silver has been coated onto various base metals by planar-magnetron (PM) sputtering to provide interlayers for diffusion welding. The vacuum-coating process consisted of two stages: (1) sputter etching of the base metals to remove surface oxide layers, followed by (2) vapor deposition of high-purity silver onto the clean surfaces. The silvers surfaces were diffusion welded at elevated pressure (207 MPa) and temperatures (483–673 K). The structures of the diffusion-welded-silver interlayer and the ‘‘as-deposited’’ coatings were determined using optical metallography and electron microscopy. The as-deposited structure consists of fine columnar grains ∼0.25 μm in diameter, perpendicular to the base-metal surface with the axes of the columns oriented along the [111] crystallographic direction. These grains contain a high density of growth twins ∼15 nm thick. The diffusion-welded-silver interlayer consists mostly of large recrystallized grains (>1 mm in diameter) containing a high density of annealing twins. However, a significant amount of the interlayer has not recrystallized. Additionally, the diffusion-welded-silver (silver-silver) interface consists of a high-angle grain boundary often formed between recrystallized and nonrecrystallized regions. For various base metals, the tensile strengths of diffusion-welded-silver joints fabricated using PM sputter deposition were found to be equal to or greater than previously reported strengths for those fabricated using brazing, electroplating, or other vapor-deposition methods. Tensile properties of diffusion-welded-silver joints fabricated using PM sputter-deposition were also found to be more reproducible than properties previously reported for joints fabricated using hot-hollow cathode (HHC) evaporation. Torsion tests of the diffusion-welded-silver joints revealed that the yield stress and strain-hardening rates in the interlayer are much higher than corresponding values for high-purity annealed bulk silver, although the maximum (steady-state) stresses are nearly identical. These results may be at least partially explained by the high twin density in the recrystallized silver and also the very fine microstructure of the nonrecrystallized regions.