Abstract

Exchange bias is a well-known and very useful effect that occurs due to an interfacial coupling between adjacent thin films of an antiferromagnet and a ferro(ferri)magnet [1]. Controlling exchange bias in an antiferromagnet/ferro(ferri)magnet thin film system by application of a gate voltage is much sought-after for spintronic applications. However, since the exchange bias effect is related to the ordering of spins in both the antiferromagnet and the ferro(ferri)magnet, and electric fields have little influence over the spin order, electrical control of exchange bias is very hard to achieve practically.Recently, ionic gating of magnetic heterostructures, where a gate voltage is used to drive ions (O2-, H+ etc.) in or out of the active layers, has yielded very interesting and promising results. Due to the size of the H+ ion, protonic gating is potentially faster and more reversible compared to gating with other ions and is also able to cause large changes in magnetic properties in the underlying active layers, such as modulation of perpendicular magnetic anisotropy in ultra-thin Co films [2]. The utility of protonic gating was further demonstrated when it was shown that ferrimagnetic order in GdCo metal thin films can be dynamically controlled by this method [3].GdCo is comprised of two sub-lattices, one made up of Gd atoms and one of Co atoms, and the magnetic moments in each sub-lattice point in opposite directions, hence resulting in ferrimagnetic order. Experimental and theoretical analyses show that the protons being inserted into the GdCo reduce the effective magnetic moment of the Gd atoms, and by starting from a Gd-dominated magnetic state, the overall magnetization can be switched 180 degrees through proton gating.Harnessing this functionality, we have shown, in this work, that exchange bias toggling is possible when we incorporate the GdCo ferrimagnet in an exchange biased heterostructure with NiO serving as the antiferromagnet. Using sputter deposition we have engineered a stack with the structure NiO(33)/Pd(1)/GdCo(10)/Pd(10)/GdOx(30)/Au(3) on thermalized Si, which has fully shifted loops, i.e. the exchange bias field is larger than the coercive field (HEB>HC), as depicted in Fig. 1(a). When a sufficiently high (>1.5V) positive voltage is applied to the Au top electrode (Pd serves as the bottom electrode), atmospheric water is split and the proton is transported through the GdOx electrolyte, into the GdCo(10)/Pd(10) layers. Starting from a Gd-dominated state of the GdCo, upon enough insertion of hydron ions after applying +1.8V, the dominant sublattice of GdCo switches from Gd to Co, as expected from previous work, but is now also accompanied by a toggling of the exchange bias direction, as shown in Fig. 1(b). The effect is fully reversible, as shown in Fig. 1(c).We can understand this phenomena when we realize that although the gating does nothing to the relative spin orientation in either the ferrimagnet or the antiferromagnet, the ferrimagnet’s magnetization direction depends not only on the spin orientation of the sublattices but also on the relative magnitude. The magnetization direction flips 180 degrees during gating but the relative spin orientation across the interface remains the same. Essentially, if initially the pinned magnetization is +Mz (meaning HEB is -ve) then after gating the pinned magnetization will be -Mz necessarily making HEB +ve. To the best of our knowledge, this is the simplest demonstration of exchange bias toggling to date, since our system does not rely on any complex materials or growth processes and works at room temperature, as opposed to some previous works [4,5]. Due to the small size of the proton, our method is highly reversible, and initially we have shown full cycling toggling of the exchange bias up to 10 times, where the hysteresis loop remains fully shifted throughout the process (Fig. 2) as evidenced by HEB/HC being over 1.We believe our simple but robust method to toggle exchange bias can serve as a platform for demonstrating more functionalities and/or devices and can have broad implications for spintronics in general.This work is supported in part by SMART, one of seven centers of nCORE, a Semiconductor Research Corporation program, sponsored by National Institute of Standards and Technology (NIST). **

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call