An application of statistics in nanotechnology: Assessing the reliability of nanosensors

Nanoscience is an exciting new area of research that is getting a lot of attention and spans across many scientific fields. Innovations in nanotechnology have already been applied to products that are currently on the market. Much attention is given to the dual problem of designing nanomaterials with novel physical properties and how such materials can be fabricated. Receiving less attention is the question of reliability. High reliability is necessary to guarantee the advancement and utilization of nanotechnology. This work focuses on modeling the behavior of one particular nanodevice, a hydrogen gas nanosensor, and then evaluating its reliability. The nanosensor is a network of ultra-small palladium nanowires which detects hydrogen gas through a change in resistivity that occurs as the gas is adsorbed onto the surface of and/or diffused into the nanowires. Unfortunately no data is available for reliability analysis for this nanosensor, so we rely on a computer model to simulate the behavior of the nanosensor. We evaluate the reliability using site percolation on different lattices while trying to resolve many of the difficulties that arise due to the small size and unique properties of the nanowires. We evaluate the reliability of the nanosensor under different assumptions on the proposed model. We also employ Approximate Bayesian Computation to obtain information about the behavior of the nanowires in an application environment.