Development of artificial hair sensors for flow measurement and haptic exploration

This work focuses on the design, fabrication and characterization of artificial hair sensors for flow measurement and haptic exploration. One requirement is to make the sensor compatible with a system that can grow and regrow the artificial hair out of an extrusion channel (not yet developed), this means that the sensors have to measure the hair deformation without being directly attached to it.

It presents three design approaches:
• Focused Electron Ion Beam (FEBID) deposited strain gauges around a hair channel (an approximately 100μm wide hole), that detect the bending of the hair. These sensors are referred to as FEBID sensors.
• Embedding strain gauges in a flexible substrate (polyimide) on top of the hair channel, so that when the hair comes out, the strain gauges bend upwards and lean on the base of the artificial hair. These sensors are referred to as polyimide (PI) sensors.
• Depositing strain gauges on a silicon membrane. In the middle of the membrane, a silicon column with a perforation in the middle serves as support for the artificial hair. The
silicon column also helps transferring the deformation from the hair to the membrane.
These sensors are referred to as membrane sensors.

After initial modeling and fabrication attempts, the FEBID sensors were discarded due to the rigidity of the mechanical design and irreproducibility of the strain gauges deposition process.

The PI sensors presented problems due to electromigration and plastic deformation of the PI flaps where the strain gauges are located, leading to cracks on the metal layer. These sensors had a strong drift and stopped working after a few measurement cycles.
On the other hand, the membrane sensors suffered no plastic deformation nor electromigration.

These have a good sensitivity for a 1 cm long hair of between 0.05 mV/mm and 1.37 mV/mm, depending on the sensor’s geometry.

The membrane sensors can be used for 2-directional flow measurement in air and for surface recognition and haptic exploration. For surfaces with regular structures, features as small as 100μm can be recognized. Measurement with sandpaper show that the sensors can differentiate between grain sizes from 201μm to 15μm.

These sensors could be used to differentiate between several types of materials and reconstruct the profile of a surface.