Thickness dependence and the role of spin transfer torque in nonlinear giant magnetoresistance of permalloy dual spin valves

Abstract

Phys. Rev. Lett. 103, 237203 ͑2009͔͒ opens up the possibility for distinct physics which extends the standard model of giant magnetoresistance. When the outer ferromagnetic layers of a dual spin valve are antiparallel, the resulting accumulation of spin in the middle ferromagnetic layer strongly modifies its bulk and interfacial spin asymmetry and resistance. Here, we report experimental evidence of the role of bulk spin accumulation in this nonlinear effect and show that interfacial spin accumulation alone cannot account for the observed dependence of the effect on the thickness of the middle ferromagnetic layer. It is also shown that spin torque acting on the middle ferromagnetic layer combined with the nonlinear effect might be useful in understanding the dynamical features associated with the nonlinear behavior.

Band-structure-dependent nonlinear giant magnetoresistance in Ni1−xFexdual spin valves

“…It can be seen that, while this is true at zero bias, at finite current, reversing either the central magnetic layer or the current direction causes a resistance change, this difference becoming more pronounced with increasing current magnitude. This is the nonlinear giant magnetoresistance effect introduced previously, [6][7][8] and it is visible in DSV devices primarily because the conventional GMR effect is canceled by the AP orientation of the outer magnetic layers. Before we consider the implications of the results shown in Fig.…”

“…A detailed description of the magnetic states in a DSV is given in Ref. 7. The most relevant configurations for this study are the antiparallel (AP) states (↑↑↓ and ↑↓↓) in which the magnetically soft FM2 layer is switched between opposite directions by a magnetic field which is too small to disturb the AP orientation of the outer layers.…”

“…These changes are too small to perturb the overall magnetic configuration and only minimally affect the associated resistance. The strongest proof in favor of the fact that the resistance changes do not originate from Oersted field effects comes from our previous 6,7 consistent observation of significant resistance modification between ↑↑↓ and ↑↓↓ states with a bias current of as low as The field distribution for a current of 1 mA with the color scale denoting the magnitude of the magnetic field generated by the current in units of mT; the maximum field (∼1.5 mT) is at the middle of the edges of the pillar, and the bulk of the layer experiences a field of less than 1 mT.…”

“…Previous experiments with such structures have revealed a current dependence of the GMR for large current densities 6,7 which requires an extension of standard VF theory to explain. Baláž and Barnaś have shown 8 theoretically that the experimental results can be reproduced by adding a linear current dependence to the values of β and γ associated with the FM2 layer.…”

“…Using standard optical lithography and Ar-ion milling, a series of 4-μm-wide tracks were processed, each connected to four large area contact pads for four terminal electrical measurements. By using a focused ion beam (FIB) microscope, nanopillar devices were fabricated in the tracks (as described elsewhere) 7,9,10 with an average dimension of 250 × 250 nm 2 . Devices were then transferred to a custom-built probe for electrical measurements.…”