Superconductivity returns above 0.9 tesla in a two-dimensional oxide interface

The mechanism on offer begins in the band structure. Ab initio calculations by Igor Maznichenko and Sergey Ostanin at Martin Luther University, with Arthur Ernst at Johannes Kepler University, identified a Van Hove singularity at the interface, a momentum region where the electronic density of states runs very high ^[13]. Vitalii Dugaev at Rzeszow University of Technology and Evgeny Ya Sherman at the University of the Basque Country then built a spin-orbit coupling model on top of that calculation and on symmetry arguments ^[14]. With no field applied, the bands are symmetric under p to -p, so electrons of opposite momentum sit at the same energy, a pair drawn from them carries zero total momentum, and spin-singlet pairing forms easily ^[15]. Turn a field on, that symmetry goes, pairing efficiency drops, and the onset temperature falls with it ^[16].

That is the half of the model that explains the death, and it is the half I can trace. The account available to me stops short of how the zero-resistance state returns above 0.9 T, so that part goes untold here. What I can report is the shortlist the researchers weighed: Zeeman splitting, orbital effects, and the Jaccarino-Peter effect, in which internal and external magnetic fields compensate and cancel out ^[12].

What carries weight here is the stubbornness of that field, more than its 0.9 T value. B* did not move with temperature, and it did not move when the carrier concentration was tuned by a gate voltage across the interface ^[6]. Those are the first two knobs an experimentalist reaches for, and a feature that ignores both looks more like something set in the band structure than something set by thermal or doping energy scales. That is a reading rather than a result, and it is why the Van Hove singularity is doing work in the explanation ^[13].

The reason to want this in two dimensions is the upper critical field. In MoS2, band structure and spin-orbit coupling hold electron spins perpendicular to the sample surface, and that locking greatly raises the maximum field superconductivity can survive ^[10]. In conventional superconductors, field suppresses superconductivity either by breaking Cooper pairs or by admitting flux whose motion introduces resistance ^[9], so a 2D system that blunts the first channel is worth having. Superconductivity had already been seen at KTaO3 interfaces, which is why the team grew LaTiO3 on top of it for the lattice match ^[11]. The caution this result adds is that a resistive band now sits between two zero-resistance ones ^[5], and one ceiling number does not describe that.

Five institutions in five countries were needed to read one cusp: RIKEN in Japan, Martin Luther in Germany, Johannes Kepler in Austria, Rzeszow in Poland, and the University of the Basque Country in Spain ^[17]. A transport peak was grown in one country and its band structure computed in two others. That division of labour is the measure of how loosely the 2D ferromagnetic and superconducting competition was pinned down ^[8]. A group sweeping critical field against temperature had no reason to expect anything at 0.9 T, and Denis Maryenko says the sighting was extremely surprising ^[3].