A current examine revealed in npj 2D Supplies and Functions explores hexagonal boron nitride (h-BN) atomristors, highlighting their notable reminiscence window, low leakage present, and minimal energy consumption. These options make them a promising candidate for energy-efficient neuromorphic computing.
Picture Credit score: Igor Petrushenko/Shutterstock.com
Background
Two-dimensional (2D) supplies, resembling graphene and transition metallic dichalcogenides, have drawn consideration for his or her distinctive electro-mechanical properties, providing benefits over conventional three-dimensional (3D) supplies. Their ultra-thin construction permits for compact, low-power machine designs. Nonetheless, manufacturing these supplies usually introduces defects that may degrade efficiency.
h-BN stands out for its robust insulating properties and mechanical stability, making it a viable resolution to a few of these challenges. This examine focuses on h-BN-based atomic-scale memristors, or atomristors, which use these properties. The researchers used a polypropylene carbonate (PPC) help layer when transferring h-BN monolayers, serving to to scale back defects and enhance machine reliability.
The Examine
The researchers created the h-BN atomristor by inserting a monolayer of h-BN between two silver (Ag) electrodes, forming a metal-insulator-metal (MIM) construction. The machine features by forming and breaking conductive bridges on the electrode interfaces. The junction space was measured at roughly 0.40 × 0.40 μm², emphasizing its atomic-scale dimensions.
To research the machine, the group used optical microscopy (OM), atomic pressure microscopy (AFM), and transmission electron microscopy (TEM) to check the morphology and crystallinity of the h-BN layers. They decided the thickness of the h-BN monolayer to be about 0.51 nm, which matches theoretical predictions.
The group additionally carried out ramped voltage stress (RVS) and pulse voltage stress (PVS) exams to evaluate the endurance and memory-switching habits of the atomristors. These exams helped decide switching thresholds (V_SET and V_RESET) and resistance states. Statistical analyses supplied perception into variability in switching parameters, together with common and normal deviation measurements for voltages and resistances. The researchers additionally examined energy consumption throughout switching, confirming the machine’s low power necessities.
Outcomes and Dialogue
The examine discovered that the h-BN atomristor achieved a reminiscence window better than 4 × 109, considerably bigger than earlier 2D atomristors. The leakage present was roughly 0.24 pA, and energy consumption throughout switching was round 3 × 10-14 W. These outcomes point out that h-BN is an efficient insulating materials with robust efficiency traits.
The machine additionally demonstrated sturdiness, sustaining over 10,000 switching cycles, reinforcing its reliability. The interface between the h-BN layer and Ag electrodes, enhanced by the PPC help layer, contributed to improved efficiency by decreasing polymer residue and guaranteeing higher contact.
Past efficiency metrics, the examine explored how these findings apply to neuromorphic computing, which requires low-power, environment friendly units. The mix of h-BN’s insulating properties and the electroactive nature of Ag electrodes suggests potential for future digital parts. Nonetheless, some challenges stay, together with device-to-device variability, which requires additional analysis to enhance consistency and scalability.
Conclusion
This examine highlights necessary developments in 2D supplies, significantly specializing in the potential of h-BN atomristors. Their giant reminiscence window, low leakage present, and minimal energy consumption make them robust candidates for integration into neuromorphic computing techniques. With demonstrated sturdiness and information retention, h-BN is a viable choice for high-performance purposes.
Nonetheless, to carry this know-how to sensible use, researchers should handle variability between units and refine fabrication strategies for higher consistency. As research proceed, additional investigation of 2D supplies like h-BN will probably be important in creating the following era of digital parts that enhance computational effectivity and higher mimic organic processes.
Journal Reference
Yang SJ., et al. (2025). Large reminiscence window efficiency and low energy consumption of hexagonal boron nitride monolayer atomristor. npj 2D Supplies and Functions. DOI: 10.1038/s41699-025-00533-9, https://www.nature.com/articles/s41699-025-00533-9
