Design and Computational Potential of Circuit-Based Multiple-Electron Network Model

Complex nonlinear physical systems can exhibit dynamic responses that provide useful resources for information processing. In this study, an electrical circuit-based multiple-electron model was developed and implemented in a random network to investigate its dynamic electrical properties and information-processing capability. The model represents discrete electron transfer and charge accumulation using tunnel junctions and charge-storage nodes and was constructed as a two-dimensional random network inspired by carbon nanotube/polyoxometalate (CNT/POM) networks. The network exhibited time-varying current responses under a constant voltage and nonlinear and hysteretic current–voltage characteristics. The hysteresis became more pronounced as the number of charge-storage nodes increased. The information-processing capability of the network was further investigated using delayed XOR and sine waveform generation tasks. The delayed XOR task was achieved using the integrated squared current response, whereas a target sine waveform was reconstructed from multiple network responses under a constant voltage input using a linear readout, yielding a coefficient of determination of 0.816. These results demonstrate that the proposed multiple-electron network exhibits nonlinear and history-dependent electrical dynamics and can support information-processing tasks.