A three-dimensional multiscale peridynamics model for Ferroelectric/Multiferroic Tunnel Junctions

本文提出了铁电/多铁隧道连接处断裂的三维多尺度周动力学模型。铁电隧道结（FTJ）是一种铁电夹层层，用作电子隧穿的电极之间的阻挡材料。通过使用铁磁电极，ftj的功能可以扩展到创建多铁隧道结（MTJs）。FTJs/MTJs广泛应用于铁电存储器、微系统、可调谐微波、二进制数据存储、铁电随机存取存储器（FRAMs）等电子器件中。FTJ/ mtj的断裂会影响其可靠性。例如，它可能导致铁电薄膜所看到的电场的减小。在铁电薄膜中，一种常见的断裂机制是由极化开关引起的，并导致剩余可切换极化的系统性损失，这是双极开关循环次数的函数。在精细尺度上，需要建立FTJ/ mtj的分子动力学模型；在粗尺度上，需要建立基于状态的热-机械-电磁PD连续体模型。长度尺度的转移是通过PD模型和MD模型的高阶梯度（HOG）模型来完成的，从而使它们表现出相同的动态色散。PD模型相对于现有模型（如XFEM）的优势在于它易于处理不连续点，因为它们是分析的自然结果。不需要裂缝拓扑的表示。此外，由于PD模型的非局部特性，它可以更自然地与MD模型耦合。该方法将通过与现有实验数据的比较来验证和验证，并用于回答与FTJ/ mtj断裂行为相关的一些紧迫问题。

In this project, a three-dimensional multiscale peridynamics (PD) model for fracture in ferroelectric/multiferroic tunnel junctions is proposed. A ferroelectric tunnel junction (FTJ) is a ferroelectric sandwich layer used as a barrier material between electrodes for tunneling of electrones. By using ferromagnetic electrodes the functionality of FTJs can be extended to create multiferroic tunnel junctions (MTJs). FTJs/MTJs have been widely used in electronic devices such as ferroelectric memories, microsystems, tunable microwaves, binary data storage, ferroelectric random access memories (FRAMs), among others. Fracture in FTJ/MTJs can influence their reliability. For instance, it might lead to the reduction of the electric field seen by the ferroelectric thin-film. One popular fracture mechanism in ferroelectric thin-films is caused by the polarization switching and results in a systematic loss of the remanent switchable polarization as a function of the number of bipolar switching cycles. A molecular dynamics (MD) model for FTJ/MTJs should be devised at the fine scale while a state-based thermomechanical-electromagnetic PD continuum model needs to be developed at the coarse-scale. The transfer of length scales is accomplished through higher order gradient (HOG) models of the PD and the MD models so that they show the same dynamic dispersion. The advantage of the PD model over existing models such as XFEM is its ease in handling discontinuities as they are a natural outcome of the analysis. No representation of the crack topology is needed. Moreover, due to the non-local character of the PD model, it can be coupled to the MD model in a more natural way. The approach will be verified and validated by comparison to available experimental data and used in order to answer some urgent questions related to the fracture behavior in FTJ/MTJs.