A recent study published in npj 2D Materials and Applications explores hexagonal boron nitride (h-BN) atomristors, highlighting their notable memory window, low leakage current, and minimal power consumption. These features make them a promising candidate for energy-efficient neuromorphic computing.
Image Credit: Igor Petrushenko/Shutterstock.com
Background
Two-dimensional (2D) materials, such as graphene and transition metal dichalcogenides, have drawn attention for their unique electro-mechanical properties, offering advantages over traditional three-dimensional (3D) materials. Their ultra-thin structure allows for compact, low-power device designs. However, manufacturing these materials often introduces defects that can degrade performance.
h-BN stands out for its strong insulating properties and mechanical stability, making it a viable solution to some of these challenges. This study focuses on h-BN-based atomic-scale memristors, or atomristors, which use these properties. The researchers used a polypropylene carbonate (PPC) support layer when transferring h-BN monolayers, helping to reduce defects and improve device reliability.
The Study
The researchers created the h-BN atomristor by placing a monolayer of h-BN between two silver (Ag) electrodes, forming a metal-insulator-metal (MIM) structure. The device functions by forming and breaking conductive bridges at the electrode interfaces. The junction area was measured at approximately 0.40 × 0.40 μm², emphasizing its atomic-scale dimensions.
To analyze the device, the team used optical microscopy (OM), atomic force microscopy (AFM), and transmission electron microscopy (TEM) to study the morphology and crystallinity of the h-BN layers. They determined the thickness of the h-BN monolayer to be about 0.51 nm, which matches theoretical predictions.
The team also performed ramped voltage stress (RVS) and pulse voltage stress (PVS) tests to assess the endurance and memory-switching behavior of the atomristors. These tests helped determine switching thresholds (V_SET and V_RESET) and resistance states. Statistical analyses provided insight into variability in switching parameters, including average and standard deviation measurements for voltages and resistances. The researchers also examined power consumption during switching, confirming the device’s low energy requirements.
Results and Discussion
The study found that the h-BN atomristor achieved a memory window greater than 4 × 109, significantly larger than previous 2D atomristors. The leakage current was approximately 0.24 pA, and power consumption during switching was around 3 × 10-14 W. These results indicate that h-BN is an effective insulating material with strong performance characteristics.
The device also demonstrated durability, sustaining over 10,000 switching cycles, reinforcing its reliability. The interface between the h-BN layer and Ag electrodes, enhanced by the PPC support layer, contributed to improved performance by reducing polymer residue and ensuring better contact.
Beyond performance metrics, the study explored how these findings apply to neuromorphic computing, which requires low-power, efficient devices. The combination of h-BN’s insulating properties and the electroactive nature of Ag electrodes suggests potential for future electronic components. However, some challenges remain, including device-to-device variability, which requires further research to improve consistency and scalability.
Conclusion
This study highlights important developments in 2D materials, particularly focusing on the potential of h-BN atomristors. Their large memory window, low leakage current, and minimal power consumption make them strong candidates for integration into neuromorphic computing systems. With demonstrated durability and data retention, h-BN is a viable option for high-performance applications.
However, to bring this technology to practical use, researchers must address variability between devices and refine fabrication techniques for better consistency. As studies continue, further investigation of 2D materials like h-BN will be essential in developing the next generation of electronic components that improve computational efficiency and better mimic biological processes.
Journal Reference
Yang SJ., et al. (2025). Giant memory window performance and low power consumption of hexagonal boron nitride monolayer atomristor. npj 2D Materials and Applications. DOI: 10.1038/s41699-025-00533-9, https://www.nature.com/articles/s41699-025-00533-9