Circular Rydberg Atoms Overview

• Circular Rydberg atoms are highly excited states with maximum angular momentum (l = |m| = n-1) that exhibit nearly classical circular orbits and hydrogenic behavior.
• Preparation protocols use microwave and RF fields to circularize low-l states, achieving high-fidelity control and extended lifetimes through engineered electromagnetic environments.
• Their strong dipole-dipole interactions and compatibility with optical trapping enable advanced applications in quantum simulation, computing, precision spectroscopy, and sensing.

Circular Rydberg atoms are highly excited atomic states in which the valence electron occupies the maximum-angular-momentum state within a fixed principal-quantum-number manifold, $l=|m|=n-1$. In this limit the electronic probability density is concentrated near a nearly classical circular orbit of radius $\sim n^2 a_0$, the relevant dipole transitions fall in the GHz-to-mm-wave domain, and the states combine unusually long radiative lifetimes with very large dipole matrix elements and extreme sensitivity to dc and microwave fields. These properties place circular Rydberg atoms at the intersection of cavity QED, precision spectroscopy, quantum simulation, quantum computing, and sensing (Meinert et al., 2020, Hölzl et al., 2024, Ramos et al., 2017).

1. Definition, hydrogenic structure, and state manifold

Circular Rydberg atoms (CRAs) are the maximally rotating states of a Rydberg manifold. In the standard spherical basis they satisfy

$l=n-1,\qquad m=l=n-1,$

so that $|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$ (Meinert et al., 2020). Their wavefunctions are strongly concentrated in a ring-like region far from the ionic core, with $\langle r\rangle \sim n^2 a_0$, and the high-$l$ character suppresses core penetration and quantum-defect effects, making the levels nearly hydrogenic even in alkali and alkaline-earth atoms (Hölzl et al., 2024, Ramos et al., 2017).

The large principal quantum number implies large orbital size, weak binding, and dense level spacings. Adjacent-manifold transition frequencies scale as $1/n^3$ and lie in the microwave regime. Representative circular-to-circular transitions quoted in the literature are $|50,C\rangle \leftrightarrow |49,C\rangle$ at $f \approx 54.3$ GHz with $\lambda \approx 2.7$ mm, and $\sim n^2 a_0$0 at $\sim n^2 a_0$1 GHz with $\sim n^2 a_0$2 cm (Meinert et al., 2020). The corresponding dipole moments between neighboring manifolds are large, scaling as $\sim n^2 a_0$3 in the hydrogenic description (Hölzl et al., 2024).

External fields reorganize the manifold in a particularly useful way. In static electric fields, high-$\sim n^2 a_0$4 states form nearly equidistant Stark ladders in $\sim n^2 a_0$5, culminating in the circular state; this ladder structure underlies essentially all efficient circularization protocols (Hölzl et al., 2024). In parabolic coordinates, which are natural for Stark physics, circular and near-circular states can be represented as special points in the $\sim n^2 a_0$6 basis, and this representation is central both to precision-spectroscopy proposals and to microwave control schemes (Ramos et al., 2017).

A recurring misconception is that CRAs are simply “long-lived Rydberg states.” More precisely, they are high-$\sim n^2 a_0$7, nearly hydrogenic states whose long lifetime derives from angular-momentum selection rules and from the fact that their dominant decay channels are microwave transitions within neighboring Rydberg manifolds rather than optical decay to low-lying states (Meinert et al., 2020).

2. Decay channels, blackbody limitation, and electromagnetic-environment engineering

At $\sim n^2 a_0$8, the dominant spontaneous decay channel of a circular state is the single dipole-allowed microwave transition

$\sim n^2 a_0$9

and the decay rate can be written as $l=n-1,\qquad m=l=n-1,$0 (Meinert et al., 2020). More generally, including thermal photons,

$l=n-1,\qquad m=l=n-1,$1

with $l=n-1,\qquad m=l=n-1,$2 and the sum extending over dipole-allowed final states (Meinert et al., 2020). Because the relevant transition frequencies are in the GHz regime, room-temperature blackbody radiation strongly populates these modes and can dominate the lifetime budget.

This blackbody sensitivity is the main qualification to the usual statement that CRAs are intrinsically long-lived. In free space at room temperature, circular-state lifetimes can be reduced by up to three orders of magnitude relative to the $l=n-1,\qquad m=l=n-1,$3 spontaneous limit, reaching values comparable to low-$l=n-1,\qquad m=l=n-1,$4 Rydberg states unless the electromagnetic mode density is engineered (Meinert et al., 2020). Experimentally, a room-temperature rubidium $l=n-1,\qquad m=l=n-1,$5 state had a free-space lifetime of about $l=n-1,\qquad m=l=n-1,$6s and was extended to about $l=n-1,\qquad m=l=n-1,$7 ms by a simple plane-parallel capacitor with one transparent electrode (Wu et al., 2022). In strontium tweezers between ITO-coated plates, the measured room-temperature lifetimes were $l=n-1,\qquad m=l=n-1,$8 ms and $l=n-1,\qquad m=l=n-1,$9 ms (Hölzl et al., 2024). More recently, Purcell suppression of blackbody modes in an ITO capacitor with $|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$0 mm enabled room-temperature lifetimes above 10 ms, including $|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$1 ms, while maintaining coherent control up to $|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$2 (Pultinevicius et al., 31 Oct 2025).

The underlying principle is mode suppression in a plane-parallel capacitor or cavity. For a plate spacing $|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$3, propagation of modes polarized parallel to the plates is inhibited when

$|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$4

so if the circular orbital plane is parallel to the plates, the dominant $|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$5-polarized $|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$6 channels are strongly suppressed (Meinert et al., 2020). In ideal cavities this can drive the lifetime back toward the spontaneous limit; with realistic indium-tin-oxide thin films, measured reflectivities of about $|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$7 below 100 GHz were sufficient to predict room-temperature lifetimes of $|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$8 ms and $|n,C\rangle \equiv |n,l=n-1,m=n-1\rangle$9 ms in a simple capacitor geometry (Meinert et al., 2020).

The practical significance is twofold. First, room-temperature operation no longer forces CRAs into the same lifetime class as low-$\langle r\rangle \sim n^2 a_0$0 Rydberg states. Second, the lifetime engineering is compatible with optical access: transparent ITO plates can simultaneously reflect GHz–THz radiation and transmit visible or near-IR light for tweezers and imaging (Meinert et al., 2020).

3. Preparation, circularization, and coherent control

CRAs are not reached by direct optical excitation from the ground state. Standard protocols first excite a low-$\langle r\rangle \sim n^2 a_0$1 Rydberg state and then transfer population through the Stark ladder by microwave or radio-frequency fields. In alkaline-earth $\langle r\rangle \sim n^2 a_0$2Sr, one demonstrated creation of $\langle r\rangle \sim n^2 a_0$3 circular states in optical tweezers by three-photon excitation to $\langle r\rangle \sim n^2 a_0$4, followed by an electric-field ramp to $\langle r\rangle \sim n^2 a_0$5 mV/cm and a $\langle r\rangle \sim n^2 a_0$6-polarized RF-driven adiabatic rapid passage at $\langle r\rangle \sim n^2 a_0$7 MHz while sweeping the dc field to $\langle r\rangle \sim n^2 a_0$8 mV/cm over $\langle r\rangle \sim n^2 a_0$9s (Hölzl et al., 2024). In rubidium tweezer arrays, an alternative sequence used optical excitation to $l$0, a resonant 64.8 GHz microwave pulse to $l$1, and then RF-assisted adiabatic transfer at 225 MHz in a static electric field, reaching the $l$2 circular level with an overall preparation efficiency of about $l$3 and state purity about $l$4 (Ravon et al., 2023).

Once prepared, circular states are natural microwave qubits. A strontium qubit encoded in

$l$5

was coherently driven by an off-resonant two-photon microwave transition near $l$6 GHz, with a measured $l$7-pulse time $l$8 ns, $l$9s from Ramsey spectroscopy, and $1/n^3$0s from spin echo (Hölzl et al., 2024). These values are already in a regime where decoherence is dominated by technical noise and motional dephasing rather than by the fundamental lifetime alone.

Circularization itself can be a bottleneck in interacting systems. A theoretical analysis of two interacting $1/n^3$1Rb atoms showed that pulses optimized for isolated atoms are degraded primarily by interaction-induced shifts of the relevant ladder transitions; by analytically adapting the pulse phase with only two linear parameters and then combining this with Krotov optimization, two atoms were circularized to $1/n^3$2 after $1/n^3$3 ns with fidelity at least $1/n^3$4 for interatomic distances down to $1/n^3$5m and all angular configurations, while satisfying experimental amplitude and frequency constraints (Hüls et al., 6 Jul 2026).

An emerging alternative route avoids ladder climbing altogether. A fully quantum-mechanical scattering analysis predicted that twisted-electron collisions can enhance the production of circular states in hydrogen, rubidium, and cesium, particularly at large opening angles and low energies, by transferring orbital angular momentum from the projectile to the bound electron (Parker et al., 21 Nov 2025).

4. Trapping, tweezer architectures, and motional dynamics

For many years, CRA experiments were associated mainly with atomic beams and microwave cavities. The optical-trapping problem is nontrivial because the Rydberg electron samples intensity variations on micron scales, and for low-$1/n^3$6 states the near-IR trapping light can induce severe photoionization or repulsive ponderomotive forces. Circular states change this balance.

A first major route used ponderomotive bottle-beam or Laguerre–Gaussian traps. In a cryogenic rubidium experiment, a Laguerre–Gaussian $1/n^3$7 beam at 1064 nm produced 2D transverse confinement of $1/n^3$8 circular atoms for up to 10 ms, with a measured trap frequency of $1/n^3$9 kHz and no measurable degradation of lifetime or coherence by the trap itself (Cortiñas et al., 2019). This established that laser trapping of CRAs is compatible with millisecond-scale operation and thousands of interaction cycles. The approach was then extended to a programmable $|50,C\rangle \leftrightarrow |49,C\rangle$0 array of optical bottle beams for individual circular rubidium atoms, with a trap depth of about $|50,C\rangle \leftrightarrow |49,C\rangle$1K, an average trap frequency of $|50,C\rangle \leftrightarrow |49,C\rangle$2 kHz, and a coarse trapping time of about 5 ms; that work also introduced the first optical, spatially and level selective detection of alkali CRAs (Ravon et al., 2023).

A second route exploits the optically active ionic core of a divalent atom. In $|50,C\rangle \leftrightarrow |49,C\rangle$3Sr, the total trapping potential for a circular state can be decomposed as

$|50,C\rangle \leftrightarrow |49,C\rangle$4

where the attractive core polarizability of Sr$|50,C\rangle \leftrightarrow |49,C\rangle$5 competes with the generally repulsive ponderomotive potential of the Rydberg electron (Hölzl et al., 2024). For $|50,C\rangle \leftrightarrow |49,C\rangle$6 in a 539.91 nm tweezer, the core term dominates sufficiently to produce stable trapping in a standard Gaussian tweezer, while the differential trap depth between $|50,C\rangle \leftrightarrow |49,C\rangle$7 and $|50,C\rangle \leftrightarrow |49,C\rangle$8 yields a measurable qubit light shift and a controlled source of motional dephasing (Hölzl et al., 2024).

The trapping problem is inseparable from motion-induced dephasing and spin-motion coupling. A two-atom theory for circular-state XXZ simulators showed that the effective spin Hamiltonian inherits the position dependence of dipole-dipole couplings; depending on trap frequency, interatomic spacing, and choice of $|50,C\rangle \leftrightarrow |49,C\rangle$9, one obtains regimes of negligible spin-motion dressing, temperature-sensitive spin dynamics usable for motional thermometry, or fully entangled spin-motion states with non-classical motional structure (Méhaignerie et al., 2023). This picture was directly connected to experiment when resonant dipole-dipole interactions between two trapped CRAs were used as a nanometer-scale distance meter to record relative oscillatory motion in optical tweezers, revealing mechanically induced spin-motion coupling from transiently populated Stark states with permanent dipoles during circularization (Méhaignerie et al., 2024).

A broader trapping concept predating the tweezer era combined magnetic Ioffe–Pritchard confinement with a static electric field. In that geometry, induced permanent dipoles of order several hundred Debye and dipolar repulsion stabilize two circular atoms at a tunable equilibrium distance, with collapse thresholds set by the competition between magnetic transverse confinement and dipole-dipole anti-confinement (Hezel et al., 2011). This suggests that “trapping” and “interaction engineering” need not be separate design problems for CRAs.

5. Interactions, effective spin models, and many-body regimes

The appeal of CRAs for many-body physics lies in combining long lifetimes with large, controllable interactions. Two main interaction classes recur in the literature. First, atoms in the same circular manifold $f \approx 54.3$0 experience van der Waals couplings scaling as $f \approx 54.3$1; second, atoms in adjacent manifolds $f \approx 54.3$2 support resonant dipole-dipole exchange scaling as $f \approx 54.3$3 (Meinert et al., 2020). Representative calculated values for side-by-side atoms at typical tweezer separations are MHz-scale van der Waals shifts and resonant exchanges of about $f \approx 54.3$4 MHz at $f \approx 54.3$5m for $f \approx 54.3$6, increasing to about $f \approx 54.3$7 MHz at the same distance for $f \approx 54.3$8 (Meinert et al., 2020). With millisecond lifetimes, such couplings support $f \approx 54.3$9–$\lambda \approx 2.7$0 coherent interaction times within a single lifetime budget (Meinert et al., 2020).

The resonant two-atom exchange between neighboring circular manifolds has now been observed directly. For two individually trapped rubidium CRAs, the effective interaction between $\lambda \approx 2.7$1 and $\lambda \approx 2.7$2 is

$\lambda \approx 2.7$3

where $\lambda \approx 2.7$4 is the interatomic distance and $\lambda \approx 2.7$5 the angle between the quantization axis and the interatomic axis (Méhaignerie et al., 2024). Microwave spectroscopy revealed the corresponding collective-state splitting, including its sign reversal between $\lambda \approx 2.7$6 and $\lambda \approx 2.7$7, its near cancellation at the magic angle $\lambda \approx 2.7$8, and its $\lambda \approx 2.7$9 scaling over distances around $\sim n^2 a_0$00–$\sim n^2 a_0$01m (Méhaignerie et al., 2024). The same interaction was strong enough to act as a distance meter with tens-of-nanometers precision and to expose relative motion in the traps (Méhaignerie et al., 2024).

These pairwise couplings motivate effective spin models. In the circular-state simulator literature, two circular levels separated by $\sim n^2 a_0$02 or $\sim n^2 a_0$03 encode a pseudo-spin-$\sim n^2 a_0$04, while dipole-dipole exchange and diagonal shifts generate XX, XY, or XXZ couplings depending on the manifold choice and field geometry (Méhaignerie et al., 2023). The long-lived-spin limit is particularly attractive for exploring regimes inaccessible to low-$\sim n^2 a_0$05 Rydberg arrays, including slow thermalization and glassy dynamics; this suggests that the main novelty of CRAs is not stronger interactions alone, but a much larger ratio of coherent interaction time to decay time (Méhaignerie et al., 2023).

The interaction problem is not restricted to resonant exchange. In a magneto-electric trap, permanent dipoles induced by static fields can stabilize two circular atoms against collapse and set an equilibrium spacing; in that case the dipole-dipole interaction itself becomes part of the confinement mechanism (Hezel et al., 2011). A plausible implication is that CRA many-body architectures may ultimately combine exchange interactions, static dipolar forces, and structured electromagnetic environments in a single device.

6. Metrology, hybrid optical control, and current directions

CRAs have long been attractive for precision measurement because they suppress nuclear-overlap and many low-lying-structure corrections while preserving high field sensitivity. A notable proposal uses cold rubidium circular and near-circular states in an intensity-modulated optical lattice to determine the Rydberg constant through Doppler-free ponderomotive spectroscopy of the $\sim n^2 a_0$06 transition. In that scheme the projected relative uncertainty in $\sim n^2 a_0$07 is $\sim n^2 a_0$08 on Earth, with the dominant uncertainty from residual lattice shifts, and a microgravity implementation is projected to approach $\sim n^2 a_0$09; a central advantage is that the determination is independent of the proton radius (Ramos et al., 2017).

Recent alkaline-earth work has added an optical-control layer unavailable in alkali circular states. In doubly excited $\sim n^2 a_0$10Sr, a metastable $\sim n^2 a_0$11 ionic-core excitation was coupled to an $\sim n^2 a_0$12 circular qubit through the electric-quadrupole interaction

$\sim n^2 a_0$13

The resulting differential qubit shift between $\sim n^2 a_0$14 and $\sim n^2 a_0$15 was measured as $\sim n^2 a_0$16 kHz by beat-node Ramsey interferometry with spin echo, and no noticeable loss of qubit coherence was observed even while the core scattered about 100 photons during continuous optical driving (Wirth et al., 2024). This establishes that weak electron-electron interactions inside a circular Rydberg atom can be accessed coherently, and it suggests a route to laser cooling, imaging, and local optical manipulation through the ionic core (Wirth et al., 2024).

A complementary solution to the optical inaccessibility of alkali CRAs uses an ancilla array. In a dual-Rydberg rubidium platform, logical qubits were encoded in $\sim n^2 a_0$17 and $\sim n^2 a_0$18, while neighboring ancilla atoms were transiently excited to low-$\sim n^2 a_0$19 Rydberg states. A Förster resonance between $\sim n^2 a_0$20 and $\sim n^2 a_0$21 produced a blockade-based quantum non-demolition measurement of the logical qubit, with ancilla-based measurement fidelities $\sim n^2 a_0$22 and $\sim n^2 a_0$23, and the same ancilla mechanism enabled local phase and spin manipulation of individual logical CRA qubits (Machu et al., 29 Sep 2025). This directly addresses a longstanding misconception that CRAs are fundamentally incompatible with optical-tweezer architectures: while direct optical transitions are absent, hybrid optical–microwave control is demonstrably possible (Machu et al., 29 Sep 2025).

Historically, CRA research moved from atomic beams and cavity QED experiments associated with Hulet, Kleppner, Haroche, and others to laser-cooled and tweezer-trapped implementations with room-temperature mode engineering (Meinert et al., 2020, Wu et al., 2022). The current trajectory is defined by three converging developments: room-temperature Purcell-suppressed lifetimes in the 1–10 ms range and beyond, programmable optical trapping of individual and interacting circular atoms, and increasingly sophisticated microwave, optical-core, and ancilla-assisted control (Pultinevicius et al., 31 Oct 2025, Méhaignerie et al., 2024). A plausible near-term implication is that the distinct CRA advantages—long coherence, strong exchange, structured-mode engineering, and hybrid optical access—will be exploited most effectively not in isolation, but in composite architectures that combine capacitors, tweezers, microwave circuitry, and optically active auxiliary degrees of freedom.