Novel nanoinformatics methods for scaling up atomistic-scale simulation of carbon nanotube synthesis

Nanotechnology is widely considered to profoundly advance multiple industrial sectors. Due to the high costs, low resolution and confounding issues with experimental studies, atomistic Molecular Dynamics (MD) and Monte Carlo (MC) simulations offer an attractive means to study and address the yield and quality issues in the synthesis of nanostructures. However, current atomistic simulations are hampered by the huge computational overheads. Consequently, they are confined to modeling only the early stages of nanomanufacturing processes. We present a novel nanoinformatics approach using a multi-step local Gaussian process model to capture the nonlinear and nonstationary dynamics of growth process to accelerate MC simulation of nanomanufacturing process, such as synthesis of carbon nanotube (CNT) in chemical vapor deposition. The key idea is that the simulated CNT growth increments were found to exhibit nonlinear, recurring near-stationary dynamics, and lifting the tube structure by the predicted growth increment can put it to near optimum position. Results show that the present approach can save more than 80% of the computational effort as in traditional MC simulation, leading to the growth of one of the largest CNTs (194 nm long) from atomistic simulations.

Novel nanoinformatics methods for scaling up atomistic-scale simulation of carbon nanotube synthesis Conference Paper