In a narrow corridor, people walk in rows rather than as a crowd. The number of rows depends on the width of the passage. In graphene, this corridor is a quantum point contact — a constriction where electrons move ballistically without scattering. Two gates with a 75 nm gap create this corridor, and a third gate changes its width, opening or closing quantum modes — discrete 'rows' for electrons. Each new mode adds a step to the conductance.

An aluminum superconducting contact is attached to this corridor. At a temperature of 10 mK, the superconductivity of aluminum influences the electrons in graphene: at the interface, Andreev reflection occurs — an electron approaching the superconductor is converted into a hole that travels back. This is like a person hitting a wall, turning around and walking back, but transferring their charge in the process. As a result, the current through the corridor is amplified: the conductance at each step increases by about a quarter.

When the bias voltage is increased, superconductivity in the contact is destroyed. This happens not gradually but abruptly: at a critical power of about 0.5 nW, comparable to the power dissipated by a grain of sand falling from a height of one centimeter per second, the contact switches to the normal state. The switching current also changes in steps, following the number of open modes. Each new mode is like an additional switch: it turns on current amplification but simultaneously brings the threshold for destroying superconductivity closer.

Previously, it was unknown how quantized modes in a ballistic channel in graphene affect superconducting transport, especially excess current and the destruction of superconductivity at large biases. Now it is clear: Andreev amplification, excess current, and switching current all change in steps as modes are sequentially filled, following the same mode structure. This was demonstrated by comparing the behavior of conductance and switching current at different gate voltages.

This platform could serve as a testbed for studying nonequilibrium superconductivity and dissipative transitions in ballistic nanostructures. It may also be useful for developing elements of quantum circuits with tunable coupling. But already now it is clear: in the narrow corridor of graphene, each row of electrons is not just a step in current but a switch that determines whether superconductivity remains on.