Planar Extrinsic Biasing Of Thin Film Shape-Memory MEMS Actuators.
- PDF / 142,772 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 45 Downloads / 198 Views
J13.9.1/Y7.9.1
Planar Extrinsic Biasing Of Thin Film Shape-Memory MEMS Actuators. D. S. Grummon1, R. Gotthardt2 and T. LaGrange2 1 Michigan State University, Dept. of Chem. Eng. and Materials Science, E. Lansing, MI 48824 2 Swiss Federal Tech. Inst. of Lausanne (EPFL), Inst. of Physics of Complex Matter, Ecublens, CH-1015, Switzerland. ABSTRACT Although slow and dissipative, sputtered thin-film shape-memory alloys like equiatomic titanium-nickel can exert a large ohmically-excited force·displacement product when deployed in photolithographically micromachined actuators. They give energy densities far exceeding those typically produced by competing microactuator materials [1], and their size can probably be scaled down to the nanometer range (where the benefits of high surface to volume ratio are best exploited for speed and efficiency). But a large, energetic, and resettable actuation stroke is possible only if some agency has imparted a non-trivial initial plastic strain, of between one and five percent, to the martensite phase. Is not always obvious how this strain is to be achieved when discrete mechanical manipulation of the active element is difficult. Furthermore, for cyclic actuation, a resetting-force that periodically re-deforms the martensite during the cooling interval must arise naturally from mechanical elements in the design. Here, several methods responding these requirements are discussed in relation to various kinematic themes. INTRODUCTION One of the earliest experiments that attempted to deploy a shape-active element made from sputtered and photolithogrqaphically micromachined nickel-titanium (for planar Si-based MEMS) resulted in a flat, TiNi thin film structure, patterned as a serpentine form, and released from the substrate along part of its length [2]. When ohmically heated, the film was observed to exuberantly distort, rising from the plane of the wafer, bending and partly coiling, but then resuming something near the original flat shape on cooling. The response was striking and unconventional, and yet the observed shape-changes, accomplished by the bending deformation of very shallow members, represented strains and forces rather smaller than are extractable from a robust TiNi shape-memory film. Built as a test bed for deposition and patterning processes, this device would have been unable to perform appreciable external work. Two additional functions (aside from attachment of loads) would need to have been incorporated in this system to make it a practical high-energy-density cyclic micromechanical actuator: First, some mechanical agency must impart, to the stable martensite, an initial plastic strain, of one to six percent. Secondly, a mechanical biasing force must cyclically communicate with the active element, to plastically redeform the martensite during the cooling segment of each actuation cycle. Often, the service load itself provides ample biasing force - an attractive alternative where possible, but for many cyclic SMA actuators1 a discrete ‘bias spring’ is desirable to control ma
Data Loading...
