Quantum non-demolition measurement of qubits

Short answer: A quantum non-demolition (QND) measurement reads an observable without introducing a back action on that observable, so repeating it gives the same answer. A Nature Nanotechnology paper reports an all-electrical QND measurement of a single electron spin in a gate-defined quantum dot, and a Physical Review Applied paper proposes a superconducting circuit for fast, high-fidelity QND readout. The second is a proposal and simulation, not a hardware result.

What QND means

An arXiv paper on QND measurements in qubit systems says measuring a quantum system to get information must disturb it, but that QND measurements allow an observable to be measured without introducing a back action on that observable from the measurement itself. The authors say such measurements explore the fundamental limits of measurement, may prove useful in gravity wave detection, telecommunications and quantum control, and are critical to key quantum information protocols such as error correction. They introduce figures of merit that apply to systems of any Hilbert space dimension, and discuss the controlled-NOT gate and an optical implementation as examples of QND devices for qubits.

An electron spin in a quantum dot

The Nature Nanotechnology paper demonstrates an all-electrical QND measurement of a single electron spin in a gate-defined quantum dot. The authors entangle the single spin with a two-electron, singlet-triplet ancilla qubit through the exchange interaction and then read out the ancilla in a single shot. They report a disturbance-free projective measurement of the single spin at a rate two orders of magnitude faster than its relaxation. Because the protocol is QND, repeating it improves the overall measurement fidelity, and they report a monotonic increase over 100 repetitions against arbitrary input states. They also observe spontaneous flips, or quantum jumps, of a single electron spin as a further sign of the QND character.

Superconducting qubit readout

The Physical Review Applied paper starts from dispersive readout, which it calls the current standard for superconducting qubits and which relies on a static transverse coupling between a qubit and a detuned cavity mode. It says this requires design trade-offs, and that the criteria for QND measurements, including reduced cavity photon numbers, further limit speed and fidelity. The authors introduce a circuit with pulse-tunable qubit-cavity and cavity-feedline couplings, so that the whole measurement can be optimised, and develop a pulsed-measurement protocol that goes beyond the rotating-wave, Jaynes-Cummings and dispersive approximations. They propose a realistic circuit architecture that they say could provide very fast, high-fidelity QND measurement with near-unit efficiency.

Limits

The spin experiment is a single-device laboratory demonstration. The superconducting paper presents a design and theory, so its efficiency figure is a prediction. The general paper is a characterisation framework, not an experiment.

FAQ

What does non-demolition mean?

The measurement does not introduce a back action on the measured observable.

Why is QND useful?

Repeating the measurement can improve fidelity. The Nature Nanotechnology paper reports a fidelity rising over 100 repetitions.

Was this shown in hardware?

For an electron spin in a quantum dot, yes, per that paper. The superconducting readout design is proposed in the Physical Review Applied paper.

How does it relate to dispersive readout?

That paper calls dispersive readout the current standard for superconducting qubits and proposes a tunable-coupling alternative.

Is it needed for error correction?

The arXiv paper says QND measurements are critical to many key quantum information protocols, such as error correction.