Exploiting Carbon Quantum Dots Synthesized by Electrochemical Exfoliation for Flexible Resistance Switching

Wearable nonvolatile memory based on flexible materials has shown an unstoppable development trend in the field of portable artificial intelligence technology. In this work, flexible electronic synapses based on carbon quantum dots (CQDs) synthesized by electrochemical exfoliation are exploited. The CQDs-based devices exhibit similar and stable current-voltage (I-V) characteristics to the initial state after undergoing multiple bending, including I-V curves, retention characteristic for 100 h, endurance characteristic for 106 bending cycles, and classical synaptic functions can be performed faster than the human brain. This is attributed to the extremely small number of defects on the surface of CQDs synthesized by electrochemical exfoliation, which ensures the density and uniformity of the carbon-based film. Interestingly, the artificial neural network model based on the Crosimm platform is used to iteratively train and recognize Modified National Institute of Standards and Technology (MNIST) handwriting, and the simulation results are between 94.2% and 94.9%, further verifying the reliability of silver (Ag)/CQDs/indium tin oxide (ITO)-based devices. Therefore, a reliable decimal arithmetic based on synaptic weight changes is proposed, and addition, multiplication, subtraction, and division can all be performed by applying high-speed pulses. This provides a path for the development of a new generation of high-speed, multifunctional memories.