Anisotropic Dicke Model: Key Aspects

• The anisotropic Dicke model is a collective light–matter Hamiltonian that couples a bosonic mode with an ensemble of two-level systems using unequal rotating and counter-rotating interactions.
• It interpolates between U(1) and Z2 symmetry regimes, enabling detailed studies of phase transitions, spectral statistics, and ergodicity in quantum systems.
• Its tunable anisotropy facilitates explorations in quantum metrology, collective-mode spectroscopy, and nonequilibrium transport across diverse experimental platforms.

Searching arXiv for recent and foundational papers on the anisotropic Dicke model to ground the article in the provided literature. The anisotropic Dicke model is a collective light–matter Hamiltonian in which a single bosonic mode is coupled to an ensemble of $N=2j$ two-level systems with unequal rotating-wave and counter-rotating-wave amplitudes. In its most common form, the anisotropy is the independent weighting of the processes $a^\dagger J_-+aJ_+$ and $a^\dagger J_++aJ_-$, so that the model interpolates between excitation-conserving $U(1)$ limits and the parity-symmetric $\mathbb Z_2$ Dicke regime. This asymmetry underlies the model’s ground-state normal-to-superradiant transition, its excited-state and ergodic structures, its driven and open-system generalizations, and its use as a platform for critical metrology, collective-mode spectroscopy, and quantum transport (Buijsman et al., 2016, Das et al., 2023, Zhu et al., 2024).

1. Hamiltonian, notation, and relation to the standard Dicke model

A widely used definition of the anisotropic Dicke model is

$$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$

where $a,a^\dagger$ are bosonic operators, $J_z,J_\pm$ are collective spin operators, $\omega$ is the boson frequency, $\omega_0$ is the atomic splitting, and $a^\dagger J_-+aJ_+$0 are the two independent couplings (Buijsman et al., 2016, Hu et al., 2020, Das et al., 2023, He et al., 27 May 2025). The model becomes the standard Dicke model when $a^\dagger J_-+aJ_+$1, and it reaches integrable limits when either coupling vanishes (Buijsman et al., 2016, Hu et al., 2020).

Different subliteratures use different anisotropy conventions. In the Dicke–Stark formulation, the rotating-wave coupling is $a^\dagger J_-+aJ_+$2 and the counter-rotating-wave coupling is $a^\dagger J_-+aJ_+$3, so the isotropic limit is $a^\dagger J_-+aJ_+$4 (Chen et al., 2024). In the driven Rydberg-array realization, the normalized anisotropy parameter is

$a^\dagger J_-+aJ_+$5

with $a^\dagger J_-+aJ_+$6 and $a^\dagger J_-+aJ_+$7, giving $a^\dagger J_-+aJ_+$8 for pure rotating-wave coupling, $a^\dagger J_-+aJ_+$9 for the isotropic Dicke model, and $a^\dagger J_++aJ_-$0 for pure counter-rotating-wave coupling (Dong et al., 27 Nov 2025). In the matter-interacting model, $a^\dagger J_++aJ_-$1 is the Tavis–Cummings limit and $a^\dagger J_++aJ_-$2 is the Dicke limit (Romero et al., 2024). The symbol $a^\dagger J_++aJ_-$3 is not universal: it denotes the ratio $a^\dagger J_++aJ_-$4 in the quantum-geometric analysis, but the anisotropy factor multiplying counter-rotating terms in the transport study (Zhu et al., 2024, Junran et al., 31 Mar 2026). This suggests that the model is most robustly identified by the structural inequality of rotating-wave and counter-rotating-wave couplings, rather than by any single parameter symbol.

Convention	Definition	Special values
$a^\dagger J_++aJ_-$5	independent RW and CRW couplings	$a^\dagger J_++aJ_-$6: Dicke; $a^\dagger J_++aJ_-$7 or $a^\dagger J_++aJ_-$8: integrable
$a^\dagger J_++aJ_-$9	RW $U(1)$0, CRW $U(1)$1	$U(1)$2: isotropic
$U(1)$3	$U(1)$4, $U(1)$5	$U(1)$6: RW only; $U(1)$7: isotropic; $U(1)$8: CRW only
$U(1)$9	coefficient of the CRW channel in the matter-interacting model	$\mathbb Z_2$0: TC; $\mathbb Z_2$1: Dicke

A recurrent benchmark for anisotropic studies is the standard isotropic Dicke model. The finite-$\mathbb Z_2$2 catastrophe-theory, symmetry-adapted-state, and fidelity analyses developed for that case are methodologically important, but that benchmark uses a single coupling multiplying co- and counter-rotating terms equally, and no anisotropy parameter appears (Nahmad-Achar et al., 2012).

2. Symmetry, integrable limits, and the basic phase structure

The generic anisotropic Dicke model has a parity symmetry

$\mathbb Z_2$3

and many finite-size studies restrict the analysis to the positive-parity sector, which contains the ground state in the parameter ranges considered (Buijsman et al., 2016, Hu et al., 2020, Das et al., 2023). In the generalized Dicke–Ising realization, the same discrete symmetry is written as

$\mathbb Z_2$4

while at $\mathbb Z_2$5 the model acquires an enhanced $\mathbb Z_2$6 symmetry with conserved total excitation number (Dong et al., 27 Nov 2025). In the generalized Dicke model with couplings $\mathbb Z_2$7 and $\mathbb Z_2$8, the same distinction is phrased as a crossover between a generic $\mathbb Z_2$9 theory and the $$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$0 Tavis–Cummings and anti-Tavis–Cummings limits (Shapiro et al., 2019).

In the thermodynamic limit, the normal-to-superradiant quantum phase transition occurs at

$$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$1

with the normal phase for $$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$2 and the superradiant phase for $$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$3 (Buijsman et al., 2016, Hu et al., 2020, Das et al., 2023). For the common resonant choice $$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$4, the critical line becomes $$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$5, and along the Dicke line $$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$6 the critical point is $$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$7 (Hu et al., 2020). The order parameter is usually the boson occupation: in the normal phase $$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$8, whereas in the superradiant phase $$H=\omega a^\dagger a+\omega_0 J_z+\frac{g_1}{\sqrt{2j}}\left(a^\dagger J_-+aJ_+\right)+\frac{g_2}{\sqrt{2j}}\left(a^\dagger J_++aJ_-\right),$$9 (Buijsman et al., 2016, Hu et al., 2020).

The integrable limits are structurally important. When $a,a^\dagger$0, the model reduces to the rotating-wave approximation, conserving the total excitation number $a,a^\dagger$1; when $a,a^\dagger$2, it maps to the $a,a^\dagger$3 case by a spin rotation and $a,a^\dagger$4 (Buijsman et al., 2016). These limits organize both the model’s symmetry classification and its later ergodic and transport behavior.

3. Critical phenomena, finite temperature, and finite-size structure

The ground-state quantum phase transition retains the Dicke-type normal-to-superradiant structure, but the anisotropic model supports additional critical phenomena. In the finite-size spectral analysis, the ground state is localized in the normal phase with inverse participation ratio near unity, while in the superradiant phase it becomes multifractal with $a,a^\dagger$5, so the transition is described as a localized-to-multifractal transition of the ground state (Das et al., 2023). The same work identifies an excited-state quantum phase transition through two characteristic energies, $a,a^\dagger$6 and $a,a^\dagger$7, extracted from jumps in von Neumann entanglement entropy; the states between these energies form a central band distinct from lower and upper spectral bands (Das et al., 2023).

At finite temperature and finite $a,a^\dagger$8, the sharp superradiant transition broadens into a fluctuational critical region of width

$a,a^\dagger$9

and the internal structure of that region depends on whether the model is effectively in a $J_z,J_\pm$0 or $J_z,J_\pm$1 regime (Shapiro et al., 2019). In the Tavis–Cummings-like sector, the universal relative fluctuation is

$J_z,J_\pm$2

whereas in the generalized Dicke regime it is

$J_z,J_\pm$3

The same study finds distinct temperature scalings of momentum squeezing,

$J_z,J_\pm$4

and identifies a minimal temperature scale $J_z,J_\pm$5 below which the zero-mode finite-temperature description breaks down and the zero-temperature quantum-critical regime takes over (Shapiro et al., 2019). A complementary finite-temperature treatment gives the thermal critical line

$J_z,J_\pm$6

valid in the superradiant regime, with mutual information serving as a numerical indicator of the thermal transition (Das et al., 2023).

A major generalization is the anisotropic Dicke–Stark model with a nonlinear Stark term and an explicit $J_z,J_\pm$7-type term,

$J_z,J_\pm$8

where $J_z,J_\pm$9 contains unequal rotating-wave and counter-rotating-wave couplings (Chen et al., 2024). In that setting, the standard isotropic Dicke model remains subject to the no-go theorem induced by the $\omega$0 term, but sufficiently strong anisotropy and/or Stark coupling can restore equilibrium superradiant phase transitions at both zero and finite temperature. The zero-temperature critical coupling is

$\omega$1

which requires $\omega$2 and $\omega$3; in particular, a finite critical coupling exists for $\omega$4 under the TRK constraint (Chen et al., 2024). The model is reported to remain in the same universality class as the conventional Dicke model, with

$\omega$5

and finite-size scaling controlled by $\omega$6 (Chen et al., 2024).

4. Spectral statistics, ergodicity, and dynamical diagnostics

A central question in the anisotropic Dicke model is whether non-ergodic behavior is tied to the normal-to-superradiant transition. Level statistics, the average ratio of consecutive level spacings, and OTOC-like probes show that the ergodic-to-non-ergodic transition is instead controlled by proximity to the integrable axes $\omega$7 and $\omega$8 (Buijsman et al., 2016). The relevant spacing benchmarks are Poisson,

$\omega$9

for non-ergodic behavior, and Wigner–Dyson/GOE,

$\omega_0$0

for ergodic behavior (Buijsman et al., 2016). That work explicitly argues that there is no intrinsic relation between the ergodic–non-ergodic transition and the precursors of the normal–superradiant quantum phase transition (Buijsman et al., 2016).

A later spectral study refines this picture by showing that the excited-state quantum phase transition and the ergodic-to-non-ergodic transition are closely related. In the central spectral band between $\omega_0$1 and $\omega_0$2, the level-spacing ratio and the participation ratio reveal a crossover from non-ergodic extended or multifractal behavior to ergodic delocalized behavior as the couplings increase (Das et al., 2023). Read together, these results suggest a distinction between two statements: the ergodic transition is not intrinsically linked to the ground-state normal-to-superradiant boundary, but it is closely organized by excited-state structure in the superradiant regime.

The out-of-time-ordered correlator provides a dynamical diagnostic of both chaos and ground-state criticality. For the anisotropic Dicke model at zero temperature,

$\omega_0$3

and the residue OTOC

$\omega_0$4

distinguishes the normal and superradiant phases: its long-time value is smaller in the normal phase and larger in the superradiant phase, and scanning it over the $\omega_0$5 plane reconstructs the phase boundary $\omega_0$6 for $\omega_0$7 (Hu et al., 2020). Finite-size analysis shows that the OTOC signature sharpens with increasing $\omega_0$8, whereas increasing temperature progressively blurs the boundary (Hu et al., 2020).

The model also supports genuinely dynamical reclassifications of the superradiant regime. One thermodynamic-limit effective description argues that the conventional superradiant phase is split by hidden exceptional points into three hierarchic subphases, $\omega_0$9, $a^\dagger J_-+aJ_+$00, and $a^\dagger J_-+aJ_+$01, characterized by harmonic-oscillator, anti-harmonic-oscillator, and inverted-harmonic-oscillator sectors in the effective Hamiltonian (He et al., 27 May 2025). In that picture, the Loschmidt echo of the trivial initial state distinguishes fully oscillatory, mixed oscillatory-decaying, and fully decaying quench dynamics, and finite-$a^\dagger J_-+aJ_+$02 exact diagonalization for $a^\dagger J_-+aJ_+$03 is reported to confirm the predicted structure (He et al., 27 May 2025).

5. Collective modes, quantum geometry, and critical metrology

The anisotropic Dicke model has become a natural setting for quantum-geometric analyses because the normal-to-superradiant transition closes an excitation gap and makes the ground state highly parameter-sensitive. The quantum geometric tensor is

$a^\dagger J_-+aJ_+$04

with quantum metric $a^\dagger J_-+aJ_+$05 and Berry curvature $a^\dagger J_-+aJ_+$06 (Zhu et al., 2024). For a nondegenerate eigenstate,

$a^\dagger J_-+aJ_+$07

so singular response near criticality follows directly from gap closing (Zhu et al., 2024). The same work defines

$a^\dagger J_-+aJ_+$08

with $a^\dagger J_-+aJ_+$09, and distinguishes two classical limits. In the classical spin limit $a^\dagger J_-+aJ_+$10, the anisotropy persists strongly and the rotating-wave coupling is more favorable for driving the transition; in the classical oscillator limit $a^\dagger J_-+aJ_+$11, the contributions of the rotating and counter-rotating terms become symmetric in the projected low-energy theory (Zhu et al., 2024). The finite-scale behavior resembles the classical-spin case, and the interplay among anisotropy ratio, spin length, and frequency ratio is reported to enhance the critical response without trade-off (Zhu et al., 2024).

The low-lying collective spectrum becomes richer when collective matter interactions are included. In the anisotropic Dicke model with

$a^\dagger J_-+aJ_+$12

the coherent-state energy surface and the Holstein–Primakoff expansion reveal phase and amplitude polariton modes, denoted $a^\dagger J_-+aJ_+$13 and $a^\dagger J_-+aJ_+$14, together with shifted critical couplings $a^\dagger J_-+aJ_+$15 and $a^\dagger J_-+aJ_+$16 (Romero et al., 2024). Without matter interactions, the Tavis–Cummings limit $a^\dagger J_-+aJ_+$17 has a Goldstone phase mode in the superradiant phase, whereas the Dicke limit $a^\dagger J_-+aJ_+$18 has a gapped, roton-like phase mode (Romero et al., 2024). With matter interactions, the normal phase can be deformed, the critical couplings shift differently in the $a^\dagger J_-+aJ_+$19- and $a^\dagger J_-+aJ_+$20-interaction sectors, and the phase mode can be suppressed in a finite coupling window if the two critical lines are sufficiently separated (Romero et al., 2024). The same study derives geometric phases for photon-number and spin contours,

$a^\dagger J_-+aJ_+$21

which are singular at the superradiant transition and can also detect a first-order transition induced when $a^\dagger J_-+aJ_+$22 (Romero et al., 2024).

6. Implementations, driven extensions, and nonequilibrium transport

Rydberg-cavity platforms have become a prominent route to realizing anisotropic Dicke physics. In a generalized Dicke–Ising model for driven Rydberg arrays coupled to an optical cavity, the normalized anisotropy parameter $a^\dagger J_-+aJ_+$23 tunes the balance between rotating-wave and counter-rotating-wave couplings, while nearest-neighbor Rydberg blockade introduces an Ising interaction (Dong et al., 27 Nov 2025). The phase diagram contains a normal phase, a superradiant phase, a checkerboard Solid-$a^\dagger J_-+aJ_+$24 phase, and a superradiant solid in which superradiance coexists with translational symmetry breaking (Dong et al., 27 Nov 2025). In that system, the normal-to-superradiant transition and the Solid-$a^\dagger J_-+aJ_+$25-to-SRS transition are second-order, whereas the Solid-$a^\dagger J_-+aJ_+$26-to-SR and SRS-to-SR transitions are first-order for any normalized anisotropy (Dong et al., 27 Nov 2025). Mean-field phase boundaries are independent of $a^\dagger J_-+aJ_+$27, but finite-size parity behavior, photon-number scaling, and the slight favoring of the SRS phase over the Solid-$a^\dagger J_-+aJ_+$28 state by stronger counter-rotating-wave fluctuations are explicitly anisotropy-dependent (Dong et al., 27 Nov 2025).

A complementary proposal uses periodic modulation of a pumping laser in a cavity-coupled Rydberg array to Floquet-engineer an anisotropic Dicke model with dipole-dipole interactions,

$a^\dagger J_-+aJ_+$29

where the effective couplings are renormalized by Bessel functions (Dong et al., 18 Mar 2025). The key result is tunability of the counter-rotating to rotating ratio from zero to infinity, enabling access to the pure rotating-wave limit, intermediate anisotropic regimes, and the pure counter-rotating-wave limit (Dong et al., 18 Mar 2025). In the adiabatic state-preparation example, increasing $a^\dagger J_-+aJ_+$30 converts parity-protected same-sector crossings into avoided crossings, opening a finite gap and improving the preparation of superradiant and superradiant-solid states (Dong et al., 18 Mar 2025).

The anisotropic Dicke model also shows sharply different responses to periodic and quasiperiodic driving. Under periodic square-wave modulation of the couplings, the dynamics reaches a plateau not followed by heating on the studied timescales, whereas under Thue–Morse and Fibonacci quasiperiodic drives a prethermal plateau is followed by heating to the infinite-temperature state (Das et al., 2023). The heating time scales as a stretched exponential in the Thue–Morse case,

$a^\dagger J_-+aJ_+$31

and exponentially in the Fibonacci case,

$a^\dagger J_-+aJ_+$32

(Das et al., 2023). The periodic drive also shifts the quantum critical boundary and extends the normal phase to larger couplings (Das et al., 2023).

Nonequilibrium heat transport adds another layer of anisotropic physics. In a reduced anisotropic Dicke system coupled to separate bosonic thermal reservoirs for the qubit and photon sectors, a dressed master equation is used to treat strong qubit–photon coupling in the eigenbasis of the interacting Hamiltonian (Junran et al., 31 Mar 2026). The central transport result is nonmonotonic in anisotropy: at weak coupling the current depends only weakly on the anisotropy factor, at moderate coupling increasing anisotropy enhances the heat current, and at strong coupling large anisotropy strongly suppresses it through multiphoton dressing (Junran et al., 31 Mar 2026). Increasing the number of qubits amplifies both the peaks and the valleys of the heat-flow characteristics, and the limiting cases $a^\dagger J_-+aJ_+$33 and $a^\dagger J_-+aJ_+$34 admit analytical thermodynamic-limit heat currents with a cotunneling transport interpretation that serve as upper boundaries for finite-size currents (Junran et al., 31 Mar 2026). The same study reports that increasing temperature bias and anisotropy enhances thermal rectification, with favorable regimes yielding rectification around $a^\dagger J_-+aJ_+$35–$a^\dagger J_-+aJ_+$36 (Junran et al., 31 Mar 2026).

A distinct but related branch of the subject is the dissipative anisotropic two-photon Dicke model, in which anisotropy and cavity loss generate pole-flip transitions, localized fixed points associated with spectral collapse, Hopf bifurcations, period-doubling cascades, chaotic dynamics, and strong phase coexistence (Li et al., 2022). Although this is not the one-photon anisotropic Dicke model, it shows how anisotropic weighting of rotating and counter-rotating channels systematically enlarges the dynamical phase portrait of Dicke-type light–matter systems (Li et al., 2022).