microwave circulator

By using magnetic flux quantum transmission in a superconducting long Josephson junction in a microwave circulator, the complexity and noise problems of existing circulator designs are solved, and low-power, frequency-tunable microwave signal transmission is realized.

CN123125607APending Publication Date: 2026-11-17ANALOG QUANTUM CIRCUITS LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202480084314.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2023-11-09
Filing Date
2024-11-07
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing microwave circulator designs suffer from large components and complex wiring, resulting in high hardware complexity and a tendency to generate harmful charge noise.

Method used

Using the magnetic flux quantum moving in the superconducting long Josephson junction as a medium, microwave signals are transmitted by conducting the magnetic flux quantum through a ring structure. A circulator is designed using a bias current source and a ring isolation layer to realize signal circulation and suppress unwanted signal transmission.

Benefits of technology

It achieves a low-power, AC-wire-free circulator design, reducing hardware complexity, lowering harmful charge noise generation, and providing a wider dynamic range and tunable frequency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123125607A_ABST
    Figure CN123125607A_ABST
Patent Text Reader

Abstract

This disclosure relates to a circulator for electromagnetic signals. The circulator includes three or more ports and a ring structure, the three or more ports being used to transmit and / or receive the electromagnetic signals, and the ring structure being used to connect the three or more ports. The ring structure includes a long Josephson junction, and the ring structure is configured to conduct moving magnetic flux quanta around the ring structure to transmit the electromagnetic signals to one of the three or more ports and suppress the electromagnetic signals at another of the three or more ports.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Cross-references to related applications

[0002] This application claims priority to Australian Provisional Patent Application No. 2023903597, filed on 9 November 2023, the contents of which are incorporated herein by reference in their entirety. Technical Field

[0003] This disclosure relates to microwave circulators. Background Technology

[0004] A microwave circulator is typically a three-port device that receives a signal at any one port and outputs the signal at an adjacent port in a clockwise or counterclockwise direction. While significant progress has been made in the design of microwave circulators, integrating existing circulators remains challenging due to the large number of components and / or wiring required.

[0005] Any discussion of documents, actions, materials, devices, articles of manufacture or the like included in this specification shall not be construed as an admission that any or all of these matters constitute part of the prior art or common knowledge in the field relating to this disclosure prior to the priority date of each appended claim. Summary of the Invention

[0006] This disclosure provides a microwave circulator that uses magnetic flux quanta moving within a superconducting long Josephson junction as a medium for transmitting microwave signals. The movement of the magnetic flux quanta induces signal circulation: the circulator primarily transports microwave signals from one port in a sequence to the next. The direction of the signal circulation is determined by the direction of the magnetic flux quanta.

[0007] A circulator for electromagnetic signals includes three or more ports for transmitting and / or receiving the electromagnetic signals; and a ring structure for connecting the three or more ports, the ring structure including a long Josephson junction, and the ring structure being configured to conduct moving magnetic flux quanta around the ring structure to transmit the electromagnetic signals to one of the three or more ports and suppress the electromagnetic signals at another of the three or more ports.

[0008] One advantage is that the circulator is a low-power device without AC wiring, which greatly reduces hardware complexity in practical operation. Second, in a circulator made of a ring of three small discrete Josephson junctions, harmful charge noise is less likely to be generated. Third, the operating frequency of the proposed circulator can be tuned dynamically by changing the flux quantum number. Finally, the dynamic range of the above circulator is greater than that of other circulators based on a single or several JJs.

[0009] In some embodiments, the ring structure comprises a single long Josephson knot along the entire ring.

[0010] In some embodiments, the Josephson junction includes an annular isolation layer located between two annular metal layers.

[0011] In some embodiments, the ring structure includes a synthetic long Josephson junction, which is formed by a plurality of discrete Josephson junctions.

[0012] In some embodiments, the circulator further includes a bias current source for generating a bias current that sustains the motion of magnetic flux quantum.

[0013] In some embodiments, the bias current is based on the frequency of the electromagnetic signal.

[0014] In some embodiments, the bias current is determined based on the relationship between the bias current and the electromagnetic signal.

[0015] In some embodiments, the circulator can be switched by changing the bias current source.

[0016] In some embodiments, the circulator further includes three or more waveguides for coupling the electromagnetic signal to the ring.

[0017] In some embodiments, the circulator further includes a current-generating source for supplying current through one electrode of the long Josephson junction, thereby generating magnetic flux quanta in the circulator structure.

[0018] In some embodiments, the circulator includes a flux injector for generating the magnetic flux quantum in the ring structure.

[0019] In some embodiments, the moving magnetic flux quantum generates a phase shift in the electromagnetic signal to produce interference that transmits the electromagnetic signal to an output port adjacent to the input port and suppresses the transmission of the electromagnetic signal to all other output ports.

[0020] In some embodiments, the interference is constructive at adjacent output ports and destructive at other ports.

[0021] A method for preparing the circulator for operation, the method comprising: Above the temperature at which the ring structure is superconducting, a magnetic field is applied to the circulator to generate the magnetic flux quantum; Cool the ring structure to maintain the magnetic flux quantum until the superconducting temperature is reached; and Disconnect the magnetic field to generate self-sustaining magnetic flux quanta in the superconducting ring structure.

[0022] A quantum computer includes multiple qubits and multiple circulators, the multiple qubits being used to perform quantum operations.

[0023] In some embodiments, the qubit is a solid-state qubit that operates at microwave frequencies, including superconducting qubits or semiconductor qubits.

[0024] In some embodiments, each circulator in the plurality of qubits is connected to one of the qubits to protect the one of the qubits from noise generated by an amplifier used to amplify the signal generated by the one of the qubits.

[0025] Throughout this specification, the word “comprise” or variations such as “comprises” or “comprising” will be understood to imply inclusion of the stated elements, integers or steps, or groups of elements, groups of integers or groups of steps, but does not exclude any other elements, integers or steps, or groups of elements, groups of integers or groups of steps. Attached Figure Description

[0026] Figure 1 The long Josephson knot is shown.

[0027] Figure 2 A microwave circulator based on a long Josephson junction is shown.

[0028] Figure 3 A method for initializing magnetic flux quanta in a microwave circulator is shown.

[0029] Figure 4 It shows when selecting (Normalized) junction length and When the dissipation constant is given, the flux quantum number of the ring LJJ is... The IV characteristic curves for different current bias values ​​are shown. The dashed line represents the effect of increasing current bias. The magnetic flux quantum velocity is close to the Swirhart value and is thus The convergence limit of the given voltage is given. Here, the current bias is... and average voltage All of them are dimensionless units.

[0030] Figure 5 The resonant frequency of a microwave signal in a toroidal Josephson junction with magnetic flux quanta is shown in diagram a) and (The solid lines marked) are relative to the magnetic flux quantum number. This will fix the knot length to It was observed that as the number of... ,frequency near (dashed line). b) Frequency of the two reverse rotation modes and (The solid line marked) and its estimate based on a linear model. and (Dashed line) relative to current bias This will fix the knot length to The magnetic flux quantum number is fixed at And the dissipation constant is fixed at Here, frequency is relative to plasma frequency. Normalization.

[0031] Figure 6 This shows the bias under (dimensionless) current. When zero, the elements of the S matrix are calculated by numerical solution of the Sine-Gordon equation (solid line) and by analytical analysis following the time-domain coupled-mode theory (label). ( (relative to driving frequency) It exhibits completely reciprocal scattering, in which and Same as a). b) Same as a), but at the resonant frequency The current bias is at its optimal value. At that time, the figure showed strong non-reciprocal scattering, in which c) At the resonant frequency Below, the elements of matrix S relative to current bias .

[0032] Figure 7 The diagram illustrates the flux quantum for three different types of magnetic flux relative to the S-matrix of an ideal circulator. S-matrix fidelity (defined in (38)) along The variation in frequency demonstrates the tunability of the proposed circulator's operating frequency.

[0033] Figure 8 The output power of a) the linear response and the nonlinear ring LJJ response is shown (by calculation). ) for input power (through a given The dependence of 1 dB on the compression point exists at approximately equal to 1 dB. b) S matrix elements ( The dependence of the input power is observed. Consistent with the 1 dB compression found in small figure a), it is observed that as the input power is increased to... Above this point, non-reciprocity begins to saturate.

[0034] Figure 9 This illustrates a) discretizing the long Josephson junction into equally spaced segments. Discrete nodes. Each node has a node flux. Related to current bias Driven, and through the critical current in the RSCJ model Parasitic capacitance is And the shunt resistance is The small Josephson junction is connected to ground. Subsequent nodes are connected via surface inductance. and surface resistance They are connected to each other. b) is the same as a), but the nodes are connected. It is also directly coupled to a current-carrying device. Waveguide.

[0035] Figure 10 (a) Definitions of voltage and current are shown. (b) Equivalent circuit of lumped elements of a transmission line.

[0036] Figure 11 This shows the termination at the load impedance. The transmission line in the middle.

[0037] Figure 12 The diagram illustrates reflection and transmission at the junction of two transmission lines with different characteristic impedances.

[0038] Figure 13 : any Port microwave network.

[0039] Figure 14 It shows having A 3 dB attenuator with matched characteristic impedance.

[0040] Figure 15 This illustrates a) direct coupling to a device with impedance The waveguide LC circuit. Here... This represents the voltage at the end of the waveguide, and specifically the voltages entering and exiting the waveguide. and sum. It is the current flowing through the waveguide. It is the current flowing through the LC capacitor, and This is the current flowing through the LC inductor. b) is the same as a), but the LC circuit uses a coupling capacitor. Capacitive coupling to the waveguide.

[0041] Figure 16 It shows a configuration with coupling to three ports. Two nearly degenerate inverse rotation modes and A schematic diagram of a three-port circulator. Ports The incoming wave at point is represented as The transmitted wave is represented as .

[0042] Figure 17 : The IV characteristic curve of the junction was measured at that time. It is the critical current. It is the gap voltage (i.e., ) and gap resistance . Detailed Implementation

[0043] The following description first provides a general explanation of microwave circulators, and then provides further details below.

[0044] Josephson knot

[0045] This disclosure utilizes a Josephson junction exhibiting the Josephson effect. A Josephson junction comprises two closely placed superconductors with a potential barrier or confinement between them. The Josephson effect is a macroscopic quantum phenomenon in which quantum mechanical effects can be observed at ordinary scales rather than atomic scales. The Josephson effect produces a current known as a supercurrent, which flows continuously across the Josephson junction without any applied voltage. The potential barrier can be a thin insulating barrier (such as a superconductor-insulator-superconductor junction, or SIS), a short segment of non-superconducting metal (SNS), or a physical contraction (ScS) that weakens superconductivity at the contact point.

[0046] This disclosure further utilizes a long Josephson junction, which has a penetration depth greater than that of the Josephson junction. A Josephson junction of one or more lengths. In terms of the basic model, a short Josephson junction is characterized by its Josephson phase. The Josephson phase is a function of time only, not coordinates; that is, the Josephson junction is assumed to be a point in space. Conversely, in a long Josephson junction, the Josephson phase can be a function of one or two spatial coordinates, i.e., or This disclosure provides an embodiment of a long and thin Josephson junction in which the second spatial coordinates can be omitted. .

[0047] Figure 1 A long Josephson junction 100 is shown, comprising a first superconductor 101 and a second superconductor 102, and an insulating layer 103 sandwiched between the first and second superconductors. The length L is designated 104 and is defined herein as the x-direction. and .

[0048] Magnetic flux quantum

[0049] The mathematical model used to describe the long Josephson junction 100 is the perturbation sine-Gordon equation as described below. The solutions to the equation include, for example... Figure 1 The cyclic current is shown as any number in 106. Therefore, the existence of the cyclic current is provided as a result of mathematical modeling of the long Josephson junction 100. It should be noted that in superconducting materials, there is no energy dissipation. Therefore, the cyclic current still exists, which is not the case in non-superconducting materials, where these currents would decay due to their energy being dissipated by the material. Since the long Josephson junction 100 can support supercurrents across the barrier 103, this cyclic current can also exist in the long Josephson junction 100. In this case, the current will circulate in the lateral direction as shown in the figure.

[0050] because Figure 1The junction in the diagram is a long junction, so a circulating current (shown as 106) can move along the long junction in the x-direction and is referred to as a flux quantum. Multiple flux quanta can coexist along the long junction and can move along the junction at the same speed. This movement can be regulated by a bias current 105, as further described below. In effect, the moving flux quanta appear to create a medium for electromagnetic signals (such as microwave signals) that preferentially propagate in the direction of flux quantum movement rather than in the opposite direction. However, more specifically, flux quanta can transmit microwave signals even if the flux quantum is stationary. For moving flux quanta, the signal propagates equally in both directions. However, due to the movement of the flux quantum, the phase of the signal is different depending on whether the signal propagates in the same direction as or in the opposite direction of the flux quantum's movement. This phase difference means that clockwise and counterclockwise signals interfere in such a way that the signal appears only at one neighboring port (constructive interference) and not at the other port (destructive interference).

[0051] The flux quantum moves along the long junction 100. In a linear long junction, the flux quantum eventually reaches one end of the long junction (e.g., the right-hand side). In a loop (ring) configuration in which the two ends of the long junction 100 are connected to each other, the flux quantum can move around the ring and continuously multiple times while the device is in operation.

[0052] Circulator

[0053] Figure 2 A ring-shaped long junction for fabricating a circulator 200 for electromagnetic signals is shown. In some instances, the circulator is configured for microwave signals, but electromagnetic signals of other frequencies can also circulate. The circulator includes a first port 201, a second port 202, and a third port 203, which are used to receive and / or transmit electromagnetic signals at any one or more of these ports. This means that each port can be used as both an input and an output port without altering the circulator 200. In other words, a signal can be applied to one of the ports, and the signal exits the circulator at the next port in a clockwise or counterclockwise direction.

[0054] Ports 201, 202, and 203 can be implemented as waveguides for coupling electromagnetic signals to ring structure 204. In other instances, more than three ports may be present. Circulator 200 further includes ring structure 204 connecting ports 201, 202, and 203. Ring structure 204 includes a long Josephson junction as explained above, having a ring-shaped first superconductor 205, a ring-shaped second superconductor 206, and a ring-shaped insulating layer 207 located between the ring-shaped first superconductor and the ring-shaped second superconductor. The ring structure is shown as circular, but other shapes are equally possible, such as oval, elliptical, triangular, square, or other shapes. These shapes are also considered rings because their ends are connected, and magnetic flux quanta can travel around the structure multiple times.

[0055] As explained above, the ring structure 204 is configured to conduct magnetic flux quanta around the ring structure 204. That is, the magnetic flux quanta move along the ring structure 204 and cause a phase difference in the electromagnetic signals between three or more ports. For example, the phase difference caused by the movement of the magnetic flux quanta results in interference with the electromagnetic signals, causing the circulator to transmit the electromagnetic signals from the first port 201 to the second port 202, but not to the third port 203. In other words, the signal is transmitted from the first port 201 to the second port 202, and is suppressed from the first port 201 to the third port 203. A current bias is used to keep the magnetic flux quanta moving.

[0056] Although the ring structure 204 in this example contains a single long Josephson knot along the entire ring, multiple individual Josephson knot segments can be constructed that together exhibit the behavior of a single long Josephson knot as described above. This is referred to as a synthetic long Josephson knot formed from multiple discrete Josephson knots.

[0057] A bias current source can be used to generate a bias current that sustains the motion of magnetic flux quantumes. Furthermore, the bias current is based on the frequency of the electromagnetic signal and is determined by the relationship between the bias current and the frequency of the electromagnetic signal. Therefore, the bias current source can be switched to reverse the circulating current, and thus the input signal exiting the circulator's port can be altered.

[0058] The circulator 200 may further include a current-generating source for supplying current through one electrode of the long Josephson junction (e.g., superconductor 205 or superconductor 206) to generate magnetic flux quanta in the ring structure.

[0059] Method for initializing magnetic flux quantum in a circulator

[0060] Figure 3A method is shown for preparing a circulator 200 for operation, and specifically for generating magnetic flux quanta as a medium for electromagnetic signal operation. Method 300 involves applying a magnetic field 301 to the circulator 200 above a temperature at which the ring structure 204 is superconducting to generate magnetic flux quanta. The magnetic field is oriented transversely to the circulator 200 such that it penetrates a long Josephson junction along the narrow width of the junction.

[0061] Method 300 then includes cooling the ring structure 204 302 to maintain the magnetic flux quantum until the superconducting temperature is reached, and finally disconnecting the magnetic field 303 to generate a self-sustaining magnetic flux quantum in the now superconducting ring structure 204.

[0062] The following description provides further details of the aforementioned method for realizing a passive on-chip superconducting circulator using the motion of a medium. Note that the term "passive" here refers to the presence of a bias current, but without AC drive, and with low power dissipation compared to other methods. This disclosure is based on a toroidal long Josephson junction (LJJ), which can support the propagation of a microwave field through the oscillation of trapped magnetic flux quanta, known as flux quanta. Due to periodic boundaries, the resonant modes of these oscillations are quantized as discrete energy levels. In the absence of perturbations, these modes are doubly degenerate, reflecting the fact that one frequency corresponds to two equivalent anti-circulation modes. A time-reversal symmetry-breaking element injects a bias current into the junction, causing the flux quantum chain to move uniformly within the loop. This subsequently undoes the doubly degeneracy of the resonant modes, resulting in a strong non-reciprocal response to an external microwave field.

[0063] In another instance, the disclosed circulator includes a flux injector that can avoid a cooling step.

[0064] Details of long Josephson junctions and flux quanta

[0065] As mentioned above, Figure 1 A schematic diagram of an LJJ (Large Junction Joint) is shown, consisting of two superconducting layers separated by a thin insulating barrier. The junction length is considered to be long. Much greater than Josephson's penetration depth And the barrier width The magnetic field can penetrate into the junction, thus generating trapped magnetic flux quanta, known as magnetic flux quanta. Physically, a magnetic flux quantum is represented by a circulating current around the junction barrier, which carries exactly one magnetic flux quantum. Under current bias When driven along the junction, the magnetic flux quantum acts as a moving medium for microwave propagation. Figure 2A ring-shaped LJJ directly coupled to three external ports (waveguides) is shown. For a certain flux quantum density and velocity, the flux quantum chain can achieve conditions for highly non-reciprocal microwave transmission, where, for example, most of the signal from port 1 is transmitted to port 2 and almost none is transmitted to port 3.

[0066] when Figure 1 The length of the Josephson knot described (along (coordinates) and its width (along (Coordinates) are respectively compared to Josephson penetration depth When the junction is much larger or much smaller, it is referred to as an LJJ. In this case, the superconducting phase (difference) across the junction can be described as a function of only one spatial coordinate. Phase The dynamics of the soliton are governed by the sinusoidal-Gordon (SG) equations, which allow for soliton solutions representing solitons with permanent wave profiles. Physically, a soliton in the LJJ corresponds to one carrying exactly one flux quantum. The magnetic flux trapped by the soliton. Therefore, the soliton is called a flux quantum and can be represented as a circulating current around the junction barrier. The physical size of the flux quantum in the LJJ is approximately the penetration depth. .

[0067] Magnetic flux quanta can be driven along the LJJ under the application of an applied current bias. Their movement speed is equivalent to the Swirhart velocity. The upper bound is defined as the effective speed of light in the LJJ medium. Unlike the Abrikosov vortex in type II superconductors and the flux quanta in Josephson junction arrays, the flux quanta in LJJs experience low damping during movement, do not produce radiation, and are fully decoupled from other electromagnetic excitations.

[0068] LJJs can be manufactured in different geometries. This disclosure provides, for example... Figure 2 The toroidal shape shown allows for the trapping of individual flux quanta and the study of their continuous motion rather than reflection from the boundary. Flux quanta in the LJJ mediate the propagation of the microwave field through their oscillations. Therefore, moving flux quanta within the LJJ can break time-reversal symmetry, and results demonstrate non-reciprocal scattering of microwaves in a linear LJJ. The moving flux quantum chain can induce destructive or constructive interference on microwave signals propagating from one port to another, depending on whether the ports are arranged in the same or opposite direction of the flux quantum motion.

[0069] As described above, this disclosure provides a design in which the driven magnetic flux quantum serves as the medium for the movement of microwaves within a ring resonator formed by a ring-shaped LJJ. Another advantage of this proposal is the fact that the propagation speed of microwaves within the LJJ is limited by the Swirhart velocity, which is typically only a few percent of the speed of light in a vacuum. This allows for a more compact implementation of the circulator on a chip.

[0070] SG equations and multiflux quantum solutions in ring-shaped LJJ

[0071] Static flux quantum. Consider a toroidal LJJ, in the absence of external perturbations and biases, the phase across the potential barrier. Satisfying the unperturbed SG equation (1) in Relative to penetration depth Normalized spatial dimensions along the knot length, and Represents relative to plasma frequency The time is normalized to the reciprocal of the given time. Further, below, a derivation of the generalized form of the SG equation is provided, which also takes into account current bias and dissipation. As is customary, from this point forward, this disclosure uses dimensionless normalized quantities other than the external drive frequency alone, which will later be converted back to the original value when analyzing the device scattering spectrum. Given that the (normalized) length of the ring LJJ is And assume that there exist objects already captured within the ring LJJ. A magnetic flux quantum, phase Satisfy boundary conditions (2) The trapping of magnetic flux quanta can be achieved by using techniques such as low-temperature scanning electron microscopy or by injecting a pair of local magnetic flux quanta currents to cool the system to the transition temperature.

[0072] The static (i.e. time-independent) multiflux quantum solution of equation (1) is given by the following equation. (3) in It uses independent variables and modulus The Jacobi elliptic sine function. Further, in the following sections, this disclosure introduces various Jacobi elliptic functions and their properties. There is also evidence that the function defined in equation (3) is... Indeed satisfied Therefore, this constitutes the true static solution of the unperturbed SG equation (1). It should also be noted that, as... The function The more explicit form is as follows (4) in It is the Jacobian amplitude function. Given the above static multiflux quantum solution... In the form of (2), the boundary conditions lead to the knot length. Magnetic flux quantum number With modulus The relationship between them (derivation below) (5) in It is a complete elliptic integral of the first kind. Solve this equation to find the solution for... and The given value Thus obtained The fully explicit form.

[0073] Moving magnetic flux quanta. The unperturbed SG equation in (1) is invariant under the Lorentz transformation (derivation below). (6) (7) in It is relative to the speed of Swirat. Normalized magnetic flux quantum velocity, and It is the Lorentz factor. Therefore, in order to obtain the dynamic multiflux quantum solution... Substitute the above transformations into (3) In other words, perform the replacement. and to obtain (8) Alternatively, the replacement can be performed according to the explicit form in (4). (9) Correspondingly, the boundary relationship in (5) is changed to (10) This reflects the relative contraction of the knot length.

[0074] Current bias and dissipation. Consider the perturbation SG equation. (11) in Represents the ratio of the critical Josephson current density Normalized current bias density, and This is the dimensionless intrinsic dissipation constant resulting from the tunneling loss of quasiparticles. When the dissipation balance comes from the energy input of the bias current, the current bias accelerates the magnetic flux quantum to its maximum critical velocity. Under this condition, a stable magnetic flux quantum velocity is achieved. Furthermore, the stable magnetic flux quantum velocity can be determined as follows, based on the current bias, in the following identity: and dissipation constant Calculation (see below) (12) in It is a complete elliptic integral of the second kind. Equations (10) and (12) form a set of two nonlinear equations, which can be numerically solved to determine the modulus. and magnetic flux quantum speed The condition is the knot length. Magnetic flux quantum number Dissipation constant and current bias Known.

[0075] IV Characteristic Curves

[0076] The multi-flux quantum solutions with different flux quantum numbers in a toroidal LJJ are characterized by their IV curves. Specifically, given the flux quantum number... and magnetic flux quantum speed The average voltage across the two ends of the ring-shaped LJJ generated by the periodic movement of magnetic flux quanta. Calculation as follows (13) in It is a total phase transition, and It is the total time it takes for the magnetic flux quantum to complete its full rotation around the junction loop. Note that voltage... It is dimensionless, and is already relative to the quantity determined by... The given voltage units are normalized. Using the results from equations (10) and (12), the voltage can be... and The following related (14) exist Figure 4 In the middle, the knot length is fixed as And the dissipation constant is fixed at . For magnetic flux quantum numbers For each value of , this figure shows the corresponding IV curve obtained by solving. It can be clearly seen that the IV curves for different flux quantum numbers are well distinguishable. Each IV curve represents the voltage. Converging to a limiting value, which is at a current bias When the magnetic flux quantum velocity becomes sufficiently large as it approaches the Swirart value, it is... Therefore, these characteristic IV curves provide a method for verifying the number of flux quanta trapped in a toroidal LJJ.

[0077] Flux quantum oscillations in a ring-shaped LJJ

[0078] This section describes the dynamics of the flux quantum and its oscillations in a toroidal LJJ. Due to the periodic boundary of the toroidal LJJ, the spectrum of flux quantum oscillations is quantized and doubly degenerate in the absence of flux quantum motion. Introducing flux quantum motion breaks this degeneracy, thereby creating a split in the frequencies of the two degenerate modes. The motion of the propagation medium realizes non-reciprocity in the same way as in conventional ferrite circulators and acoustic circulators based on large airflow.

[0079] Oscillations around the static flux quantum. Assume a static multiflux quantum solution in (15). There are small oscillations around. Therefore, the total perturbation multiflux quantum solution is (15) Will Substitute into (1), and relative to Linearization is performed to obtain the Lamé equation. (16) The potential in this equation is periodic, that is, Therefore, Bloch's theorem can be used to construct... The general solution. Specifically, will Represented as (17) in It is the wave number. With The periodicity is the same as that of the periodicity, that is ,and It is the oscillation frequency (due to its relationship with the plasma frequency). Normalization, which is dimensionless. Based on the derivations listed below, , and Given by the following formula (18) (19) (20) in , and These are the η, θ, and ζ functions of Jacobi, respectively. It is a Jacobian elliptic function, and These are complex parameters. These functions will be introduced below. In the equations above, the symbols... This means for For each value, there are two linearly independent solutions, meaning that each frequency level is degenerate twice. The Bloch solution is characterized by complex parameters. The complex parameters are not arbitrary, but limited to the following complex channels to ensure frequency. satisfy ,(twenty one) in and ,in ,This is The complementary modulus. For small oscillations, the wavenumber in equation (19) If true, then It is purely virtual. This will Limited to two encirclement roads (I) (II) ,(twenty two) Of these, only those considered Substituting the boundaries of these enclosures into equation (20) yields the frequencies. Two corresponding branches (I) (II) .(twenty three) These branches describe two types of small oscillations; namely, branch (I) represents plasma oscillations, and branch (II) represents magnetic flux quantum oscillations. Note that in... ,include and In this case, The surrounding channel produced the same frequency branch as the frequency branch in (22).

[0080] Next, the check was performed. The boundary conditions are derived from equations (2) and (15). .(twenty four) Using the form in (17) and periodicity This condition resulted in ,in Based on equation (19), the characteristic parameters were then identified. equation , (25) It is quantum constraint through the The solution determined the discrete frequency levels of small oscillations around the static flux quantum in the ring LJJ.

[0081] For the magnetic flux quantum oscillation branch In ,Discover This will be in (25) Limited to Furthermore, regarding the pattern and ,Discover And therefore This means that these modes have opposite wavenumbers, i.e., reverse rotation modes, and have the same frequency. (26) This is consistent with the previously mentioned double degeneracy. A trivial case is... In the case of or The mode represents a uniform translation mode of magnetic flux quanta. Figure 5 In section a, the first two non-trivial frequencies were plotted. and Relative to the magnetic flux quantum number It was observed that, In the largest limit, frequency near In this limit, the magnetic flux quantum in the toroidal LJJ acts as a support wave at a frequency equal to the fundamental frequency. A linear ring cavity oscillates at frequencies that are multiples of 1. Note that the wave equation in equation (16) is written in dimensionless units when the propagation speed is 1. Therefore, the linear dispersion relation is: This provides a useful estimate of the plasma frequency of a given toroidal LJJ. (Its scope can be) to The junction length can be appropriately selected to adjust the resonant magnetic flux quantum oscillation frequency. In scope Within this range, for the frequency of a superconducting quantum bit, the aforementioned range exists.

[0082] Oscillations around the moving magnetic flux quantum. Consideration is given by current bias. Drive to stable speed Small oscillations around the moving multiflux quantum solution. The small oscillations are obtained by performing a Lorentz transformation on the solution in (17) with respect to equations (6) and (7). The periodic boundary condition in (24) is then changed to (27) Based on this condition and using equations (17)-(19), the quantization constraint in (25) is changed to , (28) The velocity-related term in this equation, due to the finite current bias, breaks the pattern in (26). and The degeneracy between them. To see this, in Figure 5 In b, the drawing is... and relative to current bias This fixes the knot length to be [value]. The magnetic flux quantum number is fixed at And the dissipation constant is fixed at Clearly see and The split between them The increase indicates that the time reversal symmetry in the system has been broken. Figure 5 b also shows two approximate estimates from the linear model. and According to this model, speed Magnetic flux quantum shifts cause degeneracy of pairs and Displaced ,in It is the junction radius. It was found that when the current bias satisfies... (exist Figure 5 Select as in b When the frequency split is 0, the exact estimate and the linear estimate are in very good agreement.

[0083] Scattering of a ring LJJ directly coupled to an external waveguide

[0084] In the previous section, it was established that flux quanta in the toroidal LJJ mediate microwave propagation through their oscillations, which are doubly degenerate in the absence of flux quantum motion, corresponding to two opposite rotation modes. More importantly, it has been shown that the time-reversal symmetry in the toroidal LJJ is broken when flux quantum motion is induced by applying a finite bias current. In the following sections, the toroidal LJJ is directly coupled to an external waveguide, and this symmetry breaking is utilized to achieve a strong non-reciprocal response to microwaves.

[0085] equations of motion

[0086] Assume the ring LJJ is directly coupled to three external waveguides at three locations, and these three locations are spaced apart. ,like Figure 2 As shown. Based on classical input-output theory, the equation of motion for the phase across the junction in this case is derived below, and it is found that... (29) in It is to satisfy The coupling position, using (i.e., characteristic LJJ impedance) and (i.e., waveguide impedance) It is a dimensionless parameter representing the waveguide coupling strength. It is provided for the coupling position The input voltage, the input voltage relative to the input voltage by The given voltage units are normalized, and the δ function... express Point coupling at the location. Output voltage. It is calculated based on the input-output relationship of direct waveguide coupling. (30) In (29), waveguide coupling introduces a non-uniform dissipation term into the SG equation, which is derived from... Given the following, it is proved that this dissipation term is equivalent to the uniform dissipation term. And thus it can be replaced by the uniform dissipation term, in which case... .

[0087] S-matrix analysis

[0088] Given the input voltage For those from (29) Solve for the output voltage and then use the input-output relationship in (30) to calculate the output voltage. Elements of matrix S Defined as (31) For the simulation, assume the following set of experimental parameters: the Josephson current density per unit area is... The junction capacitance per unit area is And the inductance per unit length is These parameters can all be obtained from experiments using LJJ. These parameters correspond to the Josephson penetration depth. Plasma frequency Swirat speed (2.66% of the speed of light in a vacuum) and characteristic LJJ impedance It is also assumed that at low operating temperatures (with...) In units of ( ), compared with other parameters, the inherent dissipation constant It can be ignored, so it is set in the simulation. The goal is to make the operating frequency within a certain range. Internally, relative to the plasma frequency given above Normalized, the frequency is approximately For example from Figure 5 The analysis in section a shows that the ring-shaped LJJ is relative to Normalized resonant frequency Very close to Then select This was obtained The magnetic flux quantum number was also chosen as... This makes the first pair of degenerate static flux quantum oscillation frequencies of the ring-shaped LJJ at a frequency of 1. By setting and Numerical calculation of S matrix elements ,in and It's the external drive frequency. Also note... It contains two frequency components, one of which corresponds to the external drive frequency. And another frequency corresponds to the rotation frequency of the magnetic flux quantum around the ring LJJ. Therefore, a Fourier transform is used to extract... The intermediate frequency is The components are calculated, and the elements of the S matrix are determined.

[0089] exist Figure 6 In a and 6b, the elements of the S matrix are plotted. (Solid line), while the (normalized) current bias Around the resonance value, when they are zero and the optimal value, respectively. Change drive frequency As expected, Figure 6In a, the microwaves are scattered in completely opposite directions by the ring LJJ, where and The same applies because, in this case, the magnetic flux quantum is static, and the time-reversal symmetry of the system is not violated. Furthermore, in this case, it is found that... This means that the ring-shaped LJJ acts as a power divider, which divides the input power by a factor of 1 / 3. Transmitted to each output port and reflected to the rest .

[0090] and Figure 6 Conversely, when the current bias is finite and at its optimal value, in Figure 6 In b, at the resonant driving frequency Below Close to 1 and and A large non-reciprocal response was observed close to 0. The bandwidth at full width at half maximum (FWHM) extracted from the figure is... These findings demonstrate that, under appropriately selected operating conditions, a ring-type LJJ directly coupled to an external waveguide can operate as a high-fidelity microwave circulator over a fairly wide bandwidth.

[0091] exist Figure 6 In C, the driving frequency Fixed at the resonance value And plot the elements of the S matrix. relative to current bias It was observed that as the number of... , It gradually increased to nearly 1, while and Both decrease to near zero. near At this current bias, non-reciprocity is strongest. Under this current bias, the magnetic flux quantum velocity is: Swirart velocity. 3.57%, thus avoiding the undesirable relativistic effect in the ring-shaped LJJ. The average voltage generated by the circulation of magnetic flux quanta around the ring was also calculated using (14). Converted to the correct units, this voltage was found to be This would correspond to: through an impedance of waveguide dissipation The power output is [not specified]. Therefore, when connected to other devices, the ring-shaped LJJ operating as a microwave circulator is very unlikely to cause overheating problems.

[0092] Based on the S-matrix of the coupled-mode theory in the time domain

[0093] Note that only losses due to coupling to the external waveguide are considered here, as losses in the long Josephson junction are negligible. The scattering effect observed above can be explained by an effective three-port circulator model based on the time-domain coupled-mode theory. In this model, there exist frequencies respectively... and And the attenuation rate is and Two nearly degenerate inverse rotation modes are used. These modes are coupled to three external ports through which incoming and outgoing signals are input and output. By analyzing this model, an analytical expression for the S-matrix can be derived (derivation below). Specifically, the elements of the S-matrix... Given by the following formula (32) (33) (34) Identifying this model of a ring-type LJJ directly coupled to the waveguide reveals two nearly degenerate frequencies. and for , (35) in and These are two almost degenerate magnetic flux quantum oscillation frequencies, and It is the angular frequency of the magnetic flux quantum circulation around the ring LJJ. Due to symmetry, the attenuation rate can be assumed. and They are the same, and the attenuation rate is related to the waveguide coupling strength as follows. Related (36) This relationship is proved below. It will be given in the equation above. and Substitute these elements into equations (32)-(34) and recalculate the S matrix elements of the ring LJJ. .exist Figure 6 In the middle, a comparison was made The value is calculated using this circulator model (marked) and by numerically solving the perturbation SG equation (solid line) in (29). As can be seen, besides... Figure 6 The small region around the resonant driving frequency in a and Figure 6 Outside of the region near zero current bias in c, the two methods are very consistent with each other.

[0094] Besides approximating the elements of the S matrix, the circulator model based on temporal coupled-mode theory also allows for the acquisition of two important estimates. The first estimate is the scattering bandwidth, which is approximately... Using simulation parameters and Converted to MHz, the estimated bandwidth was found to be... This is related to Figure 6 The numerical results in b are consistent. The second estimate allows for determining the optimal current bias for maximizing the non-reciprocal scattering of microwaves by the ring LJJ. Specifically, in When, the S matrix elements in equations (32)-(34) To the greatest extent possible, it is non-reciprocal, and with respect to the toroidal LJJ system, the value is transformed into the following condition. (37) Here, , and It is current bias The implicit function; therefore, it is found that the above equations are implemented by... The ring-shaped LJJ will be provided as the optimal operating point for the high-fidelity circulator.

[0095] Tunability of operating frequency

[0096] The fact that the resonant frequency of the toroidal LJJ depends on the flux quantum number allows for dynamic tuning of the proposed circulator's operating frequency. This is achieved through repetition... Figure 6 In b, in the magnetic flux quantum number In the case of and The calculations performed prove this, and the fidelity of the S-matrix defined as follows is calculated for each case. (38) Due to the rotational symmetry of the ring-shaped LJJ, in fact .exist Figure 7 Three magnetic flux quantum numbers were plotted in the paper. fidelity And from The following is to The following clearly observes the crossing The three different corresponding operating frequencies within the range.

[0097] Power Analysis

[0098] The analysis focused on maintaining the device in such a state Figure 6 The optimal current bias value in b and such Figure 6The dependence of the simultaneous scattering performance at the resonant driving frequency in c on the input power. Figure 8 In section a, for both linear responses and the assumed nonlinear toroidal LJJ response, the output power as a function of the input power is plotted, the output power being based on... The calculated power is converted to appropriate units based on a given... It is converted to the appropriate unit. (Input power) A 1 dB compression point was identified. This input power was compared with the saturation power estimated based on the lifetime of single-photon excitation, which was determined by... Given, among which and From Figure 6 The working bandwidth extracted from b. (Found) Therefore, the value is higher than The difference is 20 dB. This indicates that the ring-shaped LJJ, operating as a microwave circulator, has a large dynamic power range.

[0099] exist Figure 8 In b, the elements of matrix S are plotted. Relative to input power. It was observed that when the input power was lower than... At this point, non-reciprocal scattering remains essentially unchanged. When the input power exceeds this value, scattering begins to saturate. Specifically, when the input power is below this value... hour, and They become very close to each other, which means that the non-reciprocity effect disappears. This analysis is consistent with... Figure 8 1 dB compression point in a Consistent.

[0100] in conclusion

[0101] Realizing microwave circulators that can be easily integrated on a chip and operate passively (affected by bias current) greatly facilitates the scaling up of quantum technologies based on solid-state systems such as superconducting circuits. This disclosure illustrates how to design such a device. Specifically, the disclosed circulator utilizes the directional motion of magnetic flux quanta in a toroidal LJJ to induce non-reciprocal scattering of microwaves. Under appropriately selected operating conditions, this system exhibits a circulation fidelity close to 1. This makes the proposed circulator superior to other on-chip designs. First, the circulator is a device that uses only a single DC control field, which greatly reduces the hardware complexity when actually operating the device. Second, in a circulator made of a ring of three small discrete Josephson junctions, it is less prone to generating harmful charge noise. Third, the operating frequency of the proposed circulator can be tuned by dynamically changing the magnetic flux quantum number. Another advantage is the large dynamic range compared to designs with discrete Josephson junctions. Long Josephson junctions have very large critical currents, so the circulator will operate at much higher power compared to devices with several single Josephson junctions.

[0102] Derivation of the sine-Gordon equation

[0103] Perturbation SG equations without waveguide coupling

[0104] Here, the perturbation SG equation in (11) is derived, which includes external current bias and dissipation. For this purpose, LJJ is discretized into a length of... Discrete unit cells, each associated with a node flux quantum, such as Figure 9 As shown in a. The Josephson critical current for each node in the RSCJ model is... Parasitic capacitance is And the shunt resistance is The small Josephson junction is connected to ground. The two consecutive nodes are connected via surface inductance. and surface resistance They are coupled together.

[0105] Consider its node flux as nodes Its Lagrangian function is given by the following equation. (39) With nodes The associated dissipation function is (40) And with nodes bias current at The associated driving function is (41) node The equation of motion at point A is given by the generalized Lagrange equation. (42) The above text gives , and Substituting into (42) yields (43) Equation (43) can actually be solved by applying Kirchhoff's current law. Figure 9 Nodes in a This is directly derived. The quantity per unit length is defined as follows: , , , and ,in , , , and These represent the junction capacitance per unit length, critical junction current density, surface inductance, surface resistance, and bias current, respectively. Flux is also converted to phase. Therefore, equation (43) can be rewritten as follows: (44) The equation, when both sides are divided by... Later obtained (45) in It is a sheet resistor. Alternatively, it can be used... (i.e., the normal conductance per unit length of the junction) to limit Dissipation terms The normal conductivity is rewritten as Take the limit Equation (45) becomes (46) Divide both sides of this equation by In order to obtain (47) in Continue with the time Relative to plasma frequency Reciprocal normalization, and position Relative to penetration depth Normalization This absorbs all the previous factors in equation (47) and simplifies it to (48) in and This represents the dissipation constant. In practice, it is related to... compared to, They are usually very small, so the terms in the above equations can be ignored. Thus, the perturbation SG equation in (11) is obtained precisely.

[0106] Perturbation SG equations using direct waveguide coupling

[0107] Assuming the node Directly coupled to current Associated waveguides, such as Figure 9 As shown in b. Flux At the node The equation of motion at point RHS is the same as that in equation (43), but RHS also includes... (49) (The derivation is shown below.) (50) (51) Convert the quantity to its per-unit length version, and convert the flux. Convert to phase And divide both sides of the equation by As accomplished in equations (44) and (45), equation (49) can be rewritten as follows: (52) in It is a dimensionless input voltage. Again... Relative to penetration depth Normalize, and will Relative to plasma frequency Normalize the reciprocal and divide all terms by This rewrites the last term on the RHS of equation (52) as follows: , And the input-output relationship in equation (51) is rewritten as follows: .

[0108] Take the continuous limit And noticed (This is the characteristic LJJ impedance), found , in and .

[0109] These explain the form of the perturbation SG equation in (29) in the presence of direct waveguide coupling and the input-output relationship in (30).

[0110] Long Josephson junction coupled with external waveguide capacitor

[0111] At the node At this point, the LJJ capacitor is coupled to the first... In the case of a single waveguide, phase and waveguide voltage The equation of motion is (53) (54) in and Already relative to Normalization. Relative to penetration depth Normalize, and will Relative to plasma frequency Normalize and divide all terms by This rewrites the last term of equation (53) as follows: .

[0112] Take the continuous limit And noticed This item becomes , in ,in At the same time, equation (54) is changed to (55) in .

[0113] The perturbation method is used for capacitive coupling. We consider weak capacitive coupling, i.e. (See below). The nodes in equation (53) The equation of motion is changed to .

[0114] Will Relative to penetration depth Normalize, and will Relative to plasma frequency Normalize and divide all terms by This rewrites the last term of equation (53) as follows: .

[0115] Set constant and Rewrite this item as (56) Jacobian function Jacobi Elliptic Function Here, the definition and properties of the Jacobian elliptic function are presented. There are three fundamental functions. , and , where u is the independent variable, and These are elliptic moduli. These functions are derived from elliptic integrals of the first kind. (57) in and It is the Jacobian amplitude, which is obtained (58) The following can be concluded (59) (60) (61) These functions are supported in MATHEMATICA, with function names JacobiSN, JacobiCN, and JacobiDN, but the modulus symbol is different. Instead Also note that when writing the Jacobian function, for example, or At that time, modulus It is usually suppressed.

[0116] The Jacobian elliptic function is a double-periodic generalization of trigonometric functions, satisfying... (62) (63) (64) and (65) (66) (67) in It is a first-kind complete elliptic integral, and in In the case of (i.e., complementary elliptic moduli), In MATHEMATICA, It is called EllipticK.

[0117] Jacobi elliptic functions also satisfy the following identities (68) (69) (70) (71) Its about The derivative is given by the following formula. (72) (73) (74) Jacobi's η ( ), θ ( ) and ζ ( )function Jacobi's η ( ), θ ( ) and ζ ( The function is defined as (75) (76) , (77) in The expression given by the following formula represents the first... A theta function [?] (78) (79) (80) and (81) In equation (77), It is an elliptic integral of the second kind, and It is a complete elliptic function of the second kind. In Mathematica, Known as EllipticTheta ,and Known as EllipticE ,in Jacobi's η ( ) and θ ( The function satisfies the periodicity condition (82) (83) Properties of multiflux quantum solutions Prove the form of the static multiflux quantum solution in (3) In the following text, it is proven that the limitation by (3) It is a true static solution to (1). Regarding Differentiating (3) reveals (84) The identity based on (3) has already been used. .about Differentiating equation (84) yields , (85) This proves It is the static solution of (1).

[0118] Prove the boundary relationship in (5)

[0119] The knot length in (5) was derived based on the boundary conditions. The number of magnetic flux quanta captured With modulus The relationship between them. Using equations (2) and (3), we obtained (86) According to equation (65), this means (87) This is exactly the relationship mentioned in (5) above.

[0120] Prove the Lorentz invariance of the SG equation

[0121] It has been proven that under the Lorentz transformation defined in equations (6) and (7), the SG equation in (1) is invariant, i.e., it shows that... or equivalent Using equations (6) and (7), we find (88) (89) Make (90) Prove the current-dissipation relationship in (12) Here, the current-dissipation relationship in (12) is proved. The current bias that provides energy to the system. Dissipation of energy in a dissipative system The power balance condition between them is given by the following equation. (91) Among them, (4) is limited to and use Represented as Here, the average value Defined as (92) in yes Quasi-period, i.e. Evaluation and And discovered (93) (94) Substituting these results into (91) yields the current-dissipation relationship in (12).

[0122] Another method to prove the current-dissipation relationship is to use another form of power balance condition. (95) in The total energy of LJJ is described by the following formula. (96) Intuitively, (95) means that the energy change in the LJJ system is equal to the sum of the energy injected from the current bias and the energy dissipated due to dissipation. When the moving magnetic flux quantum reaches the solution described above... The steady state can be specifically calculated. (97) (98) (99) These, combined with the conditions in (95), precisely produce the current-dissipation relationship.

[0123] Derive the dissipation constant caused by an effective uniform waveguide

[0124] First, we review the perturbation SG equations in (29) with waveguide coupling. We assume there is no input voltage, i.e. And there is no inherent system dissipation, i.e. Therefore, (29) is rewritten as (100) in (101) Let represent the dissipation constant caused by the non-uniform waveguide. In the long-time limit, the steady-state solution of (100) will be the same as the steady-state solution of the following equation. (102) in It is the dissipation constant caused by an effective homogeneous waveguide. In the following text, it is determined that... and The relationship between them.

[0125] First, note that (102) has the same form as (11), and therefore also has the current-dissipation relationship as in (12). (103) Then, repeat the power balance calculation from the previous subsection for (102) to establish and The relationship between them. Specifically, to make (104) because As given in equation (98), therefore only needs to be applied to evaluate And discovered (105) Therefore, we can conclude that (106) Comparing equations (103) and (106), we see... (107) Note the information given in (10) This relationship is further simplified to (108) The interpretation of this result is quite intuitive, namely the uniform dissipation constant. Equal to the length of the knot Total dissipation uniformly distributed on the upper (The factor of 3 is due to the waveguide coupling to three points).

[0126] When identifying the three-port circulator model of a ring LJJ directly coupled to the waveguide from time-domain coupled-mode theory (see below), the resonant modes in the circulator model will see a power attenuation rate equal to (Right now, Therefore, the resonant mode will see an amplitude decay rate equal to... (Right now, This explains the results in (36).

[0127] Solve the Lamé equation in (16)

[0128] Here, the general solution of equations (18)-(20) of the Lamé equation (!6) is derived. First, the small oscillations in (15) are... Written (109) And simplify (16) to (110) Note that in (3) it is given And introduce new variables In the case of Rewrite (110) as (111) in The general solution of (111) It can be found in the following forms (112) in and Let represent the Jacobian η and θ functions given below, and and These are parameters to be determined later. Take the logarithm and about Differentiation (112) yielded (113) about Differentiating this equation yields (114) The following identities were used in deriving the above two equations. (115) (116) (117) Equations (113) and (114) imply that (111) is a true solution to (112) only if the following conditions are met. (118) The equation is simplified to (119) in (120) (121) The two sides of equation (119) were analyzed around the poles, and it was found that... (122) (123) in (124) (125) review ,Discover (126) The given expression is equation (20). Using the above results, it was found that... Presented in the following forms (127) In use replace At that time, it was discovered And therefore It was changed to (128) These explain the problems in equations (18) and (19), respectively. and In the form of.

[0129] Classical Input-Output Theory

[0130] Here, we introduce the classical input-output theory.

[0131] Lumped element circuit model of transmission line

[0132] Figure 10 a shows the definitions of voltage and current for the transmission line. Figure 10 b shows a lumped element circuit model of the transmission line. Here, It is the self-inductance per unit length of the two conductors. It is the shunt capacitance per unit length generated by two conductors being brought close together. It is the resistance per unit length due to the finite conductivity of a single conductor, and It is the shunt conductance per unit length caused by the dielectric loss of the material between conductors. and Indicates loss.

[0133] according to Figure 10 b, Kirchhoff's voltage law gives... (129) Furthermore, Kirchhoff's current law gives... (130) Divide the two equations by And take the limit The telegraph operator's equation was obtained. (131) (132) For the sinusoidal steady-state condition, in the case of cosine phasors, equations (131) and (132) simplify to (133) (134) Transform equations (133) and (134) into wave equations (135) (136) in (137) It is a function of the complex propagation constant and frequency. The traveling wave solution is given by the following equation. (138) (139) in Indicates wave direction Directional propagation, and Conversely, using equations (133) and (138), we find... (140) This leads to the discovery of characteristic impedance. for (141) Termination of lossless transmission line Hypothetical form The incident wave is from It is generated by the source at that location. We know that the voltage to current ratio of this traveling wave is... This is the characteristic impedance of the wire. However, when the wire is terminated with any load... When the load is in operation, the voltage to current ratio must be... The total voltage of a lossless transmission line is (142) And the current is (143) The total voltage and current at the load are related through the load impedance, therefore... At that time, (144) right The solution was obtained (145) Then, the voltage reflection coefficient is given by the following. (146) Therefore, the voltage and current waves on the line are written as (147) (148) This indicates that voltage and current are the superposition of incident and reflected waves, meaning that the incident and reflected waves are standing waves. Only when... Only then will there be no reflected wave, and the reflected wave will be in This may occur during impedance matching.

[0134] Calculate the points along the line Time-averaged power flow at: (149) This indicates that the average power flow is constant at any point on the line, and that the total power equals the incident power. Subtract reflected power .if Then, maximum power is delivered to the load, while for It does not deliver power.

[0135] When there is a load mismatch, not all available power from the generator is delivered to the load. This loss is called return loss (RL) and is defined as follows: (150) Make the matching load ( The RL of ) is (No reflected power), while total internal reflection ( The return loss is (All incident power is reflected).

[0136] Special cases of lossless terminal wiring

[0137] Consider the case where the transmission line is short-circuited (i.e., terminated only through the wire). According to (156), it was found that... Then the voltage and current in equations (147) and (148) are... (151) (152) Consider opening routes, among which This was obtained The voltage and current on the line are... (153) (154) Consider characteristic impedance Transmission lines fed with different characteristic impedances The line, such as Figure 12 As shown. If the load line is infinitely long, or if its termination is in its own characteristic impedance such that there is no reflection from its far end, then the input impedance seen through the feed line is This makes the reflection coefficient yes (155) According to equation (147). The voltage is , (156) In the absence of reflection The voltage wave is only transmitted and can be written as , (157) Make these voltages equal to The transmission coefficients are given below. (158) The transmission coefficient between two points in a circuit is expressed in dB as insertion loss (IL): (159) Scattering matrix consider Figure 13 shown Port network, where It is incident at the port The amplitude of the voltage wave on it, and From port The amplitude of the reflected voltage wave. In the... At specific points on each port, a terminal plane is defined. and incident and The equivalent voltage and current of the wave. The terminal plane is important in providing a phase reference for the voltage and current phasors. Now, in the... At each terminal phasor, the total voltage and current are given by the following formula. (160) (161) Scattering matrix or The matrix is ​​defined relative to these incident and reflected voltage waves as (162) or (163) Specific elements of the scattering matrix are defined as (164) This means By using voltage Incident wave drive port And measure from the port The amplitude of the reflected wave Discovered. Besides the first... Except for one port, the incident wave on all other ports is set to zero, meaning that all ports should be terminated with matched loads to avoid reflections. Therefore, It is looking at the port when all other ports are terminated in a matching load. The reflection coefficient, and It is when all other ports are terminated in a matching load from the port. to port The transmission coefficient.

[0138] Consider the example circuit Figure 14 The example circuit is a 3 dB attenuator circuit. When a matching load is connected to port 2 ( The reflection coefficient observed at port 1 at the time of ( ) (165) yes With series ( / / (8.56 and (in series), so (166) This means and (Because of the symmetry of the circuit).

[0139] This is the transmission coefficient from port 1 to port 2: (167) according to The fact is that when port 2 is connected to China Times, In the aforementioned case, make and .

[0140] By applying voltage at port 1 And by using two voltage dividers, it was found that... As port 2 is located in Voltage across the load resistor: (168) Make If the input power is Then the output power is This is half the input power (-3 dB).

[0141] Reflection coefficient of an LC circuit coupled to a waveguide

[0142] In the following text, we consider LC circuits that are directly or capacitively coupled to the waveguide, such as... Figure 15 As depicted in [the text]. Given the input voltage injected into the circuit from the waveguide. The output voltage was calculated using two different methods. And the reflection coefficient. Regarding the first method, the equation of motion for the primary node flux of the coupled LC circuit is derived, and the output voltage is calculated. Based on this, the reflection coefficient was evaluated. For the second method, the reflection coefficient is calculated directly by analyzing the total load impedance. Finally, Fourier transforms show that the two methods are correlated.

[0143] LC circuits directly coupled to the waveguide

[0144] consider Figure 15 The circuit in a, where the flux node Kirchhoff's current law was generated at that point. (169) in and It also has (170) (171) This discovery (172) Then rewrite (169) as (173) Substitution And targeting Solve the above equation. In the long time limit, when... When it reaches its steady state, calculate And thus the reflection coefficient is calculated. .

[0145] The load impedance analysis presented earlier can be used to calculate differently. .calculate Figure 15 The total load impedance of the LC circuit in section a is given by the following: (174) The reflection coefficient defined in equation (146) is given by the following: (175) In fact, the above-given value can be derived from the Fourier transform of (173). The Fourier transform is given by the following equation. (176) in express Fourier transform to the frequency domain. Using this equation, regarding right Solve (177) Note (Right now, Fourier transform), found (178) in It is exactly the same as in (175).

[0146] An LC circuit with capacitive coupling to a waveguide.

[0147] consider Figure 15 The circuit in b, the circuit Kirchhoff's Current Law at Two Nodes (179) (180) in and It also has (181) (182) Using the aforementioned equations, equations (??) and (180) can be rewritten as follows: (183) (184) Substitution And targeting and Solve the above equation. In the long time limit, when... When it reaches its steady state, calculate And thus the reflection coefficient is calculated. .

[0148] The load impedance analysis presented earlier can be used to calculate differently. .Discover Figure 15 The total load impedance of the LC circuit in b is given by the following... (185) The reflection coefficient is given by the following from this equation. (186) Derivation It is given above by the Fourier transform of equations (183) and (184), which is given below. (187) (188) Using these equations, regarding right and Solve and discover (189) Note ,Discover (190) in It is exactly the same as in (186).

[0149] Perturbation Method for Capacitive Coupling. It has been seen that finding the dynamics of an LC circuit capacitively coupled to a waveguide involves solving two coupled equations of motion for the node fluxes of the two circuits, one for the LC circuit and one for the waveguide (see equations (183) and (184)). Below, weak capacitive coupling is considered, and a perturbation analysis is proposed, in which the dynamics of interest concern only the node fluxes of the LC circuit. Specifically, the weak coupling limit is given by the condition... This indicates. Here, This represents the small time delay between the input field and the reflected output field. (And the driver) Compared to the period, this delay should be very small, allowing for Markov approximation. .

[0150] Because of the search for LC circuits for external drive The steady-state response, therefore, can be assumed to be the flux at LC nodes. and output voltage Will mainly follow Oscillation. Substitution can be performed in this way. In the case described, It can be or ,and Compared to external drive It changes slowly. Substitute this value into equation (184) and replace... And in small parameters China's target Solve using a perturbation approach. It was found that at the lowest order... middle, (191) At the next level, it was discovered .

[0151] because The assumption of slow change is used to discard the last two terms on the RHS to obtain... (192) This discovery .

[0152] (193)

[0153] and (194) This changes equation (183) to (195) Temporal Coupled-Mode Theory of Circulators consider Figure 16 The three-port circulator in the image has two nearly degenerate counter-rotating modes coupled to three ports. and The dynamics of this system are described by the time-domain coupled-mode theory equations. (196) (197) Here, ( () is a clockwise (counterclockwise) rotation mode at a frequency of ( And the decay rate is ( The normalized amplitude under the condition of ) and ( ) is a port The normalized amplitude of the incoming (outgoing) wave. and This represents the coupling between the resonant mode and the wave at the port. matrix. It is the direct reflection and scattering matrix between the incoming wave and the outgoing wave.

[0154] Using constraints on time-domain coupled-mode theory in energy-saving systems, it was found that... (198) (199) These conditions and systems The combination of rotational symmetry produces (200) By driving frequency By performing a Fourier transform and solving equations (196) and (197), the scattering matrix that correlates the incoming and outgoing waves is found. for (201) Elements of the S matrix in equations (32)-(34) The parsing expression is derived from It can be deduced that, that is, .

[0155] parameter

[0156] The inductance per unit length is given by the following , (202) in This is the average distance to London. It is the barrier thickness. It is the barrier width, and The capacitance per unit length is related to the electrical energy stored in the potential barrier, and is given by the following formula: , (203) in and It is the relative permittivity of the barrier oxide. The series resistance per unit length as seen through the surface current is... (204) in It is a relaxation frequency of approximately several hundred Gcycles. The dissipative current component passing through the potential barrier is... Given, among which The unit of electrical conductivity is m.

[0157] Two dissipation coefficients and The following is given (205) , (206) in , (207) (208) In AC Scott, FYF Chu, and SA Reible, Journal of Applied Physics 47, 3272 (1976), at temperature , Several measurements were performed. The basic data for the Josephson line are shown in Table I, and its derived parameters are shown in Table II. Note that in this paper, It is considered as 0.

[0158] Table I. Basic data for the Josephson Line.

[0159] Table II. Parameters derived from the Josephson line.

[0160] Measurement can be performed at temperature The following steps will be performed. The knot length is 200. The critical current density is , And the plasma frequency is Experimental fit: and , In other instances, the simulation parameters can be: and It can be used for parameters. and Further simulations were performed on the Nb junction, and the magnetic flux quantum velocity was approximately... In yet another instance, simulation can be performed under the following conditions: In yet another instance, the experiment can be conducted under the following conditions: , , The critical current at zero field is In yet another example, the parameters can be as follows: the critical current density of the Josephson transmission line ( )for Josephson's penetration depth was... And the plasma period is That is, the plasma frequency is . This is the characteristic impedance of a long junction. The junction length is... Width is . The estimate is 0.005, and Assuming .

[0161] This may be associated with quasi-particle tunneling loss, and the junction length can be... , and The Josephson current density can be Josephson penetration depth can be Furthermore, the Josephson plasma period can be ( In another example, the knot length is 5 mm and the width is... The characteristic impedance of a long junction is .

[0162] In yet another example, the critical current density is Penetration depth is The plasma frequency is The characteristic impedance is And the knot length is The knot is 1 cm long and its width is... . , and .

[0163] In yet another example, the Josephson current density is Penetration depth is The plasma period is ( ), and the characteristic impedance is . and In another example, the penetration depth is... And the plasma period is .

[0164] In another example, the critical current density is Penetration depth is And the length of the ring knot is The sample temperature was .

[0165] In another example, the critical current under zero field is approximately Penetration depth is And the plasma frequency is .

[0166] Damping parameters

[0167] It is said It is associated with the tunneling loss of quasiparticles. Specifically, and Proportional, of which It is a temperature-dependent quasi-particle resistance (209) Here, It is the normal junction resistance when the current bias exceeds the junction critical current.

[0168] and Both can be determined from the junction's IV characteristic curve. For example, see... Figure 18 The resistance in the RCSJ model of the Josephson junction The value may be approximately ,near However, at lower temperatures, for high quality... Knot, along with Rise (for) Due to damping parameters and It is inversely proportional, therefore it can be safely assumed that at very low temperatures, .

[0169] Note that superconducting materials can be selected from a range of different materials, such as Al, Nb, YBCO, BPB, LTO, etc., while the barrier can be made from a range of different insulators, such as AlOx or MoS2.

[0170] Furthermore, the disclosed circulator can be used as part of a quantum computer, such as one using superconducting qubits or other physical platforms (nitrogen vacancies, phosphorus donors, photons, neutral atoms, trapped ions, etc.). In one example, the circulator is a three-port circulator, where one port is connected to a readout circuit. The qubit is connected to a readout resonator, and the readout signal from the qubit passes through or through the readout resonator, imprinting information about the state of the qubit. The signal is then passed through the circulator (or a series of circulators) to a higher-temperature stage and a classical amplifier. In this example, the circulator acts as an "isolation": the signal is transmitted to the amplifier with low loss from the resonators at ports 1 and 2. Noise from the amplifier is diverted from port 2 to port 3, where a matched load is connected. The matched load absorbs the noise so that it does not reach the qubit.

[0171] In another example, a quantum amplifier suitable for reflection exists. In this case, the readout signal from the qubit (containing information about its state) enters the first port of a circulator. The circulator then transfers the signal to a second port where the quantum amplifier is connected. The amplified signal is reflected from the amplifier and then transferred to a third port, which is connected to a higher-temperature stage and another amplifier.

[0172] In another example, the circulator is a four-port circulator. The first port is connected to the qubit to receive the qubit measurement result and transmits the qubit measurement result to a quantum reflection amplifier, which reflects the amplified quantum measurement result. The circulator transmits the amplified measurement result to a classical amplifier connected to the third port.

[0173] One challenge is that classical amplifiers generate significant noise that can corrupt quantum coherence. However, circulators transmit the reaction to the third port of a three-port circulator or the fourth port of a four-port circulator, where the reaction can be safely handled by readout electronics or dumped into an ambient bath. Thus, the quantum circuit is protected from the harmful noise generated by classical amplifiers or higher-temperature stages. It should be noted that not only amplifiers can be shielded, but resistors also introduce noise at higher temperatures that can corrupt quantum coherence. Quantum bits or lower-temperature stages can also be protected from those components by one or more circulators.

[0174] This demonstrates that the number of circulators increases with the number of qubits, making it desirable to construct small circulators that can be easily integrated. As discussed in this paper, the disclosed circulators containing toroidal long Josephson junctions do not require bulky AC electronics and can therefore be miniaturized and integrated into multi-qubit quantum computers.

[0175] layout

[0176] Figure 18 An example layout 1800 of a circuit that has been fabricated for experimental use is shown. Layout 1800 includes a first circulator circuit 1801, which includes a ring structure 1802 and three ports, including a first port 1803, a second port 1804, and a third port 1805, for transmitting and / or receiving electromagnetic signals. The ring structure 1802 connects the three ports and includes a ring structure comprising a long Josephson junction, the ring structure being configured to conduct moving magnetic flux quanta around the ring structure 1802 to transmit electromagnetic signals to one of the three or more ports and suppress electromagnetic signals at another of the three or more ports.

[0177] The movement of the magnetic flux quantum is controlled by a bias current provided by a first pair of bias contacts 1806, a second pair of bias contacts 1807, and a third pair of bias contacts 1808. For each pair, the bias current flows into one contact and exits through the other. In this way, the bias current can be independently controlled in each segment of the ring structure 1802 located between any two signal ports.

[0178] Current bias provides the coercive function that drives the flux quantum motion. Spatially isolated point-like bias configurations provide point-like coercive functions for flux quantum motion, while increasingly uniform bias (or a denser arrangement of current bias points around the loop) provides a more uniform and therefore spatially regular coercive function. The more spatially uniform the flux quantum force, the smoother it can be controlled, manifested as a smaller recapture current and a smooth transition of the DC junction voltage from the flux quantum plateau to zero.

[0179] Simulation results suggest that maintaining the 120-degree rotational symmetry of the loop may be advantageous for the circulation induced by the flux quantum. This would mean that all microwave ports, flux injectors, and current bias points are arranged with this symmetry. Figure 18 The figure shown depicts this.

[0180] A second circuit 1810 is present, identical to the first circuit and included as an additional test circuit. In a corner of layout 1800, a characterization circuit 1811 is present, used to measure outputs of the manufacturing process, such as the quality of the Josephson junction included in the test circuit 1811.

[0181] Note that circulators with more than three ports can be constructed. One way to achieve this is by daisy-chaining two or more circulators, where each circulator connected in the daisy chain adds two additional ports.

[0182] Measurement results

[0183] Figure 19 The data are presented as current-voltage curves in a side-by-side comparison between experiments and theory, showing the generation and movement of flux quanta at a single bias current point, with experimental data on the left and theoretical simulations on the right. These results show a comparison of the time-averaged (DC) long junction voltage (x-axis) for different bias current values ​​(y-axis). Note that for these graphs, the independent variables are on the y-axis. The left graph shows the experimental data in dimensional units, and the right graph shows the experimental simulations in dimensionless units. The different vertical asymptotes of each curve indicate different quantities of "flux quanta," which depend on the flux injector current or voltage. The results show that the flux quanta are sufficiently quantized and can be controlled in a predictable manner. Furthermore, the results demonstrate that the simulations are correct in predicting the behavior of the fabricated circuits.

[0184] Figure 20 A typical 2D scan of the injection voltage and current bias is shown: the injection voltage is fixed, then the current bias is scanned while the sensed voltage is measured. The top row shows the experimental measurements from the fabricated device, while the bottom row shows the measurements from the simulated device. These results show a Fraunhofer-like pattern (f(x) = |sin(x) / x|) for the recapture current in both experiments and simulations. The data is obtained by fixing the flux injection voltage at a certain value (x-axis) and scanning the bias current from zero to a high value to obtain the depinning current value, then returning to zero to obtain the recapture current. The results show that the simulation is correct in predicting the behavior of the fabricated circuit.

[0185] Figure 21 Simulations are shown demonstrating that three symmetrical bias points function as well as uniform bias. These graphs include theoretical calculations of the predicted flux quantum current-voltage depinning and retracing responses for different current bias configurations. These configurations, from top to bottom, are: single-point bias, two-point bias, three-point bias, and uniform bias. These units are dimensionless.

[0186] Furthermore, during the operation of the circulator, the bias current remains constant at its optimal constant value, where a specific number of flux quanta circulate at a specific velocity. The motion of these flux quanta causes a split between two resonant modes, such as... Figure 5 As shown, this promotes signal circulation. In Figure 21 In the graphs showing the measurement and simulation results, a Fraunhofer pattern as a function of the injection voltage can be observed. After each peak of the Fraunhofer pattern, for the three bias ports, as shown... Figure 21The bias current (e.g., between 0.2 and 0.3) indicated by the dashed line on the right side of the second-to-last line is a value for which the measured voltage gradually increases. This corresponds to an integer number of flux quanta inserted into the LJJ by the injector and circulating close to the Swirhart velocity, and can be used to select the appropriate injector voltage. After inserting the desired number of flux quanta, the optimal current bias can be selected using microwave measurements to identify the splitting of the two resonant modes. In other words, the bias current can be adjusted while observing the microwave measurements, and when the two resonant modes split, the appropriate bias current has been reached and can be used for the operation of the circulator.

[0187] Those skilled in the art will understand that many changes and / or modifications can be made to the above embodiments without departing from the broad overall scope of this disclosure. Therefore, the embodiments of the present invention should be considered illustrative rather than restrictive in all respects.

Claims

1. A circulator for electromagnetic signals, the circulator comprising: Three or more ports, said three or more ports being used to transmit and / or receive said electromagnetic signals; and A ring structure for connecting the three or more ports. The ring structure includes a long Josephson knot, and The ring structure is configured to conduct moving magnetic flux quanta around the ring structure to transmit the electromagnetic signal to one of the three or more ports and suppress the electromagnetic signal at another of the three or more ports.

2. The circulator of claim 1, wherein the ring structure comprises a single long Josephson knot along the entire ring.

3. The circulator of claim 2, wherein the Josephson junction comprises an annular isolation layer located between two annular metal layers.

4. The circulator according to any one of claims 1 to 3, wherein the circulator structure comprises a synthetic long Josephson junction formed by a plurality of discrete Josephson junctions.

5. The circulator according to any one of the preceding claims, wherein the circulator further comprises a bias current source for generating a bias current that maintains the motion of magnetic flux quantum.

6. The circulator of claim 5, wherein the bias current is based on the frequency of the electromagnetic signal.

7. The circulator according to claim 6, wherein the bias current is determined based on the relationship between the bias current and the electromagnetic signal.

8. The circulator according to any one of claims 6 to 8, wherein the circulator is switchable by changing the bias current source.

9. The circulator according to any one of the preceding claims, wherein the circulator further comprises three or more waveguides for coupling the electromagnetic signal to the ring.

10. The circulator according to any one of the preceding claims, wherein the circulator further comprises a current-generating source for providing current through one electrode of the long Josephson junction to generate the magnetic flux quantum in the circulator structure.

11. The circulator according to any one of the preceding claims, wherein the circulator comprises a flux injector for generating the magnetic flux quantum in the circulator structure.

12. The circulator according to any one of the preceding claims, wherein the moving magnetic flux quantum generates a phase shift in the electromagnetic signal to produce interference, the interference transmitting the electromagnetic signal to an output port adjacent to the input port and suppressing the transmission of the electromagnetic signal to all other output ports.

13. The circulator of claim 12, wherein the interference is constructive at adjacent output ports and destructive at other ports.

14. A method for preparing a circulator for operation according to any one of claims 1 to 10, the method comprising: Above the temperature at which the ring structure is superconducting, a magnetic field is applied to the circulator to generate the magnetic flux quantum; The ring structure is cooled to maintain the magnetic flux quantum until the superconducting temperature is reached; as well as Disconnect the magnetic field to generate self-sustaining magnetic flux quanta in the superconducting ring structure.

15. A quantum computer comprising a plurality of qubits and a plurality of circulators according to any one of claims 1 to 11, the plurality of qubits being used to perform quantum operations.

16. The quantum computer of claim 15, wherein the qubit is a solid-state qubit operating at microwave frequencies, including superconducting qubits or semiconductor qubits.

17. The quantum computer of claim 15 or 16, wherein each circulator in the plurality of qubits is connected to one of the qubits to protect the one of the qubits from noise generated by an amplifier, the amplifier being used to amplify the signal generated by the one of the qubits.