In International journal of nanomedicine
Background : As a key component in artificial intelligence computing, a transistor design is updated here as a potential alternative candidate for artificial synaptic behavior implementation. However, further updates are needed to better control artificial synaptic behavior. Here, an updated channel-electrode transistor design is proposed as an artificial synapse device; this structure is different from previously published designs by other groups.
Methods : A semiconductor characterization system was used in order to simulate the artificial synaptic behavior and a scanning electron microscope was used to characterize the device structure.
Results : It was found that the electrode added to the transistor channel had a strong impact on the representative transmission behavior of such artificial synaptic devices, such as excitatory postsynaptic current (EPSC) and the paired-pulse facilitation (PPF) index.
Conclusion : These behaviors were tuned effectively and the impact of the channel electrode is explained by the combined effects of the joint channel electrode and conventional gate. The voltage dependence of such oxide devices suggests more capability to emulate various synaptic behaviors for numerous medical and non-medical applications. This is extremely helpful for future neuromorphic computational system implementation.
Qi Shaocheng, Hu Yongbin, Dai Chaoqi, Chen Peiqin, Wu Zhendong, Webster Thomas J, Dai Mingzhi
artificial synapse, channel-electrode transistor, neuron behavior control, thin-film transistor