Dispersive Readout
Dispersive readout infers a superconducting qubit's state from a state-dependent shift in the frequency of a coupled microwave resonator.
The dispersive regime
Couple a qubit to a microwave resonator and detune them far apart in frequency. In this dispersive regime they do not exchange energy, but the resonator's frequency shifts by a small amount that depends on the qubit's state, up for |0> and down for |1|. Measuring the resonator therefore reveals the qubit without directly absorbing its energy, making the measurement quantum non-demolition to good approximation.
How it is done
Send a weak probe tone at the resonator. Its transmitted or reflected amplitude and phase depend on which way the resonator shifted, so the qubit state maps onto a point in the I-Q plane. Averaging or single-shot discrimination separates the two states into distinct clouds. The dispersive shift chi and the resonator linewidth set how quickly the two states can be told apart.
Amplification chain
- A Josephson parametric or traveling-wave amplifier near the quantum limit at the coldest stage
- HEMT amplifiers at intermediate temperatures
- Room-temperature digitization and demodulation into I and Q
Limits and Purcell protection
The resonator also opens a decay channel for the qubit (the Purcell effect), so a Purcell filter is added to let the readout tone through while blocking qubit decay. Too strong a probe drives the qubit out of the computational space; too weak a probe is slow. Modern dispersive readout reaches single-shot fidelities well above 99 percent in a few hundred nanoseconds, fast enough for feedback and error correction.
Dispersive readout is the standard measurement for superconducting processors and a direct application of circuit quantum electrodynamics.