[ad_1]
The memristor is a promising candidate to implement high-density reminiscence and neuromorphic computing. Based mostly on the retention time, memristors are labeled into unstable and non-volatile sorts. Nevertheless, a single memristor typically gives a particular perform primarily based on digital performances, which poses roadblocks for additional growing novel circuits. Versatile memristors exhibiting each unstable and non-volatile properties can present a number of capabilities masking non-volatile reminiscence and neuromorphic computing. Right here, a flexible memristor with unstable/non-volatile bifunctional properties was developed. Non-volatile functionalities with a storage window of 4.0 x 10^5 was obtained. In the meantime, the system can present threshold unstable functionalities with a storage window of seven.0 x 10^4 and a rectification ratio of 4.0 x 10^4. The leaky integrate-and-fire (LIF) neuron mannequin and synthetic synapse primarily based on the system have been studied. A flexible memristor allows non-volatile reminiscence, selectors, synthetic neurons, and synthetic synapses will present benefits concerning circuit simplification, fabrication processes, and manufacturing prices.
[ad_2]
