This study investigated an oxygen-vacancy-controlled bilayer TiN/TaOy/TaOx/Pt memristive synaptic device for neuromorphic computing. Multilevel characteristics of the synaptic device were observed with RESET voltage varying between −1.4 and −1.9 V. The device shows highly stable reparative 200 potentiation and depression cycles. The high nonlinearity results of αp = 0.83 for potentiation and αd = −2.03 for depression were observed with the device’s potentiation and depression functions. The device also exhibits a highly stable DC endurance of at least 1000 cycles, an AC pulse endurance of 1 M, and a steady retention of 104 s at 100 °C without any degradation. Furthermore, a Hopfield neural network (HNN) is trained to recognize a 28 × 28 pixels image as an input, representing 784 synapses. In 23 epochs, the HNN successfully identified the input image with training accuracy over 92%. This bilayer memristive device can be highly suitable for neuromorphic devices in the development of neuromorphic-computing field.

Because of the continual advancements in the artificial intelligence technology, its practical applications (e.g., in computer vision, health care, and pattern recognition) are expanding. Various architectures integrating memory cells and transistors have been used to demonstrate artificial synaptic arrays. However, designing memory cells with superior analogue switching is a major challenge in the development of neuromorphic systems. We evaluated a TiN/AlN/Cu/AlN/Pt memristive synapse for neuromorphic computing. The synaptic device exhibited multilevel characteristics with varying reset stop voltages ranging from −2.7 to −2.2 V. The device also exhibited highly stable reparative 200 potentiation and depression cycles with high nonlinearity values of approximately 2.39 for potentiation and −2.19 for depression. The device further exhibited highly stable DC endurance over 1000 cycles, AC pulse endurance (1 M), and stable retention (104 s) at 100 °C without any degradation. The experimental potentiation and depression data were used to train a Hopfield neural network to effectively identify input images with a resolution of 28 × 28 pixels, representing 784 synapses used in the simulation process. The training resulted in an accuracy of >91% in 28 epochs. Therefore, the newly designed device exhibits promising features, such as high linearity and accuracy, and is thus highly suitable for neuromorphic computing purposes.

In this study, an interfacial switching (IFS)-based bioinspired TiN/AlON/ TaON/Pt resistive random-access memory (RRAM) device was fabricated to investigate its conduction mechanism and synaptic behavior for neuromorphic computing. This device exhibited excellent dc endurance over at least 10000 cycles, an ac pulse endurance of 1 M, and long-term retention (10 6 s) at 150 °C with no degradation. The device also showed multilevel characteristics. The RESET stop voltage of the device varied from −0.9 to −1.3 V. The device was highly stable over 250 potentiation and depression cycles, which are crucial in Hopfield neural network (HNN)-based neuromorphic systems. High nonlinearity (1.13 for potentiation and −1.75 for depression) was achieved using the device potentiation and depression functions. Experimental potentiation and depression data were used to train an HNN based on the fabricated device to recognize an input image of 28 × 28 pixels that contained 784 synapses. The HNN had a training accuracy of higher than 93% in 22 iterations. The experimental results indicate that the fabricated IFSbased TiN/AlON/TaON/Pt RRAM device is highly suitable for neuromorphic devices that mimic synaptic characteristics for neuromorphic systems.