Magnetization reversal in sub-100 nm magnetic tunnel junctions with ultrathin MgO barrier biased along the hard axis

Abstract

We report on room temperature magnetoresistance and low frequency noise in sub-100 nm elliptic CoFeB/MgO/CoFeB magnetic tunnel junctions with ultrathin (0.9 nm) barriers. For magnetic fields applied along the hard axis, we observe current induced magnetization switching between the antiparallel and parallel alignments at dc current densities as low as 4 Â 10 6 A/cm 2 . We attribute the low value of the critical current to the influence of localized reductions in the tunnel barrier, which affects the current distribution. The analysis of random telegraph noise, which appears in the field interval near a magnetization switch, provides an estimate to the dimension of the pseudo pinholes that trigger the magnetization switching via local spin torque. Micromagnetic simulations qualitatively and quantitatively reproduce the main experimental observations. Slonczewski's 1 and Berger's 2 prediction that a spinpolarized current between two ferromagnets could produce spin torque (ST) and, in turn, create steady magnetization dynamics or induce a magnetization reversal has been now widely confirmed experimentally. Spin torque magnetic random access memories (ST-RAM) 3 or microwave oscillators 4 are just two examples from a number of possible future spintronic devices based on spin torque. A key step towards the implementation of magnetic tunnel junctions (MTJs) as MRAM elements is the reduction of the critical currents/voltages needed for ST switching, in order to avoid heating and back-switching. 5 Another important field of potential applications of MTJs is MRAM-like biosensing chips based on arrays of magnetic sensors, 6 scanning MTJ microscopy, 7 etc. These new applications rely on the high external field sensitivity of MTJ sensors 8 when the external bias field direction is collinear with the hard axis (HA). 9 In order to improve their spatial resolution, one should shrink these MTJ sensors in lateral size and consequently substantially decrease the MgO barrier thickness to keep the junction resistance reasonably small. This, however, may contribute to an enhanced influence of ST effects and of barrier inhomogenieties. With few exceptions, 10 spin torque switching in HA biased MTJs remains poorly understood.

Solving the stochastic Landau-Lifshitz-Gilbert-Slonczewski equation for monodomain nanomagnets : A survey and analysis of numerical techniques

“…where γ is the gyromagnetic ratio of electron, H ef f is the effective magnetic field experienced by the nanomagnet, α is the dimensionless Gilbert damping constant, M s is the saturation magnetization of the nanomagnet, I s is the applied spin current, q is the elementary charge, N s is the number of spins given as N s = 2MsV γ , and V is the volume of the nanomagnet. The nanomagnets used in spintronic devices and memories today are sub-100 nm [13,14]; for such small dimensions the macrospin approximation [15] validates the use of the monodomain model, wherein the nanomagnet is assumed to possess a single coherent magnetization (M ) over the entire domain and any spatial variation in M is not considered. This approximation is only valid for nanomagnets of dimensions smaller than the critical Stoner radius, which is of the order of a few 100s of nm for typical magnetic materials [15].…”

Low frequency noise peak near magnon emission energy in magnetic tunnel junctions

“…3(c), ∆V for this particular TLF is about 2.5 µV and doesn't change much from 13 mV to 21 mV, inconsistent with a simple defect picture, where domain rotation or channel switching leads to fixed resistance change. 19,26 In fact almost fixed ∆V can also be extracted in previous reported cases, 11,12,23 although for measurements at high temperatures TLFs are not stable and correlated with each other, thus the Lorentzian evolves to 1/f shape and detailed fitting with the activation model is not possible. How the fixed ∆V is related to magnon dynamics needs further investigation.…”

Micromagnetic Simulation of Strain-Assisted Current-Induced Magnetization Switching

“…The last two terms on the right side of (1) describe STT that tends to drag the magnetization away from its initial state to its final state. The scalar function is given by [39,40] The magnetic parameters employed in the simulations are as follows: saturation magnetization = 9.549 × 10 5 A/m [44,45], Gilbert damping parameter = 0.00439 [46], spin polarization factor = 0.5 [47], magnetic anisotropy constants 1 = 1.2 × 10 4 J/m 3 and 2 = 0 [37], elastic constants 11 = 2.57 × 10 11 Nm −2 , 12 = 1.62 × 10 11 Nm −2 , and 44 = 1.05 × 10 11 Nm −2 [37], and magnetostrictive constants 100 = 139 ppm and 111 = 22ppm [44,45]. We investigated the influence of normal substrate strains 11 and 22 on the magnetization state by assuming a zero shear strain.…”