Andreev reflection at the normal-metal/heavy-fermion superconductor $\mathrm{Ce}\mathrm{Co}{\mathrm{In}}_5$ interface

The dynamic conductance across nanoscale junctions consisting of a normal-metal (N) and a heavy-fermion superconductor (HFS) is measured over a wide temperature range $\left(60\mathrm{K}\text{to}400\mathrm{mK}\right)$. The N/HFS $(\mathrm{Au}∕\mathrm{Ce}\mathrm{Co}{\mathrm{In}}_5)$ contact is shown to be in the Sharvin limit. The background conductance develops a gradual asymmetry with decreasing temperature starting at the heavy-fermion liquid coherence temperature, $T^{*}\sim 45\mathrm{K}$, to the onset of superconducting coherence, $T_{\mathrm{c}}=2.3\mathrm{K}$. The enhanced subgap conductance observed below $T_{\mathrm{c}}$ arises from Andreev reflection. This enhancement is an order of magnitude smaller ($\sim 13.3\%$ at $400\mathrm{mK}$) than that observed for N/conventional superconductors but consistent with other N/HFS data reported. Attempts to fit to the full conductance curve as a function of temperature with extended Blonder-Tinkham-Klapwijk models, including those that account for the breakdown of the Andreev approximation and renormalizations of the Fermi momenta, clearly show that existing models cannot account for our data. We provide a theoretical framework for understanding the N/HFS Andreev conversion process.

Figure 1

Dynamic conductance spectra of a $\mathrm{Au}∕\mathrm{Ce}\mathrm{Co}{\mathrm{In}}_5$ point contact between $60\mathrm{K}$ and $400\mathrm{mK}$. Curves are shifted vertically by 0.05 for clarity.

Figure 2

(a) Fitted curves (solid lines) to normalized conductance spectra (symbols) using an $s$-wave BTK model. Curves are shifted for clarity. (b) Fitting parameters, $\Delta$ and $\Gamma$, along with BCS energy gap (solid line) and $\Gamma$ (dashed line) used for ZBC fitting in Fig. 4.

Figure 3

Comparison of best fit curves to the conductance at $400\mathrm{mK}$. Dotted (solid) curve is for $s\left(d\right)$-wave EBTK fitting.

Figure 4

Comparison of best fit curves to the ZBC vs. temperature data. Dotted (solid) curve is for $s\left(d\right)$-wave fitting.