Can circular Rydberg atoms stay stable at room temperature?

Angular momentum blocks the fastest decay channel

A Rydberg state is a highly excited electronic orbital of an atom. In the low-angular-momentum states used in most neutral-atom quantum computing, simulation and sensing work, the excited electron can decay quickly through optical and blackbody radiation, giving a lifetime on the order of 100 µs. That lifetime sets a ceiling on gate fidelities, coherence times and spectroscopic precision.

Circular Rydberg states take the opposite limit: the orbital and magnetic quantum numbers reach their maximum, |m| = l = n − 1, where n is the principal quantum number. Selection rules strongly inhibit spontaneous optical decay for these states. The remaining threat is blackbody radiation, which in earlier single-atom single-photon experiments was handled by working in a cryogenic environment. This paper asks whether that threat can instead be removed at room temperature.

A plate capacitor suppresses the damaging blackbody modes

The atom sits at the centre of an in-vacuum capacitor formed by two transparent glass plates coated on the inside with indium-tin oxide, which is highly reflective in the microwave domain. When the photon wavelength exceeds twice the plate distance, λ > 2d, the mode density at the relevant blackbody transitions is suppressed for modes polarized parallel to the plates — including the circularly polarized photons that drive the strongest decay between neighbouring circular states.

With a plate distance of d = 10.5(2) mm, suppression is expected for n ≳ 78. The authors report a more than 20-fold suppression of blackbody-induced decay, which they say would correspond to a free-space lifetime requiring a 14 K environment. The structure is not ideal: additional electrodes and conductive parts create extra resonant modes, and the paper maps these deviations rather than assuming the simple plate model.

Microwave ladders reach principal quantum number 103

The experiment starts with a single trapped 88Sr atom, promoted to the |79C⟩ circular state by optical excitation and adiabatic transfer with a circularly polarized radio-frequency drive. From there, a sequence of two-photon microwave transitions raises n in steps of two, using adiabatic Landau-Zener sweeps through each resonance. Up to 12 subsequent microwave pulses link the circular states from n = 79 to n = 103, an orbit diameter of about 1.1 µm.

Reaching these states required separating the target circular-to-circular transition from a nearby transition to an elliptical state. The two have different quadratic Stark shifts, so the sweeps are performed at electric fields of a few V cm⁻¹ chosen to avoid much stronger single-photon resonances. The average fidelity of the individual Landau-Zener steps is 94%, and the coherent transfer up to n = 101 takes about 400 µs — already longer than the free-space lifetime of the lowest circular state used.

Lifetimes pass 10 ms above n = 95

Lifetime is measured by holding the atom for a variable time and then reading out which n-manifold it occupies using state-selective field ionization. A simple exponential fit to the initial state is not enough, because population can return to the initial manifold through secondary transitions; the authors instead fit a rate-equation model across the n-manifolds.

The measured lifetime rises sharply around n ≈ 80 as the strongest circularly polarized blackbody transitions become suppressed, reaching about 5 ms near n = 85. A dip appears around n ≈ 90, which the authors attribute to π-polarized modes that the plate boundary does not suppress and that the ring electrodes enhance at certain frequencies. Above n = 95, lifetimes exceed 10 ms; for |101C⟩ the measured value is 11.5(8) ms, about 21 times the free-space lifetime of 545 µs. The authors note this exceeds previous room-temperature results at n = 60 by an order of magnitude and outperforms lifetimes obtained in a cryogenic environment by a factor of three.

Tweezer trapping lasts longer than the state itself

Long lifetimes are only useful if the atom stays in the trap. Standard Gaussian optical tweezers usually expel Rydberg atoms through the repulsive ponderomotive force on the Rydberg electron. Here the alkaline-earth strontium ion core provides an attractive potential strong enough to confine the circular state, and trapping remains effective as the electron decays into other Rydberg levels.

For atoms initialized in |97C⟩, an exponential fit to the total detected ion signal gives a 1/e trapping lifetime of 133(6) ms, about an order of magnitude longer than the initial state lifetime. With the tweezer switched off, the signal fades within a millisecond. The authors describe this trapping time as a lower bound, because the chosen ionization ramp does not detect Rydberg states above n = 104 or decay into significantly lower angular momentum states.

Coherence, not lifetime, is now the limiting timescale

The paper is explicit that the lifetime is not yet the usable coherence time. In this setup, technical sources — specifically magnetic field noise and electric field inhomogeneity — currently limit the T2 coherence time of circular-state qubits to the 100 µs scale. The authors propose active feedback, careful cancellation of gradients, laser cooling on the ionic core and dynamical decoupling as routes toward coherence times on the scale of the circular-state lifetime.

The comparison that matters is with earlier circular-state interaction control at n ≈ 50. Moving to the high n demonstrated here brings about a thousand-fold stronger van der Waals blockade and more than ten times larger dipole-exchange coupling. Those interaction strengths are what would make the longer lifetime useful for simulation, but only under the stated condition that the technical dephasing limits are addressed.

The room-temperature claim depends on the suppression structure

The lifetime enhancement is a property of the atom inside this particular electrode structure, not of circular Rydberg atoms in free space. The measured curve follows the trend of an idealized infinite capacitor with finite reflectivity, including the enhancement steps near n ≈ 80 and n ≳ 95, but the dip near n ≈ 90 only appears when the full electrode geometry is simulated with finite-element methods. The authors use an indium-tin oxide reflectivity of 96(+1/−2)% and a plate distance of 10.5(2) mm, and the shaded model bands reflect those uncertainties.

This makes the design of the blackbody-suppressing structure part of the result. Additional cavity modes can enhance the very transitions the capacitor is meant to suppress, so a different electrode layout would produce a different lifetime curve. The demonstrated range is also bounded by the state-preparation ladder: coherent control is shown up to n = 103, with the final 101→103 transition probed only spectroscopically.

The practical priority is pushing T2 toward the lifetime

For readers weighing circular Rydberg atoms against conventional low-angular-momentum Rydberg arrays, the decisive number is not the 10 ms lifetime alone but the gap between it and the 100 µs coherence time currently set by technical noise. The paper's own priority ordering follows from that gap: the lifetime barrier has been pushed two orders of magnitude out, and the next step is to bring T2 into the millisecond range through field-noise control and cooling of the ionic core.

The high-n operating point also carries a trade-off visible in the data. Going above n = 95 buys lifetimes beyond 10 ms and much stronger interactions, but it requires a longer microwave ladder — about 400 µs to reach n = 101 — and it sits where residual π-polarized cavity modes caused a lifetime dip near n = 90. Choosing an operating n therefore means balancing lifetime, interaction strength and the specific mode structure of the suppression capacitor, rather than simply maximizing n.

Sources used in this answer