Efficiency of quasiparticle evacuation in superconducting devices

The diffusion of excess quasiparticles in a current-biased superconductor strip in proximity to a metallic trap junction is studied. In particular, we have measured accurately the superconductor temperature at a near-gap injection voltage. By analyzing our data quantitatively, we provide a full description of the spatial distribution of excess quasiparticles in the superconductor. We show that a metallic trap junction contributes significantly to the evacuation of excess quasiparticles.

Figure 1

(a) Schematic of the sample design. The test, injector, and trap junctions have a tunnel barrier in between the normal-metal Cu and the superconductor Al electrodes. (b) Spatial profile of the effective superconductor temperature $T_{\text{S}}$ calculated for sample parameters at a bath temperature $T_{\text{bath}}=500$ mK. The two injectors biased at $eV/\Delta =4$ are in contact with the superconducting strip of length $L_{\text{S}}=27$ $\mu$m, and at a distance $a_1=5$ $\mu$m or $a_2=21$ $\mu$m from the test junction. (c) Scanning electron micrograph of a complete device with three injector junctions.

Figure 2

(a) Probe junction differential conductance under different injection bias voltages from the S-I-N-I-S junction. (b) Low-bias differential conductance (on a log scale) of a S-I-N-I-S test junction (solid black lines) at different injector-2 bias voltages, compared to calculated isotherms at different $T_{\text{N}}$ (red dashed lines) obtained from Eq. (1) with $\Delta =0.22$ meV. Differential conductance $dI/dV$ has been normalized by multiplying by the normal-state resistance $R_{\text{N}}$.

Figure 3

(a) Central N-metal electronic temperature $T_{\text{N}}$ dependence on injector-2 bias voltage at a bath temperature of 304 mK. The error in $T_{\text{N}}$ is smaller than the symbol size. Inset: Calibration of the extracted superconductor temperature $T_{\text{S}}$ to the measured normal-metal temperature $T_{\text{N}}$. The gray area shows the uncertainty for different values of the Kapitza coupling coefficient. (b) Corresponding extracted superconductor temperature $T_{\text{S}}$ at $x=0$. The dotted lines are fits using $D_{\text{qp}}=35\times {10}^{-4}$ m${}^2$/s.

Figure 4

Extracted superconductor temperature $T_S$ as a function of injector bias voltage at a cryostat temperature of 100 mK (squares) and 500 mK (circles). (a) and (b) correspond to an injection from injectors 1 and 2, respectively. The dotted lines show the fits with the parameter $D_{\text{qp}}=35\times {10}^{-4}$ m${}^2$/s. The red/gray and blue/dark gray dashed lines are the calculated curves with the same $D_{\text{qp}}$ value and for ${\tau }_0=12$ ns and $\infty$, respectively.