Answers are grounded in this paper and cite their sources.
Reservoir computing necessitates the development of devices exhibiting complex dynamic characteristics to enable efficient hardware implementation. This study introduces a physical reservoir computing (PRC) scheme based on series-parallel-connected Au nanogaps activated through a specialized technique. The method utilizes electromigration, driven by field emission currents across the nanogaps, to modulate the tunnel resistance. The memory capacities of physical reservoirs configured as 2 × 1 and 3 × 2 series-parallel-connected nanogaps were experimentally evaluated by using short-term memory (STM) and parity check (PC) tasks. Unlike single-nanogap systems, these series-parallel configurations do not exhibit a reservoir property region characterized by high STM capacity and low PC capacity. Instead, they demonstrate enhanced memory performance in both STM and PC tasks. Compared with systems based on single Au nanogaps, these configurations significantly improve STM and PC capabilities without increasing the number of virtual nodes, thereby preserving the processing speed inherent to single-nanogap systems. Furthermore, scanning electron microscopy revealed structural modifications within the nanogaps after reservoir operation, including the formation of single-electron transistor islands, which may enhance computational capabilities through single-electron tunneling effects. Evidence of Coulomb blockade behavior, observed as a distinct suppression of conductance near zero bias voltage in the drain current–drain voltage characteristics, further supports this enhancement. These results establish series-parallel-connected Au nanogaps as efficient physical reservoirs and present a promising approach for advancing PRC systems based on Au nanogaps subjected to this activation technique.
No takes yet. Share an insight, caveat, or question.
Sakai et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: