Observation of odd-parity superconductivity in UTe2
Symmetry properties of the order parameter are among the most fundamental characteristics of a superconductor. The pairing symmetry of recently discovered heavy fermion superconductor UTe2 featuring an exceedingly large upper critical field has attracted a great deal of attention. Even though it is widely believed that UTe2 possesses an odd-parity, spin-triplet pairing symmetry, direct evidence for it is lacking, especially at zero or low magnetic fields. We report here the selection-rule results of Josephson coupling between In, an s-wave superconductor, and UTe2. The orientation dependence of the Josephson coupling suggests very strongly that UTe2 possess an odd-parity pairing state of B_1u in zero magnetic fields. We also report the formation of Andreev surface bound states on the (1-10) surface of UTe2.
💡 Research Summary
UTe₂, a newly discovered heavy‑fermion superconductor with a modest transition temperature (Tc ≈ 2 K) but an exceptionally large upper critical field (Hc₂ ≈ 20 T along the b‑axis), has long been suspected of hosting odd‑parity, spin‑triplet pairing. However, direct evidence for such a pairing symmetry at zero or low magnetic fields has been lacking, because most experimental probes (Knight shift, 1/T₁, specific heat) can be interpreted within both singlet and triplet scenarios, especially given the strong spin‑orbit coupling (SOC) expected in this material.
The authors address this gap by exploiting the selection‑rule physics of Josephson tunneling between a conventional s‑wave superconductor (indium, In) and UTe₂. In a crystal with inversion symmetry, the order‑parameter (OP) must be either even‑parity spin‑singlet or odd‑parity spin‑triplet (finite‑momentum or odd‑frequency pairing being excluded). The D₂h point group of UTe₂ (space group Immm) admits several irreducible representations (irreps). For each irrep, the authors calculate the expected Josephson current density Jₛ on three crystallographic planes: (001), (1‑10) and (0‑11). Crucially, the odd‑parity B₁ᵤ (Γ⁻²) representation predicts a finite first‑order Josephson coupling on the (001) surface but a vanishing coupling on the (1‑10) surface, owing to the SOC‑mediated term Jₛ ∝ Im⟨Δₛ* d̂(k)·(k̂ × n̂)⟩.
Experimentally, the team prepared Josephson junctions by pressing a small In dot onto freshly cleaved UTe₂ surfaces. The natural oxide layer (InOₓ) serves as the tunnel barrier. After low‑temperature measurements, Laue diffraction identified the orientation of each junction. Three junctions (A1, A2, B) were fabricated on (001) faces, while two junctions (C1, C2) were made on (1‑10) faces.
On the (001) junctions, current–voltage (I‑V) characteristics displayed clear zero‑voltage supercurrents up to the bulk Tc of UTe₂ (≈ 1.6 K). The critical current Ic decreased smoothly with temperature, and the product Ic Rn (Rn being the normal‑state resistance) was found to be 4 %–15 % of the Ambegaokar‑Baratoff (AB) limit for two s‑wave superconductors, consistent with a first‑order Josephson coupling mediated by SOC. In contrast, the (1‑10) junctions showed no detectable supercurrent even when the measurement current was reduced to 4 µA, implying Ic ≈ 0 and an Ic Rn value at least two orders of magnitude smaller than the AB expectation. Since the normal‑state resistances of all junctions are comparable, the stark difference cannot be ascribed to junction quality; it must stem from symmetry‑based selection rules.
These observations single‑out the B₁ᵤ (Γ⁻²) odd‑parity state (or a mixture of Γ⁻¹ and Γ⁻² in the case of accidental degeneracy) as the only OP compatible with the data. The authors discuss that if an accidental near‑degeneracy between Γ⁻¹ (A₁ᵤ) and Γ⁻² (B₁ᵤ) exists, the dominant component could be B₁ᵤ, while a subdominant A₁ᵤ admixture would preserve the observed selection rule.
Beyond Josephson coupling, the study reports tunneling spectroscopy on the (1‑10) surface. Differential conductance dI/dV curves reveal a pronounced zero‑bias conductance peak (ZBCP) that emerges sharply below the UTe₂ Tc, characteristic of Andreev surface bound states (ASBS). Theoretical considerations show that for the B₁ᵤ OP, quasiparticles reflecting off the (1‑10) surface experience a sign change of the order parameter along their classical trajectory, leading to zero‑energy bound states. No such ZBCP is observed on the (001) surface, consistent with the absence of a sign change there.
The combined evidence—direction‑dependent Josephson coupling and the presence of ASBS on the (1‑10) face—provides compelling, direct confirmation that UTe₂ hosts an odd‑parity, likely spin‑triplet superconducting state of B₁ᵤ symmetry at zero magnetic field. This resolves a long‑standing debate about the pairing symmetry, aligns with prior indirect probes, and has profound implications for the material’s topological properties, potential Majorana modes, and suitability for spin‑polarized quantum devices. Future work may explore the detailed gap structure (point nodes versus fully gapped), the role of possible accidental degeneracy, and the manipulation of surface bound states for quantum information applications.
Comments & Academic Discussion
Loading comments...
Leave a Comment