Polariton-Mediated Superconductivity

• Polariton-mediated superconductivity is a mechanism where electron pairing is induced by bogolons from a Bose-condensed exciton-polariton field instead of phonons.
• It leverages hybrid Bose–Fermi systems in semiconductor heterostructures, atomically thin TMDs, and moiré flat bands to realize tunable superconducting properties.
• Dynamical screening and retarded, sign-changing interactions require non-BCS theoretical treatments to accurately predict critical temperatures.

Searching arXiv for recent and foundational work on polariton-mediated superconductivity. Polariton-mediated superconductivity denotes a class of pairing mechanisms in which electrons attract each other through the exchange of collective excitations of an excitonic or exciton-polariton condensate rather than through conventional lattice phonons. In the canonical semiconductor realization, a two-dimensional electron gas coexists with a nearby Bose-condensed exciton or exciton-polariton fluid, and the relevant exchanged bosons are Bogoliubov excitations of that condensate, often termed bogolons. The subject spans weak-coupling BCS-type proposals in semiconductor heterostructures, strong-coupling extensions involving trions and bipolarons in atomically thin materials, and moiré flat-band constructions in which polariton-induced attraction drives a non-BCS correlated regime (Laussy et al., 2011, Milczewski et al., 2023, Sun et al., 2023).

1. Definition and physical realizations

The basic physical setting is a hybrid Bose–Fermi system in which a condensate of excitons or exciton-polaritons is placed in contact with mobile electrons. In the formulation developed for semiconductor microcavities and coupled quantum wells, the bosonic condensate is electrically neutral and does not itself conduct current; rather, it mediates an effective attraction between charged electrons in a nearby two-dimensional electron gas (Laussy et al., 2011). This separates polariton-mediated superconductivity from both conventional phonon superconductivity and from proposals where cavity photons hybridize with collective modes of an already superconducting condensate, such as Higgs-polaritons or Bardasis-Schrieffer polaritons, which concern manipulation of a pre-existing superconducting state rather than the primary pairing glue (Raines et al., 2019, Allocca et al., 2018).

Two semiconductor geometries recur in the literature. In one, a microcavity contains quantum wells such that one well is $n$-doped and hosts a 2DEG, while another hosts excitons that strongly couple to the cavity photon mode, forming exciton polaritons; the polariton condensate forms at $k=0$, while the electron gas occupies a Fermi sea in the neighboring well (Laussy et al., 2011). In the other, there is no cavity: coupled quantum wells host a condensate of spatially indirect excitons, separated by a thin barrier from a quantum well containing the 2DEG (Laussy et al., 2011). Later work broadened the platform range to doped atomically thin semiconductor heterostructures with interlayer excitons (Milczewski et al., 2023), twisted moiré TMD structures coupled to a polariton condensate (Sun et al., 2023), and a single doped tungsten-based TMD monolayer embedded in a microcavity, where band inversion allows doped electrons and pumped excitons to occupy different conduction bands of the same layer (Choo et al., 13 Aug 2025).

A recurring distinction is between pure exciton-mediated and polariton-mediated cases. Formally they can be treated in the same scheme, but in the polariton case one uses the polariton dispersion $E_{\mathrm{pol}}(k)$, includes the excitonic Hopfield coefficient $X$, and obtains a more structured effective interaction in frequency; in the pure exciton case one sets $X=1$, replaces the dispersion by a quadratic exciton dispersion, and finds an interaction closer to the conventional Cooper/Bogoliubov shape (Laussy et al., 2011). This emphasizes that the photonic component does not by itself couple directly to electrons; the decisive interaction channel remains the excitonic matter component.

2. Microscopic pairing mechanism

The foundational weak-coupling picture starts from a microscopic Hamiltonian containing electrons, bosons, electron-electron Coulomb repulsion, electron-boson scattering, and boson-boson interactions. For the microcavity case, the Hamiltonian is written as

$$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$

$$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$

with $a_{\mathbf{k}}$ annihilating a polariton, $\sigma_{\mathbf{k}}$ annihilating an electron, $V_\mathrm{C}(\mathbf{q})$ the direct electron-electron Coulomb interaction, $k=0$0 the electron-exciton interaction, $k=0$1 the excitonic Hopfield coefficient, and $k=0$2 the polariton-polariton interaction (Laussy et al., 2011).

The pairing mechanism is obtained by treating the boson field as a condensate plus fluctuations: $k=0$3 with $k=0$4 the condensate density and $k=0$5 the normalization area (Laussy et al., 2011). This turns the electron-boson density-density interaction into a linear coupling between electrons and bosonic fluctuations. Diagonalization of the interacting boson sector by a Bogoliubov transformation yields elementary condensate excitations, or bogolons, and the Hamiltonian becomes

$k=0$6

where

$k=0$7

and

$k=0$8

A central consequence is that the effective coupling grows with $k=0$9, and therefore the induced attraction grows linearly with $E_{\mathrm{pol}}(k)$0 (Laussy et al., 2011). Optical tunability of $E_{\mathrm{pol}}(k)$1 is therefore one of the defining advantages of polariton-mediated proposals.

Integrating out bogolons gives a retarded effective interaction

$E_{\mathrm{pol}}(k)$2

with

$E_{\mathrm{pol}}(k)$3

When $E_{\mathrm{pol}}(k)$4, the denominator is negative, so $E_{\mathrm{pol}}(k)$5 is negative and therefore attractive; at larger exchanged frequencies it becomes repulsive (Laussy et al., 2011). In the polariton case, because $E_{\mathrm{pol}}(k)$6 and $E_{\mathrm{pol}}(k)$7 depend nontrivially on the polariton composition and dispersion, the effective interaction alternates between weak attraction, strong attraction, and strong repulsion over different frequency windows (Laussy et al., 2011). This sign-changing retardation structure is one of the characteristic departures from textbook square-well BCS kernels.

3. Interaction structure, screening, and strong-coupling extensions

A persistent theme is that the electron-mediated and boson-mediated sectors cannot be analyzed independently of screening. In the weak-coupling semiconductor treatment, the repulsive Coulomb term is kept explicitly and in 2D is modeled as

$E_{\mathrm{pol}}(k)$8

with $E_{\mathrm{pol}}(k)$9 the dielectric permittivity and $X$0 a screening constant (Laussy et al., 2011). The interaction is then averaged over the Fermi surface using

$X$1

The resulting Fermi-surface-averaged kernel $X$2 enters the gap equation (Laussy et al., 2011).

The geometry dependence is especially subtle because the electron-exciton interaction depends on transferred momentum $X$3, interwell spacing $X$4, exciton Bohr radius $X$5, electron and hole mass fractions $X$6, and any exciton dipole moment $X$7. The direct and dipolar electron-exciton matrix elements are written as

$X$8

and

$X$9

The direct term vanishes at small $X=1$0, while the dipolar term is maximal there and is naturally much larger (Laussy et al., 2011). This is why indirect excitons with intrinsic dipole moments, or cavity structures in which a dipole moment is induced by electric fields, are favored.

Later work argued that simple Fröhlich-type exchange can be qualitatively incomplete in strong-coupling regimes. In doped TMD heterostructures with interlayer excitons, the electron–exciton attraction can be strong enough to form trions, and the effective electron–exciton interaction then develops a strong frequency and momentum dependence (Milczewski et al., 2023). The theory introduces a trion field $X=1$1 via Hubbard–Stratonovich decoupling,

$X=1$2

with $X=1$3 in the contact-interaction limit (Milczewski et al., 2023). The condensate hybridizes electrons and trions, and the pairing kernel is built from off-diagonal electron–trion propagators. In this framework, the effective interaction is inseparable from the trion-dressed propagators and vertex renormalization, and the relevant control parameter is effectively $X=1$4, together with $X=1$5 (Milczewski et al., 2023). This suggests that any serious polariton generalization in strongly interacting semiconductor systems must account for the excitonic sector’s ability to support trions and polarons rather than treating polariton exchange as a weak perturbation.

A different beyond-weak-coupling route appears in a single doped tungsten-based TMD monolayer embedded in a microcavity. There the electron-polariton interaction is treated through the full electron-polariton $X=1$6-matrix,

$X=1$7

which at low energies in 2D becomes

$X=1$8

The logarithmic divergence when

$X=1$9

encodes a Feshbach-like enhancement associated with trion resonances (Choo et al., 13 Aug 2025). In this scheme, the induced interaction between electrons contains full $$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$0-matrix vertices rather than bare couplings: $$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$1 This replaces the weak-coupling $$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$2 picture by an explicitly resonant, energy-dependent kernel (Choo et al., 13 Aug 2025).

4. Gap equations and superconducting regimes

The superconducting instability is commonly formulated through a BCS-type gap equation. In the foundational semiconductor treatment the superconducting state is described by

$$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$3

or in energy form

$$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$4

with $$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$5 (Laussy et al., 2011). Because the true kernel is sign-changing, the analysis often uses stepwise approximations. For a standard square-well BCS attraction one recovers

$$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$6

which in weak coupling becomes

$$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$7

and the usual critical scale

$$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$8

When Coulomb repulsion is included as a repulsive elbow around an attractive region, the resulting analytical estimate is

$$H=\sum_{\mathbf{k}}\left[E_{\mathrm{pol}(\mathbf{k})}\,a_{\mathbf{k}}^{\dagger }a_{\mathbf{k}}+E_{\mathrm{el}(\mathbf{k})}\,\sigma_{\mathbf{k}}^{\dagger }\sigma _{\mathbf{k}}\right]$$9

showing that repulsion renormalizes but does not simply destroy superconductivity (Laussy et al., 2011).

For the more structured polariton kernel, the interaction is represented by a three-step potential with a shallow low-energy attraction up to $$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$0, a deeper attraction up to $$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$1, and a repulsive region up to $$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$2. The corresponding analytical estimate for $$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$3 is

$$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$4

One consequence is that $$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$5 depends not only on the total attractive weight but on the detailed balance of attractive and repulsive frequency windows (Laussy et al., 2011).

The more recent strong-coupling excitonic theory organizes the instability through a Bethe–Salpeter equation for the renormalized electron–trion vertex. The operational criterion is

$$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$6

which functions as a Thouless criterion (Milczewski et al., 2023). In weak coupling, $$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$7 is taken to be close to the actual 2D superfluid transition temperature, whereas in strong coupling it is the pair-formation scale and the actual superfluid transition is estimated by a BKT theory for a gas of bipolarons: $$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$8 This supports an emergent BCS–BEC crossover from weakly bound $$+\sum_{\mathbf{k}_{1},\mathbf{k}_{2},\mathbf{q}}\Bigg[ V_\mathrm{C}(\mathbf{q})\,\sigma_{\mathbf{k}_1+\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_2-\mathbf{q}}^{\dagger}\sigma_{\mathbf{k}_1}\sigma_{\mathbf{k}_2} +X V_\mathrm{X}(\mathbf{q})\,\sigma _{\mathbf{k}_{1}}^{\dagger }\sigma _{\mathbf{k}_{1}+\mathbf{q}}\,a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger }a_{\mathbf{k}_{2}} +U\,a_{\mathbf{k}_{1}}^{\dagger }a_{\mathbf{k}_{2}+\mathbf{q}}^{\dagger}a_{\mathbf{k}_{1}+\mathbf{q}}a_{\mathbf{k}_{2}} \Bigg]$$9-wave Cooper pairs to a superfluid of bipolarons (Milczewski et al., 2023).

In the resonantly enhanced single-layer TMD proposal, the instability is treated through the Thouless criterion at pair energy $a_{\mathbf{k}}$0, leading to the linearized gap equation

$a_{\mathbf{k}}$1

with explicit inclusion of the polaronic chemical-potential shift

$a_{\mathbf{k}}$2

The pairing channel is explicitly $a_{\mathbf{k}}$3-wave, and the treatment includes both the momentum dependence of the kernel and the leading one-particle renormalization (Choo et al., 13 Aug 2025).

5. Strong correlation, spinor condensates, and hybrid enhancement mechanisms

The subject broadened beyond weakly interacting parabolic-band 2DEGs when polariton-mediated attraction was transplanted into moiré flat-band settings. In a twisted TMD homobilayer near a magic angle, placed near an insulating excitonic layer inside an optical cavity, the polariton condensate induces an intervalley interaction

$a_{\mathbf{k}}$4

with

$a_{\mathbf{k}}$5

(Sun et al., 2023). Here the polariton-mediated term is enhanced by the condensate density $a_{\mathbf{k}}$6, by the electron–exciton coupling $a_{\mathbf{k}}$7, and by the excitonic content through the Hopfield coefficients. Because spin-valley locking in the moiré bands causes direct and exchange intravalley terms to nearly cancel, intervalley attraction dominates: $a_{\mathbf{k}}$8 at low density (Sun et al., 2023).

Projecting the induced interaction into a quasi-flat topological band gives

$a_{\mathbf{k}}$9

with $\sigma_{\mathbf{k}}$0 (Sun et al., 2023). In the exactly flat-band limit, the kinetic term becomes trivial and the interaction-only Hamiltonian exhibits an emergent SU(2) symmetry generated by

$\sigma_{\mathbf{k}}$1

with

$\sigma_{\mathbf{k}}$2

This yields exact paired ground states

$\sigma_{\mathbf{k}}$3

a single-particle gap of order $\sigma_{\mathbf{k}}$4, and a pseudogap scale $\sigma_{\mathbf{k}}$5, while finite-$\sigma_{\mathbf{k}}$6 superconductivity at exactly zero bandwidth is forbidden because it would require spontaneous breaking of the emergent SU(2) in 2D (Sun et al., 2023). Once a small bandwidth $\sigma_{\mathbf{k}}$7 is restored, SU(2) is broken to U(1), producing a superconducting dome with $\sigma_{\mathbf{k}}$8 and $\sigma_{\mathbf{k}}$9 in the strong-coupling regime (Sun et al., 2023).

Another direction exploits the internal spinor structure of exciton-polariton condensates. For a two-component spinor condensate, fermions couple to both density and spin Bogoliubov modes: $V_\mathrm{C}(\mathbf{q})$0 and the induced interaction is

$V_\mathrm{C}(\mathbf{q})$1

(Bighin et al., 2022). Near the miscibility threshold of the two-component condensate, the spin mode softens, $V_\mathrm{C}(\mathbf{q})$2, and on the Fermi surface the interaction reduces to

$V_\mathrm{C}(\mathbf{q})$3

with $V_\mathrm{C}(\mathbf{q})$4 (Bighin et al., 2022). This yields a resonantly amplified, long-ranged effective attraction associated with a spinodal quantum critical point rather than with a two-body trion resonance.

A distinct hybrid strategy is to combine ordinary phonon-mediated pairing with a polariton condensate. In a semiconductor–superconductor heterostructure, the bogolon channel supplements an existing phonon channel, producing a two-cutoff pairing problem. The effective two-shell couplings are

$V_\mathrm{C}(\mathbf{q})$5

leading to

$V_\mathrm{C}(\mathbf{q})$6

and

$V_\mathrm{C}(\mathbf{q})$7

(Skopelitis et al., 2018). In this class of proposals, polaritons enhance rather than replace phonon superconductivity.

6. Quantitative expectations, materials platforms, and unresolved issues

Quantitative claims vary sharply across models because the effective attraction is extremely sensitive to screening, lifetime, geometry, and whether weak- or strong-coupling physics is assumed. The foundational semiconductor theory used representative GaN-like parameters $V_\mathrm{C}(\mathbf{q})$8, $V_\mathrm{C}(\mathbf{q})$9 nm, $k=0$00 m$k=0$01, $k=0$02 m$k=0$03, $k=0$04 m, $k=0$05 meV, dipole moment $k=0$06 nm, Rabi splitting $k=0$07 meV, and $k=0$08 for polaritons, and found that increasing condensate density by roughly one order of magnitude dramatically enhances attraction and raises the estimated critical temperature (Laussy et al., 2011). The authors described the resulting temperatures as “very high” and even “record breaking” in idealized conditions, but also explicitly warned that other effects would likely intervene before such temperatures are physically realized (Laussy et al., 2011).

By contrast, a later many-body reassessment of Coulomb-coupled polariton–electron mixtures, including dynamical screening of both the 2DEG and the polariton condensate, concluded that superconductivity is feasible only in a very narrow window near a roton instability of the coupled system, with realistic critical temperatures estimated to be not larger than $k=0$09 K in the vicinity of instability (Plyashechnik et al., 2023). In that treatment the dynamically screened interaction is

$k=0$10

and the polariton-induced term is

$k=0$11

The key attraction arises from soft lower hybrid roton modes, and the authors stressed that static BCS treatments overestimate $k=0$12 because they replace the small roton energy scale by a much larger electronic energy window (Plyashechnik et al., 2023). This paper also argued that pair-bogolon or noncondensate processes are negligible once screening is included (Plyashechnik et al., 2023).

More optimistic strong-coupling estimates have emerged in atomically thin materials. In excitonic TMD heterostructures with trion physics, the maximal transition temperature can reach

$k=0$13

when the weak-coupling pair-formation scale is matched to the strong-coupling $k=0$14 side around the crossover boundary (Milczewski et al., 2023). In a single doped TMD monolayer with resonantly pumped polaritons and trion fine structure, the main prediction is that critical temperatures of order $k=0$15 K are achievable with WSe$k=0$16-like parameters, including $k=0$17, $k=0$18 meV, $k=0$19 meV, $k=0$20 m, $k=0$21 meV, $k=0$22, $k=0$23 meV, and $k=0$24 (Choo et al., 13 Aug 2025). There, $k=0$25 is nonmonotonic in electron density, with an optimum near $k=0$26, and Coulomb repulsion reduces $k=0$27 by about a factor of two but does not eliminate superconductivity (Choo et al., 13 Aug 2025).

In the moiré flat-band proposal, the strong-coupling scale is set not by a BCS exponential but by the ratio $k=0$28. QMC for a model with $k=0$29 meV gave

$k=0$30

while the flat-band bandwidth in twisted TMD bilayers is typically a few meV, implying a cautiously optimistic estimate of $k=0$31 several tens of Kelvin, with a pseudogap scale $k=0$32 K when $k=0$33 reaches tens of meV (Sun et al., 2023). This estimate is tied to the strong-correlation regime $k=0$34, not to ordinary weak-coupling BCS theory.

The choice of material platform remains unsettled. Microcavity GaN-like structures were emphasized early because polariton condensation is accessible at much higher temperatures than indirect-exciton condensation (Laussy et al., 2011). hBN microcavities were later proposed as especially favorable polaritonic platforms because monolayer hBN combines a large exciton binding energy, about $k=0$35 eV, with strong light-matter coupling $k=0$36 meV and room-temperature polariton BEC plausibility; within a Laussy-type adjacent-2DEG geometry, this was used to argue for bogolon-mediated superconductivity at a few tens of Kelvin, though Coulomb interaction and electron-phonon coupling in the 2DEG were neglected in that estimate (Yang et al., 2021). This suggests that materials with unusually robust excitons and large oscillator strength may help primarily by stabilizing the mediator condensate, while the actual electronic pairing problem remains sensitive to screening and nonadiabaticity.

Several misconceptions are repeatedly corrected across the literature. Polariton-mediated superconductivity is not simply “photon-mediated superconductivity,” because the relevant electron coupling originates from the excitonic component weighted by Hopfield coefficients (Laussy et al., 2011, Choo et al., 13 Aug 2025). Nor is it generically enough to insert a static attractive potential into a BCS equation: dynamical screening, lifetime effects, nonequilibrium pumping, and sign-changing frequency structure can qualitatively alter the result (Plyashechnik et al., 2023, Laussy et al., 2011). Finally, the existence of cavity polaritons formed from collective modes of an already superconducting state does not by itself establish a polariton-mediated pairing mechanism; those proposals concern hybridization and control of collective modes rather than the origin of the superconducting instability (Raines et al., 2019, Allocca et al., 2018).

The open problems are correspondingly clear. Equilibrium BCS-like descriptions sit uneasily with driven-dissipative polariton condensates of finite lifetime (Laussy et al., 2011). Strong-coupling regimes require explicit treatment of trions, polarons, or bipolarons (Milczewski et al., 2023, Choo et al., 13 Aug 2025). Screening remains the dominant practical limitation in realistic semiconductor heterostructures (Plyashechnik et al., 2023). And in strongly correlated moiré systems, the central challenge is no longer merely to increase the bosonic glue, but to understand the interplay between polariton-induced attraction, topological flat-band geometry, and phase coherence (Sun et al., 2023). Taken together, these works establish polariton-mediated superconductivity as a broad design framework rather than a single mechanism: electrons can pair through condensate bogolons, through resonantly enhanced polariton–electron scattering shaped by trions, or through polariton-induced attraction projected into flat or nearly flat electronic bands, but in every case the decisive quantities are the excitonic matter content, condensate density, screening environment, and the degree to which the resulting retarded interaction departs from simple weak-coupling BCS form.