Learning in colloidal polyaniline nanorods

Abstract

Liquid-based computing media are massively parallel computing devices with high fault-tolerance and self-healing capabilities. They can compute by propagating and interacting phase waves or by changing their internal coordination. Colloidal suspensions of conductive polymer nanorods are expected to be interesting candidates for developing the computing subsystem in such applications, because of their anisometry which makes them particularly susceptible to electrical fields. In this work, we investigated a suspension of polyaniline nanorods (NRs) to explore the potential of generating learning mechanisms in the colloid and applying them in the computing system of future cybernetic systems. We demonstrated that learning, as expressed in the formation of programmable conductive pathways leading to distinct states, can be implemented using Alternated Current (AC) electrical stimulation. We achieved repeatable programming of colloid resistance anisotropy that can be easily mapped into binary logic, demonstrating that this is due to the AC field effects on the hydrogen bonds that stabilise the dispersoids in the solvent as well as the charges' orientation inside the polymeric chains. We also influenced the conductivity of polyaniline (PANI) NRs by changing their molecular conformation. The findings establish robust protocols for programming future liquid robots.

Keywords

• Colloids
• Nanorods
• Learning
• Liquid robotics
• Nanomaterials