Dirac cones reshaped by interaction effects in suspended graphene

Abstract

Graphene’s linear dispersion relation makes its charge carriers behave as if they were massless. However, near the Dirac point where graphene’s valence and conduction bands meet, electron–electron interactions cause this relation to diverge, such that it becomes strongly nonlinear and the effective carrier velocity doubles. In graphene, electron–electron interactions are expected to play a significant role, as the screening length diverges at the charge neutrality point and the conventional Landau theory that enables us to map a strongly interacting electronic liquid into a gas of non-interacting fermions is no longer applicable1,2. This should result in considerable changes in graphene’s linear spectrum, and even more dramatic scenarios, including the opening of an energy gap, have also been proposed3,4,5. Experimental evidence for such spectral changes is scarce, such that the strongest is probably a 20% difference between the Fermi velocities vF found in graphene and carbon nanotubes6. Here we report measurements of the cyclotron mass in suspended graphene for carrier concentrations n varying over three orders of magnitude. In contrast to the single-particle picture, the real spectrum of graphene is profoundly nonlinear near the neutrality point, and vF describing its slope increases by a factor of more than two and can reach ≈3×106 m s−1 at n<1010 cm−2. No gap is found at energies even as close to the Dirac point as ∼0.1 meV. The observed spectral changes are well described by the renormalization group approach, which yields corrections logarithmic in n.