Fabrication of Graphene-Inserted Tunneling Device (GiTD) for Emerging Spin Devices

Abstract

Objective:Spintronics, which utilizes the spin degree of freedom of elementary particles, holds the promise for many attractive applications including non-volatile logic and storage. However, the main challenge is to achieve spin manipulation (spin injection, transport and detection) with high efficiencies. Here in this work, we focus on spin injection from the nanomagnet to another layer, which is typically degraded by any conductivity mismatch and/or interfacial imperfections. Inserting a tunnel barrier between the ferromagnetic and channel layers has shown to relieve this problem.The best way to quantitatively study spin polarization of a ferromagnetic thin film that is modified by any spin flipping at the interfaces is to directly measure the tunneling spin polarization across the ferromagnet and the tunnel barrier using superconducting tunneling spectroscopy (STS) [1]. One practical challenge in using this technique to investigate a role of any specific material of interest (any insertion layer other than the ferromagnet itself) in terms of spin polarization is to fabricate such a multi-layered device structure. The difficulty mainly arises from oxidation of underlying metallic layers.Recently, graphene has been considered as a promising material for spintronic applications. Advantages of graphene include weak spin-orbit coupling and low lattice mismatch with conventional ferromagnets such as Permalloy (Ni80Fe20). This leads researchers to re-think graphene as the long-range spin transport channel or the efficient tunnel barrier material(or both) [2]. The primary objective of this work is to demonstrate the graphene-inserted tunneling device (GiTD) that enables subsequent STS measurements to study how graphene impacts tunneling spin polarization and ultimately use it to enhance the spin injection efficiency in spin devices.New results:Fig. 1 illustrates a schematic device structure (GiTD) investigated in this work. Initially, a superconducting thin layer (5 nm) of Al was deposited by sputtering. A CVD-grown monolayer graphene was then transferred on top, which was followed by deposition of Permalloy (Py). Because a thin tunnel barrier of AlOx is easily formed by exposing the sample to ambient air, we study the role of graphene in the F/I/S device stack (F: ferromagnet, I: insulator, S: superconductor) with a native thin oxide formed at the graphene/S-layer interface. We patterned the graphene layer by photolithography and O2-plasma etching to avoid an electrical short between the top and bottom layers in the cross-bar geometry. Details of the process flow are depicted in Fig. 2, and a micrograph image of the fabricated GiTD structure (mounted on the STS holder) is shown in Fig. 3. Raman spectroscopy [3] was also conducted to confirm both the presence and absence of graphene in respective regions (Fig. 4).One important piece of innovation introduced in this work is an “offset contact” method that was developed to ensure any unnecessary oxidation of contact electrodes does not adversely impact the electrical testing. Since the STS measurement requires the sample to be mounted on a chip carrier wire-bonded to separate contact pads on the substrate, deposition of any additional metallic layer (known to facilitate the wire bonding process) may result in oxide formation on the underlying contact layer unless all constituent layers are deposited in-situ. As seen in Fig. 2 (steps (2) through (7)), we deposited first a thin film of Py and then put the S-layer (Al) with a slight offset to Py. By depositing another Py layer on top, a direct link between the top Py and bottom Py was established to bypass the oxidized Al layer in between. Fig. 5 shows the current-voltage (I-V) characteristic of the fabricated control device (Py/AlOx/Al) measured at room temperature, clearly demonstrating that (1) current flows well through the multi-layered contact pads and (2) memristor-like switching behavior is observed due to formation of a thin metal oxide (AlOx).Significance:By conducting the STS measurements that compare the control device with the graphene-inserted device (GiTD), the role of graphene in terms of tunneling spin polarization will be quantitatively studied. This will enable us to develop the novel tunnel barrier material, possibly consisting of 2D graphitic nanostructures. It will be best interfaced with both the ferromagnetic contact (spin injector or detector) and the graphene-based spin transport channel. Ultimately, this study will greatly impact the development of active spin devices such as spin FET where achieving the high spin injection still remains the biggest challenge.