A15 compounds: structure, strain and useful critical fields

Scope

A15 is a crystal-structure family, usually written $A_3B$. Its ideal undistorted prototype is primitive cubic $Pm\bar{3}n$ (No. 223), with eight atoms per cell: six A atoms on 6c sites and two B atoms on 2a sites. The body-centered arrangement of the B sublattice does not make the full crystal body-centered cubic. Three orthogonal A-atom chains help organize its electronic structure, without guaranteeing a high transition temperature. Composition, site ordering, phonons and strain remain important, and the brittle Nb–Sn phase also has to tolerate fabrication and mechanical loading. R5 R11

Nb3Sn, V3Si and Nb3Ge are distinct examples, not one table of universal constants. Matthias and colleagues reported Nb3Sn near 18 K in 1954. Gavaler's 1973 abstract reports sputtered Nb–Ge films remaining superconducting up to 22.3 K. That historical film result is neither an equilibrium bulk value nor a strict A15 temperature ceiling. “High pressure” in the deposition description refers to sputtering conditions, not a high-pressure superconductivity measurement. R1 R2

Near-stoichiometric Nb3Sn is a conventional strong-coupling benchmark, but even its coupling classification depends on composition and how the gap and transition are sampled. Godeke's review discusses interface-gap and inhomogeneity effects that can bias a gap-to-temperature ratio. A full gap is not proof that all samples have the same coupling, and a large upper critical field alone is not evidence for triplet pairing. R5 R9

Mathematical assumptions

For an orbital dirty-limit description, write $B_{c2}=\mu_0H_{c2}$ and $\ln(T/T_c)=\psi(1/2)-\psi[1/2+\hbar D B_{c2}(T)/(2\Phi_0k_BT)]$, where $D$ is an effective diffusion parameter and $\psi$ the digamma function. The Maki–de Gennes form omits Pauli, spin–orbit, strong-coupling and Fermi-surface anisotropy corrections. A useful fit to an inhomogeneous wire does not measure a unique microscopic $D$. Independent numerical roots recover the model ratio $B_{c2}(0)/(T_c|dB_{c2}/dT|_{T_c})=\pi^2e^{-\gamma}/8\simeq0.693$; this is not a universal material relation. R7

For current perpendicular to field, the pinning-force magnitude is $F_p=J_cB$. A common fixed-temperature field shape is $f(h)=h^{1/2}(1-h)^2$, with $h=B/B_{c2}$. It peaks at $h=0.2$ and produces a linear Kramer plot of $J_c^{1/2}B^{1/4}$ only under the model's assumptions. A triangular vortex lattice has $a_\Delta=(4/3)^{1/4}\sqrt{\Phi_0/B}$; the flux-shear expression includes $(1-a_\Delta/d)^{-2}$ for mean grain size $d$. At 12 T and $d=100$ nm, this factor is about 1.36, so neglecting it requires a stated tolerance. Neither the pinning peak nor a fitted Kramer intercept independently identifies the thermodynamic upper critical field. R6

A local two-component example uses the displayed $J_c=C B_{c2}^{2}h^{1/2}(1-h)^2/B$ with the same positive prefactor for both components, critical fields of 20 and 28 T, and zero current at or above each component field. Equal-area parallel addition gives $J_{\mathrm{mix}}=(J_1+J_2)/2$. Two-point secants of $J_{\mathrm{mix}}^{1/2}B^{1/4}$ over 10–14 T and 16–19 T extrapolate to 25.06 and 26.35 T; over 21–24 T only the 28 T component remains and the intercept is 28 T. These are explicitly chosen windows, not regression fits to measured data. Changing the connection to a local series bottleneck, $J_{\mathrm{series}}=\min(J_1,J_2)$, instead makes the 20 T component limiting. The example does not model current transfer, percolation or a real wire network.

Separately, Godeke considers a hypothetical uniform wire with grain size and grain-boundary density retained. Holding the prefactor fixed, the printed 24.9-to-26.9 T inputs give 2888.01 A/mm² from 2250 A/mm² at 12 T and 4.2 K, versus the paper's 2883 A/mm². If each printed field was rounded to the nearest 0.1 T, the resulting interval is approximately 2852–2924 A/mm². Compatibility with that assumed rounding does not establish the cause of the discrepancy. Neither improved value was measured. R6 R7

Experimental criteria

Godeke and colleagues used approximately 0.2–0.6 A/mm², normalized to non-Cu area, for resistive measurements of filamentary Nb3Sn in fields up to 30 T. The 10%, 50% and 90% criteria refer to the fitted normal-state resistance background, not material volume fractions. At noisy high fields, the 99% value was extrapolated from the 10–90% region while assuming comparable transition widths at low and high fields. The unresolved interlaboratory offset was about 4%; the authors treated the data as reliable within roughly ±2%, without establishing a statistical confidence interval. These low-current transitions preferentially sample high-quality conducting paths, not the full conductor. R7

The B34 wire reacted for 64 hours at 675 °C provides a concrete comparison: transport at a 10 µV/m criterion gave non-Cu $J_c=2250$ A/mm² at 12 T and 4.2 K. The Kramer field was 24.9 T, while the 99% resistive field was 26.9 T at the same temperature. Keep the non-Cu area normalization, voltage criterion and heat treatment attached to those numbers. They are not interchangeable measures of performance. R7

Gapud and colleagues studied a floating-zone V3Si single crystal, MP3, with residual-resistivity ratio 47. Their 51V NMR experiment used 6.67 T, where the reported superconducting transition was 12.8 K and the structural transition 21.1 K. Cubic and tetragonal spectra coexisted around the structural transition. The authors found no conclusive cubic-phase spectrum below the zero-field transition scale of 17 K; that is distinct from the 12.8 K transition in the actual NMR field. These observations do not establish that a martensitic transition is necessary for every A15 superconductor. R8

Limits and disputes

Elastic softening, anharmonicity and superconductivity can be correlated without a universal rule that moving closer to structural instability always improves $T_c$. Testardi's papers motivate this coupled problem; they do not remove the need for sample-specific structural and thermodynamic evidence. Changes in composition or disorder can alter several electronic and phonon parameters together. Increasing a density-of-states peak or an estimated coupling constant alone is not a validated optimization strategy. R3 R4 R9

The V3Si NMR assignments depend on orientation, spectral resolution and the local electric-field-gradient model. The reported 6.67 T experiment distinguishes its 12.8 K superconducting transition from the zero-field scale of 17 K. Broadening, overlap and Knight shifts complicate the spectra; a missing resolved cubic-phase line does not by itself establish zero cubic volume. A quantitative phase-fraction limit would require a calibrated detection threshold and relative spectral response, which this entry has not established. Similarly, a Nb3Sn resistance criterion, a Kramer extrapolation and a magnetic transition retain different sampling conditions. Record mounting and strain as well as nominal composition. R7 R8

A May 2026 doctoral-dissertation preprint explores Cu–Sn routes to Nb3Sn films for microwave dark-matter cavities. Its abstract describes a chosen cavity tested at 50 mK and 9 T, but reports a quality-factor decline in field. Neither the operating temperature nor the maximum applied field is a measured transition temperature or a demonstrated upper critical field. This is an abstract-level research pointer, not a peer-reviewed performance certification. R10

Discovery implications

Optimize a complete, conditional objective: transition temperature, useful field range, current density, mechanical tolerance and manufacturability. Preserve measured and predicted quantities separately. Compare transition criteria, composition gradients, annealing, strain, grain structure and area normalization before ranking conductors. For a proposed material, check phase stability and the coupling-weighted phonon spectrum; for a working wire or cavity, test the actual operating condition and retain negative results. R5 R6 R9 R10

Related topics

Continue with elemental superconductors, Eliashberg theory, s-wave pairing, critical fields and critical current.

Material cases

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