Impact on silicon? One may have the experience in dropping a silicon wafer on the floor, and shattering it to pieces. Or trying to dice a wafer by a diamond marker, and accidentally crunching the edges. These experiences may lead one to believe that silicon is a brittle material, useless as a constituent of machine components. However, scale the size of the machinery down to 100 um (see Fig.1-1) and polysilicon microstructures can withstand impact at 11 kHz for 72 hours without apparent damage. Thus, silicon machines that are actuated by mechanical impact are not only conceivable and feasible, but may prove to be one of the macro impacts of this technology from the micro world. In contrary to our common sense and daily experience (Fig. 1-2), silicon has exceptionally good mechanical properties, with high yield stress (6.9 GPa), high elastic constant (single crystal silicon, 190 GPa, polysilicon, 150 GPa), and high hardness as pointed out by Peterson[l]. The only problem is that single crystal silicon will cleave along the crystalline lines if and when an initial crack is formed. If the structures are orders of magnitude smaller than the macro size equivalents, the surface energy will become relatively larger and more dominant, and it will become more difficult to form an iritial crack . Moreover, if the silicon is polycrystalline, the grain boundaries further impede the propagation of cracks, providing an even tougher material for micro impact machines.
As a mechanical engineer, one may ask the question "Why use polycrystalline silicon(po1ysilicon) instead of other materials?" The objective of this dissertation is not to prove that polysilicon is the one and only material for micro mechanical machines but to utilize its mechanical properties at the micro scale plus its compatibility with integrated circuits for the development of "smart" micro systems. There obviously are many exciting new materials being developed that may have great potentials for micromachines in the future, but few have the advantage of an existing and mature fabrication technology as the silicon processes have developed by the IC industry during the past 30 years.
Polysilicon micro actuators have provided exciting new research directions in the last four years. The first polysilicon electrostatic micromotors have been made by Berkeley Researchers [3,4]. Mehregany at M.I.T.  subsequently fabricated successful micromotors and has since worked extensively on the basic properties of the design and fabrication of these motors that are about 100 um in diameter and are driven directly by electrostatic actuation. However, micromotors driven by direct electrostatic actuation have intrinsic friction barriers to overcome . Laterally driven electrostatic resonators  have provided actuation in a way that avoids friction. However, these resonators can only be actuated into oscillatory motion, as opposed to continuous forward motion that can be obtained by the electrostatic micromotors. Impact actutation provides an alternative actuator that is not limited by the friction factor and retains the advantages of continous motion.
May 31, 1992
Lee, A. P. (1992). Impact Actuation of Polysilicon Micromechanical Structures. United States: University of California, Berkeley.