Quantum metrology

Quantum metrology improves parameter estimation by controlling both the preparation and measurement of quantum states. A probe may accumulate a phase, frequency shift, displacement, or field-dependent energy change, which is then inferred from repeated observations. The quantum Fisher information provides a general measure of how sensitively a state responds to that parameter, while the quantum Cramér–Rao bound relates it to the smallest possible statistical uncertainty.¹

With N independent probes, uncertainty commonly follows the standard quantum limit, scaling as 1/√N. Entangled or otherwise nonclassical probes can in ideal conditions approach 1/N, called Heisenberg scaling.² These are asymptotic benchmarks rather than guaranteed laboratory performance: loss, noise, imperfect control, and finite measurement time can remove the advantage.

The main quantum resources are coherence, squeezing, entanglement, and long-lived internal states. Squeezed light reduces fluctuations in one measurement quadrature and is used in precision interferometry; entangled states can correlate probes so that a collective phase becomes more informative than independent measurements.³

Atomic clocks realize quantum metrology through sharply defined transitions in atoms or ions, enabling comparisons of time and frequency at extraordinary precision. Atom interferometers measure acceleration, rotation, and gravity by splitting and recombining matter waves. In quantum sensing, nitrogen-vacancy centers in diamond can detect magnetic fields locally, including in biological and materials-science settings.⁴ The measurement apparatus, calibration procedure, and estimator are part of the metrological design, not merely technical accessories.

Quantum metrology is most valuable when the parameter is weak, the measurement must be localized, or long-term stability matters. Optical interferometers use squeezed states to reduce quantum noise, while frequency standards support navigation, telecommunications, geodesy, and tests of fundamental physics.² Quantum sensors also probe magnetic and electric fields, temperature, pressure, acceleration, and gravitational potential.

Practical sensitivity is governed by more than a formal quantum bound. Decoherence can destroy the phase relationships on which an advantage depends, and entanglement may become fragile under particle loss or detector inefficiency.⁵ In many experiments, carefully optimized classical or partially nonclassical protocols outperform an ideal quantum strategy once preparation overhead, dead time, technical noise, and environmental disturbance are included.

The field includes strategies that do not simply seek the greatest instantaneous sensitivity. Adaptive measurements can change later probe settings in response to earlier results, helping resolve phase ambiguities and improve performance when the unknown parameter is poorly localized.¹ Distributed and networked sensors can also use shared quantum correlations to estimate quantities involving separated locations.

Quantum advantages are task-dependent: a state that is optimal for estimating a phase may be unsuitable for estimating several parameters simultaneously. Multiparameter estimation introduces trade-offs because the measurements that extract information about different quantities may be incompatible.⁶ Another subtlety is resource accounting. Preparation energy, particle number, bandwidth, interrogation time, and data-processing assumptions must be specified before claims of “quantum enhancement” can be compared fairly.

Quantum Fisher information

A quantity that measures how strongly a quantum state changes as an unknown parameter changes; it determines a fundamental precision bound.

Standard quantum limit

The commonly obtained 1/√N uncertainty scaling for N independent probes.

Heisenberg limit

An idealized 1/N scaling of uncertainty with probe number or another specified resource.

Squeezed state

A quantum state with reduced fluctuations in one observable, accompanied by increased fluctuations in a conjugate observable.