TY - GEN
T1 - Three-terminal spintronics devices for CMOS integration
AU - Ohno, Hideo
N1 - Funding Information:
I thank all my collaborators especially Shunsuke Fukami. Work supported in part by ImPACT from JST and by the R & D for Next-Generation Information Technology of MEXT.
Publisher Copyright:
© 2017 IEEE.
PY - 2017/8/1
Y1 - 2017/8/1
N2 - I review our recent study on realizing three-terminal spintronics devices for integration with CMOS VLSI. Three-terminal devices separate the current paths one for magnetization switching and the other for reading the state of magnetization, thereby, in principle, allow a relaxed operation window resulting in high speed operation compared to their two-terminal counterpart [1]. Of particular current interest are devices that utilize spin-orbit torque (SOT) for magnetization switching, which consist of a heavy metal layer and a target ferromagnetic structure placed on top of it. The first topic I discuss is a high-speed operation of an SOT switching with a target ferromagnetic pillar having an inplane magnetic easy axis collinear with the current flow direction in the underneath heavy-metal [2, 3]. We show that one can switch magnetization as fast as 500 ps in this structure without significant increase in switching current; this switching speed is not readily available in two-terminal devices utilizing spin-transfer torque (STT) switching without applying considerably large current - STT requires switching current inversely proportional to the switching speed in this speed range. We also show that there is a device configuration to avoid application of static magnetic field otherwise needed to observe switching. The second topic I will discuss is to use an antiferromagnetic layer as a source of spin flow as well as the source of an exchange field: The former is for the switching and the latter is for the switching in the absence of external magnetic field. It was shown in a (Co/Ni)-multilayer/PtMn structure one can switch magnetization in the absence of external magnetic field [4]. The use of antiferromagnet led us to realize an analog memory, which we recently used to demonstrate an associative memory operation in a spintronics-device based artificial neural network [5].
AB - I review our recent study on realizing three-terminal spintronics devices for integration with CMOS VLSI. Three-terminal devices separate the current paths one for magnetization switching and the other for reading the state of magnetization, thereby, in principle, allow a relaxed operation window resulting in high speed operation compared to their two-terminal counterpart [1]. Of particular current interest are devices that utilize spin-orbit torque (SOT) for magnetization switching, which consist of a heavy metal layer and a target ferromagnetic structure placed on top of it. The first topic I discuss is a high-speed operation of an SOT switching with a target ferromagnetic pillar having an inplane magnetic easy axis collinear with the current flow direction in the underneath heavy-metal [2, 3]. We show that one can switch magnetization as fast as 500 ps in this structure without significant increase in switching current; this switching speed is not readily available in two-terminal devices utilizing spin-transfer torque (STT) switching without applying considerably large current - STT requires switching current inversely proportional to the switching speed in this speed range. We also show that there is a device configuration to avoid application of static magnetic field otherwise needed to observe switching. The second topic I will discuss is to use an antiferromagnetic layer as a source of spin flow as well as the source of an exchange field: The former is for the switching and the latter is for the switching in the absence of external magnetic field. It was shown in a (Co/Ni)-multilayer/PtMn structure one can switch magnetization in the absence of external magnetic field [4]. The use of antiferromagnet led us to realize an analog memory, which we recently used to demonstrate an associative memory operation in a spintronics-device based artificial neural network [5].
UR - http://www.scopus.com/inward/record.url?scp=85028070310&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85028070310&partnerID=8YFLogxK
U2 - 10.1109/DRC.2017.7999490
DO - 10.1109/DRC.2017.7999490
M3 - Conference contribution
AN - SCOPUS:85028070310
T3 - Device Research Conference - Conference Digest, DRC
BT - 75th Annual Device Research Conference, DRC 2017
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 75th Annual Device Research Conference, DRC 2017
Y2 - 25 June 2017 through 28 June 2017
ER -