Ultrastrong Light-Matter Coupling (USC)

• Ultrastrong light–matter coupling (USC) is the regime where the interaction strength becomes a significant fraction of the bare mode and transition frequencies, leading to dressed light–matter eigenstates.
• USC challenges standard approximations like the rotating-wave approximation, necessitating new formulations in spectroscopy, dissipation, and state control.
• Diverse experimental platforms, from superconducting circuits to plasmonic and metamaterial systems, demonstrate USC’s potential in advancing quantum simulation and information processing.

Searching arXiv for authoritative USC papers and reviews to ground the article. arXiv search query: "ultrastrong light matter coupling review quantum Rabi model superconducting circuits waveguide QED" Ultrastrong light–matter coupling (USC) is the regime of cavity or waveguide quantum electrodynamics in which the interaction strength is no longer a small perturbation on the uncoupled light and matter dynamics, but instead becomes a sizable fraction of the bare mode and transition frequencies. In cavity settings the standard figure of merit is the normalized coupling $\eta \equiv g/\omega$, with $\eta \gtrsim 0.1$ conventionally marking USC and $\eta>1$ deep-strong coupling (DSC), although the $0.1$ threshold is conventional rather than fundamental; in open continua, an alternative control parameter is the radiative rate relative to the transition frequency, such as $\Gamma_1/\Delta$ (Kockum et al., 2018, Forn-Díaz et al., 2018, Forn-Díaz et al., 2016). What distinguishes USC is the breakdown of the rotating-wave approximation, the reorganization of both ground and excited states into dressed light–matter eigenstates, and the consequent need to reformulate spectroscopy, dissipation, and input-output theory in the dressed basis.

1. Defining the regime

USC is distinct from ordinary strong coupling. Weak versus strong coupling compares $g$ to losses such as $\kappa$ and $\gamma$: in weak coupling, excitation exchange is interrupted by dissipation, whereas in strong coupling, $g$ exceeds relevant linewidths and coherent vacuum Rabi oscillations occur. USC instead compares $g$ to the bare subsystem energies $\eta \gtrsim 0.1$0 and $\eta \gtrsim 0.1$1. A system can therefore be ultrastrongly coupled without being strongly coupled if losses are sufficiently large (Kockum et al., 2018).

Within the single-mode cavity setting, the regime boundaries commonly used in the literature are $\eta \gtrsim 0.1$2 for USC and $\eta \gtrsim 0.1$3 for DSC (Forn-Díaz et al., 2018). A useful refinement separates perturbative USC, $\eta \gtrsim 0.1$4, from nonperturbative USC, $\eta \gtrsim 0.1$5, reflecting whether counter-rotating effects can still be treated perturbatively or require the exact quantum Rabi structure (Forn-Díaz et al., 2018). The review literature also stresses that higher-order effects grow continuously with $\eta \gtrsim 0.1$6, so the boundary at $\eta \gtrsim 0.1$7 should be understood as a convention, not a sharp phase line (Kockum et al., 2018).

In open waveguide and continuum settings, the relevant dimensionless parameter is often not $\eta \gtrsim 0.1$8 but the spontaneous emission rate relative to the transition frequency. A superconducting flux qubit coupled directly to a one-dimensional transmission line was analyzed in terms of $\eta \gtrsim 0.1$9, with USC identified by $\eta>1$0 and nonperturbative USC by $\eta>1$1 or larger (Forn-Díaz et al., 2016). This continuum formulation is conceptually important because it shows that USC is not restricted to discrete cavity polaritons.

A broader reinterpretation has also emerged in materials physics. A 2025 survey of more than 70 materials argued that phonon-, exciton-, and plasmon-polaritons in bulk solids often realize USC or DSC intrinsically, with the operative reduced coupling written as $\eta>1$2, where $\eta>1$3 is the collective transverse mode frequency (Mueller et al., 9 May 2025). This shifts USC from a purely engineered-cavity notion toward a general regime of nonperturbative electrodynamics.

2. Canonical Hamiltonians and symmetry structure

The minimal single-emitter model of USC is the quantum Rabi model,

$\eta>1$4

which contains both excitation-conserving and counter-rotating processes (Kockum et al., 2018). In the weak-coupling, near-resonant limit, the counter-rotating terms oscillate rapidly and one recovers the Jaynes–Cummings model,

$\eta>1$5

The structural difference is decisive: the Jaynes–Cummings Hamiltonian conserves the total excitation number $\eta>1$6, whereas the quantum Rabi Hamiltonian does not (Kockum et al., 2018).

What remains conserved in the standard quantum Rabi model is parity,

$\eta>1$7

so the Hilbert space decomposes into even and odd parity sectors rather than excitation-number manifolds (Kockum et al., 2018). This parity structure governs selection rules, spectroscopy, and state engineering in USC. In generalized Rabi models relevant to superconducting circuits, the interaction may contain both longitudinal and transverse components, and parity can itself be broken, opening additional transition pathways (Kockum et al., 2018).

For many-emitter or collective matter systems, the relevant descriptions are often the Dicke and Hopfield models. The Dicke model captures $\eta>1$8 two-level systems with the familiar $\eta>1$9 collective enhancement, whereas the Hopfield model describes a bosonic matter mode $0.1$0 coupled to a cavity mode $0.1$1, with both resonant and antiresonant couplings,

$0.1$2

$0.1$3

In USC, the antiresonant $0.1$4 contribution is no longer negligible (Kockum et al., 2018).

A central subtlety is the diamagnetic or self-polarization term,

$0.1$5

with $0.1$6 for a dominant transition of frequency $0.1$7 (Kockum et al., 2018). This term is parametrically non-negligible in USC, enforces stability and gauge consistency, and sits at the center of the no-go arguments for equilibrium superradiant phase transitions when gauge invariance is treated properly (Kockum et al., 2018, Forn-Díaz et al., 2018).

In continuum USC, the effective model is instead spin-bosonic: $0.1$8 with Ohmic spectral density

$0.1$9

This mapping makes explicit that continuum USC is simultaneously a waveguide-QED and dissipative-quantum-impurity problem (Forn-Díaz et al., 2016).

3. Ground state, virtual excitations, and open-system observables

One of the defining consequences of USC is that the interacting ground state is no longer the bare vacuum $\Gamma_1/\Delta$0. In the quantum Rabi model it takes the dressed form

$\Gamma_1/\Delta$1

so the ground state contains virtual light–matter excitations bound into the vacuum of the full Hamiltonian (Kockum et al., 2018). In the DSC limit, the ground state can be approximated by a cat-like entangled form,

$\Gamma_1/\Delta$2

which makes the vacuum dressing visually explicit (Kockum et al., 2018).

These excitations are virtual, not directly radiated photons. The bare number operator can satisfy $\Gamma_1/\Delta$3, yet a correct photodetector still sees no emission from the ground state. The reason is that output photons in USC are associated with dressed positive-frequency operators, not with the bare annihilation operator. A standard construction is

$\Gamma_1/\Delta$4

which guarantees $\Gamma_1/\Delta$5 (Kockum et al., 2018).

This distinction forces a revision of open-system theory. Bare-basis master equations can predict unphysical spontaneous radiation from the USC ground state at zero temperature. The correct treatment uses dressed-state master equations or nonperturbative bosonic input-output theory, depending on the setting (Kockum et al., 2018, Forn-Díaz et al., 2018). The same point extends to single-shot monitoring: in quantum-trajectory formulations of USC, jump operators must be defined from dressed positive-frequency components rather than from bare $\Gamma_1/\Delta$6 and $\Gamma_1/\Delta$7, and this is what reveals higher-order USC processes such as the conversion of one photon into two atomic excitations (Macrì et al., 2021).

The virtual character of USC dressing also motivates dedicated detection protocols. A notable proposal is to use STIRAP in flux-based superconducting architectures to coherently amplify the conversion of virtual photon pairs in USC dressed states into a final state with two real cavity photons. In that setting the $\Gamma_1/\Delta$8-configuration was identified as the selective route because the corresponding $\Gamma_1/\Delta$9-scheme is contaminated by stray Jaynes–Cummings-like channels (Falci et al., 2017). This makes explicit that virtual excitations are physically consequential, but only when the Hamiltonian or level structure is modulated in a USC-aware manner.

4. Experimental realizations and platform diversity

The experimental USC landscape is now broad. Review articles identify the first realization of USC in intersubband polaritons with $g$0, followed by superconducting-circuit demonstrations with $g$1. Subsequent work pushed intersubband platforms to about $g$2, superconducting circuits into DSC with $g$3, and Landau polaritons to $g$4. Organic molecular systems reached $g$5, and vibrational or optomechanical settings reported $g$6 (Kockum et al., 2018). These platforms differ in microscopic detail, but all exhibit the same defining feature: the coupling is comparable to the relevant bare frequencies.

Open electromagnetic continua have provided a distinct route. In a superconducting flux qubit galvanically attached to a transmission line, the experimentally extracted coupling reached $g$7 in a fixed device and $g$8 in a tunable device, with clear evidence of bath-induced frequency renormalization consistent with the Ohmic spin-boson model (Forn-Díaz et al., 2016). A related free-space formulation was developed for multisubband plasmons in a dense two-dimensional electron gas, where the ultrastrong regime is identified by radiative broadening comparable to the matter frequency, $g$9, and where both the rotating-wave and Markov approximations become unphysical (Huppert et al., 2016). Together, these results established that optical confinement is not a prerequisite for USC.

Plasmonic and metamaterial implementations have pushed normalized splittings to unusually large values. In epsilon-near-zero coaxial nanocavities filled with SiO$\kappa$0, the reported vibrational USC metric was $\kappa$1, with a splitting about half the transverse-optical phonon frequency and a pronounced blue shift of the polariton center frequency (Yoo et al., 2020). In deeply subwavelength THz LC resonators containing a semiconductor 2DEG, the reported normalized coupling was $\kappa$2, with $\kappa$3 and only $\kappa$4 electrons participating in the lower-doped device (Jeannin et al., 2019).

Low-loss dielectric nanophotonics has also entered the field. A dual-gradient dielectric metasurface supporting quasi-bound states in the continuum was reported to reach $\kappa$5 in a two-SiO$\kappa$6-layer design, with a mode splitting equivalent to $\kappa$7 of the ENZ-mode energy, while the best centered single-layer structure yielded $\kappa$8 and $\kappa$9 meV splitting (Bau et al., 19 Feb 2025). The platform is notable because it approaches USC without resorting to plasmonic losses.

A further extension is the claim that USC is often intrinsic to solids. The bulk-material survey of more than 70 systems reported that phonon-, exciton-, and plasmon-polaritons in many solids systematically surpass cavity-based coupling strengths, with representative extracted values such as $\gamma$0 for gold nanoparticle supercrystals, $\gamma$1 for a squarylium crystal exciton, and $\gamma$2 for NaCl and MgO phonons (Mueller et al., 9 May 2025). This perspective does not replace cavity USC, but it materially broadens the domain in which USC concepts apply.

5. Quantum control, simulation, and information processing

USC has long been associated with the possibility of faster gates and qualitatively new control primitives. A concrete circuit-QED proposal used two flux qubits galvanically coupled to a resonator through a six-Josephson-junction design with flux-tunable sign-reversible coupling. In the nearly longitudinal regime, a four-step conditional-displacement protocol yields a $\gamma$3 entangler locally equivalent to a controlled-phase gate. For $\gamma$4 and $\gamma$5 GHz, the predicted gate time is $\gamma$6 ns with fidelity $\gamma$7; the same architecture supports on/off switching and sign reversal of the coupling, with switching times estimated around $\gamma$8 ns or less (Romero et al., 2011).

USC has also become a target for quantum simulation. One analog route uses a two-tone-driven circuit-QED system in the ordinary strong-coupling regime to synthesize an effective quantum Rabi Hamiltonian,

$\gamma$9

so that $g$0, $g$1, and $g$2. With realistic parameters, this yields $g$3, i.e. the DSC threshold, and reproduces the characteristic revival dynamics of the effective Rabi model (Ballester et al., 2011). A digital route used a transmon-resonator chip to implement up to 90 second-order Trotter steps and probe a combined Hilbert-space dimension $g$4, thereby observing Bell-cat-like entanglement and large photon-number buildup characteristic of deep USC (Langford et al., 2016).

Control and readout native to USC have likewise advanced. Ancilla-assisted parity-dependent protocols have been proposed for spectroscopy, state engineering, and tomography of USC polaritons by exploiting the parity symmetry of the quantum Rabi model (Felicetti et al., 2014). STIRAP-based protocols were proposed to convert virtual USC photon pairs into real output photons in a dynamically selective way (Falci et al., 2017). Experimentally, a gate-tunable gatemon coupled to a $g$5 resonator was reported at $g$6, with time-domain coherent control and coherence times $g$7 and $g$8, showing that standard qubit manipulations can survive in USC (Iglesias et al., 19 Mar 2026).

More recent work has focused on higher-order and multimode functionalities. A parity-broken superconducting-circuit model predicted two-polariton blockade under resonant single-polariton driving at $g$9, a phenomenon absent in weak and ordinary strong coupling (Ma et al., 26 Jan 2026). In Landau polaritons, a multimode slot-cavity experiment extracted $g$0 for cyclotron resonance while simultaneously coupling to finite-momentum magnetoplasmons, thereby demonstrating cavity-mediated interaction between local and nonlocal matter modes in a mixed USC/SC regime (Endo et al., 6 Sep 2025). These developments indicate that USC is increasingly used not merely as a spectroscopic label, but as a resource for engineered nonequilibrium dynamics.

6. Conceptual issues, controversies, and developing directions

Several conceptual issues remain central. Gauge consistency and the role of the $g$1 or $g$2 term are not technical afterthoughts but structural requirements of a consistent USC theory. The status of superradiant phase transitions in equilibrium models remains subtle and model-dependent once gauge invariance is enforced (Kockum et al., 2018, Forn-Díaz et al., 2018). Closely related to this are truncation issues: few-level matter descriptions and single-mode cavity models can fail in USC because additional atomic levels or cavity modes may enter quantitatively or even qualitatively.

The meaning of “photons” in USC also remains delicate. Bare cavity photons are not the physical output quanta, and ground-state photon populations are not directly extractable without auxiliary dynamics. This is why dressed-basis dissipation, dressed positive-frequency field operators, and USC-specific input-output theory are indispensable (Kockum et al., 2018). The same logic applies to measurement theory: trajectory-level observables can expose conditional dynamics that are invisible in ensemble-averaged master equations (Macrì et al., 2021).

Current directions extend USC into materials and photonic architectures that are not naturally described as single cavity plus single emitter. A theoretical framework for “quantum electrodynamical metamaterials” treats each meta-atom as a quantum Rabi unit cell and predicts optical behavior ranging from Lorentz-oscillator-like to effectively transparent as $g$3 is increased from weak to deep strong coupling (Yu et al., 2022). In parallel, the bulk-material viewpoint argues that intrinsic USC may underlie soft modes, radiative-decay suppression, and even collective ground-state instabilities associated with ferroelectricity, insulator-to-metal transitions, and exciton condensation (Mueller et al., 9 May 2025). These claims are broader than the established cavity-QED paradigm, but they indicate the growing convergence of USC with polaritonic materials science.

A persistent frontier is the direct observation of uniquely USC phenomena rather than large spectroscopic splittings alone. Review treatments repeatedly identify virtual-excitation extraction, nonadiabatic vacuum emission, deterministic higher-order processes, and properly dressed open-system measurements as the decisive next steps (Kockum et al., 2018). Taken together, the literature presents USC not simply as an incremental increase of coupling strength, but as a regime in which the operative notions of light, matter, excitation, and even vacuum must be reformulated in terms of strongly hybridized quantum degrees of freedom.