A courtroom interpreter must relay every word without being distracted by noise in the room. In the quantum world, a superconducting resonator plays that role, losing photons on schedule without straying into spurious effects. Previously, this was done with microwave pumping, but the pump itself created interference, as if the interpreter started shouting out audience remarks.

Now scientists have found a quieter way. Instead of a loud pump, they applied a DC voltage across a superconducting loop. When a pair of electrons (a Cooper pair) tunnels across the junction, it releases energy that can be precisely tuned so that two photons from the memory resonator convert into one photon in the buffer resonator. The buffer resonator quickly loses that photon, and the system settles into the desired state.

The experiment showed that the resonator loses photons in pairs at a rate of 0.51 MHz, and the collapse time of a coherent state was 50 nanoseconds. This means that within two microseconds a photon pair leaves the resonator, which is comparable to the lifetime of a single photon in the same resonator (3.1 μs).

This approach opens the way to stabilizing cat qubits, which are needed for quantum error correction. Whereas the pump used to introduce errors of its own, DC voltage now acts like an ideal interpreter: precise, quiet, and without extra words.