Simulation algorithms are developed for the prediction of effective elastic properties of NEMS (Nano-Electro Mechanical Systems) and MEMS (Micro-Electro Mechanical Systems) thin films. Finite Element Method (FEM) is used for micro-scale simulation while ab-initio Molecular Dynamics (MD) is employed for nano-scale.
A lattice model is utilized in order to simulate microstructures of thin films. The proposed method can generate a statistically equivalent microstructure to any single phase micrograph in terms of the number of grains and the grain size distribution. A desired grain size distribution (GSD) is achieved by manipulating nucleation process. Analytical functions for GSD are obtained by taking into account of the domain size and the number of grains. It is believed that nucleation and growth can be controlled by temperature and pressure. The influence of temperature and pressure on the grain size as well as the grain size distribution is investigated.
A quasi-3D mesh of the thin film is generated by employing prism elements. By applying specific boundary conditions to the quasi-3D meshed microstructure, the elastic properties of MEMS thin films are obtained through FEM analysis.
The simulation results show that stochastic distributions of grain anisotropy have a significant influence on overall elastic properties at micro-scale. A fundamental statistical methodology is adopted to characterize elastic properties of thin films.
For nano-scale simulations, the bulk modulus (and other elastic properties) can be influenced by grain boundary when grain boundary volume fraction is not negligible. Consequently, it is desirable to determine the size limit when the grain boundary begins to influence the bulk modulus significantly. The developed MD simulation algorithm found that 6nm is the critical grain size for polysilicon. Moreover, equations are derived from the simulation results for estimating bulk modulus by considering both grain and grain boundary. The developed MD simulation technique can be used to characterize bulk modulus of NEMS materials and to determine the size limit above which grain boundary can be ignored in bulk modulus simulation.