Is UTe2 a spin-triplet superconductor?

Scope

UTe2 has substantial evidence supporting odd-parity, triplet-related superconductivity, but its complete order parameter is not fixed by a single experiment. The question must specify the sample, pressure, field magnitude and direction, and superconducting phase. Spin structure, inversion parity, the number of order-parameter components, time-reversal symmetry and topology are distinct properties. This page compares selected evidence rather than declaring all of them settled together. R2 R3 R7

In a suitable intraband description of a centrosymmetric, time-reversal-invariant normal state, an even-frequency odd-parity gap has a symmetric band-doublet pairing component. With strong spin–orbit coupling, that label is not automatically a pair of freely rotating bare electron spins. The doublet's actual crystal transformation rules matter; exceptional bands need not transform as ordinary spin-1/2 pseudospins. Multiorbital and frequency-exchange assumptions must remain explicit. R1

Experimental criteria

The discovery study reported a large, anisotropic upper critical field and triplet-related interpretations. Its b-axis estimate near 40 T extrapolated beyond measurements reaching 20 T. Large critical fields motivate a triplet hypothesis, but the comparison must include orbital limiting, strong coupling, multiband structure and spin–orbit response. UTe2's paramagnetic ambient-pressure normal state also distinguishes it from uranium superconductors with established ferromagnetic order. R2

Li and colleagues' 2025 In/UTe2 junction study probes orbital symmetry through orientation-dependent Josephson coupling. Coupling appeared on the (001) and (0−11) faces, whereas two junctions on a (1−10) surface had no resolved supercurrent at a stated sensitivity of roughly 1 microampere. The authors identify this pattern with B1u among their considered representations. Their argument requires sufficiently smooth interfaces and first-order coupling; interface barriers, geometry and surface reconstruction remain part of the inference. A null signal is resolution bounded, not an exact measurement of zero. R3

Junction temperatures also need careful attribution. In that study, sample E's junction zero-resistance temperature is about 1.2 K, while the UTe2 electrode onset is about 2.0 K. These are different criteria, not two interchangeable bulk transition temperatures. This question page does not create a new material record from either value. R3

What Knight shifts establish

Kitagawa and colleagues re-examined an early sample using a small field tilt to infer the difficult a-axis response. They found a reduction and attributed the earlier nearly unchanged signal to a remaining nonsuperconducting fraction. The method uses an angular projection and a diamagnetic estimate; the proposed recovery near a 1.5 T a-axis component is an extrapolation interpreted as spin rotation. It should not be presented as a directly mapped universal rotation boundary. R5

Matsumura and colleagues measured two Te sites at approximately 0.8 T along b and 0.75 T along c, down to about 70 mK. Subtracting their shifts removes an assumed common diamagnetic contribution. The c-axis peaks overlap, and its quantitative subtraction is particularly sensitive to fitting and nearly equal hyperfine slopes. Together with earlier a-axis measurements, the work constrains a three-component d-vector description. The authors nevertheless allow higher-order momentum terms and do not uniquely distinguish a full gap from every point-node model. R4

Mathematical assumptions

In a specified spin-1/2 basis, a pure triplet example has $\Delta=(\mathbf d\cdot\boldsymbol\sigma)i\sigma_y$. Real odd $\mathbf d(\mathbf k)$ gives a symmetric spin matrix and satisfies $\Delta(\mathbf k)=-\Delta^{\mathsf T}(-\mathbf k)$. This algebra must not replace the actual symmetry matrices of a spin–orbit-coupled material. R1

For an original illustrative single-band, real-unitary, fixed-d model with isotropic spin coupling, constant density of states and zero temperature, the normalized weak-field response is

$$\frac{\chi_{ii}}{\chi_N}=1-\left\langle\frac{d_i^2}{|\mathbf d|^2}\right\rangle_{\rm FS}.$$

Direct Bogoliubov-matrix energy curvatures verify 12 directional cases using two small field steps and a finite energy cutoff. Four separate spherical averages illustrate the distinction: $\mathbf d=\mathbf k$ gives $2/3$ on all axes, whereas $(k_y,k_x,3k_xk_yk_z)$ gives approximately $(0.558,0.558,0.885)$. Both reduce all three responses, although their symmetry and nodal structures differ. These examples do not calculate UTe2's multiband susceptibility, spin–orbit matrix elements, finite-temperature response or field-induced rotation.

For the two-site extraction, write $K_j=A_j\chi+K_j^{\rm orb}+K^{\rm dia}$. Subtraction cancels only the common term. Solving for the diamagnetic contribution introduces a denominator $\Gamma-1$, where $\Gamma$ relates hyperfine slopes. The selected experiment reports ratios near 1.32 and 0.91 for different directions. Four independent synthetic examples recover the imposed common shift and verify amplification of measurement errors as the slopes approach equality; they are not a refit of the spectra. R4

Gap structure and other symmetries

Hosoi and colleagues' July 2026 thermal-transport study adds a bulk directional constraint. Its b-axis heat-current measurements reached roughly 50 mK in two clean crystals. Small residual transport and a low-field kink support a deep but finite gap minimum in the authors' interpretation. However, the zero-temperature term is extrapolated, and the zero-field temperature dependence alone explicitly leaves a restricted true-node scenario open. The finite-field model uses a semiclassical Doppler picture and does not capture all subthreshold conduction. Gap anisotropy is not a direct spin measurement. R6

Theuss and colleagues measured compression and shear moduli in single- and double-transition samples. Their absence of resolved shear discontinuities constrains a multicomponent interpretation within the elastic-coupling analysis. The paper explicitly states that this result also permits single-component even-parity states. Its additional B2u proposal is therefore distinct from the component-count constraint. Ambient-pressure findings cannot automatically settle the pressure-induced phases. R8

The reviewed Kerr study found no spontaneous signal in its tested single-transition samples. This constrains time-reversal breaking under those conditions; it does not exclude every possible triplet state. Conversely, the high-field halo experiment maps transport boundaries in a very different field regime, with a multicomponent triplet interpretation remaining a model conclusion. Neither observation alone determines the low-field bulk invariant or establishes usable Majorana modes. R7 R9

Limits and disputes

The evidence supports taking triplet-related states seriously while retaining competing representations and measurement-specific assumptions. A gap minimum, an absent shear jump, a Knight-shift change and a Josephson selection rule do not answer identical questions. Agreement should be assessed on compatible samples and phases, including surface-versus-bulk sensitivity. No independent global fit of these experiments has been performed here, and full scientific acceptance remains open. R3 R4 R6 R8

Discovery implications

Useful next tests would compare orientation-resolved junctions with bulk spin response and thermodynamic probes on well-characterized crystals, carrying uncertainties and sample fractions through the analysis. Candidate models should predict several observables under the same conditions and identify where their predictions differ. Keep claimed topology separate until the relevant gap, symmetry and electronic structure are established. This is a proposed validation program, not a certification of one microscopic pairing mechanism. R1 R3 R4

Related topics

Continue with singlet and triplet pairing, p-wave candidates, heavy-fermion superconductors, and Majorana evidence.

Material cases

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