Quantum Anomalous Hall Effect-Based Variation Robust Binary Content Addressable Memory

Electronic devices can no longer afford dimensional downscaling due to the fundamental physical limit which motivates device researchers to look for a completely new technological paradigm. Cryogenic devices are currently the most promising candidates among the alternative paradigms as they offer extremely low power and ultra-fast speed without the requirement of device miniaturization. However, the traditional von-Neumann bottleneck limits the performance throughput and power efficiency of cryogenic devices owing to the separate storage and compute elements. In this context, the concept of in-memory computing has emerged where the computations are performed inherently in the memory itself. Content addressable memory (CAM) is one such type of memory block where the memory search operation takes place inherently. We propose a Binary CAM (BCAM) based on twisted bilayer graphene moire heterostructure for cryogenic application. It provides topologically protected non-volatile quantum Hall states and exhibits variation robustness. With the appropriate circuit components and suitable bias conditions, our proposed BCAM array is capable of inherent memory search operation consuming ultra-low power of 1.2 nW/search/bit. We also perform a 1000-point Monte-Carlo variation analysis to assess the variation robustness of our proposed BCAM and find that the BCAM operates reliably even at the worst-case variation attesting to its unique variation robustness.