Figure 1

(a) Optical microscope image of a device with a Cu${}_x$Bi${}_2$Se${}_3$ microcrystal lying on six parallel metal electrodes (35 nm Pd/5 nm Cr) on a SiO${}_2$/Si substrate. The four center electrodes (labeled as 2–5) are 1 μm wide, while the two outside electrodes (labeled as 1 and 6) are 4 μm wide. For this sample, four metal electrodes are electrically connected to the superconductor microcrystal, and the corresponding data are shown in Figs. 1 and 2. (b) Schematic diagram of a superconductor microcrystal on top of three normal metal electrodes and a circuit for obtaining Andreev reflection spectra for the target N-S junction between the $I_{+}$ ($V_{+}$) electrode and the superconductor. (c)–(e) Normalized differential conductance $dI$/$dV$ (symbols) versus bias voltage $V$ at temperature $T$ = 240 mK and magnetic field $B$ = 0 for three N-S junctions and the fitting to the data (solid lines) by the BTK theory (see main text). The fitting parameters are: (c) Δ = 0.32 meV, Γ = 0.21 meV, $w$ = 61$\%$, and $Z$ = 0.7; (d) Δ = 0.35 meV, Γ = 0.13 meV, $w$ = 13$\%$, and $Z$ = 0.45; (e) Δ = 0.39 meV, Γ = 0.27 meV, $w$ = 80$\%$, and $Z$ = 0.38. The data of (c) and (d) are from the N-S junctions at electrode #2 and #3, respectively, for the sample shown in (a), and the data of (e) is from a different sample. The normalized $dI$/$dV$ is obtained by dividing the measured $dI$/$dV$ data at $T$ = 240 mK by appropriate normal-state data at $T$ above bulk $T_c$. For the data of (e), the normal state is not reached within the measured bias-voltage range, but can be determined from the temperature dependence of AR spectra (see Fig. 6 of Ref. 20).Reuse & Permissions

Figure 2

$dI$/$dV$ vs $V$ at $T$ = 240 mK and $B$ = 0 for three N-S junctions with smaller normal state resistance $R_N$. The junction of (a) is from the same sample as the junction of Fig. 1e, and the junctions of (b) and (c) are at electrode #1 and #6, respectively, for the sample of Fig. 1a.Reuse & Permissions

Figure 3

$dI$/$dV$ vs $V$ measured under different magnetic fields perpendicular to the substrate at $T$ = 240 mK (top) and at different temperatures with $B$ = 0 (bottom): (a) and (b) for the N-S junction of Fig. 1c, and (c) and (d) for the junction of Fig. 2a.Reuse & Permissions

Figure 4

$dI$/$dV$ vs $V$ measured at $T$ = 240 mK and $B$ = 0 (top); under different magnetic fields perpendicular to the substrate at $T$ = 240 mK (center); and at different temperatures with $B$ = 0 (bottom) for two N-S junctions of the same superconductor microcrystal in a device. These two N-S junctions were not electrically connected initially, but electrical connection was established after a short voltage pulse (∼10 V in magnitude) was applied to anneal the contacts.Reuse & Permissions

Figure 5

Magnetic susceptibility as a function of temperature for the bulk material of Cu${}_x$Bi${}_2$Se${}_3$.Reuse & Permissions

Figure 6

$dI$/$dV$ vs $V$ measured (a) under different magnetic fields perpendicular to the substrate at $T$ = 240 mK and (b) at different temperatures with $B$ = 0 for the N-S junction shown in Fig. 1e of the main text.Reuse & Permissions

Figure 7

Comparison between AR spectra in the first measurement [top, replots of Figs. 2b and 2c in the main text] and in the second measurement after a period of six months for the same junctions shown in Figs. 2b and 2c of the main text. The device was preserved in a desiccator during the period between the measurements. The image of the device is shown in Fig. 1a of the main text. In the second measurement, only three electrodes [labeled as #1, #2, and #6 in Fig. 1a] are still electrically connected to the superconductor crystal, which are usable for implementing the measurement by the circuit of Fig. 1b in the main text. The junction shown in the left (right) column is at electrode #1 (#6).Reuse & Permissions

Figure 8

The normalized $dI$/$dV$ data for the junctions shown in Fig. 3 of the main text. For (a) and (b), the raw $dI$/$dV$ data are divided by adjusted normal-state data, which are obtained by offset the raw $dI$/$dV$ spectrum at $T$ = 4.3 K to match the high-bias $dI$/$dV$ spectrum at $T$ = 240 mK. For (c) and (d), the raw $dI$/$dV$ data are divided directly by the raw $dI$/$dV$ spectrum at $T$ = 4.5 K.Reuse & Permissions

Figure 9

$dI$/$dV$ vs $V$ measured under different magnetic fields perpendicular to the substrate at $T$ = 240 mK (left column) and at different temperatures with $B$ = 0 (right column) for other two junctions in the same device of Fig. 4 in the main text. These two junctions were electrically connected initially without the need of voltage annealing, and they are located at the two outermost electrodes (i.e. #1 and #6) similar to the pattern shown in Fig. 1a of the main text. With four electrically connected junctions in this device, we also obtained the four-terminal resistance of the superconductor microcrystal to be 0.2 Ω above $T_c$ (typical for other devices), which is negligible compared with the junction normal-state resistance $R_N$.Reuse & Permissions