Josephson Current, Phase, Energy, and Inductance

The constitutive relations of an ideal Josephson element and the exact boundary between its nonlinear model and a bias-point small-signal inductance.

Node role: Concept.

Reader task: keep the Josephson current law, cosine energy, and operating-point inductance physically distinct.

Why it matters

A Josephson junction is not a fixed inductor. Its sine current-phase relation creates a cosine energy landscape; a linear inductance appears only after an operating point and perturbation scale have been chosen.

Keeping those layers separate lets a reviewer tell whether an implementation is solving the full nonlinear branch, using a local linear surrogate, or adding capacitance and dissipation through a larger junction model.

Follow the element law into energy, add capacitance as a separate branch, then linearize at a stated operating point. The final model comparison shows which questions those three operations can answer.

Symbols and orientation

Engineering choice. These symbols and branch directions are this page’s notation; literature with other signs must be translated consistently.

Choose one branch orientation. All electrical quantities below use that same direction.

Symbol Meaning Units
\(I\) Oriented branch current A
\(V\) Oriented branch voltage V
\(\Phi_J\) Oriented generalized branch flux Wb
\(\Phi_0=h/(2e)\) Superconducting flux quantum Wb
\(\varphi_0=\Phi_0/(2\pi)=\hbar/(2e)\) Reduced flux quantum Wb
\(I_c\) Positive critical-current amplitude A
\(E_J=\varphi_0 I_c\) Josephson energy J
\(\delta=(\Phi_J-\Phi_{J,\mathrm{offset}})/\varphi_0\) Oriented gauge-invariant phase difference dimensionless

Reversing the branch reverses \(I\), \(V\), \(\Phi_J\), and \(\delta\). A different sign convention is valid only when all four change consistently.

The ideal Josephson element

Reorganized explanation. The ideal sinusoidal current law and voltage–phase relation are the conventional Josephson-element starting model (Vool and Devoret 2017, sec. 4.1). This-page derivation: integrating that law produces the cosine energy written below.

For a conventional tunnel junction with a sinusoidal current-phase relation,

\[ \boxed{I=I_c\sin\delta} \]

and the voltage-phase relation is

\[ \boxed{V=\varphi_0\dot\delta}. \]

Because \(V=\dot\Phi_J\) and current is the derivative of stored energy with respect to branch flux,

\[ I=\frac{\partial U_J}{\partial\Phi_J} =\frac{1}{\varphi_0}\frac{\partial U_J}{\partial\delta}. \]

Integrating the current-phase relation gives

\[ \boxed{U_J(\delta)=-E_J\cos\delta}, \qquad E_J=\varphi_0 I_c. \]

The additive constant in energy is arbitrary; the curvature and energy differences are physical.

Independent parallel junction capacitance

Reorganized explanation. A tunnel junction’s capacitive branch is distinct from its ideal Josephson element (Vool and Devoret 2017, sec. 1.1). Engineering choice: independently declaring or omitting that branch makes the model input explicit; it is not a fabrication relation inferred from inductance.

The ideal Josephson element above specifies only the nonlinear constitutive branch. A capacitively shunted Josephson junction adds an independent capacitance \(C_J\) across the same two terminals and with the same branch voltage:

\[ \boxed{ I=I_c\sin\delta+C_J\dot V =I_c\sin\delta+C_J\varphi_0\ddot\delta }. \]

The added electric energy is

\[ U_{C_J}=\frac12C_JV^2. \]

\(C_J\) is a separate model input. It is not inferred from \(I_c\), \(E_J\), a zero-bias \(L_{J0}\), or a bias-point \(L_J(\delta_0)\). A fabrication model may relate junction area to both \(I_c\) and \(C_J\), but that relation needs its own declared technology evidence; the Josephson constitutive law alone does not supply it. Setting \(C_J=0\) means the parallel capacitance is explicitly omitted from this circuit model. It must not trigger an implicit estimate from \(L_J\).

Inductance belongs to an operating point

This-page derivation. Differentiate the stated sinusoidal constitutive law at the selected bias. The resulting differential inductance is local, unlike the complete cosine model (Vool and Devoret 2017, sec. 4.1).

For a small phase perturbation \(\widetilde\delta\) around a bias point \(\delta_0\),

\[ \widetilde I \approx I_c\cos\delta_0\,\widetilde\delta, \qquad \widetilde\Phi_J=\varphi_0\widetilde\delta. \]

The bias-point differential inductance is therefore

\[ \boxed{ L_J(\delta_0) =\left.\frac{\partial I}{\partial\Phi_J}\right|_{\delta_0}^{-1} =\frac{\varphi_0}{I_c\cos\delta_0} }. \]

At a zero-current minimum, \(\delta_0=0\) modulo \(2\pi\) and

\[ L_{J0}=\frac{\varphi_0}{I_c}. \]

Near \(|\cos\delta_0|=0\), the linear inductance diverges and the next nonlinear terms dominate. A negative differential inductance signals negative local curvature of the isolated Josephson potential; stability then depends on the embedding circuit rather than on the junction alone.

Four different model contracts

Model Keeps Omits Suitable question
Ideal Josephson element Full \(I_c\sin\delta\) and \(-E_J\cos\delta\). Junction capacitance and dissipation. What nonlinear current and energy follow from phase?
Capacitively shunted Josephson junction Ideal Josephson element plus an independently supplied parallel \(C_J\). Dissipation unless a separate loss element is added. What conservative plasma or nonlinear capacitive dynamics follow from the declared \(I_c\) and \(C_J\)?
Bias-point inductance \(L_J(\delta_0)\) for infinitesimal motion around one point. Large-signal curvature, phase slips, harmonic generation. What is the local resonance or small-signal response?
RCSJ-type junction Josephson element plus shunt capacitance and resistance. Microscopic quasiparticle detail unless added separately. What switching, plasma, damping, or voltage-state dynamics occur?

With a parallel resistance \(R_J\), the RCSJ branch current is

\[ I=I_c\sin\delta +C_J\varphi_0\ddot\delta +\frac{\varphi_0}{R_J}\dot\delta. \]

A physical tunnel junction has capacitance even when an ideal Josephson element is used as the conservative constitutive primitive. Naming the model and declaring \(C_J\) independently makes that abstraction visible.

This is an existing implementation description, not a freshly verified package-capability claim; current executable contracts remain package-owned.

  • The Workbench Julia JosephsonJunction lowers a supplied Josephson inductance to a nonlinear JosephsonCircuits Lj row. It is not an ordinary linear inductor.
  • The Workbench Python SQUID-frequency and \(Y_{11}\) formulas instead use \(L_{\mathrm{jun}}/2\) as the parallel combination of two identical supplied small-signal junction inductances. They contain no phase dynamics.
  • A fitted \(L_J\) must therefore state whether it is \(L_{J0}\), a bias-point differential value, or an already-combined device inductance.
NoteLimits and failure modes
  • Weak links can have a non-sinusoidal current-phase relation; then both \(U(\delta)\) and differential inductance must come from that actual relation.
  • Replacing \(L_J(\delta_0)\) by \(L_{J0}\) at finite current hides the \(\cos\delta_0\) factor.
  • Treating a linearized branch as a nonlinear Josephson element invents harmonics and mixing that the model cannot produce.
  • Treating the full Josephson element as a fixed inductor removes the very nonlinearity used for anharmonicity and parametric processes.
  • Inferring \(C_J\) from \(L_J\) without a separate technology relation invents a device model that the Josephson law does not contain.
  • Reporting \(E_J\) in “GHz” without writing \(E_J/h\) confuses energy with frequency.

Connections

References

Field Value
Status Source-backed
Used by Josephson Cosine and Quantum Anharmonicity, DC SQUID Flux Tunability, and Circuit Lagrangian, Hamiltonian, and Canonical Quantization.
Implementation links Julia junction relation, Julia lowering, and Python small-signal surrogate.
Open review question Does the page make it impossible to mistake \(L_{J0}\) for the full nonlinear element or for a finite-bias differential inductance?

Bibliography

Vool, U., and M. H. Devoret. 2017. “Introduction to Quantum Electromagnetic Circuits.” International Journal of Circuit Theory and Applications 45: 897–934. https://doi.org/10.1002/cta.2359.