betaFeedback

WorksPhysics and Astronomy

Time Domain Design of a Josephson Parametric Amplifier and Comparison with Input Output Theory

Abstract

Quantum-limited amplifiers, such as Josephson Traveling Wave Parametric Amplifiers (JTWPAs) and Josephson Parametric Amplifiers (JPAs), are essential components in quantum computers. They amplify low-power microwave signals from qubits at the 10 mK stage before further amplification at the 4 K stage using HEMT amplifiers. In JPAs, parametric amplification is based on the nonlinear properties of Josephson Junctions. While JPAs are typically designed and analyzed using input-output theory based on quantum physics, we propose an alternative approach based on an equivalent circuit model of JPAs, implemented using open-source Josephson circuit simulators. We compare the results with those obtained from input-output theory. This method enables the use of circuit optimizers for various objective functions and significantly reduces design time compared to quantum theory-based approaches.

Introduction

In the literature, Josephson parametric amplifiers have been extensively investigated [1, 2, 3, 4, 5]. They are used as quantum limited amplifiers or as a non-classical light source and have multiple modes of operation such as degenerate, non-degenerate, or 4 wave mixing and 3 wave mixing modes [6, 7, 8]. Although they are characterized as weakly nonlinear quantum harmonic oscillators rather than strongly nonlinear qubits, they essentially resemble the same architecture [9]. Input-output theory, a widely used framework in quantum optics, models the dissipative interaction of the cavity with its surrounding environment, which is treated as a set of independent harmonic oscillators [10]. This approach operates in the Heisenberg picture, enabling the derivation of equations of motion for the intra-cavity field. These can then be related to the input and output fields via boundary conditions. The formalism remains accurate even in the presence of coupled or complex cavity configurations [11], [12]. Interestingly, the semi-classical approach used in the analysis naturally lends itself to implementation with circuit simulators [13]. In this work, we use open source Josephson circuit simulators such as Jsim and JoSIM [14], [15] to characterize the device and compare the results with linearized and generalized quantum treatment. We also derive the equation of motion for JPA flux and compute it in Matlab to demonstrate the agreement

JPA Design Based on Input–Output Theory

To derive a Hamiltonian of an arbitrary circuit, we follow the steps explained in [23]. We represent the JPA as a parallel nonlinear LC circuit and use the highlighted capacitive spanning tree to derive the Hamiltonian of the circuit given in figure 1 to derive the associated Hamiltonian as in equation (1).

Figure 1. Circuit considered for quantization

H=C2Φ˙2−EJcos⁡(2πΦ0(Φ−Φext))(1)\mathcal{H} = \frac{C}{2}\dot{\Phi}^{2} - E_J \cos\left(\frac{2\pi}{\Phi_0}\left(\Phi-\Phi_{\mathrm{ext}}\right)\right) \tag*{(1)}

Where, Φ\Phi is the flux, Φ0\Phi_0 is the single flux quantum, Φext\Phi_{\mathrm{ext}} is the external flux, EJE_J is the Josephson energy. In the equation, 1st1^{st} term represents the energy associated with the capacitor, CC, and 2nd2^{nd} term represents the energy associated with the nonlinear inductor implemented by a Josephson junction. And because single junction potential is symmetric, RF driving the circuit is enough to induce the 4 wave mixing process without triggering any 3 wave mixing process which makes it convenient to drop the external flux dependence. In this case, one obtains the Hamiltonian given in equation (2).

H=C2Φ˙^2−EJcos⁡(2πΦ0Φ^)(2)\mathcal{H} = \frac{C}{2}\hat{\dot{\Phi}}^{2} - E_J \cos\left(\frac{2\pi}{\Phi_0}\hat{\Phi}\right) \tag*{(2)}

This Hamiltonian can be quantized using Dirac’s method. First, we introduce flux (Φ^)(\hat{\Phi}) and charge operators (Q^)(\hat{Q}):

(a)Φ^=Φzpf(a^†+a^)=ℏZ02(a^†+a^),(b)Q^=iQzpf(a^†−a^)=iℏ2Z0(a^†−a^).(3)\begin{aligned} \text{(a)}\quad\hat{\Phi} &= \Phi_{\mathrm{zpf}}(\hat{a}^{\dagger}+\hat{a}) = \sqrt{\frac{\hbar Z_0}{2}}(\hat{a}^{\dagger}+\hat{a}),\\ \text{(b)}\quad\hat{Q} &= iQ_{\mathrm{zpf}}(\hat{a}^{\dagger}-\hat{a}) = i\sqrt{\frac{\hbar}{2Z_0}}(\hat{a}^{\dagger}-\hat{a}). \tag*{(3)} \end{aligned}

Where Φ^zpf\hat{\Phi}_{zpf} and Q^zpf\hat{Q}_{zpf} represent the zero-point fluctuations of the flux and charge. The ladder operators a^\hat{a} and a^†\hat{a}^{\dagger} satisfy the usual bosonic commutation relation, [a^,a^†]=1[\hat{a},\hat{a}^{\dagger}]=1 and the canonical conjugate variables Φ^\hat{\Phi} and Q^\hat{Q} satisfy the commutation relation [Φ^,Q^]=iℏ[\hat{\Phi},\hat{Q}]=i\hbar. If we expand the cosine term in equation (2) and apply a rotating wave approximation assuming that the rotating terms do not have strong influence on the system, we obtain:

H=Q^22C+Φ^22LJ−EJ′4!(2πΦ^Φ0)4+O((2πΦ^Φ0)6)(4)\mathcal{H} = \frac{\hat{Q}^{2}}{2C} + \frac{\hat{\Phi}^{2}}{2L_J} - \frac{E'_J}{4!}\left(\frac{2\pi\hat{\Phi}}{\Phi_0}\right)^{4} + O\left(\left(\frac{2\pi\hat{\Phi}}{\Phi_0}\right)^{6}\right) \tag*{(4)}

Where LJL_J is the Josephson inductance. Substituting Φ^\hat{\Phi} and Q^\hat{Q} in equation (4) we obtain the so-called second quantization form as follows:

HRWA=ℏω~0a^†a^+K2a^†a^†a^a^(5)\mathcal{H}^{\mathrm{RWA}} = \hbar\widetilde{\omega}_0\hat{a}^{\dagger}\hat{a} + \frac{K}{2}\hat{a}^{\dagger}\hat{a}^{\dagger}\hat{a}\hat{a} \tag*{(5)}

Where ω~0=ω0+K\widetilde{\omega}_0=\omega_0+K is the shifted frequency arising from zero-point fluctuations (ZPF). From the second quantized Hamiltonian obtained at equation (5), one can use the Heisenberg equation of motion (HEOM) to derive the quantum Langevin equation (6), [24].

A˙=−iω~0A−iKA†AA−(γ1+γ2)A+2γ1ain(t)+2γ2bin(t)(6)\dot{A}=-i\widetilde{\omega}_0 A-iK A^{\dagger}AA-(\gamma_1+\gamma_2)A+\sqrt{2\gamma_1}a_{\mathrm{in}}(t)+\sqrt{2\gamma_2}b_{\mathrm{in}}(t) \tag*{(6)}

Here, the operators AA, aina_{\mathrm{in}}, and binb_{\mathrm{in}} represent the resonator, the input signal, and the input noise modes, respectively. Here, the signal and noise linewidths (γ1\gamma_1 and γ2\gamma_2 respectively) are defined as half width at half maximum. Considering only the pump input and neglecting any signal input for equation (6), In steady state i.e (A˙=0)(\dot{A}=0) we obtain equation (7).

K2N3+2(ω0−ωp)KN2+[(ω0−ωp)2+(γ1+γ2)2]N=2γ1pin2(7)K^{2}N^{3}+2(\omega_0-\omega_p)KN^{2}+\left[(\omega_0-\omega_p)^{2}+(\gamma_1+\gamma_2)^{2}\right]N=2\gamma_1p_{\mathrm{in}}^{2} \tag*{(7)}

As this equation accounts solely for the presence of the pump wave, NN represents the number of pump photons confined within the cavity. This parameter is crucial, as it is the pump wave that initiates and sustains the parametric amplification process. Symbol pinp_{\mathrm{in}} is pump power. Since this equation is a cubic equation in NN, it has multiple real solutions for certain drive powers. Some solutions may be in the bistable regime, which is known as the bifurcation regime, and has been treated in the literature [25], [26].

Figure 2. Normalized photon number (n) vs. normalized pump detuning (δ\delta) for various drive powers. n=N/(pin)2γ1(γ1+γ2)2n=N/(p_{\mathrm{in}})^{2}\frac{\gamma_1}{(\gamma_1+\gamma_2)^{2}} and δ=ω0−ωp\delta=\omega_0-\omega_p. To plot the pin>pcritp_{\mathrm{in}}>p_{\mathrm{crit}} case, we used arc-length method [27]

JPA enters the bistable regime as the pump power increases larger than the critical value of the pump power (pcritp_{\mathrm{crit}}) as shown in the figure 2. For JPA to operate correctly, pin<pcritp_{\mathrm{in}}<p_{\mathrm{crit}} should be satisfied. We can define the incoming field ain(t)a_{\mathrm{in}}(t) [or intracavity field A(t)A(t)] as two separate waves which are composed of pump and signal with the pump coefficient pinp_{\mathrm{in}} [or pp] being constant and real valued and the signal coefficient cin(t)c_{\mathrm{in}}(t) [or c(t)c(t)] being time dependent and complex valued.

ain(t)=[pine−i(ωpt+ϕp)cin(t)e−iωpt](8)a_{\mathrm{in}}(t)=\left[p_{\mathrm{in}}e^{-i(\omega_p t+\phi_p)}\quad c_{\mathrm{in}}(t)e^{-i\omega_p t}\right] \tag*{(8)}
A(t)=[pe−i(ωpt+ϕp)c(t)e−iωpt](9)A(t)=\left[pe^{-i(\omega_p t+\phi_p)}\quad c(t)e^{-i\omega_p t}\right] \tag*{(9)}

After determining the pump photon number in the steady state, one can solve the QLE equation (6) by substituting equations (8) and (9) in the presence of a weak signal tone cinc_{\mathrm{in}} and we arrive at the equation of motion for the signal field in the time domain:

−iωpc(t)+dcdt=−iω0c(t)−iK(c†(t)Ne−2iψB+2c(t)N)−2γ1cin(t)−(γ1+γ2)c(t)(10)-i\omega_p c(t)+\frac{dc}{dt}=-i\omega_0c(t)-iK\left(c^{\dagger}(t)Ne^{-2i\psi_B}+2c(t)N\right)-\sqrt{2\gamma_1}c_{\mathrm{in}}(t)-(\gamma_1+\gamma_2)c(t) \tag*{(10)}

Equation (10) can be solved using numerical methods, or as an integral equation. However, the most pronounced way to solve this equation in the literature is by moving into frequency domain since this is a linear equation and this approach allows one to derive the following analytical expression.

cin(ω)=12π∫−∞∞dt cin(t)eiωt,cout(ω)=12π∫−∞∞dt cout(t)eiωt,c(ω)=12π∫−∞∞dt c(t)eiωt.(11)\begin{aligned} c_{\mathrm{in}}(\omega) &= \frac{1}{\sqrt{2\pi}}\int_{-\infty}^{\infty} \mathrm{d}t\,c_{\mathrm{in}}(t)e^{i\omega t},\\ c_{\mathrm{out}}(\omega) &= \frac{1}{\sqrt{2\pi}}\int_{-\infty}^{\infty} \mathrm{d}t\,c_{\mathrm{out}}(t)e^{i\omega t},\\ c(\omega) &= \frac{1}{\sqrt{2\pi}}\int_{-\infty}^{\infty} \mathrm{d}t\,c(t)e^{i\omega t}. \tag*{(11)} \end{aligned}

With the Fourier definitions, the equation of motion (10) can be written in the frequency domain as:

i[(ω0−ωp)−ω−i(γ1+γ2)+2KN]c(ω)+iKNe−2iϕBc†(−ω)=−2γ1cin(ω)(12)\begin{aligned} i\left[(\omega_{0}-\omega_{p})-\omega-i(\gamma_{1}+\gamma_{2})+2KN\right]c(\omega)\\ +iKNe^{-2i\phi_{B}}c^{\dagger}(-\omega)=-\sqrt{2\gamma_{1}}c_{\mathrm{in}}(\omega) \tag*{(12)} \end{aligned}

To simplify the notation, the following parameters are defined:

W=i(ω0−ωp)+(γ1+γ2)+2iKNW=i(\omega_{0}-\omega_{p})+(\gamma_{1}+\gamma_{2})+2iKN
V=iKNe−2iϕB(13)V=iKNe^{-2i\phi_{B}} \tag*{(13)}
λ±=(γ1+γ2)±K2N2−(ω0−ωp+2KN)2(14)\lambda_{\pm}=(\gamma_{1}+\gamma_{2})\pm\sqrt{K^{2}N^{2}-(\omega_{0}-\omega_{p}+2KN)^{2}} \tag*{(14)}

With the definition of the following parameters, when plugged in, we obtain a complex equation. This equation can be solved by considering its hermitian conjugate pair.

(W−iω)c(ω)+Vc†(−ω)=−2γ1cin(ω)(15)(W-i\omega)c(\omega)+Vc^{\dagger}(-\omega)=-\sqrt{2\gamma_{1}}c_{\mathrm{in}}(\omega) \tag*{(15)}

The solution of a quadratic equation yields the lambda value.

c(ω)=−2γ1[(W∗−iω)cin(ω)−Vcin†(−ω)](iω−λ−)(iω−λ+)(16)c(\omega)=\frac{-\sqrt{2\gamma_{1}}\left[(W^{*}-i\omega)c_{\mathrm{in}}(\omega)-Vc_{\mathrm{in}}^{\dagger}(-\omega)\right]}{(i\omega-\lambda_{-})(i\omega-\lambda_{+})} \tag*{(16)}

This equation describes the evolution of the intracavity field. To compute output field, one may consider boundary condition given in equation (17).

cout(t)−cin(t)=2γc(t)(17)c_{\mathrm{out}}(t)-c_{\mathrm{in}}(t)=\sqrt{2\gamma}c(t) \tag*{(17)}

Using boundary condition, the gain equation reads:

cout(ω)=(1−2γ1(W∗−iω)(iω−λ−)(iω−λ+))cin(ω)+2γ1V(iω−λ−)(iω−λ+)cin†(−ω)≡G(ω)cin(ω)+M(ω)cin†(−ω)(18)\begin{aligned} c_{\mathrm{out}}(\omega)&=\left(1-\frac{2\gamma_{1}(W^{*}-i\omega)}{(i\omega-\lambda_{-})(i\omega-\lambda_{+})}\right)c_{\mathrm{in}}(\omega)\\ &+\frac{2\gamma_{1}V}{(i\omega-\lambda_{-})(i\omega-\lambda_{+})}c_{\mathrm{in}}^{\dagger}(-\omega)\\ &\equiv G(\omega)c_{\mathrm{in}}(\omega)+M(\omega)c_{\mathrm{in}}^{\dagger}(-\omega) \tag*{(18)} \end{aligned}

With equation (16), one can calculate the gain. As a result of working in the pump frame, the mode detunings are with respect to the pump frequency. Therefore, cin(ω)c_{\mathrm{in}}(\omega) represents the signal mode and the mode cin†(−ω)c_{\mathrm{in}}^{\dagger}(-\omega) is the image mode. Defining these modes, the gain is calculated as square of their coefficients.

a)Gs(ω)=∣G(ω)∣2b)Gi(ω)=∣M(ω)∣2(19)\begin{aligned} \mathrm{a)}\quad G_{s}(\omega)&=|G(\omega)|^{2}\\ \mathrm{b)}\quad G_{i}(\omega)&=|M(\omega)|^{2} \tag*{(19)} \end{aligned}

Figure 3. The solution of Equation (18) for the degenerate signal mode (ω=0\omega= 0) is used to find the optimal pump detuning for a given pump power. (a) Signal gain vs pump detuning, (b) image gain vs detuning

When one computes optimal pump frequency and power for desired gain from signal gain - pump detuning graph, we can plot signal gain graph spectrum for the device.

Figure 4. The solution of Equation (18) for optimal pump frequency. (a) signal gain vs. signal frequency (optimal Δωp\Delta\omega_{p}), (b) image gain vs. signal frequency (optimal Δωp\Delta\omega_{p})

Figure 4 illustrates both the symmetry and the gain-bandwidth tradeoff. Additionally, quantum theory predicts that a maximum gain of 38dB is theoretically attainable if the biasing is sufficiently precise. However, achieving this regime in practice is challenging because flux or charge noise induced in the JPA can easily push the device into the bifurcation regime.

JPA Design Based on Circuit Model

After calculating the gain of the JPA as explained in Section II, in this section we compute it using classical circuit and

Figure 5: (a) Linear resonator and its coupling to a transmission line. (b) Norton-equivalent model for the JPA used to extract intra-cavity voltage and current.

Linear cavity coupled to transmission line and Norton-equivalent circuit model

Figure 5. (a) Linear resonator and its coupling to a transmission line. (b) Circuit model used in time-domain analysis, where the Josephson junction replaces the inductor. [28]

electromagnetic theory. Since the JPA is a single-port device, it can be described by its one-port reflection coefficient S11S_{11}. To compute the reflection coefficient, we begin by applying Kirchhoff’s Current Law (KCL) to the circuit shown in Figure 5. to calculate the reflection coefficient, we write the wave expressions for the transmission line assuming left- and right-traveling waves:

V(z,t)=Vin(t−z/v)+Vout(t+z/v),(20)V(z,t) = V_{\mathrm{in}}(t-z/v) + V_{\mathrm{out}}(t+z/v), \tag*{(20)}
I(z,t)=1Zc[Vin(t−z/v)−Vout(t+z/v)],(21)I(z,t) = \frac{1}{Z_c}\left[V_{\mathrm{in}}(t-z/v) - V_{\mathrm{out}}(t+z/v)\right], \tag*{(21)}

where vv is the wave propagation speed and ZcZ_c is the characteristic impedance. Since the circuit elements are much smaller than the wavelength of operation, we work in the lumped limit and eliminate spatial dependence. The input and output fields are then expressed in terms of the intra-cavity voltage V(t)V(t) and current I(t)I(t):

Vin(t)=V(t)+ZcI(t)2,(22)V_{\mathrm{in}}(t) = \frac{V(t) + Z_c I(t)}{2}, \tag*{(22)}
Vout(t)=V(t)−ZcI(t)2.(23)V_{\mathrm{out}}(t) = \frac{V(t) - Z_c I(t)}{2}. \tag*{(23)}

The circuit in Figure 5 cannot be directly simulated as a wave system without using an LC ladder model or solving the telegrapher’s equations, both of which are time-consuming and unnecessary in the lumped-element limit. Therefore, we adopt the Norton equivalent circuit (Figure 5b) to obtain intra-cavity quantities. Using Eq. (23), the input and output fields follow, which allows computing the reflection coefficient. Because JoSIM operates in the time domain, we extract specific frequency components via Fourier projection:

F(ω)=1T∫0Tf(t)eiωst dt,(24)F(\omega) = \frac{1}{T}\int_{0}^{T} f(t)e^{i\omega_s t}\,\mathrm{d}t, \tag*{(24)}

which yields the desired frequency-domain signal. The reflection coefficient is defined as the ratio of output to input voltage in the frequency domain [29]:

S11(ω)=Vout(ω)Vin(ω),S11(dB)=20log⁡10∣S11(ω)∣.(25)S_{11}(\omega) = \frac{V_{\mathrm{out}}(\omega)}{V_{\mathrm{in}}(\omega)}, \qquad S_{11}(\mathrm{dB}) = 20\log_{10}\lvert S_{11}(\omega)\rvert. \tag*{(25)}

To characterize the cavity we extract its resonance frequency and quality factor. In the high-QQ limit, the impedance can be approximated as [28]:

Z(ω)=jLω02(ω0−ω)(ω0+ω)≈jLω022(ω0−ω)=12C(ω−ω0),(26)Z(\omega) = \frac{jL\omega_0^2}{(\omega_0-\omega)(\omega_0+\omega)} \approx\frac{jL\omega_0^2}{2(\omega_0-\omega)} = \frac{1}{2C(\omega-\omega_0)}, \tag*{(26)}

where LL is the Josephson inductance, ω0\omega_0 is the resonance frequency, and Z(ω)Z(\omega) is the resonator impedance. Under weak driving, the JPA behaves approximately linearly. The reflection coefficient of the resonator becomes:

Γ(ω)=Z(ω)−ZcZ(ω)+Zc≈γ−j(ω−ω0)γ+j(ω−ω0),(27)\Gamma(\omega) = \frac{Z(\omega)-Z_c}{Z(\omega)+Z_c} \approx\frac{\gamma-j(\omega-\omega_0)}{\gamma+j(\omega-\omega_0)}, \tag*{(27)}

where γ=ω0/(2Q)\gamma=\omega_0/(2Q) is the linewidth. Figure 6 shows the phase of S11S_{11} as a function of frequency. Fitting this curve yields ω0\omega_0 and γ\gamma, from which the quality factor is obtained. And to directly compare quantum and classical descriptions, we normalize the detuning (ωp−ω0)(\omega_p-\omega_0) by γ\gamma, where ωp\omega_p is the pump frequency. The gain versus normalized detuning for different pump powers is then plotted. Since the JPA is a single-port device, a positive reflection coefficient corresponds to gain. Plotting Eq. (25) yields the result in Figure 7.

Curve fitting to the reflection phase using Eq. (27)

Figure 6. Curve fitting to the reflection phase using Eq. (27).

Figure 7. JoSIM gain curve for the degenerate signal mode (ω=0\omega= 0), used to find the optimal pump detuning for a given pump power. (a) Signal gain vs. pump detuning, (b) idler gain vs. pump detuning.

Choosing ωp\omega_p at the maximum gain point for each pump power produces the optimal-frequency gain curves shown in Figure 8.

Figure 8. JoSIM gain curve for optimal pump frequency: (a) signal gain vs. signal frequency (optimal Δωp\Delta\omega_p), (b) idler gain vs. signal frequency (optimal Δωp\Delta\omega_p).

Figures 7 and 8 show that the gain increases as pinp_{\mathrm{in}} approaches pcritp_{\mathrm{crit}}, and the measured gains agree well with quantum theory apart from slight discrepancies due to the linearization of the quantum Langevin equation and the relatively low quality factor of the cavity [30].

The saturation power of individual Josephson amplifiers is lower compared to arrays of such devices [2]. Computing the 1 dB compression point at the highest gain yields −134 dBm, as shown in Figure 9. The figure also highlights the gain–bandwidth trade-off in the large input power regime.

Figure 9. Signal amplitude–signal detuning sweep for pump power 0.99 pc0.99\,p_c and ωp=3.305\omega_p = 3.305 GHz. The colormap represents signal gain in dB.

Summary and Discussion

We computed the gain of the JPA using both quantum and classical frameworks. In the quantum approach, we first derive the system Hamiltonian and model the cavity as weakly coupled to its environment, which enables us to express the intra-cavity field in terms of the input and output fields. This formulation allows for gain calculation either analytically or

numerically in the frequency domain. For the classical time-domain analysis, we employ circuit-level models of the JPA components to capture the parametric amplification behavior. Through simple transformations, the gain is extracted from lumped-element simulations. The JPA gain is calculated from the reflection coefficient of the signal wave and its image tone. In summary, we present a systematic way to tackle resonant systems and characterize their reflection coefficient using a circuit simulator, and we show that if modeled correctly, quantum input-output theory and Josephson circuit simulators can be used interchangeably. The use of circuit simulators enables designers to simplify and improve their workflow while advancing the development of complicated quantum circuits.

Acknowledgment

The authors thank Nazif Orhun Tekci for his help in editing the figures.

Appendix

In addition to the Josephson circuit based simulators for time domain analysis, we computed the signal characteristics of the JPA starting with the Josephson equations and solve the coupled differential equations by using the ODE45 solver of Matlab with strict tolerances. For this purpose, we setup the circuit shown in the Figure 10. The circuit is composed of an input current (IinI_{\mathrm{in}}), a pump current (IpI_{\mathrm{p}}), a coupling capacitor (CcoC_{\mathrm{co}}), and a basic JPA structure represented by (J1J_{1} and CC). ZcZ_{\mathrm{c}} represents the characteristics impedance of the transmission line between the input/pump currents and the JPA. The critical current of J1J_{1} is IcI_{\mathrm{c}}.

Figure 10. Josephson circuit model for time domain analysis

Starting from the Josephson equations and summing the currents at the input and output nodes of the circuit and solving the equation for the intracavity node flux (ϕint\phi_{\mathrm{int}}), we obtain equation (28)

d3ϕintdt3=−C+CcoCcoZ0Cd2ϕintdt2−IcCϕ0dϕintdtcos⁡(ϕintϕ0)+IinωCcos⁡(ωt)+IpωpCcos⁡(ωpt)−IcZ0Ccosin⁡(ϕintϕ0)(28)\frac{d^{3}\phi_{\mathrm{int}}}{dt^{3}} = -\frac{C+C_{\mathrm{co}}}{C_{\mathrm{co}}Z_{0}C}\frac{d^{2}\phi_{\mathrm{int}}}{dt^{2}} - \frac{I_{\mathrm{c}}}{C\phi_{0}}\frac{d\phi_{\mathrm{int}}}{dt}\cos\left(\frac{\phi_{\mathrm{int}}}{\phi_{0}}\right) + \frac{I_{\mathrm{in}}\omega}{C}\cos(\omega t) + \frac{I_{\mathrm{p}}\omega_{\mathrm{p}}}{C}\cos(\omega_{\mathrm{p}}t) - \frac{I_{\mathrm{c}}}{Z_{0}C_{\mathrm{co}}}\sin\left(\frac{\phi_{\mathrm{int}}}{\phi_{0}}\right) \tag*{(28)}

(28) is the equation of motion of a simple JPA capacitively coupled to a transmission line. Where, ϕ0\phi_{0} represents the single flux quantum, ωp\omega_{\mathrm{p}} is the pump frequency, and ω\omega is the signal frequency. By solving the equation, one can compute the transmission line voltages VtlV_{\mathrm{tl}} and intracavity voltage VintV_{\mathrm{int}} as follows:

Vtl=−Z0Cd2ϕintdt2−Z0Icsin⁡(ϕintϕ0)+(Iin+Ip)Z0(29)V_{\mathrm{tl}} = -Z_{0}C\frac{d^{2}\phi_{\mathrm{int}}}{dt^{2}} - Z_{0}I_{\mathrm{c}}\sin\left(\frac{\phi_{\mathrm{int}}}{\phi_{0}}\right) + (I_{\mathrm{in}}+I_{\mathrm{p}})Z_{0} \tag*{(29)}
Vint=dϕintdt(30)V_{\mathrm{int}} = \frac{d\phi_{\mathrm{int}}}{dt} \tag*{(30)}

In (29) and (30), VtlV_{\mathrm{tl}} and VintV_{\mathrm{int}} also represent the input and output voltages, respectively. If we plot them vs. time, we get the Figure 11.

Figure 11. Comparison between BDF-2 based JoSIM and RK4 Dormand-Prince based ODE45. Inset shows one period of the input and output signals.

If we compare the discrepancy between two solvers, we get a discrepancy of %0.36. So, both classical solutions fit with each other and can be used interchangeably where convenient. For the sake of simplicity, we compared the classical solution to quantum solution by using JoSIM results.

JoSIM, a time-domain Josephson circuit simulator, employs the BDF-2 method to compute equation derivatives for improved numerical stability. At its core, it solves the standard Kirchhoff equations for capacitors and inductors, with the primary distinction being that calculations are performed in the phase basis. In addition, it incorporates the Josephson relations to accurately model the behavior of Josephson junctions.

References

  1. [1]M. A. Castellanos-Beltran, K. D. Irwin, G. C. Hilton, L. R. Vale, and K. W. Lehnert, “Amplification and squeezing of quantum noise with a tunable Josephson metamaterial,” Nature Physics, vol. 4, no. 12, pp. 929–931, 12 2008.
  2. [2]L. Planat, R. Dassonneville, J. P. Martínez, F. Foroughi, O. Buisson, W. Hasch-Guichard, C. Naud, R. Vijay, K. Murch, and N. Roch, “Understanding the Saturation Power of Josephson Parametric Amplifiers Made from SQUID Arrays,” Physical Review Applied, 2019.DOI
  3. [3]L. Ranzani, G. Ribeill, B. Hassick, and K. C. Fong, “Wideband Josephson parametric amplifier with integrated transmission line transformer,” in 2022 IEEE International Conference on Quantum Computing and Engineering (QCE), 2022, pp. 314–319.arxiv.org/abs/2208.02331
  4. [4]J. Grebel, A. Bienfait, P. Dumur, H.-S. Chang, M.-H. Chou, C. R. Conner, G. A. Peairs, R. G. Povey, Y. P. Zhong, and A. N. Cleland, “Flux-pumped impedance-engineered broadband Josephson parametric amplifier,” Applied Physics Letters, vol. 118, no. 14, p. 142601, 04 2021.DOI
  5. [5]T. Elo, T. S. Abhilash, M. R. Perelstein, I. Lilja, E. V. Korostylev, and P. J. Hakonen, “Broadband lumped-element Josephson parametric amplifier with single-step lithography,” Applied Physics Letters, vol. 114, no. 15, p. 152601, 2019.DOI
  6. [6]N. Roch, E. Flurin, F. Nguyen, P. Morfin, P. Campagne-Ibarcq, M. H. Devoret, and B. Huard, “Widely Tunable, Nondegenerate Three-Wave Mixing Microwave Device Operating near the Quantum Limit,” Physical Review Letters, vol. 108, no. 14, p. 147701, 04 2012.pubmed.ncbi.nlm.nih.gov/22540823
  7. [7]J. A. Russer, M. Würth, W. Utschick, F. Bischeltsrieder, and M. Peichl, “Circuits for Josephson Parametric Amplification in Quantum Radar,” in 2023 United States National Committee of URSI National Radio Science Meeting (USNC-URSI NRSM), 2023, pp. 315–316.DOI
  8. [8]I. Siddiqi, R. Vijay, F. Pierre, C. M. Wilson, M. Metcalfe, C. Rigetti, L. Frunzio, and M. H. Devoret, “Rf-driven josephson bifurcation amplifier for quantum measurement,” Phys. Rev. Lett., vol. 93, p. 207002, Nov 2004.pubmed.ncbi.nlm.nih.gov/15600958
  9. [9]S. Boutin, D. M. Toyli, A. V. Venkatramani, A. W. Eddins, I. S. Siddiqi, and A. Blais, “Effect of Higher-Order Nonlinearities on Amplification and Squeezing in Josephson Parametric Amplifiers,” Physical Review Applied, vol. 8, no. 5, p. 054030, 2017.DOI
  10. [10]C. W. Gardiner and M. J. Collett, “Input and output in damped quantum systems: Quantum stochastic differential equations and the master equation,” Physical Review A, vol. 31, no. 6, pp. 3761–3774, 1985.pubmed.ncbi.nlm.nih.gov/9895956
  11. [11]C. Eichler and A. Wallraff, “Controlling the dynamic range of a Josephson parametric amplifier,” EPJ Quantum Technology, vol. 1, no. 1, p. 2, 2014.
  12. [12]L. Ramzani and J. Aumentado, “Graph-based analysis of nonreciprocity in coupled-mode systems,” New Journal of Physics, vol. 17, no. 2, p. 023024, 2015.
  13. [13]K. He, G. Dai, Q. Yu, Y. He, C. Zhao, J. Liu, and W. Chen, “Simulation of a flux-pumped Josephson parametric amplifier with a detailed SQUID model using the harmonic balance method,” Superconductor Science and Technology, vol. 36, no. 4, 2023.DOI
  14. [14]E. S. Fang and T. Van Duzer, “A Josephson integrated circuit simulator (JSIM) for superconductive electronics application,” in Extended Abstracts of the 1989 International Superconductive Electronics Conference (ISEC ’89), 08 1989, pp. 407–410.
  15. [15]J. A. Delport, K. Jackman, P. L. Roux, and C. J. Fourie, “JoSIM – Superconductor SPICE Simulator,” IEEE Transactions on Applied Superconductivity, vol. 29, no. 5, 2018, pp. 1–5.
  16. [23]U. Vool and M. Devoret, “Introduction to quantum electromagnetic circuits,” International Journal of Circuit Theory and Applications, vol. 45, no. 7, pp. 897–934, 2017.DOI
  17. [24]C. M. Quintana, “Josephson Parametric Amplification for Circuit Quantum Electrodynamics: Theory and Implementation,” Ph.D. dissertation, 2013.
  18. [25]R. Vijay, M. H. Devoret, and I. S. Siddiqi, “Invited Review Article: The Josephson bifurcation amplifier,” Review of Scientific Instruments, vol. 80, no. 11, p. 111101, 2009.DOI
  19. [26]I. Siddiqi, R. Vijay, F. Pierre, C. M. Wilson, L. Frunzio, M. Metcalfe, C. Rigetti, and M. H. Devoret, “The Josephson Bifurcation Amplifier for Quantum Measurements,” arXiv, 2005.DOI
  20. [27]E. Sarrouy, J. Sinou, and F. Ebrahimi, “Advances in vibration analysis research,” in Non-Linear Periodic and Quasi-Periodic Vibrations in Mechanical Systems – On the use of the Harmonic Balance Methods, 2011, pp. 419–434.
  21. [28]C. Beltran, “Development of a Josephson Parametric Amplifier for the Preparation and Detection of Nonclassical States of Microwave Fields,” Ph.D. dissertation, 2010.
  22. [29]D. M. Pozar, Microwave Engineering, 4th Edition. Wiley, 11 2011.
  23. [30]B. A. Kochetov and A. Fedorov, “Higher-order nonlinear effects in a Josephson parametric amplifier,” Phys. Rev. B, vol. 92, p. 224304, Dec 2015.DOI

Paper details

vlnc:W_d5rqnxe7
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents