A Computational Infrastructure for Continuum Analysis of Carbon Nanotubes
Award year: 2002-2003
Carbon nanotubes exhibit superior electrical properties. Metallic carbon nanotubes can carry extremely large current to serve as interconnects in nano-electronics. Semiconducting carbon nanotubes can be electrically switched on and off as field-effect transistors that are more than 500 times smaller than current devices. However, experiments have shown that electrical conductance of carbon nanotubes changes by two orders of magnitude upon mechanical deformation, i.e., metallic nanotubes become semiconducting ones after deformation. This unique electromechanical characteristic of carbon nanotubes has major implications to reliability of nanotube-based electronics because significant electrical property change due to large deformation in manufacture and operation processes may lead to device malfunction.
Nanotube-based electronics involve multiple carbon nanotubes on substrates subjected to complex deformations as in manufacture and operation processes. Atomistic studies on electrical properties, however, are limited to a single carbon nanotube with simple deformations (e.g., tension, compression and torsion). A major hurdle in atomistic studies of nanotube-based electronics is the determination of all atomic positions in deformed carbon nanotubes.
Conventional continuum analysis is very effective and robust to determine deformation of solids under arbitrary mechanical and thermal loadings, though it is not applicable at nanoscale. We propose to develop a new continuum analysis based on atomistic models, and establish a computational infrastructure for continuum analysis of carbon nanotubes. Specifically, the interatomic potential for carbon is directly incorporated into the continuum analysis via constitutive models. Such an approach retains the effectiveness and robustness of continuum analysis, and overcomes difficulties of atomistic studies to determine atomic positions in complex, deformed systems. Once atomic positions are known, tight binding calculations provide energy-dispersion relations, which govern electrical properties of deformed carbon nanotubes.
For a single carbon nanotube subject to simple deformation (e.g., uniaxial tension), our preliminary results have agreed well with atomistic studies. For multiple carbon nanotubes interacting with the environment as in the nanotube-based electronics, effective computational methods and visualization techniques are critically needed. We will work with scientists at NCSA to develop parallel computing methods that are highly efficient, robust and scalable in order to study thousands of carbon nanotubes on substrates developed for nanotube-based electronics. We will also use the visualization capabilities at NCSA to assist our study of multiple carbon nanotubes interacting with the environment. This combination of computational continuum analysis, tight binding methods and visualization techniques provides an effective and robust way to determine electrical property change of multiple, distorted carbon nanotubes due to complex deformations, which is particularly important to the reliability of nanotube-based electronics.