搜索

x
中国物理学会期刊

考虑内部偏置场非均匀分布的磁隧道结传感曲线模拟方法

Modeling of the Transfer Curves of Magnetic Tunnel Junctions Considering the Non-Uniformity of Internal Bias Field

PDF
导出引用
  • 精确模拟传感曲线对隧穿磁阻传感器设计至关重要。对于高轴比磁隧道结,受边缘效应及静磁耦合影响,其内部偏置场呈现显著的非均匀分布特性。作者前期建立的基于均匀有效场假设的各向异性磁阻模型在描述常规器件特性方面已被证实有效,本文进一步提出一种考虑内部偏置场非均匀分布的传感曲线模拟方法。通过引入离散偏置场分布,修正了对自由层微磁行为的描述。对4 μm×16 μm椭圆形磁隧道结的实验验证表明,非均匀模型有效引入了自由层磁化反转过程中的非一致性特征,消除了均匀模型预测磁阻的异常尖峰现象,精确复现了难轴附近高灵敏度工作区的传感特性。该方法显著提升了模型的物理普适性,为高灵敏度隧穿磁阻传感器优化设计提供了精准的理论工具。

     

    Accurate modeling of transfer curves is a fundamental requirement for the sophisticated design of Tunneling Magnetoresistance (TMR) sensors, which have emerged as leading candidates for weak magnetic field detection due to their superior sensitivity and wideband response. To eliminate hysteresis and achieve high linearity, TMR sensors frequently utilize magnetic tunnel junctions (MTJs) with high-aspect-ratio elliptical geometries. This design strategy leverages strong longitudinal shape anisotropy to stabilize the magnetization state; however, it inevitably introduces a significant spatial non-uniformity in the internal bias field. This non-uniformity is primarily driven by complex edge demagnetizing effects and magnetostatic coupling between the ferromagnet layers, which are often oversimplified in standard analytical models.
    While our previous anisotropic magnetoresistance model based on a uniform effective field hypothesis has proven effective for standard devices, it exhibits significant physical distortion when applied to high-aspect-ratio MTJs, resulting in inaccurate predictions in non-ideal conditions. To overcome these limitations, this paper proposes an enhanced method that incorporates the non-uniform distribution of the internal bias field. The proposed methodology centers on a deterministic discrete bias field distribution. The free layer is discretized into multiple regions represented by three distinct groups of hysterons. Each group is subjected to a unique internal bias field consisting of a static invariant component and a spatially-dependent variable component, thereby capturing the field gradient from the device center to its boundaries. Furthermore, the macroscopic resistance is derived by synthesizing the parallel tunneling conductance of these local regions. A parasitic, tunnel-independent resistance is integrated into the model to represent the non-magnetic contributions from electrodes and leads in the junctions array, ensuring a rigorous match with experimental conditions.
    Experimental verification conducted on 4 µm×16 µm elliptical MTJs demonstrates that the traditional uniform model produces anomalous resistance peak at an external field angle of 120° and broad plateaus at 140°. These discrepancies are fundamentally caused by the abrupt coherent rotation of the entire free layer, which triggered by the synchronized competition between external and uniform internal fields. In contrast, the proposed non-uniform model effectively captures the non-coherent characteristics of the magnetization reversal process. By accounting for spatial non-uniformity, the model accurately predicts the asynchronous reversal across different regions of the free layer. This leads to a phase cancellation effect that suppresses the non-physical resistance peak and reproduces the smooth transfer characteristics observed in high-sensitivity working regions of MTJs. Additionally, the model provides a unique physical insight into the slight increase in modeling error observed near the equivalent hard axis (approximately 110°), where the system physically approaches an ideal single-domain state. In conclusion, this research establishes a powerful theoretical bridge between microscopic magnetic dynamics and macroscopic sensor outputs, providing a precise theoretical tool for the optimization and performance prediction of high-sensitivity TMR sensors.

     

    目录

    /

    返回文章
    返回