A circuital model of Sb2Se3 solar cells as memdiode devices for neuromorphic applications

Abstract

Memristors are one of the four fundamental electrical components, relating electrical quantities such as current, voltage, charge (integral of current), and flux (integral of voltage). First proposed in 1971 and experimentally demonstrated in 2009, memristors have garnered significant research interest due to their unique properties. Among their many applications, neuromorphic computing and memory storage stand out. However, memristive behavior is not limited to standalone devices; it can also manifest as a parasitic effect in various systems. Despite this, research on parasitic memristive effects remains limited, especially when compared to studies on resistors and capacitors. In this work, we present experimental evidence demonstrating that a Mo/MoSe2/Sb2Se3 (10 nm)/CdS (2 nm)/indium titanium oxide (ITO) substrate configuration, originally designed as a solar cell, exhibits memristive behavior. This behavior is investigated by applying both triangular waveforms and a series of voltage pulses to the solar cell contacts while recording the current transient response. Our proposed model builds upon the standard single-diode solar cell model, with modifications to incorporate the observed memristive effects as second-order, parasitic elements. The experimental data show that the proposed model fits well with the experimental results, confirming the memristive behavior in the system.