Emission and Absorption Asymmetry in the Quantum Noise of a Josephson Junction

We measure current fluctuations of mesoscopic devices in the quantum regime, when the frequency is of the order of or higher than the applied voltage or temperature. Detection is designed to probe separately the absorption and emission contributions of current fluctuations, i.e. the positive and negative frequencies of the Fourier transformed nonsymmetrized noise correlator. It relies on measuring the quasiparticles photon assisted tunneling current across a superconductor-insulator-superconductor junction (the detector junction) caused by the excess current fluctuations generated by quasiparticles tunneling across a Josephson junction (the source junction). We demonstrate unambiguously that the negative and positive frequency parts of the nonsymmetrized noise correlator are separately detected and that the excess current fluctuations of a voltage biased Josephson junction present a strong asymmetry between emission and absorption.

Figure 1

(a) Calculated $I(V)$ of the detector under monochromatic irradiation (black curve without irradiation), with frequency indicated on the figure and high enough amplitude to have a visible PAT current, and schematic pictures of the tunneling process involved in the PAT current through the detector. Left: $|V_D|<2\Delta /e$, only emission by the source can lead to PAT. Right: $|V_D|>2\Delta /e$ the detector is mainly sensitive to absorption by the source. (b) Optical picture of the sample, showing the two junctions and the on-chip circuitry. (c) Equivalent circuit of the sample, with $R=750\Omega$, $C_C\approx 750\mathrm{fF}$.

Figure 2

(a), (b) PAT current at different detector bias (indicated on the figure) due to the ac Josephson effect ($e{V}_S<\Delta$). The curves are shifted vertically for clarity. (a) corresponds to the emission part ($V_D<2\Delta /e$) and (b) to the absorption part ($V_D>2\Delta /e$). Both exhibit a signature of the ac Josephson effect at the Josephson frequency $2e{V}_S/h$ (indicated by the dot-dashed lines). The peak near $V_S=0$ in absorption is due to the variation of the source impedance on the Josephson branch. The dashed lines are calculated curves with the transimpedance as adjustable parameter. (c) Transimpedance extracted from the data.

Figure 3

PAT current at different bias of the detector (indicated on the figure) when the source is polarized in the vicinity of $2\Delta$. The curves are shifted vertically for clarity. (a) Emission part ($e{V}_D<2\Delta$): there is a singularity at frequency $(e{V}_S-2\Delta )/h$ for $e{V}_S>2\Delta$ related to the emission of the source (indicated by the dot-dashed line). (b) Absorption part ($e{V}_D>2\Delta$): there are two singularities, one at $2\Delta -e{V}_S$ for $e{V}_S<2\Delta$ (line A) and one at $e{V}_S-2\Delta$ for $e{V}_S>2\Delta$ (line B). They are related to the absorption spectrum of the source and to a change of the environment impedance (see text). The dashed lines are simulations of the expected signal taking into account the measured value of $Z(\omega )$.

Figure 4

(a) Predicted nonsymmetrized current fluctuations of the quasiparticle current of a Josephson junction at low temperature for different bias conditions. The highest frequency the experiment is sensitive to is $2\Delta /h=115\mathrm{GHz}$. Inset: corresponding symmetrized current fluctuations. (b) Illustration of the tunneling process leading to the singularity in the emission noise ($e{V}_S>2\Delta$). (c) Tunneling processes for singularities in absorption noise.