Dissipatively Coupled Resonator Arrays

• Dissipatively coupled resonator arrays (DCRAs) are engineered networks combining coherent photon hopping and engineered dissipative channels to create non-equilibrium steady states.
• They utilize advanced models with non-Hermitian Hamiltonians and Lindblad dissipators to capture rich many-body phenomena and nonlinear dynamics.
• Applications include quantum simulation, photonic computation, and topologically robust devices relying on tunable inter-resonator couplings and nonreciprocal light transport.

Dissipatively coupled resonator arrays (DCRAs) are engineered photonic, plasmonic, or hybrid networks in which electromagnetic energy is transferred between resonator modes both via coherent (Hermitian) photon hopping and via engineered dissipative (non-Hermitian) channels. The interplay of coherent tunneling, nonlinearities (e.g., Kerr or Jaynes–Cummings), external driving, and unavoidable or intentional dissipation gives rise to rich non-equilibrium steady states, non-trivial collective phenomena, and controllable light–matter interactions. DCRAs unify quantum optical concepts with non-Hermitian many-body physics, and underpin key advances in quantum simulation, photonic computation, and topologically robust photonic devices.

1. Theoretical Foundations and Model Hamiltonians

DCRAs are most commonly modeled using non-Hermitian extensions of bosonic tight-binding Hamiltonians, augmented with Lindblad dissipators to account for photon leakage, incoherent pumping, and engineered couplings.

Generic Bose–Hubbard model with drive and loss:

$$H_{\rm BH} = \sum_i \left[\omega_c\,a_i^\dagger a_i + \frac{U}{2}\,a_i^\dagger a_i^\dagger a_i a_i\right] - J\!\!\sum_{\langle i,j\rangle} a_i^\dagger a_j + \sum_i\left[F_i\,e^{-i\omega_d t}\,a_i^\dagger + \mathrm{h.c.}\right]$$

• $a_i$ ($a_i^\dagger$): bosonic annihilation (creation) operators in resonator $i$,
• $\omega_c$: on-site resonance,
• $U$: onsite Kerr nonlinearity,
• $J$: coherent intersite coupling,
• $F_i$: external drive,
• Dissipation: Lindblad operator $\mathcal{D}[a_i][\rho]$ with photon loss rate $\kappa$ (Noh et al., 2017).

Jaynes–Cummings–Hubbard model:

$$H_{\rm JCH} = \sum_i\left[\omega_c\,a_i^\dagger a_i + \omega_a\,\sigma_i^+\sigma_i^- + g(a_i^\dagger \sigma_i^- + a_i \sigma_i^+)\right] - J \! \! \sum_{\langle i,j\rangle} a_i^\dagger a_j$$

• $\sigma_i^\pm$: two-level matter excitations; $g$: light–matter coupling (Noh et al., 2017, Grujic et al., 2012).

Dissipative coupling:

Non-Hermitian effects are introduced via complex hoppings $J_{ij}$ (e.g., $J_{ij} = |J_{ij}|\,e^{i\phi_{ij}}$) or direct dissipative Lindblad cross-terms, representing photon exchanges with a common lossy environment or through radiative channels (Ruiz-Rivas et al., 2014, Miao et al., 2024).

Non-Hermitian coupled-mode matrices:

For linear arrays,

$$H_{n,m} = (\omega_0 - i\gamma)\,\delta_{n,m} + \kappa_1\,\delta_{n,m-1} + \kappa_2\,\delta_{n,m+1}$$

with $\kappa_{1,2}$ complex-valued ("asymmetric" or "dissipative" Hopf couplings) (Longhi, 2022).

Advanced frameworks:

The coupled-quasinormal-mode (cQNM) formalism systematically describes arbitrary dissipative/coupled arrays, capturing both conservative and dissipative (time-derivative) couplings via mode overlaps of QNMs (Wu et al., 27 Jul 2025).

2. Dissipative Coupling Mechanisms

Dissipative coupling mechanisms are realized either by engineered radiative channels (common waveguides, plasmonic near-fields, auxiliary link rings) or by exploiting coupling to shared lossy baths. The effective inter-cavity coupling then comprises both coherent (Hermitian) and dissipative (anti-Hermitian) parts. In master-equation form, the dissipator reflects cross-correlation terms: $$\sum_{i,j}\frac{\Gamma_{ij}}{2}\left(2a_j\rho\,a^\dagger_i - a^\dagger_i a_j\rho - \rho a^\dagger_i a_j\right)$$ with $\Gamma_{ij}$ complex-valued dissipative coupling matrix elements, typically $$\Gamma_{ij} = \sqrt{\Gamma_i \Gamma_j}\, e^{i\phi_{ij}}$$ for radiative channels (Miao et al., 2024).

In the cQNM formalism, dissipative couplings arise as overlaps of QNM fields: $$g_{ij}^{mn} = \int_{\Omega_i} E_m^{(i)}(\vec r) \cdot E_n^{(j)}(\vec r) d^3 r$$ and appear as time-derivative couplings in evolution equations (Wu et al., 27 Jul 2025).

Dissipative couplings enable phenomena such as non-Hermitian skin modes, level attraction, zero-coupling points, and nonreciprocal light transport.

3. Non-Equilibrium and Many-Body Phenomena

Driven-dissipative dynamics, in combination with onsite and intersite nonlinearities, produce a spectrum of many-body phases and transitions unique to DCRAs (Noh et al., 2017, Grujic et al., 2012, Ruiz-Rivas et al., 2014).

• Photon fermionization: In the hardcore (large-$U$ or large-$g$) regime, photons exhibit strong anti-bunching and two-photon blockade reminiscent of fermionic statistics (Noh et al., 2017).
• Crystallization and density-wave order: Phase-engineered drives yield steady states with dimer or density-wave order—manifested as spatial bunching and reduced long-range correlations (Noh et al., 2017).
• Quantum Hall analogs: Synthetic gauge fields (complex hopping phases) permit realization of Laughlin-type fractional quantum Hall correlations, with steady-state two-photon densities exhibiting over 90% overlap with $\nu=1/2$ bosonic Laughlin states in small arrays (Noh et al., 2017).
• Finite-range collective coherence: Despite dissipation and incoherent pump, arrays develop long but finite first-order coherence, with an exponential decay determined by the photon lifetime and inter-cavity coupling. The resulting correlation length scales logarithmically or algebraically with system parameters depending on the tunneling–interaction balance (Ruiz-Rivas et al., 2014).

Summary of observable signatures:

Observable	Physical Meaning	Dissipative Array Context
$g^{(2)}(0)$	Zero-delay photon correlations	Anti-bunching (Mott/blockade), bunching (crystallization) (Noh et al., 2017, Grujic et al., 2012)
$S(\omega)$	Emission spectrum	Multiplet structure, Mollow triplets, QH excitations (Ruiz-Rivas et al., 2014)
$g^{(1)}(r)$, $C(r)$	First-order spatial coherence	Exponentially decaying in DCRAs, with tunable range (Ruiz-Rivas et al., 2014)
$S(k)$	Momentum space structure factor	Reveals population of supermodes; band populations under drive (Grujic et al., 2012)

4. Non-Hermitian Band Structure, Topology, and Skin Effects

DCRAs are inherently non-Hermitian lattices with complex-valued couplings and onsite losses/gains. The interplay of topology and non-Hermitian physics enables a range of phenomena:

• Non-Hermitian skin effect: Asymmetric or dissipative couplings cause eigenmodes to localize exponentially at array boundaries under open boundary conditions, in contrast to extended Bloch modes in Hermitian systems. The non-Hermitian dynamical matrix formalism reveals exact spectra and eigenmode structure (Longhi, 2022).
• Topological windings: In ring arrays with periodic boundary conditions, the loci of complex eigenvalues $E(k)$ as $k$ traverses the Brillouin zone define nontrivial windings in the complex plane; these “loops” label distinct mode families (lowest-loss or nested) (Hashemi et al., 30 May 2025).
• Phase-locked emission and supermode selection: By tuning boundary or edge resonator parameters, a single extended Bloch-like supermode can be isolated, with tunable phase difference and robust against both disorder and dynamical instability (Longhi, 2022).
• Level attraction and zero coupling: Non-Hermitian couplings can induce level coalescence (attraction of resonance frequencies), and analytic conditions exist for vanishing net coupling even at subwavelength separations (Wu et al., 27 Jul 2025).

5. Nonreciprocity, Bistability, and Control of Light Transport

Dissipatively coupled arrays are a fertile ground for engineered nonreciprocity—light propagation asymmetry not possible in closed, time-reversal symmetric Hermitian systems.

• Nonreciprocity via nonlinearity: Addition of Kerr nonlinearity enables strong directional asymmetry in input–output transmission, even when the linear system is strictly reciprocal. Bistability near nonlinear turning points gives rise to sharp contrasts in forward vs backward transmission (Miao et al., 2024).
• Design criteria: Optimal nonreciprocity arises with reciprocal linear couplings, nontrivial intersite phase $\phi$, and input power near the bistability threshold for one propagation direction (Miao et al., 2024).
• Figures of merit: Isolation ratio, insertion loss, and operational bandwidth quantify device performance.

6. Advanced Phenomena: Reconfigurable Soliton Combs and Photonic Applications

Arrays with carefully engineered dissipative and dispersive couplings realize reconfigurable nonlinear states not accessible in single-resonator systems.

• Non-Hermitian soliton combs: Arrays of coupled Kerr resonators with synthetic phase and dissipative couplings produce dissipative Kerr solitons (DKS) whose frequency line numbers and spacings can be tuned in situ by varying hopping phases, without any hardware redesign (Hashemi et al., 30 May 2025).
• Topological selection of comb supermodes: By controlling non-Hermitian winding, only a desired subset of array supermodes is robustly phase-locked, while all others are lossy. This enables dynamic selection of comb parameters post-fabrication.
• Microscale implementation: High-$Q$ Si$_3$N$_4$ or Hydex integrated platforms, thermal phase-tuning, and ring-based bus coupling architectures have been demonstrated compatible with the required array parameters (Hashemi et al., 30 May 2025).

7. Analytical, Numerical, and Computational Methods

The non-equilibrium quantum many-body nature of DCRAs requires advanced methods:

• Mean-field and cluster approaches: Provide simple phase diagrams and qualitative insights, but fail to capture strong correlation or quantum fluctuation regimes (Noh et al., 2017).
• Exact diagonalization: Feasible for small arrays, with explicit Liouvillian construction and quantum trajectory methods (Grujic et al., 2012).
• Matrix-product states (MPS) and TEBD: In 1D, time-evolving block decimation applied to quantum trajectories and Lindblad master equations offers scalable access to NESS in arrays up to $O(16)$ sites (Grujic et al., 2012).
• cQNM and modal approaches: Rigorous, efficient eigenmode-based modeling, closing the gap with full Maxwell simulations but enabling rapid inverse design and parametric scans (Wu et al., 27 Jul 2025).

---

Dissipatively coupled resonator arrays thus constitute a versatile, robust, and highly tunable platform for non-Hermitian quantum optics and photonics, bridging fundamental many-body physics with emergent technology for quantum simulation, optical nonreciprocity, and dynamically configurable photonic devices (Noh et al., 2017, Liu et al., 2010, Ruiz-Rivas et al., 2014, Miao et al., 2024, Wu et al., 27 Jul 2025, Longhi, 2022, Grujic et al., 2012, Hashemi et al., 30 May 2025).