Nanograin network memory with reconfigurable percolation paths for synaptic interactions

• DOI: 10.1038/s41377-023-01168-5
• Publication Date: 2023-05-15T12:13:41Z
• Authors: Hoo-Cheol Lee; Jungkil Kim; Ha-Reem Kim; Kyoung-Ho Kim; Kyung-Jun Park; Jae-Pil So; Jung Min Lee; Min-Soo Hwang; Hong-Gyu Park
• Keywords: Neuromorphic engineering

The development of memory devices with functions that simultaneously process and store data is required for efficient computation. To achieve this, artificial synaptic have been proposed because they can construct hybrid networks biological neurons perform neuromorphic However, irreversible aging these electrical causes unavoidable performance degradation. Although several photonic approaches to controlling currents suggested, suppression current levels switching analog conductance in a simple manner remain challenging. Here, we demonstrated nanograin network using reconfigurable percolation paths single Si nanowire solid core/porous shell pure core segments. control enabled the reversible adjustment persistent level, exhibiting behavior this device. In addition, behaviors erasure were through potentiation habituation processes. Photonic was achieved laser illumination on porous shell, linear decrease postsynaptic current. Furthermore, elimination emulated two adjacent interconnected nanowire. Therefore, reconfiguration conductive will pave way next-generation nanodevice technologies.