Planar Spin Glass with Topologically Protected Mazes in the Liquid Crystal Targeting for Reconfigurable Micro Security Media

Abstract The planar spin glass pattern is widely known for its inherent randomness, resulting from the geometrical frustration. As such, developing physical unclonable functions (PUFs)—which operate with device randomness—with planar spin glass patterns is a promising candidate for an advanced security systems in the upcoming digitalized society. Despite their inherent randomness, traditional magnetic spin glass patterns pose considerable obstacles in detection, making it challenging to achieve authentication in security systems. This necessitates the development of facilely observable mimetic patterns with similar randomness to overcome these challenges. Here, a straightforward approach is introduced using a topologically protected maze pattern in the chiral liquid crystals (LCs). This maze exhibits a comparable level of randomness to magnetic spin glass and can be reliably identified through the combination of optical microscopy with machine learning‐based object detection techniques. The “information” embedded in the maze can be reconstructed through thermal phase transitions of the LCs in tens of seconds. Furthermore, incorporating various elements can enhance the optical PUF, resulting in a multi‐factor security medium. It is expected that this security medium, based on microscopically controlled and macroscopically uncontrolled topologically protected structures, may be utilized as a next‐generation security system.