The source is temporarily unavailable. Showing a saved statement. Last verified:

Allender, Bray, and Bardeen (1973) proposed that conduction electrons in a thin metal film on a narrow-gap semiconductor acquire an attractive pairing interaction by exchanging virtual excitons of the semiconductor, in direct analogy with phonon-mediated BCS pairing but with a characteristic energy of order 1 eV instead of tens of meV, and calculated that this could substantially enhance the superconducting transition temperature TcT_c. The problem is whether this exciton mechanism can actually be realized: whether an admissible structure (an ABB-type metal-on-narrow-gap-semiconductor structure, one of the paper's acknowledged alternatives, or a direct modern analogue) can achieve a net exciton-mediated attraction exceeding the Coulomb repulsion (λex>μ\lambda_{\mathrm{ex}}>\mu, given the disputed local-field/umklapp enhancement and the double-counting critique), whether the exacting fabrication conditions can be met, and whether superconductivity demonstrably attributable to the exciton channel can be produced. Despite fifty years of theoretical elaboration, repeated proposals, and contested claims, the mechanism has never been unambiguously demonstrated in any material system, and the feasibility question remains open.

Background

In conventional superconductivity, described by the Bardeen-Cooper-Schrieffer (BCS) theory, electrons near the Fermi surface form Cooper pairs by exchanging virtual phonons, the quantized vibrations of the crystal lattice. Because the phonon energy scale (the Debye energy, typically tens of meV) is small compared with electronic energies, and because strengthening the electron-phonon coupling eventually destabilizes the lattice, phonon-mediated transition temperatures TcT_c are in practice limited. Shortly after BCS, W. A. Little (1964) and V. L. Ginzburg proposed that virtual excitons could replace phonons as the pairing glue. An exciton is a bound electron-hole pair, the elementary optical excitation of a semiconductor or insulator, with a typical energy of order 1 eV; an attractive interaction cutoff at such an energy would raise the BCS prefactor for TcT_c by more than an order of magnitude, in principle enabling high-temperature superconductivity. Ginzburg and Kirzhnits systematically developed this excitonic route, focusing on layered metal-dielectric sandwich structures.

The central obstacle is the Coulomb repulsion. In phonon-mediated superconductivity the slow (low-energy) phonons retard the attractive interaction relative to the essentially instantaneous Coulomb repulsion, so that in the gap equation the repulsion is renormalized to a Coulomb pseudopotential μ∗\mu^* far smaller than the bare parameter μ\mu. Excitons are fast (their energy is comparable to electronic energies), so this retardation benefit largely disappears: the exciton-induced attraction must compete with an almost unreduced repulsion, and stability arguments of Cohen and Anderson suggest the exciton coupling constant λex\lambda_{\mathrm{ex}} (the Fermi-level density of states times the Fermi-surface average of the exciton-mediated interaction) cannot exceed μ\mu. The 1973 paper of Allender, Bray, and Bardeen (ABB) analyzed the most concrete realization proposed until then: a thin metal film, roughly 10-15 Angstroms thick, in intimate contact with a narrow-gap semiconductor such as Ge, InSb, or PbTe. Metal Fermi-surface electrons tunnel about 5 Angstroms into the semiconductor gap, where they polarize the covalent bonds and exchange virtual excitons. ABB argued, by analogy with umklapp-assisted phonon pairing (umklapp processes are crystal-momentum transfers involving a reciprocal-lattice vector), that local-field effects in the microscopic dielectric response permit λex>μ\lambda_{\mathrm{ex}}>\mu without violating stability, estimated λex≈0.2\lambda_{\mathrm{ex}}\approx0.2-0.5 under favorable conditions, and solved a McMillan-type (BCS-like strong-coupling) gap equation numerically: under the stability limit λex≤μ\lambda_{\mathrm{ex}}\le\mu the pure-exciton TcT_c reaches only about 30 K, while for unrestricted λex≈1.4\lambda_{\mathrm{ex}}\approx1.4 values up to about 800 K arise (acknowledged as unrealistic); exciton-phonon cooperation with λex∼0.2\lambda_{\mathrm{ex}}\sim0.2-0.3 substantially enhances TcT_c over the phonon-only value. In the same year, Inkson and Anderson countered that such calculations double-count the excitonic polarization already contained in the screened Coulomb interaction, so the enhancement disappears; ABB replied that the critique describes a different class of semiconductor, one whose interband spectral weight disperses strongly with momentum transfer as in a metal, whereas covalent semiconductors keep their exciton spectral weight near the average gap energy ℏωg\hbar\omega_g for nearly all momenta, making large-qq (umklapp) processes decisive. This dispute was never decisively settled for a specific real material. The paper poses the establishment of the exciton mechanism as an experimental goal and lists the theoretical issues as acknowledged difficulties and approximations.

The problem is set against exacting physical requirements enumerated by ABB: intimate metal-semiconductor contact with no surface states or interfacial oxide; work functions matched within about 0.5ℏωg0.5\hbar\omega_g so the metal Fermi level sits near midgap and band bending does not suppress tunneling; a continuous ultrathin metal film; a semiconductor with a large ratio of plasma frequency to average gap; and a sufficiently sharp exciton spectral peak, since ABB found TcT_c suppressed exponentially by the exciton linewidth. Modern developments adjacent to this problem include superconductivity at oxide interfaces, exciton-polariton condensates in semiconductor microcavities (coherent macroscopic populations of hybrid light-matter excitations), and theoretical proposals of exciton-mediated pairing in two-dimensional heterostructures; none has produced an accepted demonstration of the exciton mechanism.

Problem Statement

Allender, Bray, and Bardeen's paper closes by stating that the main problem is to try to find or build suitable structures for the exciton mechanism of superconductivity. The research problem is exactly this: establish whether attractive pairing of conduction electrons by the exchange of virtual excitons can be realized in an admissible structure, the prototype being ABB's thin metallic layer in intimate contact with a narrow-gap semiconductor, with the paper's own acknowledged alternatives (metal granules in a dielectric matrix, polarizable molecular assemblies, layered organic/inorganic compounds, alkali-metal-induced band bending) and their direct modern analogues (structures in which conduction electrons acquire an attractive interaction from virtual excitons or excitonic excitations of a polarizable semiconductor or dielectric component, at a metal, doped-semiconductor, or two-dimensional interface or hybrid) equally admissible, or conversely demonstrate that no admissible structure can achieve it.

The objective has three linked facets. (1) Theoretical feasibility: a consistent microscopic treatment of a realistic interface, treating the Coulomb repulsion and the exciton-induced attraction together with the double-counting issue raised by Inkson and Anderson explicitly resolved, determining whether the net interaction at the metal Fermi surface can be attractive, that is, whether the local-field and umklapp enhancement can make the exciton coupling λex\lambda_{\mathrm{ex}} exceed the Coulomb parameter μ\mu, and what TcT_c the resulting gap equation then yields; in the source's parameterization this includes the question whether λex∼0.2\lambda_{\mathrm{ex}}\sim0.2-0.5, the range the paper estimates for its model, is actually attainable in real materials. (2) Experimental realization: for the ABB thin-film prototype, a fabricated structure meeting the physical requirements the paper enumerates, namely intimate contact free of surface states and interfacial oxide, work functions matched within about 0.5ℏωg0.5\hbar\omega_g so that the metal Fermi level lies near the semiconductor midgap and space-charge band bending does not suppress tunneling, a sufficiently thin continuous metal film, a semiconductor with a large ratio of plasma frequency to average gap, and a sufficiently sharp exciton spectral peak (the paper found TcT_c suppressed exponentially by the exciton linewidth), which exhibits superconductivity. The analogous requirements for its own geometry apply to each acknowledged alternative and direct modern analogue. (3) Attribution: evidence that the observed superconductivity is caused by the exciton channel rather than by phonons, the superconducting proximity effect, strain, or disorder.

The question therefore has two directions. Realization in an admissible structure is established by a generally accepted experimental demonstration --- one whose attribution signature has been independently reproduced --- carrying its own attribution evidence; refutation is established by a rigorous demonstration that no admissible structure can achieve a net exciton-mediated attraction exceeding the Coulomb repulsion. The reach of a refutation is fixed by the class it covers: a no-go established only for part of the admissible class bears on that part alone, whereas a no-go established at the level of the general criteria the paper enumerates, or their stated analogues for other geometries, covers the entire admissible class. As of 2026 the problem is open: no generally accepted demonstration or refutation exists in the literature.

References

  1. D. Allender, J. Bray, J. Bardeen, "Model for an Exciton Mechanism of Superconductivity", Phys. Rev. B 7, 1020 (1973). DOI: 10.1103/PhysRevB.7.1020. https://doi.org/10.1103/PhysRevB.7.1020 (primary source)
  2. J. C. Inkson, P. W. Anderson, "Comment on 'Model for an Exciton Mechanism of Superconductivity'", Phys. Rev. B 8, 4429 (1973). DOI: 10.1103/PhysRevB.8.4429. https://doi.org/10.1103/PhysRevB.8.4429
  3. D. Allender, J. Bray, J. Bardeen, "Comment on 'Model for an Exciton Mechanism of Superconductivity' --- A Reply", Phys. Rev. B 8, 4433 (1973). DOI: 10.1103/PhysRevB.8.4433. https://doi.org/10.1103/PhysRevB.8.4433
  4. W. A. Little, "Possibility of Synthesizing an Organic Superconductor", Phys. Rev. 134, A1416 (1964). DOI: 10.1103/PhysRev.134.A1416. https://doi.org/10.1103/PhysRev.134.A1416
  5. V. L. Ginzburg, D. A. Kirzhnits, "On the problem of high temperature superconductivity", Physics Reports 4, 343 (1972). DOI: 10.1016/0370-1573(72)90017-8. https://doi.org/10.1016/0370-1573(72)90017-8
  6. M. L. Cohen, P. W. Anderson, "Comments on the Maximum Superconducting Transition Temperature", AIP Conference Proceedings 4, 17 (1972). DOI: 10.1063/1.2946185. https://doi.org/10.1063/1.2946185
  7. W. A. Little, "A Look Back at the Consequences of an Innocent Question asked Fifty Years Ago on the possibility of Synthesizing as Organic Superconductor", Journal of Superconductivity and Novel Magnetism 29, 3 (2015). DOI: 10.1007/s10948-015-3261-9. https://doi.org/10.1007/s10948-015-3261-9
  8. A. Bill, H. Morawitz, V. Z. Kresin, "Electronic collective modes and superconductivity in layered conductors", Phys. Rev. B 68, 144519 (2003). DOI: 10.1103/PhysRevB.68.144519. https://doi.org/10.1103/PhysRevB.68.144519
  9. S. H. Rhim, R. Saniz, Jaejun Yu, Lin-Hui Ye, A. J. Freeman, "Interface electronic structure, two-dimensional metallicity, and possible interface superconductivity in CuCl/Si superlattices", Phys. Rev. B 76, 184505 (2007). DOI: 10.1103/PhysRevB.76.184505. https://doi.org/10.1103/PhysRevB.76.184505
  10. S. H. Rhim, R. Saniz, A. J. Freeman, "Interface two-dimensional metallicity but lack of superconductivity in heterobonded semiconductors: ZnS/Si and GaP/Si", Phys. Rev. B 79, 045313 (2009). DOI: 10.1103/PhysRevB.79.045313. https://doi.org/10.1103/PhysRevB.79.045313
  11. F. P. Laussy, A. V. Kavokin, I. A. Shelykh, "Exciton-polariton mediated superconductivity", Phys. Rev. Lett. 104, 106402 (2010). DOI: 10.1103/PhysRevLett.104.106402; arXiv:0907.2374. https://arxiv.org/abs/0907.2374
  12. Y. Kurokawa, T. Hihara, I. Ichinose, K. Sumiyama, "Possibility of Exciton Mediated Superconductivity in Nano-Sized Sn/Si Core-Shell Clusters: A Process Technology towards Heterogeneous Material in Nano-Scale", Jpn. J. Appl. Phys. 51, 070207 (2012). DOI: 10.1143/JJAP.51.070207; arXiv:1210.5027. https://arxiv.org/abs/1210.5027
  13. E. D. Cherotchenko, T. Espinosa-Ortega, A. V. Nalitov, I. A. Shelykh, A. V. Kavokin, "Superconductivity in semiconductor structures: The excitonic mechanism", Superlattices and Microstructures 90, 170 (2016). DOI: 10.1016/j.spmi.2015.12.003; arXiv:1410.3532. https://arxiv.org/abs/1410.3532
  14. P. Skopelitis, E. D. Cherotchenko, A. V. Kavokin, A. Posazhennikova, "Interplay of Phonon and Exciton-Mediated Superconductivity in Hybrid Semiconductor-Superconductor Structures", Phys. Rev. Lett. 120, 107001 (2018). DOI: 10.1103/PhysRevLett.120.107001; arXiv:1802.01348. https://arxiv.org/abs/1802.01348
  15. A. Kavokin, A. Zakhidov, "Comment on: The interpretation of room-temperature superconductivity experiments in LK-99", Materials Today 69, 1 (2023). DOI: 10.1016/j.mattod.2023.08.008. https://doi.org/10.1016/j.mattod.2023.08.008
  16. J. Cao, A. Kavokin, "The interplay between exciton- and phonon-induced superconductivity might explain the phenomena observed in LK-99", Materials Today Communications 37, 107293 (2023). DOI: 10.1016/j.mtcomm.2023.107293; arXiv:2310.09305. https://arxiv.org/abs/2310.09305
  17. J. Cao, A. Kavokin, "Exciton-enhanced superconductivity in monolayer films of aluminum", Phys. Rev. B 112, 024501 (2025). DOI: 10.1103/vhs1-59z7; arXiv:2409.12201. https://arxiv.org/abs/2409.12201
  18. S. Lin, S. Huang, M. Yang, X. Chen, H. Bi, K. Xiong, "Possible High temperature Superconductivity above 200K mediated by Bose Einstein Condensation of exciton" (2024). arXiv:2412.15003. https://arxiv.org/abs/2412.15003
  19. J. von Milczewski, X. Chen, A. Imamoglu, R. Schmidt, "Superconductivity Induced by Strong Electron-Exciton Coupling in Doped Atomically Thin Semiconductor Heterostructures", Phys. Rev. Lett. 133, 226903 (2024). DOI: 10.1103/PhysRevLett.133.226903. https://doi.org/10.1103/PhysRevLett.133.226903
  20. D. R. Harshman, A. T. Fiory, "Interfacial superconductivity in Cu/Cu2O and its effect on shielding ambient electric fields", Physica C 632, 1354600 (2025). DOI: 10.1016/j.physc.2024.1354600. https://doi.org/10.1016/j.physc.2024.1354600