In most conductors, crystal defects and phonons scatter conduction electrons and impede the flow of current. But graphene exhibits anomalously weak electron–phonon scattering and can also be prepared in such a pure state that, at temperatures of a few degrees kelvin, an electron driven by a voltage difference moves ballistically. That is, the electron travels unimpeded through the bulk of the material until it hits a boundary, deposits some momentum there, and bounces back into the bulk to move freely until it again ricochets off a boundary. At higher temperatures, conduction electrons scatter off each other rather than off phonons. Due to those interactions, the electrons collectively flow as a viscous liquid, a phenomenon called electron hydrodynamics.
That collective flow has been observed in graphene samples with constant width, but Lancaster University PhD candidate Roshan Krishna Kumar, the University of Manchester’s Andre Geim, and their international team have now added a new wrinkle by measuring the resistance of samples fabricated to have micrometer-scale constrictions such as those shown in the figure. At low temperature, the resistance was consistent with ballistic electron motion. But as the temperature was increased and collective effects became important, the investigators observed that the resistance fell. Though that so-called superballistic flow may be counterintuitive, the measured resistance drops were in accord with theoretical expectation, as was the observation that the drops were most pronounced for the thinner constrictions.
How does a fluid of electrons manage to flow through an aperture with less resistance than noninteracting, ballistically moving electrons? The key is that the flow is not uniform. Current density in the electron fluid is greatest near the centerline of the constriction and tapers off toward the sides. Because of that structure, the electron fluid is guided through the narrow channels in the graphene and avoids the boundary collisions that rob ballistic electrons of their momentum. (R. Krishna Kumar et al., Nat. Phys., in press, doi:10.1038/nphys4240.)
