Determining switching exponents with stochastic magnetic tunnel junctions*

The stability of physical systems under external perturbation is often described by local bifurcation and/or switching exponents, and access to them is essential to establish the reliable operation of solid-state devices. We investigate the exponents of spin transition torque (STT) and perpendicular magnetic field H in nanoscale magnetic tunnel junctions (MTJs) with a perpendicular easy axis [1]. By combining three dynamical measurements in a stochastic MTJ, we experimentally determine the switching exponents, which have not been experimentally accessed before [2].

We utilize an MTJ with a diameter of 34 nm and a relaxation time of ~3 ms. We investigate the thermal stability factor Δ(V,H) = Δ₀(1-V/V_C0)^nV(1±H/H_K^eff)^nH, where Δ₀ is Δ without STT and H, V_C0 the critical voltage V for the STT switching, H_K^eff the effective anisotropy field, and n_H and n_V the switching exponents of the STT and H, respectively. We determine n_V and n_H using the homodyne-detected ferromagnetic resonances, switching probabilities, and random telegraph noise measurements, and conclude that both n_H and n_V switch from 2 to 1.5 with reducing H_K^eff, which is well explained by the transition of the local bifurcation structure of the magnetic potential landscape from the pitchfork bifurcation to the saddle-node bifurcation [2].

[1] S. Ikeda et al., Nat. Mater. 9, 721 (2010). [2] T. Funatsu et al., Nat. Comm. 13, 4079 (2022).