Vibrations could flip the switch on future superconducting devices

Vibrations are a constant presence in the physical world. Atoms and molecules are constantly vibrating. The more something vibrates, the higher its temperature; and vibrations generated in a solid, or crystal structure, spread at the speed of sound.

Vibrations in solids — specifically, their smallest quantum unit, called phonons — are at the center of our latest paper,¹ published in Nature Electronics. We shed light on a novel type of superconducting switching device that could prove useful in next-generation components, for both classical and quantum computers.

Our findings offer an explanation for the still-debated microscopic processes behind the functioning of a new superconducting switch, which was first introduced in a 2018 paper.² In the original experiments, the authors found that superconductivity in a metallic nanowire was shut down after a moderate voltage was applied to a nearby gate electrode. In their interpretation, the electric field associated with the voltage was the cause of the superconductivity suppression — an explanation at odds with what textbooks teach us about the way electric fields affect superconductors and metals in general.

In a series of experiments since then, our team and others^{3, 4, 5} have found evidence for an alternative culprit: at the gate voltages where superconductivity was suppressed, weak currents of high-energy electrons leaking out of the gate were always present.

Disrupting the superconducting state

In our latest paper, we provide an explanation of how just a relatively small number of those high-energy electrons are sufficient to disrupt the superconducting state in the nanowire.

We realized a novel device geometry where a current of high energy electrons is generated without exposing the nanowire to electric fields or leakage currents. Instead, currents are produced in the vicinity of the nanowire by additional gate electrodes.

The figure below shows one of such devices. The superconducting nanowire is depicted in blue and the gate electrodes in red. Previous experiments utilized an individual gate electrode placed close to the nanowire. Applying a voltage to such a gate electrode would result in the suppression of superconductivity, as shown in the figure on the right (blue curve). The suppression of superconductivity is correlated to gate leakage currents (black curve).

In our work, we add two more gate electrodes 1 µm away from the nanowire. When a current flows among the two remote gates, the superconductivity in the nanowire is suppressed in a qualitatively similar fashion to when the nearby gate is used. In a first step, the high energy electrons trigger the emission of phonons in the silicon substrate on which both the nanowire and the gates sit. Like water waves, these phonons travel a considerable distance (several microns) and reach the superconducting nanowire, where they efficiently suppress superconductivity.

To prove the presence of these phonon waves, we etched a deep trench into the silicon substrate between the gates and the nanowire. The trench reflects the phonon waves traveling in the silicon substrate, limiting their impact on superconductivity.

It takes a nearly unmeasurable current — a few, albeit high-energy electrons — for the switching to work, possibly explaining why previous interpretations didn’t assign any significance to this leakage current and pointed to the electric field’s voltage as the cause instead.

Our findings, and specifically the role we now show phonons play, are critically important when designing devices for superconducting quantum applications. In fact, switches based on this effect would require very little power to work, and as such could be far more efficient than other methods, such as warming up a nanowire with resistive heating.

In addition, our results suggest new strategies that could be used to mitigate the well-known negative impact of phonons on the coherence of superconducting qubits. A way to circumvent that detrimental effect of phonons could be by structuring the substrate — like we did with the trench — to attenuate, or confine, the impact of the emitted phonons.

We hope the conclusions of our paper will serve as a starting point for a deeper theoretical understanding of the intricate out-of-equilibrium physics at play, and serve as a stimulus for radically different device designs where superconducting switches are activated by few high-energy phonons.