Development of Ultra-High Mechanical Damping Structures Based on the Nano-Scale Properties of Shape Memory Alloys

Abstract : We recently discovered that micro and nano pillars milled by Focused Ion Beam (FIB) on Cu-Al-Ni shape memory alloys (SMA) exhibited excellent superelastic behavior and ultra-high mechanical damping at small scale. This behavior offers an interesting potential for mechanical damping applications in Micro Electro-Mechanical Systems and other applications. The objective of the present project was to further explore the ultra-damping behavior of micro/nano pillars and arrays of such features. Several alloys of Cu-Al-Ni SMA, with the appropriate composition to exhibit superelastic effect at room temperature, were induction melted, and then oriented single crystals were grown from these alloys. Many different sizes of micro-nano pillars were milled on slides cut from the oriented single crystals, by FIB. The superelastic effect of these pillars was subjected to nano-compression tests using an instrumented nanoindenter, and the damping coefficient was measured from the load displacement curves. A high damping with a loss coefficient of about eta=0.16 has been confirmed to be stable over 100s of cycles. In order to verify the reliability of the Cu-Al-Ni pillars for mechanical damping applications, arrays of micro pillars were milled in a similar way. Nanocompression superelastic behavior was tested along with cyclic behavior, obtaining a perfectly recoverable behavior above more than 2000 cycles, with a loss coefficient eta0.1. Study of the pillars from arrays showed good reproducibility. Finally, a study of the scalability of the production process was realized and the required methodology to produce very large arrays of pillars by using the VLSI techniques employed in microelectronics was analyzed. It can be concluded that micro-nano pillars of Cu-Al-Ni SMA offer a good potential for mechanical damping applications at small scale, but the development of very large arrays and structures or stacks of arrays of pillars has to be still optimized.