Evidence for cavity-induced metallic phase from entropic electron correlation effects

Abstract: Experiments have revealed that collective strong coupling in optical cavities can drastically alter the conductivity and dielectric properties of molecular ensembles, or even trigger abrupt phase transitions (in Rayleigh scattering or dispersion force driven conformational equilibrium). Until now, the underlying physical mechanism has remained unknown. A recently proposed theory attributes these collective effects to cavity-induced electron correlations that share properties of the known spherical Sherrington-Kirkpatrick model of a spin glass. As a consequence, entropic effects can trigger a phase transition in the inter-molecular electron correlations that could potentially explain the above experiments. In the following work, we derive analytic expressions for the correlation (free-)energy-induced polarizability changes, which reveal that under certain conditions collective strong coupling can even render our molecular ensemble metallic by entropically ionizing collectively degenerate molecular orbitals. Our results are a first major step towards a holistic theory of cavity-mediated electron correlation effects.

• The paper derives non-perturbative, temperature-dependent polarizability formulas by mapping cavity-induced electron correlations onto the spherical Sherrington–Kirkpatrick spin-glass model.
• For fractionally filled, collectively degenerate states, the correlation free-energy gain can overcome Hartree–Fock excitation costs, producing divergent local polarizability and an effectively metallic phase without illumination.
• The theory offers a mechanism for reported conductivity, dielectric, and Rayleigh-scattering enhancements under vibrational strong coupling, while emphasizing that static polarizability cannot distinguish para-correlated from spin-glass phases.

Context and motivation

Experiments in polaritonic chemistry have reported that collective strong coupling (SC) of molecular ensembles to optical cavity modes can modify conductivity, dielectric response, and even trigger abrupt phase transitions — all without external illumination. The conventional polariton picture struggles here: vibrational strong coupling (VSC) cavities are tuned to IR modes that are scarcely thermally populated at room temperature, and the large-$N$ problem implies vanishing local light–matter couplings under collective scaling $\lambda \sim 1/\sqrt{N}$ (2608.16593). A recent theoretical development maps the transverse electron correlation problem of a collectively coupled ensemble onto the spherical Sherrington-Kirkpatrick (SSK) spin glass model, predicting three phases: uncorrelated, para-correlated, and spin-glass correlated, with correlations entropically stabilized by the cavity-induced degeneracy. This paper extends that framework by deriving analytic, non-perturbative expressions for the static polarizability induced by these glassy electron correlations, and shows that under fractional filling the molecular ensemble can become metallic through entropic ionization.

Theoretical framework

Starting from the Pauli-Fierz Hamiltonian within the cavity Born-Oppenheimer partitioning, the authors use a single Hartree-Fock reference $|\Phi_0\rangle$ with a full configuration interaction expansion restricted to the collectively degenerate subspace. Under a uniform probability assumption over FCI states, the transversal correlation energy is approximated at configuration interaction singles (CIS) level:

$$E_{\rm corr} \approx \langle \hat{N}_v\rangle_{\rm uniform}\, E_{\rm CIS,corr},$$

with uniform occupation number $$\langle\hat{N}_v\rangle_{\rm uniform} = N_o N_u / N_o^{\rm tot}$$, where $N_o$, $N_u$, and $N_o^{\rm tot}$ denote occupied, unoccupied, and total degenerate orbital counts. The CIS amplitude minimization problem is then equivalent to the SSK model,

$$E_{\rm corr} = -\sum_{i<j}^{N_{\rm deg}} J_{ij}\,\sigma_i\sigma_j, \qquad \sum_i \sigma_i^2 = 1,$$

where continuous spin variables $\sigma_i$ derive from normalized CIS amplitudes, and random couplings $\lambda \sim 1/\sqrt{N}$0 arise from two-electron integrals of the dipole self-energy combined with random molecular orientations relative to the cavity polarization. Assuming extensive variance $\lambda \sim 1/\sqrt{N}$1, the analytic SSK free energy per spin applies, with critical temperature $\lambda \sim 1/\sqrt{N}$2 separating the para-correlated from the spin-glass phase.

Two microscopic routes to the required extensive scaling are identified: an insulating case where $\lambda \sim 1/\sqrt{N}$3 (nearly empty or fully occupied degenerate state), and a metallic case where $\lambda \sim 1/\sqrt{N}$4 due to half-filling, implying high mobility of occupied orbitals in the degenerate subspace.

Non-perturbative polarizability derivation

The authors apply a static external field $\lambda \sim 1/\sqrt{N}$5 along the cavity polarization and solve the perturbed Hamiltonian non-perturbatively on the degenerate subspace — standard perturbation theory breaks down because of the infinite degeneracy in the thermodynamic limit. Using Slater-Condon rules (detailed in the appendix), the field-coupled problem retains the SSK structure with modified couplings $\lambda \sim 1/\sqrt{N}$6, where $\lambda \sim 1/\sqrt{N}$7 stems from one-electron transition dipole integrals. Assuming $\lambda \sim 1/\sqrt{N}$8 and $\lambda \sim 1/\sqrt{N}$9 uncorrelated, the effective disorder becomes

$|\Phi_0\rangle$0

with $|\Phi_0\rangle$1 relating intrinsic to field-induced fluctuations. Differentiating the free energy twice yields the transverse polarizability per spin:

$|\Phi_0\rangle$2

Insulating versus metallic regimes

Insulating case: with $|\Phi_0\rangle$3, the authors must impose microscopically correlated fluctuations among the transition dipole products to achieve extensive variance — a characterization they explicitly defer to future work. With collective coupling scaling, one finds $|\Phi_0\rangle$4, giving a finite local polarizability that saturates at $|\Phi_0\rangle$5 below $|\Phi_0\rangle$6. All electrons remain localized; the correlated ensemble stays insulating. Notably, no generic condition is currently known for whether the insulating correlated phase can be entered at all ($|\Phi_0\rangle$7).

Metallic case: for fractional filling, $|\Phi_0\rangle$8 diverges, so each of the $|\Phi_0\rangle$9 entropically excited orbitals carries the local free energy $$E_{\rm corr} \approx \langle \hat{N}_v\rangle_{\rm uniform}\, E_{\rm CIS,corr},$$0. Here $$E_{\rm corr} \approx \langle \hat{N}_v\rangle_{\rm uniform}\, E_{\rm CIS,corr},$$1 while $$E_{\rm corr} \approx \langle \hat{N}_v\rangle_{\rm uniform}\, E_{\rm CIS,corr},$$2, and the polarizability becomes

$$E_{\rm corr} \approx \langle \hat{N}_v\rangle_{\rm uniform}\, E_{\rm CIS,corr},$$3

where $$E_{\rm corr} \approx \langle \hat{N}_v\rangle_{\rm uniform}\, E_{\rm CIS,corr},$$4 is the Hartree-Fock excitation cost into the degenerate state. A locally divergent polarizability is the hallmark of metallic behavior: the initially insulating ensemble develops effectively free electrons when the correlation free-energy gain exceeds the excitation cost. Both the para-correlated and spin-glass phases exhibit this divergence, meaning static polarizability measurements cannot distinguish them — resolving this requires time-resolved probes sensitive to spin-glass aging effects.

Connection to experiments

The metallic scenario offers a candidate mechanism for several otherwise unexplained observations: the six-order-of-magnitude conductivity enhancement in nonconducting polymers (polystyrene, deuterated polystyrene, poly(benzyl methacrylate)) under VSC (2608.16593); dielectric constants of liquid water under VSC exceeding those of ice, which are highly sensitive to dissolved ions and thus potentially to entropic ionization; and abrupt two-order-of-magnitude increases in Rayleigh scattering intensity at a critical collective coupling strength. For the scattering experiments, the authors propose that both ensemble ordering/clustering and Thomson scattering from freed electrons may contribute, noting that ordering is likely a prerequisite for forming the collectively degenerate state that triggers the correlation phase transition. They also acknowledge that the number density of collectively ionized orbitals likely remains low, so average chemical properties should be only marginally affected while ion-sensitive observables respond strongly.

Limitations and open questions

Several assumptions constrain the results. The derivation relies on the uniform-probability ansatz over the degenerate subspace, the equivalence-class assumption of zero-mean couplings with extensive variance, and the independence of $$E_{\rm corr} \approx \langle \hat{N}_v\rangle_{\rm uniform}\, E_{\rm CIS,corr},$$5 and $$E_{\rm corr} \approx \langle \hat{N}_v\rangle_{\rm uniform}\, E_{\rm CIS,corr},$$6. For the insulating case, the microscopic origin of the correlated fluctuations needed for extensive scaling remains unspecified, as does any criterion for entering the correlated phase. In the metallic case, the density of entropically ionized electrons is undetermined, and the para/spin-glass distinction is invisible to static polarizability. More broadly, the connection between the infinitely degenerate spin glass solution and concrete chemical observables requires non-perturbative treatment beyond this work, and the theory's quantitative comparison with experimental dielectric, conductance, or scattering data remains to be established.

Conclusion

This work derives closed-form, temperature-dependent expressions for the cavity-induced transverse polarizability arising from glassy electron correlations, establishing that fractionally filled collectively degenerate states yield a locally divergent polarizability — i.e., a cavity-induced metallic phase driven purely by entropic correlation effects. The result provides a mechanistic candidate for experiments reporting dramatic conductivity, dielectric, and scattering changes under collective strong coupling, and marks a step toward connecting spin glass theory of polaritonic chemistry to measurable observables.