Magnetic graphs for cavity quantum electrodynamics

Technical Summary: Magnetic Graphs for Cavity Quantum Electrodynamics

Problem Statement
The theoretical framework for single-atom cavity quantum electrodynamics (QED), specifically the Quantum Rabi Model (QRM), faces challenges in the ultrastrong coupling (USC; $0.1 \lesssim \eta \lesssim 1$) and deep-strong coupling (DSC; $\eta \gtrsim 1$) regimes. Traditional classifications based on the dimensionless coupling strength $\eta = g/\omega_0$ rely on specific approximations, such as the Rotating-Wave Approximation (RWA), which fail in these regimes. Furthermore, gauge-invariant formulations (e.g., Coulomb gauge) introduce higher-order field operators and long-range hopping terms that obscure simple physical interpretations. Existing descriptions struggle to unify the dynamics across weak, strong, USC, and DSC regimes without invoking regime-specific approximations or complex adiabatic pictures.

Methodology
The authors propose a "magnetic graph" model to reinterpret the generalized QRM. The methodology proceeds through the following steps:

1. Generalized Hamiltonian Formulation: Starting with a two-level atom coupled to a quantized cavity field, the authors employ a gauge-invariant QRM Hamiltonian in the Coulomb gauge. This formulation accommodates vector gauge fields, broken parity, and broken time-reversal symmetries, leading to a Hamiltonian containing displacement operators $D(\pm 2i\eta)$ that induce long-range hopping between Fock states.
2. Graph Representation: The system is mapped onto a Fock-state lattice. The Schrödinger equation is transformed into a set of coupled differential equations where the hopping terms $l_{nm}^{\pm}$ (derived from associated Laguerre polynomials) represent weighted edges between nodes.
3. Floquet Rabi (FR) Graph Construction: By applying a local gauge transformation to flatten the linear variation of onsite energies, the authors construct a "Floquet Rabi graph." This graph is a complex, weighted, bipartite magnetic graph with identical sites under Floquet boundary conditions. The edges carry complex phases determined by the hopping order.
4. Spectral Analysis via Magnetic Laplacian: To quantify graph connectivity, the authors utilize the normalized magnetic Laplacian. They focus on the first eigenvalue ($\lambda_1$) of this operator. According to the generalized Cheeger's inequality, $\lambda_1$ provides a lower bound on the cost of separating a large, phase-coherent (magnetic-flux-free) subgraph from the rest of the graph.
5. Decomposition of Connectivity: To determine the physical origin of changes in $\lambda_1$, the authors disentangle two potential mechanisms: (i) the alleviation of edge-weight bottlenecks (conductance) and (ii) phase frustration induced by non-trivial magnetic fluxes in closed loops. They analyze random subgraphs to compute conductance ($C_q$) and loop weights ($W_q$), distinguishing between trivial ($\Phi_q = 0$) and non-trivial ($\Phi_q = \pi$) phases.

Key Results

• Unified Classification Metric: The study demonstrates that the coupling regimes can be classified continuously using the single metric $\lambda_1$. In the weak-coupling regime, $\lambda_1$ follows a linear law ($\lambda_1 \approx 2\eta$). As coupling increases into the USC and DSC regimes, $\lambda_1$ deviates from linearity and eventually saturates at unity. This saturation in the DSC regime signifies a high cost to isolate a nonmagnetic subgraph.
• Mechanism of Localization: The analysis reveals that the increase in $\lambda_1$ is not driven by the removal of bottlenecks (conductance remains nearly constant with respect to $\eta$) but solely by phase frustration. As $\eta$ increases, the system supports higher-order closed loops. Approximately half of these loops possess non-trivial phases ($\Phi_q = \pi$), leading to destructive interference.
• State Localization and Spectral Degeneracy: The phase frustration induced by non-trivial loops drives the localization of quantum eigenstates. This is evidenced by an increase in the Inverse Participation Ratio (IPR) and the emergence of spectral degeneracies in the DSC regime.
• Topological Nature: The phase of the loop weights is shown to be a topological invariant (quantized to 0 or $\pi$) related to the parity conversion of number states along the loop. This frustration becomes uniform across the graph in the deep-strong coupling limit.

Significance and Claims
The paper claims to provide a new framework for interpreting atom-light interactions by bridging graph theory and cavity QED. Key claims include:

• Regime-Agnostic Description: The magnetic graph model offers a continuous, single-metric classification of QED regimes that does not rely on the validity of specific approximations (like RWA) or regime-specific adiabatic pictures.
• Topological Driver: The study establishes that "phase frustration" induced by subgraph topology is the primary driver of the crossover from weak to deep-strong coupling, explaining the localization of quantum states.
• Scalability: The authors posit that this approach is scalable to many-body systems (e.g., Dicke and Rabi-Hubbard models), where long-range hopping in Fock-state lattices can similarly be transformed into large-scale magnetic graphs.
• Analytical Utility: The framework suggests that standard graph-theoretic tools, such as spectral clustering and graph neural networks, could be applied to analyze complex quantum dynamics in hybrid systems.

The authors maintain a modest tone regarding future applications, noting that while the framework is general, specific applications like graph neural networks are identified as promising directions for future research rather than immediate outcomes of the current work.