Qubit Readout
Readout converts a qubit's quantum state into a classical bit, and its speed and fidelity strongly shape overall device performance.
Measuring a qubit
Reading a qubit projects it onto |0> or |1> and yields a classical outcome. Good readout is fast, high-fidelity, and ideally quantum non-demolition, meaning it leaves the qubit in the measured state so it can be reused. Readout errors feed directly into every algorithm and are a major part of the total error budget, especially for error correction, which measures constantly.
Methods by modality
- Superconducting: dispersive readout, a probe tone reflects off a resonator whose frequency shifts with the qubit state
- Trapped ions and neutral atoms: state-dependent fluorescence, a bright or dark scattering signal
- Spin qubits: spin-to-charge conversion sensed by a nearby charge detector
- NV centers: spin-dependent optical fluorescence
Signal-to-noise
Each method must distinguish two states faster than the qubit decays. Superconducting readout uses near-quantum-limited parametric amplifiers to lift the tiny signal above noise; fluorescence readout collects enough photons to separate bright from dark with confidence. Longer integration improves fidelity but risks a T1 decay event corrupting the result.
Error modes
Assignment errors (calling |0> a |1> or vice versa), state decay during measurement, and measurement-induced transitions all degrade readout. Typical readout fidelities range from about 95 to over 99.9 percent depending on modality and effort. Because error-correcting codes measure ancilla qubits every cycle, fast and accurate readout is as important as gate quality for scaling.