Magnetoresistive sensors can deliver high sensitivity and spatial resolution in magnetic field detection, ideal to leverage applications in navigation, robotics or biomedicine. Being able to manipulate the device’s properties, as needed and during operation, has the potential to push further the limits of performance, and broaden applications of spintronic sensors. In this talk, two approaches to engineer these tuneable magnetic sensors will be introduced. One focus on local modification of the sensing layer magnetic anisotropy by nano-inclusions, with direct impact on linearity and coercivity. The second explores light as an external trigger to change the reference axis orientation and thus alter the sensitive direction. The discussion will include considerations regarding the tuning of the thin-film multilayers or of the patterned geometries to demonstrate these concepts. Such strategies open pathways to design versatile sensing devices with on-demand tuneable characteristics.
​

Magnetic tunnel junction (MTJ) serves as functional devices for various applications. Magnetic sensors and nonvolatile memories with MTJs are commercially available and its applications to unconventional computing has recently gathered a growing attention. In this seminar, I will present our recent work on MTJs to be applied for nonvolatile memory compatible with deeply-scaled semiconductor technologies and probabilistic computing that addresses computationally hard problems. For nonvolatile memory applications, it is crucial to establish a technology to exhibit high performance at ultrasmall scale down to less than 10 nm, or X nm. We have shown that utilization of shape anisotropy [1] and magneto-static coupling [2] both of which become evident in X-nm regime allows the MTJs to meet the requirements for Flash-like [2] and SRAM-like [3] applications. For probabilistic computing, MTJs are designed to generate random telegraph noise under thermal fluctuations, constituting the probabilistic bits [4]. We have shown proof-of-concepts of the probabilistic computers with MTJs performing combinatorial optimization [5], machine learning [6], and quantum simulation [7]. We have also revealed key physics to enhancing the performance of MTJs for large-scale problems [8,9]. The work has been carried out in collaboration with H. Ohno, B. Jinnai, J. Igarashi, S. Kanai, W. A. Borders, K. Hayakawa, K. Kobayashi, H. Kaneko of Tohoku University as well as S. Datta of Purdue University and K. Y. Camsari of the UC Santa Barbara. The work is supported in part by the JST-OPERA JPMJOP1611, JST-CREST JPMJCR19K3, JST-AdCORP JPMJKB2305, MEXT X-NICS JPJ011438, and JSPS KAKENHI 19H05622.
[1] K. Watanabe et al., Nat. Commun. 9, 663 (2018).​
[2] B. Jinnai et al., IEDM2020, 24.6.1 (2020).
[3] B. Jinnai et al., IEDM2021, 2.6.1 (2021).
[4] K. Y. Camsari et al., Phys. Rev. X 7, 031014 (2017).
[5] W. A. Borders et al., Nature 573, 390 (2019).
[6] J. Kaiser et al., Phys. Rev. Appl. 17, 014016 (2022).
[7] A. Grimardi et al., IEEE IEDM 2022, 22.4 (2022). ​
[8] S. Kanai et al., Phys. Rev. B 103, 094423 (2021).
​ [9] K. Hayakawa et al., Phys. Rev. Lett. 126, 117202 (2021).​


Plus d'information :https://www.spintec.fr/seminars-two-seminars-on-magnetoresistive-sensors-and-magnetic-tunnel-junctions/​