6.5 Josephson effect

The Josephson effect is a quantum phenomenon in which Cooper pairs tunnel between two superconductors separated by a thin barrier. It provides one of the clearest demonstrations of macroscopic quantum coherence in condensed matter systems and forms the foundation for superconducting electronics, precision metrology, and several leading approaches to quantum computing.

A Josephson junction consists of two superconducting electrodes coupled through a weak link. The key physics is that the superconducting order parameter on each side has a well-defined phase, and the tunneling current depends on the phase difference across the junction. The two Josephson equations capture this relationship and govern nearly everything that follows.

Fundamentals of Josephson effect

The Josephson effect arises because the superconducting state is described by a macroscopic wavefunction with a definite phase. When two such superconductors are brought close together with only a thin barrier between them, Cooper pairs can tunnel coherently across that barrier, producing a supercurrent that depends on the phase difference between the two wavefunctions.

Superconducting tunnel junctions

The classic Josephson junction is a superconductor-insulator-superconductor (SIS) structure. The insulating layer is extremely thin, typically 1–2 nm, so that the Cooper pair wavefunctions on each side overlap through the barrier. Common barrier materials include aluminum oxide ($\text{Al}_2\text{O}_3$) and niobium oxide.

• The coupling strength between the two superconductors depends on the barrier thickness and transparency.
• Thinner barriers yield larger critical currents because the wavefunction overlap is greater.
• Fabrication requires precise control of deposition processes, since even sub-nanometer variations in barrier thickness change junction parameters significantly.

Cooper pair tunneling

Unlike single-electron tunneling (which is dissipative), Cooper pair tunneling transfers charge without breaking pairs and therefore without dissipation. The tunneling rate is governed by the phase difference $\phi$ between the two superconducting wavefunctions.

Because the pairs remain coherent throughout the process, a supercurrent flows even when no voltage is applied across the junction. This is the defining signature of the Josephson effect and distinguishes it from ordinary quantum tunneling of single particles.

Weak links

A Josephson junction doesn't have to be an SIS sandwich. Any structure that provides a weak coherent coupling between two superconductors qualifies as a "weak link." Common types include:

• Tunnel junctions (SIS): the classic geometry described above
• Point contacts: a sharp superconducting tip touching a superconducting surface
• Constrictions (microbridges): a narrow neck in a superconducting thin film
• SNS junctions: a normal metal layer sandwiched between two superconductors

Each type has a characteristic critical current set by the junction geometry, material properties, and temperature.

Josephson equations

Two equations describe the fundamental behavior of any Josephson junction. Together they connect the supercurrent, the phase difference, and the voltage across the junction.

DC Josephson effect

When no voltage is applied across the junction, a DC supercurrent flows whose magnitude depends on the phase difference:

$I_s = I_c \sin(\phi)$

Here $I_c$ is the critical current, the maximum supercurrent the junction can carry. As long as the bias current stays below $I_c$, the junction remains in the zero-voltage state and carries current without dissipation. The phase difference $\phi$ adjusts itself so that $I_c \sin(\phi)$ matches the imposed current.

AC Josephson effect

When a constant voltage $V$ is maintained across the junction, the phase difference evolves linearly in time:

$\frac{d\phi}{dt} = \frac{2eV}{\hbar}$

Substituting this into the current-phase relation gives an oscillating supercurrent at the Josephson frequency:

$f_J = \frac{2eV}{h}$

The numerical coefficient is 483.6 MHz/µV. This exact relationship between voltage and frequency is what makes the Josephson effect so powerful for metrology: it links voltage directly to fundamental constants and a measurable frequency.

Current-phase relationship

For an ideal SIS tunnel junction, the current-phase relation (CPR) is sinusoidal: $I_s = I_c \sin(\phi)$. More complex junctions can have non-sinusoidal CPRs. For example:

• SNS junctions with high-transparency channels develop higher harmonics ($\sin(2\phi)$, etc.)
• Junctions involving unconventional superconductors (e.g., d-wave high-$T_c$ materials) can have CPRs reflecting the order parameter symmetry

The CPR determines the junction's Josephson energy $E_J$ through the relation $U(\phi) = -E_J \cos(\phi)$ (for the sinusoidal case), which is central to understanding junction dynamics and qubit design.

Macroscopic quantum phenomena

The Josephson effect is remarkable because it makes quantum mechanics visible at macroscopic scales. The phase difference $\phi$ is a collective variable describing billions of Cooper pairs, yet it obeys quantum mechanical rules.

Quantum interference

When a superconducting loop contains two Josephson junctions (forming a DC SQUID), the total critical current depends on the magnetic flux $\Phi$ threading the loop:

$I_c(\Phi) = 2I_{c0} \left| \cos\left(\pi \frac{\Phi}{\Phi_0}\right) \right|$

This is a direct analog of the double-slit interference pattern in optics, but for supercurrents. The periodicity in $\Phi_0$ makes SQUIDs extraordinarily sensitive magnetometers.

Flux quantization

The magnetic flux through any closed superconducting loop is quantized in units of the flux quantum:

$\Phi_0 = \frac{h}{2e} \approx 2.07 \times 10^{-15} \text{ Wb}$

This quantization arises because the superconducting wavefunction must be single-valued around the loop. The factor of $2e$ (rather than $e$) reflects the fact that the charge carriers are Cooper pairs. Flux quantization is fundamental to the operation of SQUIDs, flux qubits, and single flux quantum (SFQ) logic.

Persistent currents

Superconducting loops sustain circulating currents indefinitely to maintain flux quantization. If an external field tries to thread a non-integer number of flux quanta through the loop, a persistent current flows to make up the difference. The magnitude of this current depends on the applied flux and the loop's inductance $L$. Persistent currents are exploited in flux qubits, superconducting memory elements, and parametric amplifiers.

SQUID devices

Superconducting Quantum Interference Devices (SQUIDs) combine Josephson junctions with superconducting loops to achieve the most sensitive magnetic field measurements available. They come in two main configurations.

DC SQUID vs RF SQUID

Flux-to-voltage conversion

A DC SQUID converts magnetic flux changes into voltage changes. The voltage across the SQUID is a periodic function of the applied flux with period $\Phi_0$. In practice, a flux-locked loop feedback circuit linearizes this response by actively canceling flux changes, keeping the SQUID at its most sensitive operating point. This allows measurement of fields as small as a few femtotesla ($\sim 10^{-15}$ T).

Applications in magnetometry

• Biomagnetism: Magnetoencephalography (MEG) maps brain activity by detecting the tiny magnetic fields (~100 fT) produced by neural currents. Magnetocardiography (MCG) does the same for the heart.
• Geophysics: SQUIDs are used in mineral exploration and geological surveys to detect subtle variations in Earth's magnetic field.
• Non-destructive testing: Detecting flaws in metallic structures (aircraft fuselages, pipelines) by sensing distortions in applied magnetic fields.
• Fundamental physics: Gravitational wave antenna readout, searches for magnetic monopoles, and axion dark matter experiments.

Josephson junctions in circuits

Josephson junctions act as nonlinear, non-dissipative inductors in superconducting circuits. Their behavior depends strongly on how they are biased and what external circuit elements are present.

Current-biased configuration

A junction driven by a constant current source sits in the zero-voltage state as long as $I < I_c$. Once the current exceeds $I_c$, the junction switches to a finite-voltage state. In underdamped junctions (those with significant capacitance), this switching is hysteretic: the junction doesn't return to zero voltage until the current is reduced well below $I_c$. This hysteresis is exploited in phase qubits and SFQ logic, where the switching event represents a bit flip.

Voltage-biased configuration

When driven by a constant voltage, the junction produces an oscillating supercurrent at the Josephson frequency. There is no hysteresis in this case. Voltage-biased operation is the basis for Josephson voltage standards and can serve as a source of microwave radiation. Careful impedance matching of the external circuit at the Josephson frequency is important to avoid unwanted reflections and resonances.

Shunted vs unshunted junctions

Connecting an external resistor in parallel with the junction (shunting) damps the junction dynamics and eliminates hysteresis. The resulting behavior is well described by the resistively and capacitively shunted junction (RCSJ) model, which treats the junction as an ideal Josephson element in parallel with a resistor $R$ and capacitor $C$.

• Overdamped (shunted, $\beta_c = \frac{2eI_cR^2C}{\hbar} \ll 1$): non-hysteretic I-V curve, used in SQUIDs and some digital circuits.
• Underdamped (unshunted, $\beta_c \gg 1$): hysteretic I-V curve, used in SFQ logic and certain qubit designs.

The Stewart-McCumber parameter $\beta_c$ quantifies the damping and determines which regime the junction operates in.

Josephson junction dynamics

The time-dependent behavior of a Josephson junction is rich and connects classical nonlinear dynamics with quantum mechanics.

Washboard potential model

The dynamics of the phase $\phi$ can be visualized as a particle of "mass" $C(\hbar/2e)^2$ moving in a tilted periodic potential:

$U(\phi) = -E_J \cos(\phi) - \frac{\hbar I}{2e}\phi$

where $E_J = \frac{\hbar I_c}{2e}$ is the Josephson energy and $I$ is the bias current.

• At zero bias, the potential is a series of identical wells (a cosine).
• Increasing the bias current tilts the washboard, making the wells shallower.
• At $I = I_c$, the wells disappear and the "particle" rolls freely, corresponding to the junction switching to the voltage state.

This picture captures both classical dynamics (thermal activation over barriers) and quantum dynamics (tunneling through barriers).

Phase diffusion

In junctions where the Josephson energy $E_J$ is comparable to the thermal energy $k_BT$ or the charging energy $E_C = (2e)^2/2C$, thermal or quantum fluctuations cause the phase to diffuse between adjacent wells. This produces a small but finite voltage even for bias currents below $I_c$. Phase diffusion is particularly relevant for small-capacitance junctions and junction arrays, where fluctuations are not negligible.

Macroscopic quantum tunneling

At sufficiently low temperatures, the phase can tunnel through the potential barrier from one well to the next, rather than being thermally activated over it. This macroscopic quantum tunneling (MQT) causes the junction to switch to the voltage state at currents below the classical critical current.

MQT was one of the first experimental demonstrations that a macroscopic degree of freedom (the junction phase, involving billions of Cooper pairs) obeys quantum mechanics. It is also the operating principle behind certain superconducting qubits, where controlled tunneling between wells encodes quantum information.

High-frequency properties

The nonlinear inductance of a Josephson junction makes it a natural element for microwave and terahertz applications.

Shapiro steps

When a Josephson junction is irradiated with microwaves at frequency $f$, the I-V characteristic develops flat voltage plateaus (Shapiro steps) at:

$V_n = n\frac{hf}{2e}, \quad n = 0, \pm 1, \pm 2, \ldots$

These steps arise from phase-locking between the junction's internal Josephson oscillation and the external microwave drive. Because the step voltages depend only on $f$, $h$, and $e$, Shapiro steps provide the basis for the modern definition of the volt. Programmable Josephson voltage standards use arrays of thousands of junctions to generate precise voltages at the 1 V level and beyond.

Josephson radiation

A voltage-biased junction emits electromagnetic radiation at the Josephson frequency $f_J = 2eV/h$. For typical junction voltages (µV to mV), this falls in the microwave to terahertz range. The power output from a single junction is small (picowatts), but coherent arrays of phase-locked junctions can produce usable power levels. This is an active area of research for compact, tunable terahertz sources.

Inverse AC Josephson effect

The inverse effect occurs when external radiation at frequency $f$ is absorbed by the junction, producing DC current steps (rather than voltage steps). This happens when the photon energy $hf$ matches the energy $2eV$ associated with a Cooper pair crossing the junction. The inverse AC effect enables Josephson junctions to function as sensitive detectors and mixers at microwave and millimeter-wave frequencies.

Applications of Josephson effect

Voltage standards

Josephson voltage standards are the most accurate realization of the volt. They exploit the exact relation $V = nhf/2e$ to generate voltages traceable to the SI second (via frequency) and fundamental constants. Modern programmable Josephson voltage standard (PJVS) systems use arrays of over 250,000 junctions to produce voltages from $-10$ V to $+10$ V with relative uncertainties below $1 \times 10^{-9}$. Since the 2019 SI redefinition, the Josephson constant $K_J = 2e/h$ is fixed by definition.

Superconducting qubits

Josephson junctions provide the essential nonlinear, non-dissipative circuit element needed to build quantum bits from superconducting circuits. The three main qubit types each exploit different aspects of junction physics:

• Transmon: Operates in the regime $E_J \gg E_C$, reducing sensitivity to charge noise. Currently the most widely used design (e.g., IBM, Google processors).
• Flux qubit: Encodes information in clockwise/counterclockwise persistent current states of a superconducting loop interrupted by junctions.
• Phase qubit: Uses the lowest energy levels within a single well of the washboard potential.

Coherence times have improved from nanoseconds to hundreds of microseconds over the past two decades, driven by better materials and circuit designs.

Rapid single flux quantum logic

RSFQ is a digital logic family where information is encoded as the presence or absence of a single flux quantum $\Phi_0$ in a superconducting loop. Josephson junctions act as ultrafast switches that transfer flux quanta between loops.

• Clock speeds exceed 100 GHz, far beyond semiconductor logic.
• Power dissipation per switching event is extremely low ($\sim 10^{-19}$ J).
• Challenges include the need for cryogenic operation and the difficulty of interfacing with room-temperature CMOS electronics.

RSFQ and its energy-efficient variants (ERSFQ, eSFQ) are being developed for applications in high-performance computing, digital signal processing, and control electronics for quantum computers.

Exotic Josephson junctions

Research continues to push beyond conventional SIS junctions, exploring new materials and geometries that produce qualitatively different physics.

$\pi$-junctions

In a standard junction, the ground state has $\phi = 0$. A $\pi$-junction has a ground state at $\phi = \pi$, meaning the current-phase relation is effectively $I_s = -I_c \sin(\phi)$ (the critical current appears negative). This can be achieved by:

• Using a ferromagnetic barrier (SFS junction), where the exchange field induces a phase shift in the Cooper pair wavefunction
• Exploiting the d-wave symmetry of high-$T_c$ superconductors with appropriately oriented grain boundaries

$\pi$-junctions enable "quiet" qubits with built-in degeneracy and are used to study frustration effects in superconducting networks.

Long Josephson junctions

When one or more junction dimensions exceed the Josephson penetration depth $\lambda_J$, the phase difference varies spatially along the junction. The dynamics are described by the sine-Gordon equation:

$\lambda_J^2 \frac{\partial^2 \phi}{\partial x^2} - \frac{1}{\omega_p^2}\frac{\partial^2 \phi}{\partial t^2} = \sin(\phi)$

These junctions support topological excitations called fluxons (Josephson vortices), each carrying one flux quantum. Fluxons can propagate along the junction and interact with each other, making long junctions a rich system for studying soliton physics. Applications include flux-flow oscillators for terahertz generation and superconducting transmission lines.

Fractional vortices

In certain exotic junctions and arrays, magnetic vortices can carry a fraction of $\Phi_0$. These fractional vortices arise from:

• Frustrated geometries (e.g., loops containing both 0- and $\pi$-junctions)
• Unconventional order parameter symmetry at grain boundaries

Fractional vortices are of interest for studying topological phases of matter and have been proposed as building blocks for topologically protected qubits.

Challenges and limitations

Thermal noise effects

At finite temperature, thermal fluctuations smear out the sharp features of junction behavior. In SQUIDs, thermal noise sets a fundamental sensitivity floor. In qubits, thermal excitation of higher energy levels degrades performance. Most high-performance Josephson devices operate in dilution refrigerators at 10–20 mK to suppress thermal noise, but this adds significant system complexity and cost.

Decoherence in quantum circuits

Decoherence remains the central obstacle to large-scale superconducting quantum computing. The main noise sources are:

• Charge noise: fluctuating charges on nearby surfaces and interfaces
• Flux noise: believed to originate from unpaired surface spins
• Dielectric loss: two-level systems (TLS) in amorphous oxide layers (including the junction barrier itself) absorb and re-emit microwave photons

Mitigation strategies include using transmon designs to reduce charge sensitivity, improving surface treatments, and developing new barrier materials with fewer TLS defects. Quantum error correction provides a path forward, but requires further improvements in gate fidelity and qubit count.

Fabrication issues

Scaling up to circuits with thousands or millions of junctions demands tight control over junction parameters. The critical current $I_c$ must be uniform across an entire chip, which requires:

• Precise control of barrier thickness (sub-angstrom uniformity)
• Clean deposition environments to avoid contamination
• Reproducible lithographic patterning at the nanometer scale

Current fabrication yields are adequate for SQUID sensors and small qubit processors, but achieving the uniformity needed for large-scale quantum computers or complex RSFQ circuits remains an active engineering challenge.