Relativistic nature of carriers: Origin of electron-hole conduction asymmetry in monolayer graphene

We report electron-hole conduction asymmetry in monolayer graphene. Previously, it has been claimed that electron-hole conduction asymmetry is due to imbalanced carrier injection from metallic electrodes. Here, we show that metallic contacts have negligible impact on asymmetric conduction and may be either sample or device-dependent phenomena. Electrical measurements show that monolayer graphene based devices exhibit suppressed electron conduction compared to hole conduction due to the presence of donor impurities which scatter electrons more efficiently. This can be explained by the relativistic nature of charge carriers in a graphene monolayer and can be reconciled with the fact that in a relativistic quantum system transport cross section does depend on the sign of scattering potential in contrast to a nonrelativistic quantum system.

Figure 1

Transfer characteristics of FET based on (a) type-I and (b) type-II MLG, using different metals (Al, Cu, Au) as source-drain electrodes. In all cases, source to drain voltage (VDS) was fixed at 0.1 V. (c) Evolution of asymmetric parameter is plotted against metal (Al, Cu, and Au, respectively) work functions. Asymmetric parameter is independent of metallic work functions indicating negligible doping of MLG with metal electrodes.

Figure 2

Raman spectra of (from bottom to top) type-I and type-II MLG layers. In type-I MLG layers, Raman D peak is absent whereas, in type-II MLG layers, D peak is present. Monotonic red shift in 2D Raman peak positions which is an indication of doping can be seen in the case of type-II MLG layers. Individual spectrum is shifted on the $Y$ axis for better clarity.

Figure 3

Temperature dependence of transfer characteristics of FET based on (a) type-I and (b) type-II MLG layers. Temperature dependence of carrier mobility in (c) type-I and (d) type-II MLG layers, extracted from data given in (a) and (b). Hole and electron mobility ratio as a measure of conduction asymmetry at 300 K/10 K has been found to be 1.02/1.04 and 3.0/3.2 in the case of type-I and type-II MLG layers, respectively.

Figure 4

${\mu }_h/{\mu }_e$ as a function of impurity strength $\left(\alpha \right)$ as expressed by Eq. (6) given in text (solid line). Symbols are the experimentally observed for MLG based devices. Inset shows the dependence of normalized Raman $D$ peak intensity on the impurity strength $\left(\alpha \right)$. $Y$ axis error bars denote uncertainty in ${\mu }_h/{\mu }_e$ and $I_D/I_G$ values measured on a minimum five samples.