Topological Weyl superconductor to diffusive thermal Hall metal crossover in the B phase of UPt3
Goswami, Pallab; Nevidomskyy, Andriy H.
The recent phase-sensitive measurements in the superconducting B phase of UPt3 provide strong evidence for the triplet, chiral kz(kx±iky)2 pairing symmetries, which endow the Cooper pairs with orbital angular momentum projections Lz=±2 along the c axis. In the absence of disorder such pairing can support both line and point nodes, and both types of nodal quasiparticles exhibit nontrivial topology in the momentum space. The point nodes, located at the intersections of the closed Fermi surfaces with the c axis, act as the double monopoles and the antimonopoles of the Berry curvature, and generalize the notion of Weyl quasiparticles. Consequently, the B phase should support an anomalous thermal Hall effect, the polar Kerr effect, in addition to the protected Fermi arcs on the (1,0,0) and the (0,1,0) surfaces. The line node at the Fermi surface equator acts as a vortex loop in the momentum space and gives rise to the zero-energy, dispersionless Andreev bound states on the (0,0,1) surface. At the transition from the B phase to the A phase, the time-reversal symmetry is restored, and only the line node survives inside the A phase. As both line and double-Weyl point nodes possess linearly vanishing density of states, we show that weak disorder acts as a marginally relevant perturbation. Consequently, an infinitesimal amount of disorder destroys the ballistic quasiparticle pole, while giving rise to a diffusive phase with a finite density of states at the zero energy. The resulting diffusive phase exhibits T-linear specific heat, and an anomalous thermal Hall effect. We predict that the low-temperature thermodynamic and transport properties display a crossover between a ballistic thermal Hall semimetal and a diffusive thermal Hall metal. By contrast, the diffusive phase obtained from a time-reversal-invariant pairing exhibits only the T-linear specific heat without any anomalous thermal Hall effect.