Switching kinetics of memristive devices employing topotactic brownmillerite-perovskite phase transitions (MEMTOP)

基于氧化还原的记忆器件是未来非易失性存储器和神经形态电路中人工突触的极具吸引力的候选者。记忆行为通常是由场驱动的氧离子运动导致的局部纳米级氧化还原反应引起的。在某些钙钛矿薄膜中，如锰氧化物、铁氧体和钴矿，这种氧化还原过程导致了具有无序氧空位的导电钙钛矿(PV)相和具有有序氧空位的绝缘褐磨石(BM)相之间的拓扑定向相变。在这种情况下，氧离子传输动力学有望决定记忆装置的性能，如人工突触的切换速度、保持和可塑性，但目前对这些离子传输动力学了解甚少。我们只知道BM薄膜中的电阻转换强烈地依赖于BM薄膜的取向。在这个项目中，我们将结合各种实验和计算技术来阐明钴酸盐和铁氧体BM薄膜中与取向相关的氧离子输运与相应记忆器件的开关动力学之间的关系。具体地说，我们将为脉冲激光沉积(PLD)和金属-有机化学气相沉积(MOCVD)这两种不同的沉积技术开发不同晶体取向的制备路线。这类样品一方面将受到原位X射线衍射(XRD)和原位拉曼光谱的监测，以监测拓扑相变的动力学，另一方面，将受到18O同位素交换退火热处理和随后的二次离子质谱仪(SIMS)或拉曼光谱的分析，以获得氧在PV和BM薄膜中的示踪扩散系数。这些研究的两个分支都将得到分子动力学(MD)模拟的支持。通过这种方式，我们的目标是将拓扑定向相变的动力学与氧离子沿不同结晶方向的传输动力学联系起来。第三个阶段是通过不同的操纵面开关技术，阐明不同晶向的BM记忆器件的开关机理和动力学。将所有关于拓扑定向相变的知识放在一起，我们将得到具有改进的开关动力学的记忆器件的设计规则。

Redox-based memristive devices are highly attractive candidates for future non-volatile memories and for artificial synapses in neuromorphic circuits. The memristive behavior is usually induced by the field-driven movement of oxygen ions leading to a local nanoscale redox reaction. In thin films of certain perovskites, such as manganites, ferrites and cobaltites, this redox process results in a topotactic phase transition between a conducting perovskite (PV) phase with disordered oxygen vacancies and an insulating brownmillerite (BM) phase with ordered oxygen vacancies. In such cases, the oxygen-ion transport kinetics are expected to determine the memristive device’s performance, in terms of switching speed, retention and plasticity of artificial synapses, but at present these ion transport kinetics are poorly understood. It is only known that resistive switching in BM films strongly depends on the orientation of the BM thin film. In this project, we will combine a variety of experimental and computational techniques to elucidate the relationship between the orientation-dependent oxygen-ion transport in cobaltite and ferrite BM thin films and the switching kinetics of the corresponding memristive devices. Specifically, we will develop fabrication routes for different crystal orientations for two different deposition techniques, pulsed laser deposition (PLD) and metal-organic chemical vapour deposition (MOCVD). Such samples will be subjected, on the one hand, to in situ X-ray diffraction (XRD) and in situ Raman spectroscopy to monitor the dynamics of the topotatic phase transition, and on the other hand, to 18O isotope exchange anneals and subsequent secondary ion mass spectrometry (SIMS) or Raman spectroscopy analysis to obtain tracer diffusion coefficients of oxygen in both PV and BM thin films. Both branches of these studies will be supported by molecular dynamics (MD) simulations. In this way, we aim to link the kinetics of the topotactic phase transition to the oxygen-ion transport kinetics along different crystalline directions. The third stage consists of elucidating the switching mechanism and the kinetics of BM memristive devices with different crystalline orientations by means of different operando switching techniques. Putting all the gained knowledge on topotactic phase transitions together, we will derive design rules for memristive devices with improved switching kinetics.