A carbon nanotube bearing and Stodola rotor

Abstract

A nano-scale rotor supported on a cantilevered multi-wall carbon nanotube (MWNT) shaft (Stodola configuration) is proposed. The nanotube is also expected to function as the bearing, since individual walls of a MWNT are not strongly bonded and can slide on each other with low friction. While MWNT based rotors have been previously constructed, they have so far been limited to horizontally oriented nanotubes. The rotor uses a vertically aligned tube, which allows superior control of the rotor geometry, enabling improved rotor balancing and axisymmetric features such as electrodes or blades. The rotor is proposed as a test stand for measuring inter-wall friction in MWNTs. The low friction in nanotubes has been studied with simulations and experiments, and while it is agreed that relative motion between walls is possible, there is much debate about the qualitative nature of the friction force between walls. Furthermore the reported quantitative values of friction vary by as much as ten orders of magnitude. The proposed rotor might be used to gather new friction data on rotating MWNT bearings at higher speeds that previously attempted. In addition, identical rotors fabricated on nanotubes of varying size, type, and crystalline quality might provide a large dataset that could be used to find correlations between friction behavior and these factors. Applications for the rotor beyond a friction testing apparatus could include pumps to work with existing micro-chemical sensors, gyroscopes, energy storage flywheels, and turbomachinery for power generation. A fabrication process for the proposed rotor was developed, and is being refined. An isolated vertically aligned MWNT is grown by chemical vapor deposition (CVD), from a nickel catalyst dot defined with electron-beam lithography. A silicon dioxide sacrificial layer is applied, followed by a polysilicon layer from which to cut out the rotor.

(cont.) The rotor etch is performed by cryogenic reactive ion etching (RIE), patterned with electron-beam lithography. The rotor is released from the substrate by hydrofluoric acid vapor. One iteration of the fabrication process was completed, and further iterations are planned to complete a functional device. Actuation of the rotor would be achieved by directing jets of air at blades on the rotor, and an alternative electrostatic actuation concept is also proposed. A dynamic model of the rotor performance based on classical simple beam theory and rigid body dynamics indicates that speeds on the order of thousands to millions of revolutions per minute should be achievable, while avoiding the thirteen potential failure mechanisms analyzed.