Ionic Dynamics Model And Synaptic Simulation At Polymer Electrolyte/Semiconductor Interface

The development of artificial intelligence is in urgent need of chips with parallel computable property,high fault tolerance and low power consumption.Neuromorphic chips,as a potential candidate,have attracted substantial attention.Because of the limit of Moore's law in silicon semiconductor,it is an imminent revolution to design artificial synapses in a single component instead of an array of transistors.Currently,some issues such as the difficulty to implement modulatable plasticity,the huge disparity of intrinsic mechanism between artificial devices and biological synapses and the lack of quantitative analytic methods are still waiting for solution.In this work,we designed an artificial synapse based on polymer electrolyte/semiconductor heterojunction,mimicking the facilitation-depression interplayed plasticity,which was the classic characteristic of Schaffer collateral（SC）in rat brain.We also established a physical and a quantitative model on the level of ionic dynamics,describing the relevance of internal processes between our devices and true synapses.Our study suggested a possible means to improve synaptic computation in neuromorphic chips.The Pt/P3HT/PEO-CaTf₂/Pt multilayer cells were fabricated and their pulse responses were studied.The discharging peak,represented by the post-synaptic current（PSC）,first increases,and then decreases with increasing input number in a pulse train.The weight of the PSC decreased for low frequency stimulations,while increased for high frequency stimulations.These behaviours resembled the plasticity of Schaffer collateral.A model that the ionic polarization and doping processes interplayed at the electrolyte/semiconducting polymer interface was proposed,corresponding to the interplay of residual calcium level and vesicle fusion in presynaptic membrane.Then we developed F-D model that quantitatively described the facilitation-depression plasticity on the basis of previous study in biological synapses.The model well fit the measured data.The simulation results showed that the observed synaptic plasticity was caused by the great disparity between the recovery time constants of F and D（τ_F andτ_D）,revealing that the electrochemical doping to P3HT is much slower than the saturation of concentration polarization of PEO complex.To prove the regulatory ability of our devices,we investigated the influence of polymer structure by adjusting the composition of PEO-CaTf₂.XRD and Raman spectra showed the increase of salt concentrations,i.e.,the reduce EO/Ca²⁺,can enhance the relative content of ion-pairs and aggregates but reduce that of free ions.Consequently,STP transformed from interplayed facilitation-depression property to monotonous facilitation with salt concentration increasing.The result shows signal selectivity of our devices are able to be modulated by the types of ionic species,which might be a part of the secrets of information handling the biological body.