电极材料对NbO_(x)Mott忆阻器稳定性的影响

Effects of Electrode Material on the Stability of NbO_(x)Mott Memristor

为了改善NbO_(x)Mott忆阻器电学稳定性和一致性,提升NbO_(x)Mott忆阻器构建人工脉冲神经元的应用潜力,研究制备了通孔型NbO_(x)Mott忆阻器,并对比研究了Pt、W电极材料对器件稳定性和一致性的影响。研究结果表明,相较于常见报道的Pt电极器件,采用W电极的NbO_(x)Mott忆阻器表现出了更为优越的稳定性和一致性。此外,利用NbO_(x)Mott忆阻器搭建了振荡电路,成功实现了人工脉冲神经元的功能。基于W电极NbO_(x)Mott忆阻器的人工脉冲神经元可以稳定振荡时间超过106 s,循环耐久性可达1012次以上,其振荡波形的幅度及频率稳定性远好于基于Pt电极的人工脉冲神经元。进一步的XPS结果显示,在基于W电极的NbO_(x)Mott忆阻器中,W和NbO_(x)界面生成了一层致密的WO_(x)层,有效地阻挡了氧空位在NbO_(x)材料中的迁移。相比之下,基于Pt电极的NbO_(x)Mott忆阻器因Pt层存在大量晶界且对氧空位有较强的吸附作用,导致在电激活和阈值阻变过程中氧空位发生跳动,从而降低了器件的电学稳定性。该研究为提升NbO_(x)Mott忆阻器的稳定性和一致性提供了新的途径,有望推动其在脉冲型神经形态计算系统中的产业化应用。

To improve the electrical stability and consistency of NbO_(x)Mott memristors and boost their potential for constructing artificial spiking neurons,via-hole NbO_(x)Mott memristors were fabricated and the effects of Pt and W electrode materials on device stability and consistency were compared.The results show that the NbO_(x)Mott memristor using W electrode demonstrate better stability and consistency in comparison to the commonly reported Pt electrode devices.Furthermore,an oscillation circuit was successfully built using the NbO_(x)Mott memristor,realizing the function of artificial spiking neurons.The artificial spiking neurons based on the W electrode NbO_(x)Mott memristors exhibitstable oscillation time exceeding 106 s and good endurance of over 1012 cycles,with significantly improved amplitude and frequency stability in the oscillation waveform,compared to those based on the Pt electrode NbO x Mott memristors.In addition,XPS results reveal that in W electrode devices,a dense layer of WO_(x)forms at the interface between W and NbO_(x),effectively blocking the migration of oxygen vacancies within the NbO_(x)material.In contrast,due to the presence of numerous grain boundaries in the Pt layer and its strong affinity for oxygen vacancies,Pt electrode devices experience oxygen vacancy hopping during the process of electrical activation and threshold resistive switching,thus reducing the electrical stability of the device.In sum,this study presents a novel way to enhance the stability and consistency of NbO_(x)Mott memristors,with the potential to drive their industrial application in pulse-type neural morphological computing systems.