Deformation of blanketed and patterned bilayer thin-film microstructures during post-release and cyclic thermal loading

We study, both experimentally and theoretically, the deformation of blanketed and patterned bilayer thin film microstructures subjected to temperature cycles from room temperature to elevated temperatures following processing by surface micromachining and release from the substrate. While the theoretical treatment is general, the experimental component focuses on beam-like microstructures consisting of a 0.5 μm thick gold film on a polysilicon film that is either 1.5 μm or 3.5 μm thick. For all microstructures the underlying polysilicon film is the same size, but the gold film is patterned into a line that runs the length of the beam. Its width is varied from 0 to 100% of the width of the polysilicon. We experimentally characterize the deformation by measuring the full-field deflection of the gold/polysilicon bilayer beams as a function of temperature using a white-light interferometric microscope. From the deflection, the curvature is determined, and we report the evolution of curvature with the temperature cycling. Qualitatively the behavior is the same regardless of the linewidth. The quantitative differences can be described by a simple model incorporating an inelastic temperature-driven mechanism in addition to linear thermoelastic behavior. We show experimentally and/or analytically, how the parameters in the model vary with linewidth. The results are discussed in the context of the current understanding of microstructural evolution in thin-film metals, and in relation to anticipated thermoelastic response. We show that via a suitable thermal process, the thin film material microstructure can apparently be stabilized over a prescribed temperature range, rendering the subsequent deformation linear thermoelastic. We discuss the implications of these findings in the context of the design and fabrication of high-yield, dimensionally stable MEMS devices utilizing bilayer material systems. Although our measurements are focused on gold/polysilicon bilayer films, the concepts and associated analysis are applicable to other bilayer film systems, particularly ones with metals, although there will surely be quantitative differences.