Colossal negative magnetoresistance in dilute fluorinated graphene

Adatoms offer an effective route to modify and engineer the properties of graphene. In this work, we create dilute fluorinated graphene using a clean, controlled, and reversible approach. At low carrier densities, the system is strongly localized and exhibits an unexpected, colossal negative magnetoresistance. The zero-field resistance is reduced by a factor of 40 at the highest field of 9 T and shows no sign of saturation. Unusual staircaselike field dependence is observed below 5 K. The magnetoresistance is highly anisotropic. These observations cannot be explained by existing theories, but likely require adatom-induced magnetism and/or a metal-insulator transition driven by quantum interference.

Figure 1

(a) Comparison of Raman spectra on fluorinated (upper trace) and defluorinated (lower trace) graphene using a 514-nm excitation wavelength. Fluorination conditions are identi-cal for both samples. The Raman bands are marked in the figure. (b) A STM current image (12 nm × 12 nm) of fluorinated graphene. $V_{\mathrm{bias}}$ = 20 mV. Color scale: 2 nA. The slowly varying background is owing to the height undulation of the SiO${}_2$ surface. The orientation of the unperturbed graphene lattice is marked by white solid lines. The white dotted lines mark the signature of isolated fluorine adatoms (Ref. 15). The fluorine distribution is inhomogeneous. F adatoms within 2 nm of each other are sometimes observed. Clusters are rare. Inset: three F adatoms close to each other with overlapping $\sqrt{3}$ superstructures. Color scale: 5 nA.

Figure 2

(a) Semi-log plot of $R_s$ vs backgate $V_g$ on sample A at $T$ = 5, 10, 20, 50, 100, and 200 K (top to bottom). The backgate dopes $7\times {10}^{10}/{\mathrm{cm}}^2{V}_g$(V) and the charge neutrality point occurs at $V_g$ = 16 V. Inset: Optical image of a DFG device with the graphene piece outlined. (b) Semilog plot of $R_s$ vs $T^{-1/3}$ on sample B at the charge neutrality point, $N$ = 0.7, 1.4, and $2.5\times {10}^{12}/{\mathrm{cm}}^2$ (top to bottom). Dashed lines are fittings to the 2D VRH conduction, the slope of which yields $T_0$. The gray band corresponds to $~h/2e^2$. Inset: The $N$-dependent $T_0$ for samples A (black squares), B (red circles), and C (green triangles) in the strong localization regime.

Figure 3

(a) Normalized MR in a perpendicular magnetic field, $R_s(B)/R_s(0)$, on sample A at $T$ = 5 K and $N$ = 0.8 (black), 2.1 (red/gray), and $4.2\times {10}^{12}/{\mathrm{cm}}^2$ (blue/dark gray), labeled by their temperature dependence. (b) Normalized MR at $n=0.8\times {10}^{12}/{\mathrm{cm}}^2$ and selected temperatures. Inset: Staircaselike MR observed on another sample (B) at $N$ = 2.1 (navy/dark gray, upper) and $1.4\times {10}^{12}/{\mathrm{cm}}^2$ (orange/gray, lower). $T$ = 2 K. (c) Normalized MR on sample B at the charge neutrality point in perpendicular ($B_{\perp }$) and in-plane ($B_{\parallel }$) magnetic fields showing a large anisotropy. (d) Comparison of magnetoresistance of DFG (black, scale on the left-hand side) and defluorinated DFG (red/gray, scale on the right-hand side). For both samples, $n=1.05\times {10}^{12}/{\mathrm{cm}}^2$ and $T$ = 5 K. The defluorinated sample exhibits magnetoresistance oscillations of graphene with the filling factors marked in the figure.

Figure 4

(a) $R_s\hspace{0.5em}\mathrm{vs}\hspace{0.5em}{T}^{-1/3}$ in a semilog plot for data shown in Fig. 3b at selected fields. Solid lines are fittings to the 2D VRH conduction, the slope of which yield $T_0$. (b) $T_0$ (scale on the left-hand side) and corresponding ξ (scale on the right-hand side) vs $B$ from the data shown in (a). The blue arrow marks $B^{*}=0.21$ T.