Plasmon-mediated pairing and dynamical screening

Scope

Plasmon-mediated pairing is a proposal in which the frequency-dependent screening of Coulomb interactions contributes to a superconducting pairing kernel. A plasmon is a collective charge oscillation; it is not an exciton, a static charge-density wave or an acoustic phonon. Observing a charge mode establishes something available to couple to electrons, not the cause of their superconductivity.

Takada's early electron-gas calculation used weak coupling and a plasmon-pole approximation to examine dimensionality. Akashi and Arita later developed a dynamical superconducting density-functional kernel and predicted cooperative phonon–plasmon enhancement for compressed fcc lithium. These are calculations under specified approximations, not experimental isolation of a universal pairing mechanism. Only their publisher abstracts have been checked here. R1 R5

Mathematical assumptions

In a homogeneous scalar description, the screened interaction is $W(q,z)=V_q/\epsilon(q,z)$. Crystals and interfaces can require dielectric matrices and wavefunction matrix elements. A longitudinal collective mode is associated with a pole of the response; with damping its frequency is complex. A zero of $\operatorname{Re}\epsilon$ on the real axis is not a complete mode definition. The measured loss function $-\operatorname{Im}\epsilon^{-1}$ also contains line-shape information.

An independent Drude illustration normalizes the background dielectric constant to one and assumes plasma frequency $\omega_p$ and damping rate $\gamma$:

$\epsilon(z)=1-\frac{\omega_p^2}{z(z+i\gamma)},\qquad z_\pm=-\frac{i\gamma}{2}\pm\sqrt{\omega_p^2-\frac{\gamma^2}{4}}.$

Direct complex substitution and numerical loss maximization were checked for four damping ratios. At $\gamma/\omega_p=0.5$, the positive real-part zero is approximately $0.866\omega_p$, the real part of the pole is $0.96825\omega_p$, and the loss maximum is $0.96877\omega_p$. They are different quantities. For stronger damping a loss maximum can remain even without a real-axis dielectric zero; it does not guarantee a long-lived excitation.

With damping set to zero, a positive plasma frequency $\omega_p>0$ and nonzero real $\nu$, substituting the imaginary frequency $i\nu$ into the same illustration gives

$\frac{W(i\nu)}{V}=1-\frac{\omega_p^2}{\nu^2+\omega_p^2}=\frac{\nu^2}{\nu^2+\omega_p^2}.$

The ratio lies between zero and one for finite nonzero $\nu$. Its continuous $\nu\to0$ limit is zero; direct substitution into the singular dielectric function at zero frequency is not defined. The negative retarded term in the decomposition does not make the total interaction negative. Removing the instantaneous term would change the model. This long-wavelength expression must not be substituted for finite-momentum static screening. Whether a frequency-dependent interaction supports pairing requires the full gap equation, normal self-energy and a defensible treatment of high-energy states; the sign of one separated term is insufficient.

Geometry also changes the modes. For two identical ideal electron sheets separated by $d$, with no tunneling and a hydrodynamic density response, diagonalizing the Coulomb layer matrix gives

$\omega_\pm^2=Cq\left(1\pm e^{-qd}\right),\qquad C=\frac{2\pi n_{2D}e^2}{\kappa m}.$

Here $q$ is the in-plane wavevector magnitude, $\kappa$ the background permittivity and Gaussian electrostatic units are used. In the $qd\ll1$ limit, the in-phase mode scales as $\sqrt{2Cq}$ and the out-of-phase mode as $q\sqrt{Cd}$. Four direct matrix checks recover these limits. The illustration omits tunneling, finite thickness, carrier pressure and the particle–hole continuum. A linear branch in this ideal model neither guarantees weak Landau damping nor describes every layered crystal.

A separate lossless ionic-screening illustration uses $\epsilon_{\mathrm{tot}}/\epsilon_\infty=(\omega_L^2-z^2)/(\omega_T^2-z^2)-\omega_e^2/z^2$, where $\omega_L>\omega_T>0$ describe one polar oscillator and $\omega_e$ is the electronic plasma scale screened only by $\epsilon_\infty$. Its squared mode frequencies satisfy $x^2-(\omega_L^2+\omega_e^2)x+\omega_e^2\omega_T^2=0$. Direct root searches agree with the quadratic roots for five normalized examples. At low electronic plasma scale the lower branch approaches $\omega_e\omega_T/\omega_L$; at high scale it approaches $\omega_T$. This demonstrates hybridization, not a calculated material transition temperature.

Experimental criteria

Measure the charge response over the momentum and energy window relevant to the electrons, including linewidths, polarization, dielectric surroundings and calibration. Optical measurements near zero momentum and finite-momentum scattering constrain different parts of the problem. A pairing attribution additionally needs a screened interaction consistent with the band structure, normal response and superconducting observations on comparable samples.

Hepting and colleagues' 2018 electron-doped cuprate study used polarization and momentum dependence to identify collective charge excitations with out-of-plane dispersion. Its publisher abstract supports the mode assignment; this entry has not independently reanalyzed its spectra or extracted a pairing interaction from them. R3

Their 2022 published SLCO study reports Cu-edge RIXS at 20 K with about 40 meV energy resolution and an approximately 120 meV plasmon gap. Near $|H|\le0.005$ reciprocal lattice units, overlap with a phonon prevents unambiguous peak fitting; these points were excluded from the model fit. The authors describe an acousticlike but gapped branch and interpret its dispersion using a layered model with interlayer hopping. The measurement temperature and mode gap are not superconducting transition temperatures or direct pairing evidence. R4

Limits and disputes

Ruhman and Lee's dilute SrTiO3 proposal requires dielectric screening weaker than in the insulating material to obtain its reported pairing scale. The accepted manuscript treats the dielectric response and a high-energy renormalization parameter phenomenologically and restricts the weak-coupling calculation's energy window. Agreement with a superconducting dome therefore does not independently measure the plasmon contribution. Selected accepted-manuscript passages were checked after comparison with the earlier arXiv version; complete model reproduction remains open. R2

Davydov and colleagues compared dynamical Coulomb treatments in Eliashberg theory, superconducting density-functional theory and a hybrid approach. Their calculations show that normal-state Coulomb mass renormalization can oppose enhancement in the pairing channel. Selected published pages 3-4 and the earlier manuscript were checked. The final text also notes that neglecting the normal-state energy-shift correction can fail, for example near a doped-insulator band gap. These approximation-dependent comparisons are not a theorem excluding all plasmon mechanisms; complete version comparison and numerical reproduction remain open. R6

One must not append an attractive plasmon term to a repulsion that already includes the same screening without checking for double counting. Nor does a high collective-mode energy justify applying a phonon adiabatic approximation unchanged. Electronic scales, coupling strength, vertex corrections and competing fluctuations need explicit treatment. No independent calculation in this article solves a superconducting instability or certifies a material mechanism.

Discovery implications

For a dielectric or layered-material proposal, track how controlled density, spacing or screening changes affect both the collective response and superconducting observables. Separate changes in band filling, disorder, structural phase and phonons from the proposed charge-mode effect. Preserve negative results and uncertainty. The useful target is a testable quantitative mechanism with identifiable assumptions, not a ranking based on the largest observed plasmon energy. Full scientific acceptance remains open.

Related topics

Compare Eliashberg theory, charge fluctuations, exciton-mediated pairing and superconducting gap.

Material cases

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