Many modern quantum computers are built on superconducting circuits, where qubits - the quantum analogues of a bit - easily lose energy, and with it fragile quantum information. Researchers from the Karlsruhe Institute of Technology (Germany), led by Ioan Pop, have found a way to identify microscopic defects that "remember" lost energy and later return it to the qubit. The work was published in the journal Physical Review Letters.

It is usually assumed that energy that has left a qubit disappears irrevocably. However, some defects in surrounding materials can hold it longer than the qubit itself lives, and then give it back, returning the system to its previous state. Such memory effects are not captured by standard measurements of qubit lifetime, which record only the average rate of energy loss. Until now, physicists did not know exactly what these defects are, and they could be studied only by deliberately acting on the qubit, which slowed down computations.

Instead of active probing, Pop's team used continuous measurement and recorded spontaneous "quantum jumps" between the ground (low-energy) and excited states of the qubit caused by thermal fluctuations. The scientists then analyzed how the qubit's subsequent behavior depends on previous jumps and built a model of energy exchange with the environment. The model was tested at different qubit frequencies and electric field strengths, which made it possible to isolate individual defects.

Normally, after losing energy, a qubit needs time to make the next jump, so jumps should be distributed evenly. However, the researchers found that they often cluster. This indicates that energy given up by the qubit is stored in nearby defects and returned, passing several times between the qubit and the defect before finally dissipating. Importantly, the memory effect was observed without deliberately exciting the qubit.

Each defect appeared as a separate peak when the qubit frequency was changed and responded to the electric field, which gives a clue to its microscopic nature. The method makes it possible to investigate what these long-lived defects are and where they form, and in the future to develop error-correction schemes that take into account the failures they cause in real quantum hardware. For now this is only a first step: the scientists emphasize that the link between defects and errors has yet to be clarified.

Source: Phys.org