Modeling a new acetone sensor based on carbon nanotubes using finite elements and neural network

Abstract.

The objective of this study is to model and simulate the behavior of a new biosensor based on carbon nanotubes to detect, automatically and autonomously, the presence of acetone molecules. A 3D model, based on finite elements of SWCNTs, is proposed and developed using the FE code of the ANSYS simulator. We have developed a study on the dynamic behavior of the carbon nanotube for two clamped-clamped and clamped-free structures, by determining the frequency shift of the nanotube resonator, which varies as a function of the attached mass of acetone molecules. We found that the nanotube resonator can detect one single acetone molecule. A parametric study was conducted to highlight the influence of the nanotube diameter, the length of the nanotube, the attached molecule position and the boundary conditions on the natural frequencies of the nanotube. A multilayer neural network has been used to model the behavior of our acetone sensor. We have shown that artificial neural networks (ANNs) replace our acetone sensor because they faithfully reproduce its behavior and take into account dependence on the structural parameters. The neural network model of the sensor thus obtained was implemented on the SPICE simulator, which allowed us to simulate the operation of an electric environment and therefore assess its performance.