Josephson Traveling Wave Parametric Amplifier (JTWPA)

• JTWPA is a superconducting microwave amplifier using distributed Josephson nonlinearities to achieve exponential gain over a multi-gigahertz bandwidth.
• It employs intrinsic plasma oscillations for fine phase-matching and automatic suppression of higher harmonics, simplifying design and fabrication.
• Simulations reveal >15 dB gain, 3.5 GHz bandwidth, and minimal gain ripple, making it ideal for quantum-limited measurements and scalable qubit readout.

A Josephson Traveling Wave Parametric Amplifier (JTWPA) is a superconducting, microwave-frequency, wideband amplifier leveraging distributed Josephson nonlinearities to achieve phase-preserving amplification. Unlike resonator-based Josephson parametric amplifiers, JTWPAs employ a long chain (“lumped-element transmission line”) of Josephson junctions (JJs) or SQUIDs, and realize exponential gain over multi-gigahertz bandwidths. The Josephson Traveling Wave Parametric Amplifier with Plasma Oscillation Phase-Matching (“plasma-matched JTWPA”) employs plasma oscillations intrinsic to the Josephson elements, rather than external resonators or periodic loadings, to achieve fine phase-matching and automatic suppression of higher harmonic generation (Rizvanov et al., 2024).

1. Theoretical Foundations: Circuit Model, Nonlinearity, and Parametric Mixing

The JTWPA is modeled as a discrete transmission line formed by a periodic array of Josephson junctions. Each unit cell of length $\Delta z$ contains a JJ (critical current $I_c$, capacitance $C_{JJ}$) and shunt capacitance to ground $C_g$. The phase dynamics are governed by the Lagrangian

$$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$

where $E_J = \hbar I_c / 2e$ and $\varphi_n(t)$ is the gauge-invariant phase across the $n$-th JJ.

From the Euler–Lagrange equations, a nonlinear wave equation is obtained. The system is pumped at frequency $\omega_p$ (large amplitude) and a weak signal at $\omega_s$; mixing generates an idler at $I_c$0 such that $I_c$1 (3WM). Linearization about the strong pump yields a set of coupled-mode equations (CMEs) for the slowly-varying envelopes $I_c$2. In the undepleted-pump limit,

$I_c$3

where $I_c$4 is the gain coefficient and $I_c$5 is the total phase-mismatch, which are explicit functions of device parameters, pump strength, and bias currents (Rizvanov et al., 2024).

2. Plasma Oscillation Phase-Matching and Dispersion Engineering

Conventional JTWPAs achieve phase-matching via distributed resonators, periodic capacitance, or other dispersion-engineering techniques, but the plasma-matched design exploits the intrinsic Josephson plasma resonance for phase control. In this approach, every $I_c$6 JJ is shunted by a large capacitance $I_c$7, producing a plasma oscillation at frequency

$I_c$8

Below $I_c$9, the dispersion is continuous; as $C_{JJ}$0, $C_{JJ}$1, and for $C_{JJ}$2, $C_{JJ}$3 becomes imaginary, creating a hard stop-band for higher harmonics. The effective 1D dispersion relation is

$C_{JJ}$4

Most critically, the plasma cutoff creates a large phase-slip near $C_{JJ}$5 and forbids propagation above. By placing $C_{JJ}$6 near the second harmonic of the pump ($C_{JJ}$7 GHz in the reported work), $C_{JJ}$8 is held near zero over the instantaneous signal band ($C_{JJ}$9), thus enabling exponentially-growing gain and suppressing higher-order processes (Rizvanov et al., 2024).

3. Harmonic Suppression and Dynamic Range

Harmonic suppression is achieved intrinsically: for $C_g$0, the wavevector is imaginary and all higher harmonics, particularly the pump’s second harmonic ($C_g$1), are reflected rather than allowed to propagate. This reflection prevents up-conversion of the pump and signal tones, maximizing energy transfer to the signal/idler and avoiding depletion. In the plasma-engineered design, time-domain simulations show that power at $C_g$2 remains localized, preserving dynamic range, in contrast to homogeneous lines where leakage into harmonics grows rapidly with propagation distance. This intrinsic filtering eliminates the need for explicit external resonators or additional periodic loads beyond the every-$C_g$3 capacitive shunt (Rizvanov et al., 2024).

4. Simulation Methodologies and Reported Performance

Numerical modeling was performed with JoSIM and cross-checked in WRspice, with a 2,000-junction chain (400 unit cells; each fifth JJ shunted by $C_g$4 fF; unit cell: $C_g$5A, $C_g$6 fF, $C_g$7 fF), pumped at $C_g$8 GHz ($C_g$9–$$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$0A, $$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$1A). The resultant gain profile, bandwidth, and dynamic range are as follows:

Metric	Value	Context
Forward gain	$$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$2 dB	3.5–7 GHz band
3 dB bandwidth	3.5 GHz	$$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$3 [3.5, 7] GHz
Gain ripple	$$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$4 dB	across 3.5 GHz bandwidth
Reflection $$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$5	$$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$6 dB	across signal band
Optimal length	2,000 JJs	pump not depleted
Harmonic content	No $$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$7 propagation	for plasma-matched design

For lengths up to $$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$81,500 JJs, the signal and idler grow exponentially ($$\mathcal{L} = \sum_n \left[ \frac{1}{2} C_g (\dot V_n)^2 + C_{JJ}\frac{\Phi_0^2}{(2\pi)^2}\dot\varphi_n^2 + E_J \cos\varphi_n \right],$$9), and harmonic leakage is negligible, supporting large dynamic range in high-gain operation (Rizvanov et al., 2024).

5. Comparison with Other JTWPA Architectures

In conventional JTWPA designs, phase-matching is achieved via external resonant structures (e.g. $E_J = \hbar I_c / 2e$0/4 stubs, periodic capacitance) that create stop-bands or flatten dispersion, but these solutions increase fabrication complexity and risk introducing impedance mismatches and gain ripple (Elkin et al., 2024, Peng et al., 2022). The plasma-matched design achieves similar gain/bandwidth ($E_J = \hbar I_c / 2e$1 dB, 3.5 GHz), but requires only a single additional parallel capacitor every five junctions and avoids external resonators or spread-sensitive phase-matching elements. This simplifies fabrication and directly suppresses harmonic growth (Rizvanov et al., 2024).

Additionally, the plasma cutoff ensures automatic harmonic suppression without requiring precise tuning of element values, reducing sensitivity to statistical variations in $E_J = \hbar I_c / 2e$2, $E_J = \hbar I_c / 2e$3, or $E_J = \hbar I_c / 2e$4, a key issue for yield and reproducibility in large-scale JTWPA platforms (Kissling et al., 2023, Elkin et al., 2024). Performance matches resonant or Floquet-engineered TWPAs in gain and quantum efficiency, but with reduced design complexity (Peng et al., 2022).

6. Practical Considerations and Application Scenarios

This JTWPA design supports broad bandwidth, large dynamic range, and high gain with a minimal ripple and robust harmonic suppression inherent in the transmission line architecture. Intrinsic phase-matching and harmonic filtering are particularly attractive for scalable superconducting qubit readout, broadband quantum-limited microwave measurement, and applications requiring rapid frequency-multiplexed detection. The circuit requires only moderate critical currents and standard shunt capacitances, supporting integration with typical superconducting process flows (Rizvanov et al., 2024).

Intrinsic phase-matching via plasma oscillation also reduces sensitivity to local parameter spread and system-level fabrication nonuniformity, a major technical challenge for high-yield deployable quantum amplifier arrays (Kissling et al., 2023, Elkin et al., 2024). The approach natively avoids typical reliability bottlenecks of resonant or PCM architectures, and can be extended to higher signal bands or different frequency regimes by scaling $E_J = \hbar I_c / 2e$5 and $E_J = \hbar I_c / 2e$6 accordingly.

7. Summary of Key Results and Significance

The plasma oscillation phase-matched JTWPA achieves the following:

• Exponential gain scaling, with measured $E_J = \hbar I_c / 2e$7 dB gain and $E_J = \hbar I_c / 2e$8 GHz instantaneous bandwidth.
• Intrinsic suppression of higher harmonics above the engineered plasma frequency, automatically confining signal amplification to the fundamental manifold.
• Minimal gain ripple ($E_J = \hbar I_c / 2e$9 dB), robust to reflection and parameter mismatches across the band.
• Simplified architecture requiring no external or periodic resonant elements, facilitating fabrication and scalability.
• Dynamic range limited only by pump depletion after exponential signal growth, maximizing number of usable readout tones or multiplexed channels (Rizvanov et al., 2024).

By harnessing the native plasma resonance of the Josephson element chain, this design approach defines a new, robust methodology for scalable, high-performance, quantum-limited, wideband parametric amplification in superconducting quantum circuit platforms.