Observation of Long Spin-Relaxation Times in Bilayer Graphene at Room Temperature

We report on the first systematic study of spin transport in bilayer graphene (BLG) as a function of mobility, minimum conductivity, charge density, and temperature. The spin-relaxation time ${\tau }_s$ scales inversely with the mobility $\mu$ of BLG samples both at room temperature (RT) and at low temperature (LT). This indicates the importance of D’yakonov-Perel’ spin scattering in BLG. Spin-relaxation times of up to 2 ns at RT are observed in samples with the lowest mobility. These times are an order of magnitude longer than any values previously reported for single-layer graphene (SLG). We discuss the role of intrinsic and extrinsic factors that could lead to the dominance of D’yakonov-Perel’ spin scattering in BLG. In comparison to SLG, significant changes in the carrier density dependence of ${\tau }_s$ are observed as a function of temperature.

Figure 1

(a) AFM image of a BLG sample after MgO deposition: rms roughness $\sim 0.3\mathrm{nm}$. (b) SEM image of a BLG sample with multiple nonlocal spin valves.

Figure 2

RT data: (a) $\sigma$ vs $V_G$ and ${\tau }_s$ vs $V_G$ for BLG. (b) Nonlocal resistance as a function of the in-plane magnetic field $B_y(T)$. The horizontal arrows show the field sweep direction while the vertical arrows show the relative magnetization orientations of the injector and detector electrodes. Hanle precession measurement for a perpendicular magnetic field $B_z(T)$ sweep for (c) the same sample with $\mu \sim 2000{\mathrm{cm}}^2/\mathrm{V}\mathrm{s}$ and (d) for a sample with $\mu \sim 300{\mathrm{cm}}^2/\mathrm{V}\mathrm{s}$.

Figure 3

Results of Hanle precession measurements at RT for 17 BLG spin devices with mobility varying from $200–8000{\mathrm{cm}}^2/\mathrm{V}\mathrm{s}$. The data points with the same symbol represent different junctions on the same sample. Open and closed symbols correspond to samples with global and local MgO, respectively. (a) ${\tau }_s$ taken at $n=1.5\times {10}^{12}{\mathrm{cm}}^{-2}$ vs $\mu$ plotted on a log-log scale. (b) ${\tau }_s$ taken at the CNP vs ${\sigma }_{\min }$ for BLG samples at RT.

Figure 4

(a) ${\tau }_s$ vs $\mu$ for 8 BLG junctions at 5 K. (b) Upper panel: $R$ vs $n$ for BLG; Lower panel: ${\tau }_s$ vs $n$ for $T=300\mathrm{K}$, 50 K, and 5 K. (c) Upper panel: $R$ vs $n$ for SLG; lower panel: ${\tau }_s$ vs $n$ for SLG at RT and at 5 K. (d) Upper panel: ${\tau }_s$ vs $T$ for four densities, $n=\mathrm{CNP}$, 0.7, 1.5, and 2.2 $\times {10}^{12}{\mathrm{cm}}^{-2}$. Lower panel: ${\tau }_s$ vs $T$ for SLG at $n=\mathrm{CNP}$, 1.5 $\times {10}^{12}{\mathrm{cm}}^{-2}$.