Theory of Josephson scanning microscopy with $s$-wave tip on unconventional superconducting surface: application to Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$

Abstract

Josephson scanning tunneling microscopy (JSTM) is a powerful probe of the local superconducting order parameter, but studies have been largely limited to cases where superconducting sample and superconducting tip both have the same gap symmetry -- either s-wave or d-wave. It has been generally assumed that in an ideal $s$-to-$d$ JSTM experiment the critical current would vanish everywhere, as expected for ideal $c$-axis planar junctions. We show here that this is not the case. Employing first-principles Wannier functions for Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$, we develop a scheme to compute Josephson critical current ($I_{c}$) and quasiparticle tunneling current measured by JSTM with sub-angstrom resolution. We demonstrate that the critical current for tunneling between an s-wave tip and a superconducting cuprate sample has largest magnitude above O sites and it vanishes above Cu sites. $I_{c}$ changes sign under $\pi/2$-rotation and its average over a unit cell vanishes, as a direct consequence of the $d$-wave gap symmetry in cuprates. Further, we show that $I_{c}$ is strongly suppressed in the close vicinity of a Zn-like impurity owing to suppression of the superconducting order parameter. More interestingly, $I_{c}$ acquires non-vanishing values above the Cu sites near the impurity. The critical current modulations produced by the impurity occur at characteristic wavevectors distinct from the quasiparticle interference (QPI) analogue. Furthermore, the quasiparticle tunneling spectra in the JSTM set-up shows coherence peaks and impurity-induced resonances shifted by the $s$-wave tip gap. We discuss similarities and differences in JSTM observables and conventional STM observables, making specific predictions that can be tested in future JSTM experiments.