Chiral Floquet Selection of an Algebraic Rydberg Bound State in the Continuum

AI summary

Overview

Research area: Quantum physics — Floquet engineering of periodically driven quantum systems, bound states in the continuum (BICs), Rydberg atom arrays, and long-range dipolar spin waveguides.

Technical level: Advanced. The paper is written in the language of Floquet–Sambe space, Feshbach projection, and self-energy formalisms, though the central physical idea (symmetry forbids radiation, but the state still has a long algebraic tail) is accessible to a motivated non-specialist.

Scope (one sentence): The paper demonstrates, theoretically and numerically, that a periodically driven four-site Rydberg plaquette coupled to an axial dipolar spin waveguide hosts a nonradiating Floquet bound state whose waveguide amplitude decays only as an inverse cube of distance, with its chirality set by coherent Floquet hybridization.

What This Paper Is About

A bound state in the continuum is a localized state that survives inside a spectrum of extended, propagating states — it does not radiate even though it has somewhere to radiate to. When the coupling to that continuum is long ranged, the usual question ("is it nonradiating?") is not enough: one must also ask what shape the state takes in space. The authors construct a closed, Hermitian Rydberg system — a winding-driven four-site plaquette attached to a one-dimensional dipolar spin waveguide — in which a combined rotation–time-translation symmetry removes every propagating guide channel from one symmetry sector, and they then determine the spatial decay that the residual, nonpropagating coupling imprints on the bound state.

Key Contributions

1. A Floquet selection rule linking drive harmonics to angular channels. The authors derive the relation ℓ_eff = ℓ_s − r (mod 4), which connects the Fourier order r of the winding phase factor to the angular channel of the guide that the drive addresses. For the axial geometry the static channel is ℓ_s = 0 and r = 1 couples to ℓ_eff = −1.

2. A joint rotation–time-translation quantum number κ = m − ℓ (mod 4). This label combines the Sambe-space harmonic index m with the plaquette angular index ℓ, and the paper shows that the bound branch lives in the κ = 2 sector, whose accessible guide harmonics are m = 2 (mod 4). The nearest allowed sideband energies E_±2 fall outside the undriven guide band, while the in-band m = 0, 1 replicas belong to other sectors.

3. Identification of an algebraic, not exponential, bound state. The nonradiating waveguide amplitude follows |ψ_n| ~ n^(−3), so the corresponding probability is summable as n^(−6). The paper attributes this exponent to the combination of a direct plaquette–guide contact source and long-range propagation through the guide.

4. A microscopic Rydberg realization with pair-state input from ARC. Using the Alkali Rydberg Calculator (ARC) at 20 G, the authors evaluate 300 pair exchanges for a 25-atom device (and 1283 distinct pair-exchange calculations for the axial extension to L_w = 641), showing that the axial geometry produces a purely ℓ_s = 0 static contact to numerical precision.

Main Findings

• Radiation is excluded by symmetry, not by cancellation. The state's quasienergy sits in the global continuum, but no guide mode propagates inside its own κ = 2 sector. The paper contrasts this with Friedrich–Wintgen cancellation between open resonant paths: here the joint sector simply has no propagating guide replica.

• The tail is inverse-cube, and the exponent is measured numerically. For the L_w = 641 state, fits across three interior windows give α = 3.006–3.010. The computed profiles are compared against a distance-only geometrical reference K_geo(n) ≡ (R_{n₀}/R_n)³ ~ n^(−3), normalized at site n₀, with the Floquet snapshot normalized at n = 108 in the L_w = 321 comparison.

• Both the contact source and guide propagation set the leading coefficient. Writing the guide response as q_m and the contact source as f_m, with q_m = (E_m − H_w)^(−1) f_m and E_m = ε_B + mΩ, the off-band equation balances the direct source A_m against a propagation term J_∞ M_m. For the converged single-tone L_w = 641 state, the finite-guide ratios J_∞ M_m^(L_w)/A_m are +0.0896 for m = 2 and −0.0484 for m = −2 — so propagation can either reinforce or oppose the direct source, and exact cancellation of the two would remove the inverse-cube leading term.

• Chirality comes from coherent Floquet mixing, not from the selection rule alone. In the onsite representation the dominant components are (ℓ, m) = (2, 0), (−1, 1), and (0, 2), with bare Floquet energies 0.5731 J*, −0.2478 J*, and −0.6443 J*. The closed m = 2 guide block contributes a dispersive self-energy of 0.3530 J*, and the fundamental drive couples neighboring harmonics with magnitude 0.6655 J*. Transforming back to the phase representation, the three leading contributions to a_{−1,1} have magnitudes 0.2640, 0.1560, and 0.3159 with a common phase, giving |a_{−1,1}|² ≃ 0.61817.

• The branch is spectrally stable as the guide grows. Plaquette weight, target-chirality weight, and the inverse participation ratio (IPR) remain essentially unchanged across L_w = 21–641. At L_w = 641 the nearest extended-state spacings are Δ_ext/J* ≃ 1.09 × 10^(−4) for the single-tone drive and 6.41 × 10^(−4) for the biharmonic drive.

• The finite-guide result is rigorously tied to an infinite-guide embedded eigenstate. A residual bound at L_w = 641 with harmonic cutoff M = 16 gives full infinite-operator residuals of 5.24 × 10^(−7) J* and 5.39 × 10^(−7) J* for the two waveforms, while the distance to the κ = 2 continuum exceeds 0.818 J*. Each residual interval therefore lies inside the sector's gap and inside the continuum of another sector.

• Two robustness checks pass. Following the branch through five values of z₂ between 0 and 0.1 leaves the adjacent-state squared overlap above 0.99996. A transverse displacement δ = Δx/a_w breaks the C₄ contact equivalence with a quadratic onset, η_{C₄}(δ) = c_δ δ² + O(δ⁴), while the finite Floquet state remains continuously connected: squared overlap with the axial state is 0.918 at δ = 0.20.

• The snapshot chirality is close to but not identical to its period average. For the single-tone state, P_{χ−}(0) ≃ 0.615 and the period-averaged P̄_{χ−} ≃ 0.618, the small difference arising from interference between harmonics during a drive period.

• The device geometry is specified in atomic units. The 25-atom device has plaquette side s ≃ 12.4128 μm, axial offset h ≃ 3.9489 μm, and guide spacing a_w ≃ 12.7781 μm at B = 20 G = 0.002 T, with imposed onsite terms K_jj/J* = 1.08 − 0.01170571942455195 and (H_w)_nn/J* = 0.70. The dimensionless drive frequency is Ω/J* = 1.21, corresponding to a physical ν_drive ≃ 7.865 MHz, with z₁ = 1.1 and φ₁ = 0.49π.

Methodology in Plain English

The authors build a small model universe: four Rydberg atoms forming a plaquette, plus a line of atoms forming a one-dimensional waveguide attached along the plaquette's axis. They work in the sector with exactly one excitation, so the state is either a plaquette excitation or a waveguide excitation, and no loss or absorber is included anywhere — the system is closed and Hermitian.

The role of the drive is to make the plaquette atoms pick up time-dependent phases that wind around the plaquette (a first harmonic with amplitude z₁ and a second harmonic with amplitude z₂/2). Because the winding is tied to the atom's angular position, a drive harmonic of order r carries an angular label −r, and that is what lets the authors steer which angular channel of the waveguide the state can couple to. Combining rotation with time translation yields the conserved label κ = m − ℓ (mod 4), which is the key bookkeeping device: it tells you exactly which harmonics of the guide are reachable from a given state.

For the microscopic inputs, they do not use a toy coupling. They take the actual atomic positions and angles and run the Alkali Rydberg Calculator to get every pair exchange at B = 20 G, then assemble the static Hamiltonian from the resulting plaquette–plaquette links, guide–guide links, and plaquette–guide contacts.

To find the driven states, they numerically diagonalize the one-period evolution operator U(T) and track the branch as the guide is lengthened from L_w = 21 up to L_w = 641. To separate "does it radiate?" from "what shape does it have?", they use a Feshbach projection: project onto the plaquette (P) and the guide (Q), eliminate the guide, and obtain two separate conditions — one that the coupling to open propagating channels vanishes (Eq. 3, first line), and one self-consistent eigenvalue equation for the bound state (Eq. 3, second line). The first condition is met by symmetry; the second is solved numerically.

Finally, to prove the finite-chain result really corresponds to an infinite-guide embedded state, they bound the residual ‖(H_F^(κ) − ε_{L,M}) u_{L,M}‖ and show the resulting interval sits in a gap of the κ = 2 sector while remaining inside the continuum of other sectors.

Why This Matters

The result reframes what a bound state in the continuum can look like. Standard BIC narratives emphasize exponential confinement or delicate destructive interference between open channels. This paper shows a third route: a joint space-time symmetry empties the propagating channel set entirely, so the state is nonradiating by construction, and yet the residual long-range coupling still sculpts a power-law tail. That means a BIC's spatial profile can be a diagnostic of the long-range coupling itself — the decay exponent carries information even when no radiation escapes.

The work also connects two normally separate conversations: Floquet engineering of driven quantum matter and BIC physics in structured waveguides. Because the protection relies on a combined symmetry rather than parameter fine-tuning, the authors argue the mechanism is not tied to the specific Rydberg parameters used; the paper notes that for each system the existence of the self-consistent embedded eigenstate and its decay exponent must still be determined separately.

Potential applications (none of these are demonstrated in the paper, which is a theory and numerical study; the manuscript reports no device, experiment, or application benchmark):

• Rydberg atom arrays used as programmable quantum simulators, where robust localized excitations could serve as protected storage or as probes of long-range dipolar coupling.
• Photonic or waveguide-lattice platforms, where algebraically localized embedded states have already been demonstrated experimentally in engineered lattices and could be extended to driven, symmetry-selected designs.
• Quantum sensing and metrology schemes that exploit the extreme sensitivity of a state's spatial profile to a transverse displacement — the paper shows the unwanted angular-channel power grows as δ², which is the kind of sharp geometric turn-on that displacement sensing relies on.
• Design of driven quantum devices in which radiationless operation is required without tight geometric tolerances, since the waveform family preserves the symmetry over a continuous range of z₂.

Industry relevance. The paper is fundamental research and reports no industry partnership, fabrication, or system-level performance metric. Its relevance to quantum-hardware efforts is indirect: it offers a design principle (symmetry-selected channel exclusion plus a long-range guide) for protecting excitations in atom-array and waveguide-based architectures, and it shows how microscopic atomic parameters from ARC translate into a macroscopic, measurable spatial observable.

Future Directions

• Experimental realization and detection. The paper's predictions are numerical. Verifying the n^(−3) envelope and the predicted chirality weights (P_{χ−}(0) ≃ 0.615, |a_{−1,1}|² ≃ 0.61817) would require site-resolved readout of a driven Rydberg array, which the manuscript does not report.

• Extension beyond the single-excitation sector. The entire treatment lives in the single-excitation sector of a four-site plaquette and one-dimensional guide. Whether the joint-symmetry protection survives in the interacting, multi-excitation regime — where the guide itself would need to be treated as a driven continuum rather than the time-independent block used here — is left open. The Supplemental Material explicitly notes that if the continuum itself is driven, the Feshbach construction must use its Floquet scattering modes instead.

• Cancellation tuning and exponent control. The reported ratios J_∞ M_m^(L_w)/A_m of +0.0896 (m = 2) and −0.0484 (m = −2) show propagation can reinforce or oppose the direct source. The paper notes that exact cancellation would remove the leading inverse-cube coefficient, which raises the question of whether the decay exponent can be deliberately tuned between power laws by adjusting geometry or waveform.

• Higher-dimensional and multi-plaquette geometries. The construction depends on an axial geometry with a purely ℓ_s = 0 static channel and a C₄ plaquette. Whether analogous selection rules produce algebraic BICs in other rotational symmetries (C₃, C₆) or in coupled plaquette networks, and how robust the protection is to the transverse displacements quantified by η_{C₄}(δ), are natural extensions the paper does not address.

Target Audience

This paper is most valuable to researchers already working in Floquet engineering, BIC physics, or Rydberg quantum simulation — particularly theorists and numerical physicists comfortable with Sambe-space formalisms, Feshbach projection, and quasienergy spectra. It will also interest experimental groups building Rydberg atom arrays with individually controlled geometries and dipolar exchange, who may want to test the predicted spatial envelope. Readers seeking a conceptual introduction to BICs or Floquet driving will find the physics picture in Figure 1 accessible, but the derivations in the Supplemental Material assume graduate-level quantum mechanics and familiarity with Green's function methods.

Authors’ abstract

Bound states in the continuum (BICs) remain localized despite the presence of extended modes at the same energy. In driven systems with long-range couplings, the absence of radiation does not by itself determine how a bound state decays in space. We study a Hermitian Rydberg spin waveguide coupled to a four-site plaquette, where a winding drive phase combines rotation with time translation to forbid propagating guide channels in one symmetry sector. The resulting eigenstate remains embedded in the continuum of other sectors and carries an inverse-cube waveguide amplitude, with coherent contributions from direct plaquette--guide coupling and long-range guide propagation. Pair-state calculations determine the microscopic couplings in the axial geometry, while coherent Floquet mixing sets the weight of the dominant circulating plaquette mode. The bound state thus retains the spatial signature of the long-range coupling even when radiation is forbidden.