Exciton-mediated pairing: proposals and evidence boundaries

Scope

Exciton-mediated superconductivity proposes that electronic excitations help bind charged Cooper pairs. It is distinct from condensation of neutral electron–hole pairs and from enhanced pair correlations in a driven finite cluster. This article does not designate a verified material realization of the mechanism; that editorial boundary is not a claim that an exhaustive worldwide search has excluded one.

Little's organic-polymer proposal and the Allender–Bray–Bardeen metal–semiconductor model are historical theoretical starting points. The latter considers metal electrons tunneling into a semiconductor and coupling to virtual excitons, with interface overlap and screening essential. Their high calculated transition scales are predictions under model assumptions, not measured material records. The original publisher abstracts were checked; the full calculations remain unaudited here. R1 R2

Mathematical assumptions

A higher boson energy does not automatically raise a transition scale. An independent weak-coupling illustration assumes an electronic cutoff $E$, mode energy $\Omega\lt E$, attraction $\lambda$, and repulsion $\mu$, all couplings dimensionless:

$\mu^*=\frac{\mu}{1+\mu\ln(E/\Omega)},\qquad k_BT_{\mathrm{model}}=1.13\Omega\exp\left[-\frac{1}{\lambda-\mu^*}\right].$

This illustration requires $\lambda>\mu^*$ and separated electronic and bosonic scales; it omits self-energy, momentum structure and interface geometry. It is not a reproduction of the historical exciton model. At fixed $E,\lambda,\mu$, differentiation gives

$\frac{d\ln T_{\mathrm{model}}}{d\ln\Omega}=1-\frac{(\mu^*)^2}{(\lambda-\mu^*)^2}.$

Five finite-difference checks and an independent cutoff scan show that the prefactor gain can lose to reduced retardation of repulsion. The illustrative optimum for $E=1000$, $\lambda=0.30$, $\mu=0.25$ is $\Omega\simeq69.5$ in the same arbitrary energy unit. None of these numbers is a measured temperature. As $\Omega/E$ increases, the assumed scale separation itself becomes less credible; this formula cannot certify electronic-boson superconductivity.

For a separate symmetry illustration, take three isolated states with energies $0,3,2$ and inversion parities even, odd, even. An odd dipole operator connects the intermediate state to the other two, but has no direct matrix element between them. A drive $F\cos t$ can therefore access the final even state through two interactions. Direct time evolution, with halved time steps and conserved norm, gives a final-state probability ratio of about 15.98 when a sufficiently weak field doubles, consistent with the perturbative $F^4$ limit. At larger fields the ratio is about 15.61: the power law is an asymptote. This isolated three-state check contains neither a condensate nor a relaxation bath.

Neutral interlayer coherence can be represented by $\langle c_1^\dagger c_2\rangle$. It is invariant under a common electronic phase rotation, unlike a charged Cooper-pair amplitude. An ideal relative-phase response has a normalized stiffness matrix proportional to $\begin{pmatrix}1&-1\\-1&1\end{pmatrix}$: the common layer-field direction has zero eigenvalue, while the difference direction has finite stiffness. Three direct response checks produce equal and opposite layer currents. This illustration excludes tunneling, contacts and additional charged channels; it does not predict an entire device's electromagnetic response.

Experimental criteria

A convincing exciton-pairing attribution needs an independently characterized electronic mode, a screened coupling calculation and a controlled change of that mode that tracks pairing observables. Establish superconductivity separately through suitable transport and magnetic evidence. Then compare phonon, spin, plasmon, structural and interface alternatives on the same samples. A spectral peak, optical resonance or conductive interface alone does not select the mechanism.

Kellogg and colleagues measured separately contacted GaAs bilayers at total Landau filling one. Counterflow and parallel-current geometries show different Hall responses; the paper interprets the low-temperature counterflow trend as evidence for excitonic superfluidity. It also reports finite dissipation and no expected nonlinear signature in the examined range. These quantum-Hall conditions are not evidence for exciton-mediated charged Cooper pairing in an arbitrary van der Waals bilayer. R9

Mitrano and colleagues' earlier K3C60 manuscript reports transient superconducting-like optical features while considering a high-mobility alternative within its finite probe window. The publisher’s 24 February 2016 notice identifies a restored, previously truncated final sentence in the advance-online PDF Figure 4 caption; a full final-version comparison remains open. Budden and colleagues add calibrated electrical measurements: their author manuscript uses a 1 ns two-terminal probe and subtracts contact contributions inferred from static four-terminal measurements. At a 100 K base temperature, the reported transient resistance is compatible with zero within calibration uncertainty. This is not a direct transient four-terminal measurement, and 100 K is not entered as an equilibrium transition temperature. R3 R4

Limits and disputes

Inkson and Anderson argued that double counting in a metal–semiconductor interaction treatment removes the proposed excitonic enhancement. Allender, Bray and Bardeen replied that the criticized model differs from theirs and stressed dielectric spectral weight over a broad momentum range. Both publisher abstracts are retained. This historical exchange identifies a screening and model-consistency problem; it does not by itself establish either a universal mechanism or a universal prohibition. R7 R8

Aranzadi and colleagues' 2026 preprint, version 2, proposes a purely electronic two-photon pathway for light-enhanced pair correlations. Its 14-site one-orbital calculation reaches roughly 30 THz, still above the experimental 10 THz reference scale. In one parameter range the calculation starts in the lowest even-parity state rather than the odd ground state; initialization matters. The paper explicitly leaves relevant decay channels outside its isolated dynamics. Finite-cluster correlations and a suggested scaling trend do not establish bulk phase coherence, a metastable lifetime or the classic virtual-exciton mechanism. The source remains a preprint. R6

Ganin and colleagues' equilibrium Cs3C60 report is a separate fulleride evidence chain. Its 38 K publication title is retained as bibliographic information; this entry has not audited the complete pressure, sample and transition protocol and does not add a material record from the title. Neither that result nor photo-driven K3C60 independently identifies exciton-mediated pairing. R5

Discovery implications

Evaluate an interface proposal by its accessible electronic spectrum, wavefunction overlap and complete screened interaction, including normal-state effects. State whether an observed current is charged coflow, neutral counterflow or a pump-induced transient. Preserve contradictory measurements and preparation details. Mechanism discrimination requires predictions that survive competing explanations; a large model energy scale or a nonlinear optical response is insufficient. Full scientific, final-version and raw-data audits remain open.

Related topics

Compare charge fluctuations, Eliashberg theory, superconducting gap and the keyword catalogue for nonequilibrium and interface superconductivity.

Material cases

Values describe individual reports, not a ranking. Unknown means the checked source did not supply the field at this review scope.