Out-of-plane magnetoresistance in ferromagnet/graphene/ferromagnet spin-valve junctions

Out-of-plane spin-injection and detection through naturally stacked graphene layers were investigated in ferromagnet/graphene/ferromagnet (FGF) junctions. We obtained a maximum magnetoresistance (MR) of 4.6% at $T=4.2$ K in the junction of a four-layer graphene insertion, having a very small area$-$junction-resistance product of 0.2 $\Omega \mu$m${}^2$. According to resistance-temperature and current-voltage characteristics, the graphene layer in the FGF junction acted as a metal-like insertion rather than as an insulating barrier. A lower value for the interfacial spin asymmetry coefficient ($\gamma =0.25\pm 0.05$) obtained from the fitting of variations with interfacial resistance implies that the spin-injection efficiency along the out-of-plane direction was reduced by spin-flip scattering at graphene/ferromagnet interfaces. Our results showed that highly transparent graphene/ferromagnet interfaces with crystalline ferromagnet (FM) electrodes are required to achieve higher spin-injection efficiency through the graphene layer in a FGF junction along the out-of-plane direction.

Figure 1

(a) A conceptual drawing for spin-dependent transport in a vertically stacked FGF junction. Scanning electron microscopy image of the junction. (b) Side view of the junction. (c) Top view of the junction with the measurement configuration shown.

Figure 2

Sample fabrication processes. (a) Exfoliation of natural graphene on an oxidized Si substrate coated with a PSS/LOR layer. (b) Metal deposition (Py/Au; 30/5 nm) for the bottom electrode. (c) Dissolution of the PSS layer in water to separate the LOR layer from the Si substrate. (d) The LOR layer flipped onto a separate Si substrate. (e) Elimination of the LOR layer. (f) Metal deposition (Py/Au; 50/5 nm) for the top electrode. (g) Addition of an insulation layer with cross-linked PMMA. (h) Deposition of Au electrodes (Au; 250 nm).

Figure 3

Comparison between transparent and resistive junctions. Hysteretic magnetoresistance (MR) curves for varying the magnetic field ($H$) at $T=4.2$ K and $I=50$ $\mu$A for (a) G1-t and G1-r, (b) G2-t and G2-r, and (c) G4-t and G4-r1. Insets show the temperature dependence of the junction resistance for the corresponding samples with applied current $I=50$ $\mu$A. Transparent (G1-t, G2-t, and G4-t) and resistive (G1-r, G2-r, and G4-r1) junctions are denoted as solid and dotted lines, respectively.

Figure 4

(a) Temperature dependence of MR for samples G4-t, G4-r1, G1-t, G2-t, G3-r, and G4-r2. Solid lines show the results for linear fitting. (b) $I$-$V$ curves for the measured junctions. (c) Current bias dependence of MR for corresponding $I$-$V$ curves. The same symbol legends are used for all figures.

Figure 5

Interfacial resistance dependence of MR for various FGF junctions. Square, circle, upright-triangle, and inverted-triangle symbols indicate the MR for FGF junctions with one-, two-, three-, and four-layer graphene insertion, respectively. Best-fit curves are obtained for $\beta =0.4$ (solid) and 0.3 (dotted) for three different values of $\gamma$: $\gamma =0.3$ (upper), 0.25 (middle), and 0.2 (lower). We adopted estimated values of $l_{sf}^N\simeq 100$ nm, ${\rho }_g\simeq 1\times {10}^{-3}\Omega$cm, and ${\rho }_{\mathrm{Py}}\simeq 1.2\times {10}^{-5}\Omega$cm for the fit. The horizontal axis is plotted as a logarithmic scale of $A{R}_{\mathrm{cont}}$ for clarity. The inset shows the variation of MR for a different number of graphene layers.