Structure-property relations for manganite memristive devices

锰氧化物异质结显示出很好的电阻开关特性和多能级阻态。这使得它们成为替代非易失性存储器的理想候选者，但也可以作为神经形态计算的构建块。与更常见的丝状开关相比，锰氧化物器件已被证明在整个器件区域内均匀开关，因此在细胞到细胞和周期到周期的变化方面可能更好。此外，这些材料中的电子和离子输运可以通过改变材料的组成和微观结构来调节，这可能直接影响开关性能。虽然离子输运在开关机制中起着关键作用，但仍有许多悬而未决的问题需要了解。特别是，晶界(存在于与CMOS兼容的多晶锰氧化物器件中)显著影响了这些材料中的离子传输，但它们对电阻开关的影响尚未被直接研究。为了全面了解影响多晶锰氧化物薄膜器件开关性能的不同方面，我们将利用三个国际领先的锰氧化物离子传输小组(格勒诺布尔，亚琛)和电阻开关器件小组(Jülich)的互补专业知识。我们将通过使用双晶衬底来制备具有不同化学成分和明确晶界的外延薄膜模型系统。我们将全面研究晶界对离子输运和薄膜开关特性的影响。薄膜和器件的化学成分和结构将通过联盟内部或不同欧洲同步加速器设施(例如Soleil、ESRF和BESSY)提供的各种技术进行研究。这些技术包括表面分析技术，如扫描探针技术和光电子能谱，以及体敏感技术，如拉曼光谱或X射线吸收光谱。对开关器件进行操作光谱分析将使我们能够深入了解器件运行过程中发生的化学和结构变化。氧在不同薄膜结构中的扩散和表面交换将通过18O示踪扩散实验结合飞行时间二次离子质谱仪或拉曼光谱来研究。这些实验研究将得到氧在体相和晶界模拟单元中扩散的分子静态和分子动力学模拟的补充。这将使我们能够揭示锰氧化物记忆器件的微观结构、化学成分、离子和电子输运以及开关性能之间的复杂相互作用。在此基础上，我们将开发新的工艺路线来制造具有高可靠性和改进的开关动力学的与CMOS兼容的锰氧化物微器件。

Manganite heterostructures show very promising resistive switching characteristics and multilevel resistance states. This makes them ideal candidates for alternative non-volatile memories, but also as building blocks for neuromorphic computation. In contrast to the more common filamentary switching, manganite devices have been shown to switch homogeneously over the whole device area and might therefore be superior with respect to their cell-to-cell and cycle-to-cycle variation. Moreover, electronic and ionic transport in these materials can be tuned by varying the composition and microstructure, which could directly affect the switching performance. Although it is clear that ion transport plays a key role in the switching mechanism, many open questions are still to be understood. In particular, grain boundaries (present in CMOS-compatible polycrystalline manganite devices) significantly influence ionic transport in these materials, but their impact on resistive switching has not been directly studied yet. In order to gain an integral view of the different aspects influencing the switching properties of polycrystalline manganite thin film devices, we will utilise the complementary expertise of 3 internationally leading groups on ionic transport in manganites (Grenoble, Aachen) and resistive switching devices (Jülich). We will fabricate epitaxial thin film model systems with different chemical compositions and well-defined grain boundaries, through the use of bicrystal substrates. The impact of the grain boundaries on the ion transport and the switching properties of the films will be comprehensively studied. The chemical composition and the structure of films and devices will be investigated by the large variety of techniques available within the consortium or at different European synchrotron facilities (e.g. SOLEIL, ESRF and BESSY). These include surface analysis techniques, such as scanning probe techniques and photoelectron spectroscopy, as well as bulk sensitive techniques, such as Raman spectroscopy or X-ray absorption spectroscopy. Performing operando spectroscopy of switching devices will enable us to gain insights into the chemical and structural changes taking place during device operation. Oxygen diffusion and surface exchange in different thin-film configurations will be investigated by 18O tracer diffusion experiments in combination with time-of-flight secondary ion mass spectrometry or Raman spectroscopy. These experimental studies will be complemented by molecular static and molecular dynamics simulations of oxygen diffusion in bulk and grain-boundary simulation cells. This will enable us to uncover the complex interplay between microstructure, chemical composition, ionic and electronic transport and the switching performance of manganite memristive devices. Based on this, we will develop new routes for the fabrication of CMOS-compatible manganite micro-devices with high reliability and improved switching kinetics.