Operating superconducting channels by electron injection

Abstract

The invention relates in particular to a method of operating a superconducting channel. The method relies on an apparatus comprising: a potential superconducting material; a gate electrode; and an electrically insulating medium. The channel is defined by the potential superconducting material. The gate electrode is arranged adjacent to the channel such that an end face of the gate electrode faces a portion of the channel. The electrically insulating medium is arranged in a manner that electrically insulates the gate electrode from the channel. The channel is superconducting by cooling the apparatus. Next, a voltage difference is applied between the gate electrode and the channel to inject electrons into the channel through the electrically insulating medium to generate a gate current between the gate electrode and the channel. The electrons are injected with an average energy sufficient to change the critical current I C of the channel.

Description

BACKGROUND

[0001] The present invention relates generally to techniques for operating a superconducting channel, such as a wire of metal, for example, for switching such a channel or modulating a signal transmitted in the channel, which can be used in particular in quantum processing devices, or for tuning the resonance or cut-off frequency of a circuit comprising such a superconducting channel.

[0002] Due to their low power consumption and high speed, superconducting circuits offer great prospects as building blocks for quantum computing architectures and related cryogenic control electronics. In this context, it is desirable to develop switching devices (e.g. three-terminal devices) that can be electrically tuned between a superconducting state and a resistive state at high frequencies. Ultimately, such devices would enable novel functionalities not present in semiconductor counterparts, such as cryogenic switches, ultra-sensitive detectors, amplifiers, circulators and multiplexers. Several electrically controlled superconducting switches have been proposed, which are based on the injection of unbalanced quasi-particles in Josephson junctions. However, Josephson junctions are typically accompanied by limited source-drain critical currents and the need to operate in a magnetic field-free environment. Therefore, architectures that do not rely on Josephson junctions have been intensively investigated. Such pioneering approaches are based on three- or four-terminal devices in which an electric current, a locally generated Oersted field or heat drives a superconducting channel into a normal state.

[0003] Recent advances in quantum computing make such technologies more relevant to industrial applications. Quantum computing directly exploits quantum mechanical phenomena, such as superposition and entanglement, to perform operations on entangled qubits, i.e. states stored in qubits. For example, superconducting circuits are relatively easy to fabricate with current technology, and are therefore promising candidates for further extending quantum information technology.

[0004] Devices for switching solutions for radio frequency (RF) applications typically have one or more of the following drawbacks: they have a limited resistance in the on-state that leads to signal dissipation and / or attenuation; they have a limited on / off ratio; they are essentially electromechanical, thus leading to long-term heating and slow switching for cryogenic applications; and / or they are macroscopically large and rather expensive. SUMMARY

[0005] According to a first aspect, the invention is embodied as a method of operating a superconducting channel. The method relies on a device comprising: a potentially superconducting material; a gate electrode; and an electrically insulating medium. The channel is defined by the potentially superconducting material. The gate electrode is positioned adjacent to the channel such that an end face of the gate electrode faces a portion of the channel. The electrically insulating medium is arranged in a manner that electrically insulates the gate electrode from the channel. The device is cooled such that the channel becomes superconducting. Next, a voltage difference is applied between the gate electrode and the channel to inject electrons into the channel through the electrically insulating medium, thereby generating a gate current between the gate electrode and the channel. The electrons are injected in a sufficient amount to change a critical current I Cthe average energy injection.

[0006] The proposed solution has additional advantages and can be used for a variety of applications. For example, the present solution allows to adjust both the inductance and the resistance of the superconducting channel. Interestingly, the applied voltage difference can be positive or negative. The gate can in particular be operated with very low power consumption. The most important advantage is that the critical current I C can be adjusted (e.g. in embodiments, within less than 100 ns, down to values close to 1 ps). This in turn allows to achieve very high on / off switching frequencies. Thus, the superconducting device can advantageously be operated as a switch to locally suppress superconductivity (e.g. as a switch) and to modulate a signal through the channel, as in embodiments.

[0007] The device can for example be integrated on an existing quantum chip. When using TiN as superconducting material, the present method works satisfactorily at temperatures equal to or below 3.4 K. It can in particular be applied for signal multiplexing purposes. In other applications, the device is configured as a circuit, e.g. an LC circuit or any circuit in which the superconducting channel is configured as a wire-like inductor. In this way, by changing the inductance of the wire, the resonance frequency of the circuit can be tuned. Similarly, the resonator can form part of a frequency-tunable coupler between qubits of a quantum computer. In further applications, by changing the inductance of the wire, the cut-off frequency of a filter comprising a wire-like inductor is adjusted. For completeness, the underlying device is easy to manufacture, can have very small dimensions, and can be manufactured in an integrated fashion.

[0008] In embodiments, the applied voltage difference can reduce the critical current I C to zero and thereby locally generate a resistance in the channel at least at the level of this part of the channel. This is equivalent to locally quenching the superconductivity at the level of this part of the channel, which can advantageously be used for switching applications.

[0009] As in embodiments, the applied voltage difference can subsequently be switched off to let the potentially superconducting material self-reset to a fully superconducting state. Thus, the resistance of the channel can be locally increased in a controllable way at the level of the part of the channel opposite the electrodes and then reversed.

[0010] Preferably, the method further comprises applying a signal to the channel prior to applying the voltage difference. As a result, the resistance generated in the channel affects the transmission of the signal applied in the channel.

[0011] In preferred embodiments, the method further comprises repeatedly applying and switching off the voltage difference in order to modulate the applied signal. For example, the applied signal can be modulated at a frequency greater than 10 MHz and less than 10 GHz, although any other (lower) frequency can be used.

[0012] The applied voltage difference can be such that the intensity of the resulting gate current is at least ten times smaller than the critical current I C For example, the intensity of the gate current resulting when the voltage difference is applied can be, for example, between 10 pA and 10 nA.

[0013] Preferably, the applied voltage difference can be such that the average energy of the electrons is at least 100 times larger than the superconducting gap of the potential superconductor material.

[0014] In embodiments, the applied voltage difference corresponds to the minimum value required to switch the channel to its normal state. For this value of the applied voltage difference, the impedance of the gate electrode, as determined in particular by the insulating medium, will preferably be larger than 1 GΩ.

[0015] Preferably, the method further comprises varying the intensity of the applied voltage difference in order to controllably grow the region of the channel in which the resistance is generated. More generally, the method can comprise varying the applied voltage difference in order to, for example, adjust the properties of the channel or properties implied by the properties of the channel. In this way, the dynamic inductance L kin of the channel can be significantly adjusted. This can in turn be exploited, for example, to adjust the resonance frequency or cut-off frequency of a circuit of the device, wherein the circuit has an inductor formed by the channel, the channel being configured as a wire. In variants, the device is configured to be connected to a frequency-tunable coupler of a quantum circuit of a quantum processing device, and the voltage difference is applied in order to adjust the resonance frequency of the frequency-tunable coupler and thereby modify the coupling between the quantum circuits.

[0016] According to another aspect, the invention is embodied as a device. The latter comprises a device as described above with reference to the present method. That is, the device in particular comprises a potential superconductor material configured to define a channel (e.g., configured as a wire). The device is adapted to be operated at low temperature in order to, in operation, render the channel superconducting. The device further comprises a gate electrode directed to the channel, so that an end face of the gate electrode faces a portion of the channel, and an electrically insulating medium arranged to electrically insulate the gate electrode from the channel. The device further comprises a control unit connected to the gate electrode. The control unit is configured to apply a voltage difference between the gate electrode and the channel in order to inject electrons into the channel through the electrically insulating medium, thereby generating a gate current between the gate electrode and the channel. The control unit and the device in use allow the voltage difference to be applied to result in an average energy of the injected electrons that is sufficient to, in operation (at low temperature), change the critical current I C of the channel.

[0017] Preferably, the control unit is additionally connected to the channel, and the control unit is further configured to, in operation, both apply a signal to the channel and change the applied signal by modifying the critical current I C of the channel.

[0018] In embodiments, the control unit is further configured to, in operation, allow measuring a property of the channel or a property implied by the property of the channel. At the same time, the control unit allows controllably modifying the applied voltage difference in dependence on the measured property in order to, in operation, adjust said property.

[0019] As described herein, the channel can be configured as a wire. In embodiments, the device further comprises a circuit comprising an inductor comprising or formed by the wire. In this case, the control unit can be further configured to adjust the dynamic inductance L kin of the wire in order to adjust a resonance frequency or a cut-off frequency of the circuit.

[0020] Preferably, the device comprises a substrate forming an electrically insulating medium, and each of the gate electrode and the wire is configured on the substrate. For example, a width of the wire is between 40 nm and 300 nm, and a minimum gap between the end face of the gate electrode and the portion of the channel is between 40 nm and 300 nm. The width and the minimum gap are measured parallel to a main surface of the substrate.

[0021] The present solution can advantageously be used with various superconducting materials. For example, the potential superconducting material can comprise at least one member selected from the group consisting of Ti, TiN, Nb, Ti x Nb y N z , W, WSi, Al, and Mo.

[0022] In preferred embodiments, the device is configured as a quantum processing device, which additionally comprises a quantum circuit, and the channel of the device is connected to one or more quantum circuits. The device can in particular be configured as a switch for switching a radio frequency signal applied to said channel, as a frequency tunable coupler for modifying a coupling between two or more of said quantum circuits, or as a frequency tunable filter to selectively tune a quantum circuit (connected to said frequency tunable filter) to a coupling of the control circuit and the environment (e.g. to enhance a coupling from a control line to a readout resonator and to reduce a coupling of the control line to a transmon type of qubit).

[0023] The method and device embodying the present application will now be described by way of non-limiting examples and with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0024] In the drawings, like reference numerals in all figures indicate like or functionally similar elements, and the inclusion of like reference numerals in the claims is understood to be a recitation of the elements with those reference numerals. FIGS. 1-5 are primarily used to explain the basic embodiments of the various physical principles of the present application, while FIGS. 6-11 illustrate various embodiments and advantages of the present application, all in accordance with the present disclosure.

[0025] Figures 1A-1E Suppression of the critical current in a metallic nanowire is shown. Figure 1A is a scanning electron micrograph (with increased artificial contrast) of a device as involved in the embodiments, with an overlay schematic of the measurement setup. The micrograph shows a TiN nanowire and gate electrode fabricated on a silicon substrate. Figure 1B Differential resistance dV / dI of a nanowire is shown SD as a function of source-drain current I SD , as measured by scanning I SD upwards from very negative values. Critical current I C and re-trapping current I R are indicated. The inset shows the temperature dependence of I C (dot) and I R (squares). Figure 1C Critical current I C of a nanowire as a function of gate voltage V G is plotted. Figure 1D is a plot showing the absolute value of the gate current I G flowing between the gate and the nanowire as a function of V G . A linear component I G ~1TΩ attributed to leakage in the measurement setup is subtracted from the data. Figure 1E is a parametric plot of I C versus I G obtained from the data shown in Figure 1C and 1D .

[0026] Figure 2A and 2B show the dependence on temperature and magnetic field. Figure 2A Critical current I C of a device as a function of gate voltage V G is shown for various temperatures T. Figure 2A . Figure 2B Critical current I C of the same device as a function of out-of-plane magnetic field B ⊥ is shown.

[0027] Figures 3A-3DThis demonstrates the reproducibility of other materials. Figure 3A and 3B The diagram shows the gate voltage V for a Ti line constructed on a Si substrate. G Critical current I of the function C and gate current I G . Figure 3C and 3D The TiN lines on a 25nm SiO2 film thermally grown on a Si substrate are plotted, with the gate voltage V as the reference. G Critical current I of the function C and gate current I G .

[0028] Figures 4A-4D Spatially resolved superconductivity suppression is demonstrated. Figure 4A As used in the implementation plan Figure 1A Scanning electron micrographs of variations of the device (e.g., with artificially added contrast), with overlaid schematic diagrams of the measurement settings. Figure 4B The differential resistance dV is shown to be close to the state of complete suppression of superconductivity. i / dI SD Let i = 1, 2, 3. Then capture the current I. R Indicated by the dashed line. Figure 4C Shown as gate voltage V G1 Critical current I of the function SWi , i = 1, 2, 3, 4, 5. Figure 4D Gate current I G1 The absolute value of the gate voltage V G1 The function relationship graph. Figure 4E Is as I G1 The function of I SWi The parameter curve. Rows represent I. G1 Values, for which inhibition factors are calculated. Figure 4F The inhibition factor S is plotted as a function of the distance between the gate electrode and the nanowire segment. Black markings indicate the inhibition factor S obtained by scanning I. G1 The obtained value, while the gray marking involves scanning I G5 The obtained values. Full and empty markers indicate positive and negative gate polarity, respectively. The solid line represents the fit of an exponential function.

[0029] Figures 5A-5D The quick-switching mechanism and the self-resetting mechanism are shown. In detail, Figure 5A It shows Figure 1A The time-dependent switching characteristics of the device are used as DC gate bias V G The function. A square wave with an interpeak amplitude of 0.5V and a repetition rate of 100kHz is added to V. G . Figure 5B It shows the relationship withFigure 5A The data in the middle is measured simultaneously as V G The function of gate current I G . Figure 5C yes Figure 5A Chinese V G =4.5V (corresponding to) Figure 5A Right side and Figure 5B Line cuts to the data (shown as black marks at the bottom). Figure 5D It is near the rising transformation Figure 5C The data is magnified. The vertical dashed line represents the amplitude transition from 10% to 90%, corresponding to a rise time of 90 ns.

[0030] Figures 6A-6E This relates to an apparatus according to an embodiment. Figure 6A This is a simplified schematic diagram of a superconducting device that forms part of the device. Figure 6B A more practical representation of the superconducting device is shown (top view), where the overlapping schematics correspond to the electronic components involved in the device. Figure 6B The device shown is Figure 1A The devices are the same. Figure 6C yes Figure 6B An enlarged, rotated version of the device shown. Figure 6D The diagram shown illustrates the electron injection mechanism; it must be compared with... Figure 6C Read together Figure 6D . Figure 6E A general architecture of a quantum processing device according to an embodiment is shown, wherein a superconducting device is used as a frequency-tunable coupler connected to a qubit on a qubit platform.

[0031] Figure 7 This is a simplified representation of another superconducting device including a coplanar waveguide resonator, wherein, as in the embodiment, the gate electrode is used to tune the resonant frequency of the resonator.

[0032] Figure 8 This is a simplified representation of another superconducting device, which includes a π-section filter with an inductor as shown in the embodiment, and a gate electrode for adjusting the resonant frequency of the circuit.

[0033] Figures 9-11 This is a flowchart illustrating the higher-order steps of a method for operating a superconducting channel according to an embodiment. Figure 9 The basic steps of the method are shown. Figure 10 and 11 The steps of modulating the signal through the channel and adjusting the characteristics of the superconducting channel (or another characteristic of the device determined by the characteristics of the superconducting channel) are shown respectively.

[0034] The drawings show simplified representations of apparatuses or components thereof as referred to in the examples. The technical features depicted in the drawings are not necessarily drawn to scale. Like or functionally similar elements in the various figures are designated with the same reference signs, unless stated otherwise. DETAILED DESCRIPTION

[0035] For over sixty years, there has been active development of metallic devices capable of switching between a resistive state and a superconducting state with the application of a control signal. However, existing superconducting switches are either large in volume or require large source-drain bias currents and gate currents to operate, making them impractical for many applications and difficult to interface with conventional semiconductor electronics.

[0036] For example, devices such as those mentioned in the background section, whether based on Josephson junctions or alternative designs, typically feature low gate impedance, making large gate currents necessary for operation and drawing significant power from the drive circuit. Furthermore, such devices typically operate with large source-drain bias currents, which suppress self-reset from the normal state to the superconducting state due to Joule heating. Switching the channel back to the superconducting state then requires additional circuitry to periodically drive the source-drain current below the re-trapping current, where the transition to the superconducting state occurs. As shown recently for electrothermal devices, enabling operation at low current bias, and self-reset (from the normal state to the superconducting state) becomes possible when full suppression of the source-drain critical current is achieved.

[0037] The present invention relies on electrically controlled devices based on superconducting channels (e.g., metallic nanowires). Such devices can be used, for example, as switches or related devices, where the normal state of the channel is obtained by injecting high-energy electrons through an insulating barrier from a gate contact. Typically, the operating gate current is orders of magnitude smaller than the nanowire critical source-drain current, effectively producing a voltage-controlled device. The switching mechanism is very fast. In embodiments, the channel can self-reset from its normal state to its superconducting state, and is capable of operating in very large magnetic fields. Such devices can be advantageously used in low-power cryogenic applications, and in particular, in quantum computing applications.

[0038] The structures described below are as follows. First, a general implementation and high-order variants are described (Section 1). The next section is directed to more specific embodiments and investigates the physical basis of the invention (Section 2). The method and its variants are collectively referred to as “the method”. All reference signs S ij The method steps as shown in the flowchart of Figures 9-11 The numerical references refer to physical parts or components of the apparatus.

[0039] 1. General implementation and high-order variants

[0040] Reference is now made to Figures 6-11, which first describe an aspect of the application, which relates to a method of operating a superconducting channel 102, 112, 122. The method relies on (S10) a superconducting device 10, 11, 12, which is designed to operate at cryogenic temperatures. The device can in particular be implemented as a structured multi-layer device 10 (as shown in Figures 6A-6C Fig. 1), a coplanar waveguide resonator (as shown in Figure 7 Fig. 2) or a circuit comprising an inductor (as shown in Figure 8 Fig. 3).

[0041] In all cases, the device comprises a potentially superconducting material, which is structured to define a channel 102, 112, 122, e.g. in the form of a wire. The device further comprises a gate electrode 101, 111, 121. The gate electrode is directed towards the channel 102, 112, 122, whereby an end surface of the gate electrode faces a lateral portion of the channel. The gate electrode typically extends perpendicular to the channel, as shown in Figures 6-8. Furthermore, an electrically insulating medium 103 is provided, which electrically insulates the gate electrode 101, 111, 121 from the channel 102, 112, 122.

[0042] The electrically insulating medium 103 can in particular comprise a portion extending between the end surface of the gate electrode 101, 111, 121 and an opposite portion of the channel 102, 112, 122. That is, in this case, the electrically insulating medium can physically connect the gate electrode to the channel (i.e. it serves as a mechanical, but electrically insulating, bridge), as assumed in Figure 6B 6C and Figure 6B 6C The insulating medium 103 can in particular be provided as a substrate layer 103 supporting the superconducting material 102 and the gate electrode 101 (see

[0043] The device 10, 11, 12 needs to be cooled in S20 in order to render the channel 102, 112, 122 superconducting. Next, in S30, S32, a voltage difference is applied between the gate 101, 111, 121 and the channel 102, 112, 122 in order to inject electrons into the channel through the electrically insulating medium 103. This results in a net (albeit small) gate current between the gate electrode and the channel, the channel thus being grounded. Moreover, the electrons are injected with an average energy sufficient to modify the critical current I C of the channel.

[0044] ​​The injected electrons overcome the tunneling barrier formed by the insulating dielectric 103 through the tunneling effect or through field emission (which also involves tunneling). Note that the operation of the superconducting devices 10, 11, and 12 differs from that of a pure Josephson junction, because Josephson junctions typically operate at zero voltage difference, allowing supercurrents to flow with zero voltage drop. There are cases where a voltage difference can be applied to a Josephson junction. However, this voltage is typically on the order of the superconducting bandgap, whereas the voltage applied here is much larger.

[0045] In this context, the current flow generated when a voltage difference is applied causes a measurable leakage current between the gate electrodes 101, 111, 121 and the channels 102, 112, 122. The intensity of the resulting current depends on the applied voltage difference, typically in an exponential manner. Therefore, the operation of the device differs from simply applying an electric field to the channels.

[0046] Measurements conducted by the inventors show that a relatively small number of electrons injected at energies typically much higher than the superconducting bandgap are sufficient to trigger the generation of a large number of quasiparticles, which affects the critical current density of the superconducting material. At the level of the channel portion opposite the gate electrode, the critical current density changes at least locally. This, in turn, affects the critical current I of the channel. C Critical current I C The maximum current that can pass through channels 102, 112, and 122 without suppressing superconductivity is characterized. In summary, injecting high-energy electrons locally alters the quasiparticle density, which affects the critical current density of the superconducting material (at least locally) and, in turn, changes the critical current of the channel.

[0047] Critical current I C The critical current density can be reduced to zero to produce a non-zero resistance in the channel, as in the embodiments discussed below. If the critical current density is locally reduced to zero, the extent of the resulting normal region (i.e., the non-superconducting region) affects the total resistance of the channel. This phenomenon can be significantly used to locally desuperconductivity and, for example, to modulate electromagnetic signals through channels 102, 112, 122, as in the embodiments discussed below. Now, due to the nontrivial mechanism explained in detail later, increasing the quasi-particle density affects not only the critical current but also other properties, such as the dynamic inductance of the channel. And this additional phenomenon can be used to tune any property of the system on a macroscopic scale, as implied by a change in the state of the channel. For example, as in the embodiments discussed later, this phenomenon can be used to tune the resonant or cutoff frequency of a circuit including the channel (e.g., a linear inductor configured as an LC circuit), or to tune the resonant frequency of a tunable coupler used to couple quantum bits.

[0048] The physical phenomenon of the present invention proves to occur for various device structures and various superconducting materials. Note that, in this respect, the superconducting channel is usually made of aluminum (Al) which becomes normal at very small magnetic fields. Conversely, various superconducting materials can be used to make the device involved in the present method, including materials such as titanium nitride (TiN) which allow to maintain superconductivity in much larger magnetic fields than Al. As a result, the device involved in the present method can operate in very large magnetic fields.

[0049] The proposed solution has additional advantages and can be used for a variety of applications. For example, locally regulating the critical current density of the superconducting material allows to regulate both the inductance and the resistance of the superconducting channel. It is interesting that the applied voltage difference can be positive or negative. The channel can be operated in particular with a DC voltage at very low power (e.g. less than 1 nW in the OFF state, 0 W in the ON state). The most important advantage is that the critical current I C The speed that can be regulated (e.g. in embodiments, less than 100 ns, down to values close to 1 ps). This in turn allows to achieve very high ON / OFF switching frequencies (e.g. up to 10 GHz). Thus, the superconducting device can advantageously be operated as a switch to locally suppress superconductivity (e.g. as a switch) and to modulate a signal passing through the channel, as in embodiments.

[0050] The device can be integrated for example on an existing quantum chip (e.g. on the same platform as an amplifier or even on a qubit chip). When using TiN as superconducting material, the present method shows a satisfactory operation at temperatures lower than 3.4 K. It can be applied in particular for signal multiplexing purposes, e.g. for switching signal routing on or near a quantum chip. In other applications as described above, the device is configured as a circuit, e.g. an LC circuit or any circuit in which the superconducting channel is configured as a linear inductor. In this way, the resonance frequency of an LC oscillator or of any other oscillator comprising a linear inductor can be tuned by changing the inductance of the wire. Similarly, the resonator can form part of a coupling element of a frequency-tunable coupler between qubits of a quantum computer, whereby tuning the resonator frequency makes it possible to tune the qubit coupling. In further applications, the cutoff frequency of an LC filter, a pi filter or any other filter involving a linear inductor is regulated by changing the inductance of the wire. For completeness, the underlying device 10, 11, 12 is easy to manufacture, can have very small dimensions and can be manufactured in an integrated way. This makes it an ideal component for cryogenic applications. In addition to quantum processing and frequency tuning applications, additional applications can be envisaged (e.g. for detectors, amplifiers, circulators and multiplexers).

[0051] All of these will now be described in detail with reference to specific embodiments of the invention. First, referring to... Figures 9-10 As shown, a voltage difference can be applied between S30 and S32 to reduce the channel's critical current I. C The resistance is reduced to zero, thereby generating resistance in channels 102, 112, and 122. This resistance is generated at least locally at the level of the channel portion opposite the gate electrode. In other words, this is equivalent to locally causing superconductivity to desuperconduct at the level of portions of channels 102, 112, and 122, which can be advantageously used for switching applications.

[0052] like Figures 9-10 As further shown, the voltage difference is typically cut off at a certain point (S40), causing the superconducting materials 102, 112, and 122 to self-reset to their fully superconducting state. Note that after cooling the channel in step S20, the channel is maintained at an appropriate temperature. Therefore, the resistance of channels 102, 112, and 122 can be locally increased (e.g., returned to their normal state) at the level of the channel portion opposite the electrodes in a controllable manner, and then reversed. Thus, although hysteresis may occur, the process is reversible. As discussed later, the size of the region transitioning to normal can be controlled by the applied voltage.

[0053] Figure 9 This scheme can be particularly used to switch or even modulate signals routed via a channel, as discussed now with reference to Figures 6 and 10. That is, before applying a voltage difference in S32 to inject electrons into the channel, a signal can be applied to channel 102 in S25, for example, to route the signal through the channel. As described above, this results in a resistance in channel 102, which in turn affects (S50) the transmission of the signal applied in the channel (S25). Note that the applied signal can be an AC or DC signal. For example, it can be an RF signal, as discussed later in quantum computing applications. Therefore, this method can be particularly used for switching or modulating signals at very high frequencies, something that can be advantageously utilized in quantum processing applications. In such applications, the applied signal is typically an RF signal with a source impedance typically between 10 and 100 Ω (e.g., 50 ohms), which can be switched quickly and easily according to this method.

[0054] In particular, steps S32 and S40 can be repeatedly performed in order to repeatedly apply (S32) and cut off (S40) the voltage difference in order to modulate (S50) the signal applied to the channel 102 (S25). As described herein, the term "modulate" a signal should be understood broadly, referring to a process of changing one or more properties of the signal (e.g. typically a waveform). One can significantly change the intensity of the signal passing through the channel (e.g. by controlling the size of the high resistance region) and / or the frequency of the signal transmission (by repeatedly turning off and on the channel). Also, for example, the signal applied at step S25 can already be a modulated signal, e.g. a waveform of varying amplitude or frequency. Several applications can be considered. For example, one can switch a microwave (MW) signal and route it to different ports for multiplexing purposes.

[0055] In terms of quantity, the properties of the applied signal can be modulated (S50) for example at frequencies larger than 10 MHz and smaller than 10 GHz, which opens the door to several applications at cryogenic temperatures. Also, these frequencies represent on / off switching frequencies, which should be distinguished from the usual process of frequency modulation. Measurements performed by the inventors have demonstrated that a single operation time of less than 100 ns can be achieved. However, considering the limitations of the measurement setup they used, they concluded that the duration of a single operation can be in the order of picoseconds. Two effects are involved, namely the relaxation of the quasi-particles from the injection point and the diffusion. It has been shown that the effects of quasi-particle relaxation and diffusion are very fast processes (e.g. in the order of 20 to 50 ps or less, noting that the diffusion process can be much faster than the relaxation process) for comparable device structures, where the quasi-particle generation process is thermally triggered by the application of a heat pulse. However, in the present case, the generated quasi-particles are not in thermal equilibrium, so that the above estimates (e.g. 20 to 50 ps) can be considered as an upper limit of the actual duration of the process, which will also depend on the superconducting material used.

[0056] In the present case, modulation frequencies up to 10 GHz can still be considered in principle. Otherwise, other embodiments can require lower frequencies (e.g. in the order of 1 KHz, 1 MHz or 10 MHz or less).

[0057] Another outstanding feature of the present method is that the intensity of the gate current generated when applying (S30, S32) the voltage difference remains well below the critical current I C (e.g. at least by a factor of 10, 100 or 1000). CThe gate current does not jeopardize the state of the channel, thus making the process well controlled. In practice, however, the gate current can be several orders of magnitude smaller than the critical source-drain current. This is discussed in detail in Section 2.1. Depending on the chosen superconductor material, a gate current in the range of tens of pA to several nA can typically be used to achieve full suppression of superconductivity. That is, the strength of the gate current generated at steps S30, S32 is typically between 10 pA and 10 nA. This is in sharp contrast to the previously described devices, in which a gate current comparable to the critical current I C (e.g., several μΑ) was needed to operate the device.

[0058] A relatively small number of electrons are injected with energies typically much higher than the superconducting gap to trigger the creation of quasi-particles. For example, a voltage difference can be applied at S30, S32 so as to make the average energy at least 100 times larger than the superconducting gap of the superconductor material, see Figure 6B for an illustration. The superconducting gap is defined as the energy difference between the ground state and the lowest quasi-particle of the superconductor channel 102. In practice, the average energy of the injected electrons typically needs to be 50 to 60 000 times larger than the gap, although larger ratios can in principle be observed. Preferred ratios are in the range of 5000 to 20 000, in particular for Ti-based materials. The ratio can for example be about 10 000. That is, for a gap of 2Δ = 500 μeV, a voltage difference of about 5 eV can be applied, as Figure 6D assumed in

[0059] Note that the preferred ratio depends on the desired mode of operation. For switching applications, a very large ratio (e.g., with a small voltage difference) yields satisfactory results. Superconductors with a small gap, such as titanium, will be preferred for such applications. However, in cases where significant suppression of signal transmission when the channel is in the OFF state is desired, superconductors such as TiN with a high resistance in their normal state are preferred. Superconductor materials with a high normal state resistance are also preferred for frequency tuning applications, as they have a high dynamic inductance.

[0060] The impedance of the gate electrode, which is mainly determined by the insulating medium 103, can typically be larger than 1 GΩ. This value corresponds to the minimum of the voltage difference needed to locally switch the channel to its normal state. In practice, however, the gate impedance can be much larger. In the example discussed in Section 2.1, for example, the superconductivity can be suppressed at about 5 V / 1 nA = 5 GΩ. The gate impedance will likely be larger than 1 TΩ when the critical current density starts to be affected by the applied voltage difference. On the other hand, it can be smaller than 1 GΩ when the applied voltage is somewhat larger and the normal region in the channel starts to expand. This is a consequence of the exponential behavior of the switch: the tunnel junction makes the resistance V G / I GVaries exponentially.

[0061] In this regard, in step S34, the strength of the applied voltage difference can be changed so as to controllably grow the region of the channel 102 that generates resistance, as Figure 11 indicated in the flowchart. If desired, the voltage difference can be adjusted so as to maintain the source-drain current I SD above the re-trapping current of the superconducting material. By source-drain current I SD extending the normal region, a large OFF resistance can be achieved, which leads to uncontrolled heating effects. In this way, once a local normal region is created, all the wires turn normal due to uncontrolled heating.

[0062] More generally, in S34, the strength of the voltage difference can be changed so as to adjust any local property of the channel, or any property determined by that local property (e.g. of a component of the device or connection). To this end, the target property can be monitored in S33. For example, as previously mentioned, changing the critical current density (e.g. a local property) affects the total critical current of the channel (a non-local property). As we will see, this can further affect the dynamic inductance L kin of the channel (another non-local property), which can be used to tune additional properties, such as the resonant frequency or cut-off frequency of a circuit including the superconducting channel.

[0063] The applied voltage difference can in fact be modified to change (S34) the dynamic inductance L kin of the channel 102, 112, 122 for the following reasons. Electrons are injected into the channel, which creates a finite quasi-particle density that falls to zero away from the electron injection point. In the region of higher quasi-particle density, both the critical current density and the dynamic inductance L kin are changed. The reduction in I C results in an increase in L kin , and so is an incidental result of the increase in quasi-particle density upon injection of electrons.

[0064] As described herein, in S34, the dynamic inductance L kin may be changed so as to adjust the frequency of a circuit 120 including the superconducting channel 122, wherein the superconducting channel is configured to form a wire of an inductor 122 in the circuit. That is, the critical current I C is changed by injection of electrons, which happens to increase the dynamic inductance L kin of the channel 122, as described above. This phenomenon can now be used to adjust the frequency of the circuit. For example, the frequency can be the resonant frequency of an LC oscillator or any other oscillator including a wire-like inductor 122. By changing the inductance of the wire, the resonant frequency is modulated. In other applications, the cut-off frequency of an LC filter, for example Figure 8The hypothetical pi filter, or any other filter comprising a linear inductor, is adjusted by changing the inductance of the wire due to the gate 121. Figure 8 A pi-section filter 120 is actually shown, comprising two capacitors CI, C2 and an inductor L (e.g. a wire 122), as part of a device 12, which additionally comprises an electrode 121 separated from the wire 122 by an insulating medium (not shown).

[0065] Other applications can for example aim at tuning the resonance frequency of a coplanar waveguide (CPW) resonator 110, like Figure 7 as hypothesized in the introduction, or at tuning the resonance frequency of a frequency tunable coupler. In the example of Figure 7 The device 11 comprises a CPW resonator 110, which comprises a conductive track 112 and a pair of loop conductors 114 on both sides of the track 112. The loop conductors are separated from the conductive track by respective gaps. The gate electrode 111 is directed towards the track 112. Again, the device is voltage controlled by the gate electrode 111. The small electronic current that is generated needs to flow out; the central conductor 112 of the CPW resonator is therefore grounded, which can for example be achieved with a very thin, high dynamic inductive wire.

[0066] In other applications, the superconducting device can be used as a frequency tunable coupler for coupling quantum circuits (e.g. superconducting qubits). For example, Figure 6E A classical architecture of a quantum processing device is depicted. In this case, the quantum circuit 55 is assumed to be a Transmon type superconducting qubit, although spin and topological based qubits are also able to benefit from the present invention. The device 1 additionally comprises a number of RF controlled components, such as couplers and frequency tunable coupling elements. The device 1 will be discussed in detail below with reference to another aspect of the present invention.

[0067] Reference is now made to Figures 6-8, which describe another aspect of the present invention, which relates to a device 1. The device can in particular be a quantum processing device, or any device comprising components that operate at cryogenic temperatures. Aspects of the device have already been discussed with reference to the present method; they will only be briefly described below.

[0068] The device essentially comprises a device 10, 11, 12 comprising a potentially superconducting material. As previously described, the latter is configured to define a channel 102, 112, 122, which is preferably configured as a wire. The device is adapted to operate at cryogenic temperatures so as to render the channel 102, 112, 122 superconducting. The device further comprises a gate electrode 101, 111, 121 directed towards the channel 102, 112, 122, so that an end face of the gate electrode faces a portion of the channel. Furthermore, as previously described, an electrically insulating medium 103 is arranged to electrically insulate the gate electrode 101, 111, 121 from the channel 102, 112, 122.

[0069] As Figure 6B and 6E further shown, the device 1 also comprises a control unit 20, which is connected to the device 10 and, in particular, to the gate electrode 101. The control unit 20 can for example consist of various interconnected elements 21-24, as Figure 6B assumed in . It is typically configured to apply a voltage difference between the gate electrode 101 and the channel 102, so as to inject electrons into the channel through the electrically insulating medium 103, thereby generating a gate current between the gate electrode 101 and the channel 102. Due to the previously discussed principles, the control unit, together with the device 10, makes it possible to inject, in the operation of the device, electrons having an average energy sufficient to modify the critical current I C of the channel.

[0070] Figure 6B As Figure 11 further shown, the control unit 20 can be additionally connected to the channel 102. Moreover, the unit 20 can be configured to apply, in operation, a signal both to the channel 102 and to modify the critical current I C of the channel 102 by changing the applied signal. This makes it possible to implement applications such as those previously discussed with reference to . In particular, the control unit 20 can be configured to allow the measurement of a characteristic of the channel 102 (for example, or of any characteristic implied by this characteristic of the channel). The unit 20 can also be designed to allow the controllable modification of the applied voltage difference as a function of the measured characteristic, so as to regulate the characteristic in operation. To this end, the control unit can comprise additional equipment to monitor the characteristic to be regulated, as further discussed in section 2.1.

[0071] Figure 6B In the example of , the unit 20 comprises a source measurement unit 23, which can: apply the required voltage; and measure the current into the gate. Moreover, the unit 20 can comprise a voltmeter connected to the channel 102. The signal generator 22 is connected to the channel 102, which is itself connected to the ground 104. The various components are controlled by a logic unit 21, which can communicate with an external computer (not shown) and operate from the cloud, if necessary.

[0072] Figure 6B As 6C further shown, the device 10 comprises a substrate 103, which serves as the electrically insulating medium 103. Each of the gate electrode 102 and the wire 102 is deposited and structured on the substrate 103. The gate electrode 101 and the channel 102 are each preferably made of the same material, so as to simplify the photolithography steps (for example, a single photolithography step is needed to manufacture the device 10, as explained in section 2.2). Each of the superconducting material 102 and the gate material 101 can for example comprise Ti, TiN, Nb, Ti x Nb y Nz W, WSi, Al, Mo, combinations thereof, etc. The substrate 103 can for example comprise Si, Si02, sapphire, Si3N4, Hf02, combinations thereof, etc. Other materials can obviously be considered by the skilled person. Moreover, other structural designs can be used for the device 10. In particular, the insulating medium can be used as an intermediate layer, or as a void separating the gate 101 from the channel 102.

[0073] The width of the wire 102 is typically between 40 and 300 nm, while the minimum gap between the end face of the gate 101 and the opposite part of the channel 102 is preferably between 40 nm and 300 nm. The width of the wire 102 and the minimum gap are measured parallel to the main surface of the substrate 103 along the x-direction in Fig. 1 (e.g. in the plane, in this example transversely to the longitudinal extension direction y of the wire). However, note that the width of the channel and the minimum gap can be less than 40 nm (e.g. equal to 10 nm). Figure 6C

[0074] Figure 6D is a corresponding diagram schematically showing the density of states in the channel 102, and showing the electron injection process and the quasi-particle generation process. In this embodiment, the electrons are injected with an average energy of 5 eV. The electrons tunnel through the insulating medium. As previously explained, the high energy of the injected electrons (e.g. much larger than the superconducting energy gap, in this example 5 peV) leads to the formation of quasi-particles in the channel 102.

[0075] Figure 6E A general architecture of a quantum processing device 1 according to an embodiment is shown. Such a device 1 typically comprises one or more devices 10 and quantum circuits 55, here assumed to be of the transmon type. The superconducting channel of each device 10 is coupled to one or more quantum circuits 55, preferably via a capacitance, although inductive coupling can also be considered.

[0076] In detail, Figure 6E ​The illustrated device 1 includes several cooling stages 61-65. It particularly includes a first stage 61 which includes or is connected to a control unit 20, stage 61 can include one or more signal generators (not shown) to generate RF signals as needed to drive components 10, 55. Such signal generators can form part of control unit 20; they are at least connected thereto. The signals are generated and sent along a downward path 41 to the qubit platforms, processed as needed at the level of the qubit platforms to perform quantum computations, and output signals are sent back along an upward path 42 to control unit 20. The device also includes a second stage 65 adapted to be cooled at a much lower temperature than the first stage. The superconducting qubits 55 and apparatus 10 are disposed in the second stage 65, which can include various other components 51, 52 such as attenuators, circulators, isolators, etc. as known per se. The apparatus 1 also includes intermediate stages 62-64 between the first stage 61 and the second stage 65. Thus, the first stage, the intermediate stages 62-64 and the second stage 65 form a series (i.e. a succession) of successive stages, with each intermediate stage being adapted to be cooled at a lower temperature than any preceding stage of the series. Each platform 61-65 can include various electronic elements (e.g. attenuators, for thermalizing the RF signals from the signal generators, communication means (e.g. optical and / or electrical), cooling means, etc.). These electronic elements are known per se and are not shown in Figure 6E for simplicity. The apparatus can include additional platforms, including a room temperature platform (not shown). Figure 6E The (indicative) temperatures shown in

[0077] In the example of Figure 6E each apparatus 10 is configured as a frequency-tunable coupler. In operation, each coupler 10 is connected to two or more qubits 55 for modifying the coupling between such qubits 55. Note that the qubits 55 are computational qubits, which together with one connected coupler 10 form a coupled-qubit structure 54. Each structure 54 includes N computational qubits (e.g. fixed-frequency transmons) with N > 2, and at least one coupling element 10 which couples pairs of the N qubits. Figure 6E The apparatus 1 shown in

[0078] Figure 6E A very schematic description is provided. Of course, several arrangements of qubits 55 and couplers 10 can be considered. It is possible that a qubit 55 can be connected to one or more tunable couplers 10, for example, in order to allow the transitions between the states of the qubit to be parametrically driven by modulating the tunable coupler energy. In addition to the parametric driving of the qubits, quantum gates can also be used, in which a coupler or tunable bus is resonant with two qubits in order to exchange excitations between the qubits. Another option is to use a fixed bus, the frequency of which is changed in order to completely switch off the interaction. More generally, by tuning the frequency of the coupler 10, the multi-qubit interaction can be enabled or disabled.

[0079] The unit 20 can for example comprise or be connected to a compiler and a sequencer (not shown), which are common components of quantum processing devices. The compiler determines and schedules high-level operations, as necessary for running a quantum algorithm on the device 1. The compiler can accordingly start determining batches of operations to be performed on the qubits 55, which are then decomposed into single-cycle operations by the sequencer, which can include operations such as single-qubit operations and two-qubit operations.

[0080] In a variant, the device 10 can actually be configured as a switch 10 for switching the RF signal applied to the channel 102, for example, for multiplexing purposes, as discussed earlier. As long as superconductivity can be maintained in the respective channel 102, the device 10 can be arranged on the upper platforms 61-64 of the device 1.

[0081] In other variants, a plurality of devices 10 is used, which are configured as frequency-tunable filters, the cutoff frequency of which can be adjusted, similar to a Purcell filter for increasing the coherence of the qubits (e.g. while keeping the pulses short), for readout purposes. That is, each frequency-tunable filter is connected to a respective quantum circuit (e.g. a qubit or a readout resonator) to cut off the frequency of the signal resonating with the qubit, while allowing a control signal for readout purposes to pass.

[0082] It is noted that the device according to the embodiments can similarly comprise devices 11, 12 as depicted in, for example, Figure 7 , 8 Fig. 1 1, 12, with a control unit (not shown) connected to such devices 11, 12. In particular, the device can comprise a circuit 120 as shown in Figure 8 . This circuit has been described earlier. It basically comprises an inductor 122 formed by a channel, which is structured as a wire 122. In this case, the control unit is used to adjust the dynamic inductance L kin of the wire 122 in order to adjust the cutoff frequency of the circuit 120.

[0083] The above embodiments have been briefly described with reference to the accompanying drawings and can provide a variety of variations. Several combinations of the above features can be considered. Examples are given in the next section.

[0084] 2. Specific embodiments - technical implementation details

[0085] 2.1 Fast superconducting switch and its operation

[0086] This section describes a fast superconducting switch and its operation, where the switch has an ultra-high gate impedance (much larger than 1 GΩ) and is operated via injection of high-energy electrons through an insulating barrier, and is able to self-reset.

[0087] Figure 1A A typical device is shown in Fig. 1. The device is essentially similar to the one discussed in Section 1 with reference to Fig. 6, which comprises metal nanowires with side gates, all deposited on an insulating substrate. A gate voltage V G results in a current I G flowing between the gate and the nanowire, where the current is exponentially dependent on V G . With the onset of I G , the superconductivity in the nanowire is weakened and eventually suppressed. The measurements show that a relatively small number of electrons injected at much higher energies than the superconducting gap are sufficient to trigger the generation of a large number of quasi-particles and quench the superconductivity. This mechanism is extremely efficient, allowing the suppression of the nanowire source-drain critical current I C by tens of μΑ. In contrast to the previously demonstrated devices, where a gate current comparable to the nanowire critical current (e.g., several μΑ) is required to operate the device, full suppression of the superconductivity is achieved with gate currents between tens of pA and several nA, depending on the superconductor material used. That is, this device relies on low-energy electrons, which are injected via a metallic medium. In addition to these remarkable features, the device presented here operates similarly for both positive and negative gate voltages and for the entire range of temperatures and magnetic fields for which superconductivity persists in the nanowire, making them applicable to a wider range of environments than Josephson junctions. Finally, the superconducting switch of the invention is extremely compact, can be fabricated in a single lithography step, and is easily integrated in existing architectures, either directly on the quantum processor level as a signal router, or as a low-temperature interface between classical voltage-driven digital electronics and superconducting current-driven circuits. The basic device operation is described below in terms of the critical current and its dependence on the gate voltage, temperature, and magnetic field. Characterization is performed with different substrates and superconductors. This section further discusses how the injected quasi-particles affect the superconductivity along the channel length and demonstrate the fast electrical switching and self-resetting to the superconducting state. Experimental observations are given, followed by a discussion of the physical origin of the phenomena, and their implications for device design.

[0088] Figure 1A A superconducting device is shown, with a schematic of the measurement setup. The device consists of a 2 μm long, 80 nm wide TiN wire with a TiN side gate. The wire and side gate are separated by an 80 nm wide gap. The gate and nanowire are processed using e-beam lithography and dry etching of a TiN film deposited on an intrinsic Si substrate. Measurements are made by passing a source-drain current I SD through the nanowire and recording the resulting voltage V. A gate voltage V G is applied by a source measurement unit, which also records the current I G flowing into the gate contact. Figure 1A The second side gate, visible in Fig. 2, is not operational and is held at ground. Further details on the materials, sample fabrication, and measurement techniques are reported in Section 2.2. At V G = 0, the nanowire shows a critical current I C = 45 μA, measured when I SD is scanned upwards or downwards from zero. In contrast, when I SD is scanned from the normal state to zero, superconductivity is re-established at a re-entrance current I R below 1 μA.

[0089] Figure 1B The nanowire differential resistance dV / dI SD is shown, measured when I SD is scanned in the positive direction. The inset gives the temperature dependence of I C and I R . The large difference between I C and I R , especially at low temperatures, can be due to self-heating of the nanowire when in the normal state, and the difficulty of extracting heat through the substrate or leads. Figure 1C and 1D show I C and I G as a function of V G . For V G = ± 2.5 V, I G begins to decrease, just before I C reaches the detection level (~ 100 fA, using the measurement setup). At V G = ± 5.5 V, I G ≈ ± 1 nA, I C disappears, and the nanowire reaches its normal state resistance of 1.6 kΩ. Figure 1E The parametric plot of I C versus I G shown in Fig. 4 indicates that for small values of I G , IC Extremely rapid inhibition (approximately 50% of the inhibition occurs in I) G Within the noise level, there is a slower decay that continues until I G ~1nA.

[0090] Figure 2A and 2B The figures show the results for various temperatures T and out-of-plane magnetic fields B. ⊥ I C The gate voltage dependence. Temperature and magnetic fields do not affect it. Figure 1D I G Compared to V G This property, and leads to the same V that completely suppresses superconductivity in the wire. G The value is up to the critical temperature of 3.7 K or the critical field of 3.5 T. On the other hand, for higher temperatures, I... C The inhibition begins to systematically shift to higher V values. G (See gray arrow). More observations were made regarding B. ⊥ rely.

[0091] Utilizing more than 20 TiN-based devices with various gate shapes, nanowire widths (e.g., 40, 80, and 200 nm), nanowire lengths (e.g., 650 nm, 1 μm, and 2 μm), and gate-to-wire separation (e.g., 80 and 160 nm), the inventors were able to observe I C The inhibition, and at the same time, or at a slightly anticipated point in time, I G The initial value was higher than the detection level. Similar phenomena were further observed with devices possessing different substrate materials and superconductor materials. For example, Figure 3A and 3B They respectively show the cases for similar Figure 1A Ti nanowires, but Ti nanowires with a width of 200 nm, I C and I G Compared to V G The evolution of I. G It reaches normal operating conditions at a value as low as 40 pA. Further characterization of the nanowire reveals a critical temperature of 220 mK and a normal operating resistance of 60 Ω. Figure 3C and 3D Displayed in a similar manner Figure 1A Measurements were performed on a TiN device, despite being deposited on a 25 nm SiO2 layer thermally grown on a Si substrate. Despite V G The operating range is very different from that in Figure 1, but I C The inhibition is still related to I G The beginnings are consistent. Devices including SiO2 also show I. C Compared to VG The characteristic asymmetry of the curve, where the suppression I for negative values C Compared to a positive V G The suppression of the value is faster. Assuming the abrupt termination of the gate electrode and the large electric field achieved on the SiO2 substrate, electron extraction from the gate is expected to be easier for a negative gate bias. In the current case, the detection of small gate current asymmetry may be hindered by stray current leakage in the measurement setup for high gate bias. In summary, these results indicate that the switching mechanism is a general phenomenon, independent of specific superconducting materials. On the other hand, the data obtained also suggest that weak superconductors require less gate current for the switching mechanism to occur.

[0092] The above measurements were performed on relatively short nanowires, where a sharp transition in resistance from zero to the normal state was observed. The following describes how superconductivity decreases along the length of the nanowire from the electron injection point. Figure 4A The device shown consists of six TiN segments, each 1 μm in length and 80 nm in width. A current I is injected along the nanowires. SD Simultaneously record the voltage V applied to the first five segments. i (For example, the rightmost contact is floating). Figure 4B I was shown SD <I R The differential resistance of the first three segments is amplified. Besides the extension of the zero-resistance state, the segment far from the injection point shows an intermediate region where the resistance is greater than zero but lower than its normal state value. This intermediate region indicates that normal and superconducting domains coexist in the same segment. Figure 4C It was shown as V G1 The critical current of the function, which has the following properties: Figure 4A The markings shown; Figure 4D The corresponding gate current I is shown in the figure. G1 I C Compared to I G1 The parameter graph is shown in Figure 4E Its focus is on the recapture current I R The region below 1.4 μA. Non-local measurements highlight two distinct regions. For I SD >I R All segments switch simultaneously: once a local hot spot occurs, the Joule heating mechanism heats the surrounding TiN material in a runaway manner, and the entire channel returns to normal. For I SD <I R Sequential switching occurs: the closer the segment is to the bias gate, the smaller the gate current required to reach the normal state.

[0093] Figure 4F It shows that S = (I) R -I C ) / IC The critical current suppression factor curve is plotted as a function of the distance |Δx| measured between the current injection point and the center of each segment. Extract I G1 The data, as shown in the legend (e.g., solid marks), corresponds to I. c1 Disappearing | I G | value (for example, see Figure 4E (Vertical arrow in the image). By scanning V G5 A similar analysis was performed and is plotted in gray in Figure 4. The results are highly consistent for both gate and voltage polarity and do not show any dependence on the relative positions of the injection point, measurement point, and ground contact. In this case, the limited data range and relatively large point-to-point scattering do not allow for the determination of the functional form of S(Δx). Fitting the results to the exponential decay function exp(-Δx / λ), as shown... Figure 4F As shown by the solid line in the figure, a characteristic attenuation length λ-1.2μm is generated.

[0094] By applying a square wave with an interpeak amplitude of 500mV and a repetition rate of 100kHz to the DC gate bias V G To test something like Figure 1A The switching speed of the device. An AC source leakage current of 100 nA amplitude and 100 MHz frequency is simultaneously passed through the nanowire; a lock-in amplifier is used as a function of time to monitor the resulting AC voltage drop ΔV. When a fast signal passes through a low-temperature bias voltage -T... s When applied and measured, ΔV averages zero and reflects the change in circuit impedance between the resistive nanowires and the superconducting nanowires; a positive ΔV value reflects the normal state. Figure 5A It is shown as V G The time evolution of ΔV as a function of the time-averaged gate current. Figure 5B As shown in V. G Switching operation is achieved within a 500mV interval of approximately 4.5V, corresponding to a gate current of 1.5nA. This gate current value is consistent with the result shown in Figure 1, within a small I... C Within the limit of I SD Complete suppression. Figure 5C V is shown G The time-dependent trajectory at 4.5V highlights the rapid and reproducible switching between the two impedance states. Figure 5DAn enlargement of the approach to the rise point is shown in the middle, where the dashed lines mark the transition between 10% and 90% of the step height occurring within 90 ns. Similar results are obtained for the decay time. This transient is equal to three times the time constant of the lock-in amplifier used (e.g. 30 ns), and is therefore considered the final switching time as measurable by the measurement setup used; as discussed in section 1, it can further be considered an upper limit to the device response time. Samples specifically designed for microwave measurements and related technologies will be investigated to assess the device switching speed more precisely.

[0095] The possible origin of the observed phenomenon is now discussed. Electrons emitted from the gate arrive at the nanowire in a deeply unbalanced state, with energies on the order of e V G , much larger than the superconducting gap (e.g. 500 μeV for TiN). A large number of quasi-particles is generated within the nanowire as the electrons relax to the gap edge through inelastic scattering with other electrons and phonons. A sufficiently high concentration of quasi-particles drives the nanowire normal by unpinning the superconducting gap and suppressing the depairing critical current, which causes the channel to switch to the normal state. This behavior can lead to a superconducting nanowire single-photon detector (SNSPD), where the impact of a visible or infrared photon prompts single electrons to reach high energies, which in turn trigger the generation of a large number of quasi-particles as they relax. In the current case, high-energy electrons are provided directly through the gate current, without the need for photons. The inventors have shown that the critical current suppression is highest at the injection point and decreases with distance, with a characteristic length scale on the order of 1 μm. This length scale is believed to be related to the diffusion length of the high-energy electrons and the relaxation length of the generated quasi-particles.

[0096] The switching effect discussed here is particularly useful for the fabrication of fast electrical switches with resistive and superconducting states, which operate at extremely low gate currents. The ultra-high gate impedance and low power consumption make this switch ideal for interfaces between voltage-driven transistors and current-driven superconducting circuits. The quasi-particle relaxation length limits the extension of the segment that can be switched, and therefore limits the highest normal-state resistance of the device. In this case, it is desirable to select a superconductor with a large diffusion length and a large normal-state resistivity. Alternatively, nanowires of arbitrary length can be operated by selecting I SD > I R In this case, switching from the superconducting to the normal state requires actively resetting I SD to less than I Rvalues. For low current biases, response times below 90 ns can be achieved, which is limited by the measurement setup as discussed previously. The final switching speed is obviously determined by the quasiparticle relaxation time in the superconductor, typically below 1 ns. The devices discussed in this paper can also be used to study quasiparticle physics at unprecedented energy scales and at the limit of no current flow in the nanowire. The temperature and magnetic field studies reported in Fig. 2 show that the complete suppression of I C The required gate current is a robust property of the system, which can be advantageously used in several applications.

[0097] However, Figure 2A and 2B shows that at lower values of I G a richer physics of the quasiparticle density; the initial suppression of I C moves to increasingly high gate voltages as the temperature is increased (see grey arrows). This behavior can be considered to reflect a thermodynamic increase of the quasiparticle density in the wire, which requires the injection of more electrons before a considerable effect on I C is observed.

[0098] Other mechanisms are ruled out as sources of the suppression of I C In Figure 2A , Joule heating in the barrier (due to the dissipated power V G I G and the subsequent local temperature increase) is incompatible with the absence of a temperature dependence of the complete suppression of I C The assumption that Joule heating does not play a significant role is further tested with additional measurements. The consistent correlation between the applied very small gate voltage, the gate current, and the critical current, as well as the significant non-local response extending far beyond the gate-induced electric field (see Fig. 3), also rules out an electric field-induced suppression of superconductivity.

[0099] In summary, the inventors were able to test a number of devices in which the superconductivity was modulated or quenched by a gate current that was several orders of magnitude smaller than the source-drain critical current. The devices can operate in a non-latching mode and on a fast time scale. Due to the universality and robustness of the effects presented and the ease of device fabrication, such devices can be advantageously used in quantum processing devices and other low-temperature electronic devices.

[0100] 2.2 Method

[0101] Sample fabrication. A 20 nm TiN film was deposited on intrinsic Si. Nanowires were obtained by electron beam lithography. A 50 nm thick layer of hydrogen silsesquioxane (HSQ) based negative resist was used as a mask. After developing the resist, the unprotected TiN regions were removed by inductively coupled plasma etching in HBr plasma. That is, wire and gate were obtained from the same superconducting film in a single lithography step. After TiN etching, the HSQ was removed by immersion in a dilute hydrofluoric acid solution. Characterization of the TiN film yielded a resistivity of 68 Ω per square, a critical temperature of 3.7 K and a critical out-of-plane magnetic field of 3.5 T. These properties were preserved in the final device. Ti wires were defined by electron beam lithography using a positive tone polymethyl methacrylate mask, electron beam evaporation of a 30 nm thick Ti film and lift-off. Prior to evaporation, the chip was briefly immersed in a buffered HF solution to remove the Si native oxide. Ti evaporation was performed at a deposition rate of 1 nm s -9 at a base pressure of 10 -1 mBar. A high deposition rate was chosen to minimize contamination of the Ti film during evaporation. For both Ti and TiN wires, normal pads placed at approximately 100 pm from the active area of the device were defined by lithography and Ti / Au evaporation.

[0102] Electrical measurements. Electrical measurements were performed in a dilution refrigerator with a base temperature of 10 mK using standard lock-in techniques. Low pass and high frequency filters were installed along each wire. The DC source-drain current I SD was applied to the nanowire via a large bias resistor. The AC differential voltage V across the nanowire was then recorded with a lock-in amplifier with a 10 MΩ input impedance and used to calculate the differential resistance dV / dI SD .

[0103] Results are presented with I SD as the fast axis. Measurements were recorded from negative to positive values. This allowed the wire to be initialized in the superconducting state before starting each scan. Gate voltages were applied with a Keysight B2902A source measurement unit which also recorded the current into the gate contact. To avoid damaging the device, a compliance of ±100 nA was chosen. A linear contribution of approximately 1 pA V G associated with parasitic leakage paths in the setup was subtracted from the I -1 measurements discussed in section 2.1. First, a linear fit of the low bias I G versus V G curve was performed and the resulting line was then subtracted over the entire V G measurement range. To avoid any contribution of displacement currents, I G values were recorded, with a waiting time on the order of 30 s after a V G step.

[0104] Measurements in Figure 5 were performed via resistive bias-T s (e.g. 50 kQ resistor and 22 nF capacitor) enabling simultaneous application of DC and AC signals. The AC signal was applied through a superconducting, non- attenuating coaxial cable. The rapidly varying gate voltage was applied via a Keysight signal generator with 80 MHz bandwidth, while the nanowire resistance was measured with a Zurich instrument operating in ultra-high frequency lock-in at a base frequency of 100 MHz. The AC current of 100 nA RMS amplitude was passed through the sample, the voltage drop across the nanowire was input to the lock-in amplifier, demodulated, and plotted on a Keysight DSOX2024A oscilloscope with 200 MHz bandwidth. The lock-in was set to the shortest possible time constant (30 ns), corresponding to a measurement bandwidth of 14 MHz.

[0105] While the application has been described with reference to a limited number of embodiments, variations and alternatives, it will be appreciated that many modifications can be made within the scope of the application, and that the application can be practiced otherwise than as specifically described. In particular, features recited in given embodiments, variations or shown in the drawings (e.g. of the device class or method class) can be combined or substituted with other features in another embodiment, variation or drawing without departing from the scope of the application. Thus, various combinations of features described with respect to any of the above embodiments or variations can be envisaged which remain within the scope of the appended claims. Furthermore, many minor modifications can be made to the specific embodiments described to adapt them to the particular situation or material without departing from the scope of the application. Accordingly, the application is not limited to the specific embodiments disclosed, but rather the application is to include all embodiments falling within the scope of the appended claims. Furthermore, many other variations than those mentioned above will be apparent to those skilled in the art in the light of the overall disclosure, and can be used instead of or in addition to those described herein. For example, other materials than those explicitly mentioned can be used.

Claims

1. A method for operating a superconducting channel, the method comprising: An apparatus is provided, comprising: Channel defined by potential superconducting material, A gate electrode, positioned adjacent to the channel, such that the end face of the gate electrode faces a portion of the channel, and An electrically insulating dielectric is arranged to electrically insulate the gate electrode from the channel; The channel is made superconducting by cooling the device; and A voltage difference is applied between the gate electrode and the channel to inject electrons into the channel through the electrically insulating medium, thereby generating a gate current between the gate electrode and the channel, wherein the electrons are injected with an average energy sufficient to modify the critical current of the channel, and wherein the intensity of the gate current generated when the voltage difference is applied is between 10 pA and 10 nA.

2. The method according to claim 1, wherein Applying the voltage difference reduces the critical current to zero, thereby generating resistance in the channel at least locally at the portion of the channel.

3. The method according to claim 1 or 2, wherein the method further comprises: The voltage difference applied to the potential superconducting material is cut off to allow it to self-reset to a fully superconducting state.

4. The method according to claim 2, further comprising: Before applying the voltage difference, a signal is applied to the channel, thereby creating resistance in the channel that affects the transmission of the signal applied to the channel.

5. The method according to claim 4, wherein the method further comprises: The voltage difference is repeatedly applied and cut off to modulate the applied signal.

6. The method according to claim 5, wherein The applied signal is modulated at a frequency greater than 10 MHz and less than 10 GHz.

7. The method according to claim 1, wherein The intensity of the gate current generated when a voltage difference is applied is at least ten times smaller than the critical current.

8. The method according to claim 1, wherein A voltage difference is applied such that the average energy is at least 100 times greater than the superconducting bandgap of the potential superconducting material.

9. The method according to claim 1, wherein The applied voltage difference corresponds to the minimum value required to switch the channel to its normal state, and wherein the impedance of the gate electrode, as determined by the electrical insulating medium, is greater than 1 GΩ.

10. The method according to claim 2, further comprising: The intensity of the voltage difference is varied in order to controllably grow the region of the channel in which the resistance is generated.

11. The method according to claim 1, wherein the method further comprises: The applied voltage difference is changed in order to adjust the characteristics of the channel or the characteristics implied by the characteristics of the channel.

12. The method of claim 11, wherein The regulated characteristic of the channel is the dynamic inductance of the channel.

13. The method of claim 12, wherein The device further includes a circuit having an inductor formed by channels configured as wires; and Adjust the dynamic inductor to adjust one of the resonant frequency or cutoff frequency of the circuit.

14. The method according to claim 1, wherein The device is configured as a frequency-tunable coupler connected to the quantum circuitry of a quantum processing device; and Applying the voltage difference adjusts the resonator frequency of the frequency-tunable coupler, thereby modifying the coupling between quantum circuits.

15. An apparatus comprising: The device includes: The device is configured to operate at a cryogenic temperature that makes the channel superconducting, defined by a channel defined by a potential superconducting material. A gate electrode, positioned adjacent to the channel, such that the end face of the gate electrode faces a portion of the channel, and An electrically insulating dielectric, arranged to electrically insulate the gate electrode from the channel; and A control unit, connected to the gate electrode, is configured to apply a voltage difference between the gate electrode and the channel to inject electrons into the channel through the electrically insulating medium, thereby generating a gate current between the gate electrode and the channel, wherein the average energy of the electrons is sufficient to modify a critical current of the channel during operation of the device at the cryogenic temperature, and wherein the intensity of the gate current generated when the voltage difference is applied is between 10 pA and 10 nA.

16. The device according to claim 15, wherein The control unit is further connected to the channel, and The control unit is also configured to both apply a signal to the channel and change the applied signal by modifying the critical current during operation.

17. The device according to claim 16, wherein, The control unit is also configured to allow: During operation, the characteristics of the channel, or the characteristics implied by the characteristics of the channel, are measured, and The applied voltage difference can be controlled to adjust the characteristics of the channel or the characteristics implied by those characteristics during operation, based on the measured characteristics.

18. The device according to any one of claims 15 to 17, wherein The channel is constructed as a conductor.

19. The device according to claim 18, wherein The device further includes: Circuits including inductors, and The control unit is also configured to adjust the dynamic inductance of the wire in order to adjust one of the resonant frequency or the cutoff frequency of the circuit.

20. The device according to claim 18, wherein The device includes a substrate forming an electrically insulating dielectric, and Each of the gate electrode and the wire is constructed on the substrate.

21. The device according to claim 20, wherein The width of the wire is between 40 and 300 nm, and The minimum gap between the end face of the gate electrode and a portion of the channel is between 40 nm and 300 nm, and the width and minimum gap are measured parallel to the main surface of the substrate.

22. The device according to any one of claims 15 to 17, wherein The potential superconducting material includes at least one component selected from the group consisting of Ti, TiN, Nb, W, WSi, Al, and Mo.

23. The device according to any one of claims 15 to 17, wherein The device is configured as a quantum processing device, which further includes quantum circuitry, and The channel of the device is connected to one or more quantum circuits in the quantum circuit.

24. The device according to claim 23, wherein, The device is configured as one of the following: A switch for switching the radio frequency signal applied to the channel; or A frequency-tunable coupler is used to modify the coupling between two or more quantum circuits in the quantum circuit.

Images