Spin-orbit torque initialization of domain-wall and process damage cure in a domain-wall magnetoresistive device with a synthetic antiferromagnet

The current-induced motion of the magnetic domain wall (DW) in a synthetic antiferromagnet (SAF) has been reported to reach velocities of up to a few hundred m/s. By implementing this feature into three-terminal devices with sub-hundred nm sizes, it is possible to achieve higher density memory with SRAM-level nanosecond switching speeds. Two main hurdles for the realization of this type of device are (1) ensuring that the initial spin states on both ends of the device are anti-parallel and (2) minimizing the degradation of the SAF property during the ion beam etching process (IBE). Here, we present a new simple scheme of anti-parallel spin initialization and recovery process from the IBE damage. Employing these techniques on domain-wall magnetic tunnel junction (DW-MTJ) three-terminal devices, we validated numerous write operations with a domain wall velocity of about 73 m/s. Based on the results, when our device scales down to 100 nm in length, the expected write operation speed reaches 1.0 ns, which is comparable to SRAM’s speed. The speed of current-induced magnetic domain-wall motion in a synthetic antiferromagnet reaches a few hundred m/s, and by implementing it into memory devices, it is expected to realize higher-density memory with SRAM-level operation speeds. Two main hurdles are (1) ensuring that the initial spin states on both ends of the device are anti-parallel and (2) minimizing the property degradation during the etching process. Here, we present a new simple scheme of anti-parallel initialization and recovery process from the damage by post-annealing. Our report proves that these two obstacles can be solved, yielding a device with better performance than SOT-MRAM. Magnetoresistive random access memory (MRAM) is a promising technology for future memory devices. Researchers are exploring a type of MRAM called domain-wall magnetic tunnel junction (DW-MTJ) for its potential advantages over existing technologies. In this study, researchers developed new methods to improve DW-MTJ performance. They focused on two main challenges: initializing the device and repairing damage caused by ion beam etching. They used a synthetic antiferromagnet (SAF) structure with materials like CoFeB and MgO to achieve high performance. The team demonstrated that their methods improved device yield and reduced energy consumption. They showed that DW-MTJ could operate reliably with fast domain wall motion, which is crucial for data writing. The study concludes that these advancements make DW-MTJ a strong candidate for replacing current memory technologies like SRAM. This summary was initially drafted using artificial intelligence, then revised and fact-checked by the author.