Electrostatic analysis of nanoelectromechanical systems

We present a multiscale method, seamlessly combining semiclassical, effective-mass Schrodinger (EMS), and tight-binding (TB) theories proposed for electrostatic analysis of silicon nanoelectromechanical systems (NEMS). By using appropriate criteria, we identify the physical models that are accurate in each local region. If the local physical model is semiclassical, the charge density is directly computed by the semiclassical theory. If the local physical model is quantum-mechanical (EMS or TB model), the charge density is calculated by using the theory of local density of states (LDOS). The LDOS is efficiently calculated from Green's function by using Haydock's recursion method where the Green's function is expressed as a continued fraction based on the local Hamiltonian. Once the charge density is determined, a Poisson equation is solved self-consistently to determine the electronic properties. The accuracy and efficiency of the multiscale method are demonstrated by considering several NEMS examples.;The multiscale method can be used to compute the effect of surface and interior defects such as vacancies and broken bonds on the performance of microelectromechanical systems (MEMS). By combining multiscale electrostatic analysis with mechanical analysis, we compute the capacitance-voltage and pull-in/pull-out voltages of MEMS switches in the presence of defects in the dielectric oxide layer. Our results indicate that both surface and interior defects can change the pull-in/pull-out voltages significantly. These voltage offsets can lead to an eventual failure of the MEMS switches.;The self-consistent TB method is used to investigate carbon nanotube (CNT)-based sensors. We compute the screening effects of semiconducting and metallic single-wall carbon nanotubes (SWNTs) when water molecules and various ions pass through the nanotubes. The trajectories of ions and water molecules are obtained from molecular dynamics (MD) simulations. It is shown that metallic SWNTs have much stronger screening abilities than semiconducting SWNTs. Our results indicate that it is possible to distinctly identify different ions and also to differentiate between arm-chair and zigzag nanotubes.;The TB method is further applied to compute the electrostatic signals generated by DNA segments inside semiconducting single-wall carbon nanotubes. The electrostatic signals indicate that when defective DNA molecules pass through the CNTs, it is possible to identify the number of total base pairs and the relative positions of defective base pairs in DNA chains. Our calculations suggest that it is possible to differentiate Dickerson and hairpin DNA structures by comparing electrical signal patterns.;In addition to the multiscale method, we present an ab initio study of electronic properties of graphene flakes and nanoribbons adsorbed on clean and H-passivated (100) silicon surfaces. The graphene electronic properties are not perturbed by the H-passivated Si substrate. In contrast, the interaction between the clean Si surface and graphene can significantly change the bandstructure of graphene. This is caused by the covalent bonding formed by C and surface Si atoms, which changes the pi-band network of the graphene layer. The dispersion curves and density of states calculations show that the bonded Si-C surface states are highly delocalized and located near the Fermi energy. Adsorption energy indicates that the graphene flakes and nanoribbons are stable on clean Si(100) surfaces.