Abstract
Abstract
With the rapid expansion of the big-data era, resistive random access memory (RRAM) has attracted considerable interest for high-density data storage and neuromorphic computing applications. However, its practical deployment remains limited by significant variability in resistive switching (RS) parameters, mainly caused by the stochastic formation and rupture of conductive filaments. In this work, we reported a RRAM based on a tellurium quantum dots (Te QDs)/graphene oxide (GO) film, which effectively improves the uniformity of RS characteristics. Benefiting from the incorporation of Te QDs, the device exhibits significantly reduced parameter fluctuations, including a narrowed switching voltage distribution, highly uniform SET and RESET voltages, and fast switching speed. Based on the electrical characteristics, the embedded Te QDs are proposed to provide localized transport sites that may promote the formation of preferential carrier conduction pathways within the GO matrix. Spike-time-dependent plasticity and multiple synaptic functions are further successfully emulated. In addition, leveraging a 100 × 100 memristive array, image pattern recognition is successfully demonstrated. These results highlight a viable route toward highly reliable memristive devices and advanced artificial neural networks for intelligent applications.