What limits Tc in flat-band superconductors?

Scope

There is no single numerical temperature ceiling supplied by the phrase “flat band.” A meaningful limit specifies the Hamiltonian, density, dimensionality, interaction and current response. A bound is also different from a predicted transition: satisfying an upper bound does not establish that superconductivity occurs. This question concerns constraints on phase coherence and their relation to pair formation. R1 R2

A narrow dispersion can enhance pairing while suppressing ordinary band transport. In a multiorbital system, however, the wavefunctions and interband matrix elements can support a geometric contribution to superfluid weight. Flattening the eigenvalues alone does not determine that contribution. The relevant question is whether a particular interacting system can develop a robust total phase stiffness, with the required electromagnetic and thermodynamic limits. R3 R8

Mathematical assumptions

Define the two-dimensional phase stiffness $D_s$ by $F=(D_s/2)\int d^2r\,|\nabla\theta|^2$. Hazra and colleagues bound it by the optical spectral weight, $D_s(T)\leq\widetilde D(T)$. With a BKT transition, $k_BT_c\leq\pi\widetilde D(T_c)/2$. Replacing the right-hand side by its zero-temperature value requires an additional monotonic-stiffness assumption. Their derivation couples the vector potential to the kinetic Hamiltonian; its interaction assumptions must accompany applications. R1

For the specified two-spin parabolic band, $E_F=\pi\hbar^2n/m$ and $\widetilde D=\hbar^2n/(4m)$ give $k_BT_c\leq E_F/8$. This is a strictly two-dimensional BKT result, not a generic multiband or three-dimensional formula. In an anisotropic sheet the corresponding bound uses $\sqrt{\widetilde D_x\widetilde D_y}$. Independent density-convention and coordinate-Jacobian checks reproduce these relations; they do not determine a material's input stiffness. R1

The BKT jump concerns the renormalized total response at the transition. It does not isolate a geometric share or convert a pairing gap directly into phase stiffness. A mean-field amplitude can remain nonzero after phase coherence has been lost. The BKT article discusses current-window, finite-size and resistance-fit ambiguities that also matter when testing a proposed bound experimentally. R6

Isolated-band constraints

Verma and colleagues examine on-site attraction in a controlled hierarchy $w\ll|U|\ll E_0$, with bandwidth $w$ and isolation gap $E_0$. Projecting to localized Wannier states produces pair hopping and an interaction-dependent current operator. For their trivial-band construction, $\widetilde D_{\rm low}\leq\widetilde n|U|\Omega/2$, where $\widetilde n=\min(n,2-n)$ and $\Omega$ is the Wannier spread in the stated normalization. The Lieb-lattice symmetry sharpens the coefficient; it is not a universal constant for every lattice. R2

Their zero-temperature lower bound concerns optical spectral weight and requires the stated long-range-order and correlation assumptions. It is not a general lower bound on finite-temperature superconductivity. Topological bands require care with the Wannier basis: relaxing the Kramers-pair constraint permits localized functions and a generalized overlap bound. Extrapolating the linear interaction trend beyond the isolated-band hierarchy is unjustified; interband physics becomes important. R2

Pair formation is not pair mobility

A simple independent check makes the distinction concrete. An attractive Hubbard dimer was assembled directly in its six-dimensional two-electron Fock space. Sixteen spectra agree with the analytic ground-state energy $E_0=(-|U|-\sqrt{|U|^2+16t^2})/2$. At zero hopping the local pairs have no intersite splitting. In the strong-attraction limit, the low pair-sector splitting approaches $4t^2/|U|$, corresponding to pair hopping $2t^2/|U|$. Stronger local binding can therefore accompany slower pair motion. The dimer has no thermodynamic transition and supplies no predicted Tc. R1 R2

Flat-band geometry can change this mobility argument because overlapping Wannier orbitals generate interaction-mediated pair motion. That is a model-specific route, not permission to discard self-consistency. Peotta and Törmä's geometric response assumes a defined isolated-band and pairing regime. The Chern-trivial Lieb example also demonstrates why nonzero Chern number is not necessary for every geometric contribution. Its existence does not establish superconductivity in an arbitrary flat-band material. R3 R7

Experimental criteria

Measure a total sheet response with units, frequency, device geometry and fitting uncertainties. Tanaka and colleagues infer stiffness from microwave kinetic inductance using a circuit that includes a parallel edge contribution. Their active-region aspect ratio is an assumption in converting the measured inductance to sheet response. A comparison with a conventional estimate may support a geometric interpretation, but it does not directly measure an independently separable geometric component. R5

The selected Tanaka author manuscript distinguishes refrigerator base temperature, wired-device temperature and a half-normal-resistance transition. These are separate observables. A finite-frequency stiffness should be connected to the low-frequency, long-wavelength response before applying an equilibrium BKT criterion. Experimental comparison also needs a stated pairing scale, a temperature-dependent stiffness and evidence that sample inhomogeneity or finite size does not dominate the inferred transition. R5 R6

Limits and disputes

Multiband bounds depend on the actual band structure and occupation assumptions; an approximate estimate is not a universal material ceiling. Hazra's three-dimensional discussion does not establish a general linear stiffness bound there. A successful two-dimensional derivation cannot simply be carried into three dimensions with the same prefactor. R1

A calculated mean-field stiffness may itself be overstated when internal pairing fields are held fixed during a phase twist. Huhtinen and colleagues show a relaxation correction under their time-reversal and equilibrium assumptions. In the isolated uniform-pairing setting, the relevant integrated metric is minimized over orbital positions. A frozen-field geometric answer can therefore differ qualitatively from the relaxed response. This correction must be considered before interpreting a mean-field value as a usable bound on an experimental transition. R4

Complete scientific acceptance remains open. The selected original passages, figures and independent algebra are documented, while full projection-error control, interacting response calculations, final-author comparisons and raw experimental data have not all been reviewed. No model temperature in this page is entered as an experimental material transition. The open problem is to obtain useful, controlled limits for realistic multiband and correlated systems, rather than to maximize one isolated formula. R1 R2 R4 R5

Discovery implications

Specify the density, isolation gap, interaction scale, orbital structure and measurement convention before comparing designs. Check whether a change enhances pairing, pair mobility or both, and whether it changes the assumptions behind the calculation. Report the full relaxed response and its uncertainty alongside any decomposition. A useful prediction should identify the observation that would falsify it; a large density of states or geometric metric alone does not meet that standard. R3 R4 R8

Related topics

Continue with flat-band pairing, flat-band superconductivity, geometric superfluid weight, quantum metric, and BKT transitions.

Material cases

No sample-specific material records are included in this article. Model results are not experimental transition temperatures.