The Mott metal-insulator transition is one of the sharpest boundaries in condensed matter physics. On one side, electrons flow freely. On the other, they are locked in place by their mutual repulsion — correlation, not disorder, traps them. At the boundary between these phases, something anomalous happens: the electrical resistance hits a maximum as a function of temperature, then decreases on both sides. This peak has been observed in every bandwidth-tuned Mott system studied, and for decades nobody could agree on what it means.
Two camps formed. One said the resistance maximum is a coherence-incoherence crossover in a uniform metal — the electrons scatter so strongly that quasiparticles lose definition, and transport degrades smoothly. The other said the maximum marks the onset of spatial fragmentation — the material breaking into coexisting metallic and insulating regions, with transport controlled by the geometry of the metallic paths through the insulating matrix.
Wang et al. resolve the debate by changing the measurement. Instead of sweeping temperature and recording resistance — which averages over everything happening inside the material — they park at a fixed temperature near the resistivity maximum and watch the resistance evolve in time. What they see is random telegraph noise: the resistance switches abruptly between two discrete levels, spending random durations in each. A mesoscopic region inside the material toggles between metallic and insulating states, opening and closing the dominant conduction channel like a switch.
This is intermittency, not crossover. The material is not smoothly degrading. It is flickering. At any instant, the resistance has one of two values, corresponding to two configurations of the metallic and insulating domains. The time-averaged resistance — what a conventional sweep would measure — looks like a smooth function of temperature. But the instantaneous resistance reveals the dynamics that the average conceals.
The switching is thermally activated, with well-defined energy scales. This means the metallic and insulating configurations correspond to distinct metastable states separated by a barrier, not to continuous deformation of a single state. The resistivity maximum occurs where the material spends roughly equal time in each state — not because the electrons are maximally incoherent, but because the domain switching is maximally undecided.
The practical implication is that transport near the Mott transition is controlled by mesoscopic domain dynamics, not by single-electron physics. This is a different kind of problem — more like percolation through a fluctuating landscape than scattering in a homogeneous medium.