Title: Rising speed limits for fluxons via edge quality improvement in wide MoSi thin films

Abstract: Ultra-fast vortex motion has recently become a subject of extensive investigations, triggered by the fundamental question regarding the ultimate speed limits for magnetic flux quanta and enhancements of single-photon detectors. In this regard, the current-biased quench of a dynamic flux-flow regime - flux-flow instability (FFI) - has turned into a widely used method for the extraction of information about the relaxation of quasiparticles (unpaired electrons) in the superconductor. However, the large relaxation times $\tau_\epsilon$ deduced from FFI for many superconductors are often inconsistent with the fast relaxation processes implied by their single-photon counting capability. Here, we investigate FFI in $15$ nm-thick $182$ $\mu$m-wide MoSi strips with rough and smooth edges produced by laser etching and milling by a focused ion beam. For the strip with smooth edges we deduce, from the current-voltage ($I$-$V$) curve measurements, a factor of 3 larger critical currents $I_\mathrm{c}$, a factor of 20 higher maximal vortex velocities of 20 km/s, and a factor of 40 shorter $\tau_\epsilon$. We argue that for the deduction of the intrinsic $\tau_\epsilon$ of the material from the $I$-$V$ curves, utmost care should be taken regarding the edge and sample quality and such a deduction is justified only if the field dependence of $I_\mathrm{c}$ points to the dominating edge pinning of vortices.