Glassy Synaptic Time Dynamics in Molecular La0.7Sr0.3MnO3/Gaq3/AlOx/Co Spintronic Crossbar Devices
Andrei Shumilin,Prakriti Neha,Mattia Benini,Rajib Rakshit,Manju Singh,Patrizio Graziosi,Raimondo Cecchini,Luca Gnoli,Mirko Prezioso,Ilaria Bergenti,Valentin Alek Dediu,Alberto Riminucci
DOI: https://doi.org/10.1002/aelm.202300887
IF: 6.2
2024-04-02
Advanced Electronic Materials
Abstract:The time dynamics of the resistive state in synaptic devices is a crucial parameter of their performance. In the molecular spintronic devices studies, this time dynamics is glassy, i.e. their conductance changed logarithmically over a wide range of time scales. This shed light on the physics of the process and is indicative of a flat distribution of transport energy levels in the organic semiconductor. In the figure, the green ribbon represents the energy landscape for the impurities in the molecular semiconductor film, the blue balls represent the impurities, the blue arrows indicate migration due to the applied electric field while the red ones indicate thermally activated migration, responsible for the glassy time dynamics. The development of neuromorphic devices is a pivotal step in the pursuit of low‐power artificial intelligence. A synaptic analog is one of the building blocks of this vision. The synaptic behavior of molecular La0.7Sr0.3MnO3/tris(8‐hydroxyquinolinato)gallium/AlOx/Co spintronic devices is studied, where the conductance plays the role of the synaptic weight. These devices are arranged in a crossbar configuration, the most effective architecture for the purpose. The conductance of each cross point is controlled separately by the application of voltage pulses: when set in the high conductance potentiated state, the devices show a spin‐valve magnetoresistance, while in the low conductance depressed state, no magnetoresistance is observed. The time dependence of the resistive switching behavior is an important parameter of the synaptic behavior and is very revealing of the underlying physical mechanisms. To study the time dynamics of the resistive switching after the voltage pulses, the response of the device to trains of potentiation and depression pulses, and the time‐resolved conductivity relaxation after the pulses are measured. The results are described with the conductivity model based on impurity energy levels in the organic semiconductor's gap. A flat distribution of the activation energies necessary to move these impurities is hypothesized, which can explain the observed glassy behavior.
materials science, multidisciplinary,physics, applied,nanoscience & nanotechnology