Thermal cycling fatigue and deformation mechanism in aluminum alloy thin films on silicon

Geometric and material nonlinearity during the deformation of micron-scale thin-film bilayers subject to thermal loading

“…To this end, each test proceeds as follows. First, the structures are thermal cycled three times between room temperature and 190 • C to try to stabilize the gold microstructure since it is well-known that signiÿcant microstructural evolution can occur during the ÿrst thermal cycle after deposition (see, for example, Nix, 1989;Leung et al, 2000;Thouless et al, 1993Thouless et al, , 1996Shen and Suresh, 1995;Koike et al, 1998;Keller et al, 1999;Baker et al, 2001;Weiss et al, 2001). During this process the temperature is held constant for ≈ 3:5 min roughly every 15 • C so that thermal equilibrium is reached.…”

“…Stress relaxation during an isothermal hold has been studied by Hershkovitz et al (1985), Doerner and Nix (1988), Flinn et al (1987), Korhonen et al (1990), Shute and Cohen (1992), Shen and Suresh (1995), Thouless (1995), Koike et al (1998), and Keller et al (1999). Thouless et al (1993) measured stress relaxation of copper ÿlms on a silicon substrate at temperatures ranging from 74 • C to 187 • C. They found that the qualitative behavior was well-described by power law creep˙ = A n , but argued that a mechanistic understanding of the values of A and n that were required was lacking.…”

“…The reasons why this was necessary were not understood. Koike et al (1998) studied stress relaxation of aluminum alloy ÿlms on silicon substrates at 100 • C and 240 c C and interpreted the results to suggest three di erent deformation mechanisms occur over the temperature range of 300-650 K. The boundaries between the regions of di erent mechanisms were reasonably well described by a deformation mechanism map modiÿed to account for phenomena particular to thin ÿlms as suggested by Thouless et al (1993). Consistent with Shen and Suresh (1995) they found that stress relaxation is most pronounced during the early stages of the isothermal hold.…”

“…It is reasonably well known that metal ÿlms are often deposited in a state far from equilibrium and thus do not typically exhibit a stable microstructure in their as-deposited state. Exposing them to temperature excursions results in highly nonlinear deformation behavior during the ÿrst few cycles due to microstructural evolution such as annihilation of excess vacancies, void coalescence, and grain growth (Nix, 1989;Shen and Suresh, 1995;Koike et al, 1998;Baker et al, 2001;Legros et al, 2002). The microstructural evolution results in inelastic deformation, which at the continuum level can be treated as an inelastic strain, similar to plastic or creep strain (Zhang and Dunn, 2003).…”

Grain Boundary Properties and Grain Growth: Al Foils, Al Films

“…[28] However, for the films studied here, the minimization of these energies is not expected to play a significant role in either the initial grain growth or the eventual stagnation since the films were very strongly <111>-fiber textured even in the as-deposited condition and annealing resulted in minimal strengthening of this texture, as seen in the fiber plot of Fig 10. In addition, the Al films are in the zero stress, or lowcompressive steady-state stress state at the annealing temperature and reach this state during heating to temperature. [29][30][31] Thus, film stress and its relaxation are also not expected to play a significant role in the observed grain growth followed by stagnation.…”