Current-Induced Switching in Spin-Filter Tunnel Devices

The Internet of Things (IoT) devices needs to process a vast amount of data at high speed for smooth interfacing with other supplementary devices on the network. However, the current computational architecture is not efficient for this purpose. One of the potential solutions to tackle this issue is to use high-density nonvolatile memory (NVM). Magnetic random-access memory (MRAM) is one of the promising NVM technologies. The magnetic tunnel junction is the building block of the MRAM device. Intensive studies in the FM/B/FM MTJs, where a barrier (B) is separated by two ferromagnetic (FM) electrodes, have demonstrated that the relative orientation of two FM electrodes can be altered by either an external magnetic field or controlled by a spin-polarized current, i.e., the current-induced magnetization reversal via the spin-transfer (STT) or the field-like (FLST) components of the spin torque [1, 2]. The STT-MRAM technology has significant advantages over magnetic-field-switched MRAM, while the main challenge for implementing STT-MRAM is the requirement of high writing current for high-density and high-speed MRAM. Therefore, alternative writing and reading mechanisms for MTJs may provide a viable route towards switching energies per bit smaller compared with CMOS (~1 fJ). To achieve low-energy consumption in next-generation spintronics devices, much attention has shifted to the spin filter effect in ferromagnetic insulator europium chalcogenides [3] and perovskite oxides [4, 5]. So far, the spin-filter devices have been switched by the applied external magnetic field but to make a high-density memory solution, the current/voltage-induced switching would be required. Recently we have observed the current/voltage-dependent switching in SmSrMnO₃-based devices. The dI/dV vs. V of such spin filter device shows the large change in the dynamic conductance of the device at lower bias and low temperature (below T_cfor the ferromagnetic insulator barrier layer). The underlying mechanism of such current dependent switching is not well understood. We have conducted detailed experiments to get a comprehensive understanding of the current-driven switching dynamics in spin-filter tunnel devices. The physical understanding and theoretical model for the experimental observations are being analyzed by both first-principles calculations and a tight-binding model with a self-developed JunPy+LLG package [6, 7]. The current-driven switching in spin-filter tunnel junctions can potentially be used for high-speed and high-endurance non-volatile memory devices for AI and IoT applications.<br/> <br/> <br/><b>References:</b><br/>[1] S. S. P.Parkin et. al. Nat. Mater. 3,862 (2004).<br/>[2] S. Yusa, et. Al., Nat. Mater. 3, 868(2004).<br/>[3] J. S. Moodera et al., Phys. Today 63, 46 (2010)<br/>[4] B. Prasad et. al. Nano Lett. 14, 2789 (2014)<br/>[5] B. Prasad et. al. Adv.Mater. 27, 3079 (2015)<br/>[6] Y. -H. Tang and B. -H. Huang, J. Phys. Chem. C. 122, 20500 (2018).<br/>[7] Y. -H. Tang and B. -H. Huang, Phys. Rev. Research 3, 033264 (2021).