Jérôme Lesueur

Recent Nernst experiments in High Tc superconductors suggested that fluctuating pairs may survive well above Tc with no phase coherence. It appears that no measurement of the Nernst signal has been measured in the fluctuating regime of regular BCS superconductors. Sondhi et al recently predicted that in 2D, the Nernst signal only depends on the superconducting coherence length at a given temperature above Tc. Moreover, the superconducting contribution to the signal is orders of magnitude bigger than the normal one in diffusive metals, when the pair lifetime is longer than the normal electron scattering one.
Alexandre Pourret, Kamran Behnia and Hervé Aubin have measured the Nernst signal generated by fluctuating pairs in amorphous NbSi thin films. Indeed, they found it to be more than three orders of magnitude bigger than the normal one, such as it can be tracked far into the normal region, namely at temperature 30 times Tc. Its temperature dependence quantitatively agrees with the theoretical prediction. Moreover, when the magnetic length lB associated with the applied magnetic field becomes smaller than the superconducting coherence length, the Nernst signal is controlled by lB.