| 研究実績の概要 |
This work aimed to clarify how to decrease magnetic sensing noise in magnetic tunnel junction (MTJ) sensors by 1000 times.
(1) We analyzed the microscopic origins of magnetic noise using magnetic domain images. We developed magneto‐optical Kerr effect (MOKE) microscope with a low‐noise scientific CMOS camera and a stabilized light source. We explored the domain evolution in micro-sized permalloy islands at low magnetic fields. At a constant magnetic field, we found that the vortex-antivortex pairs in cross-tie wall form and annihilate continuously in random, with a time scale of 1 second. This type of low-frequency noise is a major contribution to sensor noise.
(2) We developed three new sensing layers (2-a,b,c) using sputtering growth and advanced micro-fabrication. (2-a) We optimized stoichiometric high‐quality Y3Fe5O12 (YIG) films on Si and Gd3Ga5O12 substrates. We made polycrystalline crack-free YIG films with roughness Ra<0.4 nm. (2-b) We optimized growth and exchange-bias in L10-PtMn antiferromagnet. We made MTJ sensors that operate at high temperatures (>250℃). (2-c) We developed vortex-type MTJ sensors. We showed that the design of vortex-layer and pin-layer diameters and relative positions can be used to increase the sensitivity.
We also used our results to: develop new algorithms for noise processing based on deep-learning approaches, and investigate spin-caloric effects in YIG/PtMn thin films.
As research output, we published: three papers (two more accepted), and four international conferences presentation (two more conf. have been canceled due to the pandemic).
|
