Impact of the La2NiO4+δ oxygen content on the synaptic properties of the TiN/La2NiO4+δ/Pt memristive devices

Abstract

The rapid development of brain-inspired computing requires new artificial components and architectures for its hardware implementation. In this regard, memristive devices emerged as potential candidates for artificial synapses because of their ability to emulate the plasticity of the biological synapses. In this work, the synaptic behavior of the TiN/La2NiO4+δ/Pt memristive devices based on thermally annealed La2NiO4+δ films is thoroughly investigated. Using electron energy loss spectroscopy, we show that annealing using reducing (Ar) or oxidizing (O2) atmospheres affects the interstitial oxygen content (δ) in the La2NiO4+δ films. Electrical characterization shows that both devices exhibit long-term potentiation/depression and spike-timing-dependent plasticity, which makes them suitable for neuromorphic applications. At the same time, the Ar annealed TiN/La2NiO4+δ/Pt device demonstrates non-volatile properties with low energy consumption during the learning process. On the other hand, in the O2 annealed TiN/La2NiO4+δ/Pt device the resistive switching behavior is more volatile and requires more energy for synaptic learning. Finally, the simulation tools show that spiking neural network architectures with unsupervised learning rules based on the experimental data achieve high inference accuracy in the digit recognition task, which proves the potential of TiN/La2NiO4+δ/Pt devices for artificial synapse applications.