Non-symmetric quantum interfaces with bilayer atomic arrays

1. Problem Statement

Quantum interfaces are essential for coupling atomic excitations to propagating photon modes, serving as the backbone for quantum memories, networks, and entanglement generation. While subwavelength atomic arrays have shown promise in achieving high-efficiency coupling via collective effects, a significant challenge arises in superwavelength regimes (where lattice spacing $a > \lambda$). In these regimes, light scatters into higher-order diffraction modes, leading to substantial radiative losses that degrade interface efficiency.

Previous approaches relied on Bragg-symmetric configurations, where the interlayer spacing $a_z$ satisfies $a_z = N\lambda/2$. In this symmetric regime, the interface efficiency is determined solely by the reflectivity ($r$), and destructive interference can cancel diffraction losses. However, the Bragg condition restricts the design space to discrete values of $a_z$ and $a$, limiting flexibility and preventing the optimization of efficiency in realistic, finite-size arrays. The authors ask: Can we achieve superior interface efficiencies by operating in non-symmetric regimes with arbitrary interlayer spacings?

2. Methodology

The authors develop a general theoretical framework to analyze bilayer atomic arrays with arbitrary interlayer spacing ($a_z$) and lattice spacing ($a$).

• Mapping to a 1D Scattering Model:
  The complex many-body problem of a 2D bilayer array is mapped onto a minimal one-dimensional (1D) quantum interface model. This model features a collective atomic excitation ($\hat{P}_q$) coupled to a target photon mode ($\hat{E}_q$) with rate $\Gamma_q$ and decaying into loss channels with rate $\gamma_{q,loss}$.
• Universal Efficiency Relation:
  They derive a universal relationship showing that the quantum interface efficiency ($r_q$) for tasks like memory storage is fully determined by classical scattering observables: reflection ($r_\pm$) and transmission ($t_\pm$).
  $$r_q = \frac{1}{2} \left[ 1 - t_\pm - r_\pm \left( \frac{c_{q\mp}}{c_{q\pm}} \right) \right]$$
  Unlike the symmetric Bragg case where $r_q = |r|$, the non-symmetric regime requires both $r$ and $t$ to characterize the efficiency.
• Bilayer Dynamics:
  The system consists of two identical 2D atomic layers (square or triangular lattices). The authors solve the Heisenberg-Langevin equations for the collective dipoles, accounting for dipole-dipole interactions mediated by free-space photons. They identify two collective eigenmodes: in-phase ($q=0$) and out-of-phase ($q=\pi$).
• Loss Mechanism Analysis:
  They explicitly calculate the loss term $\gamma_{q,loss}$, which includes individual atomic decay ($\gamma_s$) and radiative diffraction losses ($\gamma_{q,diff}$). The key insight is that $\gamma_{q,diff}$ can be engineered to zero via destructive interference between layers, even when $a_z$ is not a Bragg multiple of $\lambda/2$.

3. Key Contributions

1. Generalization Beyond Bragg Symmetry: The paper establishes that non-symmetric interfaces (arbitrary $a_z$) offer a continuous family of solutions for high-efficiency coupling, surpassing the discrete limitations of Bragg-symmetric designs.
2. Continuous Resonant Curves: The authors identify "resonant curves" in the $(a, a_z)$ parameter space where diffraction losses ($\gamma_{q,diff}$) are exactly cancelled via destructive interference. This allows for the optimization of efficiency across a continuous range of lattice spacings, rather than just discrete points.
3. Higher-Order Diffraction Cancellation: They demonstrate that by lifting the Bragg constraint, it is possible to simultaneously cancel contributions from multiple radiative diffraction orders (e.g., first and second orders) using only two layers. This is impossible in Bragg-constrained systems where phase matching for multiple orders cannot be satisfied simultaneously.
4. New Quantum Memory Protocol: They propose a quantum memory scheme based on two-level atoms (eliminating the need for auxiliary metastable states). This protocol utilizes the continuous tuning of the interlayer spacing $a_z$ to dynamically switch the coupling rate $\Gamma_q$ between a "bright" state (for storage) and a "dark" state (for storage), effectively turning the interface on and off.

4. Key Results

• Enhanced Efficiency in Superwavelength Regimes:
  Numerical simulations on finite-size arrays (using Gaussian beams) show that non-symmetric configurations on the resonant curves achieve significantly higher efficiencies than Bragg-symmetric counterparts. In specific examples, the inefficiency ($1-r_q$) was reduced by a factor of 5 compared to optimal Bragg designs.
• Scaling with System Size:
  The coupling inefficiency scales favorably with the number of atoms per layer ($N$), following a power law $1-r_q \propto N^{-1}$. This indicates that residual losses are dominated by imperfect cancellation of diffraction orders at the edges of finite arrays, which diminishes as the array grows.
• Elimination of Higher Diffraction Orders:
  For lattice spacings where two diffraction orders are radiative, the non-symmetric approach allows for the exact cancellation of both orders simultaneously, creating discrete "resonant sets" of $(a, a_z)$ that yield near-unity efficiency ($r_q \approx 1$).
• Quantum Memory Performance:
  The proposed two-level memory protocol achieves storage efficiencies ($r_f$) exceeding 0.9 for realistic parameters (e.g., $N \sim 4000$, switching times $\tau \Gamma_{1D} \sim 100$). The efficiency is bounded by the interface efficiency $r_q$ and the speed of the mechanical tuning of the layers.

5. Significance

• Flexibility in Quantum Design: This work removes the rigid constraint of Bragg symmetry, providing experimentalists with a much broader design space for optimizing light-matter interfaces. It allows for the use of larger lattice spacings (easier to fabricate with optical tweezers) without sacrificing efficiency.
• Scalability: By enabling high-efficiency coupling in the superwavelength regime, this approach facilitates the scaling of quantum networks and memories using large arrays of atoms, which are prone to diffraction losses in previous designs.
• Simplified Atomic Requirements: The proposed memory scheme relies only on two-level atoms, avoiding the complexity and decoherence issues associated with three-level systems (ladder or $\Lambda$ configurations) that require long-lived metastable states.
• Universal Characterization: The derivation linking quantum task efficiency directly to classical scattering observables ($r$ and $t$) provides a powerful, experimentally accessible tool for characterizing and optimizing any quantum interface, symmetric or not.

In conclusion, the paper demonstrates that non-symmetric bilayer atomic arrays constitute a flexible and superior platform for quantum interfaces, offering a pathway to near-unity efficiency in free-space quantum technologies by exploiting continuous interference control rather than discrete Bragg conditions.