Liquid Crystal Elastomers as Neuromorphic Muscles

Simulating a human-like learning system in soft robotics can significantly enhance the ability of robots to adapt and evolve based on external experiences. The well-established principle of muscle strengthening through training serves as the foundation for this research. In this research project, we aim to replicate this concept by developing an artificial muscle system using liquid crystal elastomers (LCEs) that exhibit memory effects similar to biological muscle fibers, enabling them to learn and adapt based on past experiences.

LCEs are crosslinked polymer networks that combine the elastic properties of rubber with the anisotropic properties of liquid crystals. LCEs show large and reversible actuation when driven by stimuli, such as temperature. As the temperature increases, liquid crystal mesogens transition from the nematic phase to the isotropic phase, leading to a notable and macroscopic deformation in the material, which can be potentially used as artificial muscles. However, this reversible shape change is not dependent on the past experience, meaning the artificial muscle shows the same contraction when exposed to the same trigger. To endow this artificial muscle with memory capabilities, we will integrate an organic electrochemical transistor (OECT) as a synaptic device, which we term the "memory element". Through voltage pulses, the conductance of the conductive polymer within the OECT can be modulated, allowing for the storage of synaptic connections as memory. The proposed architecture involves a multilayer structure with the LCE artificial muscle at the base and the memory element overlaid. In this configuration, a conductive polymer serves a dual function, acting as both the channel material for the OECT and an electrothermal heating element for the LCE layer. By tuning the conductance of the channel material, we can control the electrothermal heating, thus enabling varied degrees of muscle contraction.

We anticipate that this innovative architecture will allow for the realization of the concept of "training more, stronger muscle" in artificial systems. Through cumulative exposure to past events, the artificial muscle will exhibit increased contraction strength, mirroring the adaptive nature of biological muscles. The ultimate goal is to use biological signal to train the artificial muscle response. This research will pave the way for the development of adaptable artificial muscle systems with potential applications in robotics, prosthetics, and beyond.