Evidence of multiple nodeless energy gaps in superconducting ${\mathrm{Ba}}_{0.6}{\mathrm{K}}_{0.4}{\mathrm{Fe}}_2{\mathrm{As}}_2$ single crystals from scanning tunneling spectroscopy

We report on low-temperature scanning tunneling microscopy and spectroscopy (STM) studies of the electronic structure of single-crystalline ${\mathrm{Ba}}_{0.6}{\mathrm{K}}_{0.4}{\mathrm{Fe}}_2{\mathrm{As}}_2$. Multiple superconducting gaps are observed in the density of states (DOS) and the sizes of the two dominant gaps ${\Delta }_L$ and ${\Delta }_S$ are 7.6 and 3.3 meV, respectively. The flat bottom of the DOS spectra near zero bias indicates the nodeless feature of the gaps, while the global fitting to the spectra definitely requires the anisotropy. The nodeless gaps with finite anisotropy revealed in our STM data agree well with the expectations of an extended $s$-wave superconductivity.

Figure 1

(a) Temperature dependence of the ac susceptibility of a Ba${}_{0.6}$K${}_{0.4}$Fe${}_2$As${}_2$ single crystal used in this work. The measurements were performed with an Oxford Maglab-Exa-12 system with an ac field of 0.1 Oe and an oscillation frequency of 133 Hz. The half-height width of the imaginary component is about 1 K. (b) $133\hspace{0.5em}\text{nm}\times 133$ nm topographic STM image of the Ba${}_{0.6}$K${}_{0.4}$Fe${}_2$As${}_2$ single crystal cold-cleaved in situ. It was taken with a sample-tip voltage of 50 mV and tunneling current of 200 pA.

Figure 2

(a) Zoom in on a $17\hspace{0.5em}\text{nm}\times 17$ nm area within Fig. 1b, which is taken at 3 K. (b) Spatially resolved spectra of $\mathit{dI}/\mathit{dV}$ versus $V$ recorded along the trajectory (white line) indicated in (a) with a spacing interval of 2.7 Å. The four vertical lines are to guide the eye. (c), (d) Maps of the superconducting gap determined in the energy ranges below and above 6 meV, respectively.

Figure 3

(a) Histogram of the obtained superconducting gaps. The fit to a multipeak Gaussian function is also plotted here with a list of the fitting parameters. (b) Some spectra measured at different locations on the sample surface. Both coherence peaks of ${\Delta }_L$ and the features of a larger gap (broad peaks or humps) are indicated by short black bars. A slight kink feature contributed from ${\Delta }_S$ can still be distinguished, although for clarity it is not marked here.

Figure 4

Left column: Typical spectra measured at different positions on the sample surface (black dots). The curves are the fits to the two-band model discussed in the text. Wt.(${\Delta }_S$) means the spectral weight contributed from ${\Delta }_S$, which is equal to the parameter $\sigma$ mentioned in the text. Right column: Gap functions for the two gaps assumed in the fitting. It was noted that the adaptive gaps in these simulations should be nodeless and have a small anisotropy. The calculated coherence peaks for the larger gap (${\Delta }_L$) look more narrow than the experimental data, which may be due to the neglect of the contribution from the $~$10 meV gap, as illustrated in Fig. 3b.