Perpendicular shape anisotropy spin transfer torque magnetic random-access memory: towards sub-10 nm devices

Abstract

A new concept to increase the downsize scalability of perpendicular spin transfer torque magnetic random-access memory (p-STT-MRAM), called perpendicular shape anisotropy (PSA) STT-MRAM is presented. This approach consists of significantly increasing the thickness of the storage layer in p-STT-MRAM to values comparable to the cell diameter so as to induce a PSA in this layer which comes on top of the MgO/FeCoB interfacial anisotropy. This PSA-STT-MRAM is provided by depositing a thick ferromagnetic (FM) layer on top of an MgO/FeCoB based magnetic tunnel junction (MTJ) so that the thickness of the storage layer becomes of the order or larger than the diameter of the MTJ pillar. In contrast to conventional spin transfer torque (STT) magnetic random access memory, wherein the demagnetizing energy opposes the interfacial perpendicular magnetic anisotropy (iPMA), in these novel memory cells, both PSA and iPMA contributions favor out-of-plane orientation of the storage layer magnetization. Using thicker storage layers in these PSA-STT-MRAM has several advantages. Thanks to this robust source of bulk anisotropy, PSA-STT-MRAM offers a greatly improved downsize scalability over conventional perpendicular p-STT-MRAM. Despite the large thickness of the storage layer, PSA-STT-MRAM cells can still be written by STT provided their thermal stability factor Δ is adjusted in the same range as in conventional p-STT-MRAM, i.e. Δ of the order of 60–100 depending on the memory capacity and required bit error rate. Moreover, a low damping material can be used for the thick FM material, thus leading to a reduction of the write current. Thanks to the PSA, very high and easily tunable thermal stability factors can be achieved, even down to sub-10 nm diameters. The paper describes this new PSA-STT-MRAM concept, practical realization of such memory arrays, magnetic characterization demonstrating thermal stability factor above 200 for MTJs as small as 8 nm in diameter and the possibility to maintain thermal stability factor above 60 down to 4 nm diameter. We also show that thanks to the increased thickness of the storage layer, the anisotropy and therefore the memory retention are much less sensitive on temperature than in conventional p-STT-MRAM. This is very interesting for applications operating on a wide range of temperatures (e.g. automotive −40 °C to +150 °C), as well as to fulfill solder reflow compliance.

Current-driven transverse domain wall oscillations in perpendicular spin-valve structures

“…6). The interplay between interfacial STT, bulk STT, interlayer exchange, and magnetocrystalline anisotropy, discussed here in the context of STNOs, 5 may be additionally relevant to STT-MRAM designs employing elongated storage layers, which have recently been proposed as a route to enhancing lateral scalability of memory cells to sub-10nm scale [66][67][68] . Similar interactions are also expected to occur in analogous three-terminal systems involving a current path running horizontally through a non-magnetic heavy-metal layer connecting to the free layer in place of the polarizer-spacer duo 55,[69][70][71][72][73][74][75] .…”

Cylindrical Magnetic Nanowires Applications

“…This effect has been widely studied for NWs [103], [108], [122], [207], [208] and could now be useful for perpendicular anisotropy magnetic tunnel junction (p-MTJ) based memories [209], [210] as their packing density increases. In these p-MTJs, a thick ferromagnetic (storage) layer is deposited on top of an MTJ so that its thickness is in the order of or larger than the diameter of the MTJ [209], [211] or the thickness of the free layer of a MTJ is increased until the shape anisotropy dominates [210]. This combination of perpendicular shape anisotropy and interfacial perpendicular magnetic anisotropy increases the device's thermal stability, i.e., the minimum energy required to switch the memory between two states well above room temperature activation.…”

Memristive Devices for New Computing Paradigms

“…The emerging memory technologies include various types of memristors. In this section, four types of memristors are discussed: resistive switching, [11][12][13][14][15] phase-change, [16][17][18][19] spintronics, [20][21][22] and ferroelectric. [23][24][25] 2.1.…”