Superconducting gap: magnitude, nodes and phase

Scope

A superconducting gap is an energy scale in a specified excitation spectrum. Its magnitude, momentum dependence and relative phase answer different questions. A suppression of spectral weight is not automatically a superconducting gap, and a fitted gap does not automatically define the temperature of bulk superconductivity. This entry organizes observations and model assumptions rather than ranking materials by a single gap ratio.

Specify whether $\Delta$ denotes an order-parameter amplitude, a quasiparticle excitation threshold or a parameter extracted from a broadened spectrum. The relation between these quantities depends on the model. Keep the measured signal, the fit and the inferred state separate; otherwise apparently similar numbers can describe different observables.

Mathematical assumptions

For a clean, isotropic, single-band weak-coupling BCS model, the quasiparticle spectrum and zero-temperature limiting ratio are

$E_{\mathbf k}=\sqrt{\xi_{\mathbf k}^{\,2}+|\Delta|^2},\qquad \frac{2\Delta_0}{k_BT_c}=2\pi e^{-\gamma_E}\simeq3.528.$

Here $\xi_{\mathbf k}$ is measured from the chemical potential and $\gamma_E$ is Euler's constant. The existing BCS entry checks the gap equation and numerical limit. The ratio is not a universal threshold for identifying superconductivity or an interaction. Strong coupling and anisotropy require their own treatment; Carbotte's review abstract describes deviations from simple BCS results without supplying a universal correction. R1 R2

With a constant normal density of states $N_n$ and the same spin convention in numerator and denominator, the ideal normalized density of states at positive excitation energy is

$\frac{N_s(E)}{N_n}=\begin{cases}0,&0\le E<|\Delta|,\\ E/\sqrt{E^2-|\Delta|^2},&E>|\Delta|.\end{cases}$

The factor above the edge follows from the change of variable $E=\sqrt{\xi^2+|\Delta|^2}$ with the usual Bogoliubov spectral weights. The ideal edge is singular but integrable. A local numerical check integrates it up to a finite energy and recovers the corresponding normal-state $\xi$ interval. The check assumes no lifetime broadening, gap anisotropy or experimental convolution. R1

In this magnitude-only model, replacing a gap by its negative leaves the excitation energies unchanged. Likewise, independently summing two uncoupled spectral contributions cannot recover their relative sign. This is a limited counterexample to inferring phase from a gap magnitude; it does not say that every interference-sensitive spectrum is phase blind.

Experimental criteria

Use the following as a reporting map adapted from the pinned chapter. Each interpretation needs the stated model and a sample-specific source; these rows do not certify any material dataset. Selected passages now support the listed controls across all seven probe rows. This is a scoped methods check, not an exhaustive experimental protocol, a final-version audit or independent validation.

The measured ARPES intensity includes occupation and matrix-element factors, and quantitative analysis also requires resolution and background treatment. A leading-edge displacement therefore needs a model before it is reported as a gap amplitude. For calorimetry, Wang and colleagues separately measure addenda and discuss impurity contributions, normalization and thermal equilibration in their MgB2 sample. Heat capacity is measured; entropy requires an integration and reference choice. R12 R13

Penetration depth and heat transport require different controls. The Prozorov-Giannetta review discusses geometry, nonlocality and magnetic or surface contributions, including competing interpretations of a power law. Shakeripour and colleagues warn that nonsuperconducting regions can contribute to a residual heat-conduction term and that multiband response can obscure a particular gap. These selected methodological cautions do not import the reviews' historical material-symmetry assignments as current verdicts. R14 R15

A tunnel conductance resembles the sample density of states only under counterelectrode and matrix-element assumptions; the selected Fischer review gives the thermal convolution and explains the resolution role of lock-in modulation. Do not identify every conductance peak with $\Delta$. Optical interpretation also depends on scattering and multiband weights: the selected Dressel review compares competing fits, while Seibold and colleagues distinguish DOS peaks, excitation thresholds and collective absorption within a disordered model. These are source-scoped cautions, not a new fit to any specimen. Photoemission and thermodynamic probes constrain other aspects; magnitude-sensitive measurements alone do not determine relative phase. R3 R7 R8 R9

For point contacts, a barrier model is only part of the interpretation. Sheet and colleagues distinguish dissipative thermal contacts from a diffusive regime without significant local heating; non-ballistic does not always mean thermal. Kumar and Sheet show how multiple constrictions can mix responses in a network model. These examples motivate contact controls; they neither diagnose every observed peak nor prove that every unconventional spectrum is an artifact. R10 R11

Hardy and colleagues report a strong linear term in the penetration-depth change of YBa2Cu3O6.95 single crystals over approximately 3–25 K using a 900 MHz resonator. This is evidence relevant to low-energy excitations under their sample conditions, not a direct measurement of gap phase. The present check covers the publisher abstract. R4

By contrast, Tsuei and Kirtley use scanning SQUID measurements of NCCO/PCCO tricrystals to test phase-sensitive predictions. Their designed frustrated geometry produces a half-flux signal, while two unfrustrated NCCO geometries serve as controls. The inference depends on junction geometry and the associated model, not merely on observing a spectral dip. Selected author-manuscript passages and the control description were checked; the image fits were not independently reproduced. R6

Limits and disputes

Nodes, relative sign and topology are distinct. A low-temperature power law can constrain low-energy excitations without uniquely selecting a pairing representation. A fully open gap on each Fermi-surface sheet can coexist with a relative sign change between sheets. Disorder and sample inhomogeneity must be considered before translating a temperature dependence into a symmetry assignment. R3

A zero-bias feature alone is not a unique topology test. Identify its field, temperature, position and junction dependence, then compare explicit alternative models. Likewise, a spectroscopic gap-closing temperature must not silently replace a resistive or thermodynamic $T_c$ in a gap ratio. These are inference and reporting limits, not a claim that one alternative explains every observed feature.

Choi and colleagues' MgB2 result is an anisotropic Eliashberg calculation of distinct band-dependent gap scales. Treat those calculated gaps as theory and compare them with separately identified measurements. The pinned chapter's DOI was mismatched; the correct source is nature00898. That correction does not validate a universal multiband gap ratio or independently reproduce the calculation. R5

Discovery implications

Record the specimen, pressure or strain, temperature criterion, spectral resolution, fit model and uncertainty together. Compare probes on the same or demonstrably comparable specimens. A disagreement can expose a surface effect, background assumption or a second energy scale; it should remain visible until tested.

A useful next measurement distinguishes competing interpretations. Establish which observation would separate a change of magnitude from a change of phase, and which control would rule out a contact artifact or another ordered state. Gap measurements constrain models; they do not replace a complete evidence chain for superconductivity.

Related topics

Connect to BCS theory, multiband pairing, s± pairing, specific heat, critical fields, and zero resistance.

Material cases

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