---
title: Dissipative Dicke Lattice Dynamics
url: https://www.emergentmind.com/topics/dissipative-dicke-lattice
type: topic
---

# Dissipative Dicke Lattice Dynamics

Searching arXiv for recent and foundational papers on dissipative Dicke lattices.
A dissipative Dicke lattice is, in the strict sense, an open Dicke lattice in which each cavity mode is coupled collectively to an atomic ensemble and photons hop between neighboring cavities while also decaying out of the resonators. In a broader usage, the term also covers Dicke-type open many-body systems in which the photonic sector remains a single lossy mode while the lattice structure resides in the matter sector, as in Bose-Hubbard, Ising, or spin-chain realizations. Across these settings, the defining ingredients are collective light-matter coupling, Markovian photon loss, and nonequilibrium steady-state selection by dynamical stability rather than energy minimization [2607.00557].

## 1. Conceptual scope and relation to the Dicke model

The closed Dicke model describes a collective spin coupled to a single bosonic mode. The dissipative Dicke lattice generalizes this structure in two distinct ways. One is a literal lattice of cavity-spin units with on-site Dicke coupling, inter-cavity photon hopping, and local cavity decay. The other is a lattice-interacting matter system coupled to a single dissipative cavity mode, so that the photonic sector remains global while the matter sector carries spatial structure. The first usage is represented directly by coupled-cavity Dicke-lattice models [1405.3289, 2607.00557, 2508.10296]. The second is represented by Bose-Hubbard, Dicke-Ising, and interacting spin-chain extensions, which are not cavity arrays but are repeatedly treated as highly relevant to dissipative Dicke-lattice physics because they expose the effects of spatial correlations, local interactions, and finite-momentum matter modes on open Dicke dynamics [2012.11823, 2602.09912, 1904.01026].

This distinction matters because conclusions tied to a single global cavity coordinate do not directly carry over to a true cavity lattice without modification. A single-mode theory has one coherent-state cavity steady state and one scalar self-consistency variable, whereas a lattice theory admits momentum-dependent photonic instabilities, multiple spatial configurations, and boundary-sensitive stationary patterns. At the same time, the broader Dicke-type models supply controlled building blocks for lattice theory: fluctuation-induced heating and cooling, mode selection by susceptibilities, competition between collective order and local interactions, and dissipation-stabilized nonequilibrium attractors [2012.11823].

Related but structurally distinct generalizations include multimode disordered cavities, where photon-mediated long-range interactions and cavity loss generate normal, superradiant, and glass phases, and ordered free-space atomic arrays, where collective jump operators and geometry-dependent decay matrices realize Dicke-like many-body superradiance on a lattice of emitters [1304.5196, 2304.00093]. These systems are not Dicke lattices in the coupled-cavity sense, but they broaden the meaning of dissipative Dicke many-body physics by emphasizing multimode structure, nonlocal dissipation, and nonequilibrium criticality.

## 2. Canonical open Dicke-lattice formulations

A standard dissipative Dicke-lattice Hamiltonian is the Dicke lattice model implemented in hybrid quantum system arrays:
$$
H_{\rm DLM} =\Delta_c \sum_{\ell=1}^{N_L} a^\dagger_{\ell} a_{\ell} -t\sum_{\ell=1}^{N_L-1}(a^\dagger_\ell a_{\ell+1}+a_\ell a^\dag_{\ell+1}) + \sum_{\ell=1}^{N_L} \Delta_s  J^z_\ell + \sum_{\ell=1}^{N_L} \frac{G}{\sqrt{\mathcal N}}(J^+_{\ell}+J^-_\ell)(a_\ell+a_\ell^\dagger),
$$
supplemented by cavity loss,
$$
\dot \rho= -i[H_{\rm DLM},\rho] + \kappa \sum_\ell \left(2a_\ell \rho a^\dag_\ell- a^\dag_\ell a_\ell \rho- \rho a^\dag_\ell a_\ell\right).
$$
Each site contains a cavity mode and a collective spin representing an ensemble of NV centers, and the implementation is intrinsically driven and dissipative because the effective Dicke coupling is generated by cavity-assisted Raman processes while the dominant irreversible channel is photon leakage from the microwave cavities [1405.3289].

In momentum space, cavity hopping produces a photonic dispersion
$$
\Delta_k= \Delta_c-2t\cos k,
$$
so the normal-phase instability threshold becomes mode dependent,
$$
G_k=\sqrt{\frac{\Delta_k\Delta_s}{4}\left(1+\frac{\kappa^2}{\Delta_k^2}\right)},\qquad
G_{\rm crit}=\min_k G_k.
$$
This leads to three regimes. For small hopping, the \(k=0\) mode destabilizes first and the superradiant phase is homogeneous. For intermediate hopping satisfying
$$
0<\Delta_c-2t<\kappa<\Delta_c+2t,
$$
the first unstable mode occurs at finite wavevector \(k_c\) determined by \(\Delta_{k_c}=\kappa\), yielding
$$
k_c=\arccos\!\left(\frac{\Delta_c-\kappa}{2t}\right),\qquad
G_{\rm crit}= \sqrt{\frac{\kappa\Delta_s}{2}},
$$
and a patterned steady state
$$
\langle a_\ell\rangle \simeq \alpha \cos(\phi_0+k_c\ell).
$$
The finite-\(k\) superradiant phase is explicitly identified as a dissipation-induced feature absent in the equilibrium phase diagram [1405.3289].

A complementary strict lattice formulation is the dissipative ring of coupled Dicke units analyzed in the configuration-based treatment of a four-site Dicke lattice. There the system is a 1D ring of \(N\) coupled atom-cavity units with periodic boundary conditions, with local cavity frequency \(\omega_c\), atomic transition frequency \(\omega_a\), hopping \(\xi\), on-site collective coupling \(g\), and photon loss \(\kappa\). In the thermodynamic limit per site, \(N_a\to\infty\), the dynamics is treated semiclassically with rescaled cavity and spin variables and linear stability analysis around steady states [2607.00557].

## 3. Superradiant configurations, multistability, and nonequilibrium phase structure

The main organizing principle for strict dissipative Dicke lattices is that superradiant phases are best classified by spatial configurations of the local cavity order parameters, specifically the sign pattern of \({\rm Re}\,\langle c_i\rangle\). In the decoupled limit \(\xi=0\), each site is an independent open Dicke model with local \(\mathbb Z_2\) symmetry, so \(2^N\) superradiant sign patterns are possible above
$$
g_c=\frac{1}{2}\sqrt{\omega_a\omega_c\left(1+\frac{\kappa^2}{\omega_c^2}\right)}.
$$
With hopping, the independent local parities collapse to a global \(\mathbb Z_2\), and physically distinct patterns are grouped into configuration classes under global sign inversion and cyclic translations [2607.00557].

For the four-site dissipative ring, the representative classes are \([+\,+\,+\,+]\), \([+\,-\,-\,-]\), \([+\,+\,-\,-]\), and \([+\,-\,+\,-]\). These become four superradiant phase types: the homogeneous superradiant phase (HSRP), the staggered inhomogeneous phase ISRP1, the domain-like ISRP2, and the lower-symmetry ISRP3. The nonequilibrium phase diagram contains not only monostable regions but coexistence of two, three, and even all four superradiant phases; the paper labels a region \(J\) in which all four stable superradiant steady states coexist, so the asymptotic state depends on initial conditions [2607.00557].

For the four-site lattice, the normal phase is
$$
R_j=I_j=R_j^s=I_j^s=0,\qquad S_j^z=-\frac{1}{2},
$$
stable for \(g<g_c^{\rm NP}\), with
$$
g_c^{\rm NP} = \min_k \frac{1}{2} \sqrt{ \omega_a \omega_k \left( 1+\frac{\kappa^2}{\omega_k^2} \right) }.
$$
The superradiant branches differ by their spatial pattern:
\([+\,+\,+\,+]\) for HSRP, \([+\,-\,+\,-]\) for ISRP1, \([+\,+\,-\,-]\) for ISRP2, and \([+\,-\,-\,-]\) as representative of ISRP3. Positive hopping favors HSRP because it lowers the threshold associated with \(\omega_1=\omega_c-2\xi\), while negative hopping favors ISRP1 because it lowers the threshold associated with \(\omega_3=\omega_c+2\xi\) [2607.00557].

A central nonequilibrium result is that different dissipative configurations need not share the same universality class. Using DTWA and the on-site fluctuation
$$
\Delta_j = \langle c_j^\dagger c_j\rangle - \langle c_j^\dagger\rangle\langle c_j\rangle,
$$
the critical scaling near NP \(\to\) SRP boundaries is found to be
$$
\Delta_j^\beta \propto |g-g_c^\beta|^{-\gamma_{\rm open}},
$$
with \(\gamma_{\rm open}=-1\) for HSRP and ISRP1, but \(\gamma_{\rm open}=-\frac12\) for ISRP2. By contrast, the closed Dicke lattice selects a unique ground-state configuration by energy minimization and different equilibrium configurations share the same equilibrium universality class \(\gamma_{\rm closed}=1/2\) [2607.00557].

## 4. Finite-size chains and boundary-induced phases

Finite-size boundary conditions qualitatively reshape the stationary phase diagram. In a one-dimensional chain with Hamiltonian
$$
H = \sum_{i=1}^{N} H_{i}^{\rm Dicke} - \sum_{i=1}^{N-1} \xi \left( c_{i}^{\dagger} c_{i+1} + c_{i} c_{i+1}^{\dagger} \right) - \lambda \left( c_{1}^{\dagger} c_{N} + c_{1} c_{N}^{\dagger} \right),
$$
with \(\lambda=\xi\) for periodic boundary conditions and \(\lambda=0\) for open boundary conditions, the mean-field NP-to-SRP threshold is
$$
g_c^{P(O),NP} = \min_k \frac{1}{2} \sqrt{ \omega_a \omega_{P(O),k} \left( 1 + \frac{\kappa^2}{\omega_{P(O),k}^2} \right) },
$$
where
$$
\omega_{P,k} = \omega_c - 2\xi \cos\left[ \frac{2\pi(k-1)}{N} \right],\qquad
\omega_{O,k} = \omega_c - 2\xi \cos\left[ \frac{\pi k}{N+1} \right].
$$
The difference between traveling-wave ring modes and standing-wave open-chain modes already implies boundary-sensitive instability thresholds [2508.10296].

Under periodic boundary conditions, finite systems display the expected normal phase together with a homogeneous superradiant phase, an inhomogeneous superradiant phase, and a bistable region where both coexist. For \(N=3\), the homogeneous phase satisfies
$$
(\langle c_1\rangle,\langle S_1^-\rangle,\langle S_1^z\rangle)
=
(\langle c_2\rangle,\langle S_2^-\rangle,\langle S_2^z\rangle)
=
(\langle c_3\rangle,\langle S_3^-\rangle,\langle S_3^z\rangle).
$$
Under open boundary conditions, however, the homogeneous superradiant phase is completely absent for any finite \(N\ge 3\), and a whole zoo of inhomogeneous superradiant phases with broken translational symmetry appears [2508.10296].

The analytical reason is that a nonzero homogeneous ansatz is incompatible with the steady-state equations at edges and in the bulk. For open chains, the edge sites satisfy
$$
-\kappa\,\mathrm{Re}(\langle c_j \rangle) + (\omega_c - \xi)\,\mathrm{Im}(\langle c_j \rangle) = 0,
$$
whereas the interior sites satisfy
$$
-\kappa\,\mathrm{Re}(\langle c_j \rangle) + (\omega_c - 2\xi)\,\mathrm{Im}(\langle c_j \rangle) = 0.
$$
These relations cannot be satisfied simultaneously by a nonzero homogeneous field. Boundary mismatch therefore forbids a spatially homogeneous finite-amplitude steady state in any finite open chain [2508.10296].

For \(N=3\), the representative open-chain patterns include \(O_1\), with equal edge amplitudes and a distinct center; \(O_2\), with
$$
\langle c_1 \rangle = -\langle c_3 \rangle,\qquad \langle c_2 \rangle = 0;
$$
\(O_3\), with all three sites inequivalent; and \(O_4\), with side sites equal and opposite in sign to the center. The phase diagram contains monostable, bistable, and tristable regions. This establishes that in realistic finite arrays, boundaries are not a perturbation but a control parameter that can eliminate uniform superradiance and replace it with boundary-induced patterned order [2508.10296].

## 5. Fluctuation-selected mixed-state criticality

A central development for dissipative Dicke-lattice thinking comes from the open Dicke-type Bose-Hubbard chain coupled globally to a single lossy cavity mode. Although it is not a literal cavity lattice, it provides a controlled beyond-mean-field framework for open Dicke systems with extensive lattice matter. The model is
$$
H=H_c^{0}+H_b^{0}+H_{bc},
$$
with
$$
H_b^0=-J\sum_{j=1}^{L-1}(b_j^\dagger b_{j+1}+b_{j+1}^\dagger b_j)+\frac{U}{2}\sum_{j=1}^{L}n_j(n_j-1),
$$
$$
H_c^0=\hbar\delta\, a^\dagger a,\qquad
H_{bc}=-\frac{\hbar g}{\sqrt{L}}(a+a^\dagger)O,\qquad
O=\sum_{j=1}^{L}(-1)^j n_j,
$$
and Lindblad photon loss at rate \(\Gamma\) [2012.11823].

The key conceptual point is that mean field alone does not uniquely determine the matter steady state. The cavity relaxes to a coherent state, but the matter sector under the mean-field Liouvillian has a decoherence-free manifold: any density matrix built from eigenstates of \(H_b(\lambda)\) is stationary at mean-field level. The usual choice of the atomic ground state is therefore arbitrary in an open system. The physical steady state is selected only by fluctuations generated by
$$
\hat{\mathcal L}_1\rho=-\frac{i}{\hbar}[\delta H_{bc},\rho],
$$
which produce an effective evolution
$$
\partial_t\rho_0 \approx \left(\mathcal L_0-P_0\mathcal L_1\mathcal L_0^{-1}\mathcal L_1\right)\rho_0.
$$
Because \(\hat{\mathcal L}_1\sim 1/\sqrt L\), this becomes asymptotically exact in the thermodynamic limit [2012.11823].

When the matter subsystem thermalizes internally faster than cavity-induced heating and cooling act, the atomic density matrix is approximated by
$$
\rho^b\sim e^{-\beta H_b(\lambda)},
$$
and the effective temperature is fixed by the energy-balance condition
$$
\left\langle \frac{\partial H_b}{\partial t}\right\rangle = \frac{2\hbar g^2}{L}\int d\omega\, (1+n_B(\hbar\omega))\,\omega\,\Im\chi^R(\omega)\, \delta_\Gamma(\omega+\delta)=0.
$$
The resulting Dicke transition is therefore not simply a zero-temperature instability of matter in a coherent cavity field but generally a mixed-state transition selected by fluctuation-induced energy flow through the lossy photonic mode. The paper emphasizes that the effective temperature is already finite below threshold, with representative parameters giving \(k_B T/J\sim 1\) in the normal phase, and that this strongly shifts the critical coupling relative to zero-temperature mean field [2012.11823].

In the high-\(T\), large-\(\Gamma\) or large-\(\delta\) regime, the threshold becomes
$$
g_c \approx \frac{\delta^2+(\Gamma/2)^2}{\delta \sqrt{8n(1+n)}},
$$
whereas deep in the ordered phase the photon density saturates,
$$
\frac{\langle a^\dagger a\rangle}{L}\approx \frac12 n(1+n)\qquad (g\to\infty).
$$
The large-\(g\) asymptotics are
$$
k_B T\approx \frac{2\hbar g^2 n(1+n)\delta}{\delta^2+(\Gamma/2)^2},\qquad
\lambda\approx \frac{\delta \sqrt{2n(1+n)}}{\sqrt{\delta^2+(\Gamma/2)^2}},
$$
so stronger light-matter coupling simultaneously enhances heating and can self-limit order. The paper explicitly frames this as a useful lesson for dissipative Dicke lattices, where soft photonic modes coupled to extensive order parameters can impose universal mixed-state scales and strong-coupling saturation [2012.11823].

## 6. Driven dynamical regimes, temporal bistability, and nonstationary phases

Periodic driving and local interactions enrich dissipative Dicke systems beyond stationary superradiance. In a driven-dissipative Dicke model perturbed away from the exactly solvable collective-spin limit by adding nearest-neighbor spin interactions
$$
\hat H_{int}=J\sum_{i=1}^{N}\hat s_i^x \hat s_{i+1}^x,
$$
together with cavity loss
$$
\dot{\hat\rho}=-i[\hat H,\hat\rho]+\frac{\kappa}{2}\left(2\hat a\hat\rho \hat a^\dagger-\hat a^\dagger \hat a \hat\rho-\hat\rho \hat a^\dagger \hat a\right),
$$
the interplay of periodic bright/dark modulation of \(\lambda(t)\), dissipation, and finite-\(k\) spin waves yields stable Dicke time crystals, heating-prone regimes, dissipation-stabilized time crystals, irregular dynamics, overdamped relaxation, and metastable time crystals. The defining subharmonic response occurs at
$$
\nu=\frac{\Omega}{2},
$$
and ferromagnetic short-range interactions can reinforce the cavity-selected order rather than frustrate it, allowing stable time-crystalline behavior at non-perturbative coupling [1904.01026].

A distinct dynamical manifestation occurs in the dissipative Dicke-Bose-Hubbard system, where strong repulsive interactions make the Dicke normal-to-superradiant transition discontinuous below
$$
(J_c/U,\lambda_c/U)\approx (0.105,0.77).
$$
The resulting phase diagram contains coexistence windows MI+DW, SF+DW, and SF+SS. In real time, the coexistence region supports hysteresis and switching between SF and SS branches, while cavity loss damps this switching and drives the system to one of multiple steady-state attractors. At \(J/U=0.09\), \(\kappa/U=1.08\), and \(\lambda/U=0.88\), an initial SF state relaxes to an SF fixed point while an initial DW state relaxes to an SS fixed point; at \(\lambda/U=0.91\), both initial states relax to the same unique SS fixed point [2311.13301].

At the local level, open Dicke physics can also lose stationarity altogether. The dissipative Dicke model with spontaneous emission on each spin undergoes not only the usual superradiant phase transition
$$
g_\mathrm{c}=\sqrt{\frac{\omega\omega_{0,\gamma}}{4N}},
$$
but also a breakdown phase transition into a nonstationary heating regime with
$$
g_\mathrm{b}\approx \sqrt{\frac{\omega\omega_0}{N}}\approx 2g_c
$$
for the balanced case and sufficiently weak damping. The mechanism is a cooperative breakdown of the oscillator blockade: anti-Jaynes-Cummings pair creation together with spontaneous emission leaves one boson behind per cycle, and the heating becomes resonant when the collectively dressed manifolds align. Cooling via \(\hat L_\kappa=\sqrt{\kappa}\,\hat a\) shifts \(g_b\) upward and can suppress the breakdown entirely. The paper explicitly presents this as the single-site building block for dissipative Dicke-lattice theory [1807.06026].

## 7. Universality, implementations, and major limitations

Experimentally, dissipative Dicke lattices are closely tied to hybrid circuit-QED architectures. The implementation with arrays of superconducting microwave cavities and NV-center ensembles realizes a fully tunable model for collective light-matter interactions in the ultrastrong coupling limit through cavity-assisted Raman processes. The proposal remains viable even with substantial inhomogeneous broadening; for a Lorentzian distribution the superradiant threshold becomes
$$
G_{\rm crit} = \sqrt{\frac{\Delta_c\bar \Delta_s}{4}\left(1+\frac{\kappa^2}{\Delta_c^2}\right)\left(1+\frac{\gamma_s^2}{4\bar \Delta_s^2}\right)},
$$
and the minimum requirement can be expressed as strong collective cooperativity,
$$
\mathcal C_{\mathcal N}=\frac{2G^2}{\kappa\gamma_s}>1.
$$
The same work identifies the finite-\(k\) superradiant phase as a genuine dissipative signature and emphasizes that the array geometry is naturally feasible in microwave-resonator technology [1405.3289].

On the matter-lattice side, the dissipative Dicke-Ising model shows how the orientation of local interactions relative to the light-coupling axis changes the entire phase structure. In the transverse DIM, dissipation mainly shifts the phase diagram upward and the phase structure remains close to the ground-state one. In the longitudinal DIM, dissipation stabilizes bistable nonequilibrium steady states and induces first-order phase transitions absent in the ground-state phase diagram. The bistable phase is characterized by coexistence of superradiant and antiferromagnetic orders, and the ground-state triple point becomes a tetracritical point at
$$
g_{\mathrm{tet}}= \sqrt{\frac{\Omega(\omega^2+\kappa^2)}{2\omega}}.
$$
This is not a coupled-cavity photonic lattice, but it is a direct demonstration that spatially structured matter interactions plus photon loss can create nonequilibrium multistability and first-order Dicke transitions [2602.09912].

Two further extensions show the breadth of dissipative Dicke-lattice-type phenomena. In the multimode disordered open Dicke model, cavity loss does not destroy the quantum glass phase but modifies its universality class and produces a common low-frequency effective temperature for atoms and photons, together with characteristic fluorescence and \(g^{(2)}(\tau)\) signatures [1304.5196]. In ordered two-dimensional arrays of multilevel alkaline-earth-like atoms, collective decay matrices \(\Gamma_{jl}^a\) and collective jump operators \(\hat O_{\nu,a}\) organize a free-space Dicke-lattice problem in which superradiant bursts survive below geometry-dependent spacing thresholds and collective dissipation dynamically funnels branching into the dominant transition [2304.00093].

The principal limitations are equally clear. Strict coupled-cavity Dicke-lattice treatments are often semiclassical and rely on the thermodynamic limit per site \(N_a\to\infty\), mean-field factorization, photon loss as the only dissipation channel, and small finite lattices such as four-site or three-site rings and chains. The configuration-based universality analysis uses DTWA rather than exact Lindblad numerics for large \(N_a\) [2607.00557, 2508.10296]. Single-mode lattice-matter approaches do not capture photon hopping, photonic correlation lengths, or competition among many cavity modes, even when they provide controlled fluctuation physics [2012.11823]. This suggests that a complete theory of dissipative Dicke lattices must combine at least four elements already isolated across the literature: momentum- or configuration-resolved photonic instabilities, fluctuation-selected mixed-state steady states, finite-size and boundary sensitivity, and the possibility of genuinely dynamical attractors beyond static superradiance.

Source: https://www.emergentmind.com/topics/dissipative-dicke-lattice