CNT‐Based Artificial Hair Sensors for Predictable Boundary Layer Air Flow Sensing

Abstract

While numerous flow sensor architectures mimic the natural cilia of crickets, locusts, bats, and fish, the prediction of sensor output for given flow conditions based on the sensor properties has not been achieved. Challenges include difficulty in determining the electromechanical properties of the sensors, limited working knowledge of the boundary layer, low sensitivity to small hair deflections, and lack of models for large deflections. Within this work, hair sensors are fabricated using piezoresistive arrays of carbon nanotubes (CNTs) without traditional microelectromechanical processing. While correlating the CNT array electromechanical properties to synthesis conditions remains a challenge, a consistent, proportional, and predictable response to steady, boundary‐determined air flow is obtained using theory and measurement for various lengths of hairs. The moment sensitivity is shown to scale inversely with the CNT length and stiffness to a typical maximum of 1.3 ± 0.4% resistance change nN−1 m−1. The normalized CNT piezoresistivity is constant (1.1 ± 0.2) for a majority of the more than two dozen sensors examined despite the orders‐of‐magnitude variability in both sensitivity and CNT compressive modulus. The sensor sensitivity and noise both distinctly change as the flow transitions from steady and laminar to turbulent, suggesting the sensor may be capable of detecting flow transitions.