Toward Memristor-Like Resonant Sensors: Pinched Hysteresis Within MEMS Resonators

*Memristors*, uniquely characterized by their pinched hysteresis loop fingerprints, have attracted significant research interest over the past decade, due to their enormous potential for novel computation and artificial intelligence applications. Memristors are widely regarded as the fourth fundamental electrical component, therefore typically exhibit pinched hysteresis purely in the electrical domain, i.e. electrical input and electrical output. In broader terms, similar behavior should also exist in other physical systems across domains (e.g., physical input and electrical output), hence linking the real physical world with the digital domain (e.g., in the form of a physical sensor). In this talk, we share the first observation of **cross-domain pinched hysteresis behavior** in a micro-electro-mechanical systems (MEMS) resonator device, showing that it is viable to create resonant MEMS sensors incorporating memristor-like properties - a new class of MEMS device termed as **MemReSensor**. We will show that these MEMS sensors can have memory, i.e. the output and sensitivity depend on the previous history. We will also show that the pinched hysteresis loop characteristics can be easily tuned by an external electrical signal. We envisage that this will lay the foundations for a new way of fusing MEMS with artificial intelligence (AI), such as creating in-sensor computing, as well as in-sensor AI, e.g., multi-mode in-sensor matrix multiplication across domains.