How Rydberg Atoms Bypass Losses: A Breakthrough in Quantum Microwave Metrology ✦
Quantum sensors based on Rydberg atoms enable high-precision electric field measurements, but detection losses limit the achievement of quantum advantage. This work exploits the dual functionality of Rydberg atoms — as probes of external fields and as a resource for quantum information processing. Through dipole-dipole interactions in the ensemble, an error-prevention protocol was implemented that introduces an additional nonlinear lossy channel, paradoxically increasing Fisher information by a factor of 3.3. A sensitivity of 39 nV/cm/√Hz was achieved. The results demonstrate the possibility of enhancing quantum metrology through internal interactions of the sensor system, without the need for a universal quantum computer.
Context
Quantum metrology is one of the most promising areas of quantum technologies, where sensors can achieve accuracies unattainable with classical instruments. For instance, Rydberg atoms with their enormous dipole moments are ideal for ultra-sensitive spectroscopy of microwave fields. However, any real quantum measurement faces unavoidable detection losses, especially when using optical readout. In conventional schemes, this leads to a proportional reduction in Fisher information, the key quantity defining the precision limit according to the Cramér–Rao bound. Overcoming detection losses without resorting to complex quantum computing algorithms remains an open challenge, and the solution proposed by the Warsaw group relies on an elegant use of internal interactions within the ensemble.
Methods
The experimental setup involved a cloud of one hundred million ⁸⁷Rb atoms cooled to 78 µK in a magneto-optical trap. Using quantum optics – a probe and a coupling laser – collective Rydberg excitations, or spin waves, were created in the state |49D₅/₂⟩. Then a microwave field at 18.8 GHz drove the atoms into a superposition between this and another Rydberg state |50P₃/₂⟩, encoding the measured parameter into the Rabi angle. Next, distance-dependent dipole–dipole interactions ∼1/R³ came into play: when excitations of both types are present in the ensemble, they experience additional nonlinear losses due to mutual decoherence. This process is described by the Schrödinger equation formalism for a collective wavefunction, followed by ensemble averaging, yielding an effective Kraus operator that eliminates components with two excitations in different modes. Readout was performed by converting the spin waves back into photons and registering them with a single-photon detector of only 2% efficiency; then the rotation angle was estimated from the count distribution using maximum likelihood.
Results
The central result is a threefold increase in Fisher information per detected photon compared to the standard limit for independent probes. In the experiment, a value of F = 3.3 ± 0.3 per photon was achieved with a normalized lossless value of F₀ = 1. This corresponds to an electric field sensitivity for microwave radiation of ΔE_MW = 44 µV/cm per measurement cycle, and taking cycle duration into account, the noise spectral density was S_E = 39 nV cm⁻¹ Hz⁻½ – a result comparable to the best continuous Rydberg sensors but obtained in a pulsed mode. The key observation was super-Rabi oscillations: due to quantum decoherence induced by interactions, the mode populations showed a steeper dependence on angle θ than standard harmonic oscillations, enabling more precise angle estimation. The theoretical model, accounting for many-body effects, agrees excellently with the data, and for the simplest case of two excitations, the protocol was proven optimal among all possible quantum operations.
Implications
This work marks a transition from theoretical constructs to practical error-corrected quantum metrology schemes. It shows that internal interactions in a sensor can serve as a resource for loss protection, opening a new direction for applying quantum information in realistic settings. Crucially, the method requires no extra qubits or complex logic circuits – one only needs to carefully control the measurement process. This could greatly simplify the creation of ultra-sensitive detectors for spectroscopy, telecommunications, and even astrophysical observations, where weak signals must be distinguished from strong instrumental noise.
Future development
Further performance improvements are possible by increasing the ensemble's optical depth and the number of excited Rydberg states, which will require more powerful lasers and better cooling – down to cryogenic temperatures, where coherence times could increase by orders of magnitude. In theory, for sufficiently strong interactions, the system could discriminate between zero and one excitation, enabling operation in a fully quantum regime with single spin waves. Integration with fiber optics and quantum computing networks could lead to a new generation of distributed quantum sensor systems.
Impact
The method will find applications in quantum metrology, enabling electromagnetic field sensors with unprecedented sensitivity, as well as in optical photometry of weak signals.
Next steps
Subsequent research will aim to scale the protocol to a larger number of excitations and to study its robustness against other noise types, such as atomic thermal motion and spontaneous emission.
Key open problems
This work is directly connected to the overall challenge of achieving quantum advantage in sensing under noise. It provides a practical recipe for circumventing one of the main limitations – losses – and could be generalized to other platforms, such as NV centers or ion traps, where error correction without excessive overhead is also critical.
🎯 The radius of a Rydberg atom in the n=49 state is about 0.13 µm – thousands of times larger than an atom in the ground state. If the nucleus were the size of a pea, the electron on such an orbit would be a football field away!
Key numbers
- sensitivity: 39 nV cm⁻¹ Hz⁻½
- Fisher information increase: 3.3 times
- detection efficiency: 2%
- Rabi frequency: 2π × 0.66 MHz
- coherence time with modulation: 31 µs