Experimental demonstration of skyrmionic magnetic tunnel junction at room temperature
Sai Li, Ao Du, Yadong Wang, Xinran Wang, Xueying Zhang, Houyi Cheng, Wenlong Cai, Shiyang Lu, Kaihua Cao, Biao Pan, Na Lei, Wang Kang, Junming Liu, Albert Fert, Zhipeng Hou, Weisheng Zhao.
Science Bulletin, 2022, 67(7): 691–699
doi: 10.1016/j.scib.2022.01.016
01
简介
斯格明子是一种具有拓扑非平庸手性涡旋态的磁性结构, 因其纳米尺度和新奇电学调控特性, 在未来信息技术中有着重要应用前景. 在过去十年, 斯格明子的电学激发成核和驱动已经在实验上被证明, 然而与现行半导体工艺兼容的电学检测还没有实现, 这被认为是斯格明子应用进程中亟需填补的空白. 本文研究人员成功制备了基于斯格明子的隧道结器件, 并在实验上实现了室温下亚百纳米斯格明子的实空间观测及同步隧穿磁阻检测. 与霍尔效应相比, 斯格明子的隧穿磁阻检测方法使得信号强度提高了三个数量级以上. 该研究进一步通过脉冲电流实现了斯格明子非易失、多阻态调控, 并模拟人工突触和神经元器件功能. 该器件的成功研制将极大推进斯格明子的应用, 为高密度、超低功耗的类脑计算架构提供新的物理机制.
Fig. 1 Skyrmionic MTJ stack and experimental set-up. (a) Schematic of the MTJ stack structure with engineered interfacial DMI at Co and CoFeB interfaces respectively. (b) HAADF-STEM image of the stack and electron energy-loss spectroscopy (EELS) intensities of Co, Mg, Ta, Fe. (c) Magnetic hysteresis of the MTJ multilayer measured under the out-of-plane and in-plane magnetic fields. (d) MFM image of a typical MTJ multilayer sample taken at Typical skyrmions in the MFM images are enclosed by black dotted circles. Inset is the MFM signal (blue triangle) from a skyrmion marked by a black line. Gauss fitting (black line) indicates such a skyrmion has a diameter estimated by full width at half maximum value. (e) Scheme of the experimental set-up for TMR measurement and electrical pulse injection. A scanning electron microscope (SEM) image of the MTJ device with top and bottom electrodes. (f) Magnetoresistance curve of a MTJ device, which is confirmed by repeated measurements and different sizes of MTJ devices.
Fig. 2 Magnetic field evolution of RT skyrmions in MTJ device imaged by MFM and detected by TMR. (a) The red curve displays the TMR ratio in a magnetic field range from −100 to 100 mT. The insets i–ix are representative MFM images showing the nucleation and annihilation of skyrmions driven by the magnetic field. (b), (c) Magnetic field dependence of and of skyrmions at the negative (spin-up core) and positive field (spin-down core) range, respectively. (d), (e) Selected TMR ratios correspond to the data pointsin (b) and (c).
Fig. 3 Micromagnetic simulations of skyrmions and their TMR. (a) The schematic of the bilayer-coupled skyrmion in the MTJ. (b) Analysis of the domain–DW coupling (dotted arrows) within the two FM layers. The colored solid arrows indicated the magnetizations of the bilayer-skyrmions with the opposite chirality. (c) Out-of-plane component of the 100-nm skyrmion in Co and its corresponding stray field in CoFeB. (d) Series of the simulated magnetic domain distribution at selected fields, which indicates similar trend of magnetic textures to Fig. 2a. The color scale represents the . (e) Simulated TMR ratio based on the simulated field-dependent domain evolution in (d). The insets illustrate the spin configuration of the bilayer-skyrmions with different skyrmion number and helicity number in Co and CoFeB, where the arrow refers to the in-plane magnetic components ().
Fig. 4 Electrical characteristics and neuromorphic computing applications of the skyrmionic MTJ. (a) Measured resistance variation after one pulse injection and the induced-skyrmion density as a function of current density . (b) Measured resistance variation and the induced skyrmion density as a function of applied current pulse number (current duration is , current density is ). The insets in (a) and (b) display the MFM images of the skyrmion configuration correspondingly. (c) The MTJ array for the neuromorphic computing, where the skyrmionic MTJs behave as synapses and neurons. (d) The resistance variation is the synaptic weight and adjusted by increasing the current density at a fixed duration , mimicking the long-term potentiation of the synapse. (e) The resistance variation denotes the activation process of the neuron under a series of random current pulses. The reset of the both synapse and neuron could be performed by applying
03
本文通讯作者
侯志鹏 研究员 华南师范大学华南先进光电子研究院. 主要研究方向: 拓扑量子自旋功能材料与器件.
赵巍胜 教授 北京航空航天大学微电子学院. 主要研究方向: 自旋电子学、新型信息器件、非易失存储器等领域的交叉研究.