Quantized magnetic vortices driven by electric current determine key electromagnetic properties of superconductors. While the dynamic behavior of slow vortices has been thoroughly investigated, the physics of ultrafast vortices under strong currents remains largely unexplored. Here, we use a nanoscale scanning superconducting quantum interference device to image vortices penetrating into a superconducting Pb film at rates of tens of GHz and moving with velocities of up to tens of km/s, which are not only much larger than the speed of sound but also exceed the pair-breaking speed limit of superconducting condensate. These experiments reveal formation of mesoscopic vortex channels which undergo cascades of bifurcations as the current and magnetic field increase. Our numerical simulations predict metamorphosis of fast Abrikosov vortices into mixed Abrikosov-Josephson vortices at even higher velocities. This work offers an insight into the fundamental physics of dynamic vortex states of superconductors at high current densities, crucial for many applications.

In the framework of the time-dependent Ginzburg-Landau formalism, we study the "resistive" state of a submicron superconducting stripe in the presence of a longitudinal current. Sufficiently strong current leads to phase slippage between the leads, which is manifested as oppositely charged kinematic vortices moving in opposite directions perpendicular to applied drive. Depending on the distribution of superconducting current density the vortex-antivortex either nucleate in the middle of the stripe and are expelled laterally or enter on opposite sides of the sample and are driven together to annihilation. We distinguish between the two scenarios as a function of relevant parameters and show how the creation/annihilation point of the vortex-antivortex and their individual velocity can be manipulated by applied magnetic field and current.

International audienceSuperconducting correlations may propagate between two superconductors separated by a tiny insulating or metallic barrier, allowing a dissipationless electric current to flow(1,2). In the presence of a magnetic field, the maximum supercurrent oscillates(3) and each oscillation corresponding to the entry of one Josephson vortex into the barrier(4). Josephson vortices are conceptual blocks of advanced quantum devices such as coherent terahertz generators(5) or qubits for quantum computing(6), in which on-demand generation and control is crucial. Here, we map superconducting correlations inside proximity Josephson junctions(7) using scanning tunnelling microscopy. Unexpectedly, we find that such Josephson vortices have real cores, in which the proximity gap is locally suppressed and the normal state recovered. By following the Josephson vortex formation and evolution we demonstrate that they originate from quantum interference of Andreev quasiparticles(8), and that the phase portraits of the two superconducting quantum condensates at edges of the junction decide their generation, shape, spatial extent and arrangement. Our observation opens a pathway towards the generation and control of Josephson vortices by applying supercurrents through the superconducting leads of the junctions, that is, by purely electrical means without any need for a magnetic field, which is a crucial step towards high-density on-chip integration of superconducting quantum devices

Recent observation of unusual vortex patterns in MgB(2) single crystals raised speculations about possible "type-1.5" superconductivity in two-band materials, mixing the properties of both type-I and type-II superconductors. However, the strict application of the standard two-band Ginzburg-Landau (GL) theory results in simply proportional order parameters of the two bands-and does not support the "type-1.5" behavior. Here we derive the extended GL formalism (accounting all terms of the next order over the small τ=1-T/T(c) parameter) for a two-band clean s-wave superconductor and show that the two condensates generally have different spatial scales, with the difference disappearing only in the limit T→T(c). The extended version of the two-band GL formalism improves the validity of GL theory below T(c) and suggests revisiting the earlier calculations based on the standard model.